Surface Interactions Between Nanodiamonds and Glass in

Surface Interactions Between Nanodiamonds and Glass in

Magnetorheological Finishing (MRF)

by

Jessica Erin DeGroote

Submitted in Partial Fulfillment

of the

Requirements for the Degree

Doctor of Philosophy

Supervised by

Professor Stephen D. Jacobs

The Institute of Optics

The College

School of Engineering and Applied Sciences

University of Rochester

Rochester, New York

2007

ii

Curriculum Vitae

Jessica Erin DeGroote was born in Newark, NY on June 24, 1980. She

attended The Institute of Optics at the University of Rochester from 1998 to 2002,

and graduated with a Bachelor of Science degree in 2002. She started her graduate

studies at the University of Rochester in the fall of 2002 at The Institute of Optics

obtaining her Master of Science in Optics in 2004. She received the Frank J. Horton

Fellowship from 2003 – 2007. She pursued Ph.D. research in Magnetorheological

finishing (MRF) at the Laboratory for Laser Energetics under the direction of Dr.

Stephen D. Jacobs.

iii

Acknowledgements

My graduate career has been a long journey. Although at times it was very

difficult, I have learned from my experiences and they have been very rewarding.

There have been a great number of people that have touched my life and to them I

extend my deepest appreciation.

First of all, I would like to thank my Ph.D. thesis advisor, Dr. Stephen Jacobs.

I first started working with Dr. Jacobs as an undergraduate and his love of research

inspired me to continue on with graduate school. He has more knowledge about

optical fabrication than anyone I know, and I am truly blessed to have had the

opportunity to learn from him. One could not ask for a more supportive advisor, but

most of all I am thankful for our friendship and I look forward to continuing that for

many more years.

I came to The Institute of Optics as an undergraduate in the fall of 1998. It

gave me the opportunity to interact with many great professors. I would like to

specifically thank Dr. John Lambropoulos, Dr. James Zavislan, Dr. Thomas Brown

and Dr. Paul Funkenbusch. I would also like to extend my appreciation to the staff at

The Institute of Optics for all of their hard work and friendship, specifically Joan

Christian, Brian McIntyre and Per Adamson.

I was very fortunate to have the opportunity to conduct my research at the

Laboratory for Laser Energetics and receive funding through the Horton Fellowship.

I express my appreciation for the total support of my work provided by Harvey

Pollicove (1945-2004), founder and Director of the Center for Optics Manufacturing.

I owe a great deal of appreciation to Henry Romanofsky, Ed Fess, John Schoen,

Theresa Pfuntner and Dr. Irina Kozhinova for all of their help and friendship through

the years. I also want thank Katie Spencer, John Wilson and Amy Bishop for their

hard work on my research during their undergraduate studies at the University of

Rochester. I especially want to thank Anne Marino, I am not sure if anyone will ever

iv

compare to Anne. She is quite possibly the most intelligent and hard working person

I have ever had to opportunity to work with.

I would like to thank Alex Maltsev and Mike Kaplun for all of their work in

the optical shop. In addition to preparing all of my samples, Alex was always good

for a good laugh and a word of encouragement. My appreciation goes out to

Christine Pratt for her help with nanoindentations. I also want to thank LLE

librarians Linda Clement-Rister and Kenn Harper; I will certainly miss all of the

articles “magically” appearing in my mailbox. I would like to extend a thank you to

Jay Keck and Andrew Dillenbeck who were instrumental in preparing the force

sensor for my work. I also want to acknowledge my officemates, especially Dr. Anka

Trajkovska-Petkoska. Her friendship will never be forgotten.

Outside of the University of Rochester community I would like to

acknowledge Dr. Vitaly Slobodsky and Andrew Dominello of UK Abrasives for

supplying me with nanodiamonds for my work. I would also like to acknowledge

Hoya for donating glass samples and Tony Marino of Advanced Glass Industries for

core drilling the same glass samples. I would like to thank Dr. Ian Lee-Bennett of

Taylor-Hobson for writing the Excel macro for analyzing power spectral density to

my specifications and I would also like to thank Dr. Robert James of QED

Technologies for many helpful discussions.

Dr. Oliver Faehnle invited me to study the Fluid Jet Polishing (FJP) process

for three weeks at Fisba Optik in St. Gallen Switzerland and I am very thankful to

have had that opportunity. I learned a lot in my time there, both professionally and

personally. I also want to thank Pim Messelink for his friendship and guidance

around the lab (and Switzerland).

