In Situ Investigation of Sliding Direction Dependency on the Wear of Single Crystal Magnesium Oxide

MISSING IMAGE

Material Information

Title:
In Situ Investigation of Sliding Direction Dependency on the Wear of Single Crystal Magnesium Oxide
Physical Description:
1 online resource (50 p.)
Language:
english
Creator:
Marchman, Kellon Ryan
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Mechanical Engineering, Mechanical and Aerospace Engineering
Committee Chair:
Sawyer, Wallace G
Committee Members:
Ifju, Peter
Angelini, Thomas E.

Subjects

Subjects / Keywords:
crystal -- mgo -- wear
Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre:
Mechanical Engineering thesis, M.S.
Electronic Thesis or Dissertation
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )

Notes

Abstract:
Ionic solids are unique materials to study, offering a wide range of physical properties and varying crystallographic tendencies. Amongst this group is magnesium oxide (MgO). Magnesium oxide is a well suited ceramic in tribological testing due to its relatively high hardness, high melting temperature, and its simple cubic crystal structure. In this study, the effect of crystallographic direction on MgO's wear rate is examined. A pin-on-disk tribometer is combined with a scanning white light interferometer to provide profilometry scans of the wear track for wear rate measurements. The final wear result was determined after a sliding distance of 180 meters using a normal load of 1 Newton in a dry nitrogen environment: a distinct sinusoidal relationship between the wear rate and angular position of the wear track exists. This is indeed due to anisotropy of the material in different crystallographic directions. It is proposed that a main contributor to this phenomenon is the varying distribution of the resolved shear stress along the slip planes of the crystal. To test this result and its application to single crystal ionic solids, sodium chloride was also tested, as well as an amorphous glass specimen. A similar trend was found in sodium chloride whereas no such trend manifested itself in the noncrsytalline glass sample.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2012.
Local:
Adviser: Sawyer, Wallace G.
Statement of Responsibility:
by Kellon R Marchman.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044271:00001


This item has the following downloads:


Full Text

PAGE 1

1 IN SITU INVESTIGATION OF SLIDING DIRECTION DE PENDENCY ON THE WEAR OF SINGLE CRYSTAL MA GNESIUM OXIDE By KELLON RYAN MARCHMAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DE GREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Kellon Ryan Marchman

PAGE 3

3 To my paren ts, Ken and Rhonda Marchman, my brother, Alan Marchman and my wife, Jacquelyn Marchman for their lifelong encouragement, advice, and inspiration

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my graduate advisor Dr. W.G. Sawyer for sharing his wealth of knowledge through his teaching and mentoring, and for sharing numerous life enrich in g expe riences with me over the past three years. I would also like to thank the members of my committee Dr. Tommy Angelini and Dr. Peter G. Ifju for their time and teaching efforts through my tenure at the University of Florida I would like to thank my loving and encouraging wife Jacque lyn Marchman for her support, patience and encouragement while finishing this research. I would like to extend a special thank you to my colleague Dr. Brandon A. Krick for his help and support throughout my time in the Universi ty of Florida Tribology Laboratory. A thank you also goes to Ira Hill who aided me in automating some data retrieval for this work. I would like to thank Dr. Perry and Xueying Zhao for their help with crystallographic analysis. I would also like to than k both current and former members of the University of Florida Tribology Laboratory for their co llaborations and encouragement through my term.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 10 Experimental Practices Tribometry ................................ ................................ ......... 10 Magnesium Oxide: A Brief Overview and Previous Studies ................................ .... 12 Magnesium Oxide ................................ ................................ ............................ 12 Relevant Previous Studies ................................ ................................ ............... 13 Optical Interferometric Profilometry ................................ ................................ ........ 16 Summary of Research ................................ ................................ ............................ 17 2 EXPERIMENTAL METHODS ................................ ................................ ................. 18 Overview of Tribometer ................................ ................................ ........................... 18 Sample and Load Positioning ................................ ................................ ........... 18 Normal and Friction Force Measurement ................................ ......................... 19 In Situ Optical Profilometry ................................ ................................ ............... 22 Data Acquisition and Scan Automation ................................ ............................ 22 Environmental Control ................................ ................................ ...................... 23 Sample Selection ................................ ................................ ................................ .... 23 Experimental Techniques ................................ ................................ ....................... 24 Unce rtainty in Measurements ................................ ................................ ................. 25 3 EXPERIMENTAL RESULTS AND DISCUSSION ................................ ................... 26 4 CONCLUSIONS ................................ ................................ ................................ ..... 39 APPENDIX A UNCERTAINTY ANALYSIS ................................ ................................ .................... 40 Uncertainty in Coefficient of Friction Measurements ................................ ............... 40 Uncertainty in Wear Rate Measurements ................................ ............................... 42 B RESOLVED SHEAR STRES S CALCULATIONS ................................ ................... 45 LIST OF REFERENCES ................................ ................................ ............................... 47 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 50

PAGE 6

6 LIST OF TABLES Table page 2 1 Uncertainty analysis calculations for friction coefficient measurements ............. 25 2 2 Uncertainty analysis calculations for wear rate measurements of MgO ............. 25 3 1 Testing parameters used ................................ ................................ .................... 29 A 1 Uncertainty analys is calculations for friction coefficient measurements ............. 44 A 2 Uncertainty analysis calculations for wear rate measurements of MgO ............. 44

PAGE 7

7 LIST OF FIGURES Figure page 1 1 MgO has a simple cubic close packed structure tha t is easy to model but provides substantial wear resistance. ................................ ................................ 13 1 2 Schematic of the light paths and components of scanning white light interferometry. Adapted from [23]. ................................ ................................ ...... 17 2 1 Illustration of the tribometer used for this study. ................................ ................. 18 2 2 Exaggerated Leaf spring flexure deformation. Axes used for stiffness calculations are shown. ................................ ................................ ...................... 20 3 1 Friction coefficient, shown as a function of track location in degrees. This plot shows fairly isotropic friction values over the length of the track. ................. 26 3 2 The distribution of normal force imposed as a function of the track location in degrees. ................................ ................................ ................................ ............. 28 3 3 A schematic showing the locations of the scans around the wear track. Representative profile scans of the wear track are displayed. Each rectangle on the track represents the scaled sp atial size of the scan (583 x 437m). ....... 31 3 4 A typical cross sectional profile of the wear track of MgO. Additionally, how area l oss determinations were made. ................................ ................................ 31 3 5 Anisotropic wear behavior of single crystal MgO expressed as a function of the angular position, measured from the center of the wear track. The black arrows represent the family of sliding direction at that position. Directions were verified by LEED. ................................ ................................ ....................... 32 3 6 S lip plane notations ................................ ................................ ............................ 35 3 7 The resolved shear stresses as a function of sliding direction on the (100) plane. The solid path represents the magnitude between the dif ferent slip planes. This trend follows perfectly with the wear rate trend seen above. ..... 36 3 8 The severity of the plastic defo rmation rapidly increases over a relatively short period of sliding. The cracks began to propagate at approximately 3700 cycles. ................................ ................................ ................................ ................ 37 3 9 A comparative and control test to the MgO results above. The ionic crystalline structure produces similar dependency results while the amorphous glass shows constant wear. ................................ ............................. 38 A 1 Misalignment error in surface normal direction. Measured via optical techniques. ................................ ................................ ................................ ......... 42

