UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A Raman and scanning electron microscope analysis of magnetron sputtered thin-film carbon Laumer, Jonathan 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_february_Laumer_Jonathan.pdf [ 3.8MB ]
Metadata
JSON: 24-1.0074399.json
JSON-LD: 24-1.0074399-ld.json
RDF/XML (Pretty): 24-1.0074399-rdf.xml
RDF/JSON: 24-1.0074399-rdf.json
Turtle: 24-1.0074399-turtle.txt
N-Triples: 24-1.0074399-rdf-ntriples.txt
Original Record: 24-1.0074399-source.json
Full Text
24-1.0074399-fulltext.txt
Citation
24-1.0074399.ris

Full Text

A Raman and scanning electronmicroscope analysis of magnetronsputtered thin-film carbonbyJonathan LaumerA THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE COLLEGE OF GRADUATE STUDIES(Electrical Engineering)THE UNIVERSITY OF BRITISH COLUMBIA(Okanagan)December 2014c© Jonathan Laumer, 2014AbstractThin-film carbon coatings possess properties, such as extreme hardness,smoothness, and a nice glossy finish, that make them desirable for a varietyof industrial and military applications. This thesis examines the Ramanspectra associated with thin-films of carbon that are prepared using mag-netron sputtering. The goal is to achieve a high amount of strong bonds,i.e., sp3 bonds, as in diamond, using this inexpensive and widely availabledeposition process. Raman spectroscopy is the chosen analytical methodused for the purpose of this work, since it is non-destructive and widelyavailable. Using Raman spectroscopy, an sp3 content of up to 77 % is deter-mined. This suggests that it is possible to deposit thin-films of carbon thatapproach the properties of tetrahedral amorphous carbon, a material knownfor its excellent hardness and durability, using this inexpensive approach.A scanning electron microscope image of one of the thin-films of carbon isacquired and examined, conclusions regarding the transition between theunderlaying titanium substrate and the thin-film of carbon being drawn.Further directions for possible research are mentioned.iiTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . xiDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiChapter 1: Introduction . . . . . . . . . . . . . . . . . . . . . . . 1Chapter 2: Background . . . . . . . . . . . . . . . . . . . . . . . 122.1 Thin-film carbon and its impact on society . . . . . . . . . . 122.2 What is thin-film carbon . . . . . . . . . . . . . . . . . . . . . 142.3 Common deposition techniques . . . . . . . . . . . . . . . . . 202.4 Applications of thin-film carbon . . . . . . . . . . . . . . . . . 222.5 Raman spectroscopy in the characterization of thin-film carbon 242.6 Problems with the presence of oxygen and other contaminants 35iiiTABLE OF CONTENTSChapter 3: Experiment . . . . . . . . . . . . . . . . . . . . . . . 373.1 Deposition of thin-film carbon . . . . . . . . . . . . . . . . . . 373.2 Magnetron sputtering . . . . . . . . . . . . . . . . . . . . . . 383.2.1 Machine: Hummer XII . . . . . . . . . . . . . . . . . . 383.2.2 Chrome sputtering . . . . . . . . . . . . . . . . . . . . 423.2.3 Sample type and numbering . . . . . . . . . . . . . . . 433.2.4 Substrate preparation . . . . . . . . . . . . . . . . . . 433.2.5 RF cleaning . . . . . . . . . . . . . . . . . . . . . . . . 443.2.6 Recipe: a step-by-step guide . . . . . . . . . . . . . . . 453.2.7 Experiments . . . . . . . . . . . . . . . . . . . . . . . 453.3 Raman Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . 483.3.1 Machine: LabRAM HR . . . . . . . . . . . . . . . . . 483.3.2 Measurements . . . . . . . . . . . . . . . . . . . . . . 513.4 Scanning electron microscope . . . . . . . . . . . . . . . . . . 553.4.1 Machine . . . . . . . . . . . . . . . . . . . . . . . . . . 553.4.2 Determination of the thickness . . . . . . . . . . . . . 553.4.3 Material analysis . . . . . . . . . . . . . . . . . . . . . 57Chapter 4: Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 594.1 Experimental analysis of thin-film carbon . . . . . . . . . . . 594.2 Description and overview of the examined samples . . . . . . 624.3 Raman analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 644.3.1 Unprocessed Raman data . . . . . . . . . . . . . . . . 644.3.2 Baseline correction . . . . . . . . . . . . . . . . . . . . 664.3.3 Peak fitting . . . . . . . . . . . . . . . . . . . . . . . . 68ivTABLE OF CONTENTS4.3.4 Background subtraction . . . . . . . . . . . . . . . . . 724.4 Summary of the Raman results . . . . . . . . . . . . . . . . . 754.5 Raman interpretation . . . . . . . . . . . . . . . . . . . . . . 964.6 Film thickness and material composition profile . . . . . . . . 98Chapter 5: Conclusions . . . . . . . . . . . . . . . . . . . . . . . 108References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111Appendix A: Step by step guide for the Hummer XII . . . . . 117vList of TablesTable 1.1 A comparison of the properties of diamond and graphite. 5Table 1.2 A comparison of the different forms of amorphous andcrystalline carbon. . . . . . . . . . . . . . . . . . . . . 7Table 3.1 An overview of the sputtering experiments correspond-ing to the different sample numbers. . . . . . . . . . . 47Table 3.2 Pictures of the samples prepared and analyzed in thisthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Table 3.3 Overview of the Raman measurements and measure-ment settings. . . . . . . . . . . . . . . . . . . . . . . . 54Table 4.1 Overview of the samples with the deposition parame-ters and the Raman source identified. . . . . . . . . . . 63Table 4.2 Overview of the Raman results corresponding to thedifferent samples. . . . . . . . . . . . . . . . . . . . . . 76Table 4.3 Overview of the Raman results corresponding to thedifferent samples. . . . . . . . . . . . . . . . . . . . . . 87Table 4.4 A comparison of common Raman interpretations foundin the literature. . . . . . . . . . . . . . . . . . . . . . 99viList of FiguresFigure 1.1 The atomic distribution within diamond. . . . . . . . 2Figure 1.2 Atomic distribution within graphite and the natureof bondings. . . . . . . . . . . . . . . . . . . . . . . . 3Figure 1.3 The relationship between the sp3 content and thehardness . . . . . . . . . . . . . . . . . . . . . . . . . 11Figure 2.1 The hybridizations available to carbon atoms. . . . . 15Figure 2.2 Thin-film deposition, PVD vs. CVD. . . . . . . . . . 18Figure 2.3 Ternary phase diagram for the different kinds of thin-film carbon. This figure is from Robertson et al. [9]. . 19Figure 2.4 Scattering effects . . . . . . . . . . . . . . . . . . . . . 25Figure 2.5 General schematic of a Raman spectroscopy system . 27Figure 2.6 Raman scattering . . . . . . . . . . . . . . . . . . . . 28Figure 2.7 Different vibration modes . . . . . . . . . . . . . . . . 32Figure 2.8 The Raman interpretation of a thin-film carbon byZhao et al. [10]. . . . . . . . . . . . . . . . . . . . . . 33Figure 2.9 Raman interpretation of a thin-film carbon by Sunget al. [31]. . . . . . . . . . . . . . . . . . . . . . . . . 34Figure 2.10 Adhesion strength for different materials . . . . . . . 36viiLIST OF FIGURESFigure 3.1 Sputtering machine Hummber XII. . . . . . . . . . . 39Figure 3.2 Terminology of the sputtering vacuum chamber. . . . 41Figure 3.3 Water drop test for a titanium ring before and afterargon plasma cleaning. . . . . . . . . . . . . . . . . . 46Figure 3.4 Front side of the LabRAM HR Raman confocal mi-croscope. . . . . . . . . . . . . . . . . . . . . . . . . . 50Figure 3.5 Planar sample placement for Raman spectroscopy. . . 52Figure 3.6 Ring sample placement for Raman spectroscopy. . . . 53Figure 3.7 The SEM machine at SEMLab/UBC Okanagan . . . 56Figure 3.8 First thickness measurement of sample number 115-1. 58Figure 4.1 Unprocessed Raman spectrum of sample number 201-2. 65Figure 4.2 Unprocessed Raman spectrum of sample number 201-2 with the baseline. . . . . . . . . . . . . . . . . . . . 67Figure 4.3 Baseline corrected Raman spectrum of sample num-ber 201-2. . . . . . . . . . . . . . . . . . . . . . . . . . 69Figure 4.4 Raman interpretation of sample number 201-2 with aRaman source of 442 nm. . . . . . . . . . . . . . . . . 71Figure 4.5 Baseline corrected Raman spectrum of sample num-ber 207-1 without background compensation. . . . . . 73Figure 4.6 Raman spectrum of plain glass and flat titanium. . . 74Figure 4.7 Raman interpretation of sample number 106-2. . . . . 77Figure 4.8 Raman interpretation of sample number 111-1. . . . . 78Figure 4.9 Raman interpretation of sample number 115-3. . . . . 79Figure 4.10 Raman interpretation of sample number 121-3. . . . . 80viiiLIST OF FIGURESFigure 4.11 Raman interpretation of sample number 123-3. . . . . 81Figure 4.12 Raman interpretation of sample number 201-2 with aRaman source of 633 nm. . . . . . . . . . . . . . . . . 82Figure 4.13 Raman interpretation of sample number 204-2. . . . . 83Figure 4.14 Raman interpretation of sample number 207-1 withbackground compensation. . . . . . . . . . . . . . . . 84Figure 4.15 Raman interpretation of sample number 210-1. . . . . 85Figure 4.16 Raman interpretation of sample number 210-2. . . . . 86Figure 4.17 Comparison of Raman results with different argon gasflow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89Figure 4.18 Comparison of Raman results with glass vs. titaniumring substrate. . . . . . . . . . . . . . . . . . . . . . . 91Figure 4.19 Comparison of Raman results with glass vs. flat tita-nium substrate. . . . . . . . . . . . . . . . . . . . . . 92Figure 4.20 Comparison of Raman results to test for reproducibility. 93Figure 4.21 Comparison of Raman results with different Ramansource wavelengths. . . . . . . . . . . . . . . . . . . . 94Figure 4.22 Comparison of Raman results with and without back-ground compensation. . . . . . . . . . . . . . . . . . . 95Figure 4.23 Comparison of Raman results with DC vs. RF power. 97Figure 4.24 SEM picture of the thin-film carbon of sample number115-1. . . . . . . . . . . . . . . . . . . . . . . . . . . . 101Figure 4.25 Material line analysis of sample number 115-1. . . . . 102Figure 4.26 Thickness of sample number 115-1. . . . . . . . . . . 104ixLIST OF FIGURESFigure 4.27 Material spectrum number 14 corresponding to sam-ple number 115-1. . . . . . . . . . . . . . . . . . . . . 105Figure 4.28 Material spectrum number 16 corresponding to sam-ple number 115-1. . . . . . . . . . . . . . . . . . . . . 106Figure A.1 Step-by-step guide for thin-film carbon deposition withthe Hummer XII. . . . . . . . . . . . . . . . . . . . . 118xAcknowledgementsSpecial thanks goes to my supervisor, Dr. Stephen K. O’Leary. With-out his guidance and help, this research would not have occured. He isnot only a great researcher, but also a great mentor, helping me along myacademic path. I also would like to thank Dr. Guangrui Xia and her grad-uate students, Mr. Xiyue Li and Ms. Ye Zhu, both also being from TheUniversity of British Columbia Vancouver, for letting me use their Ramanspectroscopy machine and any help relating those measurements. My grati-tude also goes to Mr. David Arkinstall from the SEMLab at The Universityof British Columbia Okanagan for giving me the opportunity to do mea-surements with the scanning electron microscope. Thanks to the Charles E.Fipke Foundation of The University of British Columbia Okanagan, I wasalso able to acquire the use of the scanning electron microscope. Specialthanks also goes to Mr. Ben Vos, the general manager of Arnell WorkshopInc. and Mr. Roy Arnell, for supporting me in the use of their magnetronsputtering machine, this being the machine used in these depositions.xiDedicationI would like to dedicate this thesis to my brother Simon. Since my veryearliest childhood, he has been a true friend.xiiChapter 1IntroductionSince the dawn of human civilization, diamonds have captured the popu-lar imagination like no other material. Diamonds are exceptionally beautifuland possess material properties that surpass those associated with all othermaterials. The shiny look and bright appearance of diamonds makes themhighly coveted for declarations of love, i.e., as in marriage, but also in theacquisition of might, i.e., as in the crown of a king. In addition to their clearand marvellous look, diamonds possess a hardness that is second to none.Unfortunately, these positive attributes of diamond have fuelled greed, andthis has spawned human conflict. Indeed, wars have been fought over dia-monds, and countries in possession of diamonds have found them to be amixed blessing.Given diamond’s hold on human curiosity, it has been the subject of aconsiderable amount of scientific scrutiny. Initial studies into the materialproperties of diamond have focused on the distribution of atoms withinthis material. These studies demonstrated that diamond is a pure form ofcarbon that is arranged in a tetrahedral crystalline structure, as shown inFigure 1.1 [1]. Alternatively, researchers also have found that carbon cancrystallize in a trigonal form, known as graphite, this crystalline structurebeing depicted in Figure 1.2 [2]. The properties of graphite are quite distinct1Chapter 1. IntroductionFigure 1.1: A representation of the distribution of atoms within diamond.The carbon atoms are represented with the spheres, while the bonds are de-picted with the spokes. This figure is after Karamitaheri [1]. The electronicversion of this figure is in color.2Chapter 1. IntroductionFigure 1.2: A representation of the distribution of atoms within graphite.The carbon atoms are represented with spheres, while the bonds are depictedwith spokes. This figure is after Seo [2]. The electronic version of this figureis in color.3Chapter 1. Introductionfrom those associated with diamond [2]; see Table 1.1 [3]. Diamond possessesstrong tetrahedral bonds, i.e., sp3 bonds, for all axes, while graphite bondsare weak along one axis, i.e., with sp2 bonds, but also strong in the directionof the two other axes. This difference in the nature of the bonds accountsfor the differences in its material properties.In addition to these two crystalline forms of carbon, there are many othernon-crystalline forms of carbon that are available, with properties that arequite distinct from those associated with diamond and graphite. Carboncomes in both polycrystalline and amorphous forms. Further enriching thesituation, in some materials, the nature of the bonds is mixed, i.e., bothsp2 and sp3 bonds are present. Carbon may also appear in powdered form,as soot, or in glassy form. A more recent addition to the constellation ofcarbon-based materials is thin-film carbon, deposited from the vapor phase,allowing for carbon coatings on a variety of substrates. Its introduction hassubstantially increased the breadth of the possible applications for carbon.As carbon can bond in so many different forms, it is referred to as beingallotropic in character.Of course, these properties of diamonds are also desirable for a wide arrayof industrial and military applications, where resistance to wear-and-tear iscritical. The durability of tools and parts can be dramatically enhancedif diamonds are involved. As naturally occurring diamonds are rather rareand expensive, an interest in developing synthetic forms of diamond hasthus developed. Research aimed at producing synthetic diamonds foundits genesis in 1879 with the work of Hannay [4], in which hard crystals ofcarbon were formed through the heating of charcoal in a furnace. Today,4Chapter 1. IntroductionTable 1.1: A comparison of the properties of diamond with graphite. Forgraphite, the in-plane properties are mentioned first, the transverse proper-ties second. This table is from Saada [3].Property Graphite DiamondLattice constant (300 K) [A˚] 2.462, 6.708 3.567Bond length (300 K) [A˚] 1.421 1.545Atomic density [cm−3] 1.14 × 1023 1.77 × 1023Thermal conductivity [W/cm·K] 30, 0.06 25Debye temperature [K] 2500, 950 1860Electron mobility [cm2/V·sec] 20 ×103, 100 1800Hole mobility [cm2/V·sec] 15 ×103, 90 1500Melting point [K] 4200 4500Band gap [eV] -0.04 5.475Chapter 1. Introductionsynthetic diamonds are chiefly produced through the use of high pressuresand high temperatures [5]. The resultant materials have been noted for theirhardness and clarity. Other forms of synthetic diamonds, that are availablein nano-crystalline form, may be produced through the use of explosives [6].Synthetic diamonds, while in possession of many desirable material prop-erties, are expensive to fabricate, albeit this expense is less than that re-quired to purchase their naturally occurring counterparts. This has moti-vated researchers to consider the growth of thin-film carbon from the vaporphase. Thin-films of carbon, grown from the vapor phase, were first pre-pared using ion-beam deposition in 1971 by Aisenberg and Chabot, of theSpace Science Division of the Whittaker Corporation in Waltham, Mas-sachusetts [7]. Since that time, a variety of different