<%BANNER%>

Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2014-05-31.

DARK ITEM
Permanent Link: http://ufdc.ufl.edu/UFE0043903/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2014-05-31.
Physical Description: Book
Language: english
Creator: Vail, Jennifer Renee
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Jennifer Renee Vail.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Sawyer, Wallace G.
Electronic Access: INACCESSIBLE UNTIL 2014-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0043903:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043903/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2014-05-31.
Physical Description: Book
Language: english
Creator: Vail, Jennifer Renee
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Jennifer Renee Vail.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Sawyer, Wallace G.
Electronic Access: INACCESSIBLE UNTIL 2014-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0043903:00001


This item has the following downloads:


Full Text

PAGE 1

1 IN SITU MICROTRIBOLOGY OF ELASTOMERIC MATERIALS By JENNIFER VAIL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOP HY UNIVERSITY OF FLORIDA 201 2

PAGE 2

2 201 2 Jennifer Vail

PAGE 3

3 To my mother, Paula

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank the members of my committee for their support and guidance over the past four years: Prof. W. Gregory Sawyer, Prof. David Hahn, Prof. Scott Banks, and Prof. Scott Perry. I especially thank Prof. Sawyer for providing an invaluable graduate research experience. To all members of the University of Tribology lab, past and present, I thank you for your i ncomparable friendship, support and collaboration over the years. I would like to thank my family in Florida, Montana, and South Carolina for their love and support. It is impossible to express my gratitude in writing for the endless love and support my mother, father, stepmother, br other, and sister have all provided: thank you Paula, David, Barbara, Jaso n, and Abby for being my tower of strength. I would also like to thank Jessica and Bobby for the years of invaluable friendship. Finally, I or providing the soundtrack that powered this work. Finally, I wish to thank Becton Dickinson for funding my research and providing the opportunity to collaborate with Nestor Rodriguez, Paolo Mangiagalli, Bo Persson, and Brandon Krick, who have all contri buted immensely to this scientific endeavor.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 2 BACKGROUND ................................ ................................ ................................ ...... 25 2.1 Rough Surfaces ................................ ................................ ................................ 25 2.2 Contact Mechanics ................................ ................................ ........................... 26 ..................... 27 2.2.2 Adhesive Contacts: The JKR and DMT Models ................................ ...... 30 2.2.3 Modeling Between the DMT and JKR Limits ................................ ........... 33 2.2.4 Persson Model ................................ ................................ ........................ 34 2.3 Rubber Friction ................................ ................................ ................................ 38 2.4 Fluid Lubrication and Stribeck Curves ................................ .............................. 43 3 EXPERIMENTAL ................................ ................................ ................................ .... 46 3.1 Experimental Apparatus ................................ ................................ .................... 46 3.1.1 Microtribometer: Optical Module ................................ .............................. 46 3.1.2 Microtr ibometer: Motion Control Module ................................ .................. 48 3.1.3 Microtribometer: Loading Module ................................ ............................ 49 3.1.4 Measurement of Forces ................................ ................................ ........... 50 3.2 Experimental Procedures ................................ ................................ .................. 51 3.2.1 Load/Unload Testing ................................ ................................ ............... 51 3.2.2 Contact Area Variat ion with Velocity ................................ ........................ 52 3.2.3 Friction Testing ................................ ................................ ........................ 53 3.2.4 Lubricated Experiments: Stribeck Curves and Friction Testing ............... 54 3.3 Materials ................................ ................................ ................................ ........... 55 4 NITRILE RESULTS ................................ ................................ ................................ 57 4.1 Load/Unload Results ................................ ................................ ......................... 57 4.2 Contact Area Variation with Velocity ................................ ................................ 59 4.3 Friction Results ................................ ................................ ................................ 62

PAGE 6

6 5 BROM OBUTYL RESULTS ................................ ................................ ..................... 66 5.1 Load/Unload Results ................................ ................................ ......................... 66 5.2 Contact Area Variation with Velocity ................................ ................................ 69 5.3 Friction Results ................................ ................................ ................................ 73 5.4 Lubricated Results ................................ ................................ ............................ 76 5.4.1 Stribeck Curves ................................ ................................ ....................... 76 5.4.2 Lubricated Friction Results ................................ ................................ ...... 77 6 DISCUSSION ................................ ................................ ................................ ......... 80 6.1 Load/Unload ................................ ................................ ................................ ..... 80 6.2 Contact Area Variation with Velocity ................................ ................................ 83 6.3 Friction Experiments ................................ ................................ ......................... 85 6.4 Lubricat ed Testing ................................ ................................ ............................ 87 7 CONCLUSIONS ................................ ................................ ................................ ..... 90 APPENDIX A SURFACE ENERGIES ................................ ................................ ........................... 92 Bo rosilicate Glass ................................ ................................ ................................ ... 92 Cyclic Olefin Polymer ................................ ................................ .............................. 94 Bromobutyl Rubber ................................ ................................ ................................ 95 Nitrile Rubber ................................ ................................ ................................ .......... 96 B UNCERTAINTY IN CONTACT AREA MEASUREMENTS ................................ ...... 97 LIST OF REFERENCES ................................ ................................ ............................... 99 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 108

PAGE 7

7 LIST OF TABLES Table page 1 1 In situ techniques in tribology ................................ ................................ ............. 19 2 1 Plasticity index for various contact models ................................ ......................... 30 3 1 Microtribometer component sensitivities ................................ ............................. 49 3 2 Resolution of forces for various cantilevers used in experiments ....................... 50 3 3 Matrix of experiments ................................ ................................ ......................... 56 4 1 Average and standard de viation of contact area during steady state sliding of the nitrile samples run at 10 m/s and m/s ................................ ....................... 61 5 1 Average contact area and standard deviation during steady state sliding. ......... 70 6 1 Measured pull off forces for experimental systems ................................ ............. 80 6 2 Nitrile rubber friction results for all loads and counter surfaces .......................... 86 6 3 Bromobutyl results for all loads and counter surfaces ................................ ........ 87 A 1 Surface tensions of liquids used to create Zisman plot s ................................ ..... 92 A 2 Standard deviation of measured contact angles for liquids on borosilicate glass ................................ ................................ ................................ ................... 93 A 3 Standard deviation of measured contact angles fo r liquids on borosilicate glass ................................ ................................ ................................ ................... 94 A 4 Standard deviation of measured contact angles for liquids on bromobutyl rubber ................................ ................................ ................................ ................. 95 A 5 Sta ndard deviation of measured contact angles for liquids on nitrile rubber ....... 96

PAGE 8

8 LIST OF FIGURES Figure page 1 1 Classification of elastomers. ................................ ................................ ............... 15 1 2 Stress strain diagram for a natural rubber before and after vulcanization. ......... 17 1 3 Schematic of an in situ Raman tribometry experim ent. ................................ ...... 20 1 4 Schematic and image illustrating Newton's rings. ................................ ............... 22 1 5 Film thickness and friction behavior for in situ Newton's ri ngs study .................. 23 1 6 In situ SWLI tribometer. ................................ ................................ ...................... 24 2 1 The difference between real and apparent contact area for contacting surfaces. ................................ ................................ ................................ ............. 27 2 2 Schematic for Hertz's contact mechanics. ................................ .......................... 28 2 3 Comparison between Hertzian and JKR contact radii. ................................ ....... 31 2 4 Schematic of the M D model, with compressive and adhesive stresses. ........... 33 2 5 Adhesive contact map for various models ................................ .......................... 34 2 6 Surface roughness as magnification increases from a to c ................................ 35 2 7 Graphs from Persson showing the relationship between contact area, magnification and squeeze out force. ................................ ................................ 37 2 8 Schallamach's findings on the velocity and tempe rature dependence of friction ................................ ................................ ................................ ................. 39 2 9 Coefficients of f riction for spheres and cones in dry and lubricated sliding tests by Greenwood and Tabor ................................ ................................ .......... 40 2 10 Optical and mechanical schematics for the apparatus used by Barquins et al. .. 41 2 11 The contacting surface of a cylindrical sample. ................................ .................. 43 2 12 The different lubrication regimes. ................................ ................................ ....... 43 2 13 Example of a Stribeck curve. ................................ ................................ .............. 45 3 1 In situ optical tribometer. ................................ ................................ .................... 46 3 2 The optical module of the optical microtribometer. ................................ ............. 47

PAGE 9

9 3 3 The motion control module of the optical microtribometer. ................................ 48 3 4 The z stage assembly, or loading modu le, of the optical mircotribometer. ......... 49 3 5 Behavior of materials with and without adhesion. ................................ ............... 52 4 1 Contact area of nitrile rub ber against borosilicate glass (above) and cyclic olefin polymer (below) at various loads. ................................ ............................. 57 4 2 Results from the load/unload experiments with nitrile rubber ............................. 58 4 3 Individual cycle results for a sample of nitrile sliding at two velocities ............... 59 4 4 The average co ntact area and friction forces for three samples of nitrile rubber run at 10m/s and 50 m/s ................................ ................................ ..... 60 4 5 Contact images during sliding of Buna N sample one for 10 m/s (top row) and 50 m/s (botto m row). ................................ ................................ .................. 61 4 6 Coefficient of friction curves for nitrile on borosilicate glass for various loads .... 63 4 7 Averages of coefficie nts of friction for various loads of nitrile on borosilicate glass. ................................ ................................ ................................ .................. 63 4 8 Coefficient of friction curves for nitrile on cyclic olef in polymer for various loads. ................................ ................................ ................................ .................. 64 4 9 Averages of coefficients of friction for various loads of nitrile on cyclic olefin polymer. ................................ ................................ ................................ .............. 64 4 10 Break loose and gliding coefficients of friction for nitrile at normal loads of 100 mN.. ................................ ................................ ................................ ............. 65 5 1 Bromobutyl sample used for experiments ................................ ......................... 66 5 2 Bromobutyl contact area on borosilicate glass (above) and cyclic olefin polymer (below) for various normal loads. ................................ .......................... 67 5 3 Results from the load/unload experiments with bromobutyl rubber. .................. 68 5 4 Individual cycle results for a sample of bromobutyl sliding at two velocities ...... 69 5 5 The average contact area and friction forces for three samples of bromobutyl rubber run at 10m/s and 50 m/s ................................ ................................ ... 71 5 6 Contact images during sliding of the third bromobutyl sample for 10 m/s (above) and 50 m/s (below). ................................ ................................ ............. 72

PAGE 10

10 5 7 Coefficient of friction curves for bromobutyl on borosilicate glass for various loads. ................................ ................................ ................................ .................. 73 5 8 Averages of coefficients of friction for various loads of bromobutyl on borosilicate glass. ................................ ................................ ............................... 74 5 9 Coefficient of friction curves for bromobutyl on cyclic olefin polymer for various loads. ................................ ................................ ................................ ..... 75 5 10 Averages of coefficients of friction for various loads of bromobutyl on cyclic olefin polymer. ................................ ................................ ................................ .... 75 5 11 Stribeck curve for bromobutyl on borosilicate glass with 1000 cSt silicone oil lubrication. ................................ ................................ ................................ .......... 76 5 11 Stribeck curve for bromobutyl on cyclic olefin polymer with 1000 cSt silicone oil lubrication. ................................ ................................ ................................ ...... 77 5 12 Friction results for bromobutyl on lubricated glass at various normal loads and sliding velocities. ................................ ................................ .......................... 78 5 13 Friction results for bromobutyl on lubricated cyclic olefin polymer at various loads and sliding velocities ................................ ................................ ................ 79 6 1 Experimental contact area data for nitrile as compared to apparent contact areas predicted by the JKR and Hertz models. ................................ .................. 82 6 2 Experimental contact area data for bromobutyl samples as compared to apparent contact areas predicted by the JKR and Hertz models ........................ 83 6 3 Relationship between friction force and normal force. ................................ ........ 85 6 4 Schematic of material in fluid of thickness h and sample bearing area curve to relate percent of mate rial in contact to film thickness. ................................ .... 88 6 5 Relationship between the friction force, normal force, and velocity for bromobutyl on borosilicate glass with 1000 cSt silicone oil lubricatio n. .............. 89 6 5 Relationship between the friction force, normal force, and velocity for bromobutyl on borosilicate glass with 1000 cSt silicone oil lubrication. .............. 89 A 1 Zisman plot for borosilicate glass. ................................ ................................ ...... 93 A 2 Zisman plot for cyclic olefin polymer. ................................ ................................ 94 A 2 Zisman plot for cyclic olefin polymer. ................................ ................................ .. 95 A 4 Zisman plot for nitrile rubber. ................................ ................................ .............. 96

PAGE 11

11 B 1 Simulated discretization errors for contact area as a function of ................... 97

PAGE 12

12 LIST OF ABBREVIATION S AFM atomic force microscopy COC cyclic olefin polymer EDX energy dispersive X ray spectroscopy EHL elastohydrodynamic lubrication FIB focused ion beam FTIR Fourier transform infrared spe ctroscopy RMS root mean square SEM scanning electron microscope SFA surface forces apparatus SWLI scanning white light interferometer TEM transmission electron microscope XPS X ray photoelectron spectroscopy

PAGE 13

13 Abstract of Dissertation Presented to the G raduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy IN SITU MICROTRIBOLOGY OF ELASTOMERIC MATERIALS By Jennifer Vail May 2012 Chair: Wallace Gregory Sawyer Major: Mechanical Engineering Rubber systems have been studied since the early twentieth century These systems are challenging to characterize given the performance dependence on factors such as adhesion and energy dissipation. Further complicating this is the wide variety of material recipes, two of which were studied here. An in situ optical microtribometer developed to examine contact between two solids was used to garner insight into the behavior of these elastomeric materials. This work describes an experi mental study of two rubbers commonly used in sealing applications to provide insight into elastomeric tribology. Experiments were executed to study contact area of bromobutyl and nitrile rubber against glass and polymer counter surfaces at various loads and sliding velocities During loading of 5 mN, 10 mN, 15 mN, and 20 mN both materials experienced a nearly linear increase of real contact area with load and a hysteresis during unloading, resulting from the increase in energy required to separate the s ample from the counter sample. This behavior in viscoelastic materials has been explained using crack propagation theory. The contact area during sliding at 10 m/s and 50 m/s decreased with higher friction for both materials due to the higher friction force causing the material to shear more Higher friction was recorded at 10 m/s for the nitrile rubber, whereas

PAGE 14

14 the bromobutyl rubber experienced higher friction at 50 m/s The bromobutyl rubber followed previous studies that found an increase in rubb er friction with increasing velocity; for the sliding speeds used here, the nitrile undergoes behavior dominated by the crack propagation theory for viscoelastic materials. Velocity dependence of the friction behavior of the nitrile rubber was not seen un til the velocity was 1 mm/s. Friction studies were performed to study the gliding and break loose coefficients of friction for the nitrile and bromobutyl samples at normal loads of 5 mN, 10 mN, 15 mN, 20 mN, and 100 mN at a sliding velocity of 100 m/s. T he nitrile rubber coefficients of friction were consistently lower than that of the bromobutyl rubber for both counter surfaces. The bromobutyl rubber did not achieve break loose and required lubrication for gliding The presence of silicone oil significan tly reduced the break loose friction coefficient of the bromobutyl systems and allowed gliding at loads of or greater than 100 mN and sliding velocities of or greater than 100 m/s.