PAGE 8

8 B 1 The notations used in the resolved shear stress calculations. The F f vector is held constant and is swept along 360 degrees for each of the four active slip planes. ................................ ................................ ................................ .......... 45

PAGE 9

9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science IN SITU INVESTIGATION OF SLIDING DIRECTION DE PENDENCY ON THE WEAR OF SINGLE CRYSTAL MA GNESIUM OXIDE B y Kellon Ryan Marchman May 2012 Chair: W.G. Sawyer Major: Mechanical Engineering Ionic solids are unique materials to study, offering a wide range of physical properties and varying crystallographic tendencies. Amongst this group is magnesium oxide (MgO). Magnesium oxide is a well suited ceramic in tribological te sting due to its relatively high hardness, high melting temperature, and its simple cubic crystal structure. In this study, the examined. A pin on disk tribometer is combined with a scanning white light interferometer to provide profilometry scans of the wear track for wear rate measurements. The final wear result was determined after a sliding distance of 180 meters using a normal load of 1 Newto n in a dry nitrogen environment: a distinct sinusoida l relationship between the wear rate and angular position of the w ear track exists. This is indeed due to anisotropy of the material in different crystallographic directions. It is proposed that a main contributor to this phenomenon is the varying distribu tion of the resolved shear stress along the slip planes of the crystal. To test this result and its application to single crystal ionic solids sodium chloride was also tested, as well as an amorphous glass specimen. A similar trend was found in sodium chl oride whereas no such trend manifested itself in the noncrsytalline glass sample.

PAGE 10

10 CHAPTER 1 INTRODUCTION Dynamically interacting solid body surfaces are inherently d eleterious to the life of the body unless measures are taken to reduce the wear and frictio n occurring at their interface. The deformation of the solid body over a progression of cycles is known as wear. This d eformation occurs due to high contact pressures present during sliding These high contact pressures are caused by the microscopic contact areas of asperities, or peaks found on a nominally flat surface, with the interacting surface [ 1 ] For these reasons, contact mechanics of rough surfaces are of great importance in tribological wear and friction studies [ 2 ] Experimental Practices Tribo metry Tribological properties such as friction coefficients and wear rates are very important parameters to consider in engineering design. High friction coefficients can lead to minimal efficiency in dynamic systems and also increase the wear of component s. With increasing wear, numerous wear debris particles are generated which can inhibit the function of surrounding operations thus leading to a very short life span. Tribometry embodies the associated experimental techniques and methods necessary for qua ntifying these important design parameters. The friction coefficient, is defined as the ratio of the friction force to the normal force. Due to the asperity contacts as discussed previously, the magnitude of the normal and friction forces are defined a s the sum of the forces present at the asperities There are numerous tools available for obtaining the desired normal force. Some common tools include dead weight loads, piezo electric driven cantilevers, and linear

PAGE 11

11 stages. The friction forces can be measu red with linear force transducers or displacement monitoring devices such as capacitance probes. A typical tribometer configuration for such friction testing is a spherical pin on a large flat surface, or counterface. The pin material is chosen based on pr operties of high hardness and an ideally smooth surface in order to concentrate both the friction and wear on the counterface, the material of interest. The uncertainties associated with friction coefficient measurements are of importance and detailed stud ies have been done by Schmitz et al [ 3 ] f or linearly reciprocating pin on f lat tribometers and Krick et al. [ 4 ] for pin on disk rotary tribometers The deformat ion that occurs due to sliding can be measured by the wear volume of the specimen. A relationship that expresses the wear rate of a material in a tribosystem has been defined by Archard and can be reduced to: ( 1 1 ) where K is the wear rate (typically expressed in units of mm 3 /(Nm)) s is the sliding distance in P is the applied normal load, and W is the volume lost due to wear [ 5 ] Volume loss can be measured by mass loss as well as optical scanning techniques. It is important, however, to understand the amount of uncertainty associated with volume loss measurements, as this directly correlates to the quality of the calculated wear rate. Schmitz et al defines the unc ertainty of a reciprocating pin on flat for mass loss based wear rate calculations [ 6 ] Surface scanning techniques are becoming increasingly common forms of volume loss measurem ents due to their high precision and ease of implementation. For relatively large wear tracks, however, it is difficu lt to examine the

PAGE 12

12 entire wear track due to the limited spatial resolution of surface scanning devices. To avoid this problem for circular wear tracks, it is acceptable to scan the wear track in subsections at constant intervals This technique invokes uncertainty due to the non uniformity of the wear track caused by variations in the applied load [ 7 ] Colbert et al discusses the importance of the number of scans necessary in order to reduce the uncertainty of the measurement to an acceptable level [ 8 ] Magnesium Oxide : A Brief Overview a nd Previous Studies Single crystal ionic solids are of interest in this study because of their natural formation habits a nd interesting mechanical properties. Ionic solids exhibit the high strength of chemica l bonding due to the large electronegativity difference between their two oppositely charged ions. Ionic solids can have numerous crystallographic structures but one com mon factor is that most have close packed atoms. This efficient packing means that the cations will tend to fill a hole that is slightly smaller than itself. This is necessary in order to prevent the larger like charge anions from contacting each other. Magnesium Oxide Single crystal m agnesium oxide (MgO) is an ionic solid ceramic that exhibits high melting temperature a nd relatively high hardness values of T m respectively. Magnesium oxide also exhibits a high elastic modulus and yield strength, allowing higher contact stresses to be attained while avoiding excessive deformation [ 9 ] It is a val uabl e material to study because of these desirable physical properties and its simple crystal structure. The crystal structure of MgO is that of rock salt, which has a face centered cubic lattice structure as seen below in Figure 1 1 The magnesium

