Citation
Material Removal Characterization of Ceramic Materials

Material Information

Title:
Material Removal Characterization of Ceramic Materials
Creator:
Long, Julian Thomas
Publication Date:
Language:
English

Notes

Abstract:
Ceramic materials such as fused silica and yttrium aluminum garnet (YAG) are used in lasing systems as lenses, mirrors and host materials. Using Magnetic Field Assisted Finishing (MAF) to reduce the surface roughness of these materials has been shown to be an effective method of preparing them for optical applications. However, MAF falters when finishing past subnanometer roughness levels and when attempting to remove large amounts of material. To increase the effectiveness of MAF, the material removal mechanisms and their interactions with the material must be investigated. The material removal rate was investigated through various methods until optical profilometry was established as an effective method to measure material removal. Using the measured material removal rate allowed for a better understanding of the impact MAF was having on optical properties, such as laser induced damage threshold (LIDT). The surface characteristics of polycrystalline YAG were investigated through Vickers micro-hardness testing. By indenting the surface and viewing where indents sat in the grain structure, these values could be used to compare the strength of the material within the grains and at the grain boundaries. ( en )
General Note:
Awarded Bachelor of Science in Materials Science and Engineering, magna cum laude, on May 8, 2018. Major: Materials Science and Engineering. Emphasis/Concentration: Metals
General Note:
College or School: College of Engineering
General Note:
Advisor: Hitomi Greenslet. Advisor Department or School: Mechanical & Aerospace Engineering

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Julian Thomas Long. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

UFDC Membership

Aggregations:
UF Undergraduate Honors Theses

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

1 MATERIAL REMOVAL CHARACTERIZATION OF CERAMIC MATERIALS By JULIAN LONG AN HONORS THESIS PRESENTED TO THE COLLEGE OF ENGINEERING OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE ACADEMIC DISTINCTION OF MAGNA CUM LAUDE UNIVERSITY OF FLORIDA 2018

PAGE 2

2 ACKNOWLEDGMENTS First I would like to thank my research advisor Dr. Hitomi Yamaguchi Greenslet for her support and belief in my ability to be successful in a research environment. Her encouragement and consistently meaningful advice has inspired me to seek further growth as a researcher and student by pursuing further education. Additionally, I would like to thank my honors thesis committee members, Dr. Nancy Ruzycki and Dr. Simon Phillpot, for the time they spent working through this process with me. The ir support and access i bility during this time was greatly appreciated. I would also like to thank my colleagues former and present, in the Nontraditional Manufacturing Laboratory (NTML) for the support and help they provided in the laboratory I would also like to thank them for their friendshi p as both in and out of the laboratory the atmosphere was always friendly and never dull. I would like to thank the following NTML members for their support : Daniel Ross Max Stein, Pei Ying Wu Eva Hinkeldey Mingshuo Li, Richard Carl Barrington III, R azvan Vesa, Sterling Miller Bryce Alsten, Marcus Spanolios Mitchell Parker, Mikayla Lamb and Jason Ratay. I want to thank my parents, siblings, and extended family for the support and encouragement they have given me throughout my undergraduate career. I want to thank my friends outside the laboratory as well for their support during our time spent together as students This material is based upon work supported by the Air Force Office of Scientific Research (AFOSR) under Award No. FA 9550 14 1 0270. I would also like to express my thanks to Dr. Akio Ikesue and Dr. Tomosumi Kamimura for providing workpieces for this research.

PAGE 3

3 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 2 LI ST OF TABLES ................................ ................................ ................................ ........................... 4 LIST OF FIGURES ................................ ................................ ................................ ......................... 5 LIST OF ABBREVIATIONS ................................ ................................ ................................ .......... 7 ABSTRACT ................................ ................................ ................................ ................................ ..... 8 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................... 9 1.1 Fused Silica ................................ ................................ ................................ ......................... 9 1.2 Poly crystalline Yttrium Aluminum Garnet (YAG) ................................ .......................... 11 1.3 Magnetic Field Assisted Finishing (MAF) ................................ ................................ ....... 14 1.4 Research Objectives ................................ ................................ ................................ .......... 17 2 MATERIAL REMOVAL CHARACTERIZATION OF FUSED SILICA ............................ 19 2 .1 Scratch Testing ................................ ................................ ................................ ................. 19 2 .2 Diamond Stylus Profilometry ................................ ................................ ........................... 27 2 .3 Optical Profilometry ................................ ................................ ................................ ......... 29 3 DEVELOPMENT O F MATERIAL REMOVAL CHARACTERIZATION METHOD FOR POLYCRYSTALLINE CERAMICS ................................ ................................ ............ 33 3.1 Vickers Hardness Testing ................................ ................................ ................................ 33 3 .2 Indentation Load Selection ................................ ................................ ............................... 35 3.3 Indent Analysis Method ................................ ................................ ................................ .... 37 3.4 Indent Buildup Removal ................................ ................................ ................................ ... 39 4 MATERIAL REMOVAL CHARACTERIZATION OF NEODYMIUM DOPED YAG ..... 42 4.1 Sample Preparation ................................ ................................ ................................ ........... 42 4.2 Indent Analysis Results ................................ ................................ ................................ .... 44 5 CONCLUSIONS AND FUTURE WORK ................................ ................................ ............. 48 LIST OF REFERENCES ................................ ................................ ................................ ............... 50

