ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2017

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1517

Synthesis, Characterization, andEvaluation of Ag-based ElectricalContact Materials

FANG MAO

ISSN 1651-6214ISBN 978-91-554-9915-0urn:nbn:se:uu:diva-320235

Dissertation presented at Uppsala University to be publicly examined in Room 2001,Ångströmlaboratoriet, Lägerhyddsvägen 1, Uppsala, Thursday, 8 June 2017 at 09:15 forthe degree of Doctor of Philosophy. The examination will be conducted in English. Facultyexaminer: Professor Mikael Olsson (Högskolan Dalarna).

AbstractMao, F. 2017. Synthesis, Characterization, and Evaluation of Ag-based ElectricalContact Materials. Digital Comprehensive Summaries of Uppsala Dissertations from theFaculty of Science and Technology 1517. 98 pp. Uppsala: Acta Universitatis Upsaliensis.ISBN 978-91-554-9915-0.

Ag is a widely used electrical contact material due to its excellent electrical properties. Theproblems with Ag are that it is soft and has poor tribological properties (high friction and wearin Ag/Ag sliding contacts). For smart grid applications, friction and wear became increasinglyimportant issues to be improved, due to much higher sliding frequency in the harsh operationenvironment. The aim of this thesis is to explore several different concepts to improve theproperties of Ag electrical contacts for smart grid applications.

Bulk Ag-X (X=Al, Sn In) alloys were synthesized by melting of metals. An important resultwas that the presence of a hcp phase in the alloys significantly reduced friction coefficientsand wear rates compared to Ag. This was explained by a sliding-induced reorientation of easy-shearing planes in the hexagonal structure. The Ag-In system showed the best combination ofproperties for potential use in future contact applications.

This thesis has also demonstrated the strength of a combinatorial approach as a high-throughput method to rapidly screen Ag-based alloy coatings. It was also used for a rapididentification of optimal deposition parameters for reactive sputtering of a complex AgFeO2

oxide with narrow synthesis window. A new and rapid process was developed to grow lowfrictional AgI coatings and a novel designed microstructure of nanoporous Ag filled with AgI(n-porous Ag/AgI) using a solution chemical method was also explored. The AgI coatingsexhibited low friction coefficient and acceptable contact resistance. However, under very harshconditions, their lifetime is too short. The initial tribotests showed high friction coefficient ofthe n-porous Ag/AgI coating, indicating an issue regarding its mechanical integrity.

The use of graphene as a solid lubricant in sliding electrical contacts was investigated as well.The results show that graphene is an excellent solid lubricant in Ag-based contacts. Furthermore,the lubricating effect was found to be dependent on chemical composition of the counter surface.As an alternative lubricant, graphene oxide is cheaper and easier to produce. Preliminary testswith graphene oxide showed a similar frictional behavior as graphene suggesting a potential useof this material as lubricant in Ag contacts.

Keywords: electrical contact, bulk, coating, Ag-based alloys, Ag-based delafossite, AgI,graphene, graphene oxide, combinatorial material science, dc magnetron sputtering, friction,wear, hardness, contact resistance

Fang Mao, Department of Chemistry - Ångström, Inorganic Chemistry, Box 538, UppsalaUniversity, SE-751 21 Uppsala, Sweden.

© Fang Mao 2017

ISSN 1651-6214ISBN 978-91-554-9915-0urn:nbn:se:uu:diva-320235 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-320235)

Three passions, simple but overwhelmingly strong, have governed my life: the longing for love, the search for knowledge, and unbearable pity for the suffering of mankind.

Bertrand Russell

To Angelica

Preface

This work has been part of an academic and industrial collaboration, and was performed mainly at the Department of Chemistry-Ångström Laboratory, Uppsala University, Uppsala, and partly at the ABB AB, Corporate Re-search, Västerås. My project was funded by the the KIC InnoEnergy and the Swedish Centre for Smart Grids and Energy Storage (SweGRIDS). All the financial supports are highly acknowledged, including the ÅForsk and Liljewalch foundations for the conference grants. .

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Mao, F., Taher, M., Berastegui P., Andersson, A. M., Jansson

U. Tuning the tribological, mechanical and electrical properties in Ag-X (X=Al, In, or Sn) alloys. In manuscript

II Taher, M., Mao, F., Berastegui P., Andersson A. M., Jansson U. The influence of chemical and phase composition on mechani-cal, tribological and electrical properties of Ag-Al alloys. Sub-mitted to Wear

III Mao, F., Taher, M., Kryshtal, O., Kruk, A., Czyrska-Filemonowicz, A., Ottosson, M., Andersson, M. A., Wiklund, U., Jansson, U. (2016) A combinatorial study of gradient Ag-Al thin films: microstructure, phase formation, mechanical and electrical properties. ACS Applied Materials & Interface, 8 (44), 30635–30643.

IV Mao, F., Nyberg, T., Thersleff, T., Andersson, M. A., Jansson, U. (2016) Combinatorial magnetron sputtering of AgFeO2 thin films with the delafossite structure Materials & Design 91:132-142

V Sundeberg, J., Mao, F., Wiklund, U., Andersson, M. A., Jans-son, U. (2013) Solution-based synthesis of AgI coatings for low-friction applications. Journal of Materials Science 48 (5): 2236-2244.

VI Mao, F., Wiklund, U., Andersson, M. A., Jansson, U. (2015) Graphene as a lubricant on Ag for electrical contact applications Journal of Materials Science 50 (19): 6518-6525.

Reprints were made with permission from the respective publishers.

Contributions to the papers

I. Major part of the planning and synthesis, part of the characteri-

zation, discussion and writing. II. Part of the planning, experiment, discussion and writing III. Major part of the planning, experiment, discussion and writing IV. Major the part of planning, experiment, discussion and writing V. Part of the planning, experiment, discussion and writing VI. Major part of the planning, experiment, discussion and writing

Other publications

I. Zhu, D., Mao, F., Zhao, S., (2012) The influence of oxygen in

TiAlOxNy on the optical properties of colored solar absorbing coat-ings, Solar Energy Materials and Solar Cells, 98: 179-184.

II. Meng, Q.N., Wen, M., Mao, F., Nedfors, N., Jansson, U., Zheng W. T. (2013) Deposition and characterization of reactive magnetron sputtered zirconium carbide films. Surface & Coatings Technology, 232: 876-883.

III. Meng, Q.N., Malinovskis, P., Nedfors, N., Mao, F., Andersson M., Sun Y., Jansson U. (2015).Characterization of amorphous Zr-Si-C thin films deposited by DC magnetron sputtering. Surface & Coat-ings Technology, 261: 227-234.

IV. Mao, F., Lindeberg, M., Hjort, K., Klintberg, L. (2012) A polymer foil non-contact IR temperature sensor with a thermoresistor inte-grated on the back of a vertically configured thermopile. Sensors and Actuators A-Physical, 179: 56-61.

Contents

1 Introduction ......................................................................................... 15

2 Sliding electrical contacts .................................................................... 17 2.1 Definition of sliding electrical contacts .......................................... 17 2.2 Materials requirements .................................................................... 17 2.3 Electrical contact resistance ............................................................ 18 2.4 Friction and wear ............................................................................ 19

3 Material design concepts ..................................................................... 22 3.1 Ag-based bulk materials .................................................................. 22 3.2 Surface Engineering ........................................................................ 23

3.2.1 Ag-based alloy coatings ......................................................... 23 3.2.2 Ag-based compound coatings ................................................ 24 3.2.3 Designed microstructures ...................................................... 25 3.2.4 Solid lubricants ...................................................................... 25

4 Material systems .................................................................................. 28 4.1 Pure silver (Ag) ............................................................................... 28 4.2 Ag-based alloys ............................................................................... 28 4.3 AgFeO2 ............................................................................................ 31 4.4 AgI .................................................................................................. 32 4.5 Graphene and graphene oxide ......................................................... 33

5 Methods ............................................................................................... 36 5.1 Material synthesis ........................................................................... 36

5.1.1 Bulk synthesis ........................................................................ 36 5.1.2 Thin film deposition ............................................................... 36 5.1.3 Chemical synthesis ................................................................ 40

5.2 Characterization of materials .......................................................... 41 5.2.1 X-ray diffraction .................................................................... 41 5.2.2 X-ray photoelectron spectroscopy ......................................... 45 5.2.3 Raman spectroscopy .............................................................. 46 5.2.4 Additional characterization methods ..................................... 47

5.3 Evaluation of properties .................................................................. 49 5.3.1 Hardness measurements ......................................................... 49 5.3.2 Tribological testing ................................................................ 49 5.3.3 Contact resistance measurements .......................................... 50

5.3.4 Other properties measurements ............................................. 51

6 Results and discussion ......................................................................... 53 6.1 Ag-X alloys ..................................................................................... 53

6.1.1 Phase composition ................................................................. 53 6.1.2 Evaluation of properties ......................................................... 55 6.1.3 Mechanism of low-frictional behavior................................... 58

6.2 Alloy coatings ................................................................................. 60 6.2.1 Rapid material screening ....................................................... 60 6.2.2 Detailed study on Ag-Al coatings .......................................... 61

6.3 AgFeO2 compound coatings ........................................................... 67 6.3.1 Optimization of the deposition parameters ............................ 67 6.3.2 Microstructure of the AgFeO2 coatings ................................. 70 6.3.3 Properties of the AgFeO2 coatings ......................................... 72

6.4 AgI coatings .................................................................................... 73 6.5 Designed microstructures ................................................................ 77

6.5.1 Synthesis of nanoporous Ag .................................................. 78 6.5.2 Filling of AgI ......................................................................... 78

6.6 Solid lubricants ............................................................................... 80

7 Conclusions and outlook ...................................................................... 85

8 Populärvetenskaplig sammanfattning .................................................. 87

9 Acknowledgments ............................................................................... 89

10 References ........................................................................................... 92

Abbreviations

μ or CoF Friction coefficient Ff Friction force FN Normal load force A Cross sectional area of the groove V Worn volume S Sliding distance GL Few-layer of graphene flakes GO Graphene oxide fcc Face centered cubic hcp Hexagonal closed packed a.u. Arbitrary unit TCO Transparent conductive oxide DC Direct current XRD X-ray diffraction XPS X-ray photoelectron spectroscopy SEM Scanning electron microscopy EDX or EDS Energy dispersive X-ray spectroscopy TEM Transmission electron microscopy HRTEM High-resolution TEM FIB Focused ion beam SAED Selected-area electron diffraction EFTEM Energy Filtered TEM HAADF High-angle annular dark-field imaging XRF X-ray fluorescence XRR X-ray reflectivity n-porous Ag/AgI

Nanoporous Ag filled with AgI

15

1 Introduction

Electrical contacts are essential components in modern electronic devices. Common examples of electrical contacts in our daily lives are, e.g. connect-ors, chargers, switches, and so on. In general, for a contact to have a good electrical conductance, the contact materials should have low electrical re-sistance and also low contact resistance over the contact interface. There are also plenty of applications for voltage regulation in electricity transmission and distribution system, so called electrical grids, where the contact materi-als suffer from friction and wear due to sliding movements.

Fig.1.1. A schematic image of a smart grid [1].

The conventional electrical grid, which is designed based on the idea that supply follows demand, usually consist of one (large) generation source and a unidirectional distribution of the power throughout the grid in a waterfall structure. A so-called smart grid has a well-organized automation system to deliver electricity from plants (from large-scale power plants to small-scale production from e.g. solar panels in private homes) to customers in a sus-tainable, smart and energy-saving way. A smart-grid system has the ability to handle the variable power from the emerging renewable energy sources (wind, solar and wave power etc.), where production is fluctuating with time. One of its advantages is that the automation equipment and control system allow it to response accurately and deliver electricity in time to meet the

16

various demands of customers from times and zones, which also means it needs to handle multi-directional distribution of the power. A consequence of this “smartness” is that the number and frequency of the voltage regulat-ing operations increase significantly, compared to the conventional electrical grid, leading to an increased demand on switching and regulating electric contacts. A schematic image of a smart grid is shown in Fig. 1.1.

Silver (Ag) is a widely used contact material in the voltage regulation de-vices in the electrical grid, owing to its excellent electrical properties. It is the best electrical conductor among all metals. However, one of its draw-backs is poor tribological properties, leading to high friction and wear in Ag-Ag contacts. After a large number of sliding movements, high friction and wear can result in reduced lifetime, unstable performance, high risk for mal-function and/or unexpected electricity break-down. Therefore, it is important to develop new contact materials, which combine good electrical properties with good tribological properties.

One of the common methods to reduce friction and wear is to apply an oil or grease as lubricant. However, mainly due to thermal instability of these liquid lubricants there is a demand to use dry, self-lubricating and mainte-nance-free electrical contact materials instead. The benefits would be long lifetime, low friction and wear as well as low and stable contact resistance throughout the lifespan of the product. From a material science point of view, the design of such materials is complex and challenging because all these properties are difficult to obtain simultaneously.

The main objective of this thesis is to identify suitable materials to re-place pure Ag for dry sliding electrical contact applications. The research leading to this thesis has been conducted in an environment involving col-laboration between academia and industry. Therefore, the aim of this project has two parts: (i) fundamental scientific research and (ii) characterization of useful materials for industrial applications. The aim of the fundamental re-search part is to understand the composition-structure-property relationship for the investigated materials. Based on the intended application (sliding electrical contact), the thesis also focuses on the evaluation of electrical and tribological properties. Different material systems have been synthesized in the forms of coatings and/or bulk materials using different techniques. Char-acterizations of microstructure and phase composition together with the evaluation of electrical, tribological, and mechanical performance have been carried out to identify potential candidates and to study how their properties can be controlled.

17

2 Sliding electrical contacts

2.1 Definition of sliding electrical contacts An electrical contact is a junction consisting of at least two counter parts, which are intended to enable the conduction of electrical current when in contact. When a pair of contacts touches each other, they allow transfer of electrons from one contact to the other. When they are separated, the current circuit is open. Electrical contacts can be classified into two types: stationary and dynamic contacts (e.g. sliding contacts and breaking contacts). Sliding electrical contacts are moving during operation and normally suffer heavy wear and oxidation due to harsh conditions (e.g. high loads, high tempera-tures, long sliding distances, etc). Some sliding contacts conduct current during the movement, while the other types are only subjected to pure me-chanical action and conduct current in a stationary state. The sliding electri-cal contacts are important components in voltage regulating and switching devices, e.g. tap-changers, high/medium voltage breakers, disconnectors, or electrical brushes for motors, robots, and sensors. In this work, materials suited for sliding contact applications designed for conduction of electrical current (and off-load breaking) is under scrutiny. Breaking (arcing) contact applications are not considered.

2.2 Materials requirements The properties of the materials in the contact are critically important for the performance of the sliding contacts. The intrinsic electrical resistance, me-chanical properties (hardness, ductility), and chemical properties (corro-sion/oxidation resistance) of the materials in contact affect the contact re-sistance, shear strength and thus the electrical and tribological behavior of the contact system. A material that potentially could replace pure Ag as slid-ing contact material would have to satisfy, at least, the requirements listed below:

• Tribological properties: the material should have low friction

and wear rate compared to the pure Ag reference. • Electrical properties: the material has to be reasonably conduc-

tive and has a low contact resistance, similar to pure Ag.

18

• Mechanical properties: i) the hardness of the material should be enhanced since softer materials usually have high wear rates. However, soft materials will deform and increase the contact ar-ea. Therefore, the hardness should be balanced; ii) a high ductility will promote the penetration of the surface oxide layer and hence reduce the contact resistance.

• Chemical properties: a high oxidation resistance is required to minimize the influence of the poorly conducting surface oxide layer on the contact resistance. The material should also have high corrosion resistance, to ensure stable function throughout the lifetime of the contact

• Cost and environmental issues: the material should have lower price compared to pure Ag and be non-toxic.

From a material science point of view, it is a major challenge to improve the mechanical and tribological properties, while at the same time, maintaining the beneficial electrical and chemical properties of Ag. Finally, it should be noted that the contact resistance, friction, and wear are system properties and hence cannot be only attributed to a single material. They are affected also by the geometries of the two counter surfaces, applied load, testing parame-ters, the environment, etc.

2.3 Electrical contact resistance The efficiency of the contact as a current conductor is determined by the electrical contact resistance, which is the resistance of the electrical current flux over the contact junction. A brief overview of a contact interface is il-lustrated in Fig. 2.1.

It is important to note that it is not the total apparent contact area (Aapparent) that transfers the electrical current but only a small part or spots that form a the physical contact [2]. The number, size and distribution of the mechanical contact spots are determined by the contact load and the mechanical proper-ties of the contact materials. In the simplest approximation, the mechanical contact area (Amechanical) can be expressed using the contact force (Fcontact) and the hardness of the softer material (Hsoftest):

= /

However, the situation is usually more complex, since the electrically con-ductive area (Aconductive) differs from the Amechanical because of an insulating oxide and contaminations. Aconductive depends on the thickness of the oxide layer and the degree of the break-through of the surface oxide, which is de-termined by chemical and mechanical properties of the contact materials and

19

also of the load. A rough surface morphology usually decreases the contact resistance because the oxide layer is more easily broken-through, compared to a flat surface.

In summary, the contact resistance depends on many factors, such as the applied contact force, the hardness of the contact materials, the thickness of the insulating layer, and surface roughness.

Fig.2.1. Schematic image of an electrical contact (left). Insets of the contact inter-face shows the apparent area, the mechanical contact area, and the conductive area (right top); electrical current flux passes through the conductive area as seen in the cross-section view (right bottom)

2.4 Friction and wear In sliding contacts, friction and wear processes play an important role in the design of contact devices. For example, the friction between the contacts determines the force needed to move one contact at a certain contact load, and the service lifetime will depend strongly on the wear resistance.

Friction is the resistance against relative movement of two surfaces and quantified in terms of friction coefficient (μ). It is defined as: = /

where Ff is the tangential friction force required to initiate and maintain the sliding, while FN is the normal load force. There can be kinetic (dynamic) friction referring to the friction between surfaces in relative motion, or static

20

friction, which refers to the resistance to move. FN is the normal load force. In this thesis, the term friction refers to the kinetic friction.

In the simplest mode of sliding friction, the friction force arises from two sources: an adhesion force (Fadh) required to shear the areas of real contact between the surfaces, and a deformation force (Fdef) for ploughing the asperi-ties of the harder surface through the softer.

= +

Fadh is positively proportional to the shear strength (s), the sum of the real contact areas of all the asperity junctions (A’), but reversely proportional to the hardness of the softer material (H). is directly proportional to the hardness of the softer material and the cross sectional area of the groove (A) [3]. The relationships can be simplified as: ∝ /

∝

Wear is a progressive loss of material involving removal or displacement of material from its original position. In this thesis, the term wear refers to dry sliding wear since all the experiments were carried out in ambient air with appreciable humidity (30-50%) without addition of lubricating oils or greas-es. In sliding motion, wear, in mechanical terms is usually classified into two types; adhesive wear and abrasive wear. The adhesive wear refers to material transfer during the sliding process, characterized by a severely roughened surface and displacement of material. Mechanisms of abrasive wear can involve both plastic deformation and fracture. The plastic deformation is closely related to the model how ploughing relates to friction. The fracture refers to scratching by the asperities of the harder surface through the softer. It also can be caused by hard abrasive particles or the generated debris dur-ing sliding movement. Oxidation wear is a mild form of wear, which in-volves tribo-chemical actions. It occurs when the metal surface is covered with a layer of oxide, which reduces the tendency of adhesive wear. The mechanism of oxidation wear can be simply described as repeated “chemi-cal-mechanical” cycles like where parts of the original oxide layer can be removed by the asperities, leaving the underlying metal uncovered. The fresh metal will quickly react with oxygen in air to form a new oxide layer, which will eventually be scrapped off again in the following motion. The cycle repeats during the oxidation wear process.

The wear can be quantified using different parameters such as mass loss, volume loss, wear resistance, and wear rate. The most commonly used pa-rameter is the wear rate, which is the worn volume (V) divided by the sliding

21

distance (S) and the applied load ( ). The wear rate has the unit of m2N-1, but is also often given as mm3N-1m-1 or μm3N-1m-1. In this thesis the wear depth (h) and the cross-sectional area (A) of the groove, have sometimes been used for comparison of wear (Fig. 2.2).

Fig.2.2. Schematic image showing the concepts of friction (right) and the cross sec-tion view of the wear in a reciprocating contact (right).

22

3 Material design concepts

In this thesis, several different concepts have been used to improve the prop-erties of Ag as sliding electrical contacts. The concepts are illustrated in Fig. 3.1 and the selected materials combinations are listed in each concept block.

Fig. 3.1. Overview of the design concepts explored in this thesis to improve Ag in terms of better tribological properties and maintained good electrical properties. The material systems selected for each concept are listed in each concept block.

3.1 Ag-based bulk materials The poor tribological properties of Ag-Ag contacts, as represented by the high friction and severe adhesive wear during sliding, can be attributed to the

Sometimes the word thin film is used instead of coating. In some contexts thin film is used for coatings thinner than 1 μm. In the following text the word coating will be used to denote both coating and thin film.