PAGE 13

13 cations fill the octahedral holes in the lattice since the ionic radii ratio of magnesium cations to oxygen anions (r cation /r anion ) is 0.683, which is too large for tetrahedral holes and too small for cubic holes. Rock salt structure cleavage is well known and occurs along the (100) plane. Figure 1 1. MgO has a simple cubic close packed structure that is easy to model but provides substantial wear resistance. Relevant Previous Studies Wear studies on single crystal MgO date back to the 1960s and most were focused on understanding the brittle fracture and subsurface dislocation movements of MgO under wearing environments [ 9 13 ] R.J Stokes suggested that in some ceramics, although brittle fracture occurs plastic deformation can also occur at low temperatures and relatively low stresses [ 11 ] This notion was validated by R.P Steijn in 1963 when he investigated the subsurface damage of ionic crystals due to rubbing [ 12 ] Steijn conducted scratch tests on single crystal MgO and used etching techniques to verify dislocation movements around the scratch track. He found that the glide mot ions were

PAGE 14

14 Studies verifying these dislocation movement and slip plane results with electron transmission microscopy have been completed by J. Washburn et al. [ 14 ] Further work on dislocation motion was done by Stokes et al. on the dislocation movement of MgO and methods of locking dislocation movement in single crystals [ 10 ] Stokes was able to correlate some of the dislocation locking to the anisotropic plasticity of MgO. Studies were conducted on sliding speed and load dependency of wear in MgO on steel by Sugita et al. [ 15 ] Sugita found that as sliding speed increased, both the wear on the MgO single crystal slider and the steel disk decreased. He attributed this trend to a diffusion process occurring at the interface of the sliding and a reduction in the amount of brittle fr acture caused by lower sliding speeds. Amateau et al. and Dufrane et al. investigated the effect of rolling stresses on single crystal MgO [ 9 16 ] Amateau confirmed that plastic deformation incurred on the MgO sample was attributable to stresses which were similar to those calculated using Hertzian contact pressures for elastic conditions. single pass experiments based on the dependen ce of dislocation velocities due to applied stresses. In 1966, F.P Bowden and A.E Hanwell experimented with the friction of clean single crystal surfaces in very high vacuum (10 10 torr) to examine the effects of tenacious surface contaminant layers on their frictional behavior [ 17 ] Bowden experimented with single crystal MgO sliding on itself in specifically oriented crystallographic directions (sliding occurred in the [100] direction) It was found that by removing the layer of adsorbed gasses from the surface, the friction coefficients are

PAGE 15

15 greatly increased due to a lack of lubricious monolayers of oxygen and water. The contaminant layers were removed from the surface due to the rubbing action and due to the high vacuum, the recontamination rate is greatly reduced. Due to the interface of the clean surfaces, the friction is dominated by adhesive forces at the junctio n. The adhesion seen at the interface is directly related to the inherent interatomic bond strengths of the material. In a later study, Bowden et al. also investigated the frictional anisotropic behavior of MgO with a dependence on sliding direction [ 18 ] Bowden found that the fric tion of single c rystals of MgO did not depend on sliding direction if the sliding conical tip was of an angle greater than 150 (i.e. blunt or spherical tipped sliders) Knoop h ardness tests were conducted with the long axis of indention aligned in differe nt crystallographic directions. Bowden discovered that there is a hardness dependency on direction and that the hardness of the <100> family of directions is approximately half the value of the <110> directions. Bowden also found that the wear behavior of MgO produced chevron type cracks in the area directly surrounding the wear track. Through this study and previous efforts, it was deduced that these chevron patterns were formed due to dislocations occurring sub surface in response to sliding. These are th e same patterns that were found by Steijn [ 12 ] as discussed above. In situ tribology is a subset of tribological testing techniques that can be used for many purposes. Sawyer et al. used in situ profilometry and a linear reciprocating tribometer to understand the morphology of MoS 2 friction and wear [ 19 ] Argibay et al. used in situ methods to analyze sliding electrical contacts as a function of brush type and current densities [ 20 ] In [ 21 ] Wahl and Sawyer discuss the multiple applications for

PAGE 16

16 using in situ techniques. They discuss how integrating in situ in strumentation can allow for tri bologically interesting events can be correlated with physical effects of such changes Optical Interferometric Profilometry This method of surface profilometry is widely used in the metrology field due to its non destructive nature and ease of application Vertical s canning white light interferometers (SWLI) are common forms of this type of profilo meter For most applications, t hese interferometers use a Mirau interferometer coupled with a charge coupled device (CCD) image sensor to provide a topographical representation of a sample surface [ 22 ] A short coherence length light source projects a white light into a semi reflective beam splitter. One arm of the split beam is incident on a reference mirror and the other is incident on the testing sample. By varying the interferometer heig ht with a piezoelectric transducer, constructive and destructive interference fringes are created between the reflected split beams. These interference fringes are projected onto the CCD, which converts the image in to a digital array for further analysis b y software. These types of interferometers can have vertical resolutions of a few nanometers and lateral resolutions as low as 0.5 m/sample depending on the diffraction limit of the lig ht used and the magnification. Figure 1 2 depicts the light path of a SWLI similar to the model used for this study.

PAGE 17

17 Figure 1 2 Schematic of the light paths and components of scanning white light interferometry. Adapted from [ 23 ] Summary of Research In previous ionic solid research conducted over my term, interesting wear behavior manifested itself for numerous samples. One sample in particular that showed interesting wear was MgO. This study aims to further investigate the role s of crystallogr aphic structure, ionic bond strength, and sliding direction on the wear of MgO. A pin on disk tribometer is instrumented with a double leaf flexure and combined with a scanning white light interferometer for macroscopic wear testing and in situ wear track examination. A correlation with the anisotropic hardness findings of Bowden, as discussed above, and wear rate is anticipated. This work will discuss the experimental techniques, mathematical methods, and data acquisition necessary to explore this hypothes is.