Throughout my academic career I was very fortunate to have excellent

teachers and I would specifically like to thank a few that had huge impacts on my life

during my time at Williamson Central High school. I am forever indebted to Steve

Murphy, Dr. Judy DiClemente and Kate Taylor. I would not be where I am today

without these three individuals.

v

I would like to thank all of my friends who kept me balanced through it all.

Graduate school would not have been the same without the “other Jessica”, thank you

Jessica Morgan for everything. I also want to specifically thank Jennifer Steinberg,

Caitlin Marcellus, Jason Taniguchi, Matt Bolcar and Jon Watson for your continued

friendship and support. Of course I need to say thank you to the fabulous ladies of

The Yarn Hoard, thank you Sarah, Velynda, Natalie, Shannon and Kathleen. “There

is nothing that a little stitching can’t fix.”

Life has a funny way of working itself out. Fate is all I can come up with why

I was so fortunate to be able to have Phil in my life. He is not only the love of my

life, but also my best friend. Thank you, Phil, for all of your love and support.

My whole family is just absolutely fantastic. No matter what, they are always

there for me, and I am just truly blessed. I especially want to thank my brother Doug,

and my parents Dave and Jane. Throughout my entire life I have always been given

love and encouragement, and for that, I am a better person. Finally this section would

not be complete without, “I owe it all to my Dad, he is the greatest!”

Research was sponsored by the U. S. Army Armament, Research, Development and Engineering Center (ARDEC)

and was accomplished under Cooperative Agreement Number W15QKN-06-R-0501 and the U.S. Department of Energy Office of Inertial Confinement Fusion under Cooperative Agreement No. DE-FC52-92SF19460, the University of Rochester, and the New York State Energy Research and Development Authority. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of U.S. Army ARDEC or the U.S Government. The support of DOE does not constitute an endorsement by DOE of the views expressed in this article. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein.

vi

Abstract

Magnetorheological finishing (MRF) is a deterministic sub-aperture polishing

process. The process uses a magnetorheological (MR) fluid that consists of micron-

sized, spherical, magnetic carbonyl iron (CI) particles, non-magnetic polishing

abrasives, water and stabilizers. Material removal occurs when the CI and non-

magnetic polishing abrasives shear material off the surface being polished.

In this work we focus on building a better understanding of the MRF process.

Material removal is possible using a non-magnetic, polishing abrasive-free MR fluid.

We study how this occurs and how removal changes with the addition of non-

magnetic polishing abrasives. Specifically, we study nanodiamonds in the MR fluid.

Nanodiamonds can be engineered to achieve varying properties that allow us to learn

more about the MRF process.

We introduce a new MRF material removal rate model. This model contains

terms for the near-surface mechanical properties of glass, drag force, polishing

abrasive size and concentration, chemical durability of the glass, MR fluid pH and the

glass composition. We validate individual terms in our model separately and then

compare the entire model to an existing MRF material removal model. All of our

experimental data were obtained using nanodiamond MR fluids and a set of six

optical glasses.

We discuss the role of nanodiamonds in the polishing zone by examining the

residual surface texture inside MRF spots. We show evidence that the CI particles in

vii

the abrasive-free MR fluid remove material with a micro-gouging mechanism and, as

nanodiamonds are added to the system, the mechanism changes to a micro-lateral-

fracture mechanism, which is more efficient in removing material. Finally, we

modify a conventional model for surface roughness for use with MRF.