23

low hardness of Ag. Increasing the hardness of the contacting material is one of the methods used to reduce adhesive friction and wear in metal-metal contacts. It is well known that the hardness of soft metal such as Ag can be enhanced by solid solution strengthening and precipitation hardening through alloying with other elements [4, 5].

The friction and wear can also be controlled by forming an alloy with a completely different crystal structure than face centered cubic structure fcc -Ag. It has been previously reported that metals with a hexagonal close packed (hcp) structure such as lanthanum, rhenium and cobalt [6, 7] can exhibit lower friction and wear coefficients than bcc and fcc metals. The outstanding tribological properties of hcp metals were attributed to the pres-ence of easy-shearing basal slip planes. A hexagonal Ag-In alloy also has been patented as a good solid lubricant coating [8].

In this thesis, three different binary alloy systems Ag-Al, Ag-Sn and Ag-In, which all forms hexagonal intermediate phases, are studied as potential candidates for sliding electrical contact materials. Electrical, tribological and mechanical properties of these Ag alloys have been systematically investi-gated. The influence of the crystal structure on the properties for each mate-rial system was investigated in paper I, while a more detailed study focusing on the Ag-Al system is described in paper II.

3.2 Surface Engineering Surface engineering is another common method to modify the contact sur-face either by deposition of a thin coating or adding a solid lubricant. Sys-tems lubricated by oils and greases are not studied in this thesis. The purpose of the modified surface is to reduce friction and wear, without influencing the passage of current through the contact.

3.2.1 Ag-based alloy coatings Ag alloys can also be deposited as a coating on a substrate to improve the tribological properties. Compared to the bulk Ag alloys (at least on a mm scale), one of the benefits with the coatings is that less material is needed (μm scale). This, in turn, saves resources and decreases the costs. Another advantage is that the alloy coating can be deposited on different types of substrates, which provides the possibilities for other electrical contact appli-cations.

fcc refers to face centered cell lattice, however, it is commonly used as an alternate name for cubic closed packed (ccp) structure. In this thesis, we will use fcc to describe the structure.

24

Aluminum (Al), as a non-expensive and a more easily-sputtered element, was therefore selected to produce a Ag-Al alloy coating by magnetron sput-tering in paper III. There are many other elements, e.g. Nb, Cr, Zr, Ti, W, and etc, which can be sputtered to form Ag alloy coatings. However, they do not form hexagonal alloy phases and were therefore not studied in detail. A combinatorial material science approach is a high-throughput method, which can be used to rapidly discover and/or optimize new or known materials via systematic studies of the composition-structure-property relationships [9-12].This approach was therefore applied in this thesis to rapidly screen po-tential material candidates for contact applications. In paper III, the concept of Ag-Al thin films to reduce the friction and wear was investigated. In addi-tion, the strength of the combinatorial approach for rapid screening of alloy materials was also demonstrated.

In real applications, the product will be exposed to elevated temperatures and oxidation. Since Al is an easily oxidized element, the oxidation of Al could be an issue with respect to the degeneration of electrical properties. Doping with other element, e.g. Cr, may improve the properties regarding the electrical resistance as well as corrosion resistance. Therefore an anneal-ing study and also a Cr doping effect study of Ag-Al thin films was carried out as is summarized on page 65-66.

3.2.2 Ag-based compound coatings A low friction can also be obtained by depositing a non-metallic Ag-based coating on an Ag contact. It has been proposed that Ag in a tribological con-tact can react with the counter surface and produce lubricating surface ox-ides. For example, Mulligan et al. observed a low friction of a CrN/Ag con-tact which was explained by the formation of oxides including the layered delafossite phase AgCrO2 [13]. Layered compounds such as graphite and MoS2 are well-known low friction materials and the existence of layered Ag compounds suggests a potential application of delafossite coatings in sliding contacts. In paper IV, we explored the possibilities to deposit a typical delafossite phase, AgFeO2, which can be described as a natural nanolaminate of FeO2 layers separated by planar Ag layers [14, 15]. The delafossites are easy to synthesize from solution but often require post-annealing at high temperatures ≥ 700oC or high pressure to obtain the desired morphology and crystal structure for practical applications [16-20]. To our knowledge, no studies have previously been published on Ag-based delafossite thin films deposited directly by magnetron sputtering. Initial experiments showed that it was very difficult to identify the experimental parameters required to de-posit single-phase AgFeO2 coatings. In paper IV, a detailed study on the deposition of AgFeO2 thin film with delafossite structure demonstrated the strength of the combinatorial approach to rapidly and effectively identify

25

suitable process parameters for the synthesis of novel materials with narrow homogeneity ranges.

Ag can also form lubricating compounds with other non-metallic ele-ments. One example is AgI. Electrochemically deposited AgI coatings have been demonstrated and patented as a potential alternative to pure Ag con-tacts by ABB in 2003 [21]. The AgI was then formed by an electrochemical process, where a Ag anode was exposed to an iodine solution [22, 23]. In paper V, AgI coatings were synthesized using a new and rapid method.

3.2.3 Designed microstructures Another concept is to synthesize a nanocomposite with a lubricating phase inside a porous Ag phase. The lubricating component should be distributed throughout the entire structure of the materials, allowing it to continuously regenerate its low-frictional properties as the surface lubricant is worn down and give a stable low friction throughout the life-span of the component. One method to produce such a designed microstructure is to fill AgI into nanopo-rous Ag (see concept IV in Fig. 3.1). In theory, the nanoporous Ag/AgI composite should combine the excellent electrical properties of Ag with low friction behavior of AgI. The amount of AgI, initially determined by the porosity, is therefore important for the achievement of optimal properties. First of all, the pore size should not be too large since this may result in a decrease in the electrical conductivity and the mechanical stability, while a too small pore size might be challenging for filling the AgI into the Ag ma-trix. The nanoporous Ag can be obtained by de-alloying Al from an Ag-Al alloys using NaOH as a selective Al etchant. The Al concentration in the precursor alloy is thus a crucial parameter to control the pore size. Ag-Al coatings with continuous compositional gradient have been used as proof of concept as is summarized on page 78-80.

3.2.4 Solid lubricants Conventional solid lubricating materials include layered materials such as graphite, hexagonal boron nitride (h-BN), and transition metal dichalco-genide (e.g. WS2, MoS2) [24]. With the development of nanotechnology, a variety of novel low-dimensional nanomaterials have been found to have exceptional lubricating properties, including, for example, the rolling and sliding movement of zero-dimensional (0D) nanoparticles, e.g. TiO2 nano-particles [25], hollow inorganic fullerene-like MoS2 nanoparticles [26], zinc oxide nanoparticles [27], as well as one-dimensional (1D) nanotubes and nanowires [28, 29]. Two-dimensional (2D) layered materials, e.g. atomically thin sheets of graphene, h-BN, MoS2, NbSe2 have recently attracted inten-sive research interest based on their ultralow friction. This is attributed to the planar structure and low interfacial shear strengths facilitated by weak inter-

26

layer interactions, comparing to the strong in-layer chemical bonding within the layers [30, 31]. The ultralow frictional 2D layer materials have a broad range of applications from gears and engine components, to joint replace-ment in medical implants, as well as micro- and nano-electromechanical devices [32].

Fig. 3.2. Examples of 2D layer materials which can be used as lubricants in a broad range of applications. Reprinted with permission of Elsevier from Spear [32].

Graphene in particular has attracted considerable attention as one of the emerging lubricating materials [33-35]. Berman et al. demonstrated that the addition of few-layer graphene flakes to a steel surface can dramatically reduce the friction coefficient from ~ 1 to approximately 0.15 against steel in a pin-on-disk test [36, 37]. It is reasonable to assume that the tribological behavior for a graphene-coated Ag surface likewise could be significantly improved because Ag interacts weakly with graphene and forms weak Ag-C bonds, compared to a graphene-coated steel surface where stronger interac-tions are expected between Fe-C. In this thesis, the potential of graphene as a solid lubricant in sliding Ag-based electrical contacts has been investigated through a simple and quick method based on the evaporation of a few drop-lets of a commercial graphene solution (paper VI).

An alternative lubricant to graphene is graphene oxide (GO). The tribo-logical behavior of graphene oxide layers on the macro scale has been found to be not much different from that of graphene layers [38]. Compared to the high cost of graphene, GO is cheaper and easier to produce by oxidation of graphite. Therefore, GO layers have also been investigated as a lubricant on Ag surface for reducing the friction and wear. The results are unpublished but summarized on page 83-84.

The dispersion of graphene or graphene oxide on a Ag surface may also provide a good electrical contact, since they are atomically thin and only partially cover the Ag surface, making it possible for the current to easily

27

pass through the interfaces. Furthermore, graphene is a good electrical con-ductor. However, from a production point of view, cost issues are also ex-tremely important. Low-dimensional graphene and even graphene oxide are expensive and up-scaling methods for industrial production must be devel-oped.

28

4 Material systems

4.1 Pure silver (Ag) Ag is a soft, ductile, white, and shiny metal with a fcc structure. Ag has a long history as a precious metal and was used in many old monetary systems as coins. Jewelry and silverware are other traditional application areas for Ag but then often as sterling Ag (an alloy of 92.5% Ag with 7.5% Cu). Ag is also used to make symbolic objects of high social or political rank, for ex-ample the Olympic silver medal for the 1st runner-up.

From a material science and engineering point of view, Ag exhibits the highest electrical conductivity, thermal conductivity and reflectivity of any metal. It is also resistant against oxidation. All these advantages make Ag a useful metal in a range of industrial applications. For example, Ag is widely used in electrical contacts and conductors, reflective coatings in mirrors and windows. It is also used as a catalyst of chemical reactions in medicine and biology applications.

Ag is widely used in electric contacts. However, Ag in pure form has drawbacks since it is too soft (< 200 MPa) and exhibits a low mechanical wear resistance, leading to a limited lifetime. This also adds cost to the prod-uct since the price of Ag is high (~ $560/kg). Another disadvantage is that Ag is prone to materials transfer and cladding to a counter surface during sliding, which then creates an Ag-Ag contact. The friction coefficient of such a contact in a dry sliding system is high (>1). The Ag cladding and the resulting high friction can cause unstable performance of a device, or even failure due to an electrical shorting between adjacent current paths. Conse-quently, there is a need for new sliding contact materials that combine good electrical properties with good tribological properties.

4.2 Ag-based alloys Alloying with other elements is a well-known method to enhance the hard-ness of pure soft metals. However, a detailed study of the tribological prop-erties of Ag-based alloys is lacking. It is thus interesting and also meaningful to investigate the tribological properties of such Ag-based alloys both from engineering as well as from fundamental scientific point of view.

29

Fig. 4.1. The periodic table with different phases for Ag-metal alloys (group 1 and 2 metals not included). Yellow indicates that the elements are immiscible with Ag; Blue indicates the elements can form a hexagonal phase with Ag; Red shows ele-ments that can form other phases with Ag but no hexagonal phase.

Transition metals and metals in group 13-15 are typical elements which can form alloys with Ag. According to the ASM Alloy Phase Diagram Database for binary systems, only a few elements can form new phases when alloyed with Ag. As can be seen in Fig. 4.1 most of the transition metals, are gener-ally immiscible with Ag. The elements marked in yellow color are immisci-ble with Ag, meaning that, at thermodynamic equilibrium, their solubility in Ag is extremely low and these elements most likely form multiphase sam-ples when alloyed with Ag. There are three other elements marked in red which can form other but not hexagonal phases at low Ag concentrations (≤50 at.% Ag). Ti and Zr, for example, can form tetragonal phases with Ag. Hexagonal phases have previously been reported to exhibit low friction and wear in many systems. These elements are marked in blue in Fig. 4.1. The phase diagrams for Ag-Al, Ag-In, and Ag-Sn are shown in Fig. 4.2 a-c. The hcp phase in these systems are examples of Hume-Rothery compounds, or called electron compounds, in which the appearance of a particular phase requires a defined value of the ratio of valance electrons to number of atoms (e/a) [39], The e/a of equilibrium hcp structure of Ag alloys is formed at approximately 1.6. This means that Sn (e=4), In and Al (e=3) can be alloyed with Ag to form hcp structure at relatively high Ag concentrations (e.g. Ag0.8Sn0.2, and Ag0.67Al0.33), while Zn, Cd, and Hg (e=2) require much less Ag. However, metastable hcp phases can also be formed when e/a range (≈ 1.4-1.8), depending on the synthesis process and the atomic size of the alloy-ing element [40].

30

Fig. 4.2.Phase diagrams of the Ag-Al a), Ag-In b), and Ag-Sn c) systems. The ar-rows show the hcp solid solution phases [41-43]. Published on-line by ASM Interna-tional Alloy Phase Diagram Database. Copyright ASM International 2016

31

It should be noted that, for practical applications as electrical contacts, the Ag concentration in the Ag alloys should be higher than 50 at.%, in order to obtain good electrical properties. In this thesis, only Al, In, and Sn were selected for detailed studies due to their ability to form hexagonal phases at > 50 at.% Ag. Zn, Cd, and Hg was thus excluded due to the fact that they form hexagonal phases with Ag at relatively low Ag concentrations (<30 at.% Ag).

4.3 AgFeO2 Delafossites are a large family of compounds with the general formula ABO2. Typically, the monovalent A cation consists of a metal from group 8 or 9 (Pd, Pt, Ag, or Cu), while the trivalent B cation can be metals in group 3 and 13 (e.g. Sc, Y, La, Al, Ga, In, or Tl) or transition metals such as Cr, Mn, Fe, Co, and Ni [15, 44-46]. The delafossites can be divided into two groups depending on the electronic configuration of the A element. Compounds with the A-cation having a d10 electronic configuration, such as Cu and Ag, exhibit semiconducting or insulating electrical behavior, while delafossites with unfilled d-orbitals such as Pt and Pd, show metallic conductivity [15, 45, 47]. PdCoO2 is the most conductive oxide known so far. The electrical resistivity of this compound in the ab-plane is 4.69 μΩcm at room tempera-ture, which is comparable to that of copper [48].

Although delafossite-structured materials have been known for 140 years, the research interest in this type of materials has increased after the first report on the preparation of transparent p-type CuAlO2 thin films by Kawazoe and co-workers in 1997 [49]. This opened up the possibility to use delafossites as transparent conducting oxides (TCOs), since p-type TCOs are required in many practical applications based on p-n junctions. Since then, delafossites, especially those based on Cu and Ag, have received considera-ble attention due to their potential applications as e.g. p-type transparent conducting oxides [50-53], photocatalysts [54-57], anode material for Li-batteries[58], thermoelectrical materials [58-60], and gas sensors [61]. High temperature superconductivity was also suggested for doped members of this type of oxide compounds [62].

Many of the interesting properties of delafossites are related to the crystal structure with alternating planar layers of A atoms separated by slightly dis-torted edge-sharing B3+O6 octahedra. Each A atom is linearly coordinated to two oxygen atoms to form an O-A-O dumbbell unit parallel to the c-axis. Two different polytypes with a hexagonal (2H) or a rhombohedral (3R) structure can be formed depending on the stacking order of the layers. The hexagonal 2H type has a P63/mmc space group symmetry with stacking or-der of ABABAB…, while the rhombohedral 3R type has a space group of

32

R3m of ABCABC… layer stacking order. A sketch of delafossite structure of both polytypes is shown in Fig.4.3.

Fig. 4.3. Schematic representation of the delafossite structure (AgFeO2) for the 2H polytype with ABAB… stacking order along the c-axis and the 3R polytype with ABCABC… stacking sequence . The polyhedra represent Fe3+O6 distorted octahe-dra. AgFeO2 can be described as a nanolaminate with relatively weak O-Ag-O bonds introducing low interfacial shear strengths between the FeO2 layers.

4.4 AgI AgI is a well-known photosensitive material and is hence unstable under radiation of light. This made it widely employed in photography prior to the advent of digital photography. The basic mechanism involved in this process is as follows: absorption of a photon liberates an electron and a positive hole in AgI. The electron will combine with a mobile interstitial Ag+ ion, leading to a separation of the atom from the structure. Upon repeated absorption of photons, a cluster of Ag atoms can be formed, leading to color change after exposure to sunlight. AgI is also commonly used in cloud seeding [63], su-perionic conductors [64, 65], solid state batteries [66], halide ion sensors [67], photocatalyst [68] and antibacterial agent [69]. As will be shown in the thesis, AgI also yield low friction in AgI/Ag contacts.

33

Silver iodide can exist in two phases at room temperature: the β- and γ-phase. The β-phase (β-AgI, wurtzite-hexgonal) is thermodynamically stable and can be found in nature as a mineral. The γ- phase (γ-AgI, cubic zinc blende structure) is metastable and can be formed as a result of defects in the main β-phase. At a temperature of 147oC, it is well know that AgI undergoes a phase transition forming a high temperature stable α-phase (α-AgI, body-centered cubic) [70]. It is also known that the conductivity of AgI is mainly ionic and based on the migration of Ag+. In α-AgI, iodide anions form a rigid frame, in which Ag+ cations can move rapidly with high mobility. This can be considered as a “quasimolten” structure, in which the anion sublattice is “immersed” into a “cationic” melt [70, 71]. This type of material is classified as “superionic”. The mechanism behind the low friction behavior is un-known but can perhaps be related to the superionic properties described above.

4.5 Graphene and graphene oxide Graphene consists of a single monolayer of carbon atoms, tightly packed in a 2D honeycomb lattice. A sketch of the graphene structure is illustrated in Fig. 4.4. It also illustrated that graphene can be a basic building block for graphitic materials of all dimensionalities. It can be wrapped up into 0D fullerenes, rolled into 1D nanotubes, and stacked into 3D graphite [72]. Free-standing graphene was first discovered experimentally in 2004 by Geim et al. [73]. The method to produce graphene at that time was to exfoliate the layers by mechanically removing individual atomic planes using an adhesive tape. After the initial discovery, several different methods have been devel-oped to fabricate graphene, e.g. reduction of graphene oxide from graphite oxidation[74], ultrasonic exfoliation [75], epitaxial growth on carbides [76]. and chemical vapor deposition (CVD) techniques [77]. Graphene has been widely studied for its remarkable mechanical, electrical, optical, and thermal properties [78-81]. For example, it has been reported to be strongest materi-als in the world, about 200 times stronger than the strongest steel. It conducts heat and electricity very efficiently and is nearly transparent to light. The tribological properties of graphene, on the micro-scale or as a lubricating additive, have attracted intensive attention as well [31, 82, 83]. Berman et al. demonstrated in 2013 that graphene can dramatically reduce friction as a dry solid lubricant in macro-scale tribological tests [35-38].

34

Fig. 4.4. Graphene is a 2D building block for other carbon materials. It can be wrapped up into 0D fullerenes, rolled into 1D nanotube, and stacked into 3D graph-ite. Reprinted with permission of Elsevier from Geim [72].

In reality, the GO sheet is non-planar with folding and cracks. Graphene oxide layers consists of graphene-like layers with a wide variety of oxygen groups attached to the carbon atoms such as carbonyl (C=O), carboxyl (C-OOH), hydroxyl (C-OH), and epoxy (C-O-C), see Fig. 4.5. The most com-monly preparation method for GO is a top-down process called Hummers’ method where strong oxidizing agents, such as sulfuric acid H2SO4, sodium nitrate NaNO3, and potassium permanganate KMnO4, oxidize graphite and exfoliate the produced graphite oxide [84]. GO has also been prepared by using a bottom-up method in which glucose is dehydrated under hydrother-mal conditions and transferred to a substrate by dipping and lifting prior to annealing [85]. Graphene is hydrophobic and has a very low solubility in most solvents. In contrast, GO is dispersible in water and polar organic sol-vents, due to the presence of the hydrophilic functional groups. This makes it advantageous to use GO as an additive for lubrication applications [82] as well as a wide range of biological applications such as drug delivery [86], cellular imaging [87], and biosensing [88, 89].

35

Fig. 4.5 Schematic structures of graphene and graphene oxide. Reprinted with per-mission of Elsevier from Malhostra [89].

36

5 Methods

5.1 Material synthesis 5.1.1 Bulk synthesis The binary Ag-X (X=Al, In, and Sn) alloys, presented in paper I, were syn-thesized by melting small pieces of the raw elemental materials in evacuated ampoules (< 1 mTorr) at 970oC for two hours, followed by quenching in cold water. The quenching rate was approximately 200oC/s.

In the more detailed study on Ag-Al alloys (paper II), all samples were prepared by vacuum arc melting, which heats the metals via an electric arc struck between a tungsten electrode and the metals placed in a Cu crucible. Generally, several kA of DC current are used to start an arc between the two electrodes. A cooling system with cold water surrounds the crucible and also the tungsten electrode preventing them to over-heat. In this work, elemental Ag and Al powders were melted in in a laboratory-built arc furnace, previ-ously pumped to vacuum (1< Torr) and then filled with Ar up to 400 Torr, in order to avoid oxidation. A piece of Zr was used as a getter to further reduce the oxygen content prior to the melting. The homogeneity of the samples was ensured by repeating the arc melting process three times. The synthe-sized alloys were then annealed in vacuum at 500 ºC for about 24 hours. The annealing was also beneficial for reducing defects and to release stress.