PAGE 18

18 CHAPTER 2 EXPERIMENTAL METHODS Overview of Tribometer The tribometer used for this study is a custom adaptation of the in situ tribometer as described by Keith [ 24 ] This tribometer is a pin on disk apparatus that has the ability generate a wear experiment while simultaneous monitoring the evolution of wear. Figure 2 1 illustrates the tr ibometer componentry and layout as adapted for use in this work. Figure 2 1 Illustration of the tribometer used for this study. S ample and Load Positioning Samp le positioning was accomplished via a Physik Instrumente (PI) M 060 PD precision rotary stage controlled by a PI Mercury C 863 controller The stage incorporates a direct current drive servo motor with a worm gear reduction. The sample was rigidly attached to the stage using a clamp mechanism. The stage is capable of

PAGE 19

19 rotation speeds up to 90 deg/s and a unidirectional repeatability of 50 rad. Position feedback was accomplished via a rotary encoder with a resolution of 0.0018 degrees Wear track and force adjustments were achieved by OptoSigma linear ball bearing stages mounted orthogonally in the X and Z directions. The stages are micrometer driven with a resolution of 10 m. The resolution provided by these stages ensure d an accurate wear track radius position and normal force application. Normal and Friction Force Measurement This tribometer uses a 304 stainless steel monolithic biaxial force transducer to measure both the friction and normal forces [ 25 ] To compensate for sample leveling inaccuracies and to aid in providing a constant normal force, a double leaf spring flexure was incorporated into the tribometer design. The do uble stacked parallelogram l eaf spring configuration increased the torsi onal rigidity and limited out of plane rotations. The leaf spring flexure stiffness values were calculated using equations 2 1 through 2 3 [ 26 ] using the axes notations shown in Figure 2 2. ( 2 1 ) ( 2 2 ) ( 2 3 ) T he stiffness in the normal direction ( ) was calculated to be ~3.5 N/mm, while the stiffness in the lateral direction ( ) was three orders of magnitude higher at ~8,300

PAGE 20

20 N/mm. The torsional rigidity about the x axis ( ) as defined below in Figure 2 2, was calculated to be eight orders of magnitude higher than the normal stif fness at ~8.8 x 10 8 N mm/rad. These values indicate that the flexure assembly was well suited for only normal plane displacements with relatively low loads. Finite element analysis based simulations were also carried out and showed supportive values to the calculated values above. Figure 2 2. Exaggerated Leaf spring flexure deformat ion. Axes used for stiffness calculations are shown. The monolithic flexure uses two full strain gage Wheatstone bridges for both the normal and the lateral force measurements. There are two gages on each side of the flexure for each measurement. The full bridge configuration provides temperature compensation to the circuit and aids in preventing crosstalk between the signals. The gages are placed on the flexure at points furth est from the loading in order to increase sensitivity. The transducer is paired with Sensotec UV 10 in line amplifiers to provide

PAGE 21

21 direct current excitation voltage and conditions the gage outputs to be read by the data acquisition card Using the same nota tion above, the normal force stiffness ( ) and lateral stiffness ( ) of the monolithic flexure w ere found to be an order of magnitude higher than those of the leaf flexure at 71.4 N/mm and 19.0 N/mm, respectively Furthermore, experimental stiffness calculations were carried out on the entire flexure assembly. With known displacements induced on the vertical axis, the measured forces were recorded and a linear fit applied to the data. The composite stiffness for th e assembly of the leaf flexure and monolithic flexure was 3.07 N/m. This value is very close to the calculated normal direction stiffness for the leaf flexure as shown above, indicating that the monolithic flexure is considered mostly rigid in the system. The combination of this stiff flexure with the compliant leaf spring flexure ensures that the desired normal force is imposed on the sample and remains mostly constant through the wear track path. Prior to starting experiments on an instrument, it is ess ential to calibrate the strain reading flexure to obtain the calibration constant that can be used to back out forces. To calibrate the flexure assembly, the 90 degree angle bracket that the flexure assembly is mounted to was removed from the tribometer an d bolted to an optical table for rigidity. Known masses were then hung from the cantilever and voltage shifts incurred by the Wheatstone bridge were recorded. The flexure was calibrated up to two times the target nominal force to provide an acceptable rang e of calibration data. The process was repeated with the flexure rotated 90 degrees to calibrate the tangential direction stiffness. A linear fit was imposed on the data for both the normal and tangential directions and values were recorded as 1.637 N/V a nd 0.67 N/V, respectively.

PAGE 22

22 In Situ Optical Profilometry The optical profilometry is provided by a Veeco Wyko NT9100. The po sitioning of the profilometer was opposite the side of the pin su ch that the pin was unloaded from the sample and the sample can be s canned around its entire track. After each cycle interval, the disc was rotated in increments of 6 degrees and scanned, for a total of 60 optical profile measurements. The profilometer was set in vertical scanning interferometry (VSI) mode. This mode allo ws for large step sizes to be resolved while still keeping a vertical height resolution of around 3 5 nm with lateral spatial resolution of 0.5 m/sample The ins trument was equipped with a 20X Mirau objective and imaged using a field of view (FOV) of 0.55 X for a magnification of height values of approximately 11X. The measurement field for a single scan on MgO was 583m by 437m. Data Acquisition and Scan Automation Wired connections were passed through a National Instruments SCB 68 connector block and d ata acquisition was handled through a N ational Instruments PCI 6621 16 bit card. The controls programming and data storage was accomplished through LabVIEW. The acquisition rate of the force data was set to sample at a rate of 1 kHz, at which the posi tion data from the rotary stage were also recorded. The data w ere averaged every second and combined with rotary position of its associated acquisition cycle. At every cy cle stop interval, a total of 60 equally spaced optical profilometry scans were taken around the wear track which was automated for efficiency Once the prescribed cycle interval was completed and the pin was unloaded, the automation routine commanded the stage to rotate to a set number of degrees before commanding the SWLI computer

PAGE 23

23 to act ivate a new scan. Once the scan is complete, the scan dataset was saved and the process was repeated Environmental Control Many ionic crystals, such as NaCl and MgO, are largely hygroscopic. This means that the adsorption of water molecules greatly effec ts the physical and chemical properties of the sample [ 27 ] To avoid this, a dry environment was established through an impingement of dry ni trogen gas onto the surface which was then enveloped by a local environment al control container. The relati ve humidity recorded after a brief purge period was approximately 3 % RH. The relative humidity measurement was made by an Omega Thermo Hygrometer (model number RH411). The flow of nitrogen was reduced once the volume was purged and was kept constant throug h the entirety of the testing. Sample Selection Optical window samples were chosen as the running surface for this study because of their superior flatness and exceptionally smooth polished sur f aces. Prior to testing, s urface roughness measurements were taken of the MgO optical windows obtained from Crystran and averages were below R a = 2nm. All samples tested were 25mm diameter disk s of thicknesses between 2 and 3 mm. For this study, chromium oxide doped sapph ire (synthetic ruby) ball lenses 3.18 m m in diameter served as the pin Pre test SWLI surface scans of the ruby ball showed a radius of curvature of approximately 1.60 mm and a surface roughness, R a of ~4.5 m. Ruby has a hardness that is approximately 2.5 times greater than MgO with hardness measurements of ~2300 and ~910, respectively. The ruby ball lens was set on the end of a cup point set screw using a fast acting cyanoacrylate adhesive.