deposition techniques,ranging from simple methods, such as sputtering, to more sophisticated ap-proaches, such as plasma enhanced chemical vapor deposition (PECVD),have been employed in order to deposit thin-film carbon [8]. An overviewof some of the basic forms of thin-film carbon, their basic properties, andtheir possible applications, are summarized in Table 1.2 [9].Thin-films of carbon, prepared using the different available approaches,possess many different qualities, in terms of their adhesion to the under-lying substrate, their smoothness, hardness, and color. Depending uponwhich application is being considered, the right deposition technique anddeposition conditions must be selected. In this thesis, thin-films of carbon,with cosmetic jewelry applications in mind, will be fabricated and examined.The substrates primarily being considered are made of titanium, and are inthe form of rings. Unfortunately, however, the ultimate question that is the6Chapter 1. IntroductionTable 1.2: A comparison of the different forms of amorphous and crystallinecarbon. Diamond and graphite have a crystalline structure, and are consid-ered as reference materials. Their properties are compared with some formsof thin-film carbon. This table is after Robertson et al. [9].Sample sp3 (%) H (%) Density (g cm−1) Hardness (GPa) Applications/PropertiesDiamond 100 0 3.515 100 cutting, jewelryGraphite 0 0 2.267 batteries, lubricantsEvaporated C 0 0 1.9 3Sputtered C 5 0 2.2 thin-film coatingsta-C 80-88 0 3.1 80 drilling and millinga-C:H hard 40 30-40 1.6-2.2 10-20a-C:H soft 60 40-50 1.2-1.6 less than 10ta-C:H 70 30 2.4 50 low friction7Chapter 1. Introductionfocus of investigation is whether or not an inexpensive deposition technique,i.e., magnetron sputtering, is adequate for this particular application. Un-fortunately, owing to limitations in the amount of time available, this thesispresents intermediary results that are required in order to address this over-arching question. The ultimate resolution of this matter lies beyond thescope of this particular body of work and will have to be pursued in thefuture.Even though the processes whereby thin-films of carbon may be de-posited have been studied in detail, and the growing process of such filmsare now reasonably well understood, a systematic study on the influence ofthe substrate type and shape on the resultant thin-film carbon properties,and on how the deposition parameters also influence these properties, hasyet to be performed. While there have been some studies performed on theeffect of the substrate on the thin-film carbon properties, the focus of thesestudies was mostly on the impact of the substrate type and orientation.Zhao et al. [10], for example, examined the impact of depositing thin-filmcarbon on a variety of flat metal substrates. Bobzin et al. [11] depositedthin-film carbon using substrates tilted at an angle. Most studies have beenperformed on planar glass or silicon samples [12–15]. A critical comparison,in which the properties of thin-films of carbon deposited on flat substratesare contrasted with those associated with thin films of carbon deposited oncurved substrates, has yet to be performed. This is a topic that will bestudied within the context of this particular thesis.Titanium rings, such as the ones serving as the substrate for the pur-poses of the application being considered in this analysis, combine both an8Chapter 1. Introductionuncommon material and an unusual shape. For cosmetic jewelry applica-tions, the challenges are different from those encountered in many otherindustrial or military applications. The primary focus of attention lies onthe appearance and the comfort in wearing, combined with such considera-tions as long-lasting quality, as it is a luxury good. Unprocessed titaniumsubstrates are in possession of a native oxide layer, and this makes adhesiona challenge, typically militating the use of an interlayer [16]. This work willbe a first analysis of the influence of the substrate type, the substrate geom-etry, and the deposition parameters of magnetron sputtering on the sp3 tosp2 ratio of the thin-films of carbon, this ratio being a well accepted measureof thin-film carbon’s quality. This work gives insight into the dependence ofthe substrate type, shape, and the deposition parameters on the thin-filmcarbon properties. Understanding the exact influence of those parameterswill allow for the proper growth of high quality thin-films of carbon. Theproperties of thin-film carbon, like its high strength and durability, due toextreme hardness, combined with its shiny and glossy look, make them adesired feature in cosmetic jewelry. Jewelry rings, with a black thin-filmcarbon coatings, are in high demand in our society. This scientific study hasbeen accomplished in collaboration with a local company, Arnell WorkshopInc., with the eventual goal of determining the feasibility of growing highquality thin-films of carbon on titanium jewelry rings through the use of amagnetron sputtering machine.This thesis aims to accomplish five distinct goals, that will eventually beused to certify the applicability of magnetron sputtering for the purposesof preparing thin-films of carbon for cosmetic jewelry applications. First,9Chapter 1. Introductionthe production of thin-films of carbon, using this technique, will be demon-strated; given the state of the magnetron sputtering system used prior tothis work being performed, this was by no means guaranteed prior to thiswork being performed. Second, these films that arise from these depositionswill be characterized using a Raman analysis, it being demonstrated thatthe resultant sp3 to sp2 ratio is similar to that associated with other formsof thin-film carbon; the sp3 to sp2 ratio is an often referred metric of thin-film carbon quality; see Figure 1.3.1 Third, through the use of a varietyof different deposition conditions, the deposition parameter space will besampled, the resultant forms of thin-film carbon being examined throughthe use of Raman spectroscopy in order to evaluate the corresponding sp3to sp2 ratio [18]. Finally, a means of determining thin-film carbon profiles,through the use of a scanning electron microscope (SEM) analysis, will bedeveloped. Pathways, wherein these intermediary results may ultimately beused in order to assess the adequacy of magnetron sputtering for cosmeticjewelry applications, will then be defined, albeit this work will have to beperformed following the completion of this particular thesis.This thesis is organized in the following manner. In Chapter 2, the back-ground for this research is provided. Then, in Chapter 3, the experimentalmethodology employed is presented. The results obtained from these exper-iments are then analyzed in Chapter 4. Finally, conclusions are drawn andsome suggestions for future investigation are provided.1This figure shows the dependence of the hardness on the sp3 bonding content, theseresults being from Neuville et al. [17]. Most parameters within thin-film carbon have beencharacterized in terms of the sp3 bonding content.10Chapter 1. IntroductionFigure 1.3: Relationship between the amount of diamond-like chemicalbonds (sp3) in the thin-film carbon and the hardness. This figure is afterNeuville et al. [17]. The electronic version of this figure is in color.11Chapter 2Background2.1 Thin-film carbon and its impact on societyCarbon’s unusual range of physical properties stems, in large measure,from the diversity of its constituent bonds. Typically, within carbon, sp1,sp2, and sp3 bonds are present. Carbon that is purely comprised of sp2bonds in the trigonal crystalline structure, appears black, and is soft to thetouch, this form of carbon being referred to as graphite. In contrast, carbonthat is purely comprised of sp3 bonds and is in the tetrahedral crystallinestructure, is transparent and exhibits a hardness unrivalled elsewhere innature, this form of carbon being referred to as diamond. Forms of carbonthat possess both sp2 and sp3 bonds exhibit material properties that arein between these two extreme cases; sp1 bonds do occur, but are typicallyavailable in much smaller quantities than the sp2 and sp3 bonds.In general, thin-films of carbon possess both sp2 and sp3 bonds. It hasbeen found that the ratio of sp3 bonds to sp2 bonds determines many ofthe properties of the resultant thin-films. A high sp3 bond content is foundto provide these coatings with many diamond-like properties. Similarly, ahigh sp2 bond content is found to provide thin-film carbon with graphite-likecharacteristics. Thin-films of carbon are currently being deployed in a wide122.1. Thin-film carbon and its impact on societyvariety of industrial and military settings, including those in the automotive,milling, and drilling industries, in optics, and even in biomechanics [19–23].For such applications, thin-films of carbon with an abundance of sp3 bondsare preferable.In this chapter, the background material required for this study is pre-sented. Initially, the peculiar properties of thin-film carbon are accountedfor in terms of the nature of the chemical bonds that are present within thismaterial. Following an overview of the most common applications for thin-films of carbon, the societal importance of thin-film carbon is presented.Raman spectroscopy, a non-destructive tool that is to be used for the pur-poses of this analysis, is then discussed, its use in identifying the nature ofthe bonds that are present within thin-film carbon being clarified. Chal-lenges, related to the presence of oxygen and contaminants on the surfaceof the substrate, thereby influencing the adhesion between the thin-film andthe underlying substrate, are also presented.This chapter is organized in the following manner. Section 2.2 explainswhat thin-film carbon is and discusses the nature of its chemical bonds. Themost common deposition techniques, currently employed for the preparationof thin-film carbon, are then introduced in Section 2.3. Potential applica-tions for thin-films of carbon are then presented in Section 2.4. Section 2.5introduces the reader to the scientific background underlying Raman spec-troscopy. Finally, in Section 2.6, the challenges associated with the presenceof oxygen, and other contaminants, on underlying titanium substrates, isdiscussed.132.2. What is thin-film carbon2.2 What is thin-film carbonCarbon is an element with an allotropic character. It can exist in theform of graphite, where only sp2 bonds are present, and in the form ofdiamond, where only sp3 bonds are present, or somewhere in between. Itis noted that while graphite and diamond are both forms of carbon, theypossess totally different physical properties. Materials with both types ofbonds, i.e., sp2 and sp3 bonds, are found to have properties that are inbetween these extreme forms of carbon. It is natural to inquire as to whataccounts for the differences in the observed properties. It may seem odd thatthe same element, carbon, can produce so many different types of matter,with such disparate physical properties.The cause of carbon’s allotropic character lies in the diversity of thehybridization states that are available to it, i.e., sp1, sp2 and sp3 hybridiza-tions are available. Carbon based atomic orbitals, associated with thin-filmcarbon, can form new hybrid orbitals, thereby allowing for totally differenttypes of chemical bonding, and therefore, different types of material prop-erties. In diamond, the atomic bonding of the constitutive carbon atoms isentirely sp3 in character. Carbon has four valence electrons, and in the caseof pure diamond, each of them is assigned to an sp3 orbital, σ bonds beingassociated with each such orbital; see, for example, Figure 2.1 [9]. As theσ bond is one of the strongest chemical bonds known in nature, it equipsdiamond with its unique hardness.If the carbon atoms bond in the form of sp2 bonds, however, just threeout of the four valence electrons form σ bonds, the fourth valence electron142.2. What is thin-film carbonFigure 2.1: The hybridizations available to carbon atoms in the form of sp1,sp2, and sp3 bonds. This figure is after Robertson et al. [9]. The electronicversion of this figure is in color.152.2. What is thin-film carbonforming a ppi orbital with a normal orientation to the other σ bonds; thebond associated with this ppi orbital is referred as a pi bond. pi bonds areweaker than σ bonds, the bonding being weaker on the axis that includesthis ppi orbital. Graphite, however, is comprised only of sp2 bonds. In thesp1 configuration, however, the carbon atom forms two pi bonds in the yand z axes. Just as in the x axis, σ bonds are formed with the two valenceelectrons. The ratio of the number of sp3 bonds to the number of sp2 bondsis a common parameter describing the bonding character of thin-films ofcarbon, and many of its material properties may be directly related to thisratio.There are two commonly employed thin-film carbon deposition tech-niques that are widely used in order to fabricate thin-films of carbon: (1)chemical vapor deposition (CVD), and (2) physical vapor deposition (PVD).The most commonly used approach, the CVD process, is most commonlyemployed using the plasma enhanced form of this process, i.e., through theuse of PECVD. In PECVD, a voltage is applied across an inert gas in or-der to create a plasma, thereby enabling the deposition of thin-film carbon.This process allows for the gas molecules within the plasma chamber to becracked, the cracked gas molecules then depositing onto a heated substrate,thereby allowing for the growth of a thin-film of carbon. Given that it is aCVD process, the resultant thin-film coating is relatively uniform. In PVDtechniques, however, such as sputtering, an inert gas is ionized in order tocreate a plasma. The source material, carbon for example, is then bom-barded with these ions, and as a result, a stream of carbon atoms from thissource are blasted into the vacuum. These carbon atoms will then deposit162.2. What is thin-film carbononto the substrate, which has to be in a line-of-sight.In PECVD, other gases are involved in the growth of the thin-film car-bon. Therefore, in most forms of thin-film carbon, there is a fair amountof hydrogen incorporated into the film. Another characteristic of PVD pro-cesses is the shadowing effect that arises, that does not exist in CVD pro-cesses. Figure 2.2 [24] illustrates this shadowing effect. In sputtering, forexample, the source material comes from one direction only, whereas inPECVD, the gaseous source material interacts everywhere on the growingthin-film surface. While PECVD and sputtering are the dominant deposi-tion techniques used in research and industry in order to deposit thin-filmcarbon, other deposition techniques are also available for the preparation ofthis material. For example, ion beam, cathodic arc, and pulsed laser deposi-tion process have also been used in order to deposit thin-film carbon. Thesetechniques are described elsewhere in the literature [7, 9, 25].Figure 2.3 [9] shows a ternary phase diagram by Robertson et al. [9],this diagram providing a good overview of the different forms of thin-films ofcarbon that have been produced. Amorphous carbon structures with a highsp2 content are presented in the lower left hand corner of this diagram andinclude materials such as soot, glassy, and evaporated carbon. Materialswith a high sp3 content are referred to as ta-C, and placed in the uppercorner of this ternary phase diagram. Various forms of thin-film carbon, withdifferent amounts of hydrogen, can also be produced, as depicted in the lowerright hand corner. Thin-film carbon, with a fair amount of hydrogen, are,for example, hydrocarbon polymers, polyacetylene (CH), and polyethylene(CH2). Thin-film carbon, with a high sp3 content, is usually preferred due172.2. What is thin-film carbonFilm%Film%Substrate% Substrate%0.5%μm% 0.5%μm%a)% b)%Figure 2.2: Side view of a thin-film deposition in a groove. a) CVD processesare without a shadowing effect. b) PVD processes are with a shadowingeffect. This figure is after Street [24]. The electronic version of this figure isin color.182.2. What is thin-film carbonFigure 2.3: Ternary phase diagram for the different kinds of thin-film car-bon. The approximate locations for the samples considered in this workis indicated in this figure. This figure is after Robertson et al. [9]. Theelectronic version of this figure is in color.192.3. Common deposition techniquesto its diamond-like properties, such as durability and hardness. In addition,thin-films of carbon with a high sp3 content are chemically inert and fullybiocompatible. Since the atomic distribution within most forms of thin-filmcarbon is amorphous in nature, the color of the resultant material appearsblack, but with a shiny and glossy look. Its lustrous appearance makes