5.1.2 Thin film deposition In the present work, we have constructed a combinatorial platform including a combinatorial sputtering system, which can deposit thin films with large composition gradients in a single experiment. The system was designed and built in-house. The system is based on an ultra-high vacuum chamber (base pressure below 10-8 Torr) using a sputtering-up configuration, with a non-rotating substrate located above four targets. Each target has a shield and is tilted 39o relative to the substrate normal. This design together with a non-rotated substrate allows for the deposition of films with a continuous compo-sitional gradient. The sketch of the combinatorial sputtering is shown in Fig. 5.1a. An example of composition distribution of a binary gradient Ag-Al thin film, deposited on a 3” Si wafer, can be seen in Fig. 5.1b. Since four differ-ent targets can be mounted in the chamber, not only binary but also ternary

37

and even quaternary gradients are possible to obtain. In the work conducted in this thesis, the main focus was on binary systems (Paper III and IV), but addition of Cr in a gradient ternary Ag-Al-Cr system was also investigated. The results are yet unpublished but summarized on page 64-65.

Combinatorial material science (also called a materials genome ap-proach), is a high-throughput method, which can be used to rapidly explore and/or optimize new or known materials by systematic studies of the compo-sition-structure-property relationships. In this thesis, this approach has been used to screen Ag-based electrical contact thin films, mainly presented in paper III. It was also applied to rapidly identify the deposition parameters for reactive sputtering of the complex AgFeO2 oxide with narrow synthesis window, presented in paper IV.

Fig.5.1 a) Sketch of the combinatorial sputtering process, which can deposit compo-sitional gradient films; Example of compositional gradients for a Ag-Al thin film.

The combinatorial sputtering is based on a conventional magnetron sputter-ing process, which is supplied by a plasma, created by applying a potential between the cathode (target) and the anode (grounded chamber wall and magnetron shield) to discharge an introduced gas, usually Ar. For the reac-tive sputtering of AgFeO2 thin films in paper IV, a mixture of Ar and O2 was used. The plasma is self-sustained when the number of secondary electrons generated is sufficient to collide and ionize enough gas. A magnetic field is introduced by placing magnets behind the targets to trap the electrons close to the target surface and enhance the probability of the electron-Ar interac-tions, thereby increasing the density of the plasma. The densest plasma is situated in the region where the electric fields and the magnetic fields are perpendicular, resulting in effective erosion of the target and hence a race track. The whole unit comprising the target, target holder and magnets is called a magnetron. The ions in the plasma are accelerated towards target surface by an applied negative potential, and the target atoms are ejected and

38

transported in the vapor phase to the substrate, where they are deposited as a film. The chemical and phase compositions, microstructure, morphology, and properties of the deposited films are dependent on the sputtering rate of target, transport of the sputtered species, substrate temperature, and the ener-gy of the incoming species.

The number of atoms that are sputtered off the target per incident ion is called the sputtering yield. Different target materials have different sputter-ing yields. For example, Ag has very high sputtering yield (2.7 at 400 eV Ar ion energy), while Al and Fe have much lower sputtering yields (0.8 for Al and 1.0 for Fe at 400 eV Ar ion energy) [90]. The amount of material sput-tered from a target is controlled by the magnetron current. If a direct current (dc) power supply is used, the process is called dc magnetron sputtering. In the present work, one of the experimental parameters for tuning the sputter-ing rate is the target power. For a specific target, higher power generally gives a higher sputtering rate.

In reactive sputtering, the target can form an insulating layer on the target surface. In paper IV, oxide layers were formed due to the oxygen-containing atmosphere. Hysteresis curves like those shown in Fig. 5.2, can be useful to investigate the oxidation status of the targets by gradually increasing the O2 flow into the chamber. For example, the target starts with a clean metallic surface (metallic mode), which then is converted into a partially oxidized surface (transition state) and further oxidized to a fully oxidized layer when the O2 flow is sufficiently high (target poisoning, also called compound mode). Such processes can be seen in the hysteresis curves of the Ag and Fe targets shown in Fig. 5.2. As can be seen, the Ag target was in the compound mode when the flow of O2 was increased to 6 sccm, while the Fe target was poisoned starting already at a flow of 3 sccm O2. In the poisoning process, the target current can increase or decrease depending on the secondary elec-tron emission coefficients and ion bombardment on each target. For exam-ple, for the Ag target poisoning, more secondary electrons are emitted from the target surface, which increases the target current. However, for the Fe target, the situation is more complex. The minimum value of the target cur-rent could be attributed to the presence of suboxides of iron [91]. These sub-oxides may have different electronic properties, including a different sec-ondary electron coefficient, as compared to the fully oxidized material. In generally, the sputtering yield of compound target is much lower than the metal target. Fe is much more easily oxidized than the nobler Ag, and the deposition rate of Fe is thus much slower than the Ag target during the reac-tive sputtering.

39

Fig. 5.2 Hysteresis curves of Ag a) and Fe b) targets during the reactive sputtering. Black lines are for current change when the O2 flow is increased, while red lines show the current change for decrease in O2 flow.

In paper IV, a pulsed-dc power supply was used for the reactive sputtering of AgFeO2 thin films to reduce the sputtering rate of Ag, in order to obtain the stoichiometric Ag/Fe ratio. Compared with the constant dc sputtering process, pulsing of the sputtering target increases the number of energetic particles and the ion flux at the substrate, due to the increased plasma density caused by higher current density in the pulses [92]. This can also be con-firmed by the voltage and current evolution that the voltage of the Ag target decreased and the current of the Ag target increased with increasing frequen-cy (not shown here). In reactive sputtering, a higher frequency, with a fixed on-time, gives a lower power during each pulse to ensure a constant total power. As a consequence of this, the target surface may be more metallic at a lower frequency, resulting in a higher deposition rate. By adjusting the pulse frequency, it is consequently possible to change the flux of energetic particles as well as the deposition rate and thereby change the structure and properties of the growing film.

The transport of the sputtered target atoms and/or ions has also a strong influence on the deposition process. The sputtered species are subject to collisions with the gas molecules (gas scattering). These collisions will re-duce the kinetic energy of the species and hinder some of the species from reaching the substrate. It also can change the angle of the incoming species to the substrate. The gas scattering is directly dependent on the distance be-tween target and substrate, total gas pressure, and the type of target element used. In the work presented in this thesis, only the total pressure was tuned to influence the microstructure and composition of the deposited films.

The sputtered species will finally arrive at the surface of the growing film and migrate (diffuse) to energetically favorable positions. The surface diffu-sion can be increased by using higher substrate temperatures, or by applying a negative potential (so-called bias) to the substrate. The bias will increase the kinetic energy of incoming ions and provide additional energy. In the work presented in this thesis, no bias was applied and all substrates were

0 2 4 6 8 10

260

280

300

320

340

360

fully oxidized

suboxides

transition state poisoning

starts

b)

compound mode

Cu

rre

nt

(mA

)

O2 flow rate (sccm)

Fe target

metallic mode

0 2 4 6 8 10145

150

155

160

165

170

175

180

Cu

rre

nt

(mA

)

poisoningstarts

transition state

compound mode

metallic mode

Ag target

O2 flow rate (sccm)

a)

40

hence kept at the floating potential during the sputtering process. There is also the probability that species on the growing surface can be re-sputtered off by the bombardment of incoming energetic particles. The growth process by sputtering is not at thermodynamically equilibrium, owing to the high quenching rate (~ 108 oC/s), which enables the formation of metastable phas-es and also makes it difficult to form complex crystal structures with large unit cells.

In this thesis, thin films were deposited on different substrates: p-doped single-crystal Si (100) wafers (3’’ diameter), steel cylinders (10 mm diame-ter, 75 mm length), and flat Cu discs electrolytically plated with 10μm Ni coating (Cu:Ni) (3”diameter, 2 mm thickness). The Si wafers were used to investigate the composition, phase evolution, electrical resistivity and the hardness of the Ag-Al thin films. The steel cylinders were used for friction coefficient screening tests, while the Cu:Ni discs were used for contact re-sistance measurements.

5.1.3 Chemical synthesis In paper V, low-frictional AgI coatings were prepared by exposing Ag coat-ings to I2 dissolved in an ethanol-mixed water solution. The formation of AgI from elemental Ag and I2 can be described by the following redox reac-tion [93], similar as to the reaction taking place during the gaseous io-dination of Ag:

2 Ag 2 Ag+ + 2e- I2 + 2e- 2 I-

2 Ag + I2 2 AgI

According to Bose et al [94], the formation of AgI in the vapor phase can be described by a process where I2 is absorbed on the surface of Ag (and later the AgI), and where the iodine atoms are ionized by the capture of electrons from the Ag. Since it is well known that the conductivity of ionic AgI is mainly due to the migration of Ag+ ions, the Ag ions are believed to diffuse to the surface to react with I ions and form the AgI compound, leaving Ag vacancies inside the material. The formation of AgI in solution is more com-plex than in a vapor. The reaction is influenced also by the solvents and the I2 concentration. In general, AgI films can be rapidly synthesized at ambient atmosphere and room temperature in I2 dissolved solutions. However, with the assistance of ultrasound, the thickness and grain size of the resulting coating can be tuned. In an ultrasonic assisted process, expansion and implo-sion of gas bubbles occur in the solution. When the cavitation occurs in the presence of a solid surface, the implosion generates a jet stream of liquid that

41

hits the surface at a very high speed creating a local temperature and pres-sure. This effect can accelerate the involved chemical reactions [95].

In this work, low-frictional AgI coatings were synthesized by immersing precursor Ag coatings into a I2 containing solution (e.g. mixture of alcohols and water). The effect of various parameters, such as ultrasound power, treatment time, I2 concentration, and also different solvents on the synthesis process was studied. More details can be found in paper V.

AgI has also been formed inside nanoporous Ag. This was made by a se-lective etching of Al from Ag-Al thin films in basic (NaOH) solutions. In order to investigate the effect of the pore size on the filling of AgI, gradient Ag-Al thin films deposited by the combinatorial sputtering, were used as precursor coatings. The etched product was then filled with AgI by the iodi-zation of Ag in I2 containing solution or using gaseous I2.

5.2 Characterization of materials This section describes the characterization techniques used to determine the chemical composition, phase composition and microstructure of the samples. The primary methods used in this thesis are X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS) and Raman spectroscopy (Raman). These methods are therefore presented in more details. Other characterization methods such as scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray fluorescence (XRF), X-ray reflectivity (XRR) are only briefly described.

5.2.1 X-ray diffraction The crystalline structure can be determined by X-ray diffraction. The method uses coherent scattering of X-rays that have wavelengths in the same order of magnitude as the distance between the lattice planes in a crystalline struc-ture. From the diffractograms, which exhibits peaks at discrete diffraction angles, the crystalline structure and lattice parameters can be determined [96].

Standard XRD measurements are based on a so-called locked-couple, or θ-2θ measurements. The incident angle (ω) and the out-coming angle (θ) are changed simultaneously. For the coatings with highly textured orientation, it will be difficult to determine the phase composition using the θ-2θ method, since it may only show one or a few dominant peaks from lattice planes par-allel to the surface, while other reflections are not observed. However, it can provide useful indication of texture in the coatings. Such effect can be seen on a textured Ag-Al thin film, as shown in Fig.5.3.

42

Fig. 5.3 Example of a diffractogram from a θ-2θ measurement of a textured Ag-Al thin film. The diffractogram only shows one peak from planes parallel to the surface of the textured coating.

As can be seen in Fig. 5.3, a peak can be attributed to different phases. For highly textured films with only one or a few peaks are observed, this may lead to a situation where no clear phase identification is possible. In such a case, a pole figure measurement can be used. In these measurements, the 2θ angle is fixed at a constant value, to match a desired diffraction peak, while the rotation (φ) and tilt angles (ψ angle) are varied. The plane angle between the textured plane and the desired plane can then be detected. The pole fig-ure measurements, presented in this thesis, were performed on a Philips MRD X’Pert diffractometer, using CuKα radiation. The optics and set-up parameters are described in Paper II and III. Fig. 5.6 shows examples of pole figures from two selected Ag-Al thin films with different Al contents. They both show textured structures with in-plane rotation, presented as rings. Fig. 5.6a shows an example of a fcc-(200) pole figure of Ag-Al thin film with 6at% Al, showing a strong ring at 54o relative to the center. This indicates a fcc (111) texture in the thin film. In contrast, the sample in Fig. 5.6b shows a strong hcp (002) texture from a hcp-(101) pole figure (62o be-tween the (100) and (002) planes), indicating that the coating had a hcp structure with a strong c-axis oriented texture.

43

Fig. 5.4 fcc-(200) pole figure of Ag-Al thin film with (111) texture a); hcp-(101) pole figure of Ag-Al thin film with (002) texture b).

For thin films, another XRD method, called grazing incidence XRD (GIXRD), is often used to increase the signals from the thin films and de-crease the substrate contribution. For GIXRD, the incident angle is kept at a small and constant value (1-5o), while the out-coming angle is varied. This set-up can probe lattice planes of different orientations. However, it should be noted that GIXRD may in fact not show all of these when highly textured or epitaxial films are studied. A contribution from some orientations could be enhanced due to the use of the low grazing incident angles. This can be clearly seen in Fig. 5.5 where the intensity of the (103) peak was enhanced in a Ag-Al film with hexagonal structure, using an incidence angle of 5o. By using an incident angle of 10o, the contribution from the (103) plane was reduced significantly. The GIXRD measurements performed in this thesis were made using a Bruker D8 DISCOVER X-ray diffractometer or a Sie-mens D5000 instrument with parallel-beam set-up. The details regarding the measurement parameters can be found in Paper III-V.

44

Fig. 5.5 Example of the incident angle effect in GIXRD

The Bruker D8 DISCOVER X-ray diffractometer, also equipped with an automatic function on the sample holder stage, which enables X-ray scan-ning of the whole sample to investigate the crystalline phase distribution of gradient samples. The X-ray beam was line focused along the gradient direc-tion. The diffractograms obtained at different positions along the gradient direction can be plotted in a waterfall view (see Fig. 5.6a). The data can also be shown as 2D XRD maps (see Fig. 5.6b) by projection of the waterfall views, where the peak intensities are color coded. Such a 2D map clearly provides an overall view of the phase evolution over the entire composition gradient. More details regarding the measurements can be found in paper III and IV.

Fig. 5.6 Examples of a waterfall view a) and a 2D map view b) of the diffracto-grams, created by the EVA software.

hcp (103)

incident angle 10o

30 40 50 60 70 80

hcp (002)

Inte

ns

ity

(a

.u.)

Inte

ns

ity

(a

.u.)

incident angle 5o

2θ (degree)

45

5.2.2 X-ray photoelectron spectroscopy X-ray photoelectron spectroscopy (XPS), also called electron spectroscopy for chemical analysis (ESCA), is a technique based on the photoelectric ef-fect. In XPS, the binding energy can be determined by measuring the kinetic energy of the emitted electrons from the material after illumination by X-rays of known energy. Since the binding energies are specific for each ele-ment and orbital, XPS can be used for elemental identification. The relative chemical composition of the material can be determined from the intensities of each photoelectron peak, which are normalized by sensitivity factors. In the present study, the latter were either supplied by the XPS analysis soft-ware MultiPak, or obtained by calibration using reference samples with known composition. XPS spectra can also be analyzed to identify the chemi-cal bonding state of elements, since the binding energy of a specific element will be affected by the chemical bonds. This is illustrated in Fig. 5.7, show-ing the Al2p peak acquired in the wear track of a Ag-Al alloy after tribologi-cal testing. There are two visible peaks in the spectrum, which can be at-tributed to two contributions, one from Al-O bonds (~ 75.8 eV) [97] and the other one from metallic Al-Al or Al-Ag (Al-Me) bonds (~ 73.5 eV) [98].

Fig. 5.7 XPS Al2p spectrum acquired in the wear track of a Ag-Al alloy after tribo-logical testing. Unpublished data.

The analysis depth of XPS is limited by the electron mean free paths for which the photoelectrons have enough energy to escape the material surface and reach the detector. The mean free path is dependent on the material ana-lyzed and kinetic energy of the electrons, which in turn depends on the ener-gy of the X-ray source. The XPS data presented in this thesis were acquired with AlKα radiation with an energy of 1487 eV, resulting in electron mean free paths of up to a few nm [99]. Therefore, XPS only probes the outmost surface layer of the samples. In order to achieve information regarding the bulk inside the sample, depth profile analysis has to be carried out using Ar+

80 78 76 74 72 70

Al-Me

Inte

ns

ity

(a.u

.)

Binding energy (eV)

Al-O

46

ion sputtering to etch the surface of the sample. However, the energy sup-plied through the ion bombardment process will introduce defects in the analysis volume and may influence the chemical bonding state of the ana-lyzed material [100]. In the case of Ag-X alloys, different etching energies (500 eV, 2 keV, and 4 keV) were used to etch the oxide layers on the sur-face. No significant difference concerning the chemical bonding state was observed when using different ion energies. To reduce the total measurement time, a relative high etching energy was used in paper I (4 keV) and paper II (2 keV), respectively.

5.2.3 Raman spectroscopy In Raman spectroscopy, materials are studied based on their interaction with monochromic light, usually using laser light in the visible, near infrared, or near ultraviolet range. The Raman effect relies on inelastic light scattering, in which the laser light interacts with vibrational and rotational states of the materials, resulting in an energy difference, compared to the incoming light. Such energy differences are called Raman shifts, which usually are reported using the unit cm-1. It is important to note that not all materials are Raman active. Only molecular vibrations or phonons (collective vibrations) of the crystal, which are able to change the polarizability, will interact with the photons. Therefore, Raman spectroscopy is not suitable to identify elemental materials. However, it is a useful technique to identify non-metallic com-pounds and molecules. For example, the technique is widely used for charac-terization of carbon-based materials and oxides. In this thesis, Raman spec-troscopy measurements were performed on a Renishaw micro-Raman sys-tem with a 514 nm laser, and a Renishaw inVia microscope, in which 532 nm and 633 nm lasers were used.

Fig. 5.8 Raman spectra of graphene a) and graphene oxide b) flakes deposited on Ag plate. a) was adopted from paper VI, while b) depicts unpublished data

47

The Raman spectra of graphene and graphene oxide flakes can be seen in Fig. 5.8. They both show the characteristic peaks of graphite-like materials, ~ 1330 cm-1 (D peak), ~ 1600 cm-1 (G peak) and ~2650 cm-1 (2D peak) [101-104]. The D peak is caused by the breathing mode of sp2 atoms in rings, which are activated by disordered structures, e.g. edges, defects, or partial oxidation. The G peak is due to bond-stretching in rings and chains. The 2D peak gives information on the stacking order. The 2D peak in a single layer of graphene sheet is sharp and narrow and is shifted as a function of the number of graphene layers in the flakes from 2650 cm-1 in a bilayer to 2680 cm-1 in a five layer-flake and further to 2700 cm-1 in bulk graphite [103]. In addition, the shape of the 2D peak can also give information on the number of graphene layers. Typically, a broader peak indicates the presence of sev-eral graphene layers in a flake [105]. In the case of few-layer flakes, another feature is observed at ~ 1600 cm-1 (D’ peak) [103]. The Raman results, shown in Fig. 5.7a, suggests that the graphene flakes, presented in paper VI, consisted of few-layer graphene sheets. As a comparison, the Raman spec-trum for graphene oxide flakes, as seen in Fig. 5.7b, shows a very weak 2D peak, and a strong D peak together with a strong background. Similar Raman features have been observed by Ahlberg et al. and were attributed to the amorphization of graphene by the introduction of defects during the ion bombardment [106].

5.2.4 Additional characterization methods Scanning electron microscopy (SEM) was widely used in this thesis as a standard technique to obtain the morphological information for sample sur-faces and fractured cross-sections of deposited coatings on Si substrates. In SEM, an electron beam is emitted from an electron gun and focused on the sample by magnetic lenses. A common image mode used in this thesis is to detect the secondary electrons, which are ejected from the surface when the incoming electrons are inelastically scattered. The instruments used in this thesis were a LEO 1550 microscope and a Zeiss Merlin instrument, both equipped with a field emission gun as the electron source. The imaging was mainly performed using the secondary electrons and an InLens detector in the 1550 and Merlin instruments.