viii

Table of Contents

Curriculum Vitale ii

Acknowledgements iii

Abstract vi

Table of Contents viii

List of Tables xiii

List of Figures xvi

List of Symbols xxv

Chapter 1. Introduction 1

1.1. MRF Process 1

1.2. Conventional material removal and surface smoothing 3

models for grinding and polishing

1.3. CMP material removal and surface smoothing models 9

1.4. MRF material removal models 13

1.5. Overview of thesis 16

References 17

Chapter 2. Experimental approach 21

2.1. Spot Taking Machine (STM) 21

2.2. Optical glass substrates 23

2.3. Metrology 25

2.4. Fluid analysis 28

References 29

Chapter 3. Magnetorheological (MR) fluid 32

3.1. Introduction 32

3.2. Carbonyl Iron 33

ix

3.2.1. CI particle size and zeta potential 34

3.3. Nanodiamonds 38

3.3.1. Nanodiamond friability 39

3.3.2. Nanodiamond zeta potential and particle size 41

3.3.2.1. Dry nanodiamond powder (NDP) 42

3.3.2.2. UK Abrasive nanodiamonds 45

3.3.2.3. Nanodiamond properties in MR carrier 56

fluid environments

3.4. Magnetorheological (MR) fluid behavior in polishing 58

3.4.1. Abrasive free MR fluid 59

3.4.2. Nanodiamond effectiveness and friability 61

3.4.3. Nanodiamond surface charge and glass removal rate 63

References 67

Chapter 4. Glass modified surface layer 71


4.2. Nanoindentation in fluid environments 77

4.2.1. Nanohardness of LHG-8 in an Ethylene Glycol 80

environment

4.2.2. Young’s modulus in fluid environments 82

4.2.3. Nanohardness in fluid environments 83

References 86

Chapter 5. MRF material removal rate model 90


5.2. Term 1: Mechanical figure of merit 91

5.2.1. Young’s modulus 92

5.2.2. Hardness 93

5.2.3. Fracture toughness 95

x

5.2.4. Mechanical FOM term 96

5.3. Term 2: Modified Preston’s equation 98

5.3.1. Wheel speed and contact area 99

5.3.2. Drag force 100

5.3.2.1. Force sensor experimental set up 101

5.3.2.2. Drag force results 102

5.4. Term 3: Abrasive size and concentration 106

5.4.1. Nanodiamond size and concentration 106

5.4.2. CI size and concentration 111

5.5. Term 4: Glass chemical durability 112

5.5.1. Chemical durability testing protocol and removal 113

rate correlation

5.5.2. Chemical durability and aging MR fluid 116

5.6. Term 5: Glass average glass bond strength 121

5.6.1 Determination of glass average single bond strength 121

5.6.2. Glass average single bond strength results 122

5.7. MRF material removal rate model 123

References 126

Chapter 6. The role of nanodiamonds in polishing zone 131


6.2. Surface texture and the MRF material removal process 132

6.3. Surface roughness and power spectral density 137

6.3.1. Varying nanodiamond concentration 137

6.3.2. Surface roughness and drag force 146

6.3.3. Surface roughness and glass mechanical properties 151

References 155

Chapter 7. Summary 156

xi

Chapter 8. Future Work 160

Appendix A. Particle size and zeta potential 163

References 165

Appendix B. Fluid Jet Polishing 166

B.1. Fluid Jet Polishing (FJP) experimental set up 166

B.2. Experimental glass set 167

B.3. Carbonyl iron (CI) slurry 168

B.4. Silicon carbide (SiC) slurry on all seven glasses 177

(not measured over time)

B.5. Addition of UK Abrasives nanodiamonds to CI slurry 181

B.5.1. BK-7 183

B.5.2. FS 186

B.5.3. LHG-8 188

B.5.4. FCD-1 191

B.5.5. FD-60 193

B.5.6. EFDS-1 196

B.5.7. NSF-6 199

B.5.8. Experimental data for three nanodiamond slurries 201

B.6. Comparison of polishing abrasives 203

B.7. Summary 206

References 207

Appendix C. 208

C.1. Comparison of near surface and bulk FOM values 208

C.2. p-v surface roughness and drag force 210

xii

C.3. Experimental peak removal rate and surface roughness 211

data tables

C.4. Mechanical drawing for force sensor assembly 224

xiii

List of Tables

Table Title Page2.1 Actual spot times measured using the high-speed camera. 22 2.2 Optical glass mechanical properties rank ordered by increasing

mechanical figure of merit. 23

3.1 Zeta potential and particle size data for CI. 36 3.2 Zeta potential and particle size measurements for NDP

nanodiamonds in three different host solutions. 43

3.3 UK Abrasive nanodiamond properties. 46 3.4 Zeta potential and particle size measurement data for UK – Low

nanodiamonds in 10-2M KNO3, DI water and MR carrier fluid host solutions.

50

3.5 Zeta potential and particle size measurement data for UK – Medium A nanodiamonds in 10-2M KNO3, DI water and MR carrier fluid host solutions.

51

3.6 Zeta potential and particle size measurement data for UK – High nanodiamonds in 10-2M KNO3, DI water and MR carrier fluid host solutions.

