1 THE ROLE OF WATER ON THE TRIBOLOGICAL PROPERTIES OF MOLYBDENUM DISULPHIDE FILMS By RACHEL SARAH COLBERT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENT S FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Rachel Sarah Colbert
3 To my m om d ad and s ister
4 ACKNOWLEDGMENTS I would like to thank Professor W. Gregory Sawyer for his help and guidance throughout this process, without which this dissertation would not have been possible. I would also like to thank Dr. Michael T. Dugger for his mentorship, guidance and support throughout the last year of this process. I would also like to thank the other members of my committ ee, Prof. David Hahn, Prof. Scott Banks, and Prof. Scott Perry, friendship and collaboration with the members of the University o f Florida Tribology Laboratory, e specially Dr. Nicolas Argibay who has played a crucial role in the development of this dissertation. I am also very appreciative of all the support Andy Korenyi Both of Tribologix Inc. has been by providing numerous coating batches that ha ve enabled the c ompletion of this work. I would like to acknowledge the contributions of numerous staff members at Sandia National Laboratories for their input and help developing this work. In particular Robert Sorenson and Samuel Lucero for the ir advice and the use of the QCM equipment, Robert Grubbs for the use of his tube furnace, Ronald Goeke for his invaluable advice about molybdenum disulphide, Michael Brumbach for his XPS work, and Richard Grant for his SEM work I would like to thank my parents, Susan M. D ubinsky and Richard W. Colbert. Without their support and encouragement none of this would have been possible. I would also like to thank my sister, Jessica A. Worthington, for her love and support.
5 TABLE OF CONTENTS page A CKNOWLEDGMENTS ................................ ................................ ................................ 4 LIST OF TABLES ................................ ................................ ................................ ........... 7 LIST OF FIGURES ................................ ................................ ................................ ........ 8 LIST OF ABBR EVIATIONS ................................ ................................ .......................... 11 ABSTRACT ................................ ................................ ................................ .................. 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ... 14 Extraterrestr ial Environment ................................ ................................ .................. 15 History of MoS 2 as a Solid Lubricant ................................ ................................ ...... 16 2 THEORETICAL BACKGROUND ................................ ................................ ........... 23 Contact Area, Friction, and Wear ................................ ................................ ........... 23 Contact Area ................................ ................................ ................................ ... 23 Revolute Joints: Misalignment and Clearances ................................ ............... 26 Friction ................................ ................................ ................................ ............ 28 Wear Regimes ................................ ................................ ................................ 31 Wear Rates ................................ ................................ ................................ ..... 32 Quartz Crystal Microbalance Theory ................................ ................................ ...... 36 Molybdenum Disulphide Theoretical Background ................................ .................. 39 Lubri cious Properties of MoS 2 ................................ ................................ ......... 39 Water Vapor Affects ................................ ................................ ........................ 40 Solid Lubricant Dopant Additions to MoS 2 ................................ ............................. 45 Graphite ................................ ................................ ................................ .......... 47 Diamond like Carbon ................................ ................................ ...................... 50 Antimony Trioxide ................................ ................................ ........................... 52 3 EXPERIMENTAL SETUP ................................ ................................ ...................... 54 Coatings of Interest ................................ ................................ ............................... 54 Experimental Equipment ................................ ................................ ........................ 55 In Situ Linear Reciprocating Tribometer ................................ .......................... 55 Quartz Crystal Microbalance ................................ ................................ ........... 57 Experimental Setup ................................ ................................ ............................... 57 Varying Humidity Friction and Wear Experiments ................................ ............ 58 Steady State Wear Experiments ................................ ................................ ..... 60 Quartz Crystal Microbalance Experiments ................................ ....................... 61
6 Room temperature sorption ................................ ................................ ...... 61 Thermally activated desorption ................................ ................................ 63 Microscopy ................................ ................................ ............................... 65 4 EXPERIMENTAL RESULTS ................................ ................................ ................. 67 In Situ Wear Experiments ................................ ................................ ...................... 67 Steady State Wear Rate Experiments ................................ ................................ ... 70 Quartz Crystal Microbalance Experiments ................................ ............................. 71 Step Changes in Relative Humidity Experiments ................................ ............. 71 Inert Heating Experiments ................................ ................................ ............... 78 5 DISCUSSION ................................ ................................ ................................ ........ 84 Evidence of Water Absorption ................................ ................................ ............... 84 Diffusivity of Water in Molybdenum Disulphide Films ................................ ............. 85 Half Time Method ................................ ................................ ............................ 89 Moment Method ................................ ................................ .............................. 90 Limiting Slope Method ................................ ................................ ..................... 91 Diffusivity Results ................................ ................................ ............................ 93 Thermally Activated Desorption ................................ ................................ ............. 95 Continuous Mass Loss Oxidation? ................................ ............................... 95 Ratio of Water to MoS 2 in the Film ................................ ................................ .. 97 Hypothesis of Water Diffusion Locations ................................ ............................... 98 6 CONCLUSIO NS ................................ ................................ ................................ .. 100 APPENDIX A IMPINGEMENT AND OXIDATION RATES ................................ .......................... 102 B FRACTIONAL INCREASE IN OXIDATION DUE TO MASS PER UNIT AREA LOSS AT 170C ................................ ................................ ................................ .. 104 C MOLAR RATIO CALCULATIONS ................................ ................................ ........ 105 FUNDING SOURCES ................................ ................................ ................................ 107 LIST OF REFERENCES ................................ ................................ ............................ 108 BIOGRAPHICAL SKETCH ................................ ................................ ......................... 117
7 LIST OF TABLES Table page 3 1 List of the coat ings, their manufacturers, and the experiments they were used in. ................................ ................................ ................................ ...................... 54 5 1 Mass per unit area room temperature sorption values for both the thick and thin MoS 2 films, corresponding to the data sh own in Figure 4 10. ..................... 85 5 2 Calculated diffusivity values for the MoS 2 films. ................................ ................. 94 5 3 Maximum thermally desorbed mass per unit area values at 110 C .................. 97
8 LIST OF FIGURES Figure page 1 1 Typical mission critical tribological components in a satellite system that require low friction and low wear to successfully deploy. ................................ ... 15 1 2 The atomic arrangement of molybdenum disulphide which was adapted from Fleischauer et al.. ................................ ................................ .............................. 18 2 1 The difference between real and apparent area of contact. ............................... 24 2 2 The difference in the number of cycles to failure for coated standard, radiused, and chamfered bushing. ................................ ................................ .... 27 2 3 Models of surfaces in contact showing the real area of contact under elastic conditions which w as adapted from Archard ................................ ................... 30 2 4 The r elationship between crystal cut and frequency versus temperature of quartz crystals adapted from Gerber et al.. ................................ ........................ 38 2 5 The effect relative humidity has on the friction coefficient of MoS 2 ................... 41 2 6 Sorption and desorption of water vapor at 30C on MoS 2 powder adapted from Johnston and Moore ................................ ................................ ................. 43 2 7 Sorption and desorption o f water vapor at 30 C on MoS 2 powder after being degassed at 350 C and 950 C adapted from Johnston and Moore. ................ 44 2 8 The effect of exposure time on the coefficient of friction in both 50% a nd dry nitrogen environments ................................ ................................ ...................... 45 2 9 The effect of load and relative humidity on the friction coefficient adapted from Winer. ................................ ................................ ................................ ........ 46 2 10 Crystal structure of graphite adapted from Lansdown. ................................ ....... 47 2 11 The environmental dependence on the friction coefficient of graphitic films reproduced from Yen ................................ ................................ ........................ 49 2 12 The effect cycling the relative humidity from dry (0% RH) to 40% RH has on the friction coefficient which is reproduced from Voevodin et al.. ....................... 50 2 13 The effect cycling the relative humidity from dry (0% RH) to 40% RH has on the friction coefficient for a Al 2 O 3 /DLC/Au/MoS 2 film which is adapted from Baker. ................................ ................................ ................................ ................ 51 3 1 A schematic of the situ linear reciprocating tribometer with scanning w hite light objective is shown. ................................ ................................ .................... 56
9 3 2 Representative SWLI line scans taken at cycle 0 and cycle 2000 which are used to determine the area loss. ................................ ................................ ....... 59 3 3 Experimental setup for step change in relative humidity experiments. ............... 62 3 4 Experimental setup for thermal desorption experiments. ................................ ... 63 4 1 Friction coefficient and volume loss of the coatings are plotted versus normal force times sliding distance for all cycled environment experiments initiated in laboratory air. ................................ ................................ ................................ .... 68 4 2 Friction coefficient and volume loss of the coatings are plotted versus normal force times sliding distance for all cycled environment experiments initiated in dry nitrogen. ................................ ................................ ................................ ...... 69 4 3 Friction coefficient and wear rate data for 10,000 cycle experiments ................. 71 4 4 Mass per unit area data for representative step change in relative hu midity data for the 0.1 m thick pure MoS 2 film. ................................ ........................... 73 4 5 Mass per unit area data for representative step change in relative humidity data for the 0.1 m thick MoS 2 /Sb 2 O 3 /Au film.. ................................ .................. 73 4 6 Variability between three crystals coated in the same batch at similar relative humidity step changes. ................................ ................................ ...................... 74 4 7 Surface area variability be tween rough and smooth surfaces. ........................... 75 4 8 SEM micrographs of the two pure MoS 2 samples. ................................ ............. 76 4 9 Mass per unit area data for represent ative step change in relative humidity data for the 0.9 m thick MoS 2 film. ................................ ................................ ... 76 4 10 Water sorption isotherms for both 0.1 m and 0.9 m MoS 2 films and the 0.1m MoS 2 /Sb 2 O 3 /Au film. ................................ ................................ ............... 77 4 11 Relative changes in mass per unit area are measured before and after each heat/cool cycle up to 110 C for the thin MoS 2 film.. ................................ .......... 79 4 1 2 Relative changes in mass per unit area are measured before and after each heat/cool cycle up to 110 C for the thin MoS 2 /Sb 2 O 3 /Au film.. .......................... 79 4 13 Relative changes in mass per unit area are mea sured before and after each heat/cool cycle up to 170 C for the thin MoS 2 film. ................................ ........... 80 4 14 Relative changes in mass per unit area are measured before and after each heat/cool cycle up to 170 C f or the thin MoS 2 /Sb 2 O 3 /Au film.. .......................... 80
10 4 15 Relative changes in mass per unit area are measured before and after each heat/cool cycles up to 110 C and 170 C for the 0.9 m MoS 2 film. .................. 81 4 16 XPS spectral region for Mo 3p 3/2 used to determine the amount of oxidation via the ratio of Mo oxide to sulfide for the 0.1 m MoS 2 films. ............................ 83 5 1 The differential ratio for various length scales between the coated rough and polished crystals. ................................ ................................ ............................... 86 5 2 Example of calculating the time when M t /M is 0.5 by the half time method. ..... 90 5 3 Example of calculating s by the moment method. ................................ ............. 91 5 4 Example of calculating R by the moment method. ................................ ............. 92 5 5 A comparison of the diffusivity of water into various MoS 2 films ........................ 93
11 LIST OF ABBREVIATION S AFM Atomic Force Microscope H 2 O Water HRSEM High resolution scanning electron micr oscopy LPM Liter per minute Mol/area Moles per unit area MoS 2 Molybdenum d isulphid e p pm P arts per million PTFE Polytetrafluorethylene QCM Quartz crystal microbalance RH Relative humidity Sb 2 O 3 Antimony t rioxide SEM Scanning electron microscope SWLI Scannin g white light interferometer XPS X ray photoelectron spec t roscopy
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE ROLE OF WATER ON THE TRIBOLOGICAL PROPERTIES OF MOLYBDENUM DISULPHIDE FILMS By Rachel Sarah Colbert December 2012 Chair: Wallace Gregory Sawyer Major: Mechanical Engineering Molybdenum disulphide (MoS 2 ) based a dversely affected by the addition of water into the environment. The friction coefficient is known to increase with increasing relative humidity and decrease with decreasing relative humidity, which is a repeatable process. There still remains a gap in t he literature on the e ffects water has on the wear of MoS 2 films and the quantities of water up take that are associated with changes in the relati ve humidity of the environment. Both of these topic s are addressed in this work. The first studies examined the friction and wear of self mated composite MoS 2 thin films in a cycled relative humidity environment. The surface topography was evaluated as the environment was cycled through wet and dry conditions and the steady state wear rates in each environment were examined. It was found for all films the wear increased in the higher humidity environment and decreased in the dry nitrogen environment. The second portion of this work analyzed the quantity of water ab sorbed in both a pure MoS 2 and composite MoS 2 film at various relative humidities. It was found that
13 there were two distinct types of water uptake occurring, water that requires additional energy for desorption to occur and water absorbed reversibly at room temperature. The final portion of this wo rk comments on the diffusion kinetics occurring within the film. Diffusivity values for water into molybdenum disulphide films were calculated and the location of the water during the two observed sorption processes was h ypothesized.
14 CHAPTER 1 INTRODUCTI ON Sol id lubricant can be defined as a material that provides lubrication to two relatively movin g surfaces under dry conditions [ 1 ] This is generally a thin film of a low shear strength solid that lowers the friction coefficient and helps to prevent damage between the mating surfaces. Commonly used solid lubricants include graphite, molybdenum disulphide (MoS 2 ), and p olytetrafluoroethylene (PTFE). Although there is 2 and graphite dating back for centuries, the more modern uses of these solid lubrica nts dates back only to the 1930 s. There are four main forms of using a so lid lubricant: 1) loose powder rubbed onto the mating surface, 2) bonded materials attached to the surface with a binder or the use of sputtering deposition, 3) additives to greases and oils, or 4) self lubricating materials for devices such as bearings. S olid lubricants are primarily used when greases and oil lubricants cannot be used. These environments include extreme temperatures (both high and low), high vacuum, and areas of intense radiation [ 1 2 ] The benefit of solid lubricants is that they can perform in these extreme environments as well as more moderate environments. The main advantag e for solid lubricants is their stability at extreme temperatures and in irradiated environments. They are also known to be of lighter weight than oil systems and allow for placement of mechanisms closer to heat sources thus decreasing shaft lengths in as semblies. Solid lubricants also keep surfaces separated during slow sliding under high loads [ 1 ] While the advantage s are numerous, there are disadvantages to solid lubricants as well. Generally, the friction coefficients can be higher than with hydrodynamic
15 lubrication [ 1 ] Additionally, they have a finite wear li f e while oils and greases can be replenished into contacts. The wear debris is also a concern as it can contaminate otherwise clean surfaces such as satellite solar arrays Extraterrestrial Environment One of the main tasks for solid lubricants lies in moving mechanical assemblies within the aerospace industry. Spacecraft have numerous mission critical assemblies with tribological components, some of which are depicted in Figure 1 1 These components must operate in the extreme environment of space and survive terrestrial testing as well. The space environment is known for thermal cycling, ultra high vacuum, solar ultra violet radiation, atomic oxygen, and debris all of which have the potential to cause issues. The ultra high vacuum is between 10 8 t orr to 10 13 t orr depending on the orbital height. The temperatures are thought to range between 200 C up to 100 C depending on the amount of direct exposure to the sun. Figure 1 1. Typical mission critical tribological components in a satellite system that require low friction and low wear to successfully deploy.
16 Atomic oxygen is the most prevalent molecular species in lower Earth orbit and is generated by ultraviolet radiation breaking down molecular oxygen [ 3 ] Banks et al showed the highly destructive nature which oxidizes and degrades the surfaces of the coatings [ 4 ] The flux of atomic oxygen incident on a forward facing surface, as a result [ 2 3 ] These extreme environments imposed on the lubricants in space are only a portion of the conditions they must survive. The difficulty comes from the multiple environments these components must endure. Most mechanisms undergo numerous life cycle testing on ground in a relative humidity (RH ) prior to launch, then must survive the launch process, and finally operate as designed in space. Within all these environments low friction and low wearing properties are required for a successful operation to occur. History of MoS 2 as a Solid Lubrican t Molybdenum disulphide (MoS 2 ) remains one of the most common solid lubricants in use in the aerospace industry because of its low friction behavior and its low wear. Molybdenum disulphide is a naturally occurring ore which is commonly referred to as molyb denite. It is a dark blue grey or black solid which feels slippery to the touch [ 5 ] The ease of transfer to most solid surfaces and difficulty to remove it explain s its like. Both of these words referred to graphite, lead, molybdenum disulphide or other various ores up to the 1700s. There was a lack of con tinuity in the written records differentiating between these solids, in particular between graphite and molybdenum disulphide.