PAGE 24

24 Experimental Technique In order to successfully record the data necessary for tracking the rate based on the azimuthal position, several experimental steps are essential. Once the sample was clamped into the sample holder, the SWLI objective was placed over the position at which the wear track would evolve. This location was determ ined by using the known geometries of the sample holder and the calibrated wear track diameter. The pin was then lowered into contact with the sample and the normal force was set based on the 1000 Hz data displayed on the DAQ computer. Once the desired nor mal force was acquired, the next step was to set up the testing parameters on the DAQ computer. The experiments were completed such that the wearing event would take place for a defined increment of sliding cycles (a sliding cycle is one 360 degree rotatio n) after which the pin would be unloaded from the sample and the scanning cycle would commence. The scanning process was automated such that the sample would rotate six degrees, pause for approximately 20 seconds until the instrument completed the scan an d saved the data, and then the sample would rotate another six degrees and repeat the process until the sample was rotated 360 degrees (one scanning cycle) A total of 60 scans around the entire wear track were completed using this method. With the automat ed routine, one scanning cycle interval was approximately 30 minutes. Once the scanning cycle was completed, the pin was reloaded onto the surface making sure to achieve the same normal load, without having to dismount the sample or adjust the wear track position. A total of 1800 wear track scans were taken over the 3000 cycle test.

PAGE 25

25 Uncertainty i n Measurements I t is very important to develop an uncertainty analysis to validate the reported results of experimental data In tribological testing systems, uncertainties in wear rates and friction coefficients can stem from numerous sources but typically the largest source of uncertainty is derived from the precision of the instrument. Much work has been done in this area in an attempt to minimize these unce rtainties and maximize the quality of experimental measurements. Uncertainty analysis methods described by multiple previous studies [ 3 6 8] are emplo yed here to assess the quality of data supplied in this work. A complete method for calculating the u ncertainties is provided in App endix A, with summarized values in Tables 2 1 and 2 2. The uncertainty in the wear rate ranged from 3% for high volume loss to 8% for low volume losses. Table 2 1. Uncertainty analysis calculations for friction coefficient measurements Parameter Nominal Value Standard uncertainty u(x) Sensitivity, Contribution (%) F N (N) 1.01 0.025 0.20 5.65 F T (N) 0.202 0.004 0.99 94.35 Combined standard uncertainty u C 4.0 x 10 3 Table 2 2. Uncertainty analysis calculations for wear rate measurements of MgO Parameter Nominal Value Standard uncertainty u(x) Sensitivity, Contribution (%) F N (N) 1.01 0.025 0.20 25.34 V loss (mm 3 ) 1.3 x 10 4 5.61 x 10 6 5.5 x 10 3 74.65 d (m) 180 1 x 10 8 4.08 x 10 9 0 Combined standard uncertainty u C 3.6 x 10 8

PAGE 26

26 CHAPTER 3 EXPERIMENTAL RESULTS AND DISCUSSI ON As discussed previously, Bowden et al found that the friction coefficient of single crystal MgO manifested anisotropic behavior for conical pin sliders with angles less than 150 degrees but became much more isotropic with blunt pins. With sharp indenters, Bowden found that the friction coeff icient would increase by as much as three times According to this finding, a hemispherical pin should provide fairly isotropic friction behavior regardless of the sliding direction Friction coeffic ient as a function of the track location angle is shown below in Figure 3 1. Figure 3 1. Friction coefficient, shown as a function of track location in degrees. This plot shows fairly isotropic friction values over the length of the track.

PAGE 27

27 Figure 3 1 verifies that remains relatively constant throughout the circumference of the wear track. This consistency is important to the wear study in this work. A constant friction force imposes a constant and uniform shear stress distribut ion along surface of the track, which provides equal opportunity for a wear event to occur at every location along the track. Thus, any wear accumulated by the process is only attributable to the material properties instead of varying load condition s Since the va lues above were calculated based the ratio of friction force to normal force as shown here, (3 1) it is also very important to maintain a constant normal force along the circumferenc e of the wear track. A conservative normal force of 1 N was selected for these tests in an attempt to minimize brittle fracture of the specimens. With a nominally normal load of 1 N and a ruby spherical pin with a radius of 1.5875 mm, the Hertzian contact pressure for the interface was calculated to be approximately 1160 MPa. With a reported compressive strength value for MgO of 2400 MPa, the contact pressure using the above parameters is less than 50% of the compressive strength of MgO. Figure 3 2 below shows the distribution of normal forces al ong the circumference of the track for several cycles. Spatially resolved p osition data w ere acquired from the synced with force data recorded by the transducer. The normal force data is plotted in 20 degree increments in orde r to provide an accurate representation of how the normal force varies. Averaging all of the normal force position data yielded an average F N of 1.01 N. The standard deviation s of the normal force for

PAGE 28

28 several cycles were calculated and the average of these deviations per cycle was a low 0.006 N. Furthermore, an average peak to peak normal force value was found to be approximately 0.03 N. This normal force uniformity means that axial run out deviations are minimized, thus reducing the associated uncertaintie s as discussed the following section The reduction of the normal force variation was achieved by a combination of leveling techniques, parallel testing samples and implementation of a compliant flexure. Figure 3 2. The distribution of normal force imposed as a function of the track location in degrees. As discussed previously, the friction and normal forces were obtained using a monolithic cantilever transducer with a leaf spring flexure. The transducer and the con ditioner pair output a voltage that is multiplied by stiffness coefficient determined for

PAGE 29

29 the flexure in order to obtain the force acting on the flexure. The equations used to calculate the normal and friction force are shown below in equations (3 2) and ( 3 3), respectively. (3 2) (3 3) In the above, C represents the calibration constant for the flexure in N/V, and V is the output voltage of the signal conditioner. The subscripts N and T represent th e normal and tangential direction, respectively. given number of cycles. The wear studies in this work were carried out using the testing parameters summarized below in Table 3 1. Table 3 1. Testing parameters used Parameter Value s Applied nominally normal load (F N ) Nominal contact pressure Speed (angular ( ) linear (v) ) 1 N 1160 MPa 71 deg/s, 12 mm/s 3000 180m 19mm 3% Dry nitrogen Number of cycles (N) Sliding distance (d) Nominal wear track diameter (D) Relative humidity Cover gas

PAGE 30

30 For calculating the wear rate, the above parameters and geometries can be implemented based on equation (1 1). (3 4) In order to determine the volume loss by wear, sever al steps were taken. At each six degree increment along the track, an optical scan was taken of the wear track. The process was automated such that once the scan finished the dataset was saved, the stage was rotated another six d egrees and the process was repeated. Figure 3 3 depicts the scanning increments around the disk, as well as representative wear track profile scans. The wear track was centered in the image such that a profile line scan would pass through the track at the specified degree interval and normal to the tangential direction. With line scans ob tained for each degree interval, the data w ere then processed in Matlab to calculate the area below the surface that is contained by the wear track. Use of the trapezoidal integration rule was implemented to determine this area loss. To obtain the volume loss, the area loss was then integrated along the circumference of the track. A representative line scan of the MgO wear track is shown below in Figure 3 4 The dimensions of the track, however, are not constant at every scan location. Since such little wear occurred after 3000 cycles the wear track width was not always explicitly clear. For each profile scan, the wear track width was chosen by referencing the profile data with the original optical scan taken by the SWLI. This way, the correct points were chosen to integrate under for each scan thus minimizing uncertainties in the area calculations.