53

3.7 Zeta potential and particle size measurement data for UK – Medium B nanodiamonds in 10-2M KNO3, DI water and MR carrier fluid host solutions.

54

3.8 Zeta potential and particle size measurement data for UK – Medium B nanodiamonds in 10-2M KNO3, DI water and MR carrier fluid host solutions.

55

3.9 Abrasive free MR fluid data. 60 3.10 Zeta potential values of UK-Low, UK-High, UK-Medium A

and NDP nanodiamond and 0.01-vol% nanodiamond MR fluids (diluted).

66

4.2 Microhardness data measured on the bulk glass (Vickers) and nanohardness measured at a depth of 60nm into the surface (nanoindentation) in three different fluid environments.

84

5.1 Bulk and near surface Young’s modulus data for our glass set, with bulk data from the literature for comparison.

93

5.2 Measured and literature micro- and nanohardness values for our glass set. (Literature values were all made with 100gf load but according to different testing standards).

94

5.3 Bulk fracture toughness data for our six optical glasses with data from literature (when available) for comparison.

96

5.4 Mechanical FOM values calculated from Young’s modulus and hardness values measured for the bulk (B), the dry near surface (D), the near surface in DI water (W) and the near surface in MR fluid supernatant (S).

97

xiv

Table Title Page5.5 Composition data for our glass set. Values listed are weight

percentages 121

5.6 Calculated average single bond strength (sbs) [units: kJ/mol] and Term 5 (R is the gas constant [units: kJ/mol K], T is temperature = 296K and b is a unitless coefficient empirically equal to 1000.

122

6.2 0.001-vol% low friability nanodiamond MR fluid peak removal rate and surface roughness data for LHG-8.

136

6.3 0.001-vol% medium friability nanodiamond MR fluid peak removal rate and surface roughness data for LHG-8.

136

6.4 0.001-vol% high friability nanodiamond MR fluid peak removal rate and surface roughness data for LHG-8.

136

B.1 Optical glass mechanical properties rank ordered by increasing mechanical figure of merit.

168

B.2 Removal rate and surface roughness data for FJP footprints made with the CI slurry.

172

B.3 Peak removal rate and surface roughness data for FJP footprints made with SiC slurry.

179

B.4 Peak removal rate and surface roughness data for the UK-Low friability nanodiamond experiment. The first row of data for each glass is the data for the footprint made with the slurry containing only CI particles; the nanodiamonds were added after the footprints were taken.

201

B.5 Peak removal rate and surface roughness data for the UK-Medium friability nanodiamond experiment. The first row of data for each glass is the data for the footprint made with the slurry containing only CI particles; the nanodiamonds were added after the footprints were taken.

202

B.6 Peak removal rate and surface roughness data for the UK-High friability nanodiamond experiment. The first row of data for each glass is the data for the footprint made with the slurry containing only CI particles; the nanodiamonds were added after the footprints were taken.

202

C.1 0.001-vol% UK-Low friability MR fluid data. 211 C.2 0.001-vol% UK-Medium A friability MR fluid data. 212 C.3 0.001-vol% UK-High friability MR fluid data. 213 C.4 0.005-vol% UK-Low friability MR fluid data. 214 C.5 0.005-vol% UK-Medium A friability MR fluid data. 215 C.6 0.005-vol% UK-High friability MR fluid data. 216 C.7 0.01-vol% UK-Low friability MR fluid data. 217 C.8 0.01-vol% UK-Medium A friability MR fluid data. 218 C.9 0.01-vol% UK-High friability MR fluid data. 219 C.10 UK-Medium B MR fluid data. 220

xv

Table Title PageC.11 UK-Medium C MR fluid data. 221 C.12 NDP nanodiamond MR fluid data. 222 C.13 Abrasive free ramping CI data. 223

xvi

List of Figures

Figure Title Page1.1 Schematic diagram of the MRF contact zone. 2 1.2 Interferometric image of an MRF spot. Maximum removal in

the region of deepest penetration is colored purple. Typical spotting time is 2-seconds to achieve approximately 0.2µm deep spots.