Transmission electron microscopy (TEM) utilizes electrons transmitted through a thin sample and often requires complex sample preparation, usual-ly using a focused ion beam (FIB). A bright-field image is acquired by de-tecting the transmitted electrons through the sample in the same direction as the original beam. In this mode, more electrons can pass through pores and thinner or light-atom regions, and these regions will thus appear bright. In contrast, the dark-field imaging mode detects the diffracted electrons, mean-ing that pores appear dark. High-resolution TEM (HRTEM), which is based on the interference of the transmitted electrons, is very useful to obtain high

48

resolution information on the samples, e.g. identifications of amorphous or crystalline structures. Scanning TEM (STEM) provides a Z-contrast image, meaning that phases or areas with higher atomic number will appear bright. Combining this technique with HRTEM can give valuable information on atomic-resolution features. It is also possible to identify phases by using selected-area electron diffraction (SAED). Energy Filtered TEM (EFTEM) chemical maps can be acquired by filtering the transmitted electrons to only detect those with a specific energy. The TEM studies presented in this thesis were carried out with a Tecnai F30ST TEM and a FEI Titan3 G2 60−300 TEM instruments. More details on the parameters used in the TEM studies are given in in papers I, III and IV.

It is also possible to perform elemental analysis in the SEM and TEM in-struments using energy dispersive X-ray spectroscopy (EDX) (also abbrevi-ated EDS) technique. It relies on the emission of characteristic X-rays from samples when their atoms are excited by the electron beam. Information regarding the chemical compositions and elemental mappings can be ob-tained with this technique.

X-ray fluorescence (XRF) is another method to measure the chemical composition. XRF measures the intensity of the fluorescent X-ray emitted from a sample when excited by the primary X-ray beam. Each element pro-duces a set of characteristic fluorescent X-rays. The intensity of the fluores-cent X-rays from the specific element is proportional to the number of the specific atoms present in the sample, which can be quantitatively analyzed using reference samples with known thickness and density. It should be no-ticed that the compositions measured with XRF when using elemental refer-ence samples can be subject to matrix effects. However, XRF can easily be applied to provide trends in composition in a series of samples with similar compositions. The instrument used for XRF in this work was a PANalytical Epsilon 5 instrument with an automatic system, which enables it to automat-ically scan many samples in a single batch. All the samples were masked to give the same analysis area.

X-ray reflectivity (XRR) is a non-destructive analytical technique to esti-mate the thickness, density, and roughness of coatings with smooth surfaces. When X-rays are irradiated on to the sample at a certain angle, X-rays are reflected from the two interfaces of the sample: the air/coating and coat-ing/substrate interfaces. The two reflected beams with a path difference give rise to interference fringes. The periodicity of the fringes is proportional to the thickness of the film while the initial drop in intensity is proportional to the density of the film and the amplitude of the fringe is proportional to the roughness of the layers. The thickness and density can be obtained by fitting a simulated curve to the measured one. XRR and XRF are used as comple-mentary techniques. In this work, XRR was performed with a Philips MRD X'Pert diffractometer on the Ag and Fe reference samples to estimate their

49

thicknesses and densities to allow calculations of the Ag/Fe ratios based on XRF measurements.

Optical profilometry is an optical method used to study the topography of a surface without any physical contact. The incoming light is reflected by a reference mirror and the sample. The sample topography gives rise to inter-ference fringes, which are used to obtain a height profile. An optical profiler WYKO NT1100 was used to characterize the height profile of the wear tracks after the tribological tests.

5.3 Evaluation of properties 5.3.1 Hardness measurements Microhardness measurements were carried out on bulk samples, as described in paper I and II, based on the Vickers hardness test method, using a Future Tech microhardness tester. A square base pyramid shaped diamond with a load of 300g was used in all measurements. The microhardness was deter-mined by using the area of the indentation, measured by optical microscopy. The average hardness value and standard deviation were calculated based on five measurements taken on different spots on each sample.

The hardness of the coatings was determined using nanoindentation, in which the penetration depth was recorded during the indentation. The result of a nanoindentation measurement is often displayed as a load-displacement curve. For most materials, the load-displacement curve shows both elastic and plastic deformation. The measured hardness was calculated as the ratio between the maximum applied load (Fmax) and the projected contact area, which can be calculated from maximum depth (hmax), according to Oliver and Pharr [103]. In this thesis, nanoindentation was performed with a UNHT nanoindenter (Anthon Paar) using a Berkovich diamond tip, as described in paper III.

5.3.2 Tribological testing Tribological measurements were performed on different set-ups depending on the geometry of the samples. A reciprocating pin-on-disk and a rotating ball-on-disk set-up, were used to determine the friction coefficients of the Ag based samples, as presented in papers I, II, V and VI. In the measure-ments, the sample is sliding back-and forth or rotated against the static coun-ter ball or pin, which is pressed against the sample with a constant normal force. The friction force is recorded continuously. The friction coefficient is given by the ratio of the measured friction force to the normal force applied. Testing parameters, e.g. load, speed, sliding distance, the number of cycles, and the counter materials, were varied. More details are given in the papers.

50

The friction coefficients between pure Ag and compositional gradient films, deposited on steel cylinder, were directly measured by a laboratory-built cross-cylinder tribometer, presented in paper III. In this set-up, two crossed cylinders are sliding against each other in a reciprocating motion (see Fig. 5.9a). The lower moving cylinder, i.e. a sample with gradient coat-ing, is placed at 45o with respect to the sliding direction. The stationary up-per specimen, e.g. a Ag-plated Cu cylinder, is oriented perpendicular to the sliding direction and spring loaded against the moving lower cylinder. Dur-ing each stroke, every position on the gradient sample, which refers to spe-cific composition, is tested at different times and against a specific position of the counter cylinder. This allows for efficient, individual and parallel test-ing of all compositions along the sample cylinder. By plotting the friction coefficients (z-axis) versus the composition (correlated to the position) and the sliding cycle number, a 3D image can be obtained to illustrate the fric-tion evolution as is shown in Fig. 5.9b.

Fig. 5.9 Sketch of the cross-cylinder tribometer a); 3D plot of the friction evolution during a cross-cylinder experiment b).

After the tribological tests, optical and electron microscopy together with EDS, XPS, Raman and sometimes TEM were used to investigate the wear track to determine the wear rate and to study the friction and the wear mech-anisms.

5.3.3 Contact resistance measurements The contact resistances of the samples were measured based on a four-terminal resistance method, measuring the voltage drop from the probe tip to the test surface. The voltage terminal on the probe was placed as close to the contact point as possible to minimize the resistance contribution from the probe. The path of the current and voltage divided immediately after the contact point to ensure the contact resistance measurement was not affected

51

by the sheet resistance of the measured film or the internal resistance of the wires. The sketch of the contact resistance measurement is shown in Fig. 5 10a. For the gradient coatings, the contact resistance at different positions can be rapidly measured using a laboratory-built set-up, which equipped with an automatic function on the sample holder stage, giving rise to a 2D map of contact resistance for the gradient sample. An example of such a map can be seen in Fig. 5.10b.

Fig. 5.10. Sketch of the contact measurement set-up a); An example of a 2D map of contact resistance for a gradient Ag-Al sample b).

5.3.4 Other properties measurements The electrical resistance was determined with a four-point-probe instrument (CMT-SR2000N, AIT), equipped with an automatic system, which ena-bles it to rapidly scan the electrical resistance of the coatings deposited on Si wafers to yield a 2D map of electrical resistance of the gradient coating for the whole wafer. The measured value is called the sheet resistance. The elec-trical resistivity of the coating can be obtained by multiply the sheet re-sistance value by the thickness of the coating, which can be acquired from cross-sectional SEM images.

The optical properties of the AgFeO2 films, deposited on a quartz glass substrate, were measured using a Lambda 900 UV-VIS-NIR PerkinElmer spectrophotometer (see paper IV). The transmittance (T) and reflectance (R) were measured and the absorption coefficient (α) was then determined ac-cording as described below [41]. = 1 ln 1

52

Here d is the film thickness in nm. The absorbance of the quartz substrate was negligible in comparison to that of the film. The band gap of the indirect or direct transition can be determined from a plot of (αhv)n versus hv if ex-trapolating the linear portion of the plot to zero, where n=1/2 for an indirect allowed transition, and n=2 for a direct allowed transition.

53

6 Results and discussion

The aim of my thesis is to find solutions to improve the properties of Ag as a sliding electrical contact material. My studies have been focused on the se-lection of material candidates, development of new synthesis processes for suitable materials, and a scientific understanding of the relationship between the composition, structure and properties of the studied materials. In this chapter, the most important results will be described based on the outline shown in Fig. 3.1.

6.1 Ag-X alloys It has previously been reported that metals with a hcp structure exhibit lower friction and wear coefficients than metals with a fcc structure [7]. It is well known that solid solutions with a hcp structure is formed in e.g. the Ag-Al, Ag-In and Ag-Sn systems and alloys containing these materials were there-fore selected for more detailed studies. To evaluate the structure and tribo-logical properties, a series of bulk samples were prepared by melting as de-scribed in Chapter 5.1.

6.1.1 Phase composition The phase compositions of three series of the Ag-Al, Ag-In, and Ag-Sn al-loys were studied in paper I. X-ray diffractograms for the samples are shown in Fig. 6.1. It can be seen that the phase content can be controlled by tuning the chemical composition. By increasing the amount of the alloying element X, the formed phase can be varied from a solid solution with fcc-structure at about 5 at.% X for the three series, to a solid solution with hcp-structure. The hcp phase is formed at 33at. % Al and In but already at 20 at.% for Sn. This trend is consistent with the Hume-Rothery rule, which states that the structure is controlled by the valence electron concentration, yielding a hcp structure for a e/a ratio of approximately 1.6. The diffractograms also show the presence of samples with multiphase composition of hcp together with other phases, depending on the composition for each system. In general, the phase evolution is in agreement with the phase diagrams shown in Fig. 4.2. However, there are a few deviations, especially in the Ag-In system, where the hcp phase is a high temperature phase, not stable below 170 oC. The

54

presence of the hcp phase in our samples can be explained by the rapid quenching step after the melting which makes it difficult for the thermody-namically stable low-temperature phase to form preserving a metastable hcp structure at room temperature.

Fig. 6.1 XRD of Ag-X (X=Al, In, or Sn) alloys. The symbols in the diffractograms represent different phases ( hcp, fcc, ● μ, █ tetragonal AgIn2, Sn). (From Paper I)

μ+fcc

fcc

hcp

♦Ag0.95

Sn0.05

♦

Ag0.50

In0.50

Ag0.67

In0.33

Ag0.75

In0.25

Ag0.95

In0.05♦♦

♦

♦♦♦

Ag0.50

Al0.50

Ag0.67

Al0.33

♦

•••• •

♦♦♦

• •Ag

0.80Al

0.20

Ag0.95

Al0.05

hcp+Al

fcc

hcp+fcc

hcp

AgIn2+hcp

Inte

nsi

ty (

a.u

.)In

ten

sit

y (

a.u

.)In

ten

sity

(a

.u.)

30 40 50 60 70 80

Sn+hcp

fcc

hcp+fcc

hcp

Ag0.50

Sn0.50

Ag0.80

Sn0.20

Ag0.88

Sn0.12

2θ (degree)

55

6.1.2 Evaluation of properties The tribological properties are dramatically affected by the composition and thus the crystalline structure of the Ag-X alloys. As can be seen in Fig. 6.2, the friction coefficient and wear rate were significantly reduced in the pres-ence of the hcp phase for all the three alloy systems. The friction coefficients were reduced from very high values (represented by a friction threshold of 2) in the fcc region, to < 0.4 in the hcp-regions. The reduction of the wear rate shows a similar trend as seen for the friction.

Fig. 6.2 The effects of chemical composition and phase composition on the friction coefficients a) and wear rates in the Ag-Al, Ag-In and Ag-Sn systems. Friction coef-ficient a) and wear rate b) of Ag-X alloys. The tribotests were performed as dry sliding at room temperature under 5N loads, 5cm/s speed for 1000 sliding cycles. The boxes in the figures show the formation region for each phase, e.g. pink for fcc, blue for hcp and grey for other phases (Al, AgIn2. or Sn). The overlapping of the boxes represents a phase mixture. (Adapted from paper I)

It should be noticed that a single-phase hcp structure was not required to obtain both a low friction coefficients and a low wear rate. A mixture of hcp and other phases, e.g. Ag0.50Al0.50, Ag0.25In0.75, Ag0.50In0.50, Ag0.88Sn0.12, and Ag0.50Sn0.50, also leads to reduced friction and wear rates. It is also interest-

0.0

0.5

1.0

1.5

2.0

Al

hcp

fcc

μ+fcc

Ag-Al

fcc0.0

0.5

1.0

1.5

2.0fcc

hcp

AgIn2

Fri

cti

on

co

eff

icie

nt

Ag-In

0 10 20 30 40 500.0

0.5

1.0

1.5

2.0

hcp

Sn

Ag-Sn

X content (at.%)

a)

0.0

0.4

0.8

0.0

0.4

0.8

0 10 20 30 40 500.0

0.4

0.8fcc

fcc

adhesive

wear

adhesive

wear

We

ar

rate

(1

0-4m

m3N

-1⋅m

-1)

Ag-Alfcc

adhesive

wear

AgIn2

Ag-In

Sn

Al

μ+fcc

hcp

hcp

Ag-Sn

X content (at.%)

hcp

b)

56

ing to note that Ag0.80Al0.20 with a phase mixture of fcc and μ, exhibited a reduced friction coefficient but that the wear rate in this case was relatively high. This indicates that the hcp phase plays an important role in the im-provement of wear.

In the fcc regions for all the three systems, the sliding contacts showed adhesive wear associated with material transfer and Ag cladding on both contacts. This behavior can be illustrated based on the morphology of the Ag/Ag track, shown in Fig. 6.3a-b. However, in the region with the presence of a hcp phase, the wear mechanism was altered from adhesive to oxidative wear in Ag-Al system, see Fig. 3c-d, and from adhesive to abrasive wear for the Ag-In and Ag-Sn alloys, see Fig. 3e-h.

Fig. 6.3 SEM images of tracks for the selected sample Ag0.67Al0.33, Ag0.67In0.33, and Ag0.88Sn0.12 after the tribological tests performed under load 5 N, with 5 cm/s speed for 1000 sliding cycles. (From Paper I)

All the Ag-X alloys had higher contact resistances (>200 %) than the pure Ag reference sample (see Fig. 6.4). In general, the contact resistance in-

57

creased with increasing content of alloying elements. The different alloys were expected to have a tendency to form oxide layers on the surfaces, which can explain the contact resistance trend. The contact resistance of the Ag-Sn alloys was higher than for the Ag-Al alloys for similar compositions although a Ag-SnO2 composite has been reported to be a potential electrical contact material [108]. The Ag-In alloys had lower contact resistances com-pared to those of the other two systems. The most In-rich Ag-In alloy had a considerably lower contact resistance than the corresponding Ag-Sn and Ag-Al alloys. This could be due a lower oxidation rate of the AgIn2 phase giving rise to a thinner oxide, which more easily can be penetrated.

Fig. 6.4 Contact resistance of the Ag-X (X = Al, In, and Sn) alloys for different contents of the alloying elements. The symbol represents the contact resistance for the Ag reference. (From Paper I).

A more-detailed study of the Ag-Al system was performed in paper II. In this case, the samples were synthesized with arc-melting as described in Chapter 5.1. The main trend regarding the phase evolution and the change of tribological and electrical properties is similar to that described in paper I. In real applications, different loads may be required during different working conditions. A study of the contact resistance as a function of the work load was therefore carried out (see Fig. 6.5a). It can be seen that the contact re-sistances decreased with increasing load until they reached almost constant values for loads higher than 40 N. This behavior can be attributed to: (i) an increase in the real contact area at higher loads and (ii) higher loads increas-es the possibilities to rupture the aluminum oxide layer on the surfaces of the

0 10 20 30 40 500.0

0.5

1.0

1.5

2.0

X content (at.%)

Ag-In

Ag-Al

Ag-Sn

Co

nta

ct r

esis

tan

ce

(m

Ω)

58

contacts. The load effect was more significant for the alloys with higher Al contents (>25 at.% Al), which probably were covered by thicker oxide lay-ers.

In a real application, the contact resistance may be affected by the sliding movements as well. In-situ measurements during tribological experiments were therefore performed in paper II, where the contact resistance was rec-orded simultaneously during the acquisition of the friction coefficients. The results shown in Fig 6.5b represent the dynamic contact resistance of Ag-Al alloys. Stable and low contact resistances can be observed for the alloy with low Al content (<25 at.% Al) and the pure Ag reference sample, while high-er Al content (≥ 25at.% Al) resulted in higher and fluctuating contact re-sistances. This could be attributed to the increase in hardness as well as to the formation of hard and less electrically conductive oxide particles in the wear debris.

Fig. 6.5a) The effect of the load on the static contact resistance of selected Ag-Al alloys; b) In situ contact resistance measurements during dry tribologocal tests at room temperature using a 2N load, 5 cm/s speed. (Adapted from Paper II).

6.1.3 Mechanism of low-frictional behavior XPS analysis was carried out inside and outside the tracks of three selected samples; Ag0.67Al0.33, Ag0.67In0.33, and Ag0.88Sn0.12. The XPS spectra in Fig. 6.6 enable a quantitative comparison of the oxide layer thicknesses. It can be seen that a relatively high amount of aluminum oxide was present on the surface outside the track for the Ag0.67Al0.33 sample, indicating that an oxide layer was present already before the tribotests. However, almost no surface oxide could be seen for the Ag0.67In0.33 and Ag0.88Sn0.12 samples prior to the tribological tests. After the tribotests, more oxidation of the surface was ob-served in the wear track of all the three samples. However, compared to the Ag-Al sample, much shorter Ar-ion sputtering time is required to remove the oxide from the Ag-In sample. Furthermore, a small shoulder on the main Sn-Me peak, for the Ag-Sn alloy, was observed for all in-track depth profile spectra, indicating that the oxidation occurred not only on the surface but also inside the sample, probably around the grain boundaries.

59

Fig.6.6 Depth profiling Al2p (a), In3d (b), and Sn3d (c) XPS spectra obtained out-side and inside the wear tracks for the Ag0.67Al0.33, Ag0.67In0.33 and Ag0.88Sn0.12 sam-ples, respectively; A plot of the O contents versus sputtering time obtained outside and inside the tracks for all three samples (d). (From Paper I).

To understand the mechanisms behind the low friction of the hcp phases, two TEM samples were prepared by FIB from the Ag0.75In0.25 sample, which showed both low friction coefficients and wear rates. One sample was pre-pared from a region outside the wear track, and the other was prepared inside the wear track. Fig. 6.7 shows the high magnified bright field TEM images of the two samples. Compared to randomly ordered grains in the area outside the wear track, shown in Fig. 6.7a, the cross section image in the wear track showed a thin layer (~ 5 nm) with horizontally aligned hcp phase, as shown in Fig. 6.7b. The inset shows the spacing distance of the basal planes is around 2.49 Å, which indicates a hcp (002) planes. Most likely, the low friction is created by this aligned structure with easily sheared planes in the hexagonal structure.

500 496 492 488 484 480

Inte

ns

ity

(a

rb.

un

its)

Binding energy (eV)

0 10 20 30

0

20

40

60

80

100

Ag0.67Al0.33 Ag0.67In0.33 Ag0.88Sn0.12

O c

on

ten

t (a

t%)

sputter time (min)456 452 448 444 440

Inte

nsi

ty (

arb

. u

nit

s)

Binding energy (eV)

0

20

40

60

80

100

O c

on

ten

t (a

t%)

Ag0.67Al0.33 Ag0.67In0.33 Ag0.88Sn0.12

Inte

ns

ity

(a

rb.

un

its

)

Inte

nsi

ty (

arb

. u

nit

s)

d)

Inte

ns

ity

(a

rb.

un

its)

c)b) Sn3dIn3dAl2p

insideinsideinside

outsideoutsideoutside

80 78 76 74 72 70

Inte

ns

ity

(arb

. u

nit

s)

Binding energy (eV)

inside

Sn-O

Sn-MeIn-Me

In-OAl-O

Al-Meoutside

Al2pa)

60

TEM Fig.6.7 TEM cross-sectional images of the Ag0.75In0.25 sample outside a) and inside b) the weak track after the tribotest of 1000 cycles, 5N load. Compared to the randomly oriented hcp grains outside the wear track, horizontally aligned hcp planes have formed on the top surface in the wear track (~ 5 nm thick).(Adopted from pa-per I)

In summary, the microstructure and the properties of the alloys can be con-trolled by tuning the composition of the alloys. Ag-X (X=Al, In, or Sn) al-loys with the presence of a hcp phase can significantly reduce friction coef-ficient of Ag and also reduce the wear rate. The TEM results confirmed that the low frictional behavior of the Ag-X alloys with presence of hcp phase can be attributed to the sliding-induced reorientation of easy-shearing basal planes, which are parallel to the sliding direction. However, the electrical contact resistances of the alloys are more than 2 times higher than the pure Ag reference sample. This indicates that it is a true challenge to obtain Ag-based alloys to replace pure Ag bulk with improved tribological properties simultaneously maintaining the similar good electrical properties.