PAGE 31

31 Figure 3 3. A schematic showing the locations of the scans around the we ar track. Representative profile scans of the wear track are displayed. Each rectangle on the track represents the scaled spatial size of the scan (583 x 437m). Figure 3 4 A typical cross sectional profile of the wear track of MgO Additionally, how area loss determinations were made.

PAGE 32

32 Anisotropic wear behavior was found over a 3000 cycle wear experiment on a single crystal MgO specimen. The anisotropic trend can be seen below in Figure 3 5 From the figure, it is evident that t he wear rate dependency on sliding speed is manifested by a sinusoidal wave of periodicity. The wear rate data is fitted with a sinusoid that is centered about 7.73 x 10 7 mm 3 /Nm with peak to peak values ranging from as high 1. 16 x 10 7 mm 3 /Nm down to 3.90 x 10 7 m m 3 /Nm Figure 3 5 Anisotropic wear behavior of single crystal MgO expressed as a function of the angular position measured from the center of the wear track. The black arrows represent the family of sliding direction at that position. Directions were verified by LEED.

PAGE 33

33 The specific crystallograph ic sliding directions above are based on a low energy electron diffraction (LEED) analysis. The LEED analysis provides crystallographic orientation by using electron diffraction patterns. The re sults of the analysis showed that the <110> direction was approximately 23 from wear track tangential direction at the assigned 0 position. Knowing that the <110> and <100> directions must be separated by 45, the data w ere examined and it was verified t hat the direction notations above are indeed correct. The magnitudes of the wear rates seen above in Figure 3 5 are representative of a low wearing solid. The wear resistance strength of MgO can be attributed to numerous crystal chemical properties Wi th large electronegativity differences between magnesium and oxygen, the strength of the ionic bonding is high This also correlates to a high percent iconicity of approximately 68% [ 28 ] The ionic potential is an important property of an ionic solid as it describes the charge density at the surface of the ion. It also provides a sense of how strongly an ion will be elec trostatically attracted to ions of opposite charge or repelled by ions of like charge. The ionic potential as described by Erdemir [ 29 ] is the ratio of the cationic charge to the cationic radius For MgO, the ionic pot ential is calculated as approximately 2.8, where as for a similar structured 1(+), 1( ) sodium chloride, the ionic potential is much less at approximately 0.98. Determined in previous studies 1 the activation energy (expressed in eV) necessary to displace an ion from its lattice position to an interstitial site (a Frenkel defect) is of great importance to understanding how strong the ionic bonds are This 1 W.G Sawyer, B. A. Krick, K. R. Marchman, Unpublished Ionic Solid Wear, 2011.

PAGE 34

34 potential energy a geometrical surface as shown below in equation (3 5) (3 5) In the above, e is the charge of an electron (1.6022 x 10 19 C), q is the magnitude of the nstant ( 8.98755 x 10 9 J m/C 2 ). The term M S represents a geometrical represents the amount of energy necessary to remove an anion or cation from the surface based on the closeness of its nearest nei ghbors. This value was found using a s ummation of neighbors in square loops radiating from the ion of interest. It was determined that the contribution to the total energy would converge around five loops away from the ion of interest. For MgO it was found that this activation energy was high at around 40 eV, indicating a strong bonding nature of the surface. To gain an understanding of why the wear rate varies with crystallographic directions, it is important to examine the resolved shear stresses and how they are distributed along the slip planes. The slip systems of MgO are well known and o ccur at {110} <110 > orientations [ 11 15 16 18 ] In this slip system, the re are six slip planes in total; t here are four slip planes that are at a 45 angle from the cube surface ({100} planes) and two that are 90 from the cube surface. Three of the six slip planes are shown in Figure 3 6.

PAGE 35

35 I n his studies with MgO as discussed previously, Bowden determined that the shear stresses due to sliding differ substantially from the <100> directions and the <110> directions. The r esolved shear stress on a slip plane can be defined as (3 6) where is the applied stress is the angle between the resultant stress axis and the slip plane normal and is the angle between the resultant stress axis and the slip direction Figure 3 6. Slip plane notations. a) nomenclature used for resolved shear stress calculations. b) the slip planes at 90 from the (100) surface By using the notation depicted above in Figure 3 6, the resolved shearing stresses can be determined. A more detailed solution to this problem is found in Appendix B: Resolved Shear Stress Calculations. From the calculations it is evident that the shear stresses differ substantially on which way the friction force vector is oriented (i.e. the sliding direction). Since the friction coefficient was mostly uniform around the wear track as shown above, the force vector as seen in Figure 3 6 was held constant throughout its sweep. Based on the ana lysis, when sliding in the <100> directions the shear stress is mostly carried by only two of the four 45 planes. These are the planes

PAGE 36

36 that intersect the surface perpendicular to the direction of sliding. Also, the shear stress is approximately two times the magnitude of that in the <110> directions. When sliding in the <110> directions the shear stresses are more evenly distributed over all four of the 45 slip planes. The two slip planes that intersect the surface at 90 bear no shearing stress for any o f the directions. Due to the higher magnitude of shear stresses, and being more concentrated, we would expect to see higher wear in those directions. The resolved shear stresses plotted against an angular sweep on the (100) surface is shown below in Figu re 3 7. Figure 3 7. The resolved shear stresses as a function of sliding direction on the (100) plane. The solid path represents the magnitude between the different slip planes. This trend follows perfectly with the wear rate trend seen above.