2

1.3 Izumitani’s plot of polishing rate versus hardness of the hydrated layer.

5

1.4 Izumitani’s plot of polishing rate versus glass percent weight loss in water.

6

1.5 Cumbo’s surface roughness versus pH – IEP. 9 1.6 Luo and Dornfeld’s graphical model of material removal rate for

CMP. 11

1.7 MRF removal rate plotted as a function of mechanical properties.

13

1.8 Shorey’s experimental data for drag force and normal pressure versus MRF removal rate.

15

2.1 Pictures of STM before and during spot formation. 22 2.2 Optical glass (nd/vd) diagram. The refractive index (nd) at

590nm is given on the y-axis and the corresponding Abbe value (vd) is given on the x-axis. The six optical glasses used for this thesis are indicated on the diagram.

24

2.3 Photograph Mark IVxp interferometer. 25 2.4 Photographs of the Talysurf CCI 3000. 27 2.5 Diagram of an MRF spot and roughness measuring protocol. 28 3.1 SEM images of as received CI particles milled with silica. 34 3.2 SEM images of as received non-silicated CI particles. 34 3.3 CI particle distribution plot. Fused particles (not agglomerates)

make up the large end of the particle size distribution. 35

3.4 Zeta potential versus host solution pH for CI. 36 3.5 CI particle size versus host solution pH. 38 3.6 Sketches of nanodiamond (primary particle) shapes. The

average primary particle size of UDD nanodiamonds is approximately 4nm.

40

3.7 SEM image of NDP nanodiamond agglomerates. 42 3.8 Gaussian particle size distribution of NDP nanodiamonds in DI

water at pH 2.8. 43

3.9 Zeta potential versus host solution pH for NDP nanodiamonds. 44 3.10 Particle size versus host solution pH for NDP nanodiamonds. 45

xvii

Figure Title Page3.11 SEM images of (a) UK-Low, (b) UK-Medium A, (c) UK-High,

(d) UK-Medium B and (e) UK-Medium C nanodiamond agglomerates after removal from suspension, and washing.

47

3.12 Gaussian particle size distribution plot of UK nanodiamonds. 48 3.13 Zeta potential and particle size measurements for UK – Low

nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.

49

3.14 Zeta potential and particle size measurements for UK – Medium A nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.

51

3.15 Zeta potential and particle size measurements for UK – High nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.

52

3.16 Zeta potential and particle size measurements for UK – Medium B nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.

54

3.17 Zeta potential and particle size measurements for UK – Medium C nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.

55

3.18 Average zeta potential versus host solution pH for nanodiamonds suspended in MR carrier fluid.

57

3.19 Average particle size versus host solution pH for nanodiamonds suspended in MR carrier fluid.

58

3.20 Average areal rms surface roughness (left axis) and peak removal rate (right axis) experimental data for the abrasive free MR fluid spots. Surface roughness values for the initial pitch polished surfaces have been included for comparison.

60

3.21 Peak removal rate values [semi-log] for the six glasses in four MR fluids: abrasive free, 0.001-vol% UK-Low, 0.001-vol% High and 0.001-vol% UK-Medium A.

62

3.22 Peak removal rate data for abrasive free and 0.01-vol% UK-Low, 0.01-vol% UK-High, 0.01-vol% UK Medium A and 0.01-vol% NDP nanodiamond MR fluids versus surface charge measured for CI and nanodiamond particles in MR carrier fluid.

63

3.23 Peak removal rate data for abrasive free and 0.01-vol% UK-Low, 0.01-vol% UK-High, 0.01-vol% UK Medium A and 0.01-vol% NDP nanodiamond MR fluids versus nanodiamond agglomerate size measured nanodiamond particles in MR carrier fluid.

64

xviii

Figure Title Page3.24 Peak removal rate versus zeta potential of diluted MR fluid for

0.01-vol% UK-Low, 0.01-vol% UK-High, 0.01-vol% UK-Medium A, 0.01-vol% NDP and abrasive free MR fluid.

65

3.25 Sketch of nanodiamond/CI interaction hypothesis in polishing zone: (a) Attraction – more efficient removal, (b) Repulsion – less efficient removal.

67

4.1 Photograph of the Nano Indenter XP. 77 4.2 Photograph of a drop of supernatant applied to the surface of

BK-7. 79

4.3 Absolute weight loss in EG and DI water solutions for three phosphates and one silicate glass composition.

80

4.4 Average CSM nanohardness data for LHG-8 in EG and DI water solutions plotted on a semi-log plot. [Inset: average nanohardness value at 60nm depth]

81

4.5 Average CSM nanoindentation data for all six glasses. The measurements were made on the dry surfaces (shaded lines) and then again with the glass surfaces covered with a layer of MR supernatant (solid lines).