17 [ 5 ] the earliest written account that identifies molybdenum disulphide as a lubricant was written by Cramer in 1764 [ 6 ] Cramer [ 6 ] describes a black lead substan ce that is heavier than glimmer stones, feels much like soap, and lubricates surfaces when rubbed against them. It was also described as a substance that is black and shiny and upon heating loses its color and becomes softer. Lansdown [ 5 ] emphasizes two points that validate this referring to MoS 2 and not graphite. The first is that glimmer stones refer to micas which have specific gravities similar to graphite, respec tively 2.7 3.3 versus 2.0 2.3, while molybdenum is quite heavier with specific gravities on the order of 4.6 4.8. Thus, the mention of heavier glimmer stones must be MoS 2 The second point cites the change in color that occurs molybdenum in the literature. By 1916, molybdenum disulphide had become readily available commercially. Around the same ti me its crystal structure was established by Dickinson and Pauling [ 7 ] which is shown in Figure 1 2 but it still was not of great interest to the lubrication community. Awareness of molybdenum disulphide increased in the 1930s wh en technical investigations by industrial organizations began. Numerous patents started to surface containing molybdenum disulphide as either an additive, a film adhered to the surface with a binder, or production of a synthetic version. Finally in 1941, Bell and Findlay published one of the first technical papers which discussed the use of MoS 2 as a lubricant for bearings in high vacuum conditions [ 8 ] Studies by the US National Advisory Committee for Aeronautics (NACA) began in the 1940s and the first technical note on MoS 2 was published in 1948 by Johnson et al
18 [ 10 ] Solid surface films of molybdenum disulphide on steel were analyzed for contact pressure and sliding speed e ffects on the friction. Johnson et al c oncluded that MoS 2 was very effective in reducing friction at high sliding velocities. This film was chemically and thermally stable, and should have many practical applications [ 10 ] Experimental studies on MoS 2 performed by NACA and its successor, the National Aeronautics and Sp ace Administration (NASA), greatly increased the usage of MoS 2 Figure 1 2. The atomic arrangement of molybdenum disulphide which was adapted from Fleischauer et al. [ 7 9 ] The military began to use molybdenum disulphide in 1950 and the first military specification for dry powder lubrication was issued in 1952 [ 5 ] The initial applications were for anti seizure or anti galling purposes on non critical components. By 1965, the range of MoS 2 uses had grown and there were nine U.S. and five British military
19 specifications covering molybde num disulphide based powders, greases, bonded films, and anti seizure compounds [ 5 ] Technical studies were continuing and publications o f the environmental e ffects o n molybdenum disulphide films began to appear in the late 1950s. Johnson and Vaughn studied the effects of orientation of the crystallites in the sliding surface and adsorption of vapor by the MoS 2 surface. Johnson et al which explained the increase in friction coefficient after sliding was halted for a period of time [ 10 ] After every test stoppage the friction would increase, but eventually run back into a simi lar steady state value as before. This was the first discussion on how water vapor affects MoS 2 films. In 1966, Haltner and Oliver studied the effect of water vapor on the friction coefficient of a MoS 2 film [ 11 ] They confirmed that as water vapor concentrations were increased, the fric tion of MoS 2 increased, yet for graphite the friction coefficient decreased under the same conditions [ 11 12 ] An explanation for the effect of humidity on MoS 2 friction could not be thoroughly explained at this point in time. During the 1950s and 1960s the effect of temperature on Mo S 2 coatings was investigated as well. Both oxidation and sublimation temperatures were examined. Cannon discussed the lack of agreement in the research community on the melting point of MoS 2 with reported values ranging from 1300 1700 C [ 13 ] After additional testing was done, the melting point was not confirmed, but sublimation was noted to have occurred at 1250 C. Godf rey et al and Bisson et al showed that oxidation becomes appreciable above 360 C but rapidly increases above 560 C when tested in ambient environments [ 14 1 5 ] When coatings were heated in nitrogen or hydrogen
20 environments they could withstand temperatures up to 650 C and still provide satisfactory lubrication [ 16 17 ] By 1967, a new technique for applying a thin MoS 2 film was established and discussed by Spalvi ns et al [ 18 ] Prior to this time, MoS 2 was adhered through the use of binders and burnishing methods, with both techniques having some major downfalls. Binders had the potential to be undesirable due to potential chemical interactions that ed films those in which the s urface is rubbed in a particular orientation in order to align the surface layers, lack ed uniformity, tend ed to have a higher friction coefficient, and had adherence issues to the substrate. Spalvins et al established a vacuum deposition method, also kno wn as physical direct current sputtering [ 18 19 ] The films produced by this method showed an increase in wear life as well as a more stable fricti on response for the life of the films on varying substrates. The cleanliness of the substrate prior to deposition was emphasized as an important factor on performance of the film. Over the next decade the analysis of MoS 2 focused on film development, e xplanations of run in perio ds, and additional temperature a ffects. Hopkins and Campbell analyzed twenty two MoS 2 based films at temperatures ranging from 70 C up to 200 C [ 20 ] Not one film performed well at all temperatures and the conclusion was made that further developments must occur but until the n films must be matched to the specific conditions of interest. While the crystal structure of MoS 2 was investigated in the 1930s, its lubricating mechanisms explaining the run in period of the friction coefficient was under investigation during the 1970s During the early portion of this
21 mechanical stress and to form homogeneous smooth films on both sliding surfaces was discovered [ 21 22 ] Developments in MoS 2 films were important during this time. Three types of sputtering were used to produce satisfactory films in the early 1980s: radio frequency sputtering, direct current diode sputtering, and direct current triode assisted sputtering. Three essential features for a satisfactory per formance of a sputtered MoS 2 film include 1) a good substrate to achieve interfacial adhesion, 2) proper chemical composition with stoichiometry as near as possible to natural MoS 2 and 3) production of a stable structure of the films [ 23 ] Metal dopants were used to strengthen the sputtered films, as well as aid in its crystal growth and stabilize the film structure [ 23 ] New film development c ontinued to thrive in the 1990s. Modifications to the films to increase performance were done by layering or co depositing metals which promotes the formation of MoS 2 with basal plane orientation and low porosity [ 23 27 ] The tribological performance is dependent on the sputtering conditions which control the microstructure and thus control s the crystallinity, morphology, and composition [ 28 ] Spalvins and Roberts summarized various sputtering conditions that were shown to of the substrate, deposition rate, partial pressure of water, structure of the film, and final film thickness [ 29 2 ] It was noted that the adherence of the film depends s trongly on the preparation and cleanliness of the substrate surface prior to deposition [ 29 ] The deposition rate and amount of water vapor have an impact on the stoichiometry of the film; a high deposition rate and low partial pressures of water vapor allow for stoichiometric films to be deposited. If a lower rate of deposition and a higher partial pressure of water exist then
22 the films become sulfur deficient and contaminated with oxygen [ 2 29 ] The structure of sputtered MoS 2 can be varied from crystalline to amorphous with substrate temperature changes. Ambient and elevated temperatures are preferred to allow columnar structure of crystalline growth in which low shear basal planes are aligned perpendicular to the substrate surface. This structure has the tendency to not be fully dense but can be densified by low pressure deposition, ion beam bombardment durin g film growth, co sputtering with alloy dopants, or multilayer deposition. For a low friction and acceptable wear life coating an ideal film thickness between 0.2 0.6 m was suggested by Spalvins in 1992 [ 29 ] By the late 1990s, molybdenum disulphide films had been successfully used in space for many years [ 30 35 ] The friction behavior and wear perf ormance of MoS 2 were still greatly dependent on the testing environment. Thus, modifications to these films to increase performance were done by layering or co depositing metals in attempts c ments [ 23 27 ] These adaptations and additions to the coatings were made to decrease the sensitivity to atmospheric conditions. Research in the field of coating development is still an ongoing process. While there have been numerous studies of various environmental e ffects on molybdenum disulphide coatings, the role water plays on the lubricity and wear is still not fully understood. The focus of this research is to analyze the e ffect water has on MoS 2 films to gain a better understanding of the wear, associated sorption isotherms and diffusion mechanisms into the film.
23 CHAPTER 2 THEORETICAL BACKGROU ND Before understanding the origin of low friction and wear exhibited by this material, the fundamental theories associated with friction and wear must be understood. This chapter discusses a survey of topics devoted to understanding interactions between contacting bodies and the general theories applied for this research The theories associated with contact area, friction, and wear are d iscusse d in the first section. The second section goes over the theory of a quartz crystal microbalance. The final section discusses these theories as applied to molybdenum di sulphide. Contact Area, Friction, and Wear Contact Area Comprehending the difference between the real area of contact and apparent area of contact is crucial within tribological systems. While the ideal case of topographically smooth contacting bodies do es exist on rare occasions, there is in fact a significant difference between the apparent and real area of contact between two mated surfaces. This difference in contact areas stems from the fact that surfaces, whether polished or not, are all relatively rough and on a microscopic scale exhibit roughness or asperities. The apparent area of contact is the amount of area where contact would exist if ideally smooth surfaces were mated. The real area of contact is the area of the asperities of the two surfa ces that are in contact with one another which is only a fraction of the apparent area of contact. The distinction between these two areas of contact is shown in Figure 2 1. Hertz was the first to describe the elastic deformation of two contacting solid s urfaces as a function of the applied load [ 36 ] Hertzian contact stress refers to a
24 localized stress that develops as two curved surfaces of varying radii come into contact and deform slightly under the load imposed. The amount of deformation that occurs is dependent upon the modulus of elasticity of the mated materials, the geometry of the shown in Equation 2 1 which i s dependent on the normal force ( ), combined radius ( ), and the combined modulus of elasticity ( ). Figure 2 1. The difference between real and apparent area of contact. ( 2 1 ) As can be seen from the equation abov e, there exists a power law relationship between the real area of contact and the applied force, (Hertz). elastic limit was always exceeded by local stresse s on the asperity level and thus plastic deformation occurs [ 37 ] Even though the pressure over the apparent a rea of contact may be considered elastic, the force is distributed over the minimal contacting asperities of the real area of contact which plastically deform due to these applied pressures. The
25 real contact area is related to the applied load ( ) hardness ( ) but is independent of geometry for this approach. Holm found the expression for contact area assuming only plastic deformation of the surfaces which is shown in Equation 2 2 ( 2 2 ) In the 1950s Archard published an article [ 38 39 ] theory that contact between surfac es results in solely plastic deformation of asperities. Archard made the point that it is reasonable to assume plastic flow for the first cycles of the bodies contacting, but it is illogical to assume plastic flow for the entirety of a life cycle test [ 39 38 ] The point was made that the asperities may be rubbed many times without being damaged, but when plastic deformation doe s occur, the asperities must reach a steady state in which the load is supported elastically [ 38 39 ] On a microscopic scale are better equipped to describe the deformation that is occurring to form the contact area between the surfaces. Macroscopically, for a pin on disk experiment, the two surfaces can be modeled as an ideally smooth ball mated with a smooth flat surface, pressures are used due to their simplicity and popularity by providing researchers with a nominal pressure value associa ted with the system of interest [ 40 ] While, these Hertzian pressures are not necess arily the most accurate predictors of the actual pressures associated with the system, they are a good estimate to understand the magnitude of pressures being imposed.
26 Revolute Joints: Misalignment and Clearances There are additional contact mechanics to c onsider in conjunction with the contact area considerations for tribological applications. Appropriate constraints and degrees of freedom are provided with the tribological interface for most systems and especially in bushings [ 41 42 ] The design of a shaft revolving in a slightly larger journal or bushing is subject to evolutions of geometry, contact area, and pressure distribution which in turn affect the evolution in friction coefficient and torque. While this is of particular importance for bushing systems it also affects other contacting areas as well. Colbert et al. showed the importance of edge geometry on bushing systems [ 43 ] Edge geometry was studied by examining unmodified, chamfered, and radiused edges of bushings. All bushings for th is study were coated with MoS 2 /Au/Sb 2 O 3 coatings approximately 2 m thick and were tested in high vacuum environments. The friction coefficient versus sliding distance is shown for all bushing types in Figure 2 2 and a value greater than =0.27 was establ ished as coating failure. It was found that between two different standard bushings there was a three order of magnitude difference in wear life, but the variations in wear life were negligible in comparison to the chamfered and radiused bushings. Inte rrogation of the standard bushings wear scars revealed a large contrast in the behaviors of nominally identical samples. This was most likely due to concentrated loading that occurred from slight sample differences in the bushings and shaft misalignments. The sample that had almost immediate failure showed a thin wear scar at the edge of the bushing which implies a concentrated load at one edge of the bushing which caused the quickened failure of the film and thus steel on steel contact. The wear scar of the second bushing was biased towards one edge as well, but to a lesser extent.
27 This difference in behavior between the precision machined shafts and bushings implies a severe sensitivity to system misalignment. Figure 2 2. The difference in the number of cycles to failure for coated standard, sliding distance for nominally identical standard bushings coated with commercially available solid lubricant coatings. B) The cycles to failure are bushings. The contrasting frictional behavior result from a strong sensitivity to angular misalignment between the bushing and shaft. This figure is reproduced from C olbert et al. [ 43 ] The chamfered and radiused bushings were tested under identical conditions and showed a substantial improvement in wear life. As can be seen from Figure 2 2b the chamfered bushing failed after 35,000 sliding cycles and the radiused bushing ran for over 60,000 cycles without failure. Over constraints are inevitable in application, but this data shows the importance of edge modifications to mitigate adverse effects of potential misalignments in systems. Contact mechanics play an important role in all tribological systems. The contact area of the system affects the contact pressures applied to the system which in turn can affect the elastic/plastic regime being imp osed on the tribological system. While the roughness of the associated system is important, sample design plays a vital role as
28 well. Simple adjustments of the design, like applying a radius to an edge can improve the life of a system by orders of magnit ude. Friction Reduction of friction has been a topic of concern for centuries whenever mobile parts have been involved in the design of transportation, machines, or mechanical tools . When sliding surfaces come into contact the friction between them arises from two main areas. The first factor is adhesion which occurs at the real area of contact [ 44 45 ] The second factor arises from ploughing, grooving, or cracking of asperities on the mated surface. The first studies on friction coefficients were done by Leonardo da Vinci in the late 1400s in which he distinguished between sliding and rolling friction. Since his findings were only recorded in his journals and never published, the credit for the bas ic laws of dry friction went to Guillaume Amontons over 200 years later in 1699 [ 5 ] These basic laws of dry friction were stated as: 1. The friction force is independen t of the apparent area of contact 2. The friction force is proportional to the applied load Friction coefficient, 2 3 as a dimensionless ratio of friction force ( ) between the mated bodies and the appl ied normal force ( ) ( 2 3 ) In the early 1900s the frictional force was shown to be proportional to the real area of contact [ 46 48 ] The real area of contact is affected by the material s deformation
29 regime whether it is plastic, elastic, or a combination of the two. This is discussed previously in the contact area section. [ 38 44 49 ] If a single asperity is assumed to be in contact and the deformation is elastic, the area of contact, fr om Equation 2 1, is proportional to the normal force to the two thirds power, From the definition of the friction force in Equation 2 4 whic h is shown to be a function of and the effective shear strength, the proportionality between the normal force and friction force is found and shown in Equation 2 5. ( 2 4 ) ( 2 5 ) This proportionality leads to the follow ing. ( 2 6 ) This result does not portray the friction coefficient as constant and thus is inconsistent the load c an be defined as a function of and the flow pressure, ( 2 7 ) From this definition the friction coefficient is defined as ( 2 8 ) coef ficient will remain constant.