The Ag0.50In0.50 alloy with a mixture of hcp and AgIn2, shows the lowest contact resistance, and low friction and wear rate of all studied samples. This suggests that the Ag-In alloys with mixture of hcp together with other phase (AgIn2), are the optimal alloys in paper I, for the sliding electrical contact application among all the bulk samples we have synthesized. More detailed investigations on Ag-In alloys should therefore be an interesting subject for further studies.

6.2 Alloy coatings 6.2.1 Rapid material screening Ag-based alloys can also be deposited as coatings to improve the tribological properties. A combinatorial approach has been used to rapidly screen Ag coatings alloyed with e.g. Nb, Cr, Zr, Ti, W, and Al for sliding electrical contact applications. Among these elements, only Al can form a hcp alloy phase with Ag, while the other elements are immiscible with Ag. Rapid fric-

61

tion screening on gradient Ag-Me (Me=Nb, Cr, Zr, etc) coatings showed that the immiscible alloys usually have high friction coefficients and that these also exhibited severe adhesive wear. An example of such phenomenon can be seen in Fig. 6.8 a, showing a Ag-Nb coating tested against a Ag counter surface. In contrast, a significant drop in the friction coefficient could be seen in tribotests involving Ag-Al coatings, containing around 13at.% Al. The optical images of the tracks suggested adhesive wear in the high-frictional position (5-12 at.% Al). No signs of adhesive wear was observed in the low frictional region (14-18 at.% Al). A transition can be clearly ob-served at around 13 at.% Al.

Fig. 6.8 Friction screening on gradient Ag-Nb (a) and Ag-Al (b) coatings. The Ag-Al coatings show a rapid drop in the friction at about 12 at% Al, while the Ag-Nb coatings exhibit high friction values for all Nb contents. (Unpublished results).

6.2.2 Detailed study on Ag-Al coatings The phase evolution in a continuous compositional gradient (6-60 at.%Al) is shown in Fig. 6.9. The results show that the fcc-Ag phase is formed with 6-13 at.% Al. This phase is highly textured with the close-packed {111} planes parallel to the substrate surface (note that the phase identification requires pole figure measurements as described in the supplementary information to paper III). A clear phase transition can be seen at about 13 at.% Al, where hcp-Ag started to form, resulting in a mixture of fcc- and hcp-Ag at 13 - 25 at.% Al. Single-phase hcp coatings, with the close-packed {001} planes par-allel to the substrate surface, are formed between 27 and 36 at.% Al. When the Al concentration is increased further, in the range 46-54 at.% Al, a mix-ture of fcc- and hcp-phase was formed. At the highest Al concentration (60 at.%), the structure was dominated by fcc-Al with Ag in solid solution to-gether with a very small amount of hexagonal phase. The results hence showed that the μ phase could not be observed in the as-deposited coatings, although the equilibrium phase diagram of the Ag-Al system shows the ex-istence of this phase at approximately 20-25 at.% Al. This can be attributed

62

to the fact that the deposition performed at room temperature was carried out under non equilibrium conditions.

The morphology at different positions of the gradient Ag-Al coating is shown as insets in Fig. 6.9. As can be seen, at low Al concentrations (6-15 at.%Al), the grains were small and randomly oriented. In contrast, the grains in the hcp-rich region (18-22 at.% Al) were parallel to the surface. Well crystallized hexagonal grains with sharp edges can be clearly seen at 27-32 at.% Al. A further increase in the Al concentration to 37-43 at.%, gave rise to smaller and tilted grains. Finally, for the most Al-rich films with a mixture of fcc- and hcp- phases, the grains became more spherical and particle-like.

Fig. 6.9. 2D XRD map based on θ/2θ measurements on three Ag-Al gradient coat-ings deposited on Si substrates. The peak at 38.1-39.0o originates from fcc (111) and/or hcp (002). SEM images of selected positions are shown as insets to illustrate the morphology evolution as a function of the Al content. (Adapted from paper III).

The TEM results show that the Ag-Al coating at 22 at.% Al had a homoge-neous composition distribution throughout the entire coating, which was approximately 1.5 μm thick. The coating exhibits a columnar structure with precipitated Ag nanoparticles, as shown in Fig. 6.10a. The Ag nanoparticles may have come from the material itself, since the XRD results indicate the presence of a mixture of fcc-Ag and hcp phases. However, it may also partly be a result of degradation since the sample was stored in air prior to the TEM analysis. Such degradation was also observed for another Ag-Al lamella cut from the bulk sample in paper II. The electron diffraction pattern taken from a large area shows a polycrystalline structure of the film (see the inset in Fig. 6.10a). Fig. 6.10b shows the atomic structure of a grain with a clear hcp

63

structure according to SAED (inset). As can be seen in Fig. 6.10c, the basal planes in the hcp grains were almost parallel to the film surface.

Fig. 6.10 STEM-HAADF overview image of the cross-section of an Ag-Al thin film with 22at.% Al (a). The inset shows a selected area electron diffraction pattern re-vealing a polycrystalline thin film (hcp- and fcc-Ag phase). HRSTEM-HAADF images showing the atomic structure (b) and orientation of the lattice with respect to the film surface (c). The inset shows a nanodiffraction pattern of the matrix reveal-ing the hcp structure. (From Paper III)

The hardness and electrical resistivity of the Ag-Al coatings are shown in Fig. 6.11. It is seen that the hcp phase has a higher hardness than the fcc phase. The mechanism of the hardening of Ag-Al alloys can be attributed to: i) solid solution strengthening at 10-13 at.% Al [4, 5]; ii) precipitation strengthening at 15-22 at.% Al; iii) reduced number of slip systems, three in the hcp structure, compared to twelve in the fcc structure, iv) reduced grain size at 38 at.% Al, introducing more grain boundaries and hindering disloca-tion glide across the boundaries [4]. Compared to the Ag-Al coatings, the bulk sample was much softer, only 0.2-0.8 GPa at 5-50 at.%Al. The reason could be the very fine grain size (<100 nm) in the Ag-Al thin films, see Fig. 6.9. In contrast, much coarser grains (~2 μm) were present in the bulk sam-ple, as shown in paper I.

64

Fig. 6.11 Hardness (a) and electrical resistivity (b) as a function of the Al concentra-tion for the Ag-Al coatings. (Adapted from paper III).

The electrical resistivity of the Ag-Al coatings increased with increasing Al content until it reached a constant value of about 35 μΩ⋅cm, at around 25 at.% Al, where the single hcp phase started to form. The electrical resistivity was much higher than the required values for the application. The reason for the high electrical resistivity could partly be attributed to increased scattering in the solid solutions. However, other factors such as oxidation can also con-tribute.

The contact resistance for the gradient coating on the whole wafer was al-so rapidly screened using an automatic procedure. Compared with the Ag reference coating, the gradient coatings showed a higher average contact resistance (see Fig. 6.12). However, for Al contents less than about 30 at.%, the average contact resistance was relatively low, maximum 50% higher than for the Ag reference sample. The results suggest that the aluminum oxide layer formed on the Ag-Al film was penetrated under the load of 5 N. The contact resistances of the Ag-Al thin films were generally lower than for the bulk samples (presented in paper I, II) for a similar composition range. An explanation can be that the oxide layers were more easily ruptured on the harder thin films than on the softer bulk surfaces.

Fig. 6.12 2D maps for the electrical contact resistance for a pure Ag reference a) and a gradient Ag-Al film b). (From Paper III)

0 10 20 30 40 50 600

2

4

6

8

pure Ag A B C

hcp + fcchcpfcc+hcpfccH

ard

ne

ss

(G

Pa

)

Al concentration (at.%)0 10 20 30 40 50 60

0

10

20

30

40

50

pure Ag A B C

hcp + fcchcpfcc+hcp

Ele

ctr

ica

l re

sis

tiv

ity

(μΩ

⋅cm

)

Al concentration (at.%)

fcc

a) b)

65

A sliding contact in a real application could be exposed to elevated tempera-tures. This may result in phase transformations and can also accelerate the oxidation process, and thus affect the properties. An annealing study was therefore performed on the gradient Ag-Al coatings both in vacuum (<10-8

Torr) and in ambient air. The electrical resistance of the gradient coatings with different Al contents at different temperatures was plotted in a 2D man-ner, which can be integrated with the phase evolution to yield a map, illus-trating the electrical resistance evolution as a function of the phase transfor-mation as shown in Fig. 6.13. It can be clearly seen that the maximum in-crease in the electrical resistance occurred in the region involving μ phase formation, while the hcp and fcc phases were unaffected by the annealing. The trend regarding the electrical resistance for the different phases in the Ag-Al system was μ > hcp > fcc. This indicates that μ phase should be avoided due to its low electrical conductance. The electrical resistance map for the annealing samples in air shows similar ranges as annealed in vacuum, indicating that, comparing to the phase composition, oxidation is less crucial to the increase in the electrical resistance.

Fig. 6.13 2D maps of electrical resistance of the gradient Ag-Al coatings after an-nealing in vacuum a) and air b) for two hours. (Unpublished results).

As discussed above, the oxidation of Al can be an issue for the degeneration of the electrical properties of Ag-Al alloys. Doping with other element may increase the oxidation resistance. An initial study on Ag-Al-Cr coatings con-taining ternary gradients, deposited by the co-sputtering of three elemental targets, was therefore performed (unpublished results). These coatings were also annealed at elevated temperatures (500 oC) in vacuum. As can be seen in Fig. 6.14, the electrical resistances increased with increases in Cr and Al contents. This could be attributed to more electron scattering and oxidation of Al and Cr. In contrast the sample annealed at 500oC showed an obvious reduction in resistance, from 0.8 mΩ for the as-deposited sample to 0.2 mΩ (at the right top corner), after the annealing. It is also interesting to observe that the gradient direction for the electrical resistance became more directed

66

along the Y-axis, while the as-deposited coating showed a tilted gradient direction, indicating the atoms diffused and probably formed new phases, during the heat treatment process at 500oC. For the annealed sample, the region at Y=-10mm shows the highest value (0.4 mΩ) (green color), while the region at Y=10mm exhibited lower electrical resistance. Preliminary results on the as-deposited Ag-Al-Cr coatings showed the presence of the same crystalline phase as in the Ag-Al coatings, but the peaks were much broader with stronger background. The background was most likely due to an amorphous phase, the amount of which increased with the Cr content. The formation of an additional amorphous phase can be explained by the fact that the Cr is immiscible in Ag. Other characterizations, e.g. XRD, XPS, as well as contact resistance and friction screenings would also be possible to carry out rapidly on the ternary coatings.

Fig. 6.14 2D maps for the electrical resistance of the gradient Ag-Al-Cr coatings before and after annealing at 100 and 500 oC in vacuum for two hours. The positions for the Al, Ag and Cr targets are given on the wafers. The composition gradients of the Ag-Al-Cr coatings are 2-5 at.% Cr, 15-30 at.% Al, and 70-85 at.% Ag. The com-position gradients agree well with the electrical contact gradient for the as-deposited coating.

In summary, we have demonstrated that a combinatorial approach is an effi-cient method for rapid screening of new materials for sliding electrical con-tact applications. In agreement with the results from the bulk Ag-Al alloys in paper I and II, the presence of an hcp phase resulted in dramatically reduced friction coefficients and reduced adhesive wear rates. The increase in contact resistance for the Ag-Al coatings are much lower than for the Ag-Al bulk alloys which can be attributed to the hardness difference. Also, it is clear that phase transformations at higher temperatures may have detrimental effects on the electric properties. Doping with Cr increases the electrical resistance, but the effect can be reduced by heat treatment, probably due to the for-mation of a new phase.

67

6.3 AgFeO2 compound coatings The Ag-based delafossites attracted our attention as potential low-friction materials based on studies of a low-frictional CrN/Ag composite coating where AgCrO2 was proposed to be formed in a lubricating tribolayer [13]. We selected AgFeO2 as a coating candidate and tried to deposit single phase coatings with this ternary compound using magnetron sputtering. The initial experiments, however, failed and we always ended up with a mixture of different phases. This was most likely due to the narrow homogeneity range of AgFeO2. In paper IV, we demonstrated the strength of a combinatorial approach to rapidly and effectively identify suitable process parameters for the deposition of single-phase delafossite AgFeO2 coatings by reactive sput-tering.

6.3.1 Optimization of the deposition parameters It is a complex process to co-sputter a single-phase ternary compound using two elemental targets with different chemical reactivities in an O2 atmos-phere. Many parameters will affect the deposition process such as the sub-strate temperature, pulsing frequency, target power, total pressure, and O2 partial pressure, etc. A small change in one process parameter may have strong impact on the composition and structure of the deposited coating, giving rise to a narrow processing window for the growth of pure crystalline delafossite AgFeO2 coatings. We found that the optimal deposition tempera-ture range is 400-500oC. The coatings deposited at lower temperatures (≤ 400oC) were amorphous, while higher temperatures (≥ 500oC) usually lead to precipitation of Ag. The Fe target is more easily oxidized in the reactive atmosphere, resulting in very low deposition rates. The sputtering yield of Ag is also generally much higher than for Fe. In order to get stoichiometric Ag/Fe ratio, a pulsing power was therefore applied to the Ag target to reduce the Ag deposition rate, while the power of the Fe target was set to 120W without using any pulsing.

Fig. 6.15 summarizes the effect of some critical parameters, e.g. the puls-ing frequency of the Ag target, the Ag power, and the O2 partial pressure, on the Ag/Fe ratio at different positions along the gradient. It can be seen that the Ag/Fe ratio of the gradient coating decreases with increasing pulsing frequency of the Ag target, as shown in Fig.6.15a. This can be explained by the fact that, in reactive sputtering, a lower frequency leads to a higher pow-er during each pulse to ensure a constant total power. As a consequence of this, the target surface may be more metallic at lower frequency, resulting in a higher deposition rate. The optimal pulsing frequency for the specific con-dition was 150 kHz, as this produced coating with Ag/Fe ratios close to one, which is desired for a stoichiometric coating).

68

Fig. 6.15 The Ag/Fe ratio as a function of the Ag target pulsing frequency a), O2

partial pressure b), and Ag power c and d). c) for 2 mTorr total pressure and d) for 4 mTorr total pressure. The data were acquired on different positions on a Si wafer substrate, where the composition gradient is along the Y-position of the wafer. The symbol represents the edges located on the Y-position of -30 mm (Ag-rich) and represents the edges located on 30 mm (Fe-rich). The middle position of 0 mm is represented by ●. (Adapted from Paper IV)

The composition is also sensitive to the O2 partial pressure. It can be seen in Fig. 6.15b that the Ag/Fe ratio increased with increasing O2 partial pressure. This can be explained by the different surface oxidation rates of the Ag and Fe targets in the reactive atmosphere. This trend can be explained by the current behavior seen in Fig. 5.2, leading to an increase in the Ag/Fe ratio with increasing O2 partial pressure.

Furthermore, the Ag/Fe ratio in the coating increased with increasing Ag power, due to the fact that higher power gives higher Ag sputtering rates. The composition is also sensitive to the total pressure. This effect can be seen when comparing the Ag/Fe ratios in Fig. 6.15c and d. In general, the deposition rate and the relative amount of high energetic particles decrease with increasing total pressure due to collisions between the particles and gas molecules. Another important influence of the total pressure, in the O2 con-tained reactive atmosphere, is the amount of negative oxygen ions, which are formed at the target surface and subsequently accelerated with the negative target potential towards the substrate. They can be reduced by higher total pressure due to the multiple collisions between the oxygen ions and Ar. It should be noted that the effects of all experimental parameters are coupled

14 15 16 17 18 190

1

2

3

4

Y= -30mm Y= 0mm Y= 30mm

Ag

/Fe

rati

o

Ag Power (W)8 10 12 14 16 18 20 22

0

1

2

3

4

Y= -30mm Y= 0mm Y= 30mm

A

g/F

e ra

tio

Ag Power (W)

d)

0.4 0.6 0.8 1.0

0

1

2

3

4 Y= -30mm Y= 0mm Y= 30mm

A

g/F

e ra

tio

O2 partial pressure (mTorr)

c)

0 50 100 150 200

0

2

4

6

8

10

12

Y= -30mm Y= 0mm Y= 30mm

Ag

/Fe

ra

tio

Frequency (kHz)

b)a)

69

and that their influence on e.g. the sputtering rate, target oxidation, particle energy and film growth is extremely complex.

Fig. 6.16 XRD maps for gradient films deposited using different pulsing conditions: c) 0 kHz; d) 100 kHz; e) 150 kHz; f) 200 kHz. The other deposition parameters were maintained constant: Ag: 15 W; Fe: 120 W, 0 kHz; total pressure: 4 mTorr; O2 par-tial pressure: 0.8 mTorr. The symbols represent different compositions (★: AgFeO2, ▲: Ag2O, ▼: AgO, ■: Ag, ●: Si). (Adapted from Paper IV).

In order to grow stoichiometric AgFeO2 coatings with the delafossite struc-ture, all experimental parameters have to be tuned to give an Ag/Fe ratio of one. However, a correct total composition ratio of Ag/Fe may not lead to single phase AgFeO2 coatings since binary silver oxides and iron oxides in amorphous and/or crystalline forms can also be formed. In paper IV, the gradients were mapped with XRD to determine the phase compositions and the results were presented in 2D XRD maps. One example of such a XRD map is shown in Fig. 6.16 where the influence of the pulsing frequency on the Ag target was investigated. In the gradients several different crystalline phases were observed. The pink background is due to the presence of an amorphous phase giving rise to a broad background intensity in the diffrac-tograms. As can be seen, an almost single phase AgFeO2 coating was depos-ited from -30 mm to 0 mm in Fig. 6.16c.

The optimum parameters for this experiment were found to be a deposi-tion temperature of 450 oC, 15W power on the Ag target using a pulse fre-quency of 150 kHz, a constant power of 120W on the Fe target, 4mTorr total pressure, and 0.8 mTorr O2 pressure. We could see that very small changes in each parameter had a strong impact on the structure of the deposited coat-

70

ing. Similar 2D XRD maps for the influence of the O2 pressure and the Ag target power are shown in paper IV.

In summary, the AgFeO2 sputtering process is complex but a combinato-rial approach can reduce the time needed to identify the optimum conditions. As discussed above, the sputtering rate and the energy of the arriving parti-cles, and hence the growth of the coatings, are strongly dependent on the sputtering process parameters. Single phase AgFeO2 can be expected when the energy is high enough to allow surface diffusion of the atoms allowing them to find correct positions to form the rather complex crystalline structure of the delafossites. This energy is provided by the substrate temperature and/or the kinetic energy of the incoming atoms. However, too high tem-peratures could result in decomposition of the delafossite to other Ag-based oxides. This may explain the narrow temperature range for the AgFeO2 growth. Another important factor is related to the kinetic energy of the arriv-ing particles, which can be influenced by the pulsing frequency, the total pressure, etc. Too high kinetic energies may damage the growing film while too low energies will not promote the surface mobility necessary for the crystalline film growth. The deposition rate should also be controlled, since too high rates do not provide enough time for the particles to diffuse and find equilibrium condition to form the crystalline structure.

6.3.2 Microstructure of the AgFeO2 coatings The SEM images shown in Fig. 6.17 illustrate the morphology of the two best samples. Sample A represents a film with clear (00l) peaks from the AgFeO2 phase but the background also suggests the presence of a mixture of AgFeO2 and an amorphous phase. In contrast, sample B represents a film with sharp (00l) peaks and very little background intensity. Sample A shows poor crystalline morphology with very small grains. In contrast, well crystal-lized flaky features can be observed on sample B. Another important obser-vation was that the AgFeO2 films were sensitive to radiation damage when imaged in SEM at high magnifications. After prolonged exposure of the electron beam, small particles started to appear on the surface. Most likely, these particles were Ag nanocrystals formed by O-Ag-O bond breaking dur-ing the electron beam exposure.

71

Fig. 6. 17. SEM images of a) sample A, and b) sample B. The XRD diffractograms are shown as insets. The experimental parameters are given in paper IV. (From Paper IV)

A TEM study (see Fig. 6.18a) confirmed that sample A consisted of two phases, c-axis oriented AgFeO2 (green areas) and amorphous Fe2O3 (pink areas). The HRTEM image shown in Fig. 6.18b shows the c-axis oriented (but rotated in-plane) AgFeO2 phase. The spacing between the planes was determined to 0.31 nm, which is consistent with the d-spacing of (004) plane in the 2H hexagonal structure. Clearly, the AgFeO2 phase grew with a strong (00l) texture, which is remarkable since it was nucleated on an amorphous phase. This gave rise to the web-like appearance seen in Fig. 6.18a. In con-trast, the XRD results suggested that sample B was a more or less single-phase AgFeO2 thin film. This was confirmed by an energy filtered TEM mapping of the sample B shown in Fig. 6.18c, which illustrates that the ori-ented AgFeO2 phase appeared throughout the entire film. However, Ag pre-cipitates (green) can also be observed on the lamella, which were generated by the radiation beam. This illustrates the instability of the AgFeO2 delafos-site phase. A HRSTEM image with Z-contrast clearly showed the layered structure of the AgFeO2-phase, as shown in Fig. 6.18d. The spacing between the Ag layers, which is a double spacing between a Ag and a Fe layer, was approximately 0.61 nm. This is almost twice the size of the d-spacing of the (004) plane with 2H hexagonal structure.