thiscoating very attractive for cosmetic jewelry application purposes.In this thesis, a magnetron sputtering process is used in order to growa particular form of thin-film carbon. Sputtered thin-film carbon is usuallyreferred to as amorphous carbon (a-C), and most often possesses an amor-phous distribution of atoms with no grain boundaries. If the sp3 contentreaches a fairly high amount, i.e., in excess of 80% of the overall bonds, thenthe carbon is referred to as ta-C. This material, i.e., ta-C, is considered tobe one of the hardest materials, after pure diamond, and this property isbelieved to arise from its high content of sp3 bonds [9, 17].2.3 Common deposition techniquesSputtering is a commonly available PVD technique, used for a varietyof industrial application owing to its adaptability, the possibility to sputtermany different materials, and the option to easily scale up the depositionsystem [9]. An overview of the terminology of the vacuum chamber, regard-ing this deposition technique, is given in Section 3.2.1. Sputtering is a PVDprocess. For the case of thin-film carbon deposition, the source materialis a solid block of graphite that is bombarded by ionized inert gases. Theloosened graphite material then deposits onto the substrate. The deposition202.3. Common deposition techniquesoccurs in a line-of-sight, and the resultant deposition is not necessarily uni-form in thickness. In contrast, in CVD, the source material is introduced asa gas, and it reacts with the substrate surface at an elevated temperature,i.e., around 1000 ◦C. With sputtering, on the other hand, the process takesplace almost at room temperature, which allows for film depositions on plas-tic substrates, this being a considerable advantage for certain applications.For the special case of magnetron sputtering, the electrons in the plasmaare trapped close to the target through the use of magnets, which are placedbehind the sputtering source so as to increase the formation of argon ions,and hence, the bombardment rate of the target. The sputtering process isrelatively easy to control through the application of power and through theamount of gas introduced into the deposition chamber [9].PECVD is another very common deposition technique. In this approach,a plasma assists the deposition process, thereby allowing for the use of alower processing temperature; for the case of thin-film carbon depositions,temperatures as low as 700 K can be maintained. The difference with sput-tering is that the source material may be introduced in a gas instead of as asolid block. A challenge with gas-based deposition systems is that the depo-sition chambers themselves, employed for such depositions, possess a mem-ory corresponding to past depositions, i.e., trace elements, from previousdepositions, incorporate themselves into present depositions, thereby chang-ing the properties of the resultant thin-films. The advantage of PECVD isits ability to deposit relatively uniform thin-film layers, not just in a line-of-sight manner. Unfortunately, in order to get a uniform thin-film carbonlayer, the gas is not pure carbon, but part of a larger gaseous hydrocarbon,212.4. Applications of thin-film carbonsuch as methane (CH4). Thus, thin-films of carbon, grown with a PECVDsystem, possess a fair amount of hydrogen, and this undoubtedly influencesthe properties of the resultant thin-films.2.4 Applications of thin-film carbonAs described previously, thin-films of carbon, with a high sp3 content,possess exceptional qualities, such as high durability, hardness, and bio-compatibility. The smooth coating that results from such depositions alsopossess a very low coefficient of friction and, therefore, thin-films of carbonare often used for coating tools and machine parts in order to increase theirlifetime. Aside from their mechanical advantages, thin-film carbon coatingscan also be used for decorative purposes. The nice shiny black coating as-sociated with such films has a very attractive look and feel. This is whatmakes them particularly appealing for the cosmetic jewelry industry. Thisthesis investigates the influence of substrate geometry, i.e., flat titaniumas opposed to titanium rings, on the resultant thin-film carbon properties.Cosmetic jewelry rings benefit from all of the properties of a thin-film car-bon coating, and therefore, it is crucial to achieve the same quality of filmsdeposited on rings as on flat substrates.High performance applications in the automotive industry are inconceiv-able without the use of thin-film coatings. With such coatings, the pistonpins can withstand much more pressure over longer periods of time. Thus,faster and stronger cars can be produced using this technology. The Dodge2015 Challenger SRT is a good example of an automotive application for222.4. Applications of thin-film carbondiamond-like-carbon coatings. Such coatings have played a decisive rolein achieving the gains in performance that have been attained in this newmodel when contrasted with that associated with previous models.Another industry that has benefitted from the increased hardness anddurability associated with thin-film carbon is the milling and drilling indus-tries [19]. Besides a significant increase in the hardness of cutting tools, andtherefore, a corresponding enhancement in their durability, cutting toolscoated with a thin-film carbon layer have been found to reduce the sur-face roughness of the cut material [20]. This realization has substantiallycontributed to the success of these industries.Carbon can be considered to be the basis of life, as it exists abundantlyon Earth and bonds readily with other elements, such as oxygen, hydro-gen, and nitrogen, to form long and complex molecules. Carbon is thereforecompatible with living tissues. Since thin-films of carbon are biocompatible,they may also be used for implants, such as hip joint replacements. Movingparts in the body require a smooth and almost frictionless surface withoutany residual metallic wear debris. Allen et al. [23] investigated the biocom-patibility of thin-film carbon coated parts, both in vitro and in vivo. Thesetests have demonstrated that thin-films of carbon can coexist with livingtissues or organisms without causing harm. Thus, in terms of biomedicalengineering applications, thin-film carbon has a virtually unlimited poten-tial.Thin-films of carbon are also commonly applied in optics. The sensitiv-ity of long-period fibre gratings to variations in the external refractive index,for example, can be improved through the use of a thin-film of carbon [21].232.5. Raman spectroscopy in the characterization of thin-film carbonThe advantage of thin-film carbon over other types of thin-films in opticalapplications, is the enhanced resistance to mechanical impact and corrosion.For front looking infrared cameras, which are also used in military applica-tions, a thin-film carbon coating on the lens assembly has found to reducethe Narcissus effect, which is a reflection of the detector onto itself from thefront lens [22].2.5 Raman spectroscopy in the characterizationof thin-film carbonThe Raman effect was first discovered by Chandrasekhara Venkata Ra-man (1888 - 1970), who was born in southern India. On a sea voyage, goingback home to India from England, he was able to experimentally demon-strate that the blue color of the sea is not due to a reflection from the bluesky. To do this, Raman showed, through the use of a diffraction grating, thatthe maximum spectral intensity for the sea and the sky occur at differentwavelengths, and therefore, another explanation for the blue color of the seamust exist. He later discovered a frequency shift in the scattered light. Thischange in the frequency, and therefore, in the wavelength of the incidentlight due to inelastic scattering by water molecules, is the actual cause forthe blue color of the sea. The color of the sky, however, is due to Rayleighscattering, which is an elastic scattering process. Since the Rayleigh scatter-ing intensity is proportional to 1/λ4, blue light gets Rayleigh scattered themost owing to its short wavelength in the visible spectrum. The differentscattering processes are summarized in Figure 2.4 [18].242.5. Raman spectroscopy in the characterization of thin-film carbonGround'state'1st'excited'vibra2onal'state'Virtual'energy'level'Energy'Rayleigh'sca:ering'Stokes'sca:ering'an2=Stokes'sca:ering'ΔEV'='h!νV'Figure 2.4: Light scattering can be either elastic (Rayleigh scattering) or in-elastic (Stokes or anti-Stokes Raman scattering). This figure is after Bartonet al. [18]. The electronic version of this figure is in color.252.5. Raman spectroscopy in the characterization of thin-film carbonA Raman spectroscopy system consists of four basic components: (1) aRaman source, or excitation source with an incident laser beam, (2) someoptical elements in order to focus this laser beam onto the sample, (3) acollection of optical elements and filters, and finally, (4) a detector, i.e.,spectrograph, and a camera in order to collect the scattered light. Fig-ure 2.5 provides a schematic overview of a Raman spectroscopy system. Inorder to separate Raman scattering from Rayleigh scattering, a dichroicbeamsplitter, i.e., a band-pass filter, is used. Notch filters confine the laserwithin the Raman spectrum, and reduce, in this way, the strong Rayleighscattering of light.In thin-film carbon, the light associated with a laser beam gets scat-tered differently for the different hybridization states. The change in fre-quency/wavelength is usually denoted as the Raman shift (∆k in cm−1),which is the difference between the reciprocal of the excitation wavelength,λ0, from the reciprocal of the scattered wavelength, λ1, i.e.,∆k(cm−1) =(1λ0(nm)−1λ1(nm))×(107 nm)(cm). (2.1)The change in frequency is caused either by the emission or the absorption ofenergy, which results in a lower and higher vibrational mode of the molecules,respectively; see Figure 2.4. The energy emission or absorption, ∆E, isrelated to the frequency shift, ∆ω, according to∆ERaman = ~ ·∆ωRaman. (2.2)In the case of a phonon emission, the frequency shift (downshift) is termeda Stokes shift. Instead, the absorption of a phonon causes an upshift of thefrequency, and is denoted as an anti-Stokes shift; see Figure 2.6 [26]. The262.5. Raman spectroscopy in the characterization of thin-film carbonTo#Specrograph# Dichroic#Beamspli2er#Notch#Filter#Line#Filter#From#Laser#Lens#Cuve2e#Figure 2.5: General schematic overview of a Raman spectroscopy system.This figure is after Barton et al. [18]. The electronic version of this figureis in color.272.5. Raman spectroscopy in the characterization of thin-film carbonω,#k#ω',#k’#Ω,#K#ω#=#ω’#±#Ω#k#=#k’#±#K#Figure 2.6: Raman scattering: + sign for phonon emission (Stokes process)/ - sign for phonon absorption (anti-Stokes process). This figure is afterafter Kittel [26]. The electronic version of this figure is in color.282.5. Raman spectroscopy in the characterization of thin-film carbonintensity of the scattering is much smaller for the anti-Stokes shift, as thepopulation of molecules with a higher initial energy ground state is muchlower due to the Boltzmann distribution, i.e.,N ∝ exp(−EkB T). (2.3)Today, Raman spectroscopy is the most commonly used method to deter-mine the chemical nature of thin-films of carbon, and the Stokes shift is theprimary focus of analysis [10, 12, 15, 26–31]. The major advantage of Ra-man spectroscopy is its non-destructive character, and the wide availabilityof such equipment.For the specific case of thin-film carbon, Raman spectra, between 800to 2000 cm−1, are the most commonly considered; beyond this range, theRaman spectra associated with thin-film carbon is typically flat and fea-tureless. The Raman spectrum associated with a sample of thin-film carbontends to exhibit two distinct peaks, one being around 1350 cm−1, the otherbeing around 1580 cm−1. These two peaks are referred as the D and Gpeaks, respectively. The G peak, where G denotes graphite, is believed tobe related to the sp2 stretch vibrations of benzene rings, or other such molec-ular chains [15, 32]. The D peak, where D denotes disordered, is associatedwith the presence of sp3 bonding. The scattering related to the presenceof sp2 bonds causes a frequency shift, or a Raman shift, centered around1580 cm−1, and the scattering related to the presence of sp3 bonds causes aRaman shift, centered around 1350 cm−1. These two peaks in the Ramanspectra are often used in order to provide an indication as to the sp3 to sp2bonding ratio. A controversy exists in the literature as to whether it is the292.5. Raman spectroscopy in the characterization of thin-film carbonratio of the peaks or the integrated intensities of the Raman spectrum thatdetermines the sp3 to sp2 bonding ratio. For the purposes of this particularanalysis, the ratio of the integrated intensities is employed, this being themore common approach.Schwan et al. [15] suggest that the D peak in the Raman spectrum isdue to disordered structures of carbon, or also the ring stretch vibration ofbenzene rings. A Raman peak around 1180 cm−1 is suggested to be due tosp3-rich phases, although no peak around 1180 cm−1 Raman shift is foundfor the raw measured experimantal data found in this work.Several other peaks in the Raman spectra are observed besides the D-andG-peaks. Some of the key peaks, found in the literature, are summarized bySchwan et al. [15]:− 1140 cm−1: nanocrystalline diamond− 1170 cm−1: hexagonal diamond− 1180 cm−1: nanocrystalline diamond or hexagonal diamond or sp3 richphases− 1237 cm−1 and 1306 cm−1: hexagonal diamond− 1305 cm−1: hexagonal diamond− 1332 cm−1: cubic diamond− 1350 cm−1: D-Peak, microcrystalline graphite, alternating ring stretchvibration in benzene rings− 1490 cm−1: semicircle ring stretch vibration of benzene rings− 1580 cm−1: G-Peak, sp2 stretch vibration in benzene rings or sp2stretchvibration of olefinic/conjugated chains302.5. Raman spectroscopy in the characterization of thin-film carbonSome of these different vibrational modes are visually depicted in Figure 2.7[15].The most common laser source excitation wavelengths used in the liter-ature are 632.8 nm by Wu et al. [25], 325 and 532 nm by Cekada et al. [33],633 nm by Khun et al. [27] and Maharizi et al. [14], 532 nm by Woj-ciechowski et al. [13] and Zhao et al. [10], and 514.5 nm by Buijnsterset al. [34]. In this work, the excitation wavelengths 442 and 633 nm areused. 325 nm was not an option as there is not enough scattering intensityobtained for such an excitation2. The noise is found to be much greaterfor the case of 633 nm laser excitation. Therefore, 442 nm is the chosenwavelength for most of the Raman analyzes presented in this thesis.A representative Raman spectrum, from the work of Zhao et al. [10],corresponding to a thin-film of carbon, is shown in Figure 2.8 [10]. Thesp3 to sp2 bonding ratio was stated to be 0.76, and this corresponds to theratio of the integrated intensities of two Gaussian curves that fit this Ramanspectrum in a least-squares sense. In Section 4.5, a result similar to this isverified and discussed, along with other results from the literature. WhileZhao et al. [10] has a very congruent Gaussian fit with the experimentallyobtained Raman spectrum, Sung et al. [31] was not able to fit the Gaus-sian peaks as nicely; see Figure 2.9 [31], for example. It is clear that thereis a significant deviation between the fitted curve and the original data.Nevertheless, both used the same approach in interpreting the experimen-tal results, and assessed the sp3 to sp2 bonding ratio using the integrated2This is because the optical absorption coefficient at this wavelength is so pronouncedthat a very small amount of material is probed for this wavelength.312.5. Raman spectroscopy in the characterization of thin-film carbonC"C" C"C" C"C"C"C"C" C"C"C"a)" b)"c)"Figure 2.7: (a) 1588 cm−1: stretch vibrations in benzene, (b) 1486 cm−1:semicircle stretch vibrations in benzene, and (c) 1311 cm−1: alternating ringstretch vibrations in benzene. This figure is after Schwan et al. (1996) [15].The electronic version of this figure is in color.322.5. Raman spectroscopy in the characterization of thin-film carbon  a  5# TiC4 substrates              b  5# GCr15 substrates figure 1, friction coefficient vs. time  figure 2, friction coefficient vs. time of the 1# (Cr12) sample 3.5. Raman spectra The spectra of Raman can not determine the absolute amount of sp3 and sp2 fraction, but the relative quantity can be deduced. The spectra of our films were analyzed by Origin 75 using Gauss fit, and the D peak and G peak are found at 1380cm-1and 1560cm-1 respectively, which is shown in figure 3. The ratio of ID/IG is ranging from 0.68 to 0.94 for particles and the main values are focused between 0.7 and 0.8 when the data was collected from a more smooth area. This suggests that the structure of DLC film structure have no relations to substrates as expected. 