PAGE 4

4 LIST OF TABLES Table page 2 1 Scratch c reation c onditions ................................ ................................ ................................ 20 2 2 Pre p olish s cratch m easurements ................................ ................................ ....................... 23 2 3 Scratch t esting e xperimental c onditions ................................ ................................ ............ 23 2 4 Scratch p olishing c onditions ................................ ................................ .............................. 23 2 5 Post p olish s cratch m easurements ................................ ................................ ..................... 25 2 6 Experimental c onditions for m agnet c omparison ................................ .............................. 29 2 7 Polishing c onditions for m agnet c omparison ................................ ................................ ..... 30 2 8 Experimental c onditions for m aterial r emoval t esting ................................ ....................... 31 2 9 Polishing c onditions for m aterial r emoval t esting ................................ ............................. 3 2 3 1 Comparison of i ndentation l oads, m easured h ardness v alues, and i ndent s izes ................ 35 3 2 Experimental c onditions for b uildup r emoval t esting ................................ ........................ 39 3 3 Polishing c onditions for b uildup r emoval t esting ................................ .............................. 40 4 1 Experimental c onditions for m aterial r emoval c haracterization ................................ ........ 43 4 2 Polishing c onditions for m aterial r emoval c haracterization ................................ .............. 43 4 3 Comparison of V ickers h ardness v alues when d amage is o bserved ................................ .. 46

PAGE 5

5 LIST OF FIGURES Figure page 1 1 Illustration of the damage layers left by mechanical polishing techniques on fused silica ................................ ................................ ................................ ................................ .. 10 1 2 (a) Polarized image, (b) shlieren (striae) image, and (c) X ray topography image of Nd:YAG single crystals grown by the Czochralsk i method showing the core that is formed ................................ ................................ ................................ ............................... 11 1 3 Computer generated YAG unit cell with (111) plane highlighted red .............................. 13 1 4 MAF experimental equipment ................................ ................................ .......................... 14 1 5 Schematic of MAF processing principle. ................................ ................................ ........... 15 2 1 Linear slide scratching setup ................................ ................................ ............................. 19 2 2 Morphologies of each scratch condition, before polishing ................................ ............... 21 2 3 Measurement of linear scratches ................................ ................................ ....................... 22 2 4 Example of a profile used to measure depth and width of scratches ................................ 22 2 5 Morphologies of each scratch condition, after polishing ................................ .................. 24 2 6 Diamond styl us profile of polished area, including scratch 3 ................................ ........... 25 2 7 SEM image of fiber polishing pad ................................ ................................ .................... 26 2 8 Surface profile of a polished microscope slide, beginning in the unpolished region, and ending in the center of the polished region ................................ ................................ 27 2 9 Surface profile of microscope slide before (left) and after polishing (right) under condition 1 ................................ ................................ ................................ ........................ 30 2 10 Surface profile of microscope sli de before (left) and after polishing (right) under condition 2 ................................ ................................ ................................ ........................ 31 2 11 Surface profile of fused silica before (left) and after polishing (right) ............................. 32 3 1 Shape and dimensions of the diamond indenter used in Vickers hardness testing ........... 33 3 2 Buehler Micromet II Vickers hardness tester ................................ ................................ ... 34 3 3 Change in Vickers hardness with indentation load ................................ ........................... 35

PAGE 6

6 3 4 Example of an indent measured via white light interferometry ................................ ........ 37 3 5 Change in indent buildup and indent depth with polishing stages ................................ .... 40 3 6 Example of an unpolished indent (left) versus a fully polished indent (right) .................. 41 4 1 Example of an indent on an as polished surface ................................ ............................... 42 4 2 Example of an indent before (left) and after polishing (right), showing no signs of damage ................................ ................................ ................................ .............................. 44 4 3 Pre (left) and post polish indent (right) showing signs of cracks that were formed during the indentation process ................................ ................................ .......................... 45 4 4 Pre (left) and post polish indent (right) showing signs of grain dislodgement due to damage caused by the indentation process ................................ ................................ ....... 46

PAGE 7

7 LIST OF ABBREVIATIONS YAG Yttrium Aluminum Garnet MAF Magnetic Field Assisted Finishing Nd Fe B Neodymium Iron Boron Wt.% Weight Percent LIDT Laser Induced Damage Threshold CMP Chemical Mechanical Polishing

PAGE 8

8 A bstract of H onors T hesis P resented to the C ollege of E ngineering of the U niversity of Florida in P artial Fulfillment of the Requirements for the Academic Distinction of Magna Cum Laude MATERIAL REMOVAL CHARACTERIZATION OF CERAMIC MATERIALS By JULIAN LONG May 2018 Chair: Hitomi Greenslet Cochair: Major: Materials Science & Engineering Ceramic materials such as fused silica and yttrium aluminum garnet (YAG) are used in lasing systems as lenses, mirrors and host materials. Using Magnetic Field Assisted Finishing (MAF) to reduce the surface roughness of these materials has been shown to be an effective method of preparing them for optical applications However, MAF falters when finishing past sub nanometer roughness levels and when attempting to remove large amounts of material To increase the effectiveness of MAF, the material removal mechanisms and their interactions with the material must be investigated. T he material rem oval rate was investigated through various methods until optical profilometry was established as an effective method to measure material removal. Using the measured material removal rate allowed for a better understanding of the impact MAF was having on o ptical properties such as laser induced damage threshold (LIDT) The surface characteristics of polycrystalline YAG were investigated through Vickers micro hardness testing By indenting the surface and viewing where indents sat in the grain structure, these values could be used to compare the strength of the material within the grains and at the grain boundaries

PAGE 9

9 1 INTRODUCTION 1.1 Fused Silica Fused silica, SiO 2 is a common glass used for optical applications. With its high transparency and commonly available raw materials it makes for a cheap and effective material for lenses and mirrors. Of particular importance to this project is the use of the material in the lenses and mirrors of laser systems. In these systems, the silica must withstand large amounts of energy passing through the bulk of the material [1,2] The amount of energy that a material can withstand under these conditions before becoming damaged is known as the Laser Induced Damage Threshold (LI DT). Silica is popular for applications in laser systems because of its high band gap and intrinsic LIDT [ 3 5 ] However, t he LIDT of a lens can be affected by multiple factors, including surface irr egularities contaminants and subsurface damage [3 5] Surface irregularities and subsurface cracking act as light scattering points while different contaminants resulting from polishing re deposition, can absorb light and cause enough heat at the surface to cause temperature damage [3 5] Thus, these factors must be minimized to increase the LIDT, and the lifetime, of a component in a laser system. During the manufacturing process, the fused silica components are typically polished to a sub nanometer surface roughness using a combination of chemical and mechanical polishing methods. The last step in this polishing regime is a chemical mechanical polish (CMP) using ceria CeO 2, abrasive [3 5] This process accomplishes a low surface roughness but it leaves behind ceria particles embedded in the surface and causes subsurface cracking as seen in F igure 1 1 [3 5]