30 When multiple asperities are in contact on a surface under an assumption of solely elastic deformation, the load is now distributed amongst them and there is an associated area of contact for each asperity as well. Archard modeled a multiple asperity contact as spherical protuberances evenly distributed over a spherical surface which can be seen in Figure 2 3 a [ 44 ] The assumption was made that all spherical asperities were identical and thus the area of each contact was alike. The real area of contact was found to have the following proportionality. ( 2 9 ) Fig ure 2 3. Models of surfaces in contact showing the real area of contact under elastic conditions which was adapted from Archard [ 49 ] The relationship between the normal force, and the real area of contact, are as follows: A) ( B ) ( C ) ( D) (E ) ( F ) Archard continued this study by examining additional surfaces of varying asperity types. Figure 2 3 b takes the initial surface and superimposes smaller asperities. This trend of increa sing superposition of even smaller protuberances across the surface is
31 shown in Figure 2 3 c f. The proportionalities between the real area of contact and the normal force, listed in Figure 2 3 caption, become increasingly closer to unity. Archard thus c oncluded that in any real situation where elastic deformation is occurring the area of contact will be proportional to where x lies between 2/3 and 1 [ 44 45 ] This indicates that for multiple conta ct cases, even when the deformation is elastic, the real area of contact tends to be almost linearly proportional to the normal force which Wear Regimes A general definition of wear is the loss of material as loose particles from solid surfaces as a result of mechanical action [ 50 ] Severe wear is generally thought to be the cause of failure for thin films. Abrasive and adhesive wear are the two main types ass ociated with coatings. Abrasive wear occurs when a hard rough surface comes into contact with a softer asperity or surface and it results in deformation of the softer material. It is also known to remove material through ploughing, cutting, or scratching of the surface. There are two main types of abrasive wear, two body and three body. When hard debris particles are generated and they adhere to the surface that is known as two body abrasive wear. Third body abrasive wear occurs when these hard particu lates are not constrained and are free to roll and slide in the contact. Fatigue induced cracking is at times considered abrasive wear, but is limited to brittle materials which are not addressed by this research. Adhesive wear occurs when two surfaces come into contact and material transfer from one surface to another occurs due to localized bonding between the contacting asperities. This bonding can be due to cold welding or galling of the surfaces. When metallic bonding between atoms on opposing sur faces occurs, this is known as cold
32 welding. Galling on the other hand occurs from excessive friction between asperities results in localized welding which subsequently causes metal tear out or metal transfer to occur between the mated surfaces [ 51 ] While in both cases seizure of the surfac es occurs, a distinction between the two is important. The main difference is that contaminant free surfaces are required for cold welding to occur, while galling is caused by high friction between the mated surfaces. In space conditions both of these s eizures must be accounted for du ring the design process. While surfaces are not likely to become sufficiently clean for cold welding to occur during a practical mission lifetime, galling caused by loss of the lubricant is more likely. It is for this reas on that design engineers are faced with the task of over designing for multiple lifetimes to limit the possibility for failure caused by galling. Adhesion is one of the largest contributors to friction and wear of coatings. As discussed previously, surf aces of most materials are rough on the microscopic scale thus causing load between the two surfaces to be supported only at the points where the asperities touch. The real area of contact is thus minimal and the resulting pressures generated at these con tacts are high, often surpassing the yield stress of the material. These high pressures cause adhesion of the contacting asperities which in turn requires a greater lateral force, or frictional force, to allow sliding to occur [ 5 ] Wear Rates The measurement of wear is typically presented as a wear rate and may involve a single measurement at the end of an experiment or multiple measurements throughout the experiment Wear data is influenced by inherent material properties, testing conditions, surface preparation, and measurement methods [ 52 ] There are four
33 established methods to obtain wear rate values: i) analysis of the spherical cap, ii) linear displacement, iii) mass loss measurements, or iv) area loss measurements. Mass loss measurements are considered impractical for solid lubricant thin films due to material transfer that occurs from the coating to the mated surface. Similarly, linear displacement measurements of the amount of material worn from the pin do not apply either. The concept of analyzing the wear on a spherical cap of t he pin is possible, but requires an estimate of an average radius which increases the uncertainty in the measurement. The final method is an area loss measurement which is the primary methodology for these studies. Area loss measurements are done with e ither a scanning white light interferometer or profilometer. In order to estimate the wear volume with this method, the technique discussed by Williamson and Hunt [ 53 ] and Sayles [ 54 ] was employed. The technique evaluates a surface before and after deformation has occurred. Thus, an initial surface scan of the virgin material can be compared to a surface scan of the wear scar after wear has occurred and an area loss measurement can be obtained. Multiple m easurements can be made throughout the wear track to compute a best estimate of the average area loss Thus the average area can be extrapola ted over the entire length of the wear scar and a volume loss can be estimated. Archard proposed that the total wear volume is proportional to the real area of contact times the sliding distance [ 49 ] From this postulation, a wear coefficient, was introduced that has dimensionless proportionality between the real area of contact, sliding distance, and the volume of material lost, which is shown in Equation 2 10. ( 2 10 )
34 This wear coefficient is often used as an index of the severity of wear occurring in systems [ 55 ] The lower the value of the smaller the amount of wear prod uced. The specific wear rate, is a similar term to the wear coefficient, but not dimensionless which is shown in Equation 2 11. ( 2 11 ) Both of these wear descriptors are used frequent ly in literature and thus attention must be paid as to the units being addressed. The specific wear rate, with units of (mm 3 /Nm) will be the focus of the research discussed in this paper. The uncertainty associated with the wear rate is dependent upon the volume loss, normal force, and sliding distance uncertainties. The normal force uncertainty depends on the method of measurement. If the force is measured then the uncertainty associated will be dependent upon the force transducer used. If a dead weigh t loading system is in place, then the uncertainty will be associated with the calibration of the load used. For sliding distance, the main computational methods for linear stages are either the use of an LVDT or encoder. With rotary stages either an enc oder is used or the encoders, their associated uncertainty is dependent upon the uncertainty of the equipment. The uncertainty with rotary stages measuring distance from s peed and time measurements depends upon the uncertainty in the duration of the test, angular speed, and wear track radius [ 52 ] The uncertainty in volume loss according to Burris and Sawyer [ 52 ] dominate all other values. The volume loss values depend greatly on the variations in the wear track. Colbert et al describe the variability in volume loss measurements as far more significant than the uncertainties of the instruments
35 themselves [ 56 ] It is for this reason that the recommendation is to measure multiple scans throughout the wear track to account for potential variations. A useful measurement of uncertainty should con sider the number of scans taken, the variation between those scans, and the geometry of the experimental system [ 56 ] This met hod of calculating uncertainty is derived by Colbert et al and the final definition is shown below in Equation 2 12. ( 2 12 ) According to Colbert et al t he uncertainty in the volume is dependent upon the radius of the wear tra ck, variation of the scanned areas, and and the number of scans, utilized. In many instances the uncertainty of the radius, is significantly smaller than the actual radius value and thus second term of Equation 2 12 can be neglected From this knowledge, the equation can be simplified and is defined in Equation 2 13 below. ( 2 13 ) Equation 2 13 demonstrates that the uncertainties in the volume loss depend on the variation in the wear track and the number of scans utilized to measure the volume loss. A general trend of the uncertainty diminishing as more scans are performed was found. In some instances multiple scans throughout the wear track cannot be taken and the uncertainty in the instrumentati on used must be applied. While this is not an ideal case, it provides a bound to the uncertainty expected for such measurements. For this case the uncertainty in volume loss is shown in Equation 2 14 and is dependent upon
36 the total number of pixels, I, d imensions of pixels, x and y, and the surface height after cycle n at pixel i, h i,n (2 14) As the uncertainty in h i,n and h i,n+m is identical, the uncertainty in the volume loss equation can be simplified to that shown in Equation 2 15. (2 15) Quartz Crystal Microbalance Theory In order to study quantities of water up take, a quartz crystal microbalance (QCM) was necessary. Quartz crystal microbalances are used to measure the change in resonant frequency which can be dir ectly related to the change in mass per unit area. Techniques can be applied in both liquid and gaseous environments, but for the purposes of these tests, a gaseous environment is the only one of interest. The QCM technology relies on the piezoelectric pr operties of quartz. The piezoelectric effect was first discovered by Jacques and Pierre Curi e [ 57 ] The Curie brothers discovered when a mechanical stress is applied to the surfaces of certain crystals this resulted in a n electrical potential across the crystal. This allows for the acoustic resonance to be probed by electrical means which has led to the development of QCM systems as well as other standing acoustic wave devices The QCM is co mprised of a thin quartz crystal, or other piezoelectric material, which is sandwiched between two electrodes A n alternating current is applied between these two electrodes which produces vibrational motion at the crystal s resonant frequency. As mass is added to the crystal, a frequency shift occurs Typical
37 QCM systems are sensitive to the addition of less than 1 ng/cm 2 of mass and have a measurement range of resonant frequencies between 5 and 30 MHz. Sauerbrey was the first to recognize the sensitivity of these piezoelectric devices towards mass changes. He developed a method to correlate the mass change per unit area at the QCM surface to the measured change in oscillation frequency of the crystal. From these studies the change in mass per unit area, is related to the change in frequency, initial resonan t frequency, the density of the quartz crystal, and the shear modulus of the quartz crystal, T his relation is known as the Sauerbrey equation which is defined in Equation 2 1 6 ( 2 1 6 ) In order for the Sauerbrey equation to remain valid the additional mass must be treated as an extension of the thickness of the crystal and the following three criteria must be met: the additional mass must be rigid, th e additional mass must be distributed uniformly, and the change in frequency divided by the characteristic frequency, must be less than 0.02. If these conditions cannot be met, the Sauerbrey equation does not hold and other meth ods must be applied to account for acoustic impedance due to either environmental effects or viscoelastic mass being added to the crystal For the research described in the next sections the mass changes are assumed to be ideal masses that follow the thre e criteria and thus the Sauerbrey equation is valid.
38 Figure 2 4. The relationship between crystal cut and frequency versus temperature of quartz crystals adapted from Gerber et al. [ 58 ] Temperature has a strong influence on the resonating frequency of the crystals. The frequency versus temperature char acteristics of the crystal are affected by the cut of the crystal. The characteristic curves for AT, DT, SL, and XY cut crystals are shown in Figure 2 4 The plot shows the change in frequency with respect to the re s onant frequency versus the temperature difference from 25C. AT cut crystals are known to have less of a temperature dependence around room temperature, as can be seen from
39 the curve and is the primary reason for their popularity In order to account for the a temperature change a third orde r polynomial fit to heating/cooling frequency can be used to correct for the changes due to thermal gradients Molybdenum Disulphide Theoretical Background Molybdenum disulphide is the main focus of this research An understanding of its lubrication mecha nisms and how various environments have been shown to influence the tribological properties will be discussed. Lubricious Properties of MoS 2 The lubricious nature of molybdenum disulphide has long been known, but the reason for this behavior has been under debate for years. In 1967 Winer summarized the three main areas of thought on the mechanism for this low friction behavior as: 1) the easy cleavage which provides low friction coefficient, 2) the low friction is dependent on adsorbed molecules between cr ystallites, and 3) the frictional behavior is dominated by the forces acting between individual crystallites with cleavage taking place along grain boundaries as opposed to between the layers in an individual crystal [ 59 ] The mechanism which is most supported in the literature is the first which bases the frictional properties off of the nature of the crystal structure. This mechanism will be the main discussion o n the lubricity of mo lybdenum disulphide. The suggestion of easy cleavage of MoS 2 occurring due to its crystal structure was first made by Dickinson and Pauling [ 7 ] They noted that the intrinsic crystal structure was made up of lamina. Further studi es have since been performed and molybdenum disulphide has been shown to be hexagonal in nature with a layer lattice structure. Each sulfur atom is equidistant from three molybdenum atoms and each molybdenum atom is equidistant from six sulfur atoms whic h form a triangular prism
40 structure [ 5 21 22 ] Strong cov alent bonding occurs between the metal molybdenum with a high polarizing influence and the non metal sulfur which is easily polarized [ 22 ] These tri layer planes are stacked with larger distances between the adjacent sulfur layers than the thickness of the layers themselves with weak attractive van der Waals forces between them [ 21 22 ] These van der Waals forces allow for relatively weak shear forces to be required to displace lamella with respect to one another thus allowing for the low friction properties. There are three main conditions which must be satisfied for molybdenum disulphide films to exhibit low friction coefficients. First is the formation of a well adhered, uniform transfer film on the mated material. The second requirement is the reorientation of the (0001) basal planes of the MoS 2 grains within the interface to be parallel to the sliding direction. This reorientation must occur on the surface of the film the transfer film, and the third bodies formed as well [ 60 ] The last requirement is the lack of contaminants such as oxygen, carbon, and water. MoS 2 films are known to be contaminated by oxygen as molybdenum oxide or by substitution of oxygen atoms in the lamellar structure [ 61 ] Carbon contamination largely comes from the sputtering deposition process and is usually a surface contaminant of the films [ 60 ] Finally, the contaminant of the biggest concern is water. Water Vapor A ffects The role of water vapor has been a prominent topic of research in the molybdenum disulp hide community since the 1950s. As coatings were tested in varying relative humidity, the trend of increasing friction coefficient with increasing relative humidity was found [ 5 62 12 63 64 ] There reaches a point where the friction no longer
41 increases (about 40% 65% RH Figure 2 5 a and 2 5 b ), and after that point the there is a potential for the friction to decreas e which is shown in Figure 2 5 a Figure 2 5 The effect relative humidity h as on the friction coefficient of MoS 2 A ) adapted from Peterson and Johnson [ 64 ] and B ) Pritchard and Midgley [ 63 ] One theory for the rise in fric tion coeffici ent was presented by Pardee [ 65 ] in which oxidation of the surface of molybdenum disulphide films to molybdenum oxides is the cause of this rise in friction. The slight reduction in friction at the highest humidity was proposed to occur from the formation of hydrogen sulphide as a by product of the oxidation process whi ch reacts with the substrate to form a sulphide film thus providing the reduction in friction [ 65 ] Another explanation for increased friction with humidity was proposed by Gao et al [ 66 ] It was suggested that in humid environments the ease of interlamellar slip is decreased for molybdenum disulphide. The a b sorption of water vapor was theorized to soften the films thus allowing for deformation with plowing to occur, which in turn disorients the film and increases the friction.
42 A distinction between adsorption and absorption must be discussed. Adsor p tion is the adhesion of atoms or mol ecules to a surface. Absorption refers to a process in which the atoms or molecules permeate the surface into the bulk. For this discussion of molybdenum disulphide, the results leave questions as to which is occurring. Holinski and Gansheimer [ 22 ] theorized that the increase in friction coefficient with water present in the environment is due to the p olarity of the sul f ur surface which allows the water molecules to be easily adsorbed onto a run in MoS 2 film. The water molecules bond with hydrogen bonding and thus connect molybdenum d isulphide layers which causes the increase in friction coefficient. The amount of hydrogen bonding is proportional to the amount of s orbed water which is related to the humidity of the environment. Johnston and Moore studied the sorption properties of MoS 2 powder samples [ 67 ] The authors found that there was a reversible process of water up take with the MoS 2 powder sample which they describe as an adsorption process. The data can be seen in Figure 2 6 It was suggested that this reversible process, which increased rapidly at higher relative humidities, could suggest capillary condensation of water molecules well. Johnston and Moore [ 67 ] further studied the properties of MoS 2 sorption after degassing the sample at either 350 C or 950 C which can be seen in Figure 2 7 It was noted that aft er the first adsorption/desorption run, there was an irreversible amount of water that was sorbed onto the powder samples. When the second adsorption/desorption run was performed the data followed the curve of the first
43 desorption run. This again showed the reversibility of water sorption onto the powder samples after a quantity of water was irreversibly sorbed. Figure 2 6. Sorption and desorption of water vapor at 30C on MoS 2 powder adapted from Johnston and Moore [ 67 ] A transient effect in moist conditions was discovered by Haltner in which the coefficient of friction was shown to increase when the sample was exposed for a period of time [ 68 ] The test was started and a run in of the frictio n coefficient was observed, but when sliding was stopped for a period of time and then motion was resumed, the friction coefficient was shown to have increased but upon sliding dropped back down to the steady state value shown previously. After the stoppa ge time there was an increase in the friction coefficient. A similar experiment was performed in a dry nitrogen environment and for this test the friction was shown to remain at the run in value after a period of no motion This data is shown in Figure 2 8 This indicates that having moisture in the environment potentially contaminates the surface of the MoS 2 films causing the increase in friction coefficient after a period of exposure time.