PAGE 15

15 CHAPTER 1 INTRODUCTION Soft materials have been studied since the ninetee nth century and have gained significant ground in technology in recent years being used as packaging materials, adhesives, and biological materials These materials behave in a way that is often unpredictable and difficult to model. Loosely described as materials that are easily deformable, soft materials include polymers, colloids, amphiphiles, and biomaterials. The study of soft biomaterials has developed in modern years, whereas polymers, dispersed colloids and amphiphilic systems have been studied s ince the 19 th century [1] Figure 1 1. Classification of elastomers Elastomers, including rubbers, are soft polymers (Fig. 1 1) with low moduli, typically less than 10 MPa [2] Rubbers and elastomers are similar materials; they differ in their ability to recover from extension. Roughly de fined, rubbers are a subset of elastomers that have the ability to stretch to double their length or longer, then quickly retract to the original dimension. Other elastomers cannot extend to the same degree a s rubbers with full recovery [3]

PAGE 16

16 transition temperature, allowing for motion of the long polymer chains and offering the flexibility as sociated with rubbers. This motion, which produces elastic deformation, can be the result of the chains uncoiling, untwisting or straightening. When the material recovers, the chains recoil, kink or twist to their original state. The material also must b e amorphous, with natural kinks and coils; crystalline chains will not have these features and los e the ability to easily move [4] Most rubbers have crosslinked chains, althoug h the thermoplastic elastomers (such as styrene butadiene styrene) do not. Rubber compounds often contain other materials to influence properties and processing. Natural rubber and butyl rubber are two rubbers that have good mechanical properties without reinforcement but almost all rubbers require a filler material to improve these properties. Carbon blacks of diameters from 20 50nm are commonly used reinforcing fillers that reduce tears and abrasion while increasing tensile strength. Carbon blacks for m strong covalent bonds with rubber polymers due to hydroxyl end groups (basic blacks) or carboxylic acids (acid blacks). Rubbers can contain up to fifty parts per hundred parts rubber by weight of carbon black filler (tires contain 15 30 vol%) [3,4]. Pl astic fillers may also be used to reduce the cost of the rubber but often at the cost of the modulus Hydrocarbon oils are also used as fillers to reduce the cost of the rubber compound but also serve to soften the material to make processing easier. Carb on blacks and hydrocarbon oils are competitive with regards to the elastic modulus. Often the oil is added to increase the molecular weight and for polymerization, then carbon black is added to increase the subsequently reduced modulus [3].

PAGE 17

17 Curing systems and accelerators are added to rubber compounds to crosslink the system and speed up vulcanization, the irreversible process of crosslinking rubbers. Vulcanization is needed to give rubbers their usable mechanical properties; Fig. 1 2 shows the difference in performance between a vulcanized and unvulcanized rubber. Crosslinks occur in carbons that have become single bonds after vulcanization; without this process, rubbers will be soft and easily abraded. These curing systems can include sulfur, zinc oxide and free radical initiators [4]. Stabilizers are used in natural rubbers, butadiene containing, and isoprene containing rubbers because they are unsaturated and will react with oxygen. Finally, pigments can used to dye rubbers, although this is only whe n mechanical strength is not the primary concern for the material [3]. Figure 1 2 Stress strain diagram for a natural rubber before and after vulcanization Rubbers are commonly seen in tribological systems as seals and gaskets and are widespread in everyday use. Automotive tires must continue gripping road surfaces in rain and ice; in fact, it is desirable for the coefficient of friction of the tire to increase in the presence of moisture. Smooth operation of rubber parts is equally important:

PAGE 18

18 automobile wiper blades must clear water and debris from window glass without sticking and stoppers must prevent leakage in or out of a syringe and smoothly execute a vaccine. The aforementioned examples include a rubber surface slidi ng against a hard surface. This frictional behavior is of practical interest and differs from most solid material systems due to the low elastic modulus of rubbers. The nature of rubber during tribotesting has been widely speculated and could be further understood by employing in situ techniques to directly observe the rubber sample during testing. Tribological data obtained in situ offers researchers the ability to correlate interface events with friction measurements. In situ measurements are also de sirable since the sample can remain in the test environment, minimizing contamination [5] A tribometer is an instrument that measures friction forces and has been coupled with a varie ty of in situ techniques in order to move toward understanding physical phenomena during testing. Tribologists have e mployed thermal measurements [6] thermoelectric, contact resistance, and interferometric techniques in situ [7] Table 1 1 summarizes the various tools used for in situ studies. Challenges with in situ tribology include experimental apparatus design and material selection, which can be limited by the in situ technique. Microscopy and spectroscopy are two popular techniques used for in situ tribological studies. In order to use th ese, measurements must be made through an interface or on a surface parallel to sliding. When visible light is used, at least one surface must be transparent.

PAGE 19

19 Table 1 1 : In situ techniques in tribology Measurement Technique Cont act area Optical microscopy Interferometry Tribofilm formation/thickness Optical microscopy Raman microscopy Contact resistance Composition Raman microscopy ATR FTIR TEM SEM/EDX SEM+FIB XPS Wear AFM Interferometry In the early 19 90s, Belin and Martin d a tribometer with an image processor. Two local numerical images were collected: one of the evolution of the friction force and another of the electrical resistance to the interface. The variations of the sig nal were used to study degradation of thin films [8 11] The thickness of interfacial films cannot be monitored using this technique, although changes such as oxidation in the film will be detected in the measured contact resistance. Separately, Bredell et al. developed another device to measure friction and surface film resistivity in situ to determine contact as a func tion of applied normal load [12] The surface forces apparatus (SFA) is an instrument used to measure sure forces along with contact area and surface separation. The SFA was originally developed to study Van Der Waals forces in liquid and colloidal systems but has been expa nded to investigate molecular motions of systems. Two cylindrical surfaces, set 90 degrees to each other, move into contact and are retracted. One surface is attached with a cantilever whose deflection is used to determine the appl ied force. Interferome tric SFA

PAGE 20

20 use s multiple beam interferometry, sending a collimated beam of white light through a test surface which is typically mica. A metal layer on the reverse side of the mica acts as a mirror and interfero metric fringes are produced [13 16] These fringes can then be used to determine separation distance and contact area. Raman spectroscopy and x ray diffraction have been used in conjunction with SFA for in situ chemical analysis but the nature of the test sys tems SFA requires (thin sheets of mica) make this technique challenging and limited in use [17 27] Raman spectroscopy, as well as Fourier transform infrared (FTIR) spectroscopy, are popular in situ chemical analys is technique s for tribological studies [28 38] A monochromatic light source is focused on a sample and the inelastic scattering of light is measured (Fig. 1 3). The laser photons will be shifted up or down in en ergy, giving a chemical signature for the material. While the experimental setup is not as limited as SFA, considerations must be made with respect to excitation wavelength, the test Figure 1 3 Schematic of an in situ Raman tribometry experiment

PAGE 21

21 In situ Raman tribometry was used to confirm a hypothesis about the presence of boric acid in low friction systems containing boron oxide [39] Boric acid forms r eadily in humid environments and it was believed that a thin layer reduced the friction of the system until the layer was worn away. During testing of an oxidized B 4 C substrate, the Raman signal for boric acid was detected on the sliding interface when th ere was low friction and disappeared once the friction increased. Additionally, these tests demonstrated the presence of a second lubricating regime due to carbon produced during the oxidation process of the B 4 C [39] Molybdenum disulphide (MoS 2 ), a popular solid lubricant, has been studied using in situ micro Raman tribometry. MoS 2 is often used in coatings to improve friction and wear properties and it was seen that after low friction was achieved, crystalline MoS 2 formed on the sliding interface within the fir st 66 cycles [31] This study confirmed that ordered crystalline MoS2 produced low friction behavior in the amorphous Pb Mo S coating being tested. Another in situ technique uses focused ion beam (FIB) sectioning and examination by transmission electron microscopy (TEM) to capture the chemistries of wear scars. Hu et al. developed a microtribometer inside a FIB microscope with scanning electron microscopy (SEM). The wear debris was monitored via SEM and the FIB prepared in situ cross section TEM specimens. This t echnique allows insight into chemical and mechanical chan ges that occur during sliding [38] It is believed that in situ TEM studies may answ er fundamental tribological questions, such as how sliding solid lubricants accommodate motion [5]

PAGE 22

22 Optical interferometry can be used in situ to monitor transfer film formation and thi ckness, as well as real contact area of a sample [40, 41] As mentioned previously, this technique has been used with SFA [13, 17, 21, 42 44] and color interferometry has been used to monitor changes in liquid systems [45 47] seen outside the area of contact, can be used to determine separation distances and film thicknesses. The height between fringes is found by dividing the wavelength by two, and fr om this, the thickness of a film can be determined (Fig. 1 4) [48] The sensitivity of the calculated thickness is determined by the resolution of the detector, typically a charge coupled device (CCD) camera. This technique can also be limited by the index of refraction of materials and optical clarity of the contact. Figure 1 4 Schematic and image illustrating Newton's rings Figure 1 4 shows the experimental results of an in situ optical interferometry test hire ball in contact with a substrate coated with MoS 2 /Ti and the resulting optical image is used to find the tribofilm thickness. As the resulting transfer film forms and thickens, the image fringes shift inward; if the film thins, the fringes will shift outward. The friction can be monitored in correlation to transfer film thickness. The behavior of the fringes can help in understanding the relationship between transfer film formation and friction behavior.

PAGE 23

23 Figure 1 5 Film thickness and friction behavior for in situ Newton's rings study. [Reprinted with permission from K.J. Wahl and W.G. Sawyer, Observing interfacial sliding processes in solid solid contacts, MRS Bull. 33 (2008) 1163, Figure 3.] A tribometer equipped with a scanning white light interferometer can be used for in situ wear monitoring of materials. Mauntler et al. developed such a device and equipped it with a chamber to control the environment (Fig. 1 6). This technique also allows study of materials and coat ings with very low wear. The normal and friction forces are measured with load cells as a pin is loaded against a counterface. After a prescribed number of cycles, the counterface is moved under the interferometer and scans of the wear scar or transfer f ilm can be taken. This may be done cycle by cycle to watch the evolution of wear, allowing observation of behavior in both the initial and transient state of testing [5, 49] A tribometer coupled with in situ optical interferometry will offer insight into the behavi or of elastomeric materials during testing. Much has been hypothesized and modeled regarding rubber contact and friction. Most theories depend on the contact area of the rubber but use assumptions rather than actual data. Likewise, it has been suggested that the coefficient of friction changes with sliding velocity due to contact

PAGE 24

24 area changes. The recent developments of in situ tribometry can be applied to experimentally examine these theories. Figure 1 6 In situ SWLI tribomet er

PAGE 25

25 CHAPTER 2 BACKGROUND 2.1 Rough Surfaces Real surfaces are comprised of atoms that rarely lie flat and parallel to the surface. Surfaces instead are modeled using the terrace ledge kink (TLK) model. Surfaces often form terraces by atoms deviating from parallel, ledges appear when atoms are missing at the end of a terrace and create a step, and kinks result from imperfections such as an interstitial atom [50] As a result of these surface features, atoms at the sur face ha ve weaker bond strength s enabling them to react with another surface or a lubricant [51] (A schematic repre senting contact between two real surfaces is in Fig.2 1 in the next section). Surfaces are characterized using statistical parameters and multi scale techniques. Two statistical parameters are used to define a surface: the height and spatial parameters. The average roughness (R a ) is predominantly used to describe surface roughness, although the root mean square roughness (RMS or R q ) is also commonly used. The R a is an average roughness calculated for a given sampling length. A mean line through the sa mpling length is placed in order to distribute equal areas of the profile above and below the line. R a (Eqn. 2 1) is then defined as the average deviation of the surface height from this line, with y(x) representing the height above the mean line [52] (2.1) R q is the root mean square deviation from the mean line; if the surface height has a Gaussian distribution, R q =1.25 Ra.

PAGE 26

26 (2.2) These equations often result in topographical features being averaged out, thus R a and R q do not fully describe surface shapes and features. The spatial distribution of surface features is needed for a more complete topographical description. Spatial variation is often descri bed using either the autocovariance or autocorrelation function [53] The same surface can yield different values for R a or R q depending where the sampling length is taken from; li kewise, surfaces with different features may have similar values for R a due to averaging. Multi scale techniques have recently been developed in an effort to better characterize surfaces. These techniques include Fourier transforms [54, 55] wavelet transformations [56] and fractal methods [57 61] 2.2 Contact M echanics The contact area of a material will influence heat transfer, resistivity, adhesion and uation for wear of a material begins with the contacting of asperities [62] Contact mechanic s theories seek to model the behav ior of materials depending on their surface profiles, load parameters, and elastic moduli. A critical component is the real contact area versus the apparent contact area, which was investigated starting in the 1950s by Dyson and Hirst [63] Figure 2 1 demonstrates the contact between surface asperities and the real contact area. Idealized rough surfaces are modeled using spherical asperities. Simple models originally used spheres of constant radii and height. When placed into contact with another surface, it was assumed the spheres (or asperities) would deform independently [64] The normal load would then be distributed among the asperity

PAGE 27

27 contacts and this load in turn produces an ide ntical contact area for each asperity. The total contact area is the summation of the individual asperity contact areas. This contact area influences the friction of rubbers, as well as the ability of the material to squeeze out fluid and seal. Figure 2 1. The difference between real and apparent contact area for contacting surfaces The roughness of a solid material contributes, but is not solely, responsible for the contact behavior of the material property. The material properties, externally appli ed load, and adhesive behavior also influence the contact behavior of a sample. Hertz and Reine, in 1882, published their pioneering work in the field of contact mechanics, or the study of solid materials in contact with one another [65] Hertz developed the first model of contact between two bodies contact model utilizes two elastic spheres in contact, with the resulting contact area being dependent on the forces pressing the bodies into contact, the elastic properties of the materials and the geometries of the bodies [52] contact area is shown in Eqn. 2 3.