PAGE 10

10 Figure 1 1 : Illustration of the damage layers left by mechanical polishing techniques on fused silica layer thicknesses taken from [5] The addition of contaminants during the original polishing stage is due to the use of ceria as an abrasive. Ceria has a band gap E G of 3.1 eV, resulting in light absorption at the wavelengths of many Nd:YAG based lasers, one of the most common type s of laser [1,2] This makes it undesirable to have ceria remaining on the surface as t he ceria particles will reduce transmittance and cause a concentration of energy that can be harmful to the silica resulting in a lower LIDT. In order to remove these defects from the material while maintaining a low surface roughness, a different polishing technique must be used. Alternate polishing techniques, such as Magnetic Field Assisted Finishing (MAF) are being investigated as replacements. The principles behind MAF are discussed in a later section, but one of its main benefits for this application is the use of diamond abrasives which has an E G of 5.5 eV, resulting in l i ttle interaction between deposited diamond and laser light Thus, it is important to investigate the other effects of MAF on fused silica to determine its usefulness as a replacement in polishing optical components. Contamination Layer, 2 0 100 nm Subsurface Damage Layer, 100 Substrate

PAGE 11

11 1.2 Yttrium Aluminum Garnet (YAG) Y AG Y 3 Al 5 O 12 is commonly used as a solid state laser gain medium The material is doped with an active rare earth element, such as neodymium (Nd). Due to its use as a lasing host material, it must have very good optical properties to increase the efficiency of the laser. The single cryst al form of YAG was initially used due to the ease of maintaining the materials optical properties during processing. However, the Czochralski method used to produce the single crystal YAG limits the size, shape, and doping levels of the product [6] Additionally, due to a core that exists within the crystal created by the Czochralski method, as shown in F igure 1 2 only 25 % of the product is useable, resulting in a considerable amount of waste [6] In order to overcome the limitations of the Czoch ralski method, polycrystalline YAG ceramics are being researched for their potential as a replacement [6 8] Figure 1 2 : (a) Polarized image, (b) shlieren (striae) image, and (c) X ray topography image of Nd:YAG single crystals grown by the Czochralski m ethod showing the core that is formed [ 6 ]

PAGE 12

12 Pure samples of polycrystalline YAG have been shown to have almost identical optical transmission properties to single crystal samples and they were capable of being doped to almost four times the atomic percentage that is possible in single crystal YAG [6 8] Additionally, the processing time for th e polycrystalline YAG is only around 30 hrs, compared to the mont h long Czochralski method for growing single crystals [6 8] However, the presence of scatte ring centers, such as grain boundaries, pores, inclusions, secondary phases, and surface roughness, decreases the optical transmission in many of the samples and limits the efficiency of processing polycrystalline YAG [6 8] Adjustment of the processing t echniques for this material such as increased sintering times and the use of more advanced techniques such as hot isostatic pressing have limited the effects of internal scattering centers on the optical transmission of the polycrystalline material [6 8] One of the problem s remaining in the use of polycrystalline YAG is the surface roughness of the final sample [6] After processing the polycrystalline YAG ceramics have a high surface roughness that drastically reduces optical transmission due to light scattering at the surface Additionally, as in fused silica, the surface roughness of YAG ceramics can decrease the LIDT and the lifetime of the laser gain medium [9] The surface roughness must be reduced through polishing, but polishing a polycry stalline piece is much harder than polishing a single crystal. This is due to the grain structure of a polycrystalline ceramic.

PAGE 13

13 Figure 1 3 : Computer generated YAG unit cell with (111) plane highlighted red Shown for YAG in F igure 1 3, t he structure of a single crystal follows the unit cell, the building block of the material. Based on bond density the structure will have planes and directions where bond strength is weaker than others and some where bond strength is stronger than others It is possible to consistently polish single crystal YAG pieces because the orientation, and the bond strength, will not change across the surface. However, in polycrystalline YAG ceramics each grain has a different orientation This is enough to slightly change the amount of force that is required to remove material. In addition, the grain boundaries act as defects, where the amount of force required to remove material is significantly reduced due to dangling bonds. Thus the requirements for uniformly polishing polycrystalline YAG change drastically across the surface U nderstanding how the properties changes between grains and across the surface is important for developing polishing techniques t hat are capable of bridging the gap between single crystal and polycrystalline YAG. (Blue) Aluminum (Red) Oxygen (Grey) Yttrium

PAGE 14

14 1.3 Magnetic Field Assisted Finishing (MAF) The primary feature of MAF is the use of magnetic force to apply the abrasive and promote the motion of the tool Using two rare earth magnets one as the driving, or table magnet and the other as the tool magnet. To complete the polishing tool, ferrous particles are used to form a magnetic brush that aligns with the field between the two magnets though polishing pads can als o be attached to the tool magnet for specific applications. The total force of the tool magnet against the workpiece is defined by E quation 1 1. Where V is the volume of the magnetic tool or the ferrous particles, Figure 1 4 : MAF experimental equipment 100 mm Motor Sample Holder Table Magnet Clearance between Table and Tool Magnet

PAGE 15

15 Tool motion is supplied via the application of rotation to the table magnet by an electric motor. In addition to the rotation around the axis of the table magnet, the tool magnet also rotates around its own axis as seen in F igure 1 4 This is due to the differential velocity between the outside and inside edges of the tool magnet. This is referred to as self displacement of the jig, and it has the effect of increasing material removal by increasing the movement of abrasive cutting edges [10] Figure 1 5 : Schematic of MAF processing principle Another crucial characteristic of MAF arises from the combination of the applied magnetic force, the tool motion, and the friction of the abrasive. As the tool polishes and abrasive is worn down or loses cutting edges, the abrasive is cycled out and repla ced with fresh abrasive. This is caused by the loss of friction that occurs when the abrasive loses its cutting edges, at which point the rotating motion of the tool pushes the smooth abrasive out of the iron brush, while fresh Table Magnet (Below)