44 Figure 2 7. Sorption and desorption of water vapor at 30 C o n MoS 2 powder after being degassed at 350 C and 950 C adapted from Johnston and Moore [ 67 ] Research on the effects of moisture on MoS 2 coatings expanded to includ e speed, temperature, and load a ffects. Friction was found to decrease with increased sliding speed [ 10 62 68 ] Pri t chard et al. found a decrease in friction with increasing temperature and attributed it to decreasing moisture content in the film from the temperature increase [ 63 ] This theory was app lied to both the sliding speed a ffects and the effects of load shown by Barrwell as well as Milne and Gansheimer that friction de creased with increasing load [ 69 70 ] Figure 2 9 provides an example of the loading effects on MoS 2 films with varyi ng relative humidity environments. For both speed and load a ffects frictional heating is increased with increasing speed or increasing load which affects the moisture content of the film. Barry and Binkelman [ 71 ] made similar conclusions with respect to load and speed; friction was found to decrease with both increasing load and speed only when moistu re was present in the environment. When
45 the same tests were performed in a dry environment, the variation in friction with respect to speed and load was almost nonexistent. These results substantiated P r i theory of lowering the moistu re content in the film lowers the friction resistance. Figure 2 8. The effect of exposure time on the coeff icient of friction in both A) 50 % relative humidity and B) dry nitrogen environments adapted from Haltner 1964 [ 68 ] Overall, water e ffects and oxidation are still a large topic of res earch in the molybdenum disulphide community. A complete understanding of the role water plays on the film is still to be determined. What is known from the literature is that by introducin g moisture into the environment the friction coefficient of both pure and composite MoS 2 films will increase and when moisture is removed the friction coefficient decreases again. Studies of the sorption isotherms on MoS 2 powder demonstrated that after a residual quantity of water was sorbed onto the powder, the adsorp tion/desorption process was reversible as well. Solid Lubricant Dopant Additions to MoS 2 Once the effects of moisture on MoS 2 were beginning to be understood, the research community began to examine ways to limit the oxidation of the films. The
46 main ar ea of focus became the formation of composite coatings to attempt to improve the structure of the coating, increase the density, and provide greater environmental stability. While numerous dopants have been attempted with MoS 2 the main ones of interest i n this study are graphite (Gr), diamond like carbon (DLC), antimony trioxide (Sb 2 O 3 ), and gold (Au). Figure 2 9. The effect of load and relative humidity on the friction coefficient adapted from Winer [ 59 ]
47 Graphite Graphite has been in use as a solid lubricant for over 2000 years and until the mid 1900s was the most extensively used solid lubricant [ 5 ] The appearance and tribological characteristics of graphite are quite similar to molybdenum disulphide which explains the lack of differentiation between the two in early written records. Both materials are dark grey black in color and have hexa gonal lamellar, layer lattice structures [ 5 41 72 73 ] The crystal structure of graphite is shown in Figure 2 1 0 Within the layer, strong covalent bonds between the atoms form planes of high strength which provides the structural strength of the material, which is similar to the molybdenum disulphide structure. In directions normal to these planes weak bonding occurs due to the distance between atoms. This weak bonding allows for the materials to shear with greater ease in that direction. While this is similar to molybdenum disulphide, the main difference is due to each layer consisting of a repeated hexagonal arrangement of carbon atoms not differing atoms as in MoS 2 Figure 2 1 0 Crystal structure of graphite adapted from Lansdown [ 5 ] Graphite and molybdenum disulphide are quite similar in structure; but there are distinct differences between the two materials. Molybdenum disulphide is made up of two elemental species while graphite is made of only carbon atoms. This distinct difference accounts for the discrepancy between the van der Waals forces among the
48 molybdenum disulphide molecules and the coulombic attraction between the carbon atoms [ 5 ] For graphite, a tendency for interlayer bonding is increased as basal plane deformation from sliding exposes high energy edge sites. With an increase in the adhesion between layer s, the friction and wear increase unless the edge sites can be neutralized [ 74 ] Another difference comes from the environment in which each of these subs tances exhibits a low friction coefficient. Graphite exhibits a low friction characteristic in air, yet in vacuum it exhibits high friction and wear behavior. The high d isintegration of the material into a cloud of fine dust like debris [ 73 ] The low friction behavior requires the presence of water vapor or other condensable vapors which act to neutralize the energetic edge sites [ 73 75 76 ] Yen [ 73 ] found that both oxygen and water can act as neutralizers for graphitic films reducing the friction and wear rate. When a film was transitioned between dry nitrogen and dry air, with water vapor content in the chamber being held constant at 100 ppm, the friction immediately dropped from 0.4 to 0.2 and this process was shown to be reversible as can be seen in Figure 2 1 1 A This implies that the oxygen in air can inhibit the increase in friction between the graphite layers. The next experiment was a transition from wet nitrogen to wet air environments which is shown in Figure 2 1 1 b. For this ex periment the chamber was maintained at a constant 50% RH. In this instance, the coating was initially at a lower friction value around 0.15 and immediately increased to 0.2 with the addition of wet air. For this case, the introduction of oxygen into the environment increased the friction of graphite back to the value shown for the transition between dry nitrogen and air and was shown to be reversible in nature. This
49 reversible process suggests that oxygen can modify the surface chemistry of graphite throu gh either a chemisorption or physisorption process [ 73 ] While it was reversible, there was a gradual irreversible increase in friction as well. This increase is suggested to be from the gradual flattening of asperities by wear as sliding is progressed; a great er real area of contact is formed and thus higher friction is obtained [ 77 ] Figure 2 1 1 The environmental dependence on the friction coefficient of graphitic films reproduced from Yen [ 73 ] A) friction coefficient changes as the environment is varied from dry nitrogen to dry air with c onstant water vapor content of 100 ppm. B) Friction coefficient changes as the environment is varied from wet nitrogen to wet air with moisture levels maintained at 50% RH. Due to the beneficial performance of graphite in ambient conditions, composites ha ve been attempted with molybdenum disulphide to combine the beneficial vacuum/inert properties of MoS 2 with the beneficial ambient properties of graphite This has shown to improve the environmental stability in the presence of water vapor [ 78 ] An understanding of which component expresses itself at the surface during the varying environments has been under scrutiny. Zabinski et al [ 78 ] and Voevodin et al [ 34 ] showed with Raman spectroscopy on a MoS 2 /Sb 2 O 3 /graphite film that graphite becomes the primary component to express itself at the interface in ambient conditions and MoS 2 for dry sliding conditions data is depicted in Figure 2 1 2
50 Figure 2 1 2 The effect cycling the relative h umidity from dry (0 % RH) to 40% RH has on the friction coefficient which is reproduced from Voevodin et al [ 35 ] Diamond like Carbon Diamond like carbon (DLC) describes a hard carbon f ilm that is composed of amorphous materials but can include microcrystalline phases [ 79 ] Within the DLC family, there are two main types of films, hydrogenated or hydrogen free. Both types have numerous varieties that include amorphous structures and a mixture of sp 2 (diamond bonding) and sp 3 (graphite bonding) [ 80 85 ] DLC films are gaining in popularity by showing promising tribological characteristics including high hardness, chemical resistance, wear resistance, and electrical resistivity [ 79 86 ] Mechanical and tribological properties are dependent on the deposition process and thus the microstructures, ch emistry, hydrogen content, and the sp 2 /sp 3 bonded carbon ratio [ 82 87 91 ] DLC films are generally hydrogen ated when ion beam or plasma assisted chemical vapor deposition processes are used with hydrocarbon
51 precursor gases. Physical vapor deposition process use graphite as a precursor and offer the capability for hydrogen free DLC growth [ 82 ] While deposition plays a large factor in t he properties, the testing conditions and environment greatly influence the tribological properties as well. Friction coefficient values reported in literature vary substantially ranging from 0.01 to > 0.5 [ 80 ] The greatest ef fect on the friction of DLC films seems to be the moisture content in the environment, but the films do not necessarily have the same dependence on water vapor that graphite does [ 80 82 ] Hydrogen free DLC coatings tend to have improvements in the tribological properties in the presence of water vapor [ 82 ] These characteristics are proposed to be from similar mechanisms as graphite wit h intercalation or passivation of the dangling bonds, which allows for reduced shear strength. Hydrogenated DLC coatings have reduced friction and wear in inert and vacuum environments. Its degradation in moisture rich environments is due to increased frictional forces from oxygen contamination on the surface [ 82 ] Figure 2 13. The effect cycling the relative humidity from dry (0% RH) to 40% RH has on the friction coefficient for a Al 2 O 3 /DLC/Au/MoS 2 film which is adapted from Baker [ 92 ]
52 Similar to graphite, the addition of DLC has been suggested to improve environmental stability of the MoS 2 films. Baker et al. showed this improvement of frictional properties when the environment was cy cled between air (40% RH) to dry nitrogen (<1% RH) and is shown in Figure 2 1 3 [ 92 ] Chromik et al. showed that for a composite YSZ/Au/MoS 2 /C film the MoS 2 remai ns in the contact for both humid and dry environments and is the primary reason for the low friction coefficient [ 78 93 35 ] Like that of graphite, there is still debate on the role these constituents play on the Antimony T rioxide Antimony t rioxide (Sb 2 O 3 ) is an inorganic compound and probably one of the most important additives to molybdenum disulphide films for wear life improvement [ 94 96 ] The first successful applications of Sb 2 O 3 and MoS 2 were reported by both Calhoun et al. [ 94 ]  and Murphy [ 96 ] in 1965. Wh ile there has been a great deal of research into wear and friction mechanisms of MoS 2 / Sb 2 O 3 composites, a complete understanding of the mechanism has not been made. The improvement of performance is attributed to the formation of a Sb 2 O 3 MoO 3 eutectic [ 95 ] limiting of MoS 2 oxidation to MoO 3 in which the mechanism is still undetermined [ 94 ] or the protection of MoS 2 by sacrificial oxidation of Sb 2 O 3 [ 97 ] The mechanistic explanations of the early work can be split into two categories, chemical phenomena and mechanical phenomena. Lavik et al. [ 95 ] first suggest ed an explanation of chemical interactions. The friction an d wear of MoS 2 films both with and without Sb 2 O 3 additives were examined and the addition was shown to decrease the friction. From these results, a theory was introduced that the performance was improved by the formation of eutectics with molybdenum oxide s or through the
53 presence of low melting oxides which combine easily. Further studies by Nosov [ 98 ] and Yamamoto et al. [ 99 ] suggested that Sb 2 O 3 is pref erentially oxidized over MoS 2 These first two theories were rejected by Centers [ 100 ] who took the mechanistic approach that Sb 2 O 3 and other additives plastically deform a t flash temperatures on asperities which in turn help to retain MoS 2 This idea of Sb 2 O 3 forming a structural component in the films agrees with the previous studies. Harmer and Pantano [ 101 ] discussed an increase in density with the addition of Sb 2 O 3 when sintered with MoS 2 This densification of the composite helps in the structural support suggested by Centers. Porosity that develops in films affects tribological properties by providing additional mechanisms for water to diffuse into the film or potentially collapsing during contact and thus causing debris/wear of the coating [ 31 ] It has been noted that basal planes of MoS 2 which orient themselves in the slid ing direction are formed on the sliding surfaces [ 102 103 ] The structural support provided by Sb 2 O 3 al lows for the formation of these basal planes.
54 CHAPTER 3 EXPERIMENTAL SETUP Coatings of Interest A limited number of tribological materials are qualified for use in space vehicles. Soft metals such as gold, silver, lead, and indium have been used for m any years. Recent developments within the MoS 2 community have allowed for additional coatings to become qualified [ 104 107 102 108 110 ] For this series of experiments a broad variety of MoS 2 based coatings were analyzed under varying environmental conditions to determine their wear performance. They ar e listed in Table 3 1. Some of the coatings are commonly found in space applications, while others are still being developed. The Air Force Research Laboratory developed the MoS 2 / Sb 2 O 3 MoS 2 / Sb 2 O 3 / graphite, as well the MoS 2 / Sb 2 O 3 / Au Versions of both th e coating containing graphite and gold are now commercially available from Tribol o gix Inc. Hohman Coatings also deposits a version of the molybdenum disulphide film containing gold. The pure MoS 2 coating studied was deposited by Tribologix Inc The MoS 2 /Ti coating was deposited by TEER Coatings Ltd. Table 3 1. List of the coatings, their manufacturers, and the experiments they were used in. Coating Composition Coating Provider Experiments Used MoS 2 +Ni Hohman Coatings In situ MoS 2 +Ti TEER Coatings In situ MoS 2 +Sb 2 O 3 Air Force Research Labs In situ MoS 2 +Sb 2 O 3 +graphite Air Force Research Labs In situ MoS 2 +Sb 2 O 3 +Au Hohman Coatings Tribologix Inc. In situ QCM MoS 2 Tribologix Inc. QCM
55 All commercial coatings were nominally 1 m thick unless otherwise stated in this work although various deposition techniques were used to apply the coatings. MoS 2 with titanium was a layered coating created via sputtering a layer of pure titanium and then co depositing a layer of MoS 2 and titanium, and finally applying a layer of pure MoS 2 The MoS 2 and nickel, pure MoS 2 and both MoS 2 /Sb 2 O 3 /Au films were sputter deposited The addition of other constituents besides the MoS 2 was to improve adhesion to the substrate, environmental stability, as well as add toughness to the target which is made of the desired constituents resulting in a coating of roughly the same composition as the target. Experimental Equipment Two types of equipment we re used for these experiments: a tribometer and a quartz crystal microbalance. The tribometer used to perform testing was a linear reciprocator. Its functionality and design are discussed in this section. In Situ Linear Reciprocating Tribometer Wear trac k topography can vary greatly between samples which can potentially affect t he wear life of a coating. It wa s for this reason an in situ linear reciprocating tribometer depicted in Figure 3 1 was used to examine the wear occurring during experiments. T his tribometer was described in detail by Bares et al. [ 111 ] The in situ measurements were performed with a Zygo New View 5030 scanning white light interferometer (SWLI). The SWLI is capable of meas uring surface heights of less than 1 nm up to 5000 m in size. The samples for the linear reciprocating portion of the tribometer were mounted on top of a linear stage while the force transducer was placed in the load path between the
56 pin and ground with a flexural component between them. This assembly was mounted to a manual positioning stage which allowed the user to set a normal force up to a maximum value of 10 N. Figure 3 1 A s chematic of the situ linear reciprocating tribometer with scanning white light objective is shown. This tribometer is used to measure both friction and wear rates of the coatings of interest This entire tribometer was situated within an acrylic environment chamber, allowing for the gaseous environment to be varied. For the majority of experiments the environment was either a dry nitrogen cover gas or 25% relative humidity (RH) air Data acquisition was recorded via a data acquisition card and LabVIEW where the normal force, friction force, ambient temperature, temperatu re on the surface of the sample relative humidity, and stage position were all recorded.