72

Fig.6.18 a) The energy filtered TEM mapping image and b) HRTEM image of sam-ple A; c) the energy filtered TEM mapping image and b) HRSTEM image of sample B. The green particles in c) are Ag grains formed during electron beam exposure (From paper IV)

6.3.3 Properties of the AgFeO2 coatings Initial tribological tests on the AgFeO2 coating did not show significant re-duction of the the friction coefficient. They were therefore considered to be unsuitable for sliding electrical contact applications. However, they may be interesting for other applications, for example, as p-type transparent semi-conductors. The optical and electrical properties were therefore investigated using films deposited on a quartz glass substrate. XRD showed the presence of a similar c-axis oriented AgFeO2 structure as the thin film deposited on Si. The optical absorbance of the AgFeO2 thin film in the range of 0.5 - 4.2 eV is showed in Fig. 6.19a. The absorbance was calculated from the measured transmittance and reflectance spectra in the visible and near-infrared region (shown in the inset). The band gap of the indirect or direct transition can be determined from a plot of (αhv)n versus hv if the linear portion of the plot is extrapolated to zero, where n=1/2 for an indirect allowed transition, and n=2 for an direct allowed transition. It can be seen from Fig. 6.19b that the indi-rect band gap is approximately 1.7 ± 0.1 eV, and the direct band gap is ap-proximately 2.5 ± 0.1 eV. The electrical resistivity of the AgFeO2 film de-posited on a quartz substrate was too high to measure by our four-point probe equipment.

73

Fig.6.19 Optical absorbance of the AgFeO2 thin film a) as calculated from the meas-ured optical transmittance and reflectance spectra, which are shown in the inset; the band gaps of the indirect and direct transition b). (From paper IV).

6.4 AgI coatings In paper V, we demonstrated a new and rapid process to deposit low friction AgI coatings in a cost-efficient way. As discussed in Chapter 3, AgI coat-ings on Ag can be obtained by electrodeposition. An alternative route, how-ever, is to simply expose the Ag surface to I2 dissolved in a solution and to use ultrasound to provide additional energy. In the sonochemical process, expansion and implosion of plenty of bubbles occur in the solution. When the cavitation occurs in the presence of a solid surface, the implosion gener-ates a jet stream of liquid that hits the surface at a very high speed generating high local temperature and pressure. This effect can accelerate the rates of

Fig. 6.20 SEM images of AgI coatings synthesized with sonication (a, b) and with-out sonication (c, d) at 50 oC, with 120 mg/45mL I2 dissolved in a solution com-posed of 50 vol.% water and ethanol for 20 minutes. a) and c) are top-view images of the surfaces; b) and d) are cross-sectional views of fractured samples deposited on Si substrates. (Adapted from Paper V).

74

the chemical reactions [95].It was found that an AgI coating was rapidly formed on the Ag surface by dipping Ag-coated substrates into a solution containing dissolved I2. The samples were either sputtered Ag coatings (~ 0.6 μm thickness) on Si wafer, or electroplated Ag coatings (~ 20 μm thick) deposited on Cu plates. AgI was formed both with and without the use of sonication. However, the use of ultrasound increased the reaction rate. The AgI coating synthesized with sonication was thicker and the grains were larger than those obtained without sonication, as can be seen in Fig. 6.20. It should be noted that AgI is not stable under SEM beam radiation and that small Ag nanoparticles hence were formed upon prolonged electron beam exposure.

A typical AgI coating has a dull grey-brown color, compared to a metallic luster of Ag-coated substrate, as shown in the inset of Fig. 6.20. The XRD diffractorgrams of the AgI coating show peaks due to the hexagonal β-AgI phase. It should be noted that a few peaks from the γ-AgI phase overlap with some of the β-AgI peaks. It is therefore possible that small amounts of the γ-AgI phase also were present in the coating. A few extra peaks can be seen as well which originate from the Ag under the AgI coating, as seen in the cross-sectional SEM images shown in Fig. 6.19b and d.

Fig. 6.20 XRD diffractogram for a typical AgI film. Inset: Photo of a typical AgI coating and a bare Ag-coated substrate. (Adapted from Paper V)

75

Fig. 6.21 Thickness of the AgI coating as a function of proportional (vol%) of etha-nol in a mixed ethanol/water solvent a), treatment time b) and concentration of I2 c). (Adapted from Paper V)

The effects of the solvent, treatment time and the I2 concentration are present-ed in Fig. 6.21. In all cases, the I2 was added in 45mL of solvent at room tem-perature and the samples were then sonicated at 200W. In paper V, different solvents, such as methanol, ethanol, isopropanol, and water were investigated. All the alcohols behaved similarly as solvents of I2 (more than 120 mg I2 could be dissolved in 45 mL alcohols). In contrast, the solubility of I2 in water is lower, only approximately 13 mg of I2 can be dissolved in 45mL of water. Interestingly, however, all solvents resulted in about the same thickness of AgI suggesting a rather weak influence of I2 concentration on the growth process. However, a completely different behavior was observed with a mixture of ethanol and water as a solvent. A study of the influence of the ethanol/water ratio showed that 50 vol% ethanol produced the thickest AgI coating, around 2 μm, compared with 0.5 μm for a pure ethanol or water solvent. The reason for the high efficiency of the 50 vol% ethanol solution involves two opposing effects: i) the growth rate is enhanced by the increased I2 concentration; ii) the reaction rate is reversely correlated with the vapor pressure. For the mixture of ethanol and water, the vapor pressure increases with increasing ethanol con-tent.

It also can clearly see that the thickness increases with increasing time, as shown in Fig. 6.21 b). It should be noted, that the thickness of the AgI coat-ing is also limited to the thickness of the precursor Ag coating. Moreover, the effect of the I2 concentration effect was investigated by varying the added amount of I2 from 10 up to 250 mg into 45 mL of the solutions, such as pure ethanol or mixture of ethanol and water. As seen in Fig. 6.21c), the thickness is weakly dependent on the I2 concentration with pure ethanol as solvent. However, a sharp increase in thickness was observed with increasing I2 con-centration in the 50 vol% ethanol solution. The other process parameters, temperature (20-65oC) and ultrasonic power (200-750W) were also investi-gated but they were not seen to significantly affect the thickness.

0.0 0.1 0.2 0.30

500

1000

1500

2000c)

pure ethanol

50 vol.% ethanol

Th

ick

ne

ss (

nm

)

I2 content (g/45mL)

0 5 10 15 200

500

1000

1500b)

pure ethanol

50 vol.% ethanol

Th

ick

ne

ss

(nm

)

Time (min)0 20 40 60 80 100

0

500

1000

1500

2000

2500

6 min

Th

ickn

ess

(nm

)

Ethanol (vol.%)

20 min

a)

76

Fig. 6.22 Friction coefficients (a) (Adapted from Paper V) and in-situ contact re-sistance (b) for the tribotests of AgI/Ag and Ag/Ag contacts (Unpublished date)

Reciprocating tribotests showed a clear reduction of friction coefficient for the AgI coatings compared to the pure Ag coating, as seen in Fig. 6.22a. Typically, the AgI coatings exhibited a friction coefficient of around 0.4 versus a Ag counter surface. In contrast, the Ag/Ag contact shows much higher friction coefficient and the experiment with the Ag/Ag contact pair had to be stopped after a few hundred cycles after reaching a friction thresh-old of about 2. The in-situ contact resistance of the AgI/Ag contact was ini-tially high (~1.4mΩ) but dropped, rapidly to approximately 100 μΩ after a few hundred cycles of sliding tests Fig. 6.22b. It was around two times of the value for the pure Ag reference sample. This increase in contact resistance is acceptable and shows that AgI coatings can be used in practical applications.

Fig. 6.23 Friction coefficient for the rotating ball-on-disc triblogical of steel balls against the AgI coatings synthesized with sonication (a, b) and without sonication (c, d). The tests were manually stopped after 1500 or 14000 cycles. The SEM images of the tracks as well as outside track are shown on the left side. (Adapted from Paper V)

0 2000 4000 6000 8000 100000

400

800

1200

1600

Cycles

Co

nta

ct r

esis

tan

ce (μ

Ω)

Ag Reference: ~ 50 μΩ

AgI: ~ 100 μΩ

b)

0 2000 4000 6000 8000 10000

0.2

0.4

0.6

0.8

1.0

Cycles

AgI/Ag

Fri

cti

on

co

eff

icie

nt

Ag/Ag

a)

77

Other tribotests on AgI coatings using a different set-up showed similar level of friction coefficients of around 0.3 for both the sonicated and the non-sonicated coating, as shown in Fig. 6.23. The friction coefficients started to increase to approximately 0.6 after around 6000 cycles for both samples, indicating that the AgI coatings were worn through to yield Ag/steel con-tacts. This is due to the fact that AgI is very soft, which results in high wear rates under these harsh conditions. The SEM images in the insets show that, the AgI grains were smeared out on the low-frictional track after the harsh sliding movement, compare the individual grain features seen outside the track, while the AgI grains were no longer visible in the long tested track.

Raman spectra were acquired for the tracks and the counter balls after the short cycle tests (1500 cycles) and long cycle tests (14000 cycles), in order to confirm the lubricant effect of the AgI coatings. The results show the presence of AgI with a peak at around 125 cm-1, in the low-frictional track and on the ball, while almost no AgI was found in the high-frictional track and on the ball after 14000 cycles see Fig. 6.24.

In summary, AgI coatings can potentially be used to improve the tribolog-ical properties of Ag contacts, while maintaining an acceptable contact re-sistance. However, under very harsh conditions, the lifetime of the AgI coat-ing is not long enough due to its high wear rate.

Fig. 6.24 Raman spectra for the tracks and the counter balls after 1500 a) and 14000 b) cycles of ball-on-disc tribological tests. The insets are the corresponding optical images. (Adapted from Paper V).

6.5 Designed microstructures The aim of this section is to demonstrate an alternative route to synthesize a low friction surface based on microstructure design. I have investigated the possibility to produce Ag/AgI nanocomposite coatings by filling the low friction AgI compound into nanoporous Ag. The AgI should be distributed throughout the entire structure of the material, allowing the lubricant to con-

100 200 300 400 500

Raman shift (cm-1)

AgI

Inte

nsi

ty (

a.u

.)

FG

G

FE

E

b)

100 200 300 400 500

AgI

D

C

DA

B C

A

B

Inte

nsi

ty (

a.u

.)

Raman shift (cm-1)

a)

78

tinuously regenerate the low frictional properties as the surface lubricant is worn down to ensure a relatively stable low friction throughout its life-span. The process consists of two steps: i) the synthesis of nanoporous Ag and ii) the filling of AgI into the Ag network.

6.5.1 Synthesis of nanoporous Ag An Ag-Al coating, deposited on a Si substrate by the combinatorial sputter-ing process with a continuous composition gradient (12-58 at.% Al), was used as a proof-of-concept. Nanoporous Ag can be obtained by a selective etching of Al in an alkaline solution. In this process, Al is dissolved into the solution under the formation of Al(OH)3 giving rise to a pore structure where the size and number of pores depend on the Al-content of the alloy. As can be seen in Fig. 6.25, small pores (~10 nm) start to form on the boundaries of the grains at low Al concentrations (12-22 at.%). The pore size increased with increasing of Al content, until the pore size was too large and the coat-ing disintegrated and detached from the substrate. This occurred for about 40 at.% Al using the conditions employed in Fig. 6.25. I have investigated the influence of process parameters such as the NaOH concentration (1-10 M), etching time (0.5-5 minutes), temperature (RT, 50 oC), substrate (Si, Cu:Ag, FTO, steel) on the formation of the nanopores.

Fig. 6.25 SEM images for the different positions of the gradient Ag-Al coating be-fore (upper images) and after (lower images) etching with 5 M NaOH for 5 minutes. (Unpublished data).

6.5.2 Filling of AgI The filling of AgI into the nanoporous Ag was initially investigated using the same process as for the growth of the AgI coatings in paper IV. The cross-sectional SEM images show that the etched coatings have a porous columnar structure, with the pores throughout the entire thickness, for the precursors with Al contents ≥ 22at.%. However, it was found that this AgI filling approach only was able to fill the larger pores leaving the smaller pores in films with < 22 at.% Al unfilled. This can be seen in Fig. 6.26 and

79

6.27 showing top-views and cross-sections of the samples. After only one minute in the I2 solution, AgI crystals were formed on the surface blocking the growth of AgI in the pores. An alternative method was tested by expos-ing the nanoporous Ag to an I2 vapor. As can be seen in Figs. 6.26 and 6.27, AgI was in this case formed throughout the entire film after a 2 min expo-sure also for a sample with 15 at% Al. As can be seen on the cross-sectional images that the AgI-filled coating has ~ 45% increase in the thickness of the coating with 15 at.% of Al, indicating that Ag will be taken from the pore walls forming a AgI phase, and the pores possibly were completely filled with the small pores. However, with 22 and 30 at.%Al, the porous size are too big, and the pores were not completely filled.

Fig. 6.26 Cross-sectional SEM images for the nanoporous Ag and the Ag/AgI coat-ings obtained by exposing the etched Ag-Al coating to I2 solution and vapor, respec-tively. The indicated Al content refers to the precursor Ag-Al coating. The scale bar is valid for all figures. (Unpublished results).

In summary, it can be concluded that a novel nanoporous Ag/AgI composite with designed microstructure has been successfully synthesized. The etching and filling processes are rapid, simple and easy to control. The initial recip-rocating tribological tests on the nanoporous Ag/AgI coating, with around 22 at.% Al, showed a high friction coefficient, indicating an issue with the me-chanical integrity of the structure. The smaller pores could be beneficial to increase the mechanical properties. Further studies are required to identify the optimal conditions for the application.

80

Fig. 6.27 Surface SEM images for the nanoporous Ag and the nanoporous Ag/AgI coatings obtained after exposing the etched Ag-Al coating to I2 solution and gas respectively. The Al content refers to the precursor Ag-Al coating. The scale bars are valid for all figures on the same row. (Unpublished results).

6.6 Solid lubricants As described in Chapter 3.2.3, graphene has a lubricating effect on steel surfaces. In paper VI, the lubricating effect was evaluated by comparing the friction coefficients of a clean Ag surface with those for a graphene-modified Ag surface sliding against a Ag counter surface. The graphene-modified Ag surface was obtained by evaporation of three droplets of a commercial graphene-containing ethanol solution, purchased from Graphene Supermarket Inc. The results from the tribological tests under a 1 N load clearly showed the presence of a lubricating effect for the graphene layers (GL), as shown in Fig. 6.28. The friction coefficient was approximately 0.06 for the Ag/GL/Ag contact. The test was stopped manually after about 150,000 cycles of tests without loss of the lubricating effect. The SEM im-age shows a plowing pattern with the remains of greyish graphene layers in the wear track. In contrast, the friction coefficient for the Ag/Ag contact was much higher (>1), and severe adhesive wear could be observed in the track without graphene. The graphene-coated Ag surfaces exhibited even lower contact resistance compared to the pure Ag surface. The reduction in the contact resistance can be attributed to the conductive graphene flakes, which increase the real conductive area.

81

Fig. 6.28 Lifetime testing of lubricating graphene in the tribological test (1 N load; 0.02 m/s speed; 2.5 mm track radius) a); The insets show SEM images of the inside of the tracks; Contact resistance of Ag and graphene-coated Ag surfaces for different loads b). (Adapted from Paper VI).

Such a significant reduction in the friction coefficient together with the long lifetime, as well as the low contact resistance suggest that graphene is a po-tential solid lubricant for sliding electrical contacts. However, the tribologi-cal properties may dependent on the counter surface, since the metal-graphene interactions are different for the transition metals. It is well-known that the Me-C bond strength decreases with increasing number of d-electrons in the metal. For example, the W-C bond strength is rather high and several tungsten carbide phases are known such as the hexagonal WC phase. In con-trast, the Fe-C bonds are weaker and only metastable iron carbides are known. Finally, the Ag-C bonds are very weak and no stable or metastable silver carbides are known.

Fig. 6.29 Lifetime testing of the lubricating effect of graphene in the tribological tests of graphene-coated Ag versus different metals (Ag, steel and W). The tribolog-ical testing parameters were: 2N load, 0.02 m/s speed, 2.5 cm radius. (Adapted from Paper VI)

1 3 50.0

0.5

1.0

1.5

Co

nta

ct r

esi

sta

nce

(m

Ω)

Load (N)

graphene/Ag Ag

0 50000 100000 1500000.0

0.5

1.0

1.5

2.0

Ag/GL/Ag

Fri

ctio

n c

oef

fici

ent

Cycles

Ag/Ag a) b)

0 10000 20000 30000 40000 500000.0

0.2

0.4

0.6

0.8

1.0

W

Steel

Fri

cti

on

co

effi

cie

nt

Cycles

GL/Ag

Ag

82

The lifetime testing of the lubricating effect of graphene-coated Ag against three different metal counters (Ag, steel and W), shown in Fig. 6.29, con-firms that Ag/GL/Ag contact had the longest lifetime (>40000 cycles), In contrast the steel/GL/Ag contact exhibits much shorter lifetime (~2700 cy-cles), while the W/GL/Ag contact had the shortest lifetime (~500 cycles).

To confirm the lubricating effect of the graphene layers, the characteriza-tions of Raman spectroscopy and SEM were carried out inside the tracks after the tribotests. It can be seen in Fig. 6.30a that many greyish flakes were present in the Ag/GL/Ag track. The flakes could be identified as graphene layers based on the Raman spectra, which were almost identical to those for the as-deposited flakes, except for a broadening of the 2D and D+D’ peaks. This indicates that more defects were introduced during the tribological tests. For the steel/GL/Ag track, much less residual graphene could be observed and the intensity of the 2D peak decreased, indicating that the graphene lay-ers were partially consumed. No graphene flakes could be observed in the SEM image of the W/GL/Ag track, and the Raman spectrum showed peaks which could be attributed to tungsten oxides as well as very weak and broad D and G peaks. Such peaks are typical for amorphous carbon and suggest that the graphene was highly damaged or completely destroyed during the sliding test.

Fig. 6.30 Raman spectra and SEM images for the tracks on the graphene-coated Ag plate after the tribological tests using different counters: Ag a), steel b) and W c). The scale bars in the SEM images represent 100 μm. (Adapted from Paper VI).

XPS analysis was used to investigate the chemical bonding in the tracks to understand the reason for the different tribological behaviors in different metal-graphene interfaces. As shown in Fig. 6.28 a), Ag, C and O signals were detected in the all Me/GL/Ag tracks. Extra W peaks were also observed in the W/GL/Ag track. However, no Fe peaks were seen in the steel/GL/Ag track. This indicates that W was transferred to the Ag surface from the W counter ball during the sliding movement, while no such transfer occurred for the steel ball. High resolution spectra showing the W4f peak indicated that W was oxidized (data not shown here). The C1s spectra revealed a peak

1000 2000 3000

Inte

nsi

ty (

a.u

.)

Raman shift (cm-1)

Ag/GL/Ag

D+D'2D

G

D

1000 2000 3000

Inte

nsi

ty (

a.u

.)

Steel/GL/Ag

D+D'2D

G

D

Raman shift (cm-1)

b)

1000 2000 3000

Inte

ns

ity

(a.u

.)

Raman shift (cm-1)

GD

W-O

W/GL/Ag

a) c)

83

at 284.8 eV, which could be attributed mainly to the C-C bonds in graphene [106] for both the Ag/GL/Ag and steel/GL/Ag tracks. In contrast, the main C1s peak for the W/GL/Ag track could be attributed to a type of W-C inter-action. Another peak at around 289 eV, originated from a type of C-O bond, which could be observed for all three tracks. The relative intensity of the C-O contribution was slightly higher for the steel/GL/Ag pair and significantly higher for the W/GL/Ag pair. This suggests an oxidation of the graphene layers in contact with the steel and W counter surfaces. The peak shifts and increased C-O intensity for the W/GL/Ag track are in good agreement with a more or less complete decomposition of the graphene flakes and the for-mation of amorphous carbon and some kind of W-C compound.

Fig. 6.28 Survey a) and C1s b) XPS spectra for the Me/GL/Ag tracks after the tribo-logical tests. (From Paper VI)

In summary, our results clearly show that graphene is an excellent lubricant for Ag/GL/Ag contacts. In contrast, graphene showed a lubricating effect with tungsten as a counter surface but the friction coefficient was higher and the lifetime of the lubricating effect was much shorter. The steel/GL/Ag pair showed an intermediate behavior. The excellent lubricating effect of gra-phene in the Ag contact is believed to stem from its free-standing layer struc-ture and the weak interaction between graphene and Ag which results in an easy-shearing contact. A problem, however, with graphene is that the manu-facturing of pure single- or few-layer graphene is time- and cost-consuming.