PAGE 16

16 abrasive is pulled in to th e space that was left [10] This also results in the constant need for the application of fresh abrasive to replenish the amount that is flushed out. The combination of unique polishing characteristics has made MAF particularly successful in polishing ceramics. Sub nanometer surface roughness values have been achieved on a variety of workpieces including amorphous and polycrystalline ceramics. Polycrystalline ceramics are notoriously difficult to polish a t low surface roughness values However, the unique abrasive motion of MAF is thought to limit the effect of grain orientations on material removal allowing for surface roughness Sa, values below 0.5 nm Further research is required to increase understanding of the way in which material is removed from polycrystalline materials

PAGE 17

17 1.4 Research Objectives Due to its use of magnetic tools and free form brushes, MAF is highly versatile in its applications and polishing characteristics. By varying the myriad of parameters that are controlled during MAF, the surface roughness and planarization of a sample can be controlled. However, because of the complexity of its controlling parameters the mechanisms by which it interacts with the workpiece material have not been fully studied. In order to gain further understanding of the MAF process, it is necessary to t est the polishing characteristics on the materials it has been applied to The way in which the abrasive interacts with the material in both polycrystalline and amorphous materials is important to study to understand how MAF achieves fine polishing. The application of MAF to fused silica has some unique requirements based on the condition of the samples. Due to the contamination and subsurface damage layers decreasing the LIDT of the material these layers need to be removed or reduced to increase th e effectiveness of the material in optical applications MAF is capable of polishing without introducing further contamination due to its use of diamond abrasive but the material removal rate, i.e. the amount of time necessary to remove the contaminatio n that is already present is unknown. Thus it is necessary to determine the depth of material remov al during MAF on fused silica. However, no method for doing so was available so one had to be developed using the equipment available. Therefore, the fi rst objective is to establish a procedure to measure material removal of the MAF process on fused silica. The second goal is to determine the material removal on fused silica and use that value to determine the effect of MAF on the redeposited contamination and subsurface damage layers In the application of MAF to polycrystalline YAG, sub nanometer surface roughness has been achieved. However, in conditions using abrasive sizes below 0.25 m diameter, the polishing

PAGE 18

18 process began to cause grai n dislodgement, or the complete removal of a grain of the material. This was likely due to the action of the smaller abrasive at the grain boundar y but could also be due to the difference in orientation between grains. Thus the relationship between the grain structure and the physical properties at the surface had to be investigated to allow for optimization of the MAF process. One method for comparing physical properties at the surface is hardness testing. Thus, the third objective is to develop a process through which the hardness of individual grains could be measured. The fourth objective is to determine the distribution of grain hardness and the effect of grain boundaries on hardness measurements These values w ill be used to imp rove the understanding of how MAF interacts with a polycrystalline material by treating the abrasive action as comparable to that of the hardness indents.

PAGE 19

19 2 MATERIAL REMOVAL CHARACTERIZATION OF FUSED SILICA 2.1 Scratch Testing In order to determine the material removal rate of MAF on fused silica, the change in the sample weight before and after polishing is generally measured to determine the weight of material removed. However, a finer understanding of both how much, and wher e material was being removed was required for understanding the effect of MAF on LIDT measurements Thus development of a method was desired that could determine how much material was removed at different sites in the polishing zone. In this set of experiments controlled defect creation was used to create reference points. A diamond scribe was used to create scratches deep enough that the bottoms of the scratches would not be affected by the polishing process Assuming that the depth was enough to prevent abrasive action within the scratches, the depth of the scratch could be measured before and after polishing to determine the depth of the material removed by MAF. Figure 2 1 : Linear slide scratching setup Stylus Stage 100 mm Linear Slide Sample Placement Scribing Arm

PAGE 20

20 The linear slide apparatus shown in Figure 2 1 was used to create the scratches required for th e polishing experiments. Using this setup, scratches could be created that were reasonably consistent in width and depth During the scratch creation process, a 75 g weight would be attached to the scribing arm to the force for the scribe to act upon the surface The scribe wo uld then be drawn across the surface of the sample by the motion of the linear slide resulting in scratches that were similar in depth. Initial testing focused on scratch creation tested some conditions such as the effect of multiple scribe passes through the same scratch These also revealed a defect in the scribing tool that created a second scratch in close proximity to the primary scratch. This was negated by at which point the second scratch was not seen in further te sts. The scratch conditions are shown in T able 2 1, while the resulting scratch morphologies are shown in figure 2 2. Table 2 1 : Scratch c reation c onditions Scratch Run Orientation of Scribe 1 1 0 2 2 0 3 1 90

PAGE 21

21 Figure 2 2 : Morphologies of each scratch condition, before polishing White light interferometry was used to measure scratches before and after polishing The area measured by this technique is 177132 m and each scratch was measured at ten positions within the polishing area At each measurement position, ten profile lines were used to determine the average depth and width of the scratch as shown in F igure s 2 3 and 2 4. 1 2 3

PAGE 22

22 Figure 2 3 : Measurement of linear scratches Figure 2 4 : Example of a profile used to measure depth and width of scratches Using these measurement criteria, initial mea surements of the three scratches were taken Compared to the first scratch, r unning the scribe through the s econd scratch multiple times only resulted in a widening of the scratch t hough some reduction in the depth variation was also observed The third scratch showed no secondary scratch and proved to be much thinner than the previous scratches due to the lac k of that feature.