800 1000 1200 1400 1600 1800 20000500100015002000IntensityWave Number (cm-1)a  ID/IG=0.76 figure 3, Raman spectra of DLC films on all the substrates, and the spectra been deconvoluted to two bands between the wave number 1000 cm-1 and 1800cm-1. 4. Conclusion First, the structure properties and the hardness of DLC films can not be affected by the substrates. Second, the friction coefficient of DLC films will be stable if the adhesion of DLC films was strengthened. The coefficient of DLC films on all the 7 metal substrates is stable and the value is about 0.1 if with a proper interlayer.  6XUIDFH(QJLQHHULQJFigure 2.8: The Raman spectrum determined by Zhao et al. [10]. The twoGaussian peaks, selected for this peak fitting process, are depicted, as is t eirsum. The resultant integrated intensities suggest that ID/IG = 0.76 [10],where ID and IG denote the integrated intensities associated with the D-and G-peaks, respectively. The electronic version of this figure is in color.332.5. Raman spectroscopy in the characterization of thin-film carbonFigure 2.9: The Raman spectrum of a thin-film carbon sample, referred asSample A in the original article, by Sung et al. [31].342.6. Problems with the presence of oxygen and other contaminantsintensity of the two Gaussian peaks, i.e., the G- and D-peaks.2.6 Problems with the presence of oxygen andother contaminantsEven if the thin-film carbon is hard, presumably due to a high sp3 con-tent, there is no guarantee that it will adhere to the underlying substrate.Adhesion is critical to the effectivness of the resultant layer of thin-film car-bon. Good adhesion can only be achieved with a very clean and oxide-freesubstrate surface. In order to achieve proper adhesion, there are severalconditions that one must typically satisfy prior to deposition. First, be-fore deposition, the substrate surface has to be free from any dirt and dust.Second, the substrate surface must be cleaned from organic contaminationand metal oxidation, as oxygen can significantly reduce the strength of theinteratomic bonding, thereby reducing the ability of the resultant thin-filmto adhere with the underlying substrates; see Figure 2.10 [17], for example.The surface of the substrate can be cleaned through the use of soapy waterand acetone in an ultrasonic cleaning for pre-treatment. In order to re-move the oxygen layer, the substrate must be cleaned in a vacuum chamberthrough the use of a plasma. This is further discussed in Section 3.2.5.352.6. Problems with the presence of oxygen and other contaminantsFigure 2.10: Adhesion strength for different substrates and coating materi-als. This figure is after Neuville et al. [17]. The electronic version of thisfigure is in color.36Chapter 3Experiment3.1 Deposition of thin-film carbonThin-film carbon is presently being employed for a variety of industrialand military settings. It has also allowed for the development of many newand innovative products, such as high performance drills and other tools,and even pistons for a car. A number of deposition techniques that may beemployed for the fabrication of thin-film carbon are available. Each methodpresents its unique advantage and produces a different kind of thin-film car-bon. The selection as to which form of thin-film carbon should be employeddepends critically on the particular application at hand. A long-term goalof this research thrust would be the determination of a price-performancecurve for thin-film carbon, in which the trade-off between price and qualityis quantitatively established. Given the time limitations encountered duringthe writing of this thesis, this determination will be beyond the scope ofthis thesis. This thesis may be viewed as taking the first step towards therealization of this long-term goal.In this work, magnetron sputtering is used to create amorphous thin-films of carbon. Even though amorphous carbon is not the hardest materialavailable, its hardness can be readily increased through increases in the sp3373.2. Magnetron sputteringbond content, to become ta-C, at the expense of a slow growth rate [9, 12].In this thesis, the feasibility of producing thin-films of carbon with a highamount of sp3 bonds with the magnetron sputtering system that is availablewill be assessed. The answer to this question provides the first step towardsthe resolution of the ultimate question that this body of research work aimsto address, i.e., whether or not magnetron sputtering is adequate for thecosmetic jewelry applications in mind.After an introduction to the magnetron sputtering machine used in thisstudy, a description of the sample numbering and preparation means is pro-vided. The pre-cleaning process with argon plasma is also explained anddocumented. Finally, the specifications of the Raman and the SEM ma-chines, respectively, used for analyzing the samples, is given.This chapter is organized in the following manner. Section 3.2 describesthe sputtering machine used for preparing the thin-film carbon samples alongwith a description about the sample preparation, numbering, and cleaning aswell as the deposition processes. The Raman measurements, performed onthe thin-film carbon samples, are then described in Section 3.3. Finally, theSEM machine, and the approach to measure the thickness and the materialcomposition of the thin-films of carbon, are explained in Section 3.4.3.2 Magnetron sputtering3.2.1 Machine: Hummer XIIThe deposition system used for this research is a magnetron sputteringmachine manufactured by Anatech LTD. The particular machine, that is383.2. Magnetron sputteringFigure 3.1: The sputtering machine, the Hummer XII by Anatech Ltd, usedin this work for thin-film carbon depositions. The electronic version of thisfigure is in color.393.2. Magnetron sputteringemployed for the purposes of this study, is depicted in Figure 3.1. Thismachine was manufactured about 10 years ago. All control knobs areanalog and manually controlled. There is no closed-loop control system,and visual inspection during operation is required. The chamber has noload lock available in order to allow one to insert a sample without breakingthe vacuum. The top lid has to be opened in order to place the samplesin the chamber, and the chamber is exposed to oxygen and contaminantsduring every use; the same difficulties were noted by Dr. Jonathan E. Lee inhis Ph. D. Thesis for the same kind of sputtering machine [35]. Due to thelack of any closed-loop control and inaccuracies in the analog control knobs,i.e., it has a 10 kΩ potentiometer with ±10% tolerance, it is not possibleto keep all process parameters the same for each run. For most runs, anargon gas flow of 12.5 sccm is chosen, except for sample numbers 106-2,207-1, 210-1 and 210-2. To achieve a deposition pressure of 1 mTorr, theargon gas flow is set at 3.5 standard cubic centimetre per minute (sccm) forexperiment 207-1 and 4.7 sccm for experiment 106-2.The deposition parameters that are manually controllable in this set ofexperiments are the following: (1) the choice of using DC power excitation(at 400 W) or RF power excitation (at 200 W), (2) the gas flow rate, and(3) the deposition times. The throttling valve is not used during thesedepositions. The temperature can be controlled, but the heater is not usedin this setup, since, for thin-film carbon sputtering, usually the temperatureis kept low [12]. Figure 3.2 shows the setup of the vacuum chamber, withits components.403.2. Magnetron sputteringFigure 3.2: Terminology of the sputtering vacuum chamber. The electronicversion of this figure is in color.413.2. Magnetron sputtering3.2.2 Chrome sputteringThe magnetron sputtering machine, i.e., the Hummer XII, is also used forchrome sputtering. Even though this sputtering machine offers the advan-tage of multiple material depositions, there is always some inevitable cross-contamination that occurs from one deposition to the next. The chromesputtering is performed with a very similar recipe to the thin-film carbon de-positions, i.e., chrome sputtering is accomplished with a DC power of 400 Wand an argon gas flow of 12.5 sccm for about 11 minutes. The chrome mate-rial deposits not only on the substrate, but also on the walls of the vacuumchamber. This thin-film layer of chrome entraps a lot of oxygen and othercontaminant gases. The more the machine is used, the more contaminatedit becomes. It is found that it is best to grind, or wipe away, the resultantthin-film of chrome from the wall of the vacuum chamber after every use.In order to save time, however, for the purposes of those experiments, thechamber is cleaned after every 50 depositions. Tests, aimed at contrastingthe thin-film carbon that is produced following many depositions withoutcleaning with those produced immediately following a cleaning, have beenperformed. No significant change in the quality of the obtained films is ob-served. It is believed that the main problem that detracts from the thin-filmquality, in terms of its adhesion at least, is the oxygen layer on the titaniumsubstrate, which is much larger than the entrapped amount of oxygen in thecontaminated wall.423.2. Magnetron sputtering3.2.3 Sample type and numberingDifferent types and shapes of substrates are employed for the purposesof this research. These include consideration of the following:− Corning R© EAGLE XGTM AMLCD glass substrates− titanium substrates− chrome substates− flat substrates− ring substrates with a diameter of about 1 inchEach deposition run has a unique number assigned to it to describe the sub-strate, the experimental number, and the sample number considered. Thefirst number indicates if the substrate is a titanium ring (number 1) or flat(number 2). The second and third numbers are the experimental numbersin chronological order. The last number, following the dash, describes whichchamber run the sample is prepared in, there being several samples preparedfor each such run. A full list of all of the samples, their type and numbers,is provided in Chapter 4 in Section 4.2.3.2.4 Substrate preparationPrior to deposition, the titanium samples are mechanically polished witha buffing machine. It is found that a shiny and smooth surface increases theglossy look of the resultant thin-film carbon layer, and that this is importantfor cosmetic jewelry applications. Then, the substrates are cleaned in anultrasonic bath in hot water with some all purpose cleaner (TSP) for about20 minutes in order to remove any dirt accumulated from the polishing433.2. Magnetron sputteringprocess. Finally, the substrates are cleaned in another ultrasonic bath withpure acetone in order to eliminate any residual organic contaminants. Inorder to allow the substrates to dry in a streak-free manner, they are swungon a cord to dry equally in the air. The rings are never touched by humanhands during the entire cleaning process.3.2.5 RF cleaningPart of substrate preparation is the proper cleaning of the surface. Aswas described in the previous subsection, the substrates are cleaned withsoapy water and through the use of acetone in an ultrasonic bath. Afterputting the substrates in the vacuum chamber, in order to reduce as muchoxygen as possible, a vacuum pressure of about 1.3 x 10−4 Torr is created.Still, there is an oxygen and contamination layer on the substrate that can-not be removed with the cleaning bath.To clean the substrates from contamination and oxygen on a molecularlevel, an argon based plasma cleaning process is used. The vacuum chamberis filled with argon gas at a fairly high argon gas flow, i.e., 75 sccm, and ion-ized through the application of an RF power of 200 W at 13.56 MHz. Thesubstrate surface gets bombarded with argon ions, and this removes the for-eign molecules that cause contamination, and thus, no chemical reaction,or oxidation, takes place during this process. Argon RF plasma cleaningremoves many types of contamination, such as fluorine, organic contami-nation, and metal oxides [36]. By bombarding the substrate with argonions, the surface gets roughened, and the bonding or adhesion is improved.A water drop test is performed in order to verify the successful titanium443.2. Magnetron sputteringring substrate cleaning with this process. For this purpose, a water dropis dropped on an uncleaned titanium ring. With an unprocessed ring, thewater drop splits into several little drops that stick onto the surface. Afteran RF cleaning for 85 minutes, however, the water drop will slide perfectlydown the ring surface, and does not stick to the surface as with the un-cleaned ring. The comparison between this water drop test is illustrated inFigure 3.3.3.2.6 Recipe: a step-by-step guideThe overall sputtering process may be divided into four basic steps: (1) astartup phase, (2) an RF argon plasma cleaning process, (3) sputtering, and(4) the shut-down sequence. After cleaning the substrate, the samples areplaced into the chamber, and the power and gas supplies are switched on forthe machine. After pumping down to a vacuum of about 1.3 ×10−4 Torr,the RF argon plasma cleaning process can be started. After cleaning thesubstrate surface, the substrate has to be cooled down before the sputteringprocedure can take place. After executing the proper shut-down sequence,the chamber can be vented and the coating process is finished. This recipeis explained, in detail, in Appendix 5 in a chart for a step-by-step guide.3.2.7 ExperimentsTen different sputtering experiments are performed in order to partiallyprobe the deposition parameter space in terms of different argon gas flows,power selections, sputtering times, and substrate types and geometries em-ployed. Table 3.1 provides an overview of all the performed experiments.453.2. Magnetron sputteringFigure 3.3: Water drop test for a titanium ring before and after argon plasmacleaning. Left: The uncleaned ring, with water drops sticking onto thesurface. Right: An RF argon plasma cleaned ring, with water drops slidingdown the cleaned surface without residual stuck water drops. The rightimage is blurry owing to the absence of a proper focal point and photographicequipment limitation. The electronic version of this figure is in color.463.2. Magnetron sputteringTable 3.1: An overview of the sputtering experiments corresponding to thedifferent sample numbers. Deposition parameters for each sample, with thesetup and deposition time information indicated. Reference sample number201-2 is shaded.Sample Substrate RF / Power Gun-substrate RF Cleaning Pre- Sputter ArgonNo. DC distance time sputtering time106-2 Ti ring RF 200 W 60 mm 87 min 5 min 175 min 4.7 sccm111-1 Cr on Ti ring DC 400 W 60 mm 61 min 3 min 360 min 12.5 sccm115-1 Ti ring RF 200 W 60 mm 30 min 5 min 120 min 12.5 sccm115-3 Ti ring RF 200 W 60 mm 30 min 5 min 120 min 12.5 sccm121-3 Ti ring RF 200 W 60 mm 40 min 5 min 118 min 12.5 sccm123-3 Ti ring RF 200 W 60 mm 40 min 5 min 120 min 12.5 sccm201-2 Glass DC 400 W 60 mm 20 min 3 min 60 min 12.5 sccm204-2 Glass RF 200 W 60 mm 20 min 3 min 60 min 12.5 sccm207-1 Glass RF 200 W 60 mm 20 min 3 min 60 min 3.5 sccm210-1 Glass RF 200 W 60 mm 20 min 3 min 60 min 25 sccm210-2 Flat Ti RF 200 W 60 mm 20 min 3 min 60 min 25 sccm473.3. Raman SpectroscopySample number 115-1 and 115-3 are produced in the same sputtering run.The thickness of the thin-film carbon is determined for sample number 115-1 and the Raman spectrum is obtained for all other samples; the thin-filmthickness corresponding to only one thin-film is determined owing to timelimitations. Table 3.2 shows all of the sputtering samples that are preparedand analyzed for the purposes of this thesis.3.3 Raman Spectroscopy3.3.1 Machine: LabRAM HRThe machine used for this Raman analysis is a LabRAM HR Ramanconfocal microscope by Horiba Scientific, a picture of which is shown inFigure 3.4. The source laser is an HeCd laser with a 442 nm wavelengthand 2400 gratings. For all Raman measurements, an objective lens, witha 100 × magnification and a numerical aperture (NA) of 0.9 is used. Forsample 115-3, the lens is changed to 50 × magnification in order to obtaina better Raman result. The laser spot size is about 0.8 µm for this setupfor a Raman source of 442 nm. The original laser power is 2 mW/µm2 andno power filter is used for the measurements. Therefore, the total appliedpower on a spot of about 2 µm2 is around 4 mW. The spectral resolution is0.05 cm−1 over the Raman shift range, i.e., between 800 and 2000 cm−1. Twoscans for each spectrum are performed, so as to reduce the correspondingsignal-to-noise ratio. This machine also supports the source wavelengths of325 and 632.817 nm. The later will be referred to as 633 nm for the purposesof this work, as no signal is obtained for the 325 nm case.483.3. Raman SpectroscopyTable 3.2: Eleven