As an alternative lubricant, graphene oxide is cheaper and easier to pro-duce. Graphene oxide is also soluble in water due to the hydrophilic groups on the surface. In a test experiment, we applied a few droplets of graphene oxide on a Ag surface. After evaporation of the water, a tribological test was

84

carried out, which indicated very low friction values, similar to those ob-served for the graphene lubricant (see Fig. 6.29). Hence, we can conclude that the less expensive graphene oxide may be an attractive and more cost-efficient alternative in practical applications.

Fig. 6.31 Friction coefficients of bare Ag contact (Ag/Ag), graphene lubricated Ag contact (Ag/GL/Ag), and graphene oxide lubricated Ag contact (Ag/GO/Ag). The tribotests were performed using 1N load. (Unpublished results)

0 500 1000 15000.0

0.5

1.0

1.5

2.0

Ag/GO/AgAg/GL/Ag

Cycles

Fri

cti

on

co

eff

icie

nt Ag/Ag

85

7 Conclusions and outlook

The aim of my thesis is to find solutions to improve the properties of Ag as a sliding electrical contact material. My studies have been focused on the se-lection of material candidates, development of new synthesis processes for suitable materials, and a scientific understanding of the relationship between the composition, structure and properties of the studied materials.

A number of Ag-based materials, such as Ag-X (X=Al, In, or Sn) alloy bulks, Ag-X (X=Al, Nb, etc) alloy coatings, AgFeO2 and AgI coatings, gra-phene and graphene oxides as solid lubricants, and a novel designed struc-ture with nanoporous Ag filling with AgI, have been studied. These mateials have been synthesized with different methods. The bulk alloys have been synthesized by melting, while the coatings have been either deposited by dc magnetron sputtering (for Ag-X alloy and AgFeO2 coatings) or chemical methods (for the AgI coatings and the nanoporous Ag/AgI). The sources of solid lubricants, graphene and graphene oxide, were commercial products. The chemical composition, phase composition, and microstructure of the samples have been characterized by XPS, XRD, Raman, XRF, SEM, EDX, TEM, XRR, and optical profilometry. The evaluation of properties of the materials have been focused on tribological properties (friction coefficient, wear rate), electrical properties (electrical resistivity, contact resistance) and mechanical properties (hardness).

It has been shown that the microstructure and the properties of the Ag-X alloys can be controlled by tuning the composition. An important result for the Ag-X alloy systems, both for bulks and coatings is that the presence of a hcp phase can significantly reduce the friction coefficient and wear rate of Ag. However, the electrical properties of the alloys are challenging to satisfy in practical applications. The Ag-Al and Ag-Sn alloys showed high electrical contact resistances, which could be explained by surface oxidation. The best properties was obtained for the Ag0.50In0.50 alloy, containing a mixture of of hcp phase and AgIn2.This alloy showed the lowest contact resistance com-bined with a low friction and wear rate. It is possible that Ag-In alloys can be improved further for various practical contact applications and a more detailed study of this system is hence recommended.

My thesis has also demonstrated the strength of a combinatorial material science approach as a high-throughput method to rapidly screen different alloys for the selection of new material candidates. In the present work, we have constructed a combinatorial platform including a combinatorial sputter-

86

ing system, which can deposit thin films with large composition gradients in a single experiment. The coatings can then be tested using the home-built, cross-cylinder tribometer as well as the electrical contact resistance screen-ing set-up. With these tools, the friction coefficient and the contact resistance can rapidly be measured directly on the gradient coatings.

The combinatorial approach was also used for a rapid identification of the optimal deposition parameters for reactive sputtering of a complex AgFeO2 oxide with narrow synthesis window. A (00l)-textured and single-phase Ag-FeO2 coating was successfully deposited for the first time without post-annealing. The friction coefficient of the delafossite AgFeO2 coating was high and it has limited applications for sliding electrical contact applications. However, it could be interesting for other applications, for example, as a p-type transparent semiconductor. The decomposition of AgFeO2 into Ag and Fe2O3, can also be interesting for battery or other energy-related applications.

My thesis has also demonstrated a new and rapid process to deposit low friction AgI coatings in a cost-efficient way. The AgI coatings exhibited low friction coefficients and acceptable contact resistances. At present, this pro-cess is used by ABB to coat Ag contacts for some specific on-load-tap-changer (OLTC) applications requiring low torque. However, under very harsh conditions, the lifetime of the AgI coating is not long enough due to the high wear rate. An alternative concept is to fill AgI into a nanoporous Ag matrix, allowing the lubricant to be continuously regenerates, as the contact surface is worn down. We have demonstrated that the nanoporous Ag/AgI composites with the designed structure can be synthesized. The etching and filling processes are simple, rapid, and easy to control. However, the initial tribological tests showed high friction coefficients, indicating an issue with the mechanical integrity of the structure. Further studies are required to iden-tify the optimal conditions for the potential use of this composite material.

The use of graphene as a solid lubricant in sliding electrical contacts has been investigated as well. Our results show that graphene is an excellent solid lubricant in Ag-based contacts. Furthermore, the lubricant effect was found to be dependent on counter surfaces. The lifetime of the lubricant ef-fect of graphene with the three different counters studied follows the trend: Ag > steel > W. The trend has been explained by the increased Me-C inter-actions. As an alternative lubricant, graphene oxide is cheaper and easier to produce. The preliminary tribologocal tests showed a similar frictional be-havior as for graphene.

Of all the tested Ag-based contact concepts, graphene oxide showed the most promising results for practical applications. The results described in page 84 has been the motivation for further collaboration between Uppsala University and ABB within this field and several new research grants from Vinnova have recently been approved to evaluate the use of GO/Ag nano-composites as a new contact material for smart grid applications.

87

8 Populärvetenskaplig sammanfattning

Vi använder i vårt dagliga liv ett mycket stort antal elektriska kontakter, t.ex. i datorer, telefoner och TV-apparater. I princip kan en elektrisk kontakt besk-rivas som en koppling mellan två elektriskt ledande motytor som ska kon-trollera ledning av elektrisk ström. När ytorna är i kontakt leds ström över kopplingspunkten. Om ytorna däremot är separerade är strömkretsen öppen och ingen laddning kan transporteras. En bra elektrisk kontakt måste upp-fylla en rad materialkrav som hög ledningsförmåga och framförallt en låg kontaktresistans över kopplingspunkten. De två vanligaste kontaktmaterialen idag är guld (Au) och silver (Ag). Dessa ädelmetaller har god ledningsför-måga och högt motstånd mot oxidation. Detta ger låg kontaktresistans, vilket är ett krav för en fungerande kontakt. En nackdel med ädelmetaller är dock det höga priset samt det faktum att de är mjuka och har en hög nötningshas-tighet och hög friktion.

Elektriska kontakter är viktiga komponenter i den s.k. griden där de t.ex. används för att reglera spänning i transformatorer. Det vanligaste kontakt-materialet i dessa tillämpingar är Ag. I griden verkar kontakten under tuffa villkor med höga laster och hög switch-frekvens. På senare tid har begreppet ”smart grid” använts för att beskriva ett system som snabbt och effektivt kan leverera elektricitet till olika delar av energiförsörjningssystemet beroende på lokala behov, samt kunna hantera olika förnyelsebara energiformer som solenergi, vindkraft och vågkraft, där stora variationer i energiproduktion kan ske under dygnets alla timmar. Detta skapar allt större krav på att kon-trollera spänningsförändringar vilket kommer att leda till högre switch-frekvens för systemets kontakter. Nötningshastighet och friktion av kontak-ter blir därför allt viktigare parametrar i valet av kontaktmaterial. Ag som används idag behöver därför ersätts med ett alternativt kontaktmaterial som förutom låg kontaktresistans också har ett högt nötningsmotstånd samt en låg friktion. För glidande kontakter är just friktionen kritisk för att säkerställa en lång livslängd. Problemet med Ag är detta material mot andra silverytor uppvisar mycket hög friktionskoefficient.

Ett mål med denna avhandling har varit att studera nya möjliga kontakt-material baserade på Ag som en komponent men med bättre egenskaper än den rena metallen. Ett problem med detta är att många av de egenskaper som efterfrågas är svåra att kombinera i ett enda material. Om vi t.ex. ska redu-cera friktionen behöver en smörjande lågfriktionsyta skapas. Denna består ofta av komponenter som ger en ökad kontaktresistans och gör materialet

88

oanvändbart för praktisk tillämpning. Att identifiera och utveckla ett nytt kontaktmaterial är därför en mycket stor vetenskaplig och ingenjörsmässig utmaning med stor kommersiell potential.

I mitt avhandlingsarbete har jag identifierat flera olika koncept för att skapa nya kontaktmaterial baserade på Ag. Jag har undersökt möjligheten att använda nya Ag-legeringar med hexagonal struktur. Hexagonala metaller och legeringar uppvisar ofta lägre friktion än metaller med kubisk struktur. Hexagonala fasta lösningar bildas i systemen Ag-Al, Ag-Sn och Ag-In vilka har studerats i detalj i form av bulka prover framställda genom en smält-ningsprocess. Resultaten visar att hexagonala legeringar ger kraftigt sänk-ning i nötningshastighet och friktion. Tyvärr oxiderar dessa legeringar med tiden vilket ger en ökad kontaktresistans. Resultaten visar dock att Ag-In legeringar ger minst oxidationsproblem och mer detaljerade studier bör utfö-ras för att utvärdera deras potential och begränsningar i kontaktillämpningar.

I många fall vill man använda tunna beläggningar av ett kontaktmaterial. Jag har i mitt avhandlingsarbete också studerat Ag-Al legeringar framställda med s.k. magnetronsputtering vid låga tryck. För att på ett mer tidseffektivätt sätt klarlägga sambandet mellan processparametrar, struktur och egenskaper har jag använt s.k. kombinatorisk materialsyntes. Det innebär att gradient-filmer med variation i kemisk sammansättning beläggs på ett substrat i ett enda experiment. Genom att analysera olika områden på substratet kan snabbt utvärdera bästa experimentella parametrarna och optimera belägg-ningens egenskaper. Resultaten från Ag-Al skikten gav liknande resultat som de från de bulka studierna och bekräftade att hexagonala fasta lösningar ger utmärkta kontakegenskaper.

En annan möjlighet till nya kontaktmaterial är att framställa icke-metalliska ytbeläggningar. Jag har här studerat magnetronsputtering av AgFeO2, som tillhör delafossiterna, en större grupp av ternära oxider med skiktstruktur, Det har tidigare förslagits att sådana skiktade oxider ha en smörjande effekt på ytor. Genom kombinatoriska studier har jag lyckat framställa enfasiga beläggningar av AgFeO2 och karakteriserat dess egen-skaper. AgFeO2 har inte tillräckligt goda mekaniska egenskaper och är inte tillräckligt stabilt för att fungera i en elektrisk kontakt med föreningen kan ha en praktiskt användning t.ex. som p-dopad halvledare.

Avslutningsvis har jag studerat möjligheten att använda grafen och grafe-noxid som smörjmedel i elektriska kontakter. Resultaten visar på mycket låga friktionskoefficienter och utmärkt kontaktresistans. Grafenoxid som är ett billigare material uppvisar liknade egenskaper som grafen. Tyvärr, slits även grafen snabbt bort om det appliceras på Ag-ytan. Genom att skapa en komposit av grafenoxid/Ag kan en bra elektrisk kontakt som uppfyller de flesta kraven framställas.

89

9 Acknowledgments

First and foremost, I would like to thank my excellent supervisor Prof. Ulf Jansson. To me, you are definitely much more than a supervisor. You are a precise and profound scientist, a strategic and thoughtful leader, an inspiring and friendly mentor, and a stylish and literary advisor. No words can simply express my sincere appreciations for your patient and enthusiastic guidance, insightful and constructive comments, for always timely providing positively phrased feedback, for your good company during the international confer-ences, for all the discussions on cultures, histories, and books, for your sup-port not only concerning work, for your merciful listening when I was de-pression in life.

I also would like to thank my co-supervisors, Urban Wiklund, and Anna M. Andersson. Urban, thank you so much for your guidance, insights, and support in the world of tribology and nanoindentation, for your constructive discussion and comments, for your pleasant smile every time I meet with you. Anna, you have given me fantastic opportunities to explore all my ide-as, although some of them are outside-the-box of the project. I appreciated it a lot! Thanks a lot for your guidance, comments, and discussions.

A lot thanks to Pedro Berastegui and Mamoun Taher! Pedro and Ma-moun, your help means a lot to me. Thank you so much for taking care of all the technical issues for the last few weeks when I stuck in the writing of the thesis. I would like to thank Thomas Thersleff, Oleksandr Kryshtal, Adam Kruk, Aleksandra Czyrska-Filemonowicz, and Lars Riekehr. Your nice TEM works contribute a lot to our research and publications.¨

My appreciations are also sent to Tomas Nyberg and Mikael Ottosson, for all the kind guidance, explanations, discussions, and comments on sputtering and X-ray diffractions, respectively. Tomas and Mikael, although you are not my official co-supervisor, you indeed help me a lot! I also would like to thank Åke Öberg in ABB. Although you were only my co-supervisor when Anna was maternal leave, your encouragements always inspire me.

I am very thankful to all current and former members pf the department of Chemistry-Ångström, for making a pleasant working place. I am grateful to Mats and Leif, for their valuable advisors. Mats, I enjoyed the discussions with you about food, culture, life, and all the different topics. A special thank send to the thin film group members: Erik, Paulius, Stefan, Linus, Kristina, Dennis, Johan, Sebastian, for discussions around research results and, but also for all non-work related joyful activities. Thanks to Erik, for guidance

90

me a lot about work. Paulius and Stefan, thanks for your kind efforts to com-fort me when life is tough. Kristina and Linus, thanks a lot for the nice com-panies in Garmisch or San Diego.

I also enjoyed the seminar and fika time with David, Johan, Ocean, Gun-nar, Yvonne, Annika, Sarmad, Kristina, Torbjörn, Karin, Martin, Fredrik, Daniel, Kersti, Julia, Jiefang, Taha, Solveig, Charlotte, Girma, Getachew, Matthew, Habtom, Tim, Mario, William, Eric, Fabian, Reza, Pavlin, Ronnie, Viktor. Special thanks to David, for your friendship and all kind of helps. Martin, nice to discuss with you about the difference between fcc and ccp. Ocean, thanks for your kind help with phase diagram drawing. Mario, nice discuss with you about the potential use of AgFeO2 in battery. Fabian, thanks for the useful information on the thesis production. Thanks also send to our former colleagues, Jill, Nils, Matilda, Kristain, Mattis, Andreas L. Thanks to Jill, for introducing me a good start with the AgI project. Andreas, you give me a nice introduction on the art of synchrotron XPS, and thanks a lot for your company in BESSY. Liping and Qingnan, thanks for the great time when we met at Ångström. Thanks to my current and former roommates, Jonas and Stefan for all the nice time in Å3418. Many thanks to the friends on the lunch table: Tian, Shuyi, Zhang Zhen, Li Mingkai, Yiming & Lina, Shuainan & Peng, Chengjuan, Chao, Wei & Miao, Zhaohui, Jia, Bing & Kenneth, Ruijun & Qiuqiu, Song and Yue, Yanyan, Ji & Andreas, Zhigang, Dou, Yue, and Wei. I had a lot fun and enjoyed the lunch time and after-work activities.

Thanks to Andreas. I will always remember the times we worked together to build Sleipnir, our super sputter, pieces by pieces, screws by screws. I also would like to thank Anders, Gabriella, and Björgvin in Physics Department. Thank you for the helps and supports during the building process. I also would like give my thanks to the members of the tribomaterials group: Petra, Elin, Åsa, Martina, and Harald, for your kind help when I had problems on the tribological tests. Petra and Elin, thanks a lot for letting me share Elvis with you. Åsa, thanks for helping me to fix the contact resistance set-up.

Thanks a lot to Janne and Victoria for the kind helps when I had problems in clean room. Acknowledgments also send to Tomas E, for introduction of Raman and discussion on determination of optical band gap. Also thanks to Peter B and Cesar, for the knowledge, I learned during the courses. Thanks also send to Anders Lund, Eva, Tatti, Peter and Patrik, for your kind help.

Special thanks to my Uppvika friends Xia & Shuanglin, Xin & Jing, Lianghui. I will remember all the nice memories we spent together and sup-ported each other. Yu, thanks a lot for your company during my toughest time in the dark winter. Many thanks to my friends, Shuzhen & Shangfeng, Chuan & Di, Haisha & Bo, and the lovely kids. Thank you so much for your kind supports and helps when I was alone with Angelica. Thanks Valeria & Tomas, for the nice discussions and dinners. Thanks Lena, for the nice time when I worked with you in microsystem group before my PhD study. Mr.

91

Zhao, you open Ångström’s door for me and let me have chances to work in a few interesting projects in solid state Physics division. I am also appreciat-ing your recommendation for applying the PhD position. 你是我非常重要的良师益友。

Finally, I would like to thank my family. To my parents, 妈妈，谢谢你的坚强。爸爸，谢谢你照顾妈妈，让我可以坚持完我的学业。谁言寸草心，报得三春晖， 谢谢你们的爱。To my parents-in-law: 萌爷爷萌奶奶，如果不是你们这个月里来帮助我，我的论文不可能按时提交。真挚的谢谢你们！To Tao, Thanks for your love! 谢谢你十几年如一日的对待. To Angelica, my sunshine, you are the reason I became stronger and better.

92

10 References

[1] http://www.naonworks.com/old/inc_html/sub2_3.html. [2] R. Holm. Electric contacts: theory and application, Springer Science &

Business Media, 2013. [3] I.M. Hutchings. Tribology : friction and wear of engineering materials,

Edward Arnold, London, 1992. [4] W.D. Callister, D.G. Rethwisch. Materials science and engineering: An

Introduction, John Wiley & Sons, Inc., New York, 2005. [5] S. Saitoh, K. Sasaki, T. Nezu, M. Taira, Y. Araki. Hardness and

Composition of Solution-Treated Ag-Pd-Cu-Au Alloy Microstructures, MATERIALS TRANSACTIONS 49 (2008) 2360-2364.

[6] E. Rabinowicz. Friction coefficients of noble metals over a range of loads, Wear 159 (1992) 89-94.

[7] D.H. Buckley, R.L. Johnson. The influence of crystal structure and some properties of hexagonal metals on friction and adhesion, Wear 11 (1968) 405-419.

[8] A.Y. Chobee Taguchi, Hirinori Matsumoto, Nobukazu Ikeda, Akihiro Morimoto. Sliding parts and process for producing same. In: Osaka Sa, (Ed.). US005879819A, vol. 5,879,819. Japan: Japan Scientific Engineering Co., Ltd.; Fujikin Incorporated, 1999.1-6.

[9] S. Siol, T.P. Dhakal, G.S. Gudavalli, P.P. Rajbhandari, C. DeHart, L.L. Baranowski, A. Zakutayev. Combinatorial Reactive Sputtering of In2S3 as an Alternative Contact Layer for Thin Film Solar Cells, ACS Applied Materials & Interfaces 8 (2016) 14004-14011.

[10] S.M. Han, R. Shah, R. Banerjee, G.B. Viswanathan, B.M. Clemens, W.D. Nix. Combinatorial studies of mechanical properties of Ti–Al thin films using nanoindentation, Acta Materialia 53 (2005) 2059-2067.

[11] J.D. Perkins, J.A. del Cueto, J.L. Alleman, C. Warmsingh, B.M. Keyes, L.M. Gedvilas, P.A. Parilla, B. To, D.W. Readey, D.S. Ginley. Combinatorial studies of Zn-Al-O and Zn-Sn-O transparent conducting oxide thin films, Thin Solid Films 411 (2002) 152-160.

[12] T. Gebhardt, D. Music, T. Takahashi, J.M. Schneider. Combinatorial thin film materials science: From alloy discovery and optimization to alloy design, Thin Solid Films 520 (2012) 5491-5499.

[13] C.P. Mulligan, T.A. Blanchet, D. Gall. Control of lubricant transport by a CrN diffusion barrier layer during high-temperature sliding of a CrN–Ag composite coating, Surface and Coatings Technology 205 (2010) 1350-1355.

[14] R.D. Shannon, C.T. Prewitt, D.B. Rogers. Chemistry of noble metal oxides. II. Crystal structures of platinum cobalt dioxide, palladium cobalt dioxide, coppper iron dioxide, and silver iron dioxide, Inorg. Chem. 10 (1971) 719-723.

93

[15] M.A. Marquardt, N.A. Ashmore, D.P. Cann. Crystal chemistry and electrical properties of the delafossite structure, Thin Solid Films 496 (2006) 146-156.

[16] R.B. Gall, N. Ashmore, M.A. Marquardt, X. Tan, D.P. Cann. Synthesis, microstructure, and electrical properties of the delafossite compound CuGaO2, J. Alloys Compd. 391 (2005) 262-266.