85

5.1 Drawing of the top view of a Vickers indent. 96 5.2 Bulk and near surface mechanical figure of merit (FOM) values

plotted with peak removal rate data for MRF spots taken with 0.01-vol% UK-Medium A nanodiamond MR fluid.

98

5.3 Spot contact area versus peak removal rate for BK-7. Wheel speed was held constant; pump speed was varied to maintain a constant ribbon height.

99

5.4 Wheel speed versus peak removal rate for BK-7. Spot contact area was held constant and the pump speed was varied to maintain a constant ribbon height. The corresponding linear velocities with units of meters/second are given in brackets.

100

5.5 Photograph of the piezo-electric force sensor used to measure drag force, mounted on the STM.

101

5.6 An example of the force sensor output for LHG-8 pressed against an NDP nanodiamond MR fluid ribbon, plotted with Microsoft Excel.

102

5.7 Drag force plotted versus NDP nanodiamond concentration for our six optical glasses.

103

5.8 Peak removal rate plotted versus drag force for our six optical glasses in MR fluids with increasing concentrations of NDP nanodiamonds.

103

xix

Figure Title Page5.9 Peak removal rate versus drag force for four 0.01-vol%

nanodiamond MR fluids. Spots were taken on all six glasses with each fluid. The glass types are identified for the NDP points only to make the figure easier to read. The glass order is the same for the other nanodiamond fluids.

104

5.10 Peak removal rate versus Fd/Hs for four 0.01-vol% nanodiamond MR fluids. Spots were taken on all six glasses with each fluid. The glass types are identified for the NDP points only to make the figure easier to read. The glass order is the same for the other nanodiamond fluids.

105

5.11 Peak removal rate versus nanodiamond concentration for the 29nm (UK-Medium A) and 54nm (NDP) nanodiamonds for phosphate glasses LHG-8, FCD-1 and EFDS-1.

107

5.12 Peak removal rate versus nanodiamond concentration for the 29nm (UK-Medium A) and 54nm (NDP) nanodiamonds for silicate glasses FS, BK-7 and FD-60.

108

5.13 Peak removal rate data versus our third term of the MRF material removal model. The 29nm UK-Medium A and 54nm NDP nanodiamond concentration are varied. CI size and concentration are constant as described in the text. All of the linear trend lines have confidence levels greater than 99%.

109

5.14 Peak removal rate data versus the third term of the MRF material removal rate model. The experimental data for this plot includes the increasing nanodiamond concentration for the 29nm, 35nm, 44nm and 54nm nanodiamond experiments. CI size and concentration are constant as described in the text. The confidence levels are all greater than 99%.

110

5.15 Peak removal rate and out-of-field MR fluid viscosity versus CI concentration for four glasses. CI concentration was varied with the addition of DI water.

112

5.16 Photograph of the glass chemical durability testing set up. 114 5.17 Percent weight loss, Ds, versus testing solution pH for all six

optical glasses. The Ds relationships that are used in our MRF material removal model are located on the right hand side of the figure.

115

5.18 Peak removal rate versus chemical durability, Ds, for 0.01-vol% NDP nanodiamond MR fluid experimental data and Izumitani’s data for conventional CeO2 pad polishing.

116

5.19 Measured MR fluid pH values as a function of time for 0.01-vol% UK-Low, UK-Medium A and UK-High nanodiamond MR fluids. The fluids were allowed to naturally age in the STM for 9 days.

117

xx

Figure Title Page5.20 Peak removal rate versus Term 4 of our MRF material removal

model. The 0.01-vol% UK-Low nanodiamond fluid was allowed to naturally age for 9 days in the STM. All spots were made using the same operating conditions. The confidence levels for all of the linear trend lines drawn in the figure are greater than 99%.

118

5.21 Peak removal rate versus Term 4 of our MRF material removal model. The 0.01-vol% UK-Medium A nanodiamond fluid was allowed to naturally age for 9 days in the STM. All spots were made using the same operating conditions. The confidence levels for all of the linear trend lines drawn in the figure are greater than 99%.

119

5.22 Peak removal rate versus Term 4 of our MRF material removal model. The 0.01-vol% UK-High nanodiamond fluid was allowed to naturally age for 9 days in the STM. All spots were made using the same operating conditions.