PAGE 23

23 Table 2 2 : Pre p olish s cratch m easurements Scratch 1 9.7 1.7 55 7 2 8.2 1.5 70 8 3 9.0 1.7 40 7 Following the creation of the scratches that would be used as reference points, the sample was polished using an MAF based technique following the parameters outlined in T ables 2 2 and 2 3. Table 2 3 : Scratch Testing Experimental Conditions Sample Fused Silica : 40 6 mm Tool Magnet Nd Fe B 22.2 3.18 mm 0.181 T Surface Field Tool Magnet Rotation Speed 400 min 1 Rotation Speed Clearance between Table and Tool Magnet (Fig. 1 4 ) 9 .60 mm Clearance between Table and Tool Magnet Table 2 4 : Scratch p olishing c onditions Stage Medium Abrasive Time per stage 4.5 mm Line 1 Line 2 Line 3

PAGE 24

24 Following the polishing process, the sample was measured again Representative scratch morphologies after polishing are shown in F igure 2 5 The edges of the scratches had been rounded significantly in some positions, where the meeting of the rounding and the flat surface was considered the edge of the scratch for measurement purposes. Figure 2 5 : Morpholog ies of each scratch condition, a fter polishing The depth and width of the scratches post polish are shown in T able 2 5 However, d ue to rounding observed at the bottom of the scratches, it was presumed that the polishing proccess had acted on the interior of the scratches Thus these measurements could not be used to determine an accurate value for the material removal rate Table 2 5 : Post p olish s cratch m easurement 1 Fiber Polishing Pad 0 0.1 D iameter D iamond S lurry 0.1 mL every 2 min 60 min 1 2 3

PAGE 25

25 Scratch 1 7.1 1.5 59 8 2 6.4 1.5 71 9 3 7.1 1.7 45 6 In addition to the observed rounding of the bottom of the scratches, diamond stylus profilometry of the scratches post polish showed that there were variations in the surface in the immediate vicinity of the scratches. The entire depth of the scratch was not measured due to the diamond stylus being wider than the scratch, meaning it could not fit in side the scratch. Shown in figure 2 6 the entire measured area has been polished, but the area around the scratch shows increased material removal This is presumed to have been caused by the introduction of surface damage during the scratching process resulting in a weakened material in the area around the scratches. Figure 2 6 : Diamond stylus profile of polished area including S cratch 3 For all of the material removal rate experiments, a fiber polishing pad served as the medium which held the abrasive against the sample surface. An SEM micrograph of the fibers on these polishing pads is shown in F igure 2 7 The fibers have a large degree of freedom in their movement resulting in a less rigid abrasive brush. This freedom of movement was thought to -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10 0.20 20 21 22 23 24 25 Height (m) Surface Position (mm)

PAGE 26

26 have contributed to the polishing of the area inside the scratches and makes measuring mater ia l removal through controlled defects difficult a s there was no available method to make a deeper, thinner defect that might resist abrasive intrusion Figure 2 7 : SEM image of fiber polishing pad

PAGE 27

27 2.2 Diamond Stylus Profilometry Due to the observed characteristics of the fiber polishing pad, further material removal rate testing would focus on step height measurement To do so one part of the sample must be left unpolished so that it can be used as a reference in the measurement of the polished region To achieve this, a smaller magnet must be used in the polishing of the fused silica so that an unpolished area can be left around the polished zone. However, because MAF relies heavily on magnetic force, changing the size of the polishing magnet will change the polishing characteristics. In order to relate the large magnet to the small magnet, an experiment comparing the two had to be run first The details of this experiment will be discussed in S ection 2.3 Before the magnets could be compared, a method was needed that could measure over a wide area to cover both the unpolished and polished regions The first method to be tested was diamond stylus contact profilometry By drawing a diamond stylus attached to a cantilever beam across the surface, the surface profile could be measured To test this, a microscope slide was polished using the conditions in T ables 2 2 and 2 3 and the resulting profile is shown in F igure 2 8 Microscope Slide Polished Area

PAGE 28

28 Figure 2 8 : Surface profile of a polished microscope slide, beginning in the unpolished region, and ending in the center of the polished region The ultimate limitation of this measurement technique is its high error at the value range where measurements are required. In F igure 2 8 the error is as large as the difference between the polished and unpolished areas, making the material removal rate difficult to determine from measurements of this nature. -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 0 5 10 15 20 25 30 35 40 45 50 Height ( m) Surface Position (mm)

PAGE 29

29 2.3 Optical Profilometry For more accurate measurements, white light interferometry was tested as a replacement For these measurements, the zoom was reduced to the minimum for the highest measurement area, and a stitching function was used to measure over a large area by combining individual measurements How ever, even using this method, the amount of removed material was difficult to discern due to the inherent curvature of the microscope slide samples. In order to increase the amount of material removal and make it easier to measure the only parameter that could change was the polishing time. Thus, a series of experiments were run with extended polishing times. The first two experiments focused on comparing the different magnet setups by comparing their effects on identical microscope slides. The third experiment would use Condition 2 on the lens quality fused silica sample and utilize the results from the previous experiments to relate the material removal of the smaller magnet to that o f the standard magnet. The conditions for these experiments are outlined in T ables 2 6, 2 7, 2 8, and 2 9. Each sample was measured before and after the extended polish and a profile line was used to identify the areas of highest material removal.

PAGE 30

30 Table 2 6 : Experimental c onditions f or m agnet c omparison Condition 1 2 Sample Silica Microscope Slide Tool Magnet Nd Fe B 22.7 3.18 mm 0.181 T esla Nd Fe B 12.7 3 .18 mm 0.295 T Nd Fe B 12.7 3 .18 mm 0.331 T Rotation Speed 400 min 1 Clearance between Table and Tool Magnet (see Fig. 1 4 ) 9 .60 mm Table 2 7 : Polishing c onditions for m agnet c omparison Stage Medium Abrasive Time per stage 1 Fiber Polishing Pad 0 0. 1 D iameter D iamond S lurry 0.1 mL every 4 min 360 min Figure 2 9 : Surface profile of microscope slide before (left) and after polishing (right) under C ondition 1 1 mm 77 mm Y X 50 mm O T