samples are prepared and analyzed in this work. Sixtitanium ring substrates, four flat glass substrates, and one flat piece oftitanium substrate are considered. Sample number 201-2 is the referencesample. The electronic version of these figures is in color.Sample number 106-2 Sample number 111-1 Sample number 115-1Sample number 115-3 Sample number 121-3 Sample number 123-3Sample number 201-2 Sample number 204-2 Sample number 207-1(Reference)Sample number 210-1 Sample number 210-2493.3. Raman SpectroscopyFigure 3.4: Front side of the LabRAM HR Raman confocal microscope byHoriba Scienfic. The electronic version of this figure is in color.503.3. Raman Spectroscopy3.3.2 MeasurementsAll ten sputtering samples are analyzed using Raman spectroscopy. Allthe Raman measurements are performed with the following settings:− no power filter− a 100 µm hole− a 100 × magnification lens with a NA3 = 0.9− 20 seconds of acquisition time− 442 nm laser source excitation (except for sample 201-2, also 633 nm)− a spot size of 0.8 µm at 442 nm− power intensity: 2 µm2− 2400 gratings for 442 nm− spectral resolution of 0.05 cm−1 over a Raman shift range from 800 to2000 cm−1− two scans per spectrum, so as to reduce the signal-to-noise ratioThe samples are placed on microscope glass, as shown in Figure 3.5 for thereference sample (sample number 201-2), i.e., a thin-film of carbon depositedon glass using magnetron sputtering at a DC power of 400 W for 60 minutes,with a 12.5 sccm argon gas flow. The titanium rings are hung on a pen underthe lens so as to focus the laser on top of the curved area, as depicted inFigure 3.6. Table 3.3 summarizes all of the Raman measurements that areperformed in this work with the specific Raman settings.3The numerical aperture (NA) defines over what range the lens can accept or emitlight. The numerical aperture is dimensionless.513.3. Raman SpectroscopyFigure 3.5: Planar sample placement for Raman spectroscopy. The referencesample, i.e., sample number 201-2, a thin-film of carbon deposited on glassusing magnetron sputtering at a DC power of 400 W for 60 minutes, with a12.5 sccm argon gas flow, is depicted in this image. The electronic versionof this figure is in color.523.3. Raman SpectroscopyFigure 3.6: Sample placement for Raman spectroscopy. A plain titaniumring is shown as an example. The electronic version of this figure is in color.533.3. Raman SpectroscopyTable 3.3: Overview of the Raman measurements and measurement set-tings. All Raman measurements, performed in this work, with the corre-sponding measurement settings. Reference sample number 201-2 is shaded.(* without background compensation, ** with background compensation;see Section 4.3.4).Sample Raman Lens Acq. PowerNo. source magnification time filter106-2 442 nm 100× 20 sec none111-1 442 nm 100× 20 sec none115-3 442 nm 50× 20 sec none121-3 442 nm 100× 20 sec none123-3 442 nm 100× 20 sec none201-2 442 nm 100× 20 sec none201-2 633 nm 100× 20 sec none204-2 442 nm 100× 20 sec none207-1* 442 nm 100× 20 sec none207-1** 442 nm 100× 20 sec none210-1 442 nm 100× 20 sec none210-2 442 nm 100× 20 sec none543.4. Scanning electron microscope3.4 Scanning electron microscope3.4.1 MachineThe SEM machine, used in this work, was manufactured by Tescan USAInc.; it is a Mira3 XMU Field Emission SEM. This microscope is as de-picted in Figure 3.7. This machine is part of the SEMLab at The Universityof British Columbia, and it allows for high resolution imaging from a samplesurface of up to several hundred thousand times magnification. The surfaceis probed with an electron beam, and a greyscale image is thus produced.The SEM has an 80 mm2 EDS X-ray detector from Oxford Instruments inorder to detect X-ray emissions due to ionized electrons falling back into theinner shells. Each X-ray photon emission has an energy, or wavelength, thatis characteristic of the element it is emitted from. The software associatedwith this machine then matches the energy spectrum against known elemen-tal standards in order to determine the relative amounts of material. In thisway, a material compositional analysis can be performed, as described inSections 3.4.3 and 3.4.3.4.2 Determination of the thicknessIn order to determine the thickness of the thin-film carbon films, forsample number 115-1, i.e., a thin-film of carbon deposited on a titaniumring using magnetron sputtering at an RF power of 200 W for 120 minutes,with a 12.5 sccm argon gas flow, the SEM is used. The approach used inthis work is the first trial of this technique, and just sample number 115-1 is analyzed in this manner. To find the transition from the underlying553.4. Scanning electron microscopeFigure 3.7: The SEM machine from the SEMLab at The University of BritishColumbia. The electronic version of this figure is in color.563.4. Scanning electron microscopetitanium substrate to the thin-film, a side of the titanium ring sample ispolished in order to obtain a nice transition from the titanium substrate tothe thin-film carbon. The titanium ring substrate has a curved surface, andthe electron microscope detector is looking at the thin-film carbon from anon-planar point of view. Thin-film carbon has a much brighter appearancethan the underlying titanium substrate, as seen in Figure 3.8. Since thepoint of view is from an angle, the bright thin-film carbon in Figure 3.8 isnot the actual thickness, but rather a look on the surface. To find the actualthickness, i.e., the transition from substrate to the thin-film, the sample hasto be tilted, and the picture zoomed in towards the thin-film of carbon. Azoomed in picture of the thin-film carbon on the titanium ring sample isshown in Figure 4.24.3.4.3 Material analysisAs described in the previous section, it is a challenge to find the tran-sition from the titanium substrate of sample number 115-1 to the thin-filmcarbon. In order to determine the right visual interpretation of the im-age, a material compositional analysis is performed so as to distinguish thetitanium substrate from the thin-film carbon. Several spectra are taken.Spectrum number 13 and 14 analyzed the titanium substrate and spectrumnumber 15 and 16 of the thin-film carbon. Spectra 14 and 16 are shown anddiscussed in Section 4.6. A line material analysis is also performed betweenspectra 13 or 14 and 15 or 16, respectively, in order to acquire the materialcompositional transition from titanium to carbon. All the correspondingmaterial analysis results are discussed, in detail, in Section 4.6.573.4. Scanning electron microscopeFigure 3.8: SEM: The first thickness measurement of sample number 115-1, i.e., a thin-film of carbon deposited on a titanium ring using magnetronsputtering at an RF power of 200 W for 120 minutes, with a 12.5 sccmargon gas flow. The thin-film carbon is much brighter than the underlyingtitanium substrate. This view is from an angular point of view. Imageobtained with the assistance of Mr. D. Arkinstall. The electronic version ofthis figure is in color.58Chapter 4Analysis4.1 Experimental analysis of thin-film carbonForms of thin-films of carbon are currently being used for a wide varietyof industrial and military applications [8, 9, 17]. Differences in the deposi-tion technique and deposition conditions used have been found to have anenormous impact on the properties of the resultant forms of thin-film car-bon. Based on the demands of the application at hand, the right depositiontechnique and conditions must be sought. If extreme hardness is required, ahydrogen-free form of thin-film carbon should be employed [9], this typicallynot being achievable through the use of a CVD process [25]. On the otherhand, CVD processes offer uniformity in terms of thin-film deposition [24],and this is a great advantage for applications where wear-and-tear resis-tance is required, non-uniformity in the thin-film thickness and compositionhaving potentially disastrous consequences in terms of performance.In this thesis, magnetron sputtering is employed in order to deposit thin-film carbon. This is because this technique is inexpensive to use and requiresthe use of commonly available experimental equipment. As of the presentmoment, whether or not this form of thin-film carbon offers adequate qual-ity for the particular application under consideration, i.e., cosmetic jewelry,594.1. Experimental analysis of thin-film carbonremains unknown. In this chapter, results corresponding to these thin-filmcarbon samples, are presented. The ultimate purpose of this experimentalwork is to critically assess the suitability of these films for cosmetic jew-elry applications. Unfortunately, owing to limitations in the amount of timeavailable, the results presented in this thesis represent only the first steptaken in the process required in order to address this question. The qualityof the resultant films will be determined through a evaluation of the sp3 bondfraction, this usually being taken as a proxy measure for the durability andhardness of such films. As an exhaustive probing of the deposition parame-ter space lies beyond the scope of the present analysis, the approach adoptedwill be to contrast the thin-film carbon samples on a sample-by-sample ba-sis, all conditions being the same except for one; in a sense, the conditionsconsidered represent a sampling of the deposition parameter space. A sys-tematic study, in which the deposition parameter space is probed in itstotality, will be the goal of future work related to this project, but is beyondthe scope of this particular thesis.Initially, an overview of all of the experimental samples that are con-sidered in this analysis is provided, a brief description of the depositionconditions employed and the experimental procedures used being tabulated.Then, for one particular sample, which is henceforth referred to as the refer-ence sample, the means whereby unprocessed Raman spectroscopy results,i.e., raw experimental Raman results, are prepared for subsequent analy-sis, through a process of baseline correction, is detailed; this processing isused for all Raman spectra considered in this analysis, the intermediarysteps only being explicitly shown for the reference sample case. Following604.1. Experimental analysis of thin-film carbonbaseline correction, the experimental spectra are then decomposed into twodistinct peaks, the G-peak, being related to the presence of sp2 bonds, andthe D-peak, being related to the presence of sp3 bonds. Two Gaussian peaksare employed for this peak fitting purpose, and the sum of these peaks isleast-squares fit to the overall processed Raman spectra, i.e., the peak pa-rameters are selected in such a manner that the difference between the sumof the two peaks and the baseline corrected Raman spectra has the least-squares error value4. After explaining this procedure of data processing, allof the resultant results are presented and contrasted with each other. Aparticular interpretation for the Raman spectra associated with these ex-perimental results is then discussed, varying opinions, from the scientificliterature, also being presented. Finally, a preliminary analysis of one of thethin-film carbon samples, using the SEM machine, is then presented, andthe meaning of the results is discussed.This chapter is organized in the following manner. In Section 4.2, thesamples considered in this analysis are tabulated, the experimental con-ditions employed being specified. Then, in Section 4.3, the procedure ofRaman analysis is then presented. A summary of the Raman results ob-tained is then featured in Section 4.4. Further interpretation of the Ramanresults is presented in Section 4.5. Finally, in Section 4.6, the scanning elec-tron microscopy results, obtained for one of the thin-film carbon samplesconsidered in this analysis, are presented.4For the special case of very thin-films, i.e., sputtering with a low argon gas flow, aprocess of background compensation is also used.614.2. Description and overview of the examined samples4.2 Description and overview of the examinedsamplesEleven Raman measurements on ten different samples of thin-film car-bon are performed for the purposes of this analysis. Table 4.1 provides anoverview of all of the samples considered in this analysis, the depositionconditions, the substrates employed, and the Raman source wavelength be-ing tabulated for each such Raman measurement. For all of the samplesconsidered, the Raman source beam is projected onto the middle of thesample, i.e., away from its edges. All samples are probed using a Ramansource of 442 nm, except for sample number 201-2, which is measured usinga Raman source of 633 nm in order to ascertain what effect the selection ofthe Raman excitation wavelength produces. It should be noted that samplenumber 201-2 is the reference sample, this sample being produced throughmagnetron sputtering at a DC power of 400 W for 60 minutes with an argongas flow of 12.5 sccm on glass. Variations in the nature of the sputtering,i.e., the duration of the sputtering, the nature of the plasma excitation, i.e.,whether DC or RF, and the argon gas flow rates, are considered. Variationsin the type of substrate employed are also considered. For the case of RFexcitation, an RF power of 200 W is considered. The sputtering time isvaried from 60 minutes up to 360 minutes. An argon gas flow of 3.5 sccm isalmost the minimal value required for successful sputtering. The argon gasflow is increased up to double of the reference flows, i.e., 25 scm for sample210-1 and 210-2. Different substrates are considered in order to provide afeel for the influence of the substrate on the resultant film properties.624.2. Description and overview of the examined samplesTable 4.1: Overview of the samples with the deposition parameters and theRaman source identified. Ten different experimental samples were examinedwith Raman spectroscopy. Sample 201-2, the reference (shaded) sample, isanalyzed with two different Raman laser source excitation wavelengths, 442and 633 nm.Sample Substrate RF / Power Sputter Argon RamanNo. DC time source106-2 Ti ring RF 200 W 175 min 4.7 sccm 442 nm111-1 Cr on Ti ring DC 400 W 360 min 12.5 sccm 442 nm115-3 Ti ring RF 200 W 120 min 12.5 sccm 442 nm121-3 Ti ring RF 200 W 118 min 12.5 sccm 442 nm123-3 Ti ring RF 200 W 120 min 12.5 sccm 442 nm201-2 Glass DC 400 W 60 min 12.5 sccm 442 nm201-2 Glass DC 400 W 60 min 12.5 sccm 633 nm204-2 Glass RF 200 W 60 min 12.5 sccm 442 nm207-1 Glass RF 200 W 60 min 3.5 sccm 442 nm210-1 Glass RF 200 W 60 min 25 sccm 442 nm210-2 Flat Ti RF 200 W 60 min 25 sccm 442 nm634.3. Raman analysisThe following sample-by-sample comparisons are made:− DC power of 400 W (sample number 201-2) vs. RF power of 200 W(sample number 204-2)− glass substrate (sample number 207-1) vs. titanium ring substrate(sample number 106-2)− glass substrate (sample number 210-1) vs. flat titanium substrate(sample number 210-2)− low argon gas flow (sample number 210-1) vs. high argon gas flow(sample number 207-1)− test for reproducibility (sample numbers 115-3, 121-3, and 123-3)− Raman source laser excitation wavelengths of 442 vs. 633 nm (samplenumber 201-2)− thin-film carbon, prepared with exceptional low argon gas flow (samplenumber 207-1): background compensation vs. no background compen-sation4.3 Raman analysis4.3.1 Unprocessed Raman dataThe unprocessed Raman spectrum, corresponding to the reference sam-ple (sample number 201-2), i.e., a thin-film of carbon deposited using mag-netron sputtering at a DC power of 400 W for 60 minutes, with a 12.5 sccmargon gas flow, is depicted in Figure 4.1. This spectra is determined usinga 442 nm Raman source, and the spot size is about 10 µm; this spot is castupon the center at the sample, i.e., away from the edges. The spectrum644.3. Raman analysis800 1000 1200 1400 1600 1800 200005001000150020002500300035004000Raman shift (cm−1)Intensity  (a.u.)Sample number 201−2, Raman source: 442 nm  unprocessed dataFigure 4.1: The unprocessed Raman spectrum of the reference sample (sam-ple number 201-2), i.e., a thin-film of carbon deposited on glass using mag-netron sputtering at a DC power of 400 W for 60 minutes, with a 12.5 sccmargon gas flow. This spectrum is depicted between 800 and 2000 cm−1.654.3. Raman analysisbetween 800 and 2000 cm−1 is plotted, as this is the region in which peaks,relevant to this particular analysis, are found. Distinctive peaks are notedat both ∼1350 and ∼1580 cm−1 in this figure. These peaks, D and G, re-spectively, are believed to be representative of the presence of sp3 and sp2carbon bonds, respectively, as was suggested by Zhao et al. [10].4.3.2 Baseline correctionExperimental measurements are susceptible to noise. For the case ofRaman spectroscopy, the noise present will lead to the presence of a baselinein the spectrum, that must be compensated for prior to any spectral analysis.This compensation process, otherwise