57 Quartz Crystal Microbalance The QCM system used for the room temperature experiments was an Inficon system. The controller was a XTM/2 which has the capability of me asuring the frequency changes and thus associated mass per unit area values. The frequency range is 5 to 6 MHz +/ 0.1 Hz. The recorder has an update rate of 4 Hz. The crystal holder was a front load standard single sensor holder with a maximum temperat ure range of 130 C. The system used for the thermal experiments was a Maxtech TM 400 monitor. The frequency resolution was 0.03 Hz for a 6 MHz crystal. The crystal holder was an Inficon cool drawer single sensor and a thermocouple was brazed to the back of the crystal holder to allow for in situ measurements of temperature changes. The crystals were 13.97 mm diameter AT quartz crystals which are plano convex contoured. The quartz crystal was coated with gold electrodes on either side of the crystal. Two types of crystals were used for these experiments, non polished crystals from Inficon which have a base frequency range from 5.97 to 5.99 MHz and an average roughness of 1 to 5 m. The other type used were polished crystals from Tangidyne LLC which al so have a base frequency range of 5.97 to 5.99 MHz but an average roughness of better than 5 nm Experimental Setup The emphasis of this work was to gain an understa nding of the environmental factors that affect the wear of molybdenum disulphide films with a particular focus on the role water has on the tribological properties of MoS 2 The literature has demonstrated that the friction is affected by changing the relative humidity of the environment, yet there has not been a large focus on the wear of the c oatings or the
58 general quantity of water that is required to affect the tribological properties. This section discusses experiments that provided insight into how wear is affected by the addition of water into the environment and how the quantities of wat er uptake change with varying water content. All experiments investigati ng the wear and friction in a humid compared to dry environment were performed at the University of Florida Tribology Laboratory. The experiments investigating the quantity of water uptake were performed at Sandia National Laboratories. Varying Humidity Friction and Wear Experiments One of the largest influential factors on the performance of molybdenum disulphide films is water in the environment. Water contamination has been a con stant issue and its influence is not completely understood The friction response of MoS 2 based coatings has been well documented for more than 60 years yet studies on the wear process have been limited in the literature The majority of the wear experi ments have been ex situ after the coating has already failed. The goal of these experiments was to capture the evolution of the wear scar and thus the actual wear events as the test progresse d As discussed previously, a SWLI is used to measure the wear of the coatings. By taking SWLI scans at regular time intervals at a consistent location near the middle of the wear track, the area loss at that particular region for each interval was measured an example of which is depicted in Figure 3 2 As discusse d in the theory section, the majority of the error in the volume loss calculations does not originate from the equipment but in fact comes from the variation in wear that occurs throughout the wear track. Unfortunately for the in situ set of experiments, only one portion of the wear track was being measured, thus the variation in wear throughout the track cannot be
59 measured. It was assumed that the analyzed portion was a sufficient representative sample of the volume loss occurring throughout the wear tr ack. Figure 3 2. Representative SWLI line scans taken at cycle 0 and cycle 2000 which are used to determine the area loss. The samples used for the in situ experiments were rectangular coupons of 1.5 0 in x1.0 0 in x 0 .1 9 in dimensions and were made of 304 stainless steel. Prior to being shipped for coating, the flats were polished to a n average surface roughness better than 50 nm. The pin samples were 0.25 in diameter 6061 T6 aluminum spheres that were drilled and tapped to allow for a 4 40 screw. Th e spheres came polished with a mirror finish and were left unchanged prior to being sent for coating. Both coupons and pins were coated with MoS 2 based coatings, but these varied by constituents. The coatings used for these experiments are discussed in t he first section of this chapter and are summarized in Table 3 1 Prior to running any tests with this system, a calibration step was performed to verify that the objective lens of the Zygo was aligned with the wear track. A counterface and pin were moun ted to the system and the counterface was moved directly below the Zygo lens. The Zygo stage was adjusted to align the counterface normal to the
60 objective axis. After the surface was brought into focus, the stage was returned to its initial position and the pin was loaded to 5N. The sample was run for 50 cycles on a 5 mm track to allow for a visible wear scar to be achieved. The pin was then unloaded and the stage was moved 0.5 mm at a time until the center of the wear scar was in the center of the fiel d of view. This distance was then noted for future testing purposes The protocol for varying humidity environmental testing was established as a cyclic procedure. The two environments of interest were laboratory air conditions (RH>20%) and back fillin g the acrylic chamber with dry nitrogen which provides a relative humidity of less than 2%. T he sample was run for a total of 500 cycles in the dry environment with scans of the same portion of the wear track taken after cycles 0 20, 30, 40, 50, 100, 200, 300, 400, and 500 cycles. The chamber was then purged with laboratory air and the sample was run for another 500 cycles with the same scan frequency. This process was repeated for a total of 2000 sliding cycles 1000 cycles total in each environment The next wear track run follow ed the same protocol but began in a laboratory air environment and transitioned to a dry nitrogen one after wards The goal of this data was to provide insight i nto how the system accommodates for a transition between the dry and humid environments where the friction coefficients are drastically different. It also aimed at understanding if the initial run in period in either a dry or humid environment will affect the future performance of the coating. Steady State Wear Expe riments The varying humidity experiments provided insight into the evolution of wear and how the system accommodates changes in relative humidity, but it was not an ideal method for studying steady state wear of the coatings in each environment. Two
61 separ ate wear tracks were run on each sample, one for each environment. The tests were run for 10,000 cycles and SWLI scans of the wear track were taken every 1,000 cycles. For these experiments a normal force of 5 N was applied betw een the counterface and t he pin, a sliding speed of 10 mm/s, and a 5 mm wear track length was defined. Quartz Crystal Microbalance Experiments Room t emperature s orption As water has been shown to significantly affect the friction properties of molybdenum disulphide films, a better understanding of how the water is influencing these characteristics is still needed. The purpose of these experiments was to understand the quantities of water that the film sorbs during relative humidity cycling of the environment at room temperature T he approach was to provide step changes in the relative humidity of the environment to analyze the diffusion rates into the films. This was accomplished by minimizing the volume surrounding the QCM. Two 250 mL beakers were used for the dry and wet environ ments with separate flows going into both. The dry beaker had dry nitrogen flowing into it at 3 liters per minute (LPM) which allows the entire volume to be replenished in 5 seconds. The wet environment consisted of a 3 LPM dry flow that splits with a po rtion flowing through a bubble humidification system before recombining. All flows are controlled with variable area flow meters which allow ed for fine control of the desired rela tive humidity in each beaker. This system is depicted in Figure 3 3 The QC M was rigidly fixed to a relative humidity probe for all experiments. This allowed for in situ measurement of both the relative humidity and environmental temperature in the beakers. The relative humidity probe was a Vaisala DRYCAP hand
62 held dew point me ter DM70. The probe has a dew point temperature range of 50 to +60 C with an accuracy of +/ 2 C. The probe measures the relative humidity from 0 to 100% but the accuracy varies depending on the relative humidity range being read. For measurements les s than 5% RH the accuracy is +/ 0.025% RH +17% of the reading. For measurements greater than 5% the accuracy is +/ 0.8% RH +2% of the reading. It has the capability of measuring the ambient temperature from 10 to +60 C with an accuracy of +/ 2 C. Fi gure 3 3. Experimental setup for step change in relative humidity experiments. The procedure for these experiments was to mount the crystal into the holder, allow the QCM frequency to reach steady state in the ambient lab environment to verify that the el ectronics were warmed up and no electronic drift was occurring. Once a steady frequency value was read, the QCM and RH probe were placed into the dry nitrogen beaker and again allowed to equilibrate. After both the environment and the frequency had reach ed steady state (~1000+ seconds), the QCM and RH probe were removed from the dry beaker and placed immediately into the humidified environment beaker. While this was not a perfect step change in the relative humidity the environment had reached a steady state relative humidity in about 5 10 seconds. This procedure was repeated numerous times for the various relative humidities of interest
63 ( 5%, 10%, 20 %, 30 %, 40%, 50%, 60% and 70%) All experiments were performed at the ambient lab temperature which was r ecorded with the RH probe. Thermally a ctivated d esorption The next set of experiments sought to gain an understanding of the amount of water that can be removed after a thermal cycling procedure associated with the crystals after exposure to relative humid ity environments. In order to accomplish this, the crystal s w ere heated in an inert environment to remove the potential for oxidation of the film and analyze the quantit ies of water desorption from a thermal cycling process. This setup is depicted in Figu re 3 4. Figure 3 4. Experimental setup for thermal desorption experiments. The QCM system used for this set of experiments was a Maxtech TM 400 monitor The frequency resolution was 0.03 Hz for a 6 MHz crystal. The crystal holder was an Inficon cool d rawer single sensor and a thermocouple was brazed to the back of the crystal holder to allow for in situ measurements of temperature changes This assembly was welded to a CF40 feedthrough which attaches to one end of the furnace used A Thermo Electron C orporation Lindberg/Blue M tube furnace was used to heat the crystal. It has a maximum temperature of 1200 C. The QCM assembly was suspended in the middle of the tube furnace to allow for heating to be established via radiant and convective methods A flow through system was established through the
64 furnace. This consisted of a dry nitrogen flow at 4.5 LPM which was split into a pure dry flow and a wet fl ow through a bubble humidifier. Both the dry and wet flows are recombined prior to entering the cha mber and have variable area flow meters which were used for fine control of the flows, thus allowing for relative humidity regulation in the furnace. The oxygen content and relative humidity of the flow was measured on the exhaust of the furnace. The exh aust was flowed through a General Eastern Hygro M 4 chilled mirror and then through a Delta F Platinum Series Oxygen Analyzer. All heating experiments were performed under a dry nitrogen environment with oxygen content less than 5 ppm A LabVIEW program was used to read in the temperature and crystal resonant frequency data. T he relative humidity and oxygen content was recorded by the user at regular intervals. The experimental procedure was to mount the crystal into th e chamber, e stablish a dry nitroge n flow through in the environment and allow reach a steady state value. Once a steady state frequency was established the furnace was turned on to heat the crystal to 110 C. After holding the system at 110 C for 3 10 hou rs, the furnace was turned off and allowed to cool down back to the ambient lab temperature. This was repeated 1 2 times to understand if additional water could be removed after the first heating/cooling cycle. T he last portion of this experimental proc edure was to follow a similar protocol as that of the 110 C heating, but heat the crystal to 170 C to observe if additional mass loss will occur at the higher temperature.
65 Microscopy Upon completion of the QCM experiments, various surface characterizatio n techniques were performed on the crystals Extreme high resolution s canning electron microscopy ( HRSEM ) was utilized to characterize the surface microstructure of the deposited films. The HR SEM work was performed at Sandia National Laboratories by Rich ard P. Grant on a FEI Magellan 400 X ray photoelectron spectroscopy (XPS) was performed at Sandia National Laboratories by Michael T. Brumbach. XPS was performed using a Kratos Axis Ultra DLD Instrument with a monochromatic Al K (1486.6 eV) source opera ted at 225 W and a base pressure less than 5x10 9 t orr. Full survey spectra were collected with an analyzer pass energy of 160 eV and step size of 900 1000 meV. High resolution spectra were collected with an analyzer pass energy of 40 eV and step size of 0.1 eV. The analyzed area was an ellipse with dimensions on the order of 300 m by 700 m. All spectra were analyzed using CasaXPS software. The quantifications were obtained from CasaXPS using the built in relative sensitivity factors. The peaks were fit using a Shirley background and a Gaussian(70%)/Lorentzian(30%) line shape The Mo 3p 3/2 peak was fit using two unconstrained components. The low binding energy component is designated as MoS 2 while the higher binding energy component is associated w ith MoO 3 The final characterization technique utilized an atomic force microscope (AFM) to acquire topographical images to determine the surface area of the films. This work was performed by Anna Trujillo and Tony Ohlhausen at Sandia National Laboratories on a Bruker Dimension Icon Peak force tapping mode was performed with a 2 nm nominal
66 radius TESP SS tip cantilever Two scan sizes were analyzed, 5.0 x 5.0 m and 500 x 500 nm.
67 CHAPTER 4 EXPERIMENTAL RESULTS In Situ Wear Experiments The in situ chang ing environment experiments were performed to gai n a better understanding how the wear of molybdenum disulphide (MoS 2 ) coating s develop with changes in the water content in the environment. It has been shown throughout the literature that increasing the r elative humidity of the environment leads to an increase in the friction coefficient. The goal of these experiments was to observe the wear with v arying relative humidity and correlate the results to the previously understood friction coefficient trends. The method used to estimate the wear volume occurring in these systems was discussed in Chapter 2 For this particular set of experiments, due to the nature of the experimental equipment, only one cross sectional scan of the wear track could be made dur ing testing. While the assumption was made that the wear was uniform throughout the wear scar, the uncertainty in the measurements is lower than if the variation in the wear track could have been taken into account The experimental data for all cyclic te sts initiated in laboratory air environments can be seen in Figure 4 1. For most of the coatings the volume loss did not significantly change with the changing environment, but as expected the friction coefficient changed considerably. The only coating w hich exhibited a noticeable difference in volume loss during the tests was the MoS 2 /Ni film. For this film, the wear rate was less in the dry nitrogen environment as compared to the humid air portions of the test.
68 Figure 4 1. Friction coefficient and v olume loss of the coatings are plotted versus normal force times sliding distance for all cycled environment experiments initiated in laboratory air.
69 Figure 4 2. Friction coefficient and volume loss of the coatings are plotted versus normal force times sliding distance for all cycled environment experiments initiated in dry nitrogen.
70 Similar results can be seen from the tests initiated in dry nitrogen, which is depicted in Figure 4 2. Again, the only coating demonstrating a repeatable difference in wear rates between the environments is the MoS 2 /Ni film. While wear rate values could be estimated for each envi ronment during these 2000 cycle tests, the quantity of volume loss was minim al for some of the films. This increases the uncertai nty in the measu rements as the SWLI used has an estimated repeatability of 10 nm and some of the films had a final wear depth of only 100 200 nm for the 2000 cycle tests It is for this reason that further studies were required to look at the stea dy state wear rate of ea ch film. Steady State Wear Rate Experiments In order to study the steady state wear rate of each coating, tests were performed for 10,000 cycles in a single environment. Tests were performed on all of the composite films tested in the previous in situ sec tion and the results are summarized in Figure 4 3. The aim of the 10,000 cycle experiments was to explore if the humidity in the environment affect ed the steady state wear rates of the composite films. I n all instances, the friction coefficient was higher in humid air than dry nitrogen which follows the trends from the literature The wear rate had a similar trend for all films, but some films had a more pronounced difference with the change in environment than others. The MoS 2 with Sb 2 O 3 and graphite fi lm had the smallest difference between the humid air and dry nitrogen environments. This is attribu ted to the addition of graphite which as discussed in Chapter 2 exhibits its lubricious nature in a humid environment [ 109 82 93 ]
71 Figure 4 3. Friction coefficient and wear rate data for 10,000 cycle exp eriments in both humid air and dry nitrogen environments on composite MoS 2 films. The error bars for the friction coefficient are the associated standard deviation of the data. The wear rate error bars could not be displayed as the associated instrumenta l uncertainty is negligible on these scales. Quartz Crystal Microbalance Experiments Quartz crystal microbalance experiments were performed to gain an understandin g of the quantities of water up take associated with changes in the environmental relative hu midity. Isotherms of water sorption on MoS 2 powder have been shown in the literature, but studies h ave not been performed on thin films. The goal of these experiments was to determine the quantity of water sorbed by the MoS 2 film. Step Changes in Relativ e Humidity Experiments The first data of interest was the room temperature water uptake of the films due to step changes in the relative humidity of the environment. The data is presented in
72 units of mass per unit area which was determined via the Sauerbr ey equation applied to the raw resonant frequency data. I n order to grasp a better mental model of what a mass per unit area quantity meant these values were converted to monolayers of water sorbed. This data is not meant to lit erally assume that the wat er up take is in a monolayer fashion, but rather provide a mental model for the reader to comprehend the magnitude of these numbers. Two methods were determined to calculate the number of monolayers of water sorbed The first relies on the density and thic kness. The thickness of a monolayer of water is assumed to be 3.1 x10 10 m and the density of water is 1 x10 6 g/m 3 The measured mass per unit area is divided by both the density and thickness, as shown in Equation 4 1 (4 1) The second method takes into account the area number density, molecular weight n, is assumed to be 10 15 molecules/cm 2 the molecular weight M, N A is 6.023 x10 23 mole cules/mol. Th is method is shown in Equation 4 2 Both methods provide results on the same order of magnitude, thus th e first method is chosen to calculate the estimated monolayers of water for all tests. (4 2) Experimental data for a 0.1 m thick pure MoS 2 film deposited on a rough substrate is depicted in Figure 4 4 a This graph shows the mass per unit area changes with time as the relative humidity of the environment is changed for a subset of the data.
73 The step changes can then be normalized with respect to time and compared to one another, which is depicted in Figure 4 4 b. The sorption isotherm for each coating is measured f rom the individual step changes in mass per unit area Figure 4 4. Mass per unit area data for rep resentative step change in relative humidity data for the 0.1 m thick pure MoS 2 film. A) Mass per unit area versus time. B) mass per unit area and monolayers of water versus normalized time for the step change portions of data. Figure 4 5. Mass per unit area data for representative step change in relative humidity data for the 0.1 m thick MoS 2 /Sb 2 O 3 /Au film. A)Mass per unit area versus time. B) Mass per unit area and monolayers of water versus normalized time for the step change portions of data.
74 A similar representative data set for a 0.1 m composite MoS 2 /Sb 2 O 3 /Au film is shown in Figure 4 5 Again the sorption isotherm is determined from the step change in mass per unit area data Three different crystals all coated during the same deposition process were cycled thro ugh a similar step change in relative humidity to analyze the repeatability in the measurements. The normalized mass per unit area data is shown in Figure 4 6. Significant variability was shown in the mass per unit area for these three crystals. As all c oatings were deposited in the same deposition run, this variability is assumed to be due to the roughness of the quartz crystal substrate. Figure 4 6. Variability between three crystals coated in the same batch at similar relativ e humidity step changes. The se crystals were analyzed using a scanning e lectron microscope ( SEM ) and a scanning white light interferometer ( SWLI ) Variability in the average roughness values was found between each of the crystals depicted in Figure 4 6. T his implies the surface area for each crystal varies and thus t he sites available for water up take to occur vary as well In addition SEM images showed that there were relatively large trenches formed during the deposition of the coatings on these rough surfaces. Between the roughness of the substrate and the trenches in the coating there are additional surface
75 area sites and thus additional pathways for wate r up take to occur. Due to the surface area variability between samples the pure MoS 2 and compos ite MoS 2 /Sb 2 O 3 /Au coatings cannot be compared at this time, but the general magnitude of values can be noted for each film. Figure 4 7. Surface area variability between rough and smooth surfaces. A) depicts a surface that has an average roughness of 8 nm from a SWLI (left) and SEM (right) and B) depicts a surface with an average roughness of 350 nm from a SWLI (left) and SEM (right). The large surface roughness and voids formed during the deposition process led to the purchase of polished QCM crystals t o limit these surface area differences as well as improve the uniformity of the coating during the deposition process. The previous crystals had an average roughness of approximately 250 nm. The polished crystals had
76 an average roughness of approximately 8 nm. Representative images from a SWLI and SEM of a rough coated crystal and a polished coated crystal are depicted in Figure 4 7. Figure 4 8. SEM micrographs of the two pure MoS 2 samples. A) Depicts the 0.1 m Further analysis of the coatings microstructural differences using high resolution SEM techniques exemplified the difference between the 0.1 m and the 0.9 m thick MoS 2 phy are shown in Figure 4 8. Figure 4 9. Mass per unit area data for representative step change in relative humidity data for the 0.9 m thick MoS 2 film. A) Mass per unit area versus time. B) Mass per unit area and monolayers of water versu s normalized time for the step change portions of data.
77 The next set of experiments explored the difference in mass uptake for the variable thickness coatings. The samples of interest were the previously studied 0.1 m thick MoS 2 coating on the rough crystal and a 0.9 m thick MoS 2 coating on a polished crystal. A representative portion of the step change in relative humidity data for the 0.9 m thick MoS 2 film is shown in Figure 4 9. The data from all of the step change experiments is summarized in Figure 4 10 This figure depict s the sorption isot herms for the 0.9 m thick MoS 2 sample, and the 0.1 m thick MoS 2 and MoS 2 /Sb 2 O 3 /Au samples. This illustrates that the two 0.1 m coatings had similar magnitudes of water up take as compared to the 0 .9 m coating whose water sorption val ues increased by an order of magnitude. Figure 4 10. Water sorption isotherms for both 0.1 m and 0.9 m MoS 2 films and the 0.1m MoS 2 /Sb 2 O 3 /Au film.