PAGE 31

31 After polishing the microscope slide with C ondition 1, it exhibited the surface profile shown on the right of F igure 2 9 The center of the polishing zone, where the magnetic field is the strongest, had the highest material removal and was used to calculate the maximum material removal rate. This process was repeated with a second microscope slide polished using C ondition 2, as shown in F igure 2 1 0 Figure 2 1 0 : Surface profile of microscope slide before (left) and after polishing (right) under C ondition 2 Based on the surface profiles before and after polishing, the material removal during the polish was determined to be 0.8 m and 0.5 m for C onditions 1 and 2 respectively. These values were used to form a direct numerical comparison between C ondition 1 a nd 2 and it was concluded that C ondition 2 removes 62.5% of the material that C ondition 1 is able to remove. Table 2 8 : E xperimental c onditions for m aterial r emoval t est ing Condition 2

PAGE 32

32 Sample Fused Silica : 40 6 mm Tool Magnet Nd Fe B 12 .7 3 .18 mm 0.295 T Nd Fe B 12 .7 3 .18 mm 0.331 T Rotation Speed 400 min 1 Clearance between Table and Tool Magnet (see Fig. 1 4 ) 9 .60 mm Table 2 9 : Polishing c onditions for m aterial r emoval t esting Stage Medium Abrasive Time per stage 1 Fiber Polishing Pad 0 D iameter D iamond S lurry 0.1 mL every 2 min 360 min With the comparison between the magnets established, the lens grade fused silica was polished using C ondition 2. The fused silica samples had more consistent planarization, due to the commercial polishing received during processing, as shown in F igure 2 12 (left). The resulting polished area was similar to that seen in the previous experiments. Figur e 2 1 1 : Surface profile of fused silica before (left) and after polishing (right)

PAGE 33

33 Based on the profile of the post polish surface, the material removal rate on this sample was determined to be 0.3 m. By using the numerical comparison of the material removal rates of the two conditions, the theoretical material removal of C ondition 1 o n lens grade fused silica was calculated to be 0.48 m. Using this value, the material removal rate of the standard MAF process was calculated to be 1.3 nm/min 3 DEVELOPMENT OF MATERIAL REMOVAL CHARACTERIZATION METHOD FOR POLYCRYSTALLINE CERAMICS 3.1 Vickers Hardness Testing The study of fused silica focused primarily on material removal rate because the surface properties of an amorphous material were assumed to be relatively uniform However, this is not the case with polycrystalline YAG ceramics. T o determine the impact of grain structure on the effectiveness of MAF in finishing a polycrystalline YAG surface, it was necessary to characterize the differences between grains at the surface. To do so, Vickers micro hardness testing was used to measure the hardness of the material at various points on the surface. Vickers hardness testing uses a diamond indenter with a square shape as shown in F igure 3 1

PAGE 34

34 Figure 3 1 : Shape and dimensions of the diamond indenter used in Vickers hardness testing [11] After pressing the indenter against the surface with a set amount of force, the diagonals of the resulting indent could be measured and used to calculate the Vickers hardness value using E quation 3 1 and 3 2 (3 1) (3 2 ) Where d 1 and d 2 are the indent diagonals in mm d is the average indent diagonal, F is the indentation load in kgf, and HV is the Vickers hardness value [11,12] Vickers micro hardness uses small loads to create indents that are measured in microns using a 40 X objective The equipment used to indent samples and measure indents is shown in F igure 3 2.

PAGE 35

35 Figure 3 2 : Buehler Micromet II Vickers hardness tester 3.2 Indentation Load Selection Using Vickers micro hardness indentation to measure hardness, it is possible to measure the hardness of individual grains through the application of a properly sized load. Thus, for the purpose of the primary experiments, in which the measurement of individual grains is one of the primary goals, it was necessary to first determine the indentation load that would result in a properly sized indent Each available indent load was tested for indent size and resulting hardness measurements. These results are summarized in T able 3 1 and F igure 3 3. Table 3 1 : Comparison of i ndentation l oads, m easured h ardness v alues, and i ndent s izes Indent Load (g) Hardness Value, HV Indent Size, d 25 4010 3.4 50 3364 5.3 100 3169 7.7 Sample Holder Dia mond Indenter Objective (10 X ) Objective (40 X ) 50 mm

PAGE 36

36 200 2281 12.8 300 2214 15.9 500 1686 23.5 1000 1682 33.2 Figure 3 3 : Change in Vickers hardness with indentation load Based on F igure 3 3, a trend can be seen between the indent load and the resulting hardness measurements. As the indent load is decreased, the measured hardness value more than doubles. This means that at lower indentation loads, the measurement value is extremel y sensitive to changes in indent size, which has a large impact on error in the resulting measurements. However, due to expected grain sizes between 10 15 m, an indent load less than 100 g is desired. An assumption was made that, since hardness is a relative value, the large values resulting from smaller indentation loads would not be a hindrance to comparison within the same material. Thus, the 50 g indentation load was selected based off of the dimensions of its resulting indent. According to ASTM E384 11 this load size can be used for micro hardness testing, but it requires a well polished surface [11] Although the sample used to test indentation loads had already been 0 500 1000 1500 2000 2500 3000 3500 4000 4500 0 200 400 600 800 1000 Vickers Hardness, HV Indentation Load (g)

PAGE 37

37 polished, this requirement would mean that all further testing would have to wait until the high quality samples had been prepared by MAF. 3.3 Indent Analysis Method One of the drawbacks of the Vickers micro hardness tester was the optical measurement tool. Due to the transparency of the samples, identifying grain structure while making the indents was impossible. Thus, another method would be required for det ermining the position of indents within the grain structure after indentation. To make this possible, i ndents were made in a grid with the goal of placing enough of them within grain s that a viable data set would be achieved. Another measurement method could then be used to identify which of the indent s lay within a grain. The measurement tool that would be used for all indent characterization was a n optical profilometer using white light interferomet ry A CMP using colloidal silica had already been