known as baseline correction, is acommon practice amongst researchers in this field. The process of baselinecorrection is explained in the following paragraph.Baseline correction is typically achieved by subtracting a baseline fromthe unprocessed data, this baseline arising as a consequence of these “noisy”processes. In Figure 4.2, the baseline is depicted for the Raman spectrumcorresponding to the reference sample (sample number 201-2), i.e., a thin-film of carbon deposited on glass using sputtering at a DC power of 400 W for60 minutes with a 12.5 sccm argon flow, this being the same sample as thatconsidered in Figure 4.1. The baseline corresponding to this particular dataset is chosen to span from 1000 to 1800 cm−1, and is drawn by connectingthe data points corresponding to these limits; this range for the baselinecorrection was suggested by Zhao et al. [10]. Other researchers use differentmeans of determining this baseline, but the essence remains the same as thatdepicted in Figure 4.2, these other techniques being further discussed in the664.3. Raman analysis800 1000 1200 1400 1600 1800 200005001000150020002500300035004000Raman shift (cm−1)Intensity (a.u.)Sample number 201−2, Raman source: 442 nm  unprocessed databaselineFigure 4.2: The unprocessed Raman spectrum of the reference sample (sam-ple number 201-2), i.e., a thin-film of carbon deposited on glass using mag-netron sputtering at a DC power of 400 W for 60 minutes, with a 12.5 sccmargon gas flow, depicted in Figure 4.1, with a baseline, connecting the datapoints corresponding to 1000 and 1800 cm−1, being shown with the dashedline. The electronic version of this figure is in color.674.3. Raman analysisliterature. The data is baseline corrected by removing this baseline from theunprocessed spectrum. For the case of the reference sample (sample number201-2), the resultant baseline corrected data is depicted in Figure 4.3.4.3.3 Peak fittingThere is a general consensus in the field that the amount of sp2 andsp3 bonds within a given sample of thin-film carbon is proportional to theintegrated intensity of the corresponding peaks. That is, the ratio of theintegrated intensity of the ∼1350 cm−1 peak (D-peak) and the ∼1580 cm−1peak (G peak) provides a measure of the sp3 to sp2 bonding ratio within thethin-film of carbon. Accordingly, peak fitting is critical to the subsequentanalysis. If one accepts that the presence of sp2 bonds leads to the existenceof a peak at ∼1580 cm−1 and that the presence of sp3 bonds leads to theexistence of a peak at ∼1350 cm−1, it is usually assumed that the overallbaseline corrected spectrum is the sum of these two peaks. Thus, the taskbecomes identifying the peak parameters that best fit the resultant baselinecorrected data.For the purposes of this analysis, each peak is assumed to be in the formof a Gaussian function, i.e., each peak is assumed to be of the formA exp[−(x−m)2σ2], (4.1)where A denotes the peak amplitude, m represents the peak location, andσ provides a measure of the peak breadth. The total fit of the Ramanspectrum is therefore a fit toAG exp[−(x−mG)2σG2]+AD exp[−(x−mD)2σD2], (4.2)684.3. Raman analysis1000 1100 1200 1300 1400 1500 1600 1700 1800050010001500200025003000Raman shift (cm−1)Intensity (a.u.)Sample number 201−2, Raman source: 442 nm  baseline correctedFigure 4.3: The baseline corrected Raman spectrum of the reference sample(sample number 201-2), i.e., a thin-film of carbon deposited on glass usingmagnetron sputtering at a DC power of 400 W for 60 minutes, with a 12.5sccm argon gas flow, between 1000 and 1800 cm−1.694.3. Raman analysiswhere AG and AD denote the peak amplitudes of the G- and D-peaks,respectively, mG and mD represent their peak locations, and σG and σDrepresent their breadth. Assuming that the baseline corrected data corre-sponds to the sum of two appropriately selected Gaussian peaks, as in Eq.(4.2), the approach employed is to select the model parameters, i.e., AG,AD, mG, mD, σG, and σD, so that the difference between this sum of twopeaks and the baseline corrected experimental data is minimized; this anal-ysis is performed within the framework of a least-squares error analysis. Forthe reference sample, i.e., sample number 201-2, for the baseline correctedRaman experimental data, depicted in Figure 4.3, the two peaks that arisefrom this peak decomposition process, and the sum of the two peaks, are alldepicted in Figure 4.4. The integrated peaks, for this particular fit, suggestthat the sp3 to sp2 bonding ratio is 3.21 ± 0.07 for this sample of thin-film carbon. This is close to that typically found in high-quality thin-filmcarbon [12].It should be noted that, as has been hinted earlier, some researchersinstead assert that it is the peak magnitude that should be looked at, ratherthan the peak integrations; see, for example, Tai et al. [29]. Given thatthe majority of researchers employ the integrated intensity, this will be theprocedure employed for the purposes of this particular analysis, whereID =∫ 1800 cm−11000 cm−1AD exp[−(x−mD)2σ2D], (4.3)for the D-peak, andIG =∫ 1800 cm−11000 cm−1AG exp[−(x−mG)2σ2G], (4.4)for the G-Peak. This issue is further discussed in Section 4.5.704.3. Raman analysis1000 1100 1200 1300 1400 1500 1600 1700 1800050010001500200025003000Raman shift (cm−1)Intensity (a.u.)Sample number 201−2, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.4: The peak fitting corresponding to the baseline corrected Ramanspectrum of the reference sample (sample number 201-2), i.e., a thin-filmof carbon deposited on glass using magnetron sputtering at a DC power of400 W for 60 minutes, with a 12.5 sccm argon gas flow, between 1000 and1800 cm−1. The two Gaussian peaks selected for this peak fitting processare depicted, as is the sum. The resultant integrated intensities suggest thatID / IG = 3.21 ± 0.07. The fitting parameters are found to be AD = 2042and AG = 2174, mD = 1388 cm−1 and mG = 1588 cm−1, σD = 197 cm−1and σD = 57.4 cm−1. The root mean square error is 69.74. The electronicversion of this figure is in color.714.3. Raman analysis4.3.4 Background subtractionOne of the samples considered in this analysis, sample number 207-1, i.e.,a thin-film of carbon deposited on glass using magnetron sputtering at anRF power of 200 W for 60 minutes, with a 3.5 sccm argon gas flow, appearsto be very thin due to the low argon gas flow that is employed for thisparticular deposition. As this sample is thinner, the number of interactionswith the Raman source will be reduced. Thus, while most Raman spectraconsidered in this thesis reach a maximum of thousands of counts per Ramanshift interval, as is seen in Figure 4.1, the Raman spectrum correspondingto sample number 207-1 only reaches a maximum of a couple of hundredscounts per Raman shift interval, as may be seen in Figure 4.5.As this sample is very thin, interactions with the underlying substrateand background are much more likely to influence the resultant Ramanspectrum. These interactions should be compensated for. To do this, theRaman spectrum corresponding to a pure glass substrate is measured; seeFigure 4.6. The Raman spectrum of sample 207-1, without backgroundcompensation, is given in Figure 4.5. The ratio of the integrated intensityof the two Gaussian peaks is very different if the background is taken intoconsideration, i.e., ID/IG = 3.35 ± 0.13. With background compensation,the true ratio of sp3 to sp2 is determined as ID/IG = 2.24 ± 0.05. The least-squares error of the Gaussian fit is lower after background compensation;see Figure 4.14.724.3. Raman analysis1000 1100 1200 1300 1400 1500 1600 1700 18000100200300Raman shift (cm−1)Intensity (a.u.)Sample number 207−1, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.5: The baseline corrected Raman spectrum without backgroundcompensation of sample number 207-1, i.e., a thin-film of carbon depositedon glass using magnetron sputtering at an RF power of 200 W for 60 minutes,with a 3.5 sccm argon gas flow. The two Gaussian peaks selected for thispeak fitting process are depicted, as is the sum. The resultant integrated in-tensities suggest that ID/IG = 3.35 ± 0.13. The fitting parameters are foundto be AD = 203.2 and AG = 206, mD = 1407 cm−1 and mG = 1585 cm−1,σD = 197.3 cm−1 and σD = 57.81 cm−1. The root mean square error is11.80. The electronic version of this figure is in color.734.3. Raman analysis800 1000 1200 1400 1600 1800 2000050100150200250300350400450500Raman shift (cm−1)Intensity (a.u.)Plain samples of glass and titanium, Raman source: 442 nm  plain glassplain flat titaniumFigure 4.6: The Raman spectrum of a piece of plain glass and flat titanium.The electronic version of this figure is in color.744.4. Summary of the Raman results4.4 Summary of the Raman resultsThe other thin-film carbon samples considered in this analysis, tabulatedin Table 4.2, have also had their Raman spectra measured. The results ofthis peak fitting, on the baseline corrected Raman spectra, are depictedin Figures 4.7, 4.8, 4.9, 4.10, 4.11, 4.12, 4.13, 4.14, 4.15, and 4.16for sample numbers 106-2, 111-1, 115-3, 121-3, 123-3, 201-2 with a Ramansource of 633 nm, 204-2, 207-1, 210-1, 210-2, respectively. Table 4.2 gives anoverview of all of the experimental samples considered in this work and thecorresponding Raman results for the sp3 to sp2 bonding ratio, i.e., ID/IG. Itis noted that all samples have sp3 to sp2 bonding ratios that are close to thatof high-quality thin-film carbon [12]. The error is based on a Matlab R© resultwithin a 95% confidence interval. Matlab R© uses the method of least-squareswhen peak fitting data. The error is basically a range of possible resultswithin 95% of the best fit. For a specification of the thin-film carbon quality,in terms of hardness and durability, often the sp3 content is employed. InTable 4.3, the Raman results, in terms of the sp3 to sp2 ratio, ID/IG, isgiven along with the corresponding sp3 content, i.e., ID/(ID + IG). The sp3content shows minimal variation, i.e., the differences between these valuesare quite slight. Since in the literature Raman results are mostly expressedin terms of the sp3 to sp2 ratio, i.e., ID/IG, this work will also present thisbonding ratio as the final result, rather than just the sp3 content. Also, inTable 4.3, the fitting parameters, AD and AG, mD and mG, σD and σG,according Eq. (4.2) from Section 4.3.3, are given.754.4. Summary of the Raman resultsTable 4.2: Experimental parameters for each sample with the Raman resultsobtained for the sp3 to sp2 bonding ratio, i.e., ID/IG. Reference samplenumber 201-2 is shaded. (* without background compensation, ** withbackground compensation; see Section 4.3.4).Sample Substrate RF / Power Sputter Argon Raman ID/IG Figure /No. DC time source Page106-2 Ti ring RF 200 W 175 min 4.7 sccm 442 nm 3.30 ± 0.07 4.7 / 77111-1 Cr on Ti ring DC 400 W 360 min 12.5 sccm 442 nm 2.77 ± 0.06 4.8 / 78115-3 Ti ring RF 200 W 120 min 12.5 sccm 442 nm 3.20 ± 0.06 4.9 / 79121-3 Ti ring RF 200 W 118 min 12.5 sccm 442 nm 3.39 ± 0.06 4.10 / 80123-3 Ti ring RF 200 W 120 min 12.5 sccm 442 nm 3.20 ± 0.07 4.11 / 81201-2 Glass DC 400 W 60 min 12.5 sccm 442 nm 3.21 ± 0.07 4.4 / 71201-2 Glass DC 400 W 60 min 12.5 sccm 633 nm 3.18 ± 0.10 4.12 / 82204-2 Glass RF 200 W 60 min 12.5 sccm 442 nm 2.91 ± 0.06 4.13 / 83207-1* Glass RF 200 W 60 min 3.5 sccm 442 nm 3.35 ± 0.13 4.5 / 73207-1** Glass RF 200 W 60 min 3.5 sccm 442 nm 2.24 ± 0.05 4.14 / 84210-1 Glass RF 200 W 60 min 25 sccm 442 nm 3.01 ± 0.06 4.15 / 85210-2 Flat Ti RF 200 W 60 min 25 sccm 442 nm 2.88 ± 0.05 4.16 / 86764.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 1800050010001500200025003000Raman shift (cm−1)Intensity (a.u.)Sample number 106−2, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.7: The baseline corrected Raman spectrum of sample number 106-2, i.e., a thin-film of carbon deposited on a titanium ring using magnetronsputtering at an RF power of 200 W for 175 minutes, with a 4.7 sccm argongas flow. The two Gaussian peaks selected for this peak fitting process aredepicted, as is the sum. The resultant integrated intensities suggest thatID/IG = 3.30 ± 0.07. The fitting parameters are found to be AD = 1992and AG = 1936, mD = 1389 cm−1 and mG = 1588 cm−1, σD = 199 cm−1and σD = 61.78 cm−1. The root mean square error is 57.80. The electronicversion of this figure is in color.774.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 180005001000150020002500300035004000Raman shift (cm−1)Intensity (a.u.)Sample number 111−1, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.8: The baseline corrected Raman spectrum of sample number 111-1, i.e., a thin-film of carbon deposited on a chrome layer on a titanium ringusing magnetron sputtering at a DC power of 400 W for 360 minutes, witha 12.5 sccm argon gas flow. The two Gaussian peaks selected for this peakfitting process are depicted, as is the sum. The resultant integrated inten-sities suggest that ID/IG = 2.77 ± 0.06. The fitting parameters are foundto be AD = 2506 and AG = 3228, mD = 1372 cm−1 and mG = 1589 cm−1,σD = 190.9 cm−1 and σD = 53.38 cm−1. The root mean square error is102.01. The electronic version of this figure is in color.784.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 1800050010001500200025003000350040004500Raman shift (cm−1)Intensity (a.u.)Sample number 115−3, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.9: The baseline corrected Raman spectrum of sample number 115-3, i.e., a thin-film of carbon deposited on a titanium ring using magnetronsputtering at an RF power of 200 W for 120 minutes, with a 12.5 sccm argongas flow. The two Gaussian peaks selected for this peak fitting process aredepicted, as is the sum. The resultant integrated intensities suggest thatID/IG = 3.20 ± 0.06. The fitting parameters are found to be AD = 2507and AG = 2952, mD = 1395 cm−1 and mG = 1590 cm−1, σD = 218.6 cm−1and σD = 57.42 cm−1. The root mean square error is 89.78. The electronicversion of this figure is in color.794.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 1800020040060080010001200Raman shift (cm−1)Intensity (a.u.)Sample number 121−3, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.10: The baseline corrected Raman spectrum of sample number 121-3, i.e., a thin-film of carbon deposited on a titanium ring using magnetronsputtering at an RF power of 200 W for 118 minutes, with a 12.5 sccm argongas flow. The two Gaussian peaks selected for this peak fitting process aredepicted, as is the sum. The resultant integrated intensities suggest thatID/IG = 3.39 ± 0.06. The fitting parameters are found to be AD = 751.9and AG = 761.7, mD = 1400 cm−1 and mG = 1583 cm−1, σD = 218.1 cm−1and σD = 62.86 cm−1. The root mean square error is 21.18. The electronicversion of this figure is in color.804.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 18000200400600800100012001400160018002000Raman shift (cm−1)Intensity (a.u.)Sample number 123−3, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.11: The baseline corrected Raman spectrum of sample number 123-3, i.e., a thin-film of carbon deposited on a titanium ring using magnetronsputtering at an RF power of 200 W for 120 minutes, with a 12.5 sccm argongas flow. The two Gaussian peaks selected for this peak fitting process aredepicted, as is the sum. The resultant integrated intensities suggest thatID/IG = 3.20 ± 0.07. The fitting parameters are found to be AD = 1128and AG = 1201, mD = 1394 cm−1 and mG = 1590 cm−1, σD = 202.2 cm−1and σD = 58.94 cm−1. The root mean square error is 39.78. The electronicversion of this figure is in color.814.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 1800020040060080010001200Raman shift (cm−1)Intensity (a.u.)Sample number 201−2, Raman source: 633 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.12: The baseline corrected Raman spectrum (with 633 nm excita-tion wavelength) of the reference sample (sample number 201-2), i.e., a thin-film of carbon deposited on glass using magnetron sputtering at a DC powerof 400 W for 60 minutes, with a 12.5 sccm argon gas flow. The two Gaussianpeaks selected for this peak fitting process are depicted, as is the sum. Theresultant integrated intensities suggest that ID/IG = 3.18± 0.10. The fittingparameters are found to be AD = 913.2 and AG = 702.6, mD = 1339 cm−1and mG = 1582 cm−1, σD = 170.8 cm−1 and σD = 69.63 cm−1. The rootmean square error is 58.70. The electronic version of this figure is in color.824.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 1800050010001500200025003000Raman shift (cm−1)Intensity (a.u.)Sample number 204−2, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.13: The baseline corrected Raman spectrum of sample number 204-2, i.e., a thin-film of carbon deposited on glass using magnetron sputteringat an RF power of 200 W for 60 minutes, with a 12.5 sccm argon gasflow. The two Gaussian peaks selected for this peak fitting process aredepicted, as is the sum. The resultant integrated intensities suggest thatID/IG = 2.91 ± 0.06. The fitting parameters are found to be AD = 1842and AG = 2172, mD = 1389 cm−1 and mG = 1594 cm−1, σD = 193.6 cm−1and σD = 56.15 cm−1. The root mean square error is 63.26. The electronicversion of this figure is in color.834.