PAGE 28

28 Figure 2 2. Schematic for Hertz's contact mechanics (2.3) Where a is the radius of the contact circle, P is the applied l oad, r is the equivalent radius of the spheres (Eqn. 2 Eqn.2 5: (2.4) (2.5) Where r is the radius sphere, with 1 and 2 indicating each sphere. The Hertzian model is for an ideal system and does not account for roughness of the two contacting materials. Greenwood and Williamson proposed a statistical approach to consider a rigid plate in contact with a deformable surface [66] The contact area and load can be computed using the separation distance between surfaces and height distribution of asperities. The key parameters of the Greenwood Williamson model are the curvature of asperities and standard deviation o f surface roughness. The height distribution must be statistically valid and asperities are assumed to be spherical

PAGE 29

29 with a constant curvature. Deformations are considered independent from asperity to asperity [67] Equations 2. 6 2.8 show the contact area, normal load supported, and asperity height dis (2.6) (2.7) (2.8) given by Eqn. 2 9, to describe deformation on a rough surface. (H is the indentation hardness will be flattened. (2.9) Later studies refined the Greenwood Wi lliamson model, primarily focusing on the assumptions of spherical asperities acting independently. Mikic suggested the contact between materials would depend on the slope of the asperity and modified the plasticity index accordingly [68] Table 2 1 shows the plasticity index as it evolved with refined contact mode ls [53]

PAGE 30

30 is the asperity curvature of the harder surface, p s is a pressure, and R q is the RMS surface roughness. As new contact models were proposed, the model asserted by Greenwood Williamson was shown to be a sound approximation for rough surfaces in contact. Table 2 1. Plasticity index for various contact models Surface contact model Plasticity index Greenwood Williamson (GW) Whitehouse and Archard (W A) Bower and Johnson (B J) 2.2.2 Adhesive C ontacts: The JKR and DMT Models To determine the force required to separate two rigid spheres in contact, Bradley solved for the potential between the spheres using the Lennard Jones potential to govern surface forces. The potential between atoms will vary as x n over a distance x; the gap between the spheres will vary with d (n+5) From his solution, he found that when [69] where R is the equivalen t radius. The work of adhesion, or surface energy, is and depends on the individual surface energies as well as the energy between the two surfaces (Eqn. 2.10). (2.10) s arose in rapid succession to address deformable contact between spheres. These theories incorporated adhesion into Hertzian contact. Johnson, Kendall and Roberts proposed

PAGE 31

31 what is now referred to as the JKR model in 1971 and Derjaguin, Muller and Tpolov presented their DMT model in 1975 [67] Johnson, Kendall and Roberts used surface energy to describe contact between elastic materials [70] They demonstrated that at light loading, attractive forces between surfaces play a significant role in contact mechanics. When zero surface forces are applied, the contact radius is equi valent to that given by Hertz. The contact radius (a) will increase with an attractive force between surfaces; this contact is adhesive. Figure 2 3. Comparison between Hertzian and JKR contact radii This new radius was defined as a 1 (Eqn. 2.11), where instead of using the applied load F n the apparent load F 1 (Eqn. 2.12) is used and elastic material properties are taken into account. (2.11) (2.12) The JKR model uses elastic parameters k 1 and k 2, which combine for the parameter K:

PAGE 32

32 (2.13) (2.14) In the presence of surface forces, the total energy of the system must be considere d: stored elastic energy, mechanical energy, and surface energy. In equilibrium, the apparent load is shown to be larger than the applied load. When the surface energy is reduced to zero, Eqn. 2.15 is reduced to the Hertz equation. (2.15) The force to pull the surfaces apart is shown in Eqn. 2.16, with R again being the equivalent radius. (2.16) Derjaguin et al. approached contact deformations in a different manner than JKR and utilized thermodyn amics [71] JKR modified the Hertzian contact profile, whereas the DMT model maintains it. It assumes inte ractions outside this region are due to Van der Waals forces. The result was a pull off force which differed from that of the JKR model: (2.17) The parameter (Eqn. 2.18) is used to determine if the JKR or DMT model applies to a given system by comparing the surface forces (adhesion) to the amount of elastic deformation. The parameter depends on the equivalent radius, composite modulus, surface energy, and the distance between the surface z o Small values of (<0.1) apply

PAGE 33

33 to small, rigid solids where elastic deformation is negligible and the DMT model is used [72] (2.18) 2.2.3 Modeling Between the DMT and JKR Limits The JKR and DMT mod els two extremes of contact; a closed for m solution developed by Maugis spans the intermediate regime. This analysis used the Dugdale approximation that states until a certain separation distance is reached, the intensity of the adhesive force is constant The Maugis Dugdale model is comprised Hertzian contact (compression, P 1 ) and Dugdale adhesive stress (p a ) [73] Figure 2 4. Schematic of the M D model, with co mpressive and adhesive stresses Two parameters are used to express the M which is related to the Tabor coefficient () [74] (2.19) (2.20)

PAGE 34

34 determine what model is appropriate to use and is shown in Fig. 2 5 [74] Fig ure 2 5. Adhesive contact map for various model s. [Reprinted with permission from K.L. Johnson, Mechanics of adhesion, Tribol. Int. 31 (1998) 414, Figure 3 ] The Maugis Dugdale can be cumbersome to solve and was simplified by the Carpick Ogletree Salmero n (COS) method. The contact radius is defined in Eqn. 2.21 w to to the JKR case ( is a transition The intermediate regimes (between 0 and 1) can be solved using the COS approximation for the contact radius [75] where a 0 is the radius at zero load, L is the applied load, and L c is the negative critical load. (2.21) 2.2.4 Persson Model Many contact mechanics theories utilize the Hertzian model of spherical surface asperities and neglect elastic coupling between contact regions. For this to be valid, the normal load must be low en ough that the distance between contact regions is large enough to be able to decouple the regions. In this case, the real area of contact will be

PAGE 35

35 much smaller than the nominal area. Models such as Greenwood Williamson use this method but a model is neede d to address contact mechanics at high loads. The contact mechanics theory of Persson aims to cover all normal loads, or squeezing, forces, and particularly holds true for the cases when the real contact area approaches the nominal contact area and may be applied when adhesion is present [76] The surface roughness is represented using a Fourier transform power spec trum found using Eqn. 2.22. The height of the surface is z, where z=h(x) is the height of the surface at x, < > represents the ensemble average and q is a wavevector. Surfaces are described as self affine fractals, meaning the statistical properties are unchanged for the surface at any scale [76] (2.22) contact. At low magnifications, contact will appear to occur at the macroasperity level but at higher magnifications it is seen that there is only partial contact occurring (F ig. 2 6). Figure 2 6. Surface roughness as magnification increases from a to c The stress distribution P stress at the interface and A 0 is the nominal contact area. This equation can be rewr itten to solve for the real area of contact (Eqn. 2.24).

PAGE 36

36 (2.23) (2.24) The theory assumes this equ ation will also be satisfied for partial contact. The two boundary conditions necessary to solve this equation for elastic contacts are: nce there will be no tensile stresses present in that case. Figure 2 7 shows the dependence of contact area with respect to magnification and effect of the area on squeeze out. If the squeeze out force is small enough that the real contact area is much l n predicting that the contact area will linearly depend on the normal load as other theories such as Archard [77] have also predicted. This approach accounts for elastic coupling and potentially remains valid for all magnifications, whereas other theories break dow n at high magnifications [76] For the case of a nominally flat elastic solid against a rough, hard surface wit h no equating the load to the contact pressure integrated over the apparent con tact area.) The total projected area in this case will be equal to the sum of the elastic contact area

PAGE 37

37 and non contact area. An equation to describe the flow of the area from contact to non contact as magnification increases can be written as Eqn. 2.25 [76] (2.25) Figure 2 7. Graphs from Persson showing the relationship between cont act area, magnification and squeeze out force. [Reprinted with permission from B.N.J. Persson, Contact mechanics for randomly rough surfaces, Surf. Sci. Rep. 61 (2006 ) 208, Figure 10.] The inclusion of adhesion to this model changes the boundary conditio ns of the adhesion a The stress a Crack theory is used to calculate the detachment stress:

PAGE 38

3 8 (2.26) The normal force for the a 0 P( a the force on the contact cannot equal the applied external load which is only possible when there is adhesive attraction acting between the surfaces. The Persson model shows that the adhesive forc es exist not only in the contact area, but outside as well. This theory is critical t o understanding rubber friction 2.3 Rubber Friction The coefficient of friction is defined as =F n /F f where F n is the applied normal force and Ff is the frictional for ce. Pioneering work in rubber friction began in the early 1900s; Ariano in 1930 showed that as sliding velocity increases, the coefficient of friction of rubber materials will also increase [78, 79] The later work of Roth et al. in 1942 [80] and Schallamach [81] in 195 3 confirmed rubber friction as a rate dependent quantity. Studies by Grosch (1962), where a variety of rubbers slid against hard surfaces, found that the velocity dependence of rubber friction has a maximum; for low velocities, he suggested the friction was independent of the velocity. It was also suggested that when the sliding velocity was above a few centimeters per second, frictional heating must be considered [82] theory that the velocity and temperature dependence of friction behavior obeys the same laws as viscoe lastic properties shown in Figure 2 8 is molecular dimensions, F represents the friction force, infinite rate of deformation, N 0 is the number of available bonding sites,

PAGE 39

39 ti shown relate the friction force to only sliding velocity (the solid line) and temperature dependence (dotted line). From these studies, two mechanisms were thought to cont ribute to rubber friction: interfacial adhesion and ploughing (or deformation) of the rubber surface [83] Grosch was the first to suggest that the coefficient of friction of rubber materials is a bulk property. Figure 2 8. Schallamach's findings on the velocity and temperature dependence of friction [Reprinted with permission from A. Schallamach, A theory of dynamic rubber friction, Wear 6 (1963) 380, Figure 2.] Material deformatio n, or ploughing, was introduced as an origin of rubber friction by Greenwood and Tabor (1957). Using hard spheres against lubricated rubber counter samples, large deformations were observed as a result of hysteresis losses in the rubber due to internal fr iction [84] As asperities slide against a counter sample, oscillating forces produce cyclic deformation that results in energy dissipation [85] The authors likened it to rolling friction and noted that when interfacial adhesion is low, friction is dominated by deformation losses.

PAGE 40

40 Figure 2 9. Coefficients of f riction for spheres and cones in dry and lubricated sliding tests by Greenwood and Tabor [Reprinted with permission f rom J.A. Greenwood and D. Tabor, The friction of hard sliders on lubricated rubber: The importance of deformation losses, Proc. Phys. Soc. 71 (1957) 997. Figure 6 ] The dependence of rubber friction on pressure was a topic of dissention among early studie s. Rolling friction will increase as pressure increases [86] but Schallamach believed elastic ma terials under low pressure would have higher coefficients of friction The coefficient of friction of an unlubricated sample increased as the normal load (W) decreased. Greenwood and Tabor found that the friction due to deformation losses depends on W 4/3 Subsequent studies revealed that when the normal load is very low or very high, the coefficient of friction will be independent of load. At low loads, adhesion is significant; at high loads, the real contact area approaches the apparent contact area [85] Greenwood and Tabor also asserted that for elastic materials, the real contact area is critical to the outcome of overall friction beh avior. This area will determine the friction from shearing and interfacial adhesion and becomes more important as the load is reduced. An experimental setup with a glass hemisphere against a natural rubber counter sample was used to determine contact are a [87] An optical microscope was area calculations.

PAGE 41

41 The experimental results were compared to both Hertzian theory and the JKR theory and ultimately verified the JKR model. Figure 2 10. Optical and mechanical schematics for the apparatus used by Barquins et al. [Adapted with permissi on from M. Barquins and R. Courtel, Rubber friction and the rheology of viscoelastic contact, Wear 32 (1975) 134 135, Figures 1,2.] The same apparatus was used to observe at the contact during sliding at slow speeds (approximately a few micrometers per s econd). Upon sliding, the area of contact moved as a whole, which the authors suggested was due to adhesive forces maintaining closeness between surfaces. A line traced was traced on the rubber surface perpendicular to the sliding direction; when sliding ceased, the line continued moving, indicating interfacial relaxation. The authors noted at the conclusion of this study that experiments using rubber balls on various counter samples were crucial to understanding rubber behavior. They later published up dated work with an inverted experimental setup and used a glass lens, wavy glass, and abrasive paper as counter samples against a natural rubber [88]

PAGE 42

42 contact area calculations. Another apparatus was used for friction studies wit h the wavy glass: a rubber pad was held on a lever arm and the counter sample mounted on a turntable. Tests were designed to study the influence of velocity, normal load and radius of curvature. Various radii of rubber balls were used and the coefficient of friction increased with system showed little variation with velocity at low speeds. At higher sliding speeds, the coefficient of friction increased with velocity and was higher for the smooth glass than the wavy glass or abrasive paper [88] The variation in velocity data may be due to changing interfacial shear strength or time dependent adhesion. The contact area strays from Hertzian theory for higher velocities, indicating possible hysteresis loss or contracting rubber a t these speeds. As the velocity increases, the elastic properties will stiffen and reduce contact area [85] Rubber flats did not show s ensitivity to velocity, indicating that the geometry of contact is important to behavior. To break contact of curved surfaces, the force required depends on the material properties in a manner dependent on the geometry. The pull off force of spherical co ntacts has been found to be independent of elastic properties (K) as long as the work of adhesion is constant. Cylindrical contacts see a pull of force proportional K 1/3 [89] The cylindrical contact is conformal along the lateral axis and has a smaller contact width at separation compared to the cylindrical contact (Fig. 2 11). Compar ed to hard solids, the real contact area of rubber is much larger and will grow with time. Hard surfaces will form junctions of contacts resulting from randomly

PAGE 43

43 distributed asperities. These junctions are elastically deformed and will pop apart when sepa rated. Rubber contacts will instead elongate before eventually breaking apart; the chains will also constantly move to minimize energy. At low velocities, the rubber will fill out surface cavities on the short wavelength surface roughness due to adhesive forces, thus increasing friction. If adhesion can be reduced, friction will follow suit [85, 90] Figure 2 11. The contacting s urface of a cylindrical sample [Reprinted with permission from M.K. Chaudhury and T. Weaver, Adhesive contact of cylindrical lens and a flat sheet, J. Appl. Phys. 80 (1996) 31, Figure 1.] 2.4 Fluid Lubricati on and Stribeck Curves In fluid lubricated systems, there are three primary regimes: boundary, elasto hydrodynamic, and hydrodynamic lubrication (Fig. 2 12). Boundary lubrication occurs at low sliding velocities or under high contact pressures. A thin fi lm of lubricant is not maintained under these conditions and asperities will remain in contact, resulting in higher coefficients of friction [52] Figure 2 12. The different lubrication regimes