[17] Z. Deng, X. Zhu, R. Tao, W. Dong, X. Fang. Synthesis of CuAlO2 ceramics using sol-gel, Mater. Lett. 61 (2007) 686-689.

[18] J. Ding, Y. Sui, W. Fu, H. Yang, S. Liu, Y. Zeng, W. Zhao, P. Sun, J. Guo, H. Chen, M. Li. Synthesis and photoelectric characterization of delafossite conducting oxides CuAlO2 laminar crystal thin films via sol–gel method, Appl. Surf. Sci. 256 (2010) 6441-6446.

[19] N. Tsuboi, Y. Takahashi, S. Kobayashi, H. Shimizu, K. Kato, F. Kaneko. Delafossite CuAlO2 films prepared by reactive sputtering using Cu and Al targets, J. Phys. Chem. Solids 64 (2003) 1671-1674.

[20] B.L. Stevens, C.A. Hoel, C. Swanborg, Y. Tang, C. Zhou, M. Grayson, K.R. Poeppelmeier, S.A. Barnett. DC reactive magnetron sputtering, annealing, and characterization of CuAlO2 thin films, J. Vac. Sci. Technol., A 29 (2011) 011018.

[21] A.S.S. B. Electrical contact element and use of the contact element. In: AB A, (Ed.). US, vol. 6,565,938 BI. Sweden, 2003.

[22] S. Arnell, G. Andersson. Silver iodide as a solid lubricant for power contacts. Proceedings of the Forth-Seventh IEEE Holm Conference on Electrical Contacts (IEEE Cat. No.01CH37192), (200)1.239-244.

[23] J. Lauridsen, P. Eklund, J. Lu, A. Knutsson, M. Odén, R. Mannerbro, A.M. Andersson, L. Hultman. Microstructural and Chemical Analysis of AgI Coatings Used as a Solid Lubricant in Electrical Sliding Contacts, Tribology Letters 46 (2012) 187-193.

[24] W. Guo, J. Yin, H. Qiu, Y. Guo, H. Wu, M. Xue. Friction of low-dimensional nanomaterial systems, Friction 2 (2014) 209-225.

[25] Q. Xue, W. Liu, Z. Zhang. Friction and wear properties of a surface-modified TiO2 nanoparticle as an additive in liquid paraffin, Wear 213 (1997) 29-32.

[26] L. Cizaire, B. Vacher, T. Le Mogne, J.M. Martin, L. Rapoport, A. Margolin, R. Tenne. Mechanisms of ultra-low friction by hollow inorganic fullerene-like MoS2 nanoparticles, Surface and Coatings Technology 160 (2002) 282-287.

[27] M. Kachoei, F. Eskandarinejad, B. Divband, M. Khatamian. The effect of zinc oxide nanoparticles deposition for friction reduction on orthodontic wires, Dental Research Journal 10 (2013) 499-505.

[28] J. Cumings, A. Zettl. Low-Friction Nanoscale Linear Bearing Realized from Multiwall Carbon Nanotubes, Science 289 (2000) 602.

[29] L. Joly-Pottuz, F. Dassenoy, J.M. Martin, D. Vrbanic, A. Mrzel, D. Mihailovic, W. Vogel, G. Montagnac. Tribological properties of Mo–S–I nanowires as additive in oil, Tribology Letters 18 (2005) 385-393.

[30] C. Lee, Q. Li, W. Kalb, X.-Z. Liu, H. Berger, R.W. Carpick, J. Hone. Frictional Characteristics of Atomically Thin Sheets, Science 328 (2010) 76.

[31] T. Filleter, J.L. McChesney, A. Bostwick, E. Rotenberg, K.V. Emtsev, T. Seyller, K. Horn, R. Bennewitz. Friction and Dissipation in Epitaxial Graphene Films, Physical Review Letters 102 (2009) 086102.

94

[32] J.C. Spear, B.W. Ewers, J.D. Batteas. 2D-nanomaterials for controlling friction and wear at interfaces, Nano Today 10 (2015) 301-314.

[33] K.-S. Kim, H.-J. Lee, C. Lee, S.-K. Lee, H. Jang, J.-H. Ahn, J.-H. Kim, H.-J. Lee. Chemical Vapor Deposition-Grown Graphene: The Thinnest Solid Lubricant, ACS Nano 5 (2011) 5107-5114.

[34] V. Eswaraiah, V. Sankaranarayanan, S. Ramaprabhu. Graphene-Based Engine Oil Nanofluids for Tribological Applications, ACS Applied Materials & Interfaces 3 (2011) 4221-4227.

[35] D. Berman, A. Erdemir, A.V. Sumant. Graphene as a protective coating and superior lubricant for electrical contacts, Applied Physics Letters 105 (2014) 231907.

[36] D. Berman, A. Erdemir, A.V. Sumant. Reduced wear and friction enabled by graphene layers on sliding steel surfaces in dry nitrogen, Carbon 59 (2013) 167-175.

[37] D. Berman, A. Erdemir, A.V. Sumant. Few layer graphene to reduce wear and friction on sliding steel surfaces, Carbon 54 (2013) 454-459.

[38] D. Berman, A. Erdemir, A.V. Sumant. Graphene: a new emerging lubricant, Materials Today 17 (2014) 31-42.

[39] U. Mizutani, H. Sato, M. Inukai, E.S. Zijlstra. Theoretical Foundation for the Hume-Rothery Electron Concentration Rule for Structurally Complex Alloys, Acta Physica Polonica A 126 (2014) 531-534.

[40] T.B. Massalski. Lattice spacing trends in close-packed hexagonal phases based on the noble metals, J. Phys. Radium 23 (1962) 647-654.

[41] A.J. McAlister. The Ag−Al (Silver-Aluminum) system, Bulletin of Alloy Phase Diagrams 8 (1987) 526-533.

[42] W. Gierlotka, Y.-C. Huang, S.-W. Chen. Phase Equilibria of Sn-Sb-Ag Ternary System (II): Calculation, Metallurgical and Materials Transactions A 39 (2008) 3199.

[43] Z. Bahari, M. Elgadi, J. Rivet, J. Dugué. Experimental study of the ternary Ag–Cu–In phase diagram, Journal of Alloys and Compounds 477 (2009) 152-165.

[44] W.C. Sheets, E. Mugnier, A. Barnabé, T.J. Marks, K.R. Poeppelmeier. Hydrothermal Synthesis of Delafossite-Type Oxides, Chem. Mater. 18 (2005) 7-20.

[45] A.N.Banerjee. Reactive-Sputtered Wide-Bandgap p-Type Semiconducting Spinel AB2O4 and Delafossite ABO2 Thin Films for ‘Transparent Electronics’. in: D.Depla SMa, (Ed.). Reactive Sputter Deposition, vol. 12. Springer, NY, USA, (2008) 413-484.

[46] B.V. Beznosikov, K.S. Aleksandrov. Predictions of compounds in the family of delafossites, J. Struct. Chem. 50 (2009) 102-107.

[47] R.D. Shannon, D.B. Rogers, C.T. Prewitt, J.L. Gillson. Chemistry of noble metal oxides. III. Electrical transport properties and crystal chemistry of ABO2 compounds with the delafossite structure, Inorg. Chem. 10 (1971) 723-727.

[48] M. Tanaka, M. Hasegawa, H. Takei. Growth and Anisotropic Physical Properties of PdCoO2 Single Crystals, J. Phys. Soc. Jpn. 65 (1996) 3973-3977.

[49] H. Kawazoe, M. Yasukawa, H. Hyodo, M. Kurita, H. Yanagi, H. Hosono. P-type electrical conduction in transparent thin films of CuAlO2, Nature 389 (1997) 939-942.

95

[50] R. Nagarajan, N. Duan, M.K. Jayaraj, J. Li, K.A. Vanaja, A. Yokochi, A. Draeseke, J. Tate, A.W. Sleight. p-Type conductivity in the delafossite structure, Int. J. Inorg. Mater. 3 (2001) 265-270.

[51] H. Hosono, H. Ohta, M. Orita, K. Ueda, M. Hirano. Frontier of transparent conductive oxide thin films, Vacuum 66 (2002) 419-425.

[52] J. Tate, M.K. Jayaraj, A.D. Draeseke, T. Ulbrich, A.W. Sleight, K.A. Vanaja, R. Nagarajan, J.F. Wager, R.L. Hoffman. p-Type oxides for use in transparent diodes, Thin Solid Films 411 (2002) 119-124.

[53] C.G. Granqvist. Transparent conductors as solar energy materials: A panoramic review, Sol. Energ. Mat. Sol. C. 91 (2007) 1529-1598.

[54] S. Bassaid, M. Chaib, S. Omeiri, A. Bouguelia, M. Trari. Photocatalytic reduction of cadmium over CuFeO2 synthesized by sol-gel, J. Photochem. Photobiol., A 201 (2009) 62-68.

[55] A. Stadler. Transparent Conducting Oxides-An Up-To-Date Overview, Materials 5 (2012) 661-683.

[56] Z.H. Chonco, A. Ferreira, L. Lodya, M. Claeys, E. van Steen. Comparing silver and copper as promoters in Fe-based Fischer-Tropsch catalysts using delafossite as a model compound, J. Catal. 307 (2013) 283-294.

[57] Y. Ma, X. Zhou, Q.B. Ma, A. Litke, P. Liu, Y. Zhang, C. Li, E.J.M. Hensen. Photoelectrochemical Properties of CuCrO2: Characterization of Light Absorption and Photocatalytic H-2 Production Performance, Catal. Lett. 144 (2014) 1487-1493.

[58] T. Nozaki, K. Hayashi, T. Kajitani. Electronic Structure and Thermoelectric Properties of the Delafossite-Type Oxides CuFe1–xNixO2, J. Electron. Mater. 38 (2009) 1282-1286.

[59] Y. Ono, K. Satoh, T. Nozaki, T. Kajitani. Structural, magnetic and thermoelectric properties of Delafossite-type oxide, CuCr1-xMgO2 (0≤x≤0.5), Jpn. J. Appl.Phys. 46 (2007) 1071-1075.

[60] K. Hayashi, K. Sato, T. Nozaki, T. Kajitani. Effect of Doping on Thermoelectric Properties of Delafossite-Type Oxide CuCrO2, Jap. J. Appl. Phys. 47 (2008) 59.

[61] C. Li, J.H. Hsieh, J.C. Cheng, C.C. Huang. Optical and photoelectrochemical studies on Ag2O/TiO2 double-layer thin films, Thin Solid Films 570, Part B (2014) 436-444.

[62] A. Nakanishi, H. Katayama-Yoshida. Computational materials design for superconductivity in hole-doped delafossite CuAlO2: Transparent superconductors, Solid State Commun. 152 (2012) 24-27.

[63] A. Zipori, D. Rosenfeld, J. Shpund, D.M. Steinberg, Y. Erel. Targeting and impacts of AgI cloud seeding based on rain chemical composition and cloud top phase characterization, Atmospheric Research 114–115 (2012) 119-130.

[64] T.Y. Vergent'ev, E.Y. Koroleva, D.A. Kurdyukov, A.A. Naberezhnov, A.V. Filimonov. Behavior of the low-frequency conductivity of silver iodide nanocomposites in the superionic phase transition region, Physics of the Solid State 55 (2013) 175-180.

[65] H.M. Zhang, T. Tsuchiya, C.H. Lang, K. Terabe. Size-Controlled Agl/Ag Heteronanowires in Highly Ordered Alumina Membranes: Superionic Phase Stabilization and Conductivity, Nano Letters 15 (2015) 5161-5167.

[66] Y.-G. Guo, Y.-S. Hu, J.-S. Lee, J. Maier. High-performance rechargeable all-solid-state silver battery based on superionic AgI nanoplates, Electrochemistry Communications 8 (2006) 1179-1184.

96

[67] D. James, T.P. Rao. Potentiometric sensing of iodide using polymeric membranes of microwave stabilized β-AgI, Electrochimica Acta 66 (2012) 340-346.

[68] N. Kakuta, N. Goto, H. Ohkita, T. Mizushima. Silver Bromide as a Photocatalyst for Hydrogen Generation from CH3OH/H2O Solution, The Journal of Physical Chemistry B 103 (1999) 5917-5919.

[69] Y. Xu, S. Huang, H. Ji, L. Jing, M. He, H. Xu, Q. Zhang, H. Li. Facile synthesis of CNT/AgI with enhanced photocatalytic degradation and antibacterial ability, RSC Advances 6 (2016) 6905-6914.

[70] G. Burley. Kinetics and Mechanism of the Low-Cubic to Hexagonal Phase Transformation of Silver Iodide1, The Journal of Physical Chemistry 68 (1964) 1111-1114.

[71] O.M. Shabanov, R.T. Kachaev, A.A. Iskakova, S.M. Gadzhiev, A.P. Pashaev. Conductivity of α-AgI and its melt in high-intensity pulsed electric fields, Russian Journal of Electrochemistry 45 (2009) 356-359.

[72] A.K. Geim, K.S. Novoselov. The rise of graphene, Nature Materials 6 (2007) 183-191.

[73] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V. Grigorieva, A.A. Firsov. Electric Field Effect in Atomically Thin Carbon Films, Science 306 (2004) 666-669.

[74] S. Stankovich, D.A. Dikin, R.D. Piner, K.A. Kohlhaas, A. Kleinhammes, Y. Jia, Y. Wu, S.T. Nguyen, R.S. Ruoff. Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide, Carbon 45 (2007) 1558-1565.

[75] E. Widenkvist, D.W. Boukhvalov, S. Rubino, S. Akhtar, J. Lu, R.A. Quinlan, M.I. Katsnelson, K. Leifer, H. Grennberg, U. Jansson. Mild sonochemical exfoliation of bromine-intercalated graphite: a new route towards graphene, Journal of Physics D: Applied Physics 42 (2009) 112003.

[76] Interface structure of epitaxial graphene grown on 4H-SiC(0001), Physical Review B 78 (2008) 205424.

[77] Q. Yu, J. Lian, S. Siriponglert, H. Li, Y.P. Chen, S.-S. Pei. Graphene segregated on Ni surfaces and transferred to insulators, Applied Physics Letters 93 (2008) 113103.

[78] S. Stankovich, D.A. Dikin, G.H.B. Dommett, K.M. Kohlhaas, E.J. Zimney, E.A. Stach, R.D. Piner, S.T. Nguyen, R.S. Ruoff. Graphene-based composite materials, Nature 442 (2006) 282-286.

[79] C. Lee, X. Wei, J.W. Kysar, J. Hone. Measurement of the Elastic Properties and Intrinsic Strength of Monolayer Graphene, Science 321 (2008) 385-388.

[80] Y. Gao, P. Hao. Mechanical properties of monolayer graphene under tensile and compressive loading, Physica E: Low-dimensional Systems and Nanostructures 41 (2009) 1561-1566.

[81] A.K. Geim. Graphene: Status and Prospects, Science 324 (2009) 1530-1534.

[82] S. Choudhary, H.P. Mungse, O.P. Khatri. Dispersion of alkylated graphene in organic solvents and its potential for lubrication applications, Journal of Materials Chemistry 22 (2012) 21032-21039.

[83] H. Kinoshita, Y. Nishina, A.A. Alias, M. Fujii. Tribological properties of monolayer graphene oxide sheets as water-based lubricant additives, Carbon 66 (2014) 720-723.

97

[84] W.S. Hummers, R.E. Offeman. Preparation of Graphitic Oxide, Journal of the American Chemical Society 80 (1958) 1339-1339.

[85] L. Tang, X. Li, R. Ji, K.S. Teng, G. Tai, J. Ye, C. Wei, S.P. Lau. Bottom-up synthesis of large-scale graphene oxide nanosheets, Journal of Materials Chemistry 22 (2012) 5676-5683.

[86] Graphene oxide-based drug delivery vehicles: functionalization, characterization, and cytotoxicity evaluation, Journal of Nanoparticle Research 15 (2013).

[87] S.W. Prateek, S.B. Shashwat, J.-B. Archana, R.M. Russel, R.Z. Khushbu, K. Jayant, R.D.K. Misra. Cellular imaging using biocompatible dendrimer-functionalized graphene oxide-based fluorescent probe anchored with magnetic nanoparticles, Nanotechnology 23 (2012) 415101.

[88] L. Feng, Y. Chen, J. Ren, X. Qu. A graphene functionalized electrochemical aptasensor for selective label-free detection of cancer cells, Biomaterials 32 (2011) 2930-2937.

[89] B.D. Malhotra, S. Srivastava, S. Augustine. Biosensors for Food Toxin Detection: Carbon Nanotubes and Graphene, MRS Proceedings 1725 (2015).

[90] N. Laegreid, G.K. Wehner. Sputtering Yields of Metals for Ar+ and Ne+ Ions with Energies from 50 to 600 ev, Journal of Applied Physics 32 (1961) 365-369.

[91] D. Depla, X.Y. Li, S. Mahieu, R.D. Gryse. Determination of the effective electron emission yields of compound materials, Journal of Physics D: Applied Physics 41 (2008) 202003.

[92] P.J. Kelly, A.A. Onifade, Y. Zhou, G.C.B. Clarke, M. Audronis, J.W. Bradley. The Influence of Pulse Frequency and Duty on the Deposition Rate in Pulsed Magnetron Sputtering, Plasma Process. Polym. 4 (2007) 246-252.

[93] Z. Zheng, A.R. Liu, S.M. Wang, B.J. Huang, X.M. Ma, H.X. Zhao, D.P. Li, L.Z. Zhang. In situ fabrication of AgI films on various substrates, Materials Research Bulletin 43 (2008) 2491-2496.

[94] S.K. Bose, S.C. Sircar. Estimation of hole conductivity of AgI through tarnishing studies of silver in iodine vapour at about room temperature, Journal of Materials Science 11 (1976) 129-133.

[95] K.S. Suslick. Sonochemistry, Science 247 (1990) 1439-1445. [96] M. Birkholz. Principles of X-ray Diffraction. Thin Film Analysis by X-Ray

Scattering. Wiley-VCH Verlag GmbH & Co. KGaA, (2006) 1-40. [97] L. Rosenberger, R. Baird, E. McCullen, G. Auner, G. Shreve. XPS analysis

of aluminum nitride films deposited by plasma source molecular beam epitaxy, Surface and Interface Analysis 40 (2008) 1254-1261.

[98] T.J. Sarapatka. Palladium-induced charge transports with palladium/alumina/aluminum interface formation, The Journal of Physical Chemistry 97 (1993) 11274-11277.

[99] S. Hufner. Photoelectron Spectroscopy: Principles and Applications. Third ed., Springer, Berlin, 2003.

[100] E. Lewin, M. Gorgoi, F. Schäfers, S. Svensson, U. Jansson. Influence of sputter damage on the XPS analysis of metastable nanocomposite coatings, Surface and Coatings Technology 204 (2009) 455-462.

[101] A.C. Ferrari, J. Robertson. Interpretation of Raman spectra of disordered and amorphous carbon, Physical Review B 61 (2000) 14095-14107.

98

[102] I. Pócsik, M. Hundhausen, M. Koós, L. Ley. Origin of the D peak in the Raman spectrum of microcrystalline graphite, Journal of Non-Crystalline Solids 227–230, Part 2 (1998) 1083-1086.

[103] A.C. Ferrari. Raman spectroscopy of graphene and graphite: Disorder, electron–phonon coupling, doping and nonadiabatic effects, Solid State Communications 143 (2007) 47-57.

[104] A.C. Ferrari, J.C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K.S. Novoselov, S. Roth, A.K. Geim. Raman spectrum of graphene and graphene layers, Physical Review Letters 97 (2006).

[105] M.-S. Won, O.V. Penkov, D.-E. Kim. Durability and degradation mechanism of graphene coatings deposited on Cu substrates under dry contact sliding, Carbon 54 (2013) 472-481.

[106] P. Ahlberg, F.O.L. Johansson, Z.-B. Zhang, U. Jansson, S.-L. Zhang, A. Lindblad, T. Nyberg. Defect formation in graphene during low-energy ion bombardment, APL Materials 4 (2016) 046104.

[107] W.C. Oliver, G.M. Pharr. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments, Journal of Materials Research 7 (1992) 1564-1583.

[108] Y. Zhu, J. Wang, L. An, H. Wang. Preparation and Study of Nano-Ag/SnO2 Electrical Contact Material Doped with Titanium Element, Rare Metal Materials and Engineering 43 (2014) 1566-1570.

[109] M. Klinger, A. Jager. Crystallographic Tool Box (CrysTBox): automated tools for transmission electron microscopists and crystallographers, Journal of Applied Crystallography 48 (2015) 2012-2018.

[110] D. Ferrah, J. Penuelas, C. Bottela, G. Grenet, A. Ouerghi. X-ray photoelectron spectroscopy (XPS) and diffraction (XPD) study of a few layers of graphene on 6H-SiC(0001), Surface Science 615 (2013) 47-56.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1517

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally throughthe series Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.(Prior to January, 2005, the series was published under thetitle “Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-320235

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2017