78 Inert Heating Experiments During the step change in relative humidity experiments the mass per unit area would not always approach the initial minimum value for a dry nitrogen environment. From this knowledge the following heating experiments were performed to determine if there was any additional water in the films that is not removed by solely removing the moisture in the environment. The initial heating experiments were performed on the 0.1 m thick pure MoS 2 and MoS 2 /Sb 2 O 3 /Au films. Both samples went through numerous heat/cool cycles up to 110 C. The mass per unit area data is sho wn in conjunction with the associated temperature data versus time for the MoS 2 film and the MoS 2 /Sb 2 O 3 /Au film in Figure 4 11 and Figure 4 12 respectively In both figures, the temperature change is shown on the top plot and the mass per unit area is sho wn on the bottom plot. Due to the previously discussed temperature effects on the QCM crystal, only the relative changes in mass per unit area at room temperature, before and after heating, are compared. These changes in mass are labeled throughout the figure s. It was noted for both the MoS 2 and the MoS 2 /Sb 2 O 3 /Au films, after going through one heat/cool cycle at 110 C a large portion of mass was shown to be removed from the film. When the crystals went through another heat/cool cycle with the environme nt remaining dry ( O 2 <5 ppm ) only negligible mass loss was observed. This implies that at 110 C it was possible to thermally desorb a maximum amount of water with one heat/cool cycle. The next set of experiments analyzed the potential for additional water removal by heating to 170 C instead of 110 C. The mass per unit area and temperature data is
79 again plotted versus time for both the 0.1 m MoS 2 and the MoS 2 /Sb 2 O 3 /Au film s This data is shown in Figures 4 13 and 4 14 respectively. Figure 4 11. Rela tive changes in mass per unit area are measured before and after each heat/cool cycle up to 110 C for the thin MoS 2 film. A) depicts the temperature versus time and B) depicts the mass per unit area versus time. Figure 4 12. Relative changes in mass pe r unit area are measured before and after each heat/cool cycle up to 110 C for the thin MoS 2 /Sb 2 O 3 /Au film. A) depicts the temperature versus time and B) depicts the mass per unit area versus time.
80 Figure 4 13. Relative changes in mass per unit area are measured before and after each heat/cool cycle up to 170 C for the thin MoS 2 film. A) depicts the temperature versus time and B) depicts the mass per unit area versus time. Figure 4 14. Relative changes in mass per unit area are measured before and aft er each heat/cool cycle up to 170 C for the thin MoS 2 /Sb 2 O 3 /Au film. A) depicts the temperature versus time and B) depicts the mass per unit area versus time. Additional mass loss was observed upon heating to 170 C for both of the 0.1 m films. In both instances mass changes of about 0.1 g/cm 2 were observed for all heat/cool cycles at 170 C once the initial large quantity of thermally desorbed mass was removed. For the MoS 2 film, the first 170 C heat/cool cycle which was held for ~3 hours prior to co oling had a mass loss of 0.13 g/cm 2 Upon the next heat/cool cycle,
81 the temperature was held at 170 C for ~15 hours and only observed a mass loss of 0.14 g/cm 2 The MoS 2 /Sb 2 O 3 /Au film exhibited identical mass loss values of 0. 1 2 g/cm 2 when held at 1 70 C for 7 hours and then at 4.5 hours which was the same order of magnitude as the MoS 2 film The final experiment performed a similar heat/cool procedure up to 110 C and 170 C for the 0.9 m thick MoS 2 film. Numerous cycles could not be performed du e to time constraints with the equipment, thus a heat/cool cycle to 110 C immediately followed by a heat/cool cycle to 170 C was performed. The heat/cool data for the 0.9 m film is shown in Figure 4 15. Figure 4 15. Relative changes in mass per unit a rea are measured before and after each heat/cool cycles up to 110 C and 170 C for the 0.9 m MoS 2 film A) depicts the temperature versus time and B) depicts the mass per unit area versus time. As additional mass loss was shown to occur at elevated temper atures, verification that oxidation was not contributing to the mass loss values was required. High
82 resolution x ray photoelectron spectroscopy ( XPS ) was performed on three crystals for each 0.1 m film, one that was as deposited, one film that had seen r elative humidity exposures, and a film that had seen relative humidity exposures and gone through thermal cycles. The goal of the high resolution XPS was to quantify the degree of degradation that has occurred in the films for either relative humidity exp osure and/ or thermal cycling. XPS data was acquired around the Mo 3p region, fits were performed on the data for the corresponding oxide and sulfide for Mo binding energies, and finally the data was deconvoluted into the respective oxide and sulfide speci es contributions. This same process could be used on the S 2p peak, but as MoO 3 is not volatile, the Mo 3p peak provides a more reliable measurement of the quantity of oxidation as compared to the S 2p region. The data is shown in Figure 4 16. It was fou nd that there was not a strong trend of oxidation occurring in the films. It hard to conclude whether the increase in oxidation seen was due to the environments to whic h the samples were exposed or simply variability within the samples This data further supports the idea that if oxidation is occurring, it i s minimal and thus the majority if not all of the mass per unit area values are due to water removal.
83 Figure 4 16. XPS spectral region for Mo 3p 3/2 used to determine the amount of oxidation via the ratio of Mo oxide to sulfide for the 0.1 m MoS 2 films. A1) A MoS 2 coated crystal that was as deposited. A2) A MoS 2 coated crystal is depicted that had experienced st ep changes in relative humidity. A3) A MoS 2 coated crystal is depicted that had experienced both step changes in relative humidity and heat/cool cycling. B1) A composite MoS 2 coated crystal is depicted that was as deposited. B2) A composite MoS 2 coated cry stal is depicted that had experienced step changes in relative humidity. B3) A composite MoS 2 coated crystal is depicted that had experienced both step changes in relative humidity and heat/cool cycling.
84 CHAPTER 5 DISCUSSION This chapter describes the th ree major conclusions of the water sorption experiments and presents a hypothesis explaining the phenomenological difference in the sorption mechanism for room temperature desorbable water and thermally desorbable water. The first section is a discussion of the evidence that water absorbs into, rather than adsorbing onto, the film. The second section presents various methods for the calculation of diffusion coefficients of water into films of MoS 2 The third section presents a discussion of the thermally desorbable water. Lastly, the fourth section examines hypothetical phenomenological mechanisms by which room temperature and elevated temperature desorption of water occurs. Evidence of Water Absorption A difference in the amounts of water uptake was obs erved at room temperature for the thick and thin pure MoS 2 films. The coating thickness was increased by a factor of 9 from 0.1 m to 0.9 m. The room temperature mass per unit area of sorbed water at various relative humidit ies are summarized in Table 5 1. Two factors affect water uptake in the films: film thickness and surface roughness where surface roughness exists on multiple length scales Figure 5 1 summarizes the ratio of the rough to smooth crystal surface areas. As seen in Table 5.1, the ratio of water uptake between rough and polished crystal s was approximate ly 1:40 The ratio of the rough to polished crystal surface area at the nanometer scale was quantified via AFM measurements and was found to be 1:4. This ratio is under estimating the sur face area, as can be seen by the difference in resolution between the SEM image and AFM image shown in Figure 5 1C, but provide a general approximation. If the diffusion of
85 water was solely a film surface (adsorption) phenomenon then film thickness would not affect the mass uptake and only the surface roughness would affect the measurement If this was the case, the mass up take should scale by 1:2 as compared to the measured 1:40 ratio measured Thus, the film thickness must be influencing the water up take into the film. If water permeation into the film (absorption) is assumed to occur, then the ratio of rough/thin to polished/thick coated crystal surface is greater than 1:18. This data suggests that the water is diffusing into the film (permeating t he film). Table 5 1. Mass per unit area room temperature sorption values for both the thick and thin MoS 2 films, corresponding to the data shown in Figure 4 10. Diffusivity of Water in Molybdenum Disulphide Films With the knowledge that the water absor be d at room temperature permeates into the film, the kinetics associated with the rate of reversible water uptake are of interest. From the step change experiments performed, the diffusivity of water into these films can be determined.
86 Figure 5 1. The differential ratio for various length scales between the coated rough and polished crystals. A) depicts the difference in film thickness, B) SEM and AFM images are shown which illustrate the difference in micrometer scale roughness which is measured usin g a both a SWLI and AFM to provide the 2:1 ratio. C) Both AFM and higher resolution SEM images are shown to exemplify the difference in nanometer scale microstructural roughness which was measured using an AFM.
87 Diffusion coefficients of gases or vapors int o solids can be measured via a number of gravimetric analysis methods such as indirectly measuring changes in mass as a function of the resonant frequency of a quartz crystal with a quartz crystal microbalance (QCM) [ 112 114 ] In order for accurate measurements to be made t he sample must have a well defined geometry, typically a thin film and the weight gain of the sample is measured with respect to time in a ma intained atmosphere of penetrant at constant temperature and pressure. Crank derived solutions for one dimensional diffusion through a plane sheet [ 114 ] The derivation starts with which is shown in Equation 5 1. (5 1) The boundary conditions and initial condition are shown in Equations 5 2 through 5 4. (5 2) (5 3) (5 4) The initial condition, Equation 5 2, describes that at t=0 the film is at a uniform concentration of C 1 The first boundary condition, Equation 5 3, describes the concentration at the surface of the film, x=L, as being constant, C 0 at any time after the beginning of diffusion into the film. In order for this boundary condition to be satisfied, the environme nt surrounding the apparatus, in this case the quartz crystal microbalance, must be changed instantaneously and maintained at a constant value with no pressure or temperature changes. Equation 5 4 requires that there is no transport beyond the bottom of the film, at x=0. The solution of Equation 5 1 subject ed
88 to these boundary and initial conditions is described by Crank and shown in Equation 5 5 [ 115 116 ] This solution assumes that the diffusion coefficient is constant and the volume remains constant (the film does not swell). (5 5) E quation 5 5 assumes that the up take is considered a diffusion process which is controlled by a constant diffusion coefficient, D. In this equation M t is the total amount of vapor absorbed by the sheet at time t, and M is the equilibrium sor ption attained. The key assumptions made in this derivation by Crank are that the diffusion of the penetrant into the film occurs in one dimension, the concentration of the penetrant at the surface of the film is maintained at a constant value, and the di ffusion coefficient is a constant [ 115 ] The first assumption is valid as the thickness is small in comparison to the surface area of the coated quartz crystal. The film thicknesses ranged from 0.1 to 0.9 m while the diameters of the coatings were over 8 mm. The second assumption is valid as the concentration of the environment surrounding the QCM system was kept constant. The surrounding volume was minimized and the flow rates allowed the environment to reach a steady state value in less th an 5 seconds and be maintained to the concentration of interest. The final assumption is verified by the experimental data. Obtaining a sorption curve from the QCM frequency data is a typical application of a QCM system. As the penetrant diffuses into th e film of interest, a computer is used to record the resonant frequency of the crystal as a function of time. As was shown in Figures 4 4, 4 5, and 4 9 there is a frequency response measured with a step change in
89 relative humidity. From the measured fre quency response data, the ratio of M t /M can be calculated and graphs of M t /M versus time can be generated. The mass change of penetrant is proportional to the associated frequency change, as was shown from the Sauerbrey equation in Equation 2 16. The ratio of M t /M can be expressed as the ra tio of the frequency changes which is shown in Equation 5 6. The mass after time t, is defined as the difference between the frequency of the crystal at time t after the penetrant is allowed into the chamber, f t and the frequency of coated crystal prior to addition of penetrant, f 1 The mass at equilibrium is defined as the difference between the frequency of the crystal at equilibrium, f and the frequency prior to the addition of penetrant. (5 6) From the M t /M data the associated diffusion coefficients can be estimated through numerous methods. The three methods of interest for this work are the half time, limiting slope, and moment method [ 112 113 116 117 ] Half Time Method The half time method is the simplest of the three that will be described for estimating the diffusion coefficient. The data is plotted in a similar fashion as the representative da ta shown in Figure 5 2 From this plot, the time at which M t /M = 0.5 is determined The diffusion coefficient is then found from Equation 5 7 [ 116 117 ] (5 7) This equation can be simplified to solve for the diffusion coefficient as shown in Equation 5 8 [ 117 ]
90 (5 8) Figure 5 2 Example of calculating the time when M t /M is 0.5 by the half time method. The associated error for Equation 5 8 was calculated via the law of propagation of uncertainty which is shown in Equation 5 9. (5 9) Moment Method The moment method requires the use of the entire sorption curve. The first moment, s is calculated via numerical integration as shown in Equation 5 10 (5 10) The first moment is determine d from the plot of 1 M t /M versus time a representative graph is shown in Figure 5 3 The first moment is measured via numerical integration of the area under the curve, shown in Figure 5 3. The equation
91 for the average diffusion coefficient is then fou nd from Equation 5 1 1 which is derived in detail by Willoughby [ 116 ] (5 11) Figure 5 3 Example of calculating s by the moment method. The associated error for Equation 5 11 was calculated via the law of propagation of uncertainty which is shown in Equation 5 12. (5 12) Limit ing Slope Method The limiting slope method [ 112 ] is based off of the first term in the series from Equation 5 5 at suffici ently long times. This approximation is simplified and shown in Equation 5 1 3 (5 1 3 )
92 At large times, a plot of ln(1 M t /M ) versus time will approach a straight line and the diffusion coefficient, D, can be calculated by Equation 5 1 4 An example of determining the limiting slope is shown in Figure 5 4 (5 1 4 ) In E quation 5 1 3 R represents the li miting slope of the previously described plot. This limiting slope it typically found by performing a linear regression from M t /M = 0.98 0 0.999. Figure 5 4 Example of calculating R b y the moment method. A) is an example of the thick sample in which the method is valid and B) is an example of the thin sample in which the resolution of the instrument does not allow for this method to be used. The limiting slope method is not a good fit for this data set. This is due to the increase in the relative error as there is an increased need for finer resolution with these relatively small mass change values in comparison to literature results. The values of the mass change per unit area become too small to resolve when M t /M is on the order of 0.9 8 for thinner film data sets as is seen Figure 5 4B
93 Diffusivity Results Data from all three films w ere studied to compare the difference in diffusivity between sample thickness and composition. For reasons discussed previously on ly the half time and moment methods are used to calculate the diffusion coefficients of water into MoS 2 Table 5 2 summarize s the diffusivity results found from these experiments The two methods provided values on the same order of magnitude, but the un certainty in the moment method was greater thus for this di scussion the half time method diffusivity values were used and are summarized in Figure 5 5 Figure 5 5. A comparison of the diffusivity of water into various MoS 2 films. The error bars repre sent the uncertainty in the data due to the use of the half time method. The diffusion coefficient is not dependent upon the relative humidity of the environment which validates the assumption of a constant diffusion The composite and pure 0.1 m MoS 2 fi lms were found to be of similar magnitudes while the diffusivity of water in the thicker pure MoS 2 film was found to have a faster rate
94 Table 5 2. Calculated diffusivity values for the MoS 2 films. A) is the 0.9 m pure sample, B) is the 0.1 m pure sampl e, and C) is the 0.1 m composite. The topographical differences, as shown in the SEM micrographs depicted in Figure 4 8, is hypothesized to be the potential cause in the change in diffusivity values between the varying thickness samples The surface are a availability for water up take in the thick sample with the thin columnar like topographical structure would be larger than that of the topography sh own for the thin sample. While at this point in time t his
95 there is potential f or further studies to analyze the surface area/internal porosity of the films using BET sorption isotherm methods. Thermally Activated Desorption Continuous Mass Loss Oxidation? During the inert heating experiments it was determined that relatively large changes in mass per unit area were observed after the first heat/cool cycle. This change in mass per unit area has since been referred to as thermally activated desorption. Two regimes of mass per unit area loss were observed, one at 110 C and another at 170 C. Upon heating to 110 C, a considerable amount of mass per unit area loss was measured approximately twice the quantity observed during the room temperature absorption experiments After further heat/cool cycles, minimal additional quantities of mass removal per unit area were found ~0.007g/cm 2 Yet, when the same procedure was repeated for 170 C continuous larger changes in mass per unit area were observed ~0.12 g /cm 2 As these temperatures are not high enough to be affecting the quart oxidizing causing these mass losses, with the rate of oxidation being less at 110 C as compared to 170 C. In order to determine the potential for environmental impurit ies being the culprit for this continuous loss of mass per unit area the impingement rate of water on to the coating surface was studied. The detailed analysis is described in Appendix A, a brief explanation follows here. This method is an attempt to pro vide an upper bound for the potential for environmental oxidation that could occur in the film.