PAGE 44

44 Fewer asperi ties make contact as the fluid film thickens. The system undergoes hydrodynamic lubrication when a thick fluid film separates sliding surfaces. The viscous forces in the film provide pressure to support the normal load. As the speed or load in the system is increased during hydrodynamic sliding, the coefficient of friction will increase due to drag [52] The regime between boundary and hydrodynamic lubrication is known as elastohydrodynamic (EHL) lubrication. Non conforming surfaces produce higher dependence on pressure is important. Rubber materials undergo what is referred to as viscosity of the lubricant remains constant. Hard EHL applies to materials with a higher elastic modulus. The fluid film thickness in hard EHL is typically thinner than that of hydrodynamic lubrication and on the order of tenths of micrometers to a few micrometers. Depending on the surface (2 27) When >3, the surfaces are separated by a full flui d film. Between 1< <3, EHL is mixed. The Stribeck curve (Fig. 2 13) is a tool used to determine the lubrication regime when fluid is present. The coefficient of friction is plotted against the product of the viscosity and sliding speed divided by the n ormal load ( ). This curve can be created for materials by varying the applied normal load and/or the sliding speed. Stribeck

PAGE 45

45 originally created this curve by running experiments on lubricated journal bearings, using two hard su rfaces present with fluid lubrication. A Stribeck curve involving a system of rubber against a hard surface is a useful tool for modeling and predicting behavior in applications such as road tires and rubber seals [91] Figure 2 13. Example of a Stribeck curve

PAGE 46

46 CHAPTER 3 EXPERIMENTAL 3.1 Experimental Apparatus A custom built in situ optical microtribometer (micro Newton force resolution), shown in Figure 3 1, was used to perform the experiments described in this doc ument. The tribometer operates in a linear reciprocating mode with the option of obtaining in situ imaging during experimentation. The tribometer is comprised of three modules: motion control, loading, and optical. Figure 3 1 In situ optical tribomet er 3.1.1 Microtribometer: Optical M odule The optical module was designed by CSM instruments and is shown below (Fig. 3 2) and is comprised of an LED, mirror, beam splitter, objective, and detector. A collimated beam of light shines through the objective onto the underside of the sample

PAGE 47

47 which is in contact with the counter sample. An orange LED with a wavelength of 596nm with a full width half max of 16nm is used for these experiments to create a crisp distinction between the LED and dark contact regions A typical microscope objective is used to magnify the contact and a digital video camera is used to record the image. The CCD camera has square pixels and can record up to fifteen frames per second. Using a 10x objective, the images have a micrometer t o pixel ratio of 0.350 m/pixel. Figure 3 2. The optical module of the optical micro tribometer The optical microtribometer utilizes the same principles as described in Chapter 1. The beam of light will pass through the glass lens and into the glass ai r (or glass lubricant) boundary. The index of refraction will increase here, causing no phase shift to the light and it will reflect against the sample in contact. As the reflected light travels back through the air glass boundary to the glass, the index of refraction decreases, causing a phase shift. This phase shift will result in the contact regions appearing dark.

PAGE 48

48 For lighting hitting the sample when it is separated from the surface, interference fringes occur and the height separation can be found using h 3.1.2 Microtribometer: Motion Control M odule The motion control module (Fig. 3 3) provides the reciprocating motion to the counter sample in the x direction and the loading of the sample in the z direction via piezoelectric stages. The piezoelectric stage in the z direction (normal force) has a travel range of 100m over a total of 10V with a resolution of 0.2 0.4nm and repeatability of 1nm. The x stage piezo travels up to 1500m with a resolution of 2 3nm and repeatability of 14nm. Figure 3 3 The motion control module of the optical micro tribometer Reciprocating motion of the counter sample can be done at speeds of a few micrometers per second up to a few millimeters per second. This counter sample must fit in a 50 mm diameter hold and is p ositioned beneath the sample using micrometers in

PAGE 49

49 the x and y directions. One revolution is equivalent to 0.5 mm of linear motion and can be done in increments of 1 m. The sample can be manually lowered to the counter sample using a micrometer attache d to the z stage assembly to manually lower the sample into or near contact with the counter sample. Figure 3 4. The z stage assembly, or loading module of the optical mirco tribometer 3.1.3 Microtribometer: Loading M odule The z stage assembly, or the l oading module, contains the cantilever assembly shown in Figure 3 4. This assembly contains a flexural cantilever to which the sample is mounted, two capacitance probes in the x and z directions, and an aluminum target box to create a charge for the pro bes to measure. The capacitance probes operate in a voltage range of 10 V with a resolution of 300 V. A summary of instrument sensitivity is shown in Table 3 1. Table 3 1. Microtribometer component sensitivities C omponent L imit R esolution X stage 1500 m 2 3 nm (repeatability of 14nm) Z stage 100 m 0.2 0.4 nm (repeatability of 1nm) M icrometers varies per micrometer 0.5 m C apacitance probes 50 m 15 nm

PAGE 50

50 3.1.4 Measurement of F orces When the z piezo is activated, the cantilever assembly moves the s ample into contact and begins loading, resulting in deflection of the cantilever. Reported forces are determined from the measured cantilever tip displacements; coefficients of friction are determined from the ratio of tip deflections in the lateral and n ormal directions. The target box, mounted to the cantilever tip, provides a measurable voltage change upon deflection. The signal conditioner for the capacitance probes interprets a displacement of 5 m/V; the range of the probes is 50 m with a resolu tion of 15 nm. A variety of cantilevers have been created to address a wide range of test conditions. Low load cantilevers are used for force conditions in micro Newtons and the highest load cantilevers measure forces in Newtons. This range results from the individual stiffness values of each cantilever. These experiments will utilize high load cantilevers operating with applied normal loads in milli Newtons. To eliminate parasitic motion in load/unload tests, two cantilevers can be vertically stacked u sing spacers between cantilevers. The cantilevers used in these experiments are detailed in Table 3 2 (References [92 95] detail force/displacement measurements using capacitance probes and flexural cantilevers.) Table 3 2. Resolution of forces for various cantilevers used in experiments E xperiment C antilever stiffness (mN/m) F orce resolution (mN) Load/unload (all materials) Normal : 7.974 Lateral : 2.304 N ormal: 0.120 Lateral : .035 0 Area/ velocity &friction stu dies : bromobutyl Normal : 0.654 Lateral : 0.929 N ormal: 0.010 Lateral : 0.014 A rea / velocity & friction studies : nitrile Normal : 2.376 Lateral : 5.217 N ormal: 0.036 Lateral : 0.078 High load (100 mN and above) friction studies (all materials) Normal : 82.2 73 Lateral : 117.51 N ormal: 1.234 Lateral : 1.763

PAGE 51

51 The force resolution of the cantilevers can be used to determine the uncertainty in measurements. In most cases, the error bars are not visible in figures as the error is smaller than the symbol used to re present the data point. The law of propagation of uncertainty is used to determine the square of the uncertainty to measured coefficient of friction: (3.1) 3.2 Experimental Procedures 3.2. 1 Load /Unload Testing Contact area and adhesion of the rubber samples against both glass and polymer discs are studied using load/unload tests. The rubber sample (butyl or Buna N) is mounted on a double cantilever and loaded against the optically transparent counter sample (borosilicate glass or cyclic olefin polymer). The sample is loaded to a prescribed normal force and then unloaded until the sample has been completely removed from the counter sample. The loads used ar e 5 mN, 10 mN, 15 mN, and 20 mN and the sample was loaded at a rate of 2.75 m/s. This loading profile allows repeatability to be seen: for example, during the 20 mN test, as the sample passes through the lower loads, the area should match the area seen with the previous 5, 10, and 15 mN tests and all loading scenarios shoul d see a linear increase of area with load. During this process, images are taken at prescribed intervals in order to determine the real contact area for a known normal load. Boxcar averaging using a three point boxcar was used to smooth the images to cre ate a more accurate real area calculation. A background image taken prior to testing was subtracted out of all subsequent images.

PAGE 52

52 Next, three images were averaged together in order to eliminate any higher order fringes that may be present. The final comp osite image was then converted to a black and white image, with white pixels representing contact. The total number of white pixels was summed and converted to micrometers using a measured scale factor of 0.347 micrometers/pixel for this system. Figure 3 5. Behavior of materials with and without adhesion 3.2.2 Contact Area Variation w ith Velocity Imaged sliding was performed to investigate the influence of sliding velocity on contact area. Borosilicate glass was used as the counter sample in these expe riments. Three samples each of nitrile and bromobutyl rubber were loaded to 20 mN and slid over a distance of 800 m. The maximum sliding speed to obtain blur free images is 50 m/s and this was used as the fast velocity in these experiments. The slower sliding velocity used was 10 m/s. Each sample was run three times at the slow velocity, followed by the fast velocity, with images being taken at approximately every 0.15 s. The contact areas obtained for the three cycles were averaged to report the tre nd of

PAGE 53

53 contact area versus velocity for both velocities. (The standard deviation of the three calculated contact areas for each velocity will also be reported.) Comparisons of overall contact area between the slower and faster velocities were observed, as well as any changes in the contact area during sliding. To verify the sample contact area was not changing as a result of running these six cycles, the sample was run again at the slow speed after the testing was complete. When the resulting contact are a was compared to the initial contact area from the first 10 m/s cycle, it was the same or similar, illustrating that the variations in contact area were due only to the change in velocity and not changes in the samples themselves. 3.2.3 Friction T esting The loads used for the load/unload testing were repeated for single cycle friction tests to study trends in the change of the coefficient of friction with increasing load at these contact pressures. A higher load of 100 mN was also used in order to obse rve the coefficient of friction when the load increases by an order of magnitude. The sliding distance for these experiments was 800 m and the reciprocating speed was 100 m/s. Both rubber materials were used and slid against borosilicate glass discs an d cyclic olefin polymer discs. The gliding coefficient of friction and break loose coefficient of friction were observed for three individual tests labeled a, b, and c (The break loose point in sliding occurs where the sample overcomes the lateral forc es and is free to slide ; this point is observed as a peak early on in the coefficient of friction versus sliding distance.) Data was acquired using LabView, with a sampling period of 0.0 05 seconds. Nitrile samples at 100 mN of normal load were slid at 1 00 m/s and 1000 m/s to see if the order of magnitude change in velocity would result in higher coefficients of

PAGE 54

54 friction. (This could not be done for the bromobutyl rubber as the samples did not reach gliding at speeds of 1000 m/s without lubrication pr esent.) 3.2.4 Lubricated Experiments: Stribeck Curves a nd Friction Testing Stribeck curves were created by varying the applied normal load and the sliding velocity. The sample was loaded lightly and slid at a prescribed velocity as the load is increased. This was repeated for multiple sliding velocities in order to create a curve that has boundary through hydrodynamic lubrication. The sliding velocities used were 50 m/s, 100 m/s, 200 m/s, 500 m/s, 1000 m/s and 2000 m/s. The resulting coefficients of friction were plotted with respect to the ratio of the velocity to normal load. The exponents of the velocity and normal load were 0.65 and 0.21, respectively, as previously determined by Rennie et al [96] Upon completion of the Stribeck curves, velocities and loads were chosen to simulate mixed lubrication conditions and hydrodynamic lubrication conditions for lubricated friction tests. The normal load used was 100 mN at velocities of 100 m/s and 1000 m/s. Tests were also done at normal loads of 1800 mN to achieve an apparent contact pr essure of approximately 1 MPa at these same velocities. The samples were studied when slid immediately upon loading, as well as a timed delay from six seconds to 6000 seconds to study the influence of storage time on friction behavior. The oil present al lows the bromobutyl rubber to break loose early in sliding and glide at a low coefficients of friction not attainable in the dry system. (Dry friction tests at these loads and velocities did not yield gliding of the rubber due to break loose not occurring .)

PAGE 55

55 3.3 Materials Nitrile and bromobutyl rubber samples will be used in these studies. The different rubber compositions were chosen to study differences in adhesion and coefficients of friction. Commercial ly available Buna N spheres of 4.8 mm diameter we re used for the nitrile rubber and cut in half using a razor blade. The modulus of the Buna N sphere is approximately 5.5 Mpa. A scanning white light interferometer (Veeco Wyko NT9100) was used to determine the root mean square (RMS) roughness of the sp heres which was found to be 5. 110.15 m for ten samples The bromobutyl samples are manufactured by Datwyler for syringe plungers. Sections of the plunger were cut using a razor blade to produce samples of approximately 1.5 mm in length, width, and he ight. The samples have a radius of curvature in the x direction of 0.5 mm and a radius of curvature of 3 mm in the y direction. The modulus of the bromobutyl rubber is approximately 4 MPa. The bromobutyl samples had and RMS roughness of 5.250.35 m for ten samples A silicone lubricant with a kinematic viscosity of 1000 cSt was used with the bromobutyl rubber for some experiments. Borosilicate float glass optical windows of 3 mm thickness and cyclic olefin polymer windows of 2 mm thickness were used as the counter surfaces. Both window materials were 25 mm in diameter. The modulus of the glass is approximately 64 GPa and the modulus of the polymer is approximately 3 GPa. The RMS roughness for the windows was measured with a stylus profilometer (Veeco Dektak 8). The RMS roughness of the glass windows was 2.06 nm, more than three orders of magnitude smoother than the nitrile rubber. The RMS roughness of the polymer windows was 257 nm An experimental matrix, including materials and parameters, is shown in Table 3 3.

PAGE 56

56 Table 3 3. Matrix of experiments Experiment Sample Counter sample Parameters Load/unload N itrile B romobutyl Borosilicate glass Cyclic olefin polymer F n = 5 mN, 10 mN, 15 mN, 20 mN Area/velocity N itrile B romobutyl Borosilicate glass Fn =20 mN v= 10 m/s, 50 m/s Friction testing N itrile B romobutyl B orosilicate glass C yclic olefin polymer F n = 5 mN, 10 mN, 15 mN, 20 mN, 100 mN v= 100 m/s Stribeck curves B romobutyl Borosilicate glass with oil Cyclic olefin polymer with oil F n =10 220mN v=10 3000um/s Lubricated friction B romobutyl Borosilicate glass with oil Cyclic olefin polymer with oil F n =100 mN, 1800 mN V= 100 m/s, 1000 m/s

PAGE 57

57 CHAPTER 4 NITRILE RESULTS Nitrile is a synthetic rubber and is a copolymer of butadiene and acrylonit rile; butadiene offers elasticity and acrylonitrile provides strength, hardness and resistance to oil and fuel. This combination of properties has led nitrile rubber to be a popular material to use for sealing applications (such as o rings) [97] 4.1 Load /Unload Results For many of its applications it is important that the nitrile component pro vide an impenetrable seal, blocking chemicals from flowing through the rubber part. Imaging at known loads allows insight into the sealing of the material, with a continuous network on contact representing a full seal. Various contact images throughout a loading process are shown for a nitrile sphere (surface e counter surfaces; the apparent contact area is shown via the circle encasing the real contact area. Figure 4 1. Contact area of nitrile rubber against bor osilicate glass (above) and cyclic olefin polymer (below) at various loads.