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 18000100200300400Raman shift (cm−1)Intensity (a.u.)Sample number 207−1, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.14: The baseline corrected Raman spectrum (with backgroundcompensation) of sample number 207-1, i.e., a thin-film of carbon depositedon glass using magnetron sputtering at an RF power of 200 W for 60 minutes,with a 3.5 sccm argon gas flow. The two Gaussian peaks selected for thispeak fitting process are depicted, as is the sum. The resultant integrated in-tensities suggest that ID/IG = 2.24 ± 0.05. The fitting parameters are foundto be AD = 290.6 and AG = 288.5, mD = 1356 cm−1 and mG = 1583 cm−1,σD = 178.4 cm−1 and σD = 80.2 cm−1. The root mean square error is 9.75.The electronic version of this figure is in color.844.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 1800050010001500200025003000Raman shift (cm−1)Intensity (a.u.)Sample number 210−1, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.15: The baseline corrected Raman spectrum of sample number210-1, i.e., a thin-film of carbon deposited on glass using magnetron sput-tering at an RF power of 200 W for 60 minutes, with a 25 sccm argongas flow. The two Gaussian peaks selected for this peak fitting process aredepicted, as is the sum. The resultant integrated intensities suggest thatID/IG = 3.01 ± 0.06. The fitting parameters are found to be AD = 1844and AG = 2092, mD = 1385 cm−1 and mG = 1594 cm−1, σD = 194.2 cm−1and σD = 56.6 cm−1. The root mean square error is 59.68. The electronicversion of this figure is in color.854.4. Summary of the Raman results1000 1100 1200 1300 1400 1500 1600 1700 1800010002000300040005000Raman shift (cm−1)Intensity (a.u.)Sample number 210−2, Raman source: 442 nm  baseline correctedG−Peak (sp2)D−Peak (sp3)Two peak Gaussian fitFigure 4.16: The baseline corrected Raman spectrum of sample number210-2, i.e., a thin-film of carbon deposited on a flat piece of titanium usingmagnetron sputtering at an RF power of 200 W for 60 minutes, with a25 sccm argon gas flow. The two Gaussian peaks selected for this peak fittingprocess are depicted, as is the sum. The resultant integrated intensitiessuggest that ID/IG = 2.88 ± 0.05. The fitting parameters are found tobe AD = 3209 and AG = 3802, mD = 1391 cm−1 and mG = 1591 cm−1,σD = 197.8 cm−1 and σD = 57.62 cm−1. The root mean square error is94.75. The electronic version of this figure is in color.864.4. Summary of the Raman resultsTable 4.3: Comparison of the Raman results obtained for the sp3 to sp2bonding ratio, i.e., ID/IG vs sp3 content. Reference sample number 201-2 is shaded. (* without background compensation, ** with backgroundcompensation; see Section 4.3.4).Sample Raman ID/IG sp3 AD mD σD AG mG σG Root mean Figure /No. source content square error Page106-2 442 nm 3.30 ± 0.07 0.77 ± 0.004 1992 1389 199 1936 1588 61.78 57.80 4.7 / 77111-1 442 nm 2.77 ± 0.06 0.73 ± 0.004 2506 1372 190.9 3228 1589 53.38 102.01 4.8 / 78115-3 442 nm 3.20 ± 0.06 0.76 ± 0.004 2507 1395 218.6 2952 1590 57.42 89.78 4.9 / 79121-3 442 nm 3.39 ± 0.06 0.77 ± 0.003 751.9 1400 218.1 761.7 1583 62.86 21.18 4.10 / 80123-3 442 nm 3.20 ± 0.07 0.76 ± 0.004 1128 1394 202.2 1201 1590 58.94 39.78 4.11 / 81201-2 442 nm 3.21 ± 0.07 0.76 ± 0.004 2042 1388 197 2174 1588 57.4 69.74 4.4 / 71201-2 633 nm 3.18 ± 0.10 0.76 ± 0.006 913.2 1339 170.8 702.6 1582 69.63 58.70 4.12 / 82204-2 442 nm 2.91 ± 0.06 0.74 ± 0.004 1842 1389 193.6 2172 1594 56.15 63.26 4.13 / 83207-1* 442 nm 3.35 ± 0.13 0.77 ± 0.007 203.2 1407 197.3 206 1585 57.81 11.80 4.5 / 73207-1** 442 nm 2.24 ± 0.05 0.69 ± 0.005 290.6 1356 178.4 288.5 1583 80.2 9.75 4.14 / 84210-1 442 nm 3.01 ± 0.06 0.75 ± 0.004 1844 1385 194.2 2092 1594 56.6 59.68 4.15 / 85210-2 442 nm 2.88 ± 0.05 0.74 ± 0.003 3209 1391 197.8 3802 1591 57.62 94.75 4.16 / 86874.4. Summary of the Raman resultsThe root mean-square error is defined asRMSE =√√√√ 1nn∑i=1wi(yˆi − yi)2, (4.5)where n is the number of data points, required to calculate the error minusthe number of fitted coefficients, yˆi is the ith fitted data point, and yi isthe ith true measurement data point. In the following figures, i.e., Figures4.17, 4.18, 4.19, 4.20, 4.21, 4.22, and 4.23, the Raman results are comparedwith each other. The aim is to compare samples with just one parameterchanged in the deposition parameters selected. Due to time limitations, itis not possible to fully probe the deposition parameter space. Nevertheless,a preliminary comparison, with significant differences in the Raman resultsobserved, can be presented. The sp3 to sp2 bonding ratio is notably higherfor higher argon gas flows. With a gas flow of 25 sccm, the sp3 to sp2 bondingratio is 3.01 ±0.06 compared with a ratio of 2.24 ±0.05 with a low gas flowof 3.5 sccm, as is seen in Figure 4.17. Schwan et al. [12] achieved a highsp3 content, up to 87 %, in thin-film carbon, using magnetron sputtering.The present work aims to replicate this same process and result. Therefore,a deposition pressure of 1 mTorr is targeted. Due to inaccuracies in thesputtering system used in this particular work, the argon gas flow had to beadjusted in order to achieve the desired deposition pressure. For experimentnumber 207-1, i.e., a thin-film of carbon deposited on glass using magnetronsputtering at an RF power of 200 W for 60 minutes, the argon gas flow was3.5 sccm and the sp3 to sp2 bonding ratio is 2.24 ±0.05. For experimentnumber 106-2, i.e., a thin-film of carbon deposited on a titanium ring usingmagnetron sputtering at an RF power of 200 W for 175 minutes, the argon884.4. Summary of the Raman results00.511.522.533.54I D / I GThin−film carbon samples25 sccm argon gas flow3.5 sccm argon gas flowFigure 4.17: Raman analysis comparison of sample number 210-1 on theleft and 207-1 (with background compensation) on the right. For both sam-ples: glass substrate, RF power of 200 W, 60 min sputtering, Raman source442 nm. The electronic version of this figure is in color.894.4. Summary of the Raman resultsgas flow was 4.7 sccm in order to achieve a deposition pressure of 1 mTorrand the corresponding sp3 to sp2 bonding ratio is found to be 3.30 ±0.07.Figure 4.18 shows a comparison of those two experiments. At this point, itis not possible to state if the substrate type or the deposition time is thecause for the differences in the nature of the chemical bonding.Very similar sp3 to sp2 bonding ratios are found for planar glass andplanar titanium substrate (sp3 to sp2 bonding ratio of 3.01 ±0.06 vs. 2.88±0.05, respectively), but with a higher argon gas flow, as depicted in Fig-ure 4.19. The implications are that differences in the type of the substratehave a small influence on the nature of the chemical bonding, but rather onthe deposition time. A more systematic study should be employed in orderto determine exactly the impact of the substrate geometry or the depositiontime.To verify the reproducibility of the employed sputtering deposition sys-tem, three independent runs are employed; see Figure 4.20. It seems like thesecond experiment has a higher sp3 to sp2 bonding ratio with just 2 minutesdifference in the deposition time. This difference in the sp3 to sp2 bondingratio is most probably due to variations in the system and not due to the 2minutes difference in the deposition time.In Figure 4.21, the variation of the sp3 to sp2 bonding ratio with respectto the Raman source is given. There is much more noise related to a Ramansource of 633 nm. However, the Raman result is very similar if a Ramansource of 442 nm is used.A proper consideration of the background for samples, prepared withunusually low argon gas flows, is very critical. As shown in Figure 4.22, the904.4. Summary of the Raman results00.511.522.533.544.5I D / I GThin−film carbon samplesGlass substrate60 min sputtering3.5 sccm argon gas flowTi ring substrate175 min sputtering4.7 sccm argon gas flowFigure 4.18: Raman analysis comparison of sample number 207-1 (glass sub-strate, 60 min sputtering, 3.5 sccm argon gas fow, with background compen-sation) on the left and 106-2 (titanium ring substrate, 175 min sputtering,4.7 sccm argon gas flow) on the right (similar gas flows produced the samevacuum pressure of about 1 mTorr). For both samples: RF power of 200 Wand Raman source 442 nm. The electronic version of this figure is in color.914.4. Summary of the Raman results00.511.522.533.54I D / I GThin−film carbon samplesGlass substrate Flat Ti substrateFigure 4.19: Raman analysis comparison of sample number 210-1 on theleft and 210-2 on the right: For both samples: RF power of 200 W, 60 minsputtering, 25 sccm argon gas flow, Raman source 442 nm. The electronicversion of this figure is in color.924.4. Summary of the Raman results00.511.522.533.544.5I D / I GThin−film carbon samples120 min sputtering118 min sputtering120 min sputteringFigure 4.20: Raman analysis comparison of sample number 115-3 on theleft, 121-3 in the middle and 123-3 on the right. For all samples: titaniumring substrate, RF power of 200 W, 12.5 sccm argon gas flow, Raman source442 nm, about 120 min sputtering. The electronic version of this figure isin color.934.4. Summary of the Raman results00.511.522.533.54I D / I GThin−film carbon samplesRaman source442 nm Raman source633 nmFigure 4.21: Raman analysis comparison of the reference sample 201-2 onthe left and also 201-2 on the right. For these samples: glass substrate, DCpower of 400 W, 60 min sputtering, 12.5 sccm argon gas flow. The electronicversion of this figure is in color.944.4. Summary of the Raman results00.511.522.533.544.5I D / I GThin−film carbon sampleswithoutbackground subtractionwithbackground subtractionFigure 4.22: Raman analysis comparison of sample number 207-1 on the leftand also 207-1 on the right. For these samples: glass substrate, RF power of200 W, Raman source 442 nm, 60 min sputtering, 3.5 sccm argon gas flow.The electronic version of this figure is in color.954.5. Raman interpretationdifference in the obtained Raman results is significant. It seems like the typeof power applied, i.e., RF of 200 W or DC of 400 W, does not effect the sp3to sp2 ratio, as depicted in Figure 4.23.Based on the spot tests, performed in this thesis, four basic trends areobserved; (1) it seems like RF power of 200 W with a medium amount ofgas flow of 12.5 sccm and a long deposition time of 120 minutes producesa high sp3 to sp2 bond ratio of 3.2 as experienced with sample numbers115-3, 121-3 and 123-3. (2) If the gas flow is not too high the adhesion isbetter on glass substrates and the thin-film carbon does not easily rub off thesubstrate surface and (3) low gas flows of 3.5 sccm on glass and 4.7 sccm ontitanium rings, respectively, produces very low adhesion, i.e., the thin-filmcarbon can easily be rubbed off. Finally (4), DC power seems to producea form of thin-film carbon with a deeper black color, probably due to thehigher power of 400 W that produces a thicker film.4.5 Raman interpretationMeans of interpreting the spectrum associated with a sample of thin-filmcarbon has been the subject of controversy over the years, with a varietyof different approaches being suggested [10, 29, 31, 37–39]. Raman spec-troscopy relates to the interaction of the source light with the vibrationalmodes, for the material under consideration. In essence, the source lightexcites these vibrational modes, and the energy is lost due to the emissionof phonons, i.e., the quanta of vibration, is what is measured, this corre-sponding to the observed shifts in the Raman spectrum. Accordingly, as964.5. Raman interpretation00.511.522.533.54I D / I GThin−film carbon samplesDC power of 400 WRF power of 200 WFigure 4.23: Raman analysis comparison of the reference sample 201-2 onthe left and 204-2 on the right. For both samples: glass substrate, 60 minsputtering, 12.5 sccm argon gas flow, Raman source 442 nm. The electronicversion of this figure is in color.974.6. Film thickness and material composition profilethe vibrational character of a material may be expressed in terms of thenature of the chemical bonds that are present, from a chemist’s perspective,the Raman results may be accounted for in terms of the chemical bonds thatare present. While this may lead to an oversimplification in the analysis, andwhile the finer details associated with the Raman spectrum are beyond thescope of this elementary point of view, it provides a very intuitive perspec-tive of the source of the peaks that are observed in the Raman spectrum.The most common approach is to fit the Raman data with two Gaussiancurves and create the ratio of the integrated peak intensities for a measureof the sp3 to sp2 ratio. Another way to interpret the Raman data, employedby some in the literature [29, 38], is to determine the ratio of the intensitypeak values. Table 4.4 lists some literature references with their publishedsp3 to sp2 ratio and stated approach for the Raman data interpretation,along with an estimate of the peak ratio and integrated intensity ratios, i.e.,area ratio, respectively, based on their published figures. This compilationshows that most researchers use the ratio of the integrated peak intensitiesand therefore, this approach is used in this work as well.4.6 Film thickness and material compositionprofileIn order to determine the thickness of a magnetron sputtered sampleof thin-film carbon, a SEM is employed. While other approaches may beemployed, the SEM offers an ideal opportunity to study the obtained filmson a micron scale. Furthermore, it may be used in circumstances where984.6. Film thickness and material composition profileTable 4.4: A comparison of common Raman interpretations found in the lit-erature. Six different Raman studies on thin-film carbon and a verificationof their resultant ratio of sp3 to sp2 bonds. The peak ratio and integratedintensity, i.e., area, ratio was estimated from the presented figures and com-pared with the claimed sp3 to sp2 ratio. The last column draws a conclusionhow the Raman spectra might have been interpreted.Author / Year ID/IG ID/IG ID/IG Peak /Peak Area Claim AreaZhao et al. [10] / 2008 0.468 0.776 0.76 AreaTai et al. [29] / 2006 0.556 0.851 1.72 PeakMa et al. [37] / 2002 0.732 1.32 1.3 AreaSung et al. [31] / 1997 Sample 1: 1.28 Sample 1: 4.85 Sample 1: 2.9 AreaSample 2: 0.54 Sample 2: 2.16 Sample 2: 1.7Sheeja et al. [38] / 2001 1.68 3.89 1.76 PeakZhang et al. [39] / 2000 Sample 1: 0.581 Sample 1: 1.81 Sample 1: 1.8 AreaSample 2: 0.722 Sample 2: 2.19 Sample 2: 2.1994.6. Film thickness and material composition profiletraditional means of determining film thicknesses are frustrated by practi-cal constraints; determining the film thickness on, for example, rings. Inaddition, the use of the SEM for this purpose, while admittedly overblown,offers other tangential pieces of information that allow one to glean furtherinsights into the nature of these films. In order to demonstrate the use of anSEM for determining the thickness of the thin-film carbon on a substrate,the sample number 115-1, i.e., a thin-film of carbon deposited on a titaniumring using magnetron sputtering at an RF power of 200 W for 120 minutes,with a 12.5 sccm argon gas flow, is considered. The primary challenge is todetermine where the transition from substrate (titanium ring) to the thin-film carbon occurs. It is critical to analyze the thin-film perpendicular toits surface, as rings possess cylindrical geometry, that pose challenges forplanar focusing. Figure 4.24 provides a representative view of this sampleof thin-film carbon on a titanium ring from a certain angle. The materialline capabilities of the SEM machine available at The University of BritishColumbia allow one to distinguish the thin-film from the substrate. In Fig-ure 4.25, a zoomed-in version of the line, depicted in Figure 4.24, is shown.The material content across this line is given in the sub-figure immediatelybelow. From this information, it is clear that the dark section in the imagecorresponds to the thickness of the thin-film carbon, because the contentof carbon increases linearly due to the angular point of view. Another ap-proach is to determine the thickness of the thin-film carbon based on thismaterial line analysis. The carbon content increases from around 5.5 upto around 9.5 µm. Since this view is from an angle, the thickness appearsto be bigger than in actual reality. Therefore, 4 µm must be too much, as1004.6. Film thickness and material composition profileFigure 4.24: An SEM picture of sample number 115-1, i.e., a thin-film