120

5.23 Peak removal rate versus Term 5 with the average single bond strength for five MR fluids.

123

5.24 Experimental peak removal rate data for six glasses with various MR fluids versus our MRF material removal rate model, incorporating mechanics, polishing particle properties and chemistry. The terms A, v, φCI, CCI, R and T are all constant.

124

6.1 False color surface maps and areal roughness values (all in nm) of the LHG-8 surfaces. No grooving is observed on the pitch polished surface (upper left). The paths taken by abrasives across the part surface within spots are seen as grooves extending from top to bottom in all other images. The field of view is 0.35mm x 0.35mm. Nanodiamond concentration is 0.001-vol%. [MR fluid flow direction indicated].

133

6.2 Average areal rms surface roughness versus elapsed time for LHG-8.

135

6.3 Peak removal rate versus elapsed time for LHG-8 corresponding to surface roughness data shown in Figure 6.2.

135

6.4 Average p-v surface roughness inside MRF spots made with NDP nanodiamond MR fluid at various nanodiamond concentrations.

138

6.5 1-D PSD analyzed in the horizontal direction (perpendicular to the MRF grooves) inside LHG-8 MRF spots made with 0, 0.001 and 0.03-vol% NDP nanodiamond MR fluids and compared to the initial pitch polished surface.

140

xxi

Figure Title Page6.6 1-D PSD analyzed in the vertical direction (parallel to the MRF

grooves) inside LHG-8 MRF spots made with 0, 0.001 and 0.03-vol% NDP nanodiamond MR fluids and compared to the initial pitch polished surface.

141

6.7 Average rms surface roughness inside MRF spots made with NDP nanodiamond MR fluid at various nanodiamond concentrations.

143

6.8 1-D PSD analyzed in the horizontal direction (perpendicular to the MRF grooves) inside EFDS-1 MRF spots made with 0, 0.007 and 0.1-vol% NDP nanodiamond MR fluids and compared to the initial pitch polished surface.

144

6.9 1-D PSD analyzed in the vertical direction (parallel to the MRF grooves) inside EFDS-1 MRF spots made with 0, 0.007 and 0.1-vol% NDP nanodiamond MR fluids and compared to the initial pitch polished surface.

145

6.10 Average areal p-v surface roughness inside MRF spots made with 0 – 0.01 vol% NDP nanodiamond MR fluid versus the corresponding drag force.

146

6.11 Average areal rms surface roughness inside MRF spots made with 0 – 0.01 vol% NDP nanodiamond MR fluid versus the corresponding drag force.

147

6.12 Average areal p-v surface roughness inside MRF spots made with 0 – 0.01 vol% NDP nanodiamond MR fluid versus the corresponding drag force. Linear trend lines are drawn for the individual glass types.

148

6.13 Average areal rms surface roughness inside MRF spots made with varying NDP nanodiamond concentration (0 – 0.01 vol%) MR fluid versus the corresponding drag force. Linear trend lines are drawn for the individual glass types.

149

6.14 Average areal p-v surface roughness inside MRF spots made with varying glass type versus the corresponding drag force. Linear trend lines are drawn for the different NDP nanodiamond concentrations (0–0.01-vol%) in the MR fluid.

150

6.15 Average areal rms surface roughness inside MRF spots made with varying glass type versus the corresponding drag force. Linear trend lines are drawn for the different NDP nanodiamond concentrations (0-0.01-vol%) in the MR fluid.

151

6.16 Average areal p-v surface roughness values versus E1/2/Hv. The spots were taken using 0.01-vol% NDP MR fluid. The mechanical properties were measured on the bulk glass.

152

xxii

Figure Title Page6.17 Average areal p-v surface roughness values versus Es

1/2/Hs. The spots were taken using 0.01-vol% NDP MR fluid. The near surface mechanical properties were measured in MR fluid supernatant environment.

153

6.18 Average areal p-v surface roughness versus (Es1/2/Hs)⋅(1/Fd).

Spots were taken using the 0.01-vol% NDP fluid. 154

8.1 Peak removal rate for FS and BK-7 versus nanodiamond friability. The three MR fluids contained 0.01-vol% nanodiamonds.

160

A.1 Drawing of a negatively charged particle in suspension. 164 B.1 FJP experimental set up. Typically 8L of slurry is continuously

circulated through this system. The delivery system and nozzle are part of a prototype Zeeko FJP machine.

167

B.2 FJP footprint shape and orientation for surface roughness measurements using the CCI 3000.