PAGE 58

58 The force displacement curves for the nitrile rubber reveal little to no adhesion present for the counter faces as seen by the measured pull off forces. In the case of the COC cou nter surface, there is a slight pull off force (Figure 4 2). The force displacement curves for increasing loads overlap almost perfectly with the curves produced at small loads for both counter surfaces. Figure 4 2. Results from the load/unload experi ments with nitrile rubber. A) Force displacement for bromobutyl on borosilicate glass. B) Contact area for bromobutyl against borosilicate glass. C) Force displacement for bromobutyl on cyclic olefin polyer. D) Contact area for bromobutyl on cyclic olefi n polymer.

PAGE 59

59 4.2 Contact Area Variation with Velocity The individual cycle results for the first sample of nitrile for the contact area and velocity experiments are shown in Figure 4 3 In all cases, there is an increase in contact until steady state slidin g occurs. At 10 m/s, the contact area for all three cycles was very similar. During cycle a, the contact area decreased while it increased slightly during cycle c. The contact area of cycle b saw little change during steady state sliding. When run at 5 0 m/s, cycles a and c produced nearly identical results, while cycle b yielded slightly higher contact for much of the sliding duration. The contact area during the sliding of cycles a and c slightly increased; cycle b saw decreased contact area during s liding, indicating a rise in the lateral force (friction force) over sliding. The standard deviation between the areas calculated from cycles a,b, and c was less than 150m 2 for 10 m/s and less than 500 m 2 for 50 m/s for all samples. Figure 4 3. Indi vidual cycle results for a sample of nitrile sliding at two velocities. A) Sliding velocity was 10 m/s. B) Sliding velocity was 50 m/s. The average friction force is presented alongside the contact area for all three samples in Figure 4 4 The contac t area of all samples increased with the higher

PAGE 60

60 velocity. Correspondingly, the average friction force for the higher velocity was less than that of the lower velocity. The higher friction force can account for the smaller contact area seen at 10 m/s: as a stronger lateral force pushes against the material, the rubber will deform more, reducing the contact area. Figure 4 4 The average contact area and friction forces for three samples of nitrile rubber run at 10m/s and 50 m/s. A) Sample 1 contact a rea. B) Sample 1 friction forces. C) Sample 2 contact area. D) Sample 2 friction forces. E) Sample 3 contact area. F) Sample 3 friction forces

PAGE 61

61 The average contact areas for the three samples are shown in Table 4 1, along with the standard deviations of the average contact area during sliding. The increase in contact area from 10 m/s to 50 m/s was approximately 1400 m 2 for the first sample, 2800 m 2 for the second, and 1426 m 2 for the third. Table 4 1 Average and standard deviation of contact area during steady state sliding of the nitrile samples run at 10 m/s and m/s S ample Velocity (m/s) Average contact area (m 2 ) Standard deviation of contact area (m 2 ) 1 10 10727.87 44.83 50 12121.83 154.51 2 10 11093.05 35.48 50 13926.00 393.89 3 10 13632.12 174.98 50 15058.01 410.76 Figure 4 5. Contact images during sliding of Buna N sample one for 10 m/s (top row) and 50 m/s (bottom row). The glass counter surface was sliding from right to left.

PAGE 62

62 The contact area during sliding of the firs t sample is shown in Figure 4 5. Images labeled (a) were taken when the sample loaded, images (b) at the onset of sliding (the glass counter surface traveled from right to left), and images (c) during steady state sliding. For 10 m/s, the contact grew by 785 m 2 from (a) to (c); for 50 m/s, the area increased by 521m 2 during sliding. While the contact area grew during sliding, the general shape of the contact remained similar from loading to sliding. When compared to the images taken at 10 m/s, the i mages from the 50 m/s sliding are visibly darker as a result of the higher pixel count from the increased contact area. 4.3 Friction R esults The results of the friction testing for nitrile rubber on glass at loads of 5, 10, 15, and 20 mN are shown in Figu re 4 6. For all loads, the break loose coefficient of friction occurred around 0.8 and the subsequent gliding coefficient of friction was between 0.6 and 0.8. The curves are generally flat after break loose occurs, with a very steady gliding coefficien t of friction. Each load was run three times and the average of these tests is shown in Figure 4 7. The similar friction coefficient values can be seen in the overlap of the cycles for all loads. The averages of the break loose coefficients of friction an d gliding coefficients of friction for the three tests at each load is shown in Figure 4 7b. As the normal load increased, there was not a significant variation in either the break loose coefficient of friction or gliding coefficient of friction for this system at these loads. In general, there is a slight decrease in the break loose coefficient of friction as the normal load increased for the nitrile polymer system (Figure 4 9). The gliding coefficient of friction leveled off at around 0.8 for most loa ds after increasing slightly from 5 mN to

PAGE 63

63 10 mN. Both break loose and gliding coefficients of friction were higher for the polymer counter surface than seen in the glass counter surface. Figure 4 6. Coefficient of f riction curves for nitrile on borosil icate glass for various loads. A) Fn=5 mN. B) Fn=10 mN. C) Fn=15 mN. D ) Fn= 20 mN Figure 4 7 Averages of coefficients of friction for various loads of nitrile on borosilicate glass. A) A verage coefficient of friction curves for each load B) A verag e break loose and gliding coefficient of friction values for each load

PAGE 64

64 Figure 4 8. Coefficient of f riction curves for nitrile on cyclic olefin polymer for various loads. A) Fn=5 mN. B) Fn=10 mN. C) Fn=15 mN. D ) Fn= 20 mN Figure 4 9. Averages of c oefficients of friction for various loads of nitrile on cyclic olefin polymer. A) A verage coefficient of friction curves for each load B) A verage break loose and gliding coefficient of friction values for each load

PAGE 65

65 The average of six tests of nitrile ru bber at 100 mN sliding at both 100 m/s and 1000 m/s is shown in Figure 4 10. (The error bars represent the standard deviation of these values). While the coefficient of friction for 100 m/s was similar to that of 10 m/s to 100 m/s, there was a notab le increase when the sliding velocity is 1000 m/s. Figure 4 10 Break loose and gliding coefficients of friction for nitrile at normal loads of 100 mN A) Borosilicate glass counter surface. B) Cyclic olefin polymer counter surface.

PAGE 66

66 CHAPTER FIVE BRO MOBUTYL RESULTS Bromobutyl rubber, copolymer of isobromobutylene and isoprene, is popular due to its impermeability to gases. It is used in a variety of applications, from chewing gum to air tight sealing components. The bromobutyl rubber used in these experiments is a bromobutyl rubber produced by mixing with bromine. The rubber is manufactured by Datwyler (formerly Helvoet Pharma) in the form of a plunger to be used in medical syringes. Individual ribs were sectioned off for all testing. These ribs a re critical in syringe performance: they serve to keep medicine from leaking out of the syringe while ensuring foreign contaminants are kept out. Figure 5 1. Bromobut yl sample used for experiments. A section of a rib was used for testing 5.1 Load/Unload R esults The geometry of the bromobutyl rubber (a single rib from a syringe stopper) yields a different contact footprint than the spherical nitrile. In this case, the contact area again is comprised of groups of asperity contacts, but an elliptical appar ent contact area is produced. The medicine within a syringe will not leak out as long as the real contact area reaches percolation (continuous network of in contact regions).

PAGE 67

67 Figure 5 2. Bromobutyl contact area on borosilicate glass (above) and cyclic o lefin polymer (below) for various normal loads. The force displacement curves for the bromobutyl rubber reveal adhesion present for both counter surfaces through the measured pull off forces. In the case of the COC counter surface, the pull off force is s tronger (F=1.06 mN) than the borosilicate glass (F=0.91 mN). The force displacement curves for increasing loads overlap almost perfectly with the curves produced at small loads for both counter surfaces. The curves for glass and COC are nearly identical, except for the pull off forces.

PAGE 68

68 Figure 5 3. Results from the load/unload experiments with bromobutyl rubber. A) Force displacement for bromobutyl on borosilicate glass. B ) Contact area for bromobutyl against borosilicate glass. C) Force displacement f or bromobutyl on cyclic olefin polyer. D) Contact area for bromobutyl on cyclic olefin polymer. Adhesion can also be seen in Figure 5 3b and Figure 5 3d: during unloading, the samples are still in contact after zero load has been reached. The COC counter sample required more force to pull the entire sample from the counter surface than in the glass

PAGE 69

69 case. The contact area increase with load was more linear with the glass counter sample although in both cases, the contact area was repeatable for given load s. 5.2 Contact Area Variation with Velocity Figure 5 4 shows the individual cycle results for the first sample of bromobutyl rubber run at 10 m/s and 50 m/s. After it is loaded, the bromobutyl rubber undergoes a decrease in contact area until the materi al overcomes what is referred to as the break loose force, or the force at which the rubber begins to move freely. In the cases of sliding at 10 m/s, the contact area after this event remained relatively unchanged and steady state sliding was reached aft er the glass disc had slid about 200 m. At 50 value. The friction force (Figure 5 5) for this sample sliding at 50 m/s, was seen to continuously increase until th e final 200 m of sliding, explaining why the contact area decreased during this time. In all cases, the standard deviation between the area calculated for the three cycles was 2000 m 2 or less for 10 m/s and 5000 m 2 or less for 50 m/s. Figure 5 4. In dividual cycle results for a sample of bromobutyl sliding at two velocities. A) Sliding velocity was 10 m/s. B) Sliding velocity was 50 m/s.

PAGE 70

70 The first two bromobutyl samples run at 10 m/s followed the behavior noted above. After decreasing contact a rea prior to breaking free, the average contact area for the three cycles quickly became relatively stable for the duration of sliding. Correspondingly, the average friction force was also relatively constant. In the case of the second sample, a slight d ip in the friction force at 10 m/s is seen; likewise, at the same point in sliding, there is a small increase in the contact area. The first sample yielded similar contact areas at both sliding speeds; for the final quarter of the motion profile, the av erage contact area was the same for the 10 m/s and 50 m/s cycles. The contact area of 50 m/s would likely have been less than 10 m/s had sliding continued, which would be expected as the friction force for the 50 m/s was greater than that of 10 m/s at this point in the sliding. The second and third samples saw the same correlation between friction force and contact area as observed in the nitrile samples: for higher friction forces, there was a smaller contact area. The first and third samples had a higher friction forces at the higher sliding speed whereas the second sample saw nearly the same friction forces towards the end of sliding at both speeds. Table 5 1. Average contact area and standard deviation during steady state sliding. S ample Veloci ty (m/s) Average contact area (m 2 ) Standard deviation of contact area (m 2 ) 1 10 11965.06 1708.286 50 14800.72 2003.054 2 10 29241.69 1029.394 50 34273.41 3712.68 3 10 34347.73 6184.55 50 16354.00 4317.52

PAGE 71

71 Figure 5 5. The average contact are a and friction forces for three samples of bromobutyl rubber run at 10m/s and 50 m/s A) Sample 1 contact area. B) Sample 1 friction forces. C) Sample 2 contact area. D) Sample 2 friction forces. E) Sample 3 contact area. F) Sample 3 friction forc es.

PAGE 72

72 The contact area during sliding of the third bromobutyl sample is shown in Figure 5 6. The sample was loaded in image (a), the onset of sliding began at image (b) and image (c) was taken toward the end of the 800 m for both velocities. For 10 m/s, t he contact decreased by 31156 m 2 from (a) to (c); for 50 m/s, the area decreased 40053 m 2 during sliding. The shape of the contact visibly changes due to deformation from the friction force, which is higher than that of the nitrile rubber which maintai ned the same general contact shape during sliding. Figure 5 6. Contact images during sliding of the third bromobutyl sample for 10 m/s (above) and 50 m/s (below). The glass counter surface was sliding from right to left.