ofcarbon deposited on a titanium ring using magnetron sputtering at an RFpower of 200 W for 120 minutes, with a 12.5 sccm argon gas flow. Originaland colored picture (colors are artificial). Image obtained with the assistanceof Mr. D. Arkinstall. The electronic version of this figure is in color.1014.6. Film thickness and material composition profileFigure 4.25: Material line analysis of sample number 115-1, i.e., a thin-filmof carbon deposited on a titanium ring using magnetron sputtering at anRF power of 200 W for 120 minutes, with a 12.5 sccm argon gas flow. Imageobtained with the assistance of Mr. D. Arkinstall. The electronic version ofthis figure is in color.1024.6. Film thickness and material composition profileshown with a visual inspection in Figure 4.26. Furthermore, the transitionis fringed due to insufficient polishing. For future measurements, this hasto be improved. Nevertheless, the order of magnitude for the thickness isconsistent.The thickness can be more accurately identified by tilting the sampleuntil the surface is in line with the viewer. Based on the resolution ofthe microscope image picture, the thickness of the thin-film carbon can bemeasured; see Figure 4.26. For sample number 115-1, the thickness is to beabout 1.7 µm.Another feature of the SEM machine is the possibility of measuring thecontent of the material. A high energy electron ionizes electrons from theinner shell of the material, and when they transition back to their normalstate, an X-ray is emitted. The SEM has an X-ray detector, and is thereforeable to determine the number of X-rays and their energy. The programmatches those energy peaks with known elemental standards. The amountthat is being presented normalized to 100% of the total measured X-rays.Therefore, the x-axis of the spectra images is in keV (each material emits adifferent energy). Figure 4.27 shows the material analysis spectrum 14; seeFigure 4.24 on sample number 115-1, i.e., a thin-film of carbon depositedon a titanium ring using magnetron sputtering at an RF power of 200 Wfor 120 minutes, with a 12.5 sccm argon gas flow. The titanium content is64.5 wt% and shows clearly that this part of the sample is the substrateitself, i.e., the titanium ring. Spectrum 16 in Figure 4.28 shows a titaniumcontent of just 8.6 wt%, and a carbon content of 73.8 wt% and confirms theexistence of thin-film carbon. There is also a fair amount of oxygen evident1034.6. Film thickness and material composition profileFigure 4.26: SEM: Thickness of sample number 115-1, i.e., a thin-film ofcarbon deposited on a titanium ring using magnetron sputtering at an RFpower of 200 W for 120 minutes, with a 12.5 sccm argon gas flow. Imageobtained with the assistance of Mr. D. Arkinstall. The electronic version ofthis figure is in color.1044.6. Film thickness and material composition profileFigure 4.27: Material spectrum 14 corresponding to thin-film carbon sput-tering sample number 115-1, i.e., a thin-film of carbon deposited on a ti-tanium ring using magnetron sputtering at an RF power of 200 W for 120minutes, with a 12.5 sccm argon gas flow. (Spectrum 14 in Figure 4.24).1054.6. Film thickness and material composition profileFigure 4.28: Material spectrum 16 corresponding to thin-film carbon sput-tering sample number 115-1, i.e., a thin-film of carbon deposited on a ti-tanium ring using magnetron sputtering at an RF power of 200 W for 120minutes, with a 12.5 sccm argon gas flow. (Spectrum 16 in Figure 4.24).1064.6. Film thickness and material composition profilein both spectra. As presented in Chapter 2, in Section 2.6, the presenceof oxygen can effect the adhesion immensely. Since the thin-film carbonassociated with most of the samples does not adhere very well, and is prettyeasy to buff or polish off, the conclusion can be made that contaminationand the oxygen layer will probably prevent good adhesion properties.With this X-ray detector, it is not possible to detect the amount ofhydrogen, because hydrogen has just one valence electron. The amount ofhydrogen in the thin-film carbon, produced in this thesis, is assumed to below, since the samples were prepared with magnetron sputtering, and nohydrogen based gases were employed.107Chapter 5ConclusionsMagnetron sputtering provides an inexpensive means of depositing thin-films of carbon using widely available equipment. In this thesis, experimentswere performed on thin-films of carbon prepared using magnetron sputter-ing. The resultant thin-films were then examined using Raman and SEManalyzes. Four distinct goals were accomplished as a result of this body ofwork. First, the production of thin-film carbon, using this technique, wasdemonstrated, the resultant sp3 to sp2 bonding ratio being comparable tothat observed in other forms of high-quality ta-C [12]. Second, the thin-filmswere characterized using Raman spectroscopy, the sp3 to sp2 bonding ratioscorresponding to these thin-films being determined using the experimentalobtained spectra corresponding to these thin-films. Third, through the useof a variety of deposition conditions, the deposition parameter space wassampled, the resultant forms of thin-film carbon being examined throughthe use of Raman spectroscopy in order to evaluate the corresponding sp3to sp2 ratio. Further, a means of determining the thin-film carbon profiles,through the use of an SEM analysis, was developed.The sp3 content, achieved in the deposited thin-films of carbon, wasfound to reach up to 77%. Therefore, the thin-films of carbon producedin this work can be almost considered a form of ta-C, which requires an108Chapter 5. Conclusionssp3 content minimum of 80% [9]; ta-C is viewed as being the best qualitytype of thin-film carbon. Unfortunately, a fair amount of oxygen, between11 to 15%, is found within these samples of thin-film carbon. As oxygendetracts from good adhesion, it is expected that the adhesion of these filmsis limited, although this has yet to be put to the experimental test. Ahigh argon gas flow of 25 sccm is found to produce thin-films of carbonwith a significantly higher ratio of sp3 to sp2 bonding ratio, i.e., 3.01 ±0.06,than with a low argon gas flow of 3.5 sccm, where the sp3 to sp2 bondingratio is found to be 2.24 ±0.05. With a low gas flow of around 3.5 sccmor 4.7 sccm (a process pressure of 1 mTorr was employed), the sp3 to sp2bonding ratio is different from glass substrate to titanium ring substrate fordifferent deposition times, namely 2.24 ±0.05 for 60 minutes and 3.30 ±0.07for 175 minutes, respectively. Further experiments are required in order toinvestigate if the type of substrate, or the duration of the deposition, isthe reason for the differences in the sp3 to sp2 bonding ratio. Not muchdifference in the sp3 to sp2 bonding ratio was detected between planar glassand planar titanium substrate for high argon gas flows of 25 sccm. TheRaman source, of either 442 or 633 nm, seems not to have a significantimpact on the sp3 to sp2 ratio, even though the noise related to a Ramansource of 633 nm is much bigger; the sp3 to sp2 bonding ratio of 3.21 ±0.07vs. 3.18 ±0.10. It was also discovered, that the background of samples,prepared with a low argon gas flow, has to be considered in the Raman dataanalysis. Without background subtraction, the sp3 to sp2 bonding ratio is3.35 ±0.13, whereas with background subtraction it is 2.24 ±0.05.There are a number of important matters that could be further studied109Chapter 5. Conclusionsrelated to this topic. First, the analysis presented herein focused on contrast-ing samples prepared through a variety of different means with each other.A systematic probing of the deposition parameter space was not performed,owing to time limitations. This is a critical deficiency of this work. How theselection of substrate influences the properties of the obtained films, i.e., beit the material employed (glass or titanium) or the shape of the substrate(ring or planar), should also be examined in a more systematic manner.Finally, how the sample of thin-film carbon that have been fabricated forthe purposes of this particular analysis contrast with those associated withother research groups, would also be worth of further investigation. Theseimportant topics, that build up on the results presented in this thesis, willbe addressed at a later point. Ultimately, these future studies will allowone to answer as to whether or not magnetron sputtering is adequate forcosmetic jewelry applications. This thesis represents the fist step in thisdeeper research problem.110References[1] H. Karamitaheri, “Thermal and thermoelectric properties of nanostruc-tures,” Ph. D. Dissertation, 2013.[2] S.-C. Seo, “The characterization of diamond-like carbon films depositedusing unbalanced magnetron sputtering at Ohio Universtity,” Ph. D.Dissertation, 1996.[3] D. Saada, “Ion implantation into diamond and the subsequent graphi-tization,” Ph. D. Dissertation, 2000.[4] J. B. Hannay, “On the artificial formation of the diamond,” Proceedingsof the Royal Scociety of London, vol. 30, no. 200-205, pp. 450–461, 1879.[5] (2014, November). [Online]. Available:http://www.diamondlab.org/80-hpht synthesis.htm[6] K. Iakoubovskii, M. Baidakova, B. Wouters, A. Stesmans, G. Adri-aenssens, A. Vul’, and P. Grobet, “Structure and defects of detonationsynthesis nanodiamond,” Diamond and Related Materials, vol. 9, pp.861–865, 2000.[7] S. Aisenberg and R. Chabot, “Ion-beam deposition of thin films of111References Referencesdiamondlike carbon,” Journal of Applied Physics, vol. 42, no. 7, pp.2953–2958, 1971.[8] A. Grill, “Diamond-like carbon: state of the art,” Diamond and RelatedMaterials, vol. 8, pp. 428–434, September 1998.[9] J. Robertson, “Diamond-like amorphous carbon,” Materials Scienceand Engineering R, vol. 37, pp. 129–281, 2002.[10] D. C. Zhao, N. Ren, Z. J. Ma, G. J. Xiao, and S. H. Wu, “Study on themechanic properties of DLC films on different metal substrates,” KeyEngineering Materials, vol. 373-374, pp. 117–121, 2008.[11] K. Bobzin, N. Bagcivan, N. Goebbels, and K. Yilmaz, “Effect of thesubstrate geometry on plasma synthesis of DLC coatings,” Plasma Pro-cesses and Polymers, vol. 6, no. 1, pp. S425–S428, 2009.[12] J. Schwan, S. Ulrich, H. Roth, H. Ehrhardt, S. R. P. Silva, J. Robertson,R. Samlenski, and R. Brenn, “Tetrahedral amorphous carbon films pre-pared by magnetron sputtering and dc ion plating,” Journal of AppliedPhysics, vol. 79, no. 3, pp. 1416–1422, 1995.[13] K. Wojciechowski, R. Zybala, R. Mania, and J. Morgiel, “DLC lay-ers prepared by the PVD magnetron sputtering technique,” Journalof Achievements in Materials and Manufacturing Engineering, vol. 37,no. 2, pp. 726–729, 2009.[14] M. Maharizi, O. Segal, E. Ben-Jacob, Y. Rosenwaks, T. Meoded,N. Croitoru, and A. Seidman, “Physical properties of a:DLC films and112References Referencestheir dependence on parameters of deposition and type of substrate,”Diamond and related materials, vol. 8, pp. 1050–1056, 1998.[15] J. Schwan, S. Ulrich, V. Batori, H. Ehrhardt, and S. R. P. Silva, “Ramanspectroscopy on amorphous carbon film,” Journal of Applied Physics,vol. 80, no. 1, pp. 440–447, 1996.[16] L. F. Bonetti, G. Capote, L. V. Santos, E. J. Corat, and V. J. Trava-Airoldi, “Adhesion studies of diamond-like carbon films deposited onTi6Al4V substrate with a silicon interlayer,” Thin Solid Films, vol. 515,pp. 375–379, September 2006.[17] S. Neuville and A. Matthews, “A perspective on the optimisation ofhard carbon and related coatings for engineering applications,” ThinSolid Films, vol. 515, pp. 6619–6653, June 2007.[18] A review of Raman for multicomponent analysis, vol. 9129. SPIE,2014.[19] J. G. Buijnsters, P. Shankar, W. Fleischer, W. J. P. van Enckevort, J. J.Schermer, and J. J. ter Meulen, “CVD diamond deposition on steelusing arc-plated chromium nitride interlayers,” Diamond and RelatedMaterials, vol. 11, pp. 536–544, 2002.[20] E. Gariboldi, “Drilling a magnesium alloy using PVD coated twistdrills,” Journal of Materials Processing Technology, vol. 134, pp. 287–295, 2002.[21] M. Smietana, M. L. Korwin-Pawlowski, W. J. Bock, G. R. Pickrell, andJ. Szmidt, “Refractive index sensing of fiber optic long-period grating113References Referencesstructures coated with a plasma deposited diamond-like carbon thinfilm,” Measurement Science and Technology, vol. 19, 2008.[22] Low Reflectance DLC Coatings on Various IR Substrates, vol. 8353.SPIE, P.O. Box 10, Bellingham, WA 98227-0010, United States, 2012.[23] M. Allen, B. Myer, and N. Rushton, “In vitro and in vivo investiga-tions into the biocompatibility of diamond-like carbon (DLC) coatingsfor orthopedic applications,” Journal of Biomedical Materials Research,vol. 58, pp. 319–328, 2001.[24] R. A. Street, Hydrogenated amorphous silicon. Cambridge UniverstityPress, New York, 1991.[25] A. Wu, J. Sun, X. Shen, N. Xu, Z. Ying, Z. Dong, and J. Wu, “Diamond-like carbon thin films prepared by ECR argon plasma assisted pulsedlaser deposition [online],” Diamond and Related Materials, vol. 15, pp.1235 – 1241, 2005.[26] C. Kittel, Introduction to Solid State Physics. John Wiley and Sons,Inc., Berkeley, 2005.[27] N. W. Khun, E. Liu, and M. D. Krishna, “Structure, adhesive strengthand electrochemical performance of nitrogen doped diamond-like car-bon thin films deposited via dc magnetron sputtering,” Journal ofNanoscience and Nanotechnology, vol. 10, no. 7, pp. 4752–4757, 2010.[28] G. Speranza and L. Minati, “Characterization of C-based materials:The evaluation of sp2 and sp3 hybrids,” Diamond and Related Materi-als, vol. 16, pp. 1321–1324, January 2007.114References References[29] F. C. Tai, S. C. Lee, C. H. Wei, and S. L. Tyan, “Correlation betweenID/IG ratio from visible Raman spectra and sp2/sp3 ratio from XPSspectra of annealed hydrogenated DLC film,” Materials Transactions,vol. 47, pp. 1847–1852, July 2006.[30] A. C. Ferrari and J. Robertson, “Interpretation of Raman spectra ofdisordered and amorphous carbon,” Physical Review B, vol. 61, no. 20,pp. 14 095–14 107, Mai 2000.[31] S. Sung, X. Guo, K. Huang, F. Chen, and H. Shih, “The strengtheningmechanism of DLC film on silicon by MPECVD,” Thin Solid Films,vol. 315, pp. 345–350, September 1997.[32] J. Robertson, “Properties of diamond-like carbon,” Surface and Coat-ings Technology, vol. 50, pp. 185–203, August 1992.[33] M. Cekada, M. Kahn, P. Pelicon, Z. Siketic, I. B. Radovic, W. Wald-hauser, and S. Paskvale, “Analysis of nitrogen-doped ion-beam-deposited hydrogenated diamond-like carbon films using ERDA/RBS,TOF-ERDA and Raman spectroscopy,” Surface and Coatings Technol-ogy, vol. 211, pp. 72–75, September 2011.[34] J. Buijnsters, P. Shankar, W. van Enckevort, J. Schermer, and J. terMeulen, “The effect of nitriding on the diamond film characteristicson chromium substrates,” Diamond and Related Materials, vol. 11, pp.1760–1768, June 2002.[35] J. E. Lee, “A quantitative analysis of tunable long period grating tech-nology and its application,” Ph. D. Dissertation, 2007.115[36] S. D. Szymanski, “Argon plasma cleaning of fluorine, organic and oxidecontamination using an advanced plasma treatment system,” MarchPlasma Systems, Concord, California, U.S.A., Tech. Rep., 2007.[37] F. Ma, Q. Chen, X. Cai, G. Li, and H. Ma, “DLC film fabricated bya composite technique of unbalanced magnetron sputtering and PIII,”Materials Transactions, vol. 43, pp. 1398–1402, 2002.[38] D. Sheeja, B. Tay, L. Yu, and S. Lau, “Low stress thick diamond-like carbon films prepared by filtered arc deposition for tribologicalapplications,” Surface and Coatings Technology, vol. 154, pp. 289–293,December 2001.[39] Q. Zhang, S. Yoon, J. Ahn, Rusli, H. Yang, B. Gan, C. Yang, F. Watt,E. Teo, and T. Osipowice, “Low stress thick diamond-like carbon filmsprepared by filtered arc deposition for tribological applications,” Dia-mond and Related Materials, vol. 9, pp. 1758–1761, 2000.116Appendix A: Step by stepguide for the Hummer XIIFigure A.1 provides a step-by-step guide for the operation of the HummerXII, in terms of thin-film carbon depositions.117Appendix A: Step by step guide for DLC magnetron sputtering with theHummer XIIShut%down%Spu+ering%RF%cleaning%Startup%Fit$rings$onto$spindle$Close$chamber$properly$Turn$on$nitrogen$and$argon$gas$Plug$in$both$power$cables$Turn$on$cooling$water$Turn$on$roughing$pump$Turn$on$turbo$pump$at$2.0$x$10>4$Pump$down$to$1.3$x$10>4$Set$argon$gas$flow$to$75$sccm$Turn$on$RF$power$$(200$W$/$13.56$MHz)$AOer$25$min:$turn$power$off$Reduce$argon$gas$flow$to$20$sccm$Cool$down$to$about$40˚$C$Set$argon$gas$flow$to$12.5$sccm$Pre>spuSering$for$about$2$min$Open$gun$shuSer$SpuSering$process$Turn$off$power$supply$Turn$off$turbo$pump$AOer$15$min:$turn$off$roughing$pump$Turn$off$nitrogen$and$argon$gas$VenWng$of$the$chamber$FigureA.1:Stepbystepguideforthin-filmcarbondepositionusingtheHummerXII.118

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0074399/manifest

Comment

Related Items