PAGE 38

38 applied to the test sample of low quality YAG to expo se the grains, so optical profilometry was able to identify them based on the height difference between them. Amongst the grain structure, the indents were highly visible due to a large surface buildup caused by the indentation process, seen in F igure 3 4 Figure 3 4 : Example of an indent measured via white light interferometry The surface buildup on the indents averaged 0.3 m and was primarily skewed towards the top left and top right of the indent. Due to its height relative to the surface, t his feature made viewing the position of the indent in the grain structure difficult A polishing process was designed to prepare the indent for measurement by removing the obscuring surface buildup and clarifying the surrounding grain structure by providing higher relief between grains

PAGE 39

39 3.4 Indent Buildup Removal To use polishing to clarify the positions of indents and the grain structure required a process that could meet two criteria. First, it had to fully remove the surface buildup without significantly removing material in areas around the indent Second, it must provide high relief between grains so that the position of indents within the grain structure can be more easily identified. To accomplish these goals a two step process was decided upon that included a mechanical polishing stage followed by a CMP stage. The mechanical polishing stage uses 0.25 m diameter diamond slurry as an abrasive. By using a fine abrasive, the material removal was aggressive but not enough to significantly alter the surface outside of the indent However, in initial testing, it removed the distinction between grains that had been established by a previous CMP. Thus a second CMP stage was required to follow

PAGE 40

40 the first stage. In the second stage, a colloidal silica suspension was applied to the sample using a polishing pad. The details of both stages are shown in T ables 3 2 and 3 3. Table 3 2 : Experimental c onditions for b uildup r emoval t esting Sample Polycrystalline YAG Ceramic Number of Indents 12 in a 3 4 grid, centered at origin Tool Magnet Nd Fe B 12.7 3.18 mm 0.295 T R otation Speed 400 min 1 Clearance Between Table and Tool Magnet (see Fig. 1 4 ) 1 0.5 mm Table 3 3 : Polishing c onditions f or b uildup r emoval t esting Stage Medium Abrasive Time per stage 1 44 Iron Particles 0 S lurry 0.1 mL every 2 min 2 0 min 2 Iron Particles 3 wt. % Colloidal Silica Suspension 1 mL every 5 min 3 0 min The sample indents were measured before and after the two polishing stages and the resulting impact on the indent buildup and depth is shown in F igure 3 5 As expected, the first stage acted to remove indent buildup, while the second stage appears to have had no impact on the buildup though its results are clear when looking at the surface through optical profilometry seen in F igure 3 6.

PAGE 41

41 Figure 3 5 : Change in indent buildup and indent depth with polishing stages Figure 3 6 : Example of an unpolished indent (left) versus a fully polished indent (right) One of the main concerns with the results shown in F igures 3 5 and 3 6 is that the material removal of the first stage was significant Therefore, t he structure of the indent was removed entirely Additionally, the grain structure was not well revealed For the remaining tests on higher 0.150 0.003 0.003 0.402 0.434 0.449 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0 1 2 Height ( m) Polishing Stage Indent Buildup Indent Depth

PAGE 42

42 quality, Nd: YAG samples, the polishing process would s ee changes in polishing times to mitigate these issues 4 MATERIAL REMOVAL CHARACTERIZATION OF NEODYMIUM DOPED YAG 4.1 Sample Preparation Using the method established in C hapter 3, a sample of 1% Nd:YAG was prepared for testing. After the sample was polished using a pre established MAF process, indents were made in a 77 grid. Each indent was separated from the next one by 1.5 mm resulting in a coverage of 99 mm The samples were then measured by optical profilometry to gain a baseline to which the post polish surface could be compared. An example of an indent made on the polished surface is shown in F igure 4 1.

PAGE 43

43 Figure 4 1 : Example of an indent on an as polished surface The rest of the experimental conditions remained consistent with previous tests done with YAG samples, which are shown in T able 4 1. However, the polishing conditions, shown in table 4 2, were changed to mitigate some of the material removal issues seen i n Chapter 3 The polishing time on S tage 1 was halved to 10 min to reduce its effect on the internal structure of the indents while the polishing time of S tage 2 was doubled to 60 min to increase grain exposure. Table 4 1 : Experimental c onditions for m aterial r emoval c haracterization. Sample Laser Grade Polycrystalline 1 % Nd:YAG Number of Indents 49 in a 7 7 grid Separation 1.5 mm Load 50 g Load Time 10 s Tool Magnet Nd Fe B 12.7 3.18 mm 0.295 T Rotation Speed 400 min 1 5 mm 35 mm Y X 57 mm O

PAGE 44

44 Clearance between Table and Tool Magnet (see Fig. 1 4) 10.5 mm Table 4 2: Polishing c onditions for m aterial r emoval c haracterization Stage Medium Abrasive Time per stage 1 44 0 slurry 0.1 mL every 2 min 10 min 2 3 wt. % Colloidal Silica solution 1 mL every 5 min 60 min 4.2 Indent Analysis Results After polishing, each indent was again measured using optical profil ometer Using these images, the position of indents within the grain structure could be determined. This was done by two steps. The first step was to determine whether the surrounding grains would be large enough to contain an indent. The second was to look at the indent itself and any surface damage that was present. Indents with visible cracking or grain dislodgement, where a grain has been visibly pushed out of its position, were taken to have landed at or near grain boundaries. Only the indents that showed no signs of external damage were within a grain. Some examples of the damage types are seen in F igures 4 2, 4 3, and 4 4.