PAGE 73

73 5.3 Friction R esults Figur e 5 7 shows the results for bromobutyl rubber sliding against borosilicate glass windows. Unlike the nitrile rubber, this system showed a decrease in both break loose coefficient of friction and gliding coefficient of friction as the normal load increased In general, break loose again occurred early on in sliding; the exception is at a normal load of 20 mN, where a double break loose event appears. This event was found to be repeatable showing that as the normal load increases, this event becomes more significant. Figure 5 7. Coefficient of f riction curves for bromobutyl on borosilicate glass for various loads. A) Fn=5 mN. B) Fn=10 mN. C) Fn=15 mN. D ) Fn= 20 mN

PAGE 74

74 The average break loose coefficient of friction decreased from approximately 1.75 at a normal load of 5 mN to below 1.5 at 20 mN. The average gliding coefficient of friction decreased from 1.5 to below 1.25 (Figure 5 8 ). Figure 5 8. Averages of coefficients of friction for various loads of bromobutyl on borosilicate glass. A) A verage c oefficient of friction curves for each load B) A verage break loose and gliding coefficient of friction values for each load As with the nitrile rubber, the bromobutyl rubber experienced an increase in both the break loose and gliding coefficients of fric tion when slid against the polymer counter surface (Figure 5 9) The coefficients of friction decreased with increased normal load, as observed with the bromobutyl on glass system. Against both counter surfaces, the bromobutyl samples reached break loose later in the sliding profile than the nitrile rubber, which can be attributed to the stronger adhesive forces present in the bromobutyl systems. When the friction curves were averaged together, the occurrence of where break loose occurs can be compared f or all normal loads. While the break loose coefficient of friction was higher at the lower loads (5 and 10 mN), it occurred sooner in sliding than at subsequent higher normal loads (15 and 20 mN). This was also seen in the bromobutyl/glass system, althou gh not as pronounced as the bromobutyl/coc polymer

PAGE 75

75 system. While adhesion appears to have a stronger influence on the coefficient of friction at lower loads, when more of the sample is in contact (at higher loads), it takes longer for the sample to break loose and glide freely. Figure 5 9. Coefficient of f riction curves for bromobutyl on cyclic olefin polymer for various loads. A) Fn=5 mN. B) Fn=10 mN. C) Fn=15 mN. D ) Fn= 20 mN Figure 5 10. Averages of coefficients of friction for various loads of bromobutyl on cyclic olefin polymer. A) A verage coefficient of friction curves for each load B) A verage break loose and gliding coefficient of friction values for each load

PAGE 76

76 5.4 Lubricated R esults The bromobutyl samples are typically used in syringes wit h a form of lubrication so extra experiments were done with this material to garner further insight into syringe tribology. The lubrication used in these experiments was silicone oil with a viscosity of 1000 cSt. With the introduction of lubrication, Str ibeck curves were created to determine lubrication regimes and subsequently, loads and sliding velocities to use to operate is specific lubrication regimes. These curves were created for silicone oil lubricated borosilicate glass and cyclic olefin polymer counter sample s. 5.4.1 Stribeck C urves The Stribeck curve for bromobutyl rubber on borosilicate glass in 1000 cSt silicone oil is shown in Figure 5 11 At some sliding velocities (such as 200 m/s) the sample can undergo multiple lubrication regimes dep ending on the normal load. At other sliding velocities (such as 2000 m/s), the sample was in hydrodynamic lubrication for all normal loads. At 50 m/s and 100 m/s, the sample never reached hydrodynamic lubrication. Figure 5 11. Stribeck curve for bro mobutyl on borosilicate glass with 1000 cSt silicone oil lubrication

PAGE 77

77 The Stribeck curve for bromobutyl rubber on cyclic olefin polymers in 1000 cSt silicone oil is shown in Figure 5 12 For this counter surface, velocities of 10 m/s and 20 m/s were use d in addition to the other prescribed velocities to complete this curve. Figure 5 11. Stribeck curve for bromobutyl on cyclic olefin polymer with 1000 cSt silicone oil lubrication 5.4.2 Lubricated Friction Results The friction test results are shown in Figure 5 12 for bromobutyl on lubricated glass In practice, a syringe stopper does not reciprocate so only the first 800 m of sliding were studied. For all tests, the normal load of 1800 mN yielded lower coefficients of friction at break loose and s moother gliding as compared to the tests with a normal load of 100 mN. Both the break loose coefficients of friction and gliding coefficients of friction increased with hold time, indicating that the longer the stopper is inside a syringe, the higher the friction forces will be upon motion.

PAGE 78

78 Figure 5 12. Friction results for bromobutyl on lubricated gla s s at various normal loads and sliding velocities. A) F n =100 mN, v=100 m/s. B) F n =1800 mN, v=100 m/s. C) F n =100 mN, v=1000 m/s. D) F n = 1800 mN, v=1000 m/s The friction test results for the COC counter surface are shown in Figure 5 13 As was typical with the cyclic olefin polymer counter surface for all dry testing, the coefficients of friction of the lubricated polymer system are higher than those of the glass system. In the cases of longer hold times, the break loose coefficient of friction neared tha t of the dry break loose coefficients of friction seen previously. Also notable

PAGE 79

79 of the lubricated polymer tests are the sticking and slipping seen at a load of 1800 mN sliding at 100 m/s. This stick slip was not seen at the higher sliding speed or when the bromobutyl was run against borosilicate glass under the same conditions. Figure 5 13. Friction results for bromobutyl on lubricated cyclic olefin polymer at various loads and sliding velocities. A) Fn=100 mN, v=100 m /s. B) Fn=1800 mN, v=100 m/s C) Fn =100 mN, v=1000 m/s D) Fn=1800 mN, v=1000 m/s

PAGE 80

80 CHAPTER SIX DISCUSSION The optical in situ tribometer provided insight into the behavior of the rubbers in contact with the counter surfaces. While limited by the optical resolution, real contact areas could be calculated for experiments involving loading and sliding. Trends in contact area and shape could be observed to garner a further under standing of elastomeric tribology. 6.1 Load/U nload The bromobutyl rubber demonstrated stro nger adhesion via measured pull off force (Table 6 1 ) for both counter surface materials than the nitrile rubber, which may be due in part to the rougher surface of t he nitrile spheres. Adhesion due to van der Waals forces is usually strong but difficult to observe on rough surface due to the small range of action. If the roughness of a material is on a longer scale than the van der Waals range, these interactions ar e diminished between the materials, lowering the measured adhesion [98, 99] Table 6 1. Measured pull off forces for experimental systems S ystem Pull off force (mN) N itrile/borosilicate glass 0 N itrile/COC poly mer 0.54 B romobutyl/borosilicate glass 0.91 B romobutyl/COC polymer 1.06 For all cases, the unloading of samples showed a strong hysteresis. This trend has been explained by likening the loading to closing a crack, while unloading is akin to opening a crack [100, 101] More energy is required to open the crack (unload the rubber) than is required to close the crack, giving rise to the hysteresis effect seen. Rubber exp eriences a time dependent deformation that is governed by the crack tip

PAGE 81

81 velocity; for a crack opening quickly, an effective interfacial energy is produced and enhanced whereas a closing crack reduces this energy. There are two classic contact mechanic theo ries that apply to a soft material (like rubber) in contact with a hard material (such as the glass or polymer): JKR and Hertz. These models can be used to predict the apparent contact area of the nitrile sphere but will overestimate the real contact area (The apparent contact area is found by fitting a circle to the real contact area, as shown previously in Figure s 4 1 and 5 2 and the real contact area is much smaller than an area that would fill the entire circle.) Using Equation 2 16 with the measur ed pull off force, the work of adhesion for the apparent contact area can be found, and the predicted JKR area is shown in Figure 6 1 for nitrile rubber along with the area predicted by Hertz and the measured real and apparent contact areas. With the abse nce of measured adhesion in the nitrile/glass case (Fig ure 6 1 a), the JKR theory reduces to the Hertz contact theory. The two theories are slightly different due to the pull off force present for the COC polymer case (Fig ure 6 1 b). The apparent contact are a is almost exactly modeled by the JKR theory for the COC case but is slightly off in the case of the glass counter sample. When a small pull off force is used in the JKR model, the measured apparent contact area is very similar to the predicted JKR area, which indicates that some adhesion is present in the system. It is likely the pull off force is too small to be measured with the cantilever used in these experiments; were there no adhesion present, the apparent contact area would have been closely mode led by the Hertz theory, which is not the case.

PAGE 82

82 Figure 6 1. Experimental contact area data for nitrile as compared to apparent contact areas predicted by the JKR and Hertz models A) B orosilicate glass counter surface. B) C yclic olefin polymer counter surface. While these classic contact models can be used with the nitrile sphere, they are less useful for the bromobutyl samples tested here. The bromobutyl samples have two radii of curvature (0.5mm in the x direction, 3mm in the y direction). Classi c contact models only take into account one radius of curvature per sample and assume a circular contact region, whereas the bromobutyl sample produces an elliptical contact. The mismatch between the classic JKR and Hertz models with the experimental data is shown in Figure 6 2 Classical contact mechanic models can be used with these materials to garner an understanding of the apparent area and load relationship, which can be useful in predicting apparent contact pressures. Models to represent the contact behavior of rubbers are still being developed. While Hertz and JKR are intended for use with an elastic material, modeling the contact region as simple circular areas does not take into account the clusters of asperities in and out of contact. The geome try, surface

PAGE 83

83 roughness of the samples, and viscoelastic behavior of the material impact the resulting real contact area. Recent work by Persson has attempted to tackle the modeling of complex rubber systems [85, 102 104] and further progress in this area may allow contact area modeling of complex geometries such as the bromobutyl samples used here. Meanwhile, the optical microtribometer allows the study the real contact area produced by samples of various geometrie s. Figure 6 2. Experimental contact area data for bromobutyl samples as compared to apparent contact areas predicted by the JKR and Hertz models. A) B orosilicate glass counter surface. B) C yclic olefin polymer counter surface. 6.2 Contact Area Variatio n with Velocity The contact area results of both materials show a decrease in contact area with an increase in friction force. Previous studies relate the change in contact area to changes in sliding speeds and suggest that real contact area decreases wit h increasing sliding velocity [88, 99, 101, 105 108] This has been attributed to stiffening of the rubber as sliding speed increases but can be explained by energy storage at the interface. Frictional shear stress will build up elastic deformation energy which is stored at the interface. At high enough velocities, this energy breaks adhesive bonds and non

PAGE 84

84 adhesive sliding results [106, 107] However, this assumes that th e friction force increases as sliding velocity increases. While this trend was seen in the bromobutyl sample, the nitrile rubber experienced the opposite: friction force decreased with increasing sliding velocity for the speeds used here, resulting in cont act area decreasing with increasing sliding speed. A second mechanism can be used to explain the increase in contact area with sliding speed seen by the nitrile rubber. Returning again to likening adhesive bonding to the theory of interfacial cracks, th e boundary between contact and non contact during sliding will be closing cracks on the front side while opening cracks on the exit side. As noted previously, viscoelastic materials may see an enhancement in the effective interfacial energy at crack openi ngs [101] This in turn may increase the adhesive interaction and ultimately the contact area. For materials with a hi gh glass transition temperature, this effect has been found to be important as these materials are highly dissipative for low sliding velocities. The energy of the nitrile rubber (Tg= 26 C) will be higher than the bromobutyl rubber (Tg= 73 C) due to it s higher glass transition temperature. Another likely contribution to this mechanism in the nitrile rubber is the roughness of the material. An increase of roughness will lead to a large number of contact spots and provide more regions between contact and non contact, enhancing the crack opening mechanism. As this mechanism becomes more pronounced, an increase in contact area will result with changing velocity. The nitrile rubber spheres used are rougher than the bromobutyl rubber used here and the mater ials used in previous studies which found a decrease in contact area with an increase in sliding velocity.

PAGE 85

85 6.3 Friction E xperiments The influence of adhesion on the coefficient of friction of the materials can be seen by plotting the relation between the f riction and normal forces for the various loading scenarios. Multiple points of the sliding curves were taken at the break loose event and during gliding for each normal load. The results are shown in Figure 6 3 with a linear fit through the data: when adhesion is not present, this line will intercept at zero. Figure 6 3. Relationship between friction force and normal force. A) B r omobutyl on borosilicate glass. B) B rom obutyl on cyclic olefin polymer. C) Nitrile on borosilicate glass. D) N itrile on cyclic olefin polymer In the cases where the coefficient of friction decreased with normal load, the y intercept does not occur at zero, but instead inte rcepts at the value of the pull off force.

PAGE 86

86 At higher loads, the friction force is significant enough to overcome this adhesive force; at lower loads, the adhesion is more pronounced. This was further studied by running samples at normal loads of 100 mN for both rubber materials and counter surfaces. The results are shown in Table s 6 2 and 6 3 along wi th the real and apparent contact pressures for these systems. Table 6 2. Nitrile rubber friction results for all loads and counter surfaces C ounter surface N ormal load (mN) G liding coefficient of friction B reak loose coefficient of friction R eal contact pr essure (MPa) A pparent contact pressure (MPa) Glass 5 0.600 0.004 0.7140.043 0.963 0.0867 10 0.691.012 0.8870.990 1.171 0.095 15 0.6080.032 0.8170.135 1.112 0.142 20 0.5450.052 0.6840.169 1.162 0.166 100 0.6300.018 0.5950.046 1.330 0.135 COC 5 1.0680.134 0.6610.069 1.620 0.069 10 1.0110.178 0.8150.053 1.671 0.120 15 0.9010.051 0.8180.047 1.810 0.152 20 0.9790.082 0.7800.039 2.094 0.176 100 0.8250.059 0.9710.164 2.121 0.230 For the case of the nitrile rubber sliding aga inst borosilicate glass, little adhesion has been seen, and as a result, gliding coefficients of friction are not significantly different for load of 5 mN to 100 mN. The cyclic olefin polymer system shows more adhesion, and the gliding coefficient of fric tion for 100 mN was decreased when compared to the lower loads. In all cases, the apparent contact pressure is significantly smaller than the real contact pressure as a result of the apparent contact area being much larger than the real contact area. The decrease in gliding coefficient of friction with increasing load was more pronounced for the bromobutyl systems, which was expected given the stronger adhesion in these systems when compared to the nitrile systems.

PAGE 87

87 Table 6 3. Bromobutyl results for all loa ds and counter surfaces C ounter surface N ormal load (mN) G liding coefficient of friction B reak loose coefficient of friction R eal contact pressure (Pa) A pparent contact pressure (Pa) Glass 5 1.5060.01229 1.7910.104 0.433 0.123 10 1.3030.0598 1.6680. 138 0.530 0.175 15 1.1920.0421 1.4080.116 0.591 0.181 20 1.1270.078 1.4300.266 0.627 0.217 100 0.9750.072 1.2820.017 1.672 0.254 COC 5 2.2510.036 2.7890.044 0.675 0.133 10 1.8700.053 2.4230.149 0.727 0.170 15 1.7050.065 2.2400.290 0. 850 0.215 20 1.7900.192 2.2530.341 0.945 0.286 100 1.4890.074 1.4490.141 1.981 0.376 As mentioned previously, past studies involving rubber systems have found an increase in coefficient of friction with an increase of velocity that was not observ ed with the nitrile system at the sliding velocities used to study the contact area. However, when this velocity was increased by over an order of magnitude, the coefficient of friction did increase. At slower sliding velocities (less than 1 mm/s), the c oefficient of friction of nitrile rubber likely depends more on adhesion than at faster sliding velocities (greater than 1 mm/s), where the adhesive bonds are broken, giving way to the more classic friction behavior observed in previous studies. 6.4 Lubric ated T esting From the Stribeck curve, a relationship between friction force, normal force, and sliding velocity can be obtained. If the fluid shearing is considered a negligible contribution to friction, the height of the sample in lubricated contact is t raditionally related to velocity and normal load by Equation 6.1 Utilizing bearing area curves, the fraction of the sample in contact ( ) is inversely proportional to the film thickness

PAGE 88

88 Figure 6 4. Schematic of material in flu id of thickness h and sample bearing area curve to relate percent of material in contact to film thickness (6.1) The real contact area is the fraction in contact of the nominal contact area; the nominal contact area can be fou nd from the contact pressure (B) and normal load. (6.2) (6.3) Substituting the relationship between and h into Equation 6.1 then plugging into Equation 6.3 the followin g relationship between real contact area, normal load, and velocity is found: (6.4) Likewise, the real contact area can be determined from the shear stress: (6.5) By subtituting Equation 6.5 into Eq uation 6.4 the relationship between friction force, normal load, and velocity is determined. (6.6) Figures 6 5 and 6 6 show the experimental data of the bromobutyl rubber lubricated with silicone oil on glass and polymer count er surfaces.