96 The quantity of impurities in the ultra high purity nitrogen gaseous flow is less than 5 ppm and thus a value of 1 ppm of water was assumed to interact with the coating for these upper bound calculations. The partial pressure of impurities in the environment is calculated from which the flux of impurities impacting per area per second is found. This flux can then be used to determine the total number of molecule s of oxygen landing on the surface per area for the duration of the heat cycle If the balanced chemical equation for oxidation of MoS 2 is analyzed, which is shown in Equation 5 1 4 3.5 impacting impurities are required to convert one MoS 2 molecule to MoO 3 [ 118 ] (5 1 4 ) These calculations assume that every 3.5 impurities impacting the surface will lead to the oxidation of one MoS 2 to MoO 3 This is an upper bound on the problem as it is not likely that every i mpact will participate in the oxidation chemistry. The film is also assumed to be completely dry after being heated to 110 C and thus only environmental impurities are the source for oxidation. The change in mass per unit area associated with this uppe r bound method was on the order of 10 00 g/ cm 2 The values thought to be due to oxidation that were observed during testing were at most 0.14 g/ cm 2 Thus, it is very possible that environmental oxidation could be an explanation for this continuous mass loss occurring at both 110C and 170 C. An attempt at quantifying the degradation of the film was made using x ray photoelectron spectroscopy (XPS). As discussed previously in the experimental results section, the XPS data did not show a strong trend of oxidation of the films. Also, as a single crystal was not used for all XPS measurements variability was found and no strong statements as to the degree of oxidation can be made. However, as there was
97 not a significant increase in the MoO 3 area of the Mo 3p peak, this suggests that the amount of oxidation occurring i s minimal to nonexistent. To further exemplify this fact calculations were made to determine what the change in fractional percentage of MoO 3 would be required for the measured mass loss per u nit area at 170 C. The detailed analysis is shown in Appendix B. It was found that the fractional percentage of MoO 3 would increase by ~3% with a mass loss of 0.12 g/cm 2 for a film that is 0.1 m thick For the 0.9 m thick film, the fractional percen tage of MoO 3 drops to 0.3%. With such minimal changes in the quantity of oxidation, the variability of the oxidation between crystals masks the potential to measure the amount of oxidation occurring within the films. Ratio of Water to MoS 2 in the Film A comparison of the maximum quantity of mass per unit area thermally desorbed at 110 C for all three films was made. Table 5 3 summarizes the results found during the expe rimentation for both of the 0.1 m films T hese values were shown to be repeatable if the coating had been exposed for a sufficient time Table 5 3. Maximum thermally desorbed mass per unit area values at 110 C The thermally desorbed quantities of mass per unit area for the two 0.1 m films were on the same order of magnitude while th e thicker film had a value 10 times greater. The thermally desorbed mass per unit area scaled with the thickness, similar to the room temperature absorption values discussed earlier. It was of interest to
98 determine what moles per area (mol/area) of H 2 O t o mol/area of MoS 2 this would be associated with. The method for determining the ratio of the mol/area for water to molybdenum disulphide is described in detail in Appendix C. This ratio will be referred to as the molar ratio for the purposes of this disc ussion. The mol/area of MoS 2 is determined from the molecular weight of MoS 2 density of MoS 2 and thickness of the film. The mol/area of water was determined from the measured mass per unit area loss and the molecular weight of water. The molar ratios for 0.1 m and 0.9 m pure MoS 2 films were found to be similar, 0.14 and 0.16 respectively. This constant molar ratio of water to MoS 2 further supports the idea that the quantity of the thermally desorbed water scales with the film thickness. Hypothesi s of Water Diffusion Locations With the knowledge that there are both room temperature sorption and thermally activated desorption processes occurring within these films, the hypothesis is that there are two diffusion processes taking place. The thermally desorbed quantities of water are the first to diffuse into a completely dry film. The water preferentially phy s isorb s to defects within the film essentially capping the sites, which allows the water to be at a lower energy state than in the vapor phase l eading to the necessity for additional energy to remove the water molecules. The next diffusion process occurs once the defects sites are capped and the water begins to stack up within the film. The water absorbed during this process does not require add itional energy to be remove d thus the mass per unit area of water within the film will change with partial pressure of water changes. As the 0.9 m and 0.1 m MoS 2 films were deposited by the same process, it would be anticipated that the ratio of the num ber of defect sites to the volume of MoS 2
99 deposited would scale with the film thickness. This appears to be a valid assumptio n as the molar ratios are both approximately 0.15 water to MoS 2 It is known from microprobe analysis that the films are not stoi chiometric but closer to MoS 1.6 Thus, on top of the typical defects that occur during deposition sulfur vacancies could be another potential site for the physisorbed water that requires an energetic input to be removed
100 CHAPTER 6 CONCLUSIONS Water has b een shown to significantly impact the tribological properties of molybdenum disulphide films. Both the friction coefficient and wear rates of the films were shown to increase with an increase in the moisture in the environment. While the wear rates follo wed the same trend, it was not to the same magnitude as the increase seen in the friction coefficient results. Through the use of a quartz crystal microbalance water was shown to be absorbed into the film. Two different quantities of water were found, ro om temperature sorption and water that require d an energetic input in order for desorption to occur. The thermally activated desorption quantities were found to have constant molar ratios of water to molybdenum disulphide for both film thicknesses. The ro om temperature sorption values linearly increased with an increasing film thickness, but the molar ratios were not constant. The films were found to be sub stoichiometric and thus further studies on a stoichiometric film would be of interest to determine i f limiting the number of sul f ur deficiencies will minimize the quantity of water that requires thermal energy to be removed. It is also of interest to determine how a tribologi cally run in film, with oriented surface layers, would affect the diffusivity o f water into the film. Another study of interest would be to have comparable thickness pure and composite MoS 2 films de posited on polished crystals to compare their change in diffusivity and water uptake quantities and potentially relate the differences ba ck to the density differences of the films. Finally, determining the internal surface area through BET isotherm measurements would allow
101 a direct comparison between the films and water uptake without having to account for microstructural differences.
102 APP ENDIX A IMPINGEMENT AND OXID ATION RATES This section present s a derivation of the impingement rates of impurities from the gas stream on the molybdenum disulphide coating surface during the heating experiments. These experiments were run with ultra high p urity nitrogen gas flow through system. As ultra high purity nitrogen is known to have up to 5% impurities present in the gas, a conservative estimate of 1 part per million (ppm) of these impurities having the potential to lead to oxidation of the coatin g is made It is then assumed that one tenth of these oxidizing impurities would impact the surface of the coating From this the partial pressure of water P H2O present in the nitrogen can be calculated from atmospheric pressure, P atm and the fraction of impurities, n This calculation is shown in Equation A 1 (shorthand is used in the remaining derivation to refer to equations, such as A 1 for Equation A 1). (A 1) The nitrogen gas was flowing through a heated tube furnace t hus the relative partial pressure of the impurities ,P 2 is calculated from the pressure at room temperature found via A 1, P 1 and the associated temperatures, T 1 and T 2 which are input into the ideal gas law. (A 2)
103 The number of impurity molecules impacting the surface per unit area per second f, is determined from the relative pressure, molecular weight of water, M, and the temperature of the gas stream and is shown in A 3. (A 3) Assuming a period o f time, t, and the knowledge that it will take a minimum of 3.5 impurities to convert one molybdenum disulphide molecule to one molybdenum trioxide molecule the nu mber of moles oxidized per area, m, is found from A 4. (A 4) The associated grams per area is then calculated for both molybdenum disulphide and molybdenum trioxide by taking m and dividing by the molecular weight M, of either molybdenum disulphide or molybdenum trioxide shown in A 5 and A 6 respectively (A 5) (A 6) The change in mass per area is then found by the subtraction of the two values and converting to g/cm 2 which is shown in A 7 (A 7)
104 APPENDIX B FRACTIONAL INCREASE IN OXID ATION DUE TO MASS PE R UNIT AREA LOSS AT 170C The derivation of the fractional change of molybdenum trioxide (MoO 3 ) is based on the continuous mass per unit area loss observed at 170 C The fraction of MoO 3 prior to heating is designated as f 1 and the fr action of MoO 3 after heating is f 2 The densities of the system before and after heating are defined in Equation B 1 and B 2 respectively (shorthand is used in the remaining derivation to refer to equations, such as B 1 for Equation B 1). (B 1) (B 2) The differen ce between the initial and final system densities would equal the measured change in mass per unit area dm/A, observed divided by the thickness of the film L, which is shown in B 3. (B 3) From B 3 the change in the fractional p ercentage of MoO 3 can be determined as shown in B 4 (B 4)
105 APPENDIX C MOLAR RATIO CALCULAT IONS Th is section present s a derivation of the ratio of moles of water to the moles of molybdenum disulphide MoS 2 from the mass per un it area, m/A, measurements made. The theoretical moles per area m ol / A within the film is found from the density MoS2 and molecular weight of MoS 2 M MoS2 and the thickness of the film, L. This is shown for the thick film in Equation C 1 (shorthand is used in the remaining derivation to refer to equations, such as C 1 for Equation C 1) and thin film in C 2 ( C 1) (C 2) The moles per area of water for the thick and thin samples is calculated from the associated meas ured mass change per area, m/A, and the molecular weight of water, M H2O This is shown in C 3 and C 4 respectively (C 3 ) (C 4 ) The ratio of the moles per water to the moles of molybdenum disulphide is calculated by taking the ratio of C 3 to C 1 which is shown in C 5 and C 4 to C 2 which is shown in C 6
106 (C 5) (C 6)
107 FUNDING SOURCES Financial support for this work was provided in p art by Harris Corporation, an AFOSR MURI grant, and Sandia National Laboratories. Any opinions, finding, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the Air Force Offi ce of Scientific Research. Sandia National Laboratories is a multi program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for y Administration under contract DE AC04 94AL85000.
108 LIST OF REFERENCES 1. Campbell, M.E.: Solid Lubrication Technology a Review. Mechanical Engineering 90 (2), 28 & (1968). 2. Roberts, E.W.: Thin Solid Lubricant Films in Space. Tribo logy International 23 (2), 95 104 (1990). 3. Krick, B.A., Sawyer, W.G.: Space Tribometers: Design for Exposed Experiments on Orbit. Tribology Letters 41 (1), 303 311 (2011). doi:10.1007/s11249 010 9689 y 4. Banks, B.A., de Groh, K.K., Miller, S.K.: Low Eart h Orbital Atomic Oxygen Interactions with Spacecraft Materials. In: Materials Research Society, Boston, Massachusetts, November 29 December 3, 2004 2004. NASA TN (2004) 5. Lansdown, A.R.: Molybdenum disulphide lubrication, 1st ed. Elsevier, Amsterdam ; New York (1999) 6. Cramer, J.A.: Elements of the Art of Assaying Metals. London (1764) 7. Dickinson, R.G., Pauling, L.: The crystal structure of molybdenite. Journal of the American Chemical Society 45 1466 1471 (1923). 8. Bell, M.E., Findlay, J.H.: Molyden ite as a New Lubricant. Physical Review 59 (922) (1941). 9. Fleischauer, P.D., Lince, J.R., Bertrand, P.A., Bauer, R.: Electronic structure and lubrication properties of molybdenum disulfide: a qualitative molecular orbital approach. Langmuir 5 (4), 1009 10 15 (1989). doi:10.1021/la00088a022 10. Johnson, R.L., Godfrey, D., Bisoon, E.: Friction of solid films on steel at high sliding velocities. In. (1948) 11. Haltner, A.J., Oliver, C.S.: Effect of Water Vapor on Friction of Molybdenum Disulfide. Industrial & Engineering Chemistry Fundamentals 5 (3), 348 & (1966). 12. Peterson, M.B., Johnson, R.L.: Friction and Wear Investigation of Molybdenum Disulfide I Effect of Moisture. In: NACA (ed.). Lewis Flight Propulsion Laboratory, Cleveland, (1953) 13. Cannon, P.: M elting Point and Sublimation of Molybdenum Disulphide. Nature 183 (4675), 1612 1613 (1959). 14. Godfrey, D., Nelson, E.C.: Oxidation Characteristics of Molybdenum Disulfide and Effect of such Oxidation on its Role as a Solid Film Lubricant. In: NACA (ed.). Lewis Flight Propulsion Laboratory, Cleveland, (1949)
109 15. Bisson, E.E., Johnson, R.L., Swikert, M.A.: Friction, Wear, and Surface Damage of Metals as Affected by Solid Surface Films: A Review of NACA Research. In: NACA (ed.). NACA TN, London, (1957) 16. R owe, G.W.: Some observations on the frictional behaviour of boron nitride and of graphite. Wear 3 (4), 274 285 (1960). doi:Doi: 10.1016/0043 1648(60)90292 1 17. Amateau, M.F., Krause, H.H., Orcutt, F.K., Glaeser, W.A., Allen, C.M.: Development of Lubricants for Positive Displacement Power Sources from 1200F to 1500F. In: USAF (ed.), vol. 61 478. USAF TR, (1961) 18. Spalvins, T.: Deposition of Mos2 Films by Physical Sputtering and Their Lubrication Properties in Vacuum. ASLE Transactions 12 (1), 36 & (1969). 19. Spalvins, T., Przybyszewski, J.: Deposition of Sputtered Molybdenum Disulfide Fimls Friction Characteristics of such Films in Vacuum NASA Technical Note(D 4269) (1967). 20. Hopkins, V., Campbell, M.: Friction and Wear Life of Selected Solid Lubrican t Films at 100 F, Rt, and 400 F. Lubrication Engineering 25 (11), 430 & (1969). 21. Farr, J.P.G.: Molybdenum disulphide in lubrication. A review. Wear 35 (1), 1 22 (1975). doi:Doi: 10.1016/0043 1648(75)90137 4 22. Holinski, R., Gansheim.J: Study of Lubrica ting Mechanism of Molybdenum Disulfide. Wear 19 (3), 329 & (1972). 23. Stupp, B.C.: Synergistic Effects of Metals Co Sputtered with Mos2. Thin Solid Films 84 (3), 257 266 (1981). 24. Spalvins, T.: Frictional and Morphological Properties of Au Mos2 Films Sp uttered from a Compact Target. Thin Solid Films 118 (3), 375 384 (1984). 25. Roberts, E.W., Price, W.B.: Advances in Molybdenum Disulfide Film Technology for Space Applications. In: Sixth European Space Mechanisms & Tribology Symposium, Zurich 1995, pp. 27 3 278 26. Liu, H.W., Zhang, X.S., Yu, D.Y., Wnag, X.P.: The Enhancement of Wear Life and Moisture Resistance of Sputtered Mos2 Films by Metal Ion Implantation. Wear 173 (1 2), 145 149 (1994). 27. Renevier, N.M., Fox, V.C., Teer, D.G., Hampshire, J.: Coati ng Characteristics and Tribological Properties of Sputter Deposited MoS2/Metal Composite Coatings Deposited by Close Field Unbalanced Magnetron Sputter Ion Plating. Paper presented at the ICMCTF, San Diego,