PAGE 37

37 The large dislocations observed by numerous previous studies on MgO typically manifest themselves as chevron type cracks emanating from the wear track. This cracking is a result of both high load and sliding distance. In order to obtain a wear track that produces a ccurate area loss data, this realm of plastic deformation was avoided Wear studies and measurements were only carried out prior to this severe cracking. In Figure 3 8 a before and afte r micrograph of the wear track shows how the sample can go from little plastic deformation to se vere plastic deformation in a span of ~700 cycles. Figure 3 8 The severity of the plastic deformation rapidly increases over a relatively short period of sliding. The cracks began to propagate at approximately 3700 cycles. Al ongside the MgO wear testing, two other comparative tests were run to validate the effect of crystallographic sliding direction on wear rates. Sodium chloride (NaCl) is of the same structure as MgO but has ionic charges of 1(+), 1( ). Based on the trend fo und in MgO, the NaCl wear rates should follow suit. The expected difference is that the wear rate magnitude will increase by at least a couple orders of magnitude. To test the trend in the opposite direction, a control test was administered on amorphous so da lime

PAGE 38

38 glass. The experimental procedures for the MgO experiment were followed precisely for these two comparative tests. The results for this set of tests are reported in Figure 3 9 below. Figure 3 9 A comparative and control test to the MgO results above. The ionic crystalline structure produces similar dependency results while the amorphous glass shows constant wear. The orientation for the NaCl crystal in the above data is not well defined at this point as it was for MgO. Currently, LEED data is being taken for future studies. It is evident, however that the trend is still there and the crystallographic directions are expected to match those defined for MgO.

PAGE 39

39 CHAPTER 4 CONCLUSION S A novel pin on disk tribometer was fitted to a scanning white light interferometer for an in situ approach at wear monitoring of single crystal magnesium oxide. The objective of this work was to determine if single crysta l magnesium oxide would wear dependently on the crystallographic sliding direction. To validate the measurements taken, t he combined standard uncertainties were determined and were on the order of a f ew percent. Using six degree scan increments for a total of 60 data points, the wear volume was calculated. The wear rates showed a strong crystallographic sliding direction dependency that is not induced by a varying normal force. This suggests that there is some sort of mechanical or atomically energetic proc ess that explains the anisotropic wear behavior. It is suggested that the resolved shear stresses on the slip planes are substantially different when sliding the <100> and <110> directions. In fact, it was found that when sliding in the <110> directions, t he shear stress acting on the active slip planes was half the magnitude of that when sliding in the <100> directions.

PAGE 40

40 APPENDIX A UNCERTAINTY ANALYSIS The combined standard uncertainty of a measured value can be found by applying the law of propagation of uncertainty, shown below in equation (3 4 ) (A 1) The combined standard uncertainty represents one standard deviation of the wear rate measurement. (3 5 ) (A 1) This root sum square method incorporates the pa d their associated uncertainties Uncertainty in Coefficient of Friction Measurements Equ ation (A 1) can be applied to equation (3 1) in order to determine the uncertainty in the friction force measurement. (A 2) In order to determine the uncertainty in the F N and F f components, the root sum square method is then applied to equations (3 2) and (3 3). The combined standard uncertainty equations for the normal and tangential forces are

PAGE 41

41 (A 3) (A 4) For the abov e equations, u(C N ) and u(C T ) are still unknown. These values are the uncertainties associated with the calibration constants in the normal and tangential directions. A subsequent application of the root sum square method to these parameters yielded calibra tion constant uncertainties of 7.3 x 10 3 and 1.7 x 10 2 N/V for the normal and tangential directions, respectively. The measured voltage uncertainties of the calibration sequence, V N and V T scale as a function of the number of bits, n, that the data acquisition card possess and is equal to The card used in this study was a 16 bit card yielding an uncertainty in voltage readings of 15 V. The calculated u ncertainty in the normal force based on the above equation was 0.005N. The fluctuation of the normal force, however, is approximately 0.025 N and therefore dominates over the calculated uncertainty value. Along with the residual uncertainties associated with measurement para meters as discussed above, misalignment errors can be a large source of uncertainty as well. As described in [ 3 ] the error fraction for obtaining accurate friction coefficient measurements is a function of the angular misalignment of the no rmal and tangential force axes. Using a manufacturing tol erance misalignment angle less than 2 as

PAGE 42

42 discussed below, and an average friction coefficient of 0.2 the error fraction associated with the friction coefficient measurement was approximately 4 % Based on the pin misalignment nomenclature shown in Figure (A 1 is less than 2. The cosine uncertainty of the normal force due to normal direction misalignment then comes out to approximately 0.0001 N [ 6 ] The largest contributor to the coefficient of friction measurement was found to be the friction force measurement itself. The results of the combined uncertainty analysis for friction coefficient measurements are summarize d below in Table A 1 Figure A 1 Misalignment error in surface normal direction. Measured via optical techniques. Uncertainty in Wear Rate Measurements Wear rate measurements also incur uncertainties based on the components of the measurand, K. Apply in g equation (3 4) to equation (A 1 ) yields the combined standard uncertainty associated with the wear rate measurement:

PAGE 43

43 (A 5) The method used here for determining the uncertainty in volume loss of the wear track was adopted from [ 8 ] which depends heavily on the number of cross sectional scans (N) used in the area measurements. The volume loss uncertainty also depends on the track diameter (D) and the standard deviation of the areas ( A ). (A 6) To minimize the uncertainty associated with the volume loss measurements, thus increasing the validity of the wear rate measurement, 60 total scans were taken around the wear track and analyzed for area loss. A depict ion of the sample, wear track, and scan orientat ions are shown Figure 3 3. The tilt of the sample in the tangential direction to the pin was determined from the SWLI scan to be approximately 0.06. This value is small enough to neglect its effect on the ov erall sliding distance. Therefore, t he uncertainty associated with the sliding distance, u (d), is d ominated by the uncertainty in the radius measurement. This measurement was made using a Veeco Dektak stylus profilometer, with X and Y axis resolutions of 0 .01 m Due to the uncertainties involved in making area loss measurements, it was determined that the volume loss measurement was the largest contributor to the wear rate uncertainty. The combined uncertainties for the wear rate measurements are summarized in Table A 2.