PAGE 89

89 Figure 6 5. Relationship between the friction force, normal force, and velocity for bromobutyl on borosilicate glass with 1000 cSt silicone oil lubrication In the case of the polymer counter surface, there is almost a complete overlap in th e data for all sliding demonstrating the validity of the derived relationship between these three parameters Figure 6 5. Relationship between the friction force, normal force, and velocity for bromobutyl on borosilicate glass with 1000 cSt silicone oil lubrication

PAGE 90

90 CHAPTER 7 CONCLUSIONS This work utilized an in situ microtribometer to examine rubber friction. This tribometer made it possible to obtain real contact area measurements during loading, unloading, and sliding. The rubbers studied were brom obutyl and nitrile; the counter surfaces were borosilicate glass and cyclic olefin polymer. The resulting contact area data offered insight into the adhesion present in the rubber systems, as well as mechanisms influencing contact area during sliding. Th e friction data acquired by the microtribometer allowed a study in trends in friction behavior with various loads and velocities. The following summarizes the critical findings of this study: Rubber materials in contact with another surface can be though t of as cracks propagating: more energy is required to remove a rubber sample from a counter surface than is required to put the rubber into contact. Contact area decreases with increasing friction force. Two mechanisms control the contact area of rubber materials; one mechanism explains the decrease in contact area at faster sliding speeds and the other explains why the area decreases at slower sliding speeds. At faster speeds, energy storage at the interface can break adhesive bonds and the friction forc e will become velocity dependent instead of adhesion dependent. This mechanism was found to take effect at different velocities for different materials; in the case of the nitrile rubber, the sliding velocity had to increase by two orders of magnitude in o rder for the coefficient of friction to increase. The second mechanism, modeling the rubber surface as a crack opening and closing during sliding, describes the behavior of the nitrile rubber at velocities less than 1000 m/s. The stronger the adhesio n is in a system the more influence it has on the coefficient of friction at low loads. Loading needs to be significant enough such that the friction force can overcome adhesion rather than coupling into it. At certain loads and speeds, some rubber syst ems require lubrication to achieve gliding. In this case, Stribeck curves are useful tools to determine the proper test

PAGE 91

91 conditions to achieve specific lubrication regimes. For silicon e oil with kinematic viscosity of 1000 cSt, the friction force was foun d to be proportional to Fn 1.5 *V 0.5

PAGE 92

92 APPENDIX A SURFACE ENERGIES The surface energies of the discs (borosilicate glass and cyclic olefin polymer) and rubbers ( nitrile and bromobutyl ) were determined using contact angles to create Zisman plots. Zisman plo ts are created using liquids of known surface tensions plotted against the cosine of the contact angle; Zisman found a rectilinear relationship between the surface energy (or tension in liquids) and the cosine of the contact angle. A curve fit will extrap olate to cos( )= 1 at the critical surface energy of the solid material. While the critical surface energy of solids is well defined using a homologous series of liquids, all liquids generally follow a linear trend. When a variet y of chemical classes are used to determine the surface energy, only the solid in question is characterized, making this extrapolated intercept valuable [109, 110] For these measurements hexane, hexadecane, heptan e, ethylene glycol, glycerol, and water were selected from a list of commonly used liquids for Zisman plots [109] The surface energies of these liquids are shown in Table A 1 Table A 1. Surface tensions of liquids used to create Zisman plots (from www. surface tension.de) L iquid S urface tension (mN/m) n hexane 18.43 n heptane 20.14 n hexadecane 27.47 ethylene glycol 47.70 glycerol 64.00 water 72.80 Borosilicate G lass Bo 283 mN/m [111] The glass was plasma cleaned prior to experimental measurements. Due to reflectivity of the surface, subsequent image processing to determine the contact angle was difficult

PAGE 93

93 for many of the liq uids. The plot below ( Figure A 1 ) shows data obtained from the measurable images of 3 6 droplets per liquid using hexadecane, ethylene glycol, glycerol, and water; the standard deviations of these measurements are shown in Table A 2 Figure A 1. Zisman plot for borosilicate glass. Extrapolating a linear fit through the data shown, the corresponding value for the surface tension when the cos( )=1 is around 276 mN/m. As a check, this value falls within the published range Table A 2. S tandard deviation of measured contact angles for liquids on borosilicate glass L iquid S tandard deviation (angles) S n hexadecane 2.0412 0.0124 Ethylene glycol 2.7749 0.0136 glycerol 4.9322 0.0298 water 3.0550 0.0207

PAGE 94

94 Cyclic Olefin Polymer The cyclic olefin polymer surface energy was determined by using n hexane, n heptane, n hexadecane, et hylene glycol and glycerol. The resulting Zisman plot is shown in Figure A 2 and the standard deviations of these measurements are shown in Table A 3 For each liquid, 3 6 droplets were measured. (The surface was wetted by the n hexane and n heptane, re sulting in zero contact angle and are not included in Table A 3 .) Figure A 2. Zisman plot for cyclic olefin polymer The experimental critical surface energy for the cyclic olefin polymer used in these ( polytetrafluoroethylene has a surface energy of around 19 mN/m [109, 112] and polymers fall between 10 30 mN/m. Table A 3 S tandard deviation of measured contact a ngles for liquids on borosilicate glass L iquid S tandard deviation (angles) S N hexadecane 0.5773 0.0014 Ethylene glycol 3.7237 0.0627 G lycerol 2.8867 0.0491

PAGE 95

95 Bromobutyl R ubber Three droplets of n hexane, n hexadecane, ethylene g lycol, glycerol, and water were used to measure the critical surface energy of the butyl rubber used in these experiments. The Zisman plot is shown in Figure A 3 and the standard dev iations of the measurements in T able A 4 Figure A 2. Zisman plot for cyclic olefin polymer The resulting critical surface energy for this butyl rubber is 18.9 mN/m. Published values of butyl rubbers (CAS # 9003 27 4 from Accu Dyne Test) range from 19 mN/m 35 mN/m; as rubber recipes vary from manufacturer to manufacturer this range is expected. Table A 4 S tandard deviation of measured contact angles for liquids on bromobutyl rubber L iquid S tandard deviation (angles) S N hexadecane 3.2145 0.0543 Ethylene glycol 3.4640 0.0583 G lycerol 0.0092 1 .0000 W ater 0.0368 2.121 0

PAGE 96

96 Nitrile Rubber A flat piece of the nitrile rubber used in the experiments was ordered to make the surface energy measurements. A minimum of three droplets of n hexane, n heptane, n hexadecane, ethylene glycol, glycerol, and wa ter were used. The critical surface ene 34 mN/m [113] ). Figure A 4. Zisman plot for nitrile rubber The standard deviations for the measured angles are shown in Table A 5 As previously noted, n hexane and n heptane are not included as the contact angles were zero for all measurements. Table A 5 S tandard deviation of measured contact angles for liquids on nitrile rubber L iquid S tandard deviation (angles) S N hexadecane 5.354 0 0.0347 Ethylene glycol 0 .0000 0 .0000 G lycerol 4.9328 0.0860 W ater 12.2474 0.1986

PAGE 97

97 APPENDIX B UNCERTAINTY IN CONTACT AREA MEASUREMENTS Fregly and Sawyer developed a non dimensional area variable to predict w orst case discretization errors in contact area calculations for different aspect ratios of elliptical contacts [114] The parameter, is the ratio of the number of perimeter elements to the total number of elements in contact, or phrased in another way: the ratio of edge contact pixels to total contact pixels. The experimental error for various aspect ratios collapsed onto a single curve and they found that w orst case discretization err ors followed a power law of (The aspect ratios used were 1, 2, 5, 10, and 20.) Figure B 1. Simulated discretization errors for contact area as a function of [Adapted with p ermission from B.J. Fregly and W.G. Sawyer, Estimation of discretization errors in contact pressure measurements, J. Biomech. 36 (2003) 311, Figure 2.] The contact areas in these experiments were such that perimeter pixels were a small fraction of the tot al contact, resulting in a small error. For example, a contact area

PAGE 98

98 patch which contained 4050 square pixels contained only 289 edge pixels, yielding a of 0.07. Plugging this into the power law, an error of less than 1% is pr oduced. For all ratios of perimeter pixels to total pixels in contact, the percent error in the contact area measurements fall in the lower left of Figure B 1.

PAGE 99

99 LIST OF REFERENCES [1] G. Gompper and M. Schick, An Introduction to Sof t Matter, Volume 1: Polymer Melts and Mixtures, first ed., Wiley VCH Verlag GmbH & Co., Weinheim, Germany, 2006 [2] D. Tabor, Bulk modulus of rubber, Polymer 35 (1994) 2759 2763. [3] S.L. Rosen, Fundamental Principles of Polymeric Materials, second ed ., John Wiley & Sons Incorporated, New York, 1993 [4] W. Callister, Materials Science and Engineering: An Introduction, sixth ed., John Wiley, New York, 2001 [5] K.J. Wahl and W.G. Sawyer, Observing interfacial sliding processes in solid solid contacts MRS Bull. 33 (2008) 1159 1167. [6] F.E. Kennedy, Thermal and thermomechanical effects in dry sliding, Wear 100 (1984) 453 476. [7] F.P. Bowden and D. Tabor, The Friction and Lubrication of Solids, Clarendon Press, Oxford, 1964 [8] M. Belin, J. Lo pez and J.M. Martin, Triboscopy, a quantitative tool for the study of the wear of a coated material, Surface and Coatings Technology 70 (1994) 27 31. [9] M. Belin and J.M. Martin, Triboscopy, a new approach to surface degradations of thin films, Wear 1 56 (1992) 151 160. [10] M. Belin and J.M. Martin, Triboscopy, a new quantitative tool for microbiology, Wear 168 (1993) 7 12. [11] K.J. Wahl, M. Belin and I.L. Singer, A triboscopic investigation of the wear and friction of MoS2 in a reciprocating sli ding contact, Wear 214 (1998) 212 220. [12] L.J. Bredell, L.B. Johnson and D. Kuhlmannwilsdorf, Teaming measurements of the coefficient of friction and of contact resistance as a tool for the investigation of sliding interfaces, Wear 120 (1987) 161 173 [13] M. Heuberger, J. Vanicek and M. Zach, The extended surface forces apparatus. II. Precision temperature control, Rev. Sci. Instrum. 72 (2001) 3556 3560. [14] M. Heuberger, M. Zach and N.D. Spencer, Sources and control of instrumental drift in th e surface forces apparatus, Rev. Sci. Instrum. 71 (2000) 4502 4508.

PAGE 100

100 [15] M. Zach and M. Heuberger, Statistical and systematic errors of the surface forces apparatus, Langmuir 16 (2000) 7309 7314. [16] M. Zach, J. Vanicek and M. Heuberger, The extende d surface forces apparatus. Part III. High speed interferometric distance measurement, Rev. Sci. Instrum. 74 (2003) 260 266. [17] D. Gourdon, M. Yasa, A.R.G. Alig, Y.L. Li, C.R. Safinya and J.N. Israelachvili, Mechanical and structural properties of BaC rO4 nanorod films under confinement and shear, Advanced Functional Materials 14 (2004) 238 242. [18] Y. Golan, M. Seitz, C. Luo, A. Martin Herranz, M. Yasa, Y.L. Li, C.R. Safinya and J. Israelachvili, The x ray surface forces apparatus for simultaneous x ray diffraction and direct normal and lateral force measurements, Rev. Sci. Instrum. 73 (2002) 2486 2488. [19] S.C. Bae, H. Lee, Z.Q. Lin and S. Granick, Chemical imaging in a surface forces apparatus: Confocal Raman spectroscopy of confined poly(dime thylsiloxane), Langmuir 21 (2005) 5685 5688. [20] S.C. Bae, J.S. Wong, M. Kim, S. Jiang, L. Hong and S. Granick, Using light to study boundary lubrication: Spectroscopic study of confined fluids, Phil. Trans. R. Soc. A 366 (2008) 1443 1454. [21] P.M. McGuiggan, M.L. Gee, H. Yoshizawa, S.J. Hirz and J.N. Israelachvili, Friction studies of polymer lubricated surfaces, Macromolecules 40 (2007) 2126 2133. [22] A. Mukhodpadhyay, J. Zhao, S.C. Bae and S. Granick, An integrated platform for surface forces measurements and fluorescence correlation spectroscopy, Rev. Sci. Instrum. 74 (2003) 3067 3072. [23] F. Mugele, T. Becker, A. Klinger and M. Salmeron, Two dimensional observation of drainings and layering transitions in confined liquids, Colloids and S urfaces A 206 (2001) 105 113. [24] D. Tabor, F.R.S. Winterton and R.H.S. Winterton, The direct measurement of normal and retarded van der Waals forces, Proc. Roy. Soc. A 312 (1969) 435 450. [25] J.N. Israelachvili and D. Tabor, Measurement of van der Waals diserpsion forces in the range 1.4 to 130nm, Proc. Roy. Soc. Lond. Ser. A 331 (1972) 19 28. [26] J.N. Israelachvili and D. Tabor, The shear properties of molecular films, Wear 24 (1973) 386 390.

PAGE 101

101 [27] A.M. Homola, J.N. Israelachvili, P.M. McGui ggan and L. Gee, Fundamental (1990) 65 83. [28] C.U.A. Cheong and P.C. Stair, In situ studies of the lubricant chemistry and frictional properties of perfluoropolyalkyl ethers at a sliding contact, Tribol. Lett. 10 (2001) 117 126. [29] S.D. Devorak, K.J. Wahl and I.L. Singer, Friction behavior of boric acid and annealed boron carbide coatings st udied by in situ Raman tribometry, Tribol. Trans. 45 (2002) 354 362. [30] I.L. Singer, S.D. Dvorak, K.J. Wahl and T.W. Scharf, Role of third bodies in friction and wear of protective coatings, J. Vac. Sci. Technol. A 232 (2003) 263 272. [31] S.D. Dvo rak, K.J. Wahl and I.L. Singer, In situ analysis of third body contributions to sliding friction of a Pb Mo S coating in dry and humid air, Tribol. Lett. 28 (2007) 263 2. [32] T.W. Scharf and I.L. Singer, Monitoring transfer films and friction instabili ties with in situ Raman tribometry, Tribol. Lett. 14 (2002) 3 8. [33] T.W. Scharf and I.L. Singer, Thickness of diamond like carbon coatings quantified with Raman spectroscopy, Thin Solid Films 440 (2003) 138 144. [34] R.R. Chromik, C.C. Baker, A.A. Voevodin and K.J. Wahl, In situ tribometry of solid lubricant nanocomposite coatings, Wear 262 (2007) 1239 1252. [35] R.R. Chromik, A.L. Winfrey, J. Luning, R.J. Nemanich and K.J. Wahl, Run in behavior of nanocrystalline diamond coatings studied by in s itu tribometry, Wear 265 (2008) 477 489. [36] T.W. Scharf and I.L. Singer, Role of third bodies in friction behavior of diamond like nanocomposite coatings studied by in situ tribometry, Tribol. Trans. 45 (2002) 363 371. [37] T.W. Scharf and I.L. Sin ger, Quantification of the thickness of carbon transfer films using Raman tribometry, Tribol. Lett. 17 (2003) 3 8. [38] J.J. Hu, R. Wheeler, J.S. Zabinski, P.A. Shade, A. Shiveley and A.A. Voevodin, Transmission electron microscopy analysis of Mo W S Se film sliding contact obtained by using focused ion beam microscope and in situ microtribometer, Tribol. Lett. 32 (2008) 49 57.