110 28. Didziulis, S.V., Hilton, M.R., Bauer, R., Fl eischauer, P.D.: Thrust Bearing Wear Life and Torque Tests of Sputter Deposited MoS2 Films. In: Corporation, T.A. (ed.). El Segundo, (1992) 29. Spalvins, T.: Lubrication with sputtered MoS2 films: Principles,operation, and limitations. Journal of Materials Engineering and Performance 1 (3), 347 351 (1992). 30. Lince, J.R., Fleischauer, P.D.: Chapter 7: Solid Lubricants. In: Conley, P.L. (ed.) Space Vehicle Mechanisms: Elements of Successful Design. John Wiley & Sons, Inc., New York (1998) 31. Hilton, M.R., Fleischauer, P.D.: Applications of Solid Lubricant Films in Spacecraft. Surface & Coatings Technology 55 (1 3), 435 441 (1992). 32. Loewenthal, S.H., Chou, R.G., Hopple, G.B., Wenger, W.L.: Evaluation of Ion Sputtered Molybdenum Disulphide Bearings for Spa cecraft Gimbals. Tribology Transactions 37 (3), 505 515 (1994). 33. Hopple, G.B., Keem, J.E., Loewenthal, S.H.: Development of Fracture Resistant, Multilayer Films for Precision Ball Bearings. Wear 162 919 924 (1993). 34. Voevodin, A.A., Fitz, T.A., Hu, J.J., Zabinski, J.S.: Nanocomposite tribological coatings with "chameleon" surface adaptation. Journal of Vacuum Science & Technology a Vacuum Surfaces and Films 20 (4), 1434 1444 (2002). doi:Doi 10.1116/1.1487875 35. Voevodin, A.A., Zabinski, J.S.: Laser s urface texturing for adaptive solid lubrication. Wear 261 (11 12), 1285 1292 (2006). doi:DOI 10.1016/j.wear.2006.03.013 36. Hertz, H.: ber die Berhrung fester elastischer Krper (On Contact Between Elastic Bodies). Journal fr die reine und angewandte Mat hematik 92 156 171 (1882). 37. Holm, R.: Electric Contacts: Theory and Application, 4th ed. Springer Verlag, New York (1967) 38. Archard, J.F., Hirst, W.: An Examination of a Mild Wear Process. Proceedings of the Royal Society of London Series a Mathemat ical and Physical Sciences 238 (1215), 515 & (1957). 39. Archard, J.F., Hirst, W.: The Wear of Metals under Unlubricated Conditions. Proceedings of the Royal Society of London Series a Mathematical and Physical Sciences 236 (1206), 397 & (1956). 40. Yong, S., Sang Gook, K.: A lateral, self cleaning, direct contact MEMS switch. In: Micro Electro Mechanical Systems, 2005. MEMS 2005. 18th IEEE International Conference on, 30 Jan. 3 Feb. 2005 2005, pp. 195 198
111 41. Stachowiak, G., Batchelor, A.: Engineering Tri bology, 2nd ed. Butterworth Heinemann, (2001) 42. Williams, J.A.: Engineering Tribology, vol. 488. Cambridge University Press, (2005) 43. Colbert, R.S., Alvarez, L.A., Hamilton, M.A., Steffens, J.G., Ziegert, J.C., Burris, D.L., Sawyer, W.G.: Edges, cleara nces, and wear: Little things that make big differences in bushing friction. Wear 268 (1 2), 41 49 (2010). doi:DOI 10.1016/j.wear.2009.06.030 44. Archard, J.F.: Elastic Deformation and the Laws of Friction. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences 243 (1233), 190 205 (1957). 45. Bowden, F.P., Tabor, D.: Friction Lubrication and Wear a Survey of Work during Last Decade. British Journal of Applied Physics 17 (12), 1521 & (1966). 46. Adam, N.K.: In: The Physics and Chemistry of Surfaces. p. 232. Oxford University Press, (1938) 47. Holm, R.: Wiss. Veroff. Siemens Werken 7 217 258 (1938). 48. Ernst, E., Merchant, M.E.: Conference on Friction and Surface Finish. In: Conference on Friction and Surface Finish, Camb ridge, Mass.: MIT 1940, p. 76 49. Archard, J.F.: Single Contacts and Multiple Encounters. Journal of Applied Physics 32 (8), 1420 & (1961). 50. Antler, M.: Sliding Wear of Metallic Contacts. Ieee Transactions on Components Hybrids and Manufacturing Techno logy 4 (1), 15 29 (1981). 51. Dursch, H., Spear, S.: On orbit coldwelding fact or friction? LDEF, 69 months in space : first post retrieval symposium 3134 1565 (1991). 52. Burris, D.L., Sawyer, W.G.: Measurement Uncertainties in Wear Rates. Tribology Let ters 36 (1), 81 87 (2009). doi:DOI 10.1007/s11249 009 9477 8 53. Williams.Jb, Hunt, R.T.: Asperity Persistence and Real Area of Contact between Rough Surfaces. Proceedings of the Royal Society of London Series a Mathematical and Physical Sciences 327 (1569), 147 & (1972). 54. Sayles, R.S.: How two and three dimensional surface metrology data are being used to improve the tribological performance and life of some common machine elements. Tribology International 34 (5), 299 305 (2001). 55. Blau, P.J.: ASM Han dbook, Volume 18 Friction, Lubrication, and Wear Technology. In., p. 21. ASM International, (1992)
112 56. Colbert, R.S., Krick, B.A., Dunn, A.C., Vail, J.R., Argibay, N., Sawyer, W.G.: Uncertainty in Pin on Disk Wear Volume Measurements Using Surface Scanni ng Techniques. Tribology Letters 42 (1), 129 131 (2011). doi:DOI 10.1007/s11249 010 9744 8 57. Ward, M.D., Buttry, D.A.: In Situ Interfacial Mass Detection with Piezoelectric Transducers. Science 249 (4972), 1000 1007 (1990). doi:10.1126/science.249.4972.100 0 58. Gerber, E.A., Sykes, R.A.: A Quarter Century of Progress in the Theory and Development of Crystals for Frequency Control and Selection. In: 25th Annual Symposium on Frequency Control. 1971, 1971 1971, pp. 1 45 59. Winer, W.O.: Molybdenum disulfide a s a lubricant: A review of the fundamental knowledge. Wear 10 (6), 422 452 (1967). 60. Martin, J.M., Donnet, C., Le Mogne, T., Epicier, T.: Superlubricity of molybdenum disulphide. Physical Review B 48 (14), 10583 (1993). 61. Lince, J.R., Hilton, M.R., Bom mannavar, A.S.: Oxygen Substitution in Sputter Deposited Mos2 Films Studied by Extended X Ray Absorption Fine Structure, X Ray Photoelectron Spectroscopy and X Ray Diffraction. Surface & Coatings Technology 43 4 (1 3), 640 651 (1990). 62. Peterson, M.B., J ohnson, R.L.: Friction of Possible Solid Lubricants with Various Crystal Structures. In: NACA (ed.). Lewis Flight Propulsion Laboratory, Cleveland, (1954) 63. Pritchard, C., Midgley, J.W.: The effect of humidity on the friction and life of unbonded molybde num disulphide films. Wear 13 (1), 39 50 (1969). 64. Peterson, M.B., Johnson, R.L.: Factors Influencing Friction and Wear with Solid Lubricants. Lubrication Engineering 11 325 330 (1955). 65. Pardee, R.P.: Effect of Humidity on Low Load Frictional Proper ties of a Bonded Solid Film Lubricant. ASLE Transactions 15 (2), 130 & (1972). 66. Gao, C., Bredell, L., Kuhlmann Wilsdorf, D., Makel, D.D.: Micromechanics of MoS2 lubrication. Wear 162 164 (Part 1), 480 491 (1993). doi:Doi: 10.1016/0043 1648(93)90533 r 67. Johnston, R.R.M., Moore, A.J.W.: The burnishing of molybdenum disulphide on to metal surfaces. Wear 7 (6), 498 512 (1964). 68. Haltner, A.J.: An evaluation of the role of vapor lubrication mechanisms in MoS2. Wear 7 (1), 102 117 (1964). doi:Doi: 10.1016/00 43 1648(64)90082 1
113 69. Barwell, F.T., Milne, A.A.: Use of Molybdenum Disulphide in Association with Phosphated Surfaces. Industrial Lubrication and Tribology 3 (9), 10 14 (1951). 70. Gansheimer, J.: Neue Erkenntnisse uber die Wirkungsweise von Molybdandisu lfid als Schmierstoff. Schmiertechnik 5 271 280 (1964). 71. Barry, H.F., Binkelma.Jp: Mos2 Lubrication of Various Metals. Lubrication Engineering 22 (4), 139 & (1966). 72. Shankara, A., Menezes, P., Simha, K., Kailas, S.: Study of solid lubrication with MoS2 coating in the presence of additives using reciprocating ball on flat scratch tester. Sadhana 33 (3), 207 220 (2008). 73. Yen, B.K.: Influence of water vapor and oxygen on the tribology of carbon materials with sp(2) valence configuration. Wear 192 (1 2), 208 215 (1996). 74. Lancaster, J.K.: A review of the influence of environmental humidity and water on friction, lubrication and wear. Tribology International 23 (6), 371 389 (1990). doi:Doi: 10.1016/0301 679x(90)90053 r 75. Savage, R.H.: Graphite Lu brication. Journal of Applied Physics 19 (1) (1948). 76. Savage, R.H., Schaefer, D.L.: Vapor Lubrication of Graphite Sliding Contacts. Journal of Applied Physics 27 (2), 136 138 (1956). 77. Lancaster, J.K.: Friction and Wear of Non Graphitic Carbons. ASLE Transactions 20 (1), 43 54 (1977). 78. Zabinski, J.S., Bultman, J.E., Sanders, J.H., Hu, J.J.: Multi environmental lubrication performance and lubrication mechanism of MoS2/Sb2O3/C composite films. Tribology Letters 23 (2), 155 163 (2006). doi:DOI 10.1007/s 11249 006 9057 0 79. Grill, A.: Review of the tribology of diamond like carbon. Wear 168 (1 2), 143 153 (1993). doi:Doi: 10.1016/0043 1648(93)90210 d 80. Erdemir, A., Eryilmaz, O.L., Fenske, G.: Synthesis of diamondlike carbon films with superlow friction a nd wear properties. Journal of Vacuum Science & Technology A 18 (4), 1987 1992 (2000). 81. Bull, S.J.: Tribology of carbon coatings: DLC, diamond and beyond. Diamond and Related Materials 4 (5 6), 827 836 (1995). doi:Doi: 10.1016/0925 9635(94)05325 1 82. Mu ratore, C., Voevodin, A.A.: Chameleon Coatings: Adaptive Surfaces to Reduce Friction and Wear in Extreme Environments. Annual Review of Materials Research 39 (1), 297 324 (2009). doi:doi:10.1146/annurev matsci 082908 145259
114 83. Cuomo, J.J., Doyle, J.P., Bru ley, J., Liu, J.C.: Sputter Deposition of Dense Diamond Like Carbon Films at Low Temperature. Applied Physics Letters 58 (5), 466 468 (1991). 84. Wei, R., Wilbur, P.J., Erdemir, A., Kustas, F.M.: The Effects of Beam Energy and Substrate Temperature on the Tribological Properties of Hard Carbon Films on Aluminum. Surface & Coatings Technology 51 (1 3), 139 145 (1992). 85. Ronkainen, H., Koskinen, J., Likonen, J., Varjus, S., Vihersalo, J.: Characterization of wear surfaces in dry sliding of steel and alumina on hydrogenated and hydrogen free carbon films. Diamond and Related Materials 3 (11 12), 1329 1336 (1994). doi:10.1016/0925 9635(94)90147 3 86. Vanhulsel, A., Velasco, F., Jacobs, R., Eersels, L., Havermans, D., Roberts, E.W., Sherrington, I., Anderson, M. J., Gaillard, L.: DLC solid lubricant coatings on ball bearings for space applications. Tribology International 40 (7), 1186 1194 (2007). doi:DOI: 10.1016/j.triboint.2006.12.005 87. Harris, S.J., Weiner, A.M., Meng, W.J.: Tribology of metal containing diamo nd like carbon coatings. Wear 211 (2), 208 217 (1997). 88. Erdemir, A., Switala, M., Wei, R., Wilbur, P.: A Tribological Investigation of the Graphite to Diamond Like Behavior of Amorphous Carbon Films Ion Beam Deposited on Ceramic Substrates. Surface & Co atings Technology 50 (1), 17 23 (1991). 89. Donnet, C., Grill, A.: Friction Control of Diamon like Carbon Coatings. Surface & Coatings Technology 94 95 (1 3), 456 462 (1997). 90. Grill, A., Patel, V., Meyerson, B.: Tribological Behavior of Diamond like Car bon Effects of Preparation Conditions and Annealing. Surface & Coatings Technology 49 (1 3), 530 536 (1991). 91. Koskinen, J., Schneider, D., Ronkainen, H., Muukkonen, T., Varjus, S., Burck, P., Holmberg, K., Scheibe, H.: Microstructural Changes in DLC Fi lms Due to Tribological Contact. Surface & Coatings Technology 108 (1 3), 385 390 (1998). 92. Baker, C.C., Hu, J.J., Voevodin, A.A.: Preparation of Al2O3/DLC/Au/MoS2 chameleon coatings for space and ambient environments. Surface & Coatings Technology 201 (7 ), 4224 4229 (2006). doi:DOI 10.1016/j.surfcoat.2006.08.067 93. Chromik, R.R., Baker, C.C., Voevodin, A.A., Wahl, K.J.: In situ tribometry of solid lubricant nanocomposite coatings. Wear 262 (9 10), 1239 1252 (2007). doi:DOI 10.1016/j.wear.2007.01.001 94. C alhoun, S.F., Meade, F.S., Murphy, G.P., Young, R.L.: Factors Affecting Performance of Resin Bonded Solid Film Lubricants. Lubrication Engineering 21 (3), 97 & (1965).
115 95. Lavik, M.T., Hubbell, R.D., Mcconnell, B.D.: Oxide Interaction Concept for Improve d Performance with Molybdenum Disulfide. Lubrication Engineering 31 (1), 20 & (1975). 96. Murphy, G.P.: Corrosion Inhibiting Resin Bonded Solid Film Lubricant. U.S. Patent, 97. Campbell, M., Hopkins, V.: Development of Polyimide Bonded Solid Lubricants. L ubrication Engineering 23 (7), 288 & (1967). 98. Nosov, M.I.: Lubricating Properties of Mixtures of Molybdenum Disulfide with Antimony Oxides and Sulfides. Chemistry and Technology of Fuels and Oils 14 (7 8), 518 519 (1978). 99. Yamamoto, E., Wada, K., Fuk uzuka, T., Shimogori, K., Fujiwara, K., Tsuji, K.: Lubricating Films to Prevent Galling of Stainless Steel Threaded Parts. Lubrication Engineering 40 (10), 588 597 (1984). 100. Centers, P.W.: The Role of Oxide and Sulfide Additions in Solid Lubricant Compa cts. Tribology Transactions 31 (2), 149 156 (1988). 101. Harmer, R.S., Pantano, C.G.: Synergistic Mechanisms of Solid Lubricants. In: Force, A. (ed.). Wright Patterson Air Force BAse, Dayton, (1978) 102. Scharf, T.W., Kotula, P.G., Prasad, S.V.: Friction a nd wear mechanisms in MoS2/Sb2O3/Au nanocomposite coatings. Acta Materialia 58 (12), 4100 4109 (2010). doi:DOI 10.1016/j.actamat.2010.03.040 103. Zabinski, J.S., Donley, M.S., Mcdevitt, N.T.: Mechanistic Study of the Synergism between Sb2o3 and Mos2 Lubrica nt Systems Using Raman Spectroscopy. Wear 165 (1), 103 108 (1993). 104. Aouadi, S.M., Paudel, Y., Luster, B., Stadler, S., Kohli, P., Muratore, C., Hager, C., Voevodin, A.A.: Adaptive Mo2N/MoS2/Ag tribological nanocomposite coatings for aerospace applicati ons. Tribology Letters 29 (2), 95 103 (2008). doi:DOI 10.1007/s11249 007 9286 x 105. Martin, J.M., Le Mogne, T., Boehm, M., Grossiord, C.: Tribochemistry in the analytical UHV tribometer. Tribology International 32 (11), 617 626 (1999). 106. Muratore, C., V oevodin, A.A.: Molybdenum disulfide as a lubricant and catalyst in adaptive nanocomposite coatings. Surface & Coatings Technology 201 (7), 4125 4130 (2006). doi:DOI 10.1016/j.surfcoat.2006.08.014 107. Roberts, E.W.: Ultralow Friction Films of MoS2 for Space Applications. Thin Solid Films 181 461 473 (1989).
116 108. Tagawa, M., Muromoto, M., Hachiue, S., Yokota, K., Ohmae, N., Matsumoto, K., Suzuki, M.: Hyperthermal atomic oxygen interaction with MOS2 lubricants and relevance to space environmental effects in low earth orbit effects on friction coefficient and wear life. Tribology Letters 18 (4), 437 443 (2005). doi:10.1007/s11249 004 3594 1 109. Voevodin, A.A., Zabinski, J.S.: Nanocomposite and nanostructured tribological materials for space applications. Com posites Science and Technology 65 (5), 741 748 (2005). doi:DOI 10.1016/j.compscitech.2004.10.008 110. Fleischauer, P.D., Hilton, M.R., Raju, B.K., Vedachalam, N., Wei, L.Y., Huang, K.W., Nishimura, M.: International Applications of Space Tribology. Tribolog y International 23 (2), 135 147 (1990). 111. Bares, J.A., Argibay, N., Mauntler, N., Dudder, G.J., Perry, S.S., Bourne, G.R., Sawyer, W.G.: High current density copper on copper sliding electrical contacts at low sliding velocities. Wear 267 (1 4), 417 424 (2009). doi:10.1016/j.wear.2008.12.062 112. Balik, C.M.: On the extraction of diffusion coefficients from gravimetric data for sorption of small molecules by polymer thin films. Macromolecules 29 (8), 3025 3029 (1996). 113. Felder, R.M.H., G.S.: Methods of Experimental Physics. Academic Press, New York (1980) 114. Crank, J.P., G. S.: Diffusion in Polymers. Academic Press, New York (1968) 115. Crank, J.: Some Methods of Deducing the Diffusion Coefficient and its Concentration Dependence from Sorption Experim ents Transactions of the Faraday Society 51 (12), 1632 1641 (1955). doi:10.1039/tf9555101632 116. Willoughby, S.L.: Effective Diffusion Coefficients for Toluene in Calcium Carbonate Filled Poly(Vinyl Acetate) from Quartz Crystal Microbalance Sorption Experi ments. Cameron University, 184 (1998). 117. Crank, J.: The Mathematics of Diffusion. Oxford University Press, Oxford (1975) 118. Abdel Rehim, A.M.: Thermal analysis and X ray diffraction of roasting of Egyptian molybdenite. Journal of Thermal Analysis and Calorimetry 57 (2), 415 431 (1999).
117 BIOGRAPHICAL SKETCH Rachel Sarah Colbert was born in Gainesville, Florida. She graduated in 2003 from the International Baccalaureate Program at Eastside High School. She was accepted at the University of Florida with a full scholarship from Florida Bright Futures Program. In 2007 she graduated with a Bachelor of Science degree in Mechanical Engineering. She was then admitted to the graduate program at the University of Florida where she completed her Master of S cience (2009) and Doctor of Philosophy (2012) in the field of mechanical engineering. She researched under the direction of Professor W. Gregory Sawyer from 2007 2012 where her research focus was on the environmental effects on the tribological properti es of solid lubricants. From July 2011 until December 2012 she interned at Sandia National Laboratories under the mentorship of Dr. Michael T. Dugger. During this time she continued to study solid lubricants but began to focus primarily on the role of wa ter. Rachel was also very active in societies. She was a founding member of the Graduate Student Council for the Mechanical and Aerospace Engineering Department, serving as chair of the council from 2009 2010. She continued her involvement as a member o f the Society of Tribologists and Lubrication Engineers where she took on leadership positions in the Materials Tribology and the Early Career Tribologists Committees.