PAGE 44

44 Table A 1 Uncertainty analysis calculations for friction coefficient measurements Parameter Nominal Value Standard uncertainty u(x) Sensitivity, Contribution (%) F N (N) 1.01 0.025 0.20 5.65 F T (N) 0.202 0.004 0.99 94.35 Combined standard uncertainty u C () 4.0 x 10 3 Table A 2 Uncertainty analysis calculations for wear rate measurements of MgO Parameter Nominal Value Standard uncertainty u(x) Sensitivity, Contribution (%) F N (N) 1.01 0.025 0.20 25.34 V loss (mm 3 ) 1.3 x 10 4 5.61 x 10 6 5.5 x 10 3 74.65 d (m) 180 1 x 10 8 4.08 x 10 9 0 Combined standard uncertainty u C (K ) 3.6 x 10 8

PAGE 45

45 APPENDIX B RESOLVED SHEAR STRES S CALCULATIONS Figure B 1. The notations used in the resolved shear stress calculations. The F f vector is held constant and is swept along 360 degrees for each of the four active slip planes. Vector equations slip plane normal : slip direction :

PAGE 46

46 friction force resultant R: alpha : lambda : resolved shear stress :

PAGE 47

47 LIST OF REFERENCES 1. Bhushan, B., Contact mechanics of rough surfaces in tribology: multiple asperity contact. Tribology Letters, 1998. 4 (1): p. 1 35. 2. Bowden, F.P. and D. Tabor, Friction, lubrication and wear: a survey of work during the last decade. British Journal of Applied Physics, 1966. 17 (12): p. 1521. 3. Schmitz, T.L., et al., The Difficulty of Measuring Low Friction: Uncertainty Analysis for Friction Co efficient Measurements. Journal of Tribology, 2005. 127 (3): p. 673 678. 4. Krick, B. and W. Sawyer, A Little Analysis of Errors in Friction for Small Wear Tracks. Tribology Letters, 2010. 39 (2): p. 221 222. 5. Archard, J.F. and W. Hirst, The Wear of Metals under Unlubricated Conditions. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 1956. 236 (1206): p. 397 410. 6. Schmitz, T.L., et al., Wear Rate Uncertainty Analysis. Journal of Tribology, 2004. 126 (4): p. 8 02 808. 7. Guicciardi, S., et al., On data dispersion in pin on disk wear tests. Wear, 2002. 252 (11 12): p. 1001 1006. 8. Colbert, R., et al., Uncertainty in Pin on Disk Wear Volume Measurements Using Surface Scanning Techniques. Tribology Letters, 2011. 42 (1): p. 129 131. 9. Amateau, M.F. and J.W. Spretnak, Plastic Deformation in Magnesium Oxide Crystals Subjected to Rolling‐Contact Stresses. Journal of Applied Physics, 1963. 34 (8): p. 2340 2345. 10. Stokes, R.J. and C.H Li, Dislocation configurations and the initiation of yielding in magnesium oxide. Discussions of the Faraday Society, 1964. 38 : p. 233 242. 11. Stokes, R.J., Microstructure and Mechanical Properties of Ceramics 1963, Office of Naval Research: Hopkins, Minnesota. 12. Steijn, R.P., Sliding and Wear in Ionic Crystals. Journal of Applied Physics, 1963. 34 (2): p. 419 &. 13. Chang, R., Dislocation Relaxation Phenomena in Oxide Crystals. Journal of Applied Physics, 1961. 32 (6): p. 1127 1132.

PAGE 48

48 14. Washburn, J. et al., Electron microscope observations of deformed magnesium oxide. Philosophical Magazine, 1960. 5 (58): p. 991 999. 15. Sugita, T., K. Suzuki, and S. Kinoshita, Wear characteristics of magnesium oxide single crystals on steel. Wear, 1977. 45 (1): p. 57 73. 16. Dufrane, K.F. and W.A. Glaeser, Rolling contact deformation of MgO single crystals. Wear, 1976. 37 (1): p. 21 32. 17. Bowden, F.P. and A.E. Hanwell, The Friction of Clean Crystal Surfaces. Proceedings of the Royal Society of London. Series A. M athematical and Physical Sciences, 1966. 295 (1442): p. 233 243. 18. Bowden, F.P. and C.A. Brookes, Frictional Anisotropy in Nonmetallic Crystals. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 1966. 295 (1442): p. 244 258. 19. W. G. Sawyer, S.S.P., M. A. Hamilton, N. Arigbay, L. A ALvarrez, N. Mauntler, A. A Voevodin. Wear Testing and In Situ Profilometry in World Tribology Congress 2009. 20. Argibay, N., J.A. Bares, and W.G. Sawyer, Asymmetric wear behavior of self mated copper fiber brush and slip ring sliding electrical contacts in a humid carbon dioxide environment. Wear, 2010. 268 (3 4): p. 455 463. 21. K. Wahl, W.G.S., Observing Interfacial Sliding Processes in Solid Solid Contacts. MRS Bulletin, 2008. 22. Wyant, J.C. White light interferometry 2002. SPIE. 23. Kurganskaya, I. The Application of VSI(Vertical Scanning Interferometry) to the Study of Crystal Surface Processes 2009 07/13/2009 [cited 2012 01/31/2012]; Available from: http://cnx.org/content/m22326/latest/ 24. Keith, J.H., Design of a Pin On Disk Tribometer with In Situ Optical Profilometry U.o. Florida, Editor 2010: Gainesville. 25. Krick, B. and W. Sawyer, Space Tribometers: Design for Exposed Experiments on Orbit. Tribology Letters, 2011. 41 (1): p. 303 311. 26. N. Mauntler, T.S., M. Hamilton, J. Steffens, N. Argibay, J. Bares, W.G Saywer, Tribometers for In Situ Profilometry 2007. 27. Xu, L., H. Bluhm, and M. Salmeron, An AFM stu dy of the tribological properties of NaCl (100) surfaces under moist air. Surface Science, 1998. 407 (1 3): p. 251 255.

PAGE 49

49 28. E. W Bucholz, C.S.K., K. R. Marchman, W.G Sawyer, S. R Phillpot, S. B Sinnott, K. Rajan, Data driven Model for Estimation of Friction Coefficient via Informatics Methods. Tribology Letters, 2012. 29. Erdemir, A., A crystal chemical approach to lubrication by solid oxides. Tribology Letters, 2000. 8 (2): p. 97 102.

PAGE 50

50 BIOGRAPHICAL SKETCH Kellon Marchman was born in 1988 in Fort Walt on Beach, Florida. He grew up in Niceville, Florida where he attended Niceville High School. He captained the NHS varsity soccer team his senior year in 2006. He continued his education at the University of Florida starting in the summer of 2006 where he w as initially enrolled in nuclear engineering. After determining his passion, he switched to mechanical engineering where he remained and finished his undergraduate coursework in less than 4 years. He started undergraduate research in the University of Flor ida Tribology Laboratory under Professor W. Gregory Sawyer in the summer of 2009. He graduated with honors in the summer of 2010 with his Bachelor of Science degree in mechanical engineering. He remained in the Tribology Laboratory for his graduate studie s where he attended and presented at numerous Society of Tribologists and Lubrication Engineers (STLE) conferences. His main graduate work was in the study of wear of mineralogical and ionic solids, however, he helped with numerous other studies. He graduated in the spring of 2012 with his Master of Science degree in mechanical engineering.