PAGE 102

102 [39] S.D. Dvorak, K.J. Wahl and I.L. Singer, Friction behavior of boric acid and annealed boron carbide coatings studied by i n situ Raman tribometry, Tribol. Trans. 45 (2002) 354 362. [40] A. Ovcharenko, G. Halperin and I. Etsion, In situ and real time optical investigation of junction growth in spherical elastic plastic contact, Wear 264 (2008) 1043 1050. [41] A. Ovcharen ko, G. Halperin, I. Etsion and M. Varenburg, A novel test rig for in situ and real time optical measurement of the contact area evolution during pre sliding of a spherical contact, Tribol. Lett. 23 (2006) 55 63. [42] M.L. Gee, P.M. McGuiggan, J.N. Israe lachvili and A.M. Homola, Liquid to solid like transitions of molecularly thin films under shear, J. Chem. Phys. 93 (1990) 1895 1906. [43] R. Tadmor, N.H. Chen and J.N. Israelachvili, Thickness and refractive index measurements using multiple beam inter ference fringes (FECO), J. Colloid Interface Sci. 264 (2003) 548 553. [44] H.E. Sliney, Dynamics of solid lubricration as observed by optical microscopy, ASLE Trans. 21 (1977) 109 117. [45] P.M. Cann, H.A. Spikes and J. Hutchinson, The development of a spacer layer imaging method (SLIM) for mapping elastohydrodynamic contacts, Tribol. Trans. 39 (1996) 915 921. [46] R.P. Glovnea, A.K. Forrest, A.V. Olver and H.A. Spikes, Measurement of sub nanometer lubricant films using ultra thin film interferomet ry, Tribol. Lett. 15 (2003) 217 230. [47] P.M. Cann and H.A. Spikes, In contact IR spectroscopy of hydrocarbon lubricants, Tribol. Lett. 19 (2005) 289 297. [48] K.J. Wahl, R.R. Chromik and G.Y. Lee, Quantitative in situ measurement of 736. [49] N. Mauntler, A. Rennie, T. Schmitz and W.G. Sawyer, Design and testing of a tribometer for in situ microscopy, Proc. ASPE Annual Meeting (2006) [50] D.P. Woodruff, The Solid Liquid Interface, Cambridge University Press, London, 1973 [51] E.A. Gulbransen, The role of minor elements in the oxidation of metals, Corrosion 12 (1956) 61 67.

PAGE 103

103 [5 2] I.M. Hutchins, Tribology: Friction and Wear of Engineering Materials, Arnold, London, 1992 [53] G. Stachowiak and A.W. Batchelow, Engineering Tribology, third ed, Elsevier, Oxford, 2005 [54] W.P. Dong, P.J. Sullivan and K.J. Stout, Comprehensive st udy of parameters for characterizing three dimensional topography IV: Parameters for characterizing spatial and hybrid properties, Wear 178 (1994) 45 60. [55] Z. Peng and T.B. Kirk, Two dimensional fast Fourier transform and power spectrum for wear part icle analysis, Tribol. Int. 30 (1997) 583 590. [56] L. Starck, F. Murtag and A. Bijaoui, Image Processing and Data Analysis: The Multiscale Approach, Cambridge University Press, New York, 1998 [57] R.S. Sayles and T.R. Thomas, Surface topography as a non stationary random process, Nature 271 (1987) 431 434. [58] J.I. McCool, Relating profile instrument measurements to the functional performances of rough surfaces, J. Tribol. 109 (1987) 264 270. [59] A. Majumdar and B. Bhushan, Role of fractal geo metry in roughness characterization and contact mechanics of surfaces, J. Tribol. 112 (1990) 313 327. [60] A. Majumdar and C.L. Tien, Fractal characterization and simulation of rough surfaces, Wear 136 (1990) 459 484. [61] M.V. Berry and Z.V. Lewis, On the Weierstrass Mandelbrot fractal function, Proc. Roy. Soc. Lond. A 337 (1974) 379 394. [62] J.F. Archard, Contact and rubbing of flat surfaces, J. Appl. Phys. 24 (1953) 981 988. [63] J. Dyson and W. Hirst, The true contact area between solids, P roc. Roy. Soc. Lond. A 243 (1954) 309 313. [64] J.F. Archard, Elastic deformation and the laws of friction, Proc. Roy. Soc. Lond. A 243 (1957) 190 205. [65] H. Hertz and J. Reine, Uber die Beruhrung fester elasticher Korper, Agnew. Math. 92 (1882) 15 6 171. [66] J.A. Greenwood and J.B.P. Williamson, Contact of nominally flat surfaces, Proc. Roy. Soc. Lond. A 295 (1966) 300 319.

PAGE 104

104 [67] K.L. Johnson, Contact Mechanics, Cambridge University Press, London, 1987 [68] B.B. Mikic, Analytical studies of t he contact of nominally flat surfaces: Effect of previous loading, J. Lubrication Technol. 93 (1971) 451 459. [69] J.A. Greenwood, Adhesion of elastic spheres, Proc. Roy. Soc. Lond. A 453 (1997) 1277 1297. [70] K.L. Johnson, K. Kendall and A.D. Rober ts, Surface energy and the contact of elastic solids, Proc. Roy. Soc. Lond. A 324 (1971) 301 313. [71] B.V. Derjaguin, V.M. Muller and Y.P. Toporov, Effect of contact deformations on the adhesion of particles, J. Colloid Interface Sci. 53 (1975) 314 32 6. [72] D. Tabor, Surface forces and surface interactions, J. Colloid Interface Sci. 58 (1977) 2 13. [73] D. Maugis, Adhesion of Spheres: The JKR DMT Transition Using a Dugdale Model, J. Colloid Interface Sci. 150 (1992) 243 269. [74] K.L. Johnson, Mechanics of adhesion, Tribol. Int. 31 (1998) 413 418. [75] R.W. Carpick, D.F. Ogletree and M. Salmeron, A general equation for fitting contact area and friction vs load measurements, J. Colloid Interface Sci. 211 (1999) 395 400. [76] B.N.J. Persson Contact mechanics for randomly rough surfaces, Surf. Sci. Rep. 61 (2006) 201 227. [77] R.A. Onions and J.F. Archard, The contact of surfaces having a random structure, J. Phys. D: Appl. Phys. 6 (1973) 289 304. [78] R. Ariano, The coefficients of f riction between rubber and various materials: part II Gripping friction of rubber belting, Rubber Chem. Technol. 3 (1930) 286 293. [79] R. Ariano, The relation between forces and deformations with special reference to rubbers, Rubber Chem. Technol. 3 (1930) 62 67. [80] F.L. Roth, R.L. Driscoll and W.L. Holt, Frictional properties of rubber, J. Research Natl. Bur. Standards 28 (1942) 439 462. [81] A. Schallamach, Surface condition and electrical impedence in rubber friction, Proc. Phys. Soc. Lond. Sect. B 66 (1953) 817 826.

PAGE 105

105 [82] K.A. Grosch, The relation between the friction and visco elastic properties of rubber, Proc. Roy. Soc. Lond. A 274 (1963) 21 39. [83] A. Schallamach, A theory of dynamic rubber friction, Wear 6 (1963) 375 382. [84] J.A. Greenwood and D. Tabor, The friction of hard sliders on lubricated rubber: The importance of deformation losses, Proc. Phys. Soc. 71 (1957) 989 1001. [85] B.N.J. Persson, On the theory of rubber friction, Surf. Sci. 401 (1998) 445 454. [86] D. T abor, The mechanism of rolling friction II: The elastic range, Proc. Roy. Soc. Lond. A 29 (1955) 198 220. [87] M. Barquins and R. Courtel, Rubber friction and the rheology of viscoelastic contact, Wear 32 (1975) 133 150. [88] M. Barquins and A.D. Rob erts, Rubber friction variation with rate and temperature: some new observations, J. Phys. D: Appl. Phys. (1986) 547 563. [89] M.K. Chaudhury and T. Weaver, Adhesive contact of cylindrical lens and a flat sheet, J. Appl. Phys. 80 (1996) 30 37. [90] B .N.J. Persson, O. Albohr, C. Creton and V. Peveri, Contact area between a viscoelastic solid and a hard, randomly rough, counter sample, J. Chem. Phys. 120 (2004) 8779 8793. [91] R. Stribeck, Die wesentlichen Eigenschaften der Gleit und Rollenlager (Th e basic properties of sliding and rolling bearings), Zeitschrift des Vereins Deutscher Ingenieure 36 (2002) 1341 1348, 1432 1438, and 1463 1470. [92] O. Mollenhauer, S.I.U. Ahmed, F. Spiller and H.Haefke, High precision positioning and measurement system s for microtribotesting, Tribotest. 12 (2006) 189 199. [93] W.A. Bonin, Multi Dimensional capacitive transducer, patent no. 579,996 (1997) [94] P.L. Dickrell, S.B. Sinnott, D.W. Hahn, N.R. Raravikar, L.S. Schadler, P.M. Ajayan and W.G. Sawyer, Frict ional anisotropy of oriented carbon nanotube surfaces, Tribol. Lett. 18 (2005) 59 62. [95] H. Liu and B. Bhushan, Adhesion and friction studies of microelectromechanical systems/nanoelectromechanical systems using a novel microtriboapparatus, Papers fro m the 49th International Symposium of the American Vacuum Society (2003)

PAGE 106

106 [96] A.C. Rennie, P.L. Dickrell and W.G. Sawyer, Friction coefficient of soft contact lenses: measurements and modeling, Tribol. Lett. 18 (2005) 499 504. [97] R.L. Hudson, O Ri ng Design and Materials Guide, R.L. Hudson & Company, 2011 [98] K. Kendall, The adhesion and surface energy of elastic solids, J. Phys. D: Appl. Phys. 4 (1971) 1186 1196. [99] A.D. Roberts, Surface charge distribution in rubber adhesion and friction, J. Phys. D: Appl. Phys. 10 (1977) 1801 1819. [100] B.N.J. Persson, O. Albohr, G. Heinrich and H. Ueba, Crack propragation in rubber like materials, J. Phys. Condens. Matter 17 (2005) R1071 1142. [101] B.N.J. Persson and E.A. Brener, Crack propagation in viscoelastic solids, Phys. Rev. E 71 (2005) [102] B.N.J. Persson and A.I. Volokitin, Rubber friction on smooth surfaces, Eur. Phys. J. E Soft Matter 21 (2006) 69 80. [103] B.N.J. Persson, Theory of rubber friction and contact mechanics, J. Chem. Phys. 115 (2001) 3840 3861. [104] B.N.J. Persson, Rubber friction and tire dynamics, J. Phys.: Condens. Matter 23 (2011) 1 14. [105] A. Chateauminois, T.N. Danh, P. Paolino, M.C. Audry, C. Fretigny, Y.L. Chenadec, M. Portigliatti and E. Barthel, Sur face pressure and shear stress fields within a frictional contact on rubber, J. Adhes. 87 (2011) 235 250. [106] A.R. Sakvoor and G.A.D. Briggs, Effect of tangential force on contact of elastic solids in adhesion, Proc. Roy. Soc. A 356 (1977) 103 114. [107] K. Vorvolakos and M.K. Chaudhury, The effects of molecular weight and temperature on the kinetic friction of silicone rubbers, Langmuir 19 (2003) 6778 6787. [108] F. Wu Bavouzet, J. Clain Burckbuchler, A. Buguin, P.G.D. Gennes and F. Brochard Wya rt, Stick slip: wet versus dry, J. Adhes. 83 (2007) 761 784. [109] A.W. Adamson, Physical Chemistry of Surfaces, fifth ed., John Wiley & Sons, Inc., New York City, 1990

PAGE 107

107 [110] W.A. Zisman, Relation of equilibrium contact angle to liquid and solid const itution, in R.F. Gould (Eds.), Contact Angle, Wettability, and Adhesion, American Chemical Society, Washington DC, 1964, pp. [111] J. Kraxner, M. Liska, R. Klement and M. Chromeikova, Surface tension of borosilicate melts with the composition close to t he e glass, Ceramics 53 (2009) 141 143. [112] J.Y. Shin, H.Y. Park, C. Liu, J. He and S.C. Kim, Chemical structure and physical properties of cyclic olefin copolymers, Pure Appl. Chem. 77 (2005) 801 814. [113] G.M. Bartenev and L.A. Akopyan, Surface free energy of polymers, Fiziko Khimicheskaya Mekhanika Materialov 5 (1969) 670 676. [114] B.J. Fregly and W.G. Sawyer, Estimation of discretization errors in contact pressure measurements, J. Biomech. 36 (2003) 609 613.

PAGE 108

108 BIOGRAPHICAL SKETCH Jennif er Vail was born in 1984 in Jacksonville, Florida. She was raised in nearby Ponte Vedra Beach where she attended Allen D. Nease High School. After spending her childhood attending football games at the University of Florida, she entered the university In August 2002. She worked as an undergraduate teaching assistant for two courses and interned with General Electric before receiving her Bachelor of Science in mechanical engineering in December 2006. Jennifer spent the spring of 2007 working at Walt Disn ey World before deciding to return to the University of Florida for her graduate education. Working in the Tribology Laboratory under Professor Greg Sawyer, Jennifer received her Master of Science in 2009 and Doctor of Philosophy in 2012 in mechanical eng ineering. She spent two years studying polymer composites before shifting focus to elastomeric materials. During her time at the University of Florida, the football team won two national championships and the basketball team won back to back championship s.