PAGE 45

45 Figure 4 2 : Example of an indent before (left) and after polishing (right), showing no signs of damage In F igure 4 2, the indent shows no signs of cracking or grain dislodgement and was placed within a n area containing large grains that have boundaries adjacent to the indent. Thus, this indent and those like it were considered to be within grains The hardness values measured from these indents were used to calculate the average hardness value for grains of this material. 25 m

PAGE 46

46 Figure 4 3 : Pre (left) and p ost polish indent (right) showing signs of cracks that were formed during the indentation process An example of an indent with cracks present is shown in F igure 4 3. These cracks were caused during the indentation process when an indent landed along a grain boundary and the stress separated the grains on either side of the boundary. During the polishing process, the cracks are polished as well, resulting in the wide, rounded shapes seen in the bottom image 25 m

PAGE 47

47 Figure 4 4 : Pre (left) and p ost polish indent (right) showing signs of grain dislodgement due to damage caused by the indentation process An example of an indent with extreme grain dislodgements is shown in F igure 4 4. The grain dislodgement is typically a result of the combined actions of the indentation and polishing process. When an indent lands at a grain boundary it has the potential to displace material and force that displaced material under an adjacent grain, resulting in the displacement of that grain. During the polis hing process, that grain is then pulled out by passing abrasives leaving a large pit in its place. In the case of F igure 4 4, grain dislodgement occurred before the polishing stage which was seen in few other indents. Grain dislodgement was seen as ind icative of an indent that landed at the grain boundary of small or shallow grains Table 4 3 : Comparison of V ickers h ardness v alues when d amage is o bserved Damage Type Observed Number of Indents Average Measured Hardness, HV Cracking 24 3615 4 3 0 Grain Dislodgement 28 3538 680 Indent only 6 3655 1000 25 m

PAGE 48

48 In T able 4 3, the average hardness values measured from indents with each type of damage are compared. U ndamaged indents showed higher average hardness value than those that were damaged. This could be because the cracking or grain dislodgement acts as a stress reliever moving nearby material to accommodate a larger indent which is measured as a lower hardness value. However, each average hardness value also has a very high standard deviation. This is the result of two factors ; the first being the di fference in grains with different crystallographic orientations and the second being the sensitivity of the Vickers hardness tester. As discussed in S ection 3.2, Vickers micro hardness indentation with small loads is extremely sensitive to changes in the indent dimensions. With a load of 50 g a difference of 2 m in the d value result s in a difference of approximately 2000 HV resulting in the large distributions of the measur ed values

PAGE 49

49 5 CONCLUSIONS AND FUTURE WORK From the work done in S ection 2.3, the material removal rate was established to be 1.3 nm/min on fused silica. The contamination layer left behind by ceria polishing on the same material is known to be between 20 100 nm [5] Thus, MAF can be expected to remove that level of material in 15 75 min of polishing. In order to guarantee the removal of the contamination layer, future polishing methods should proceed for up to 75 min. Future work on this to pic should include a further investigation of the effect of time polishing on the characteristics of the fiber polishing pad. If the pad is decomposing or otherwise decreasing in effectiveness as time passes, then the material removal rate determined through lengthy polishing times may not hold true for shorter polishing times The investigation of the surface characteristics of polycrystalline 1% Nd: YAG yielded results that showed there is a difference in the hardness of a grain when compared to the grain boundary. However, due to the small number of indents that landed within grains, the average hardness values cannot be used for inter grain comparisons. Still, the measurements showed that there is a decrease in the material hardness in the grain boundary regions. This will cause material removal to be greater at the grain boundaries than within the bulk grain when using more gentle process ing conditions due to intrusion of small abrasive particles into the grain boundaries. Future work on this subject should continue to investigate other dopant levels of Nd:YAG However, the use of Vickers micro hardness testing for this purpose could be replaced by nano indentation. This would allow for more accurate small scale measuremen ts, which could quantify the differences between individual grains Material characterization methods such as x ray diffraction could be used to determine the existence of inclusions or secondary phases that may weaken the material and change the effect o f MAF at localized areas. Additionally, the method

PAGE 50

50 developed in C hapter 2 to determine material removal rate could be applied as a means of comparing different dopant levels of Nd:YAG

PAGE 51

51 LIST OF REFERENCES [ 1] Kuzuu Nobu, et al "Laser Induced Bulk Damage of Various Types of Silica Glasses at 266 Nm." Japanese Journal of Applied Physics 43.10 (2004): 7174 175. Web. [ 2] Kuzuu, Nobu, et al "Laser Induced Bulk Damage of Various Types of Silica Glasses at 532 and 355 Nm." Japanese Journal of Applied Physics 43.5A (2004): 2547 548. Web. [ 3] He, Xiang, et al "Subsurface Defect Characterization and Laser induced Damage Performance of Fused Silica Optics Polished with Colloidal Silica and Ceria." Chinese Physics B 25.4 (2016): 048104. Web. [ 4] Kamimura, Tomosumi, et al "Ion Etching of Fused Silica Glasses for High Power Lasers." Japanese Journal of Applied Physics 37.Part 1, No. 9A (1998): 4840 841. Web. [ 5] Kamimura, Tomosumi, et al "Enhancement of Surface Damage Resistance by Removing Subsurface Damage in Fused Silica and Its Dependence on Wavelength." Japanese Journal of Applied Physics 43.No. 9A/B (2004): n. pag. Web. [ 6] Ikesue, A et al. Progress in Ceramic Lasers Annual Review of Materials Research vol 36, no. 1, 2006, pp. 397 429., doi:10.1146/annurev.matsci.36.011205.152926. [ 7] Sanghera, J. et al., Materials vol. 5, 2012, pp. 258 277., doi:10.3390/ma5020258. [ 8] Optical Materials vol. 19, no. 1, 2002, pp. 183 187., doi:10.1016/s0925 3467(01)00217 8. [ 9] Fu, Yuelong, et al "Influence of Surface Roughness on Laser induced Damage of Nd:YAG Transparent Ceramics." Ceramics International 41.10 (2015): 12535 2542. Web.

PAGE 52

52 [ 10] Yamaguchi, Hitomi, et al "Study of Finishing of Wafers by Magnetic Field Assisted Finishing." Journal of Advanced Mechanical Design, Systems, and Manufacturing 3.1 (2009): 35 46. Web. [ 11] ASTM Standard E384 11 2009 for Knoop and Vickers Hardness of Materials ASTM International, West Conshohocken, PA, 2009 [ 12] Kamm, Janice, and George Vander Voort. "An Introduction to Microindentation Methods." Tech Notes 1.6 (2015): 1 4. Web.