170

B.3 Peak removal rate versus elapsed time for the CI slurry. 170 B.4 Normalized peak removal rate versus elapsed time for the CI

slurry. 171

B.5 Average areal p-v surface roughness for footprints taken with the CI slurry.

172

B.6 False color images of the initial pitch polished surfaces and the surfaces inside the FJP footprints (1st set of footprints).

174

B.7 Average areal rms surface roughness for footprints taken with CI slurry.

175

B.8 Average areal rms surface roughness for the pitch polished surfaces and the first FJP footprints versus 1/Hv

1/2. 176

B.9 PSD plots for LHG-8 and FS measured inside the CI slurry footprint and the adjacent pitch polished surface.

177

B.10 Peak removal rate versus glass mechanical figure of merit for SiC and CI slurries. SEM images of SiC and CI are included to the right of the plot.

178

B.11 Average areal p-v surface roughness measured inside SiC footprints compared to CI footprints and the pitch polished surface.

179

B.12 False color surface images inside SiC footprints. 180 B.13 Average areal rms surface roughness versus 1/Hv

1/2. 180 B.14 Slurry pH as a function of time. 182 B.15 SEM images of FJP slurries. 182 B.16 Peak removal rate data on BK-7 with 3 slurries. 183 B.17 Normalized peak removal rate for BK-7 with 3 slurries. 183 B.18 Average areal p-v surface roughness for BK-7 with 3

nanodiamond/CI slurries. 185

xxiii

Figure Title PageB.19 Average areal rms surface roughness for BK-7 with 3


B.20 Peak removal rate for FS with 3 nanodiamond/CI slurries. 186 B.21 Normalized peak removal rate for FS with 3 nanodiamond/CI

slurries. 187

B.22 Average areal p-v surface roughness for FS with 3 nanodiamond/CI slurries.

187

B.23 Average areal rms surface roughness for FS for 3 nanodiamond/CI slurries.

188

B.24 Peak removal rate for LHG-8 with 3 nanodiamond/CI slurries. 189 B.25 Normalized peak removal rate for LHG-8 with 3


B.26 Average areal p-v surface roughness for LHG-8 with 3 nanodiamond/CI slurry footprints.

190

B.27 Average area rms surface roughness for LHG-8 with 3 nanodiamond/CI slurry footprints.

190

B.28 Peak removal rate data for FCD-1 footprints made with 3 nanodiamond/CI slurries.

191

B.29 Normalized peak removal rate data for FCD-1 footprints made with 3 nanodiamond/CI slurries.

192

B.30 Average areal p-v surface roughness data inside FCD-1 footprints made with 3 nanodiamond/CI slurries.

192

B.31 Average areal rms surface roughness data inside FCD-1 footprints made with 3 nanodiamond/CI slurries.

193

B.32 Peak removal rate data for FD-60 footprints made with 3 nanodiamond/CI slurries.

194

B.33 Normalized peak removal rate data for FD-60 footprints made with 3 nanodiamond/CI slurries.

194

B.34 Average areal p-v surface roughness data inside FD-60 footprints made with 3 nanodiamond/CI slurries.

195

B.35 Average areal rms surface roughness data inside FD-60 footprints made with 3 nanodiamond/CI slurries.

195

B.36 Peak removal rate data for EFDS-1 footprints made with 3 nanodiamond/CI slurries.

197

B.37 Normalized peak removal rate data for EFDS-1 footprints made with 3 nanodiamond/CI slurries.

197

B.38 Average areal p-v surface roughness data inside EFDS-1 footprints made with 3 nanodiamond/CI slurries.

198

B.39 Average areal rms surface roughness data inside EFDS-1 footprints made with 3 nanodiamond/CI slurries.

198

B.40 Peak removal rate data for NSF-6 footprints made with 3 nanodiamond/CI slurries.

199

xxiv

Figure Title PageB.41 Normalized peak removal rate data for NSF-6 footprints made

with 3 nanodiamond/CI slurries. 200

B.42 Average areal p-v surface roughness data inside NSF-6 footprints made with 3 nanodiamond/CI slurries.

200

B.43 Average areal rms surface roughness data inside NSF-6 footprints made with 3 nanodiamond/CI slurries.

201

B.44 Peak removal rate data ranges for the FJP footprints taken on the set of seven glasses with the five different slurries. The nanodiamond/CI slurries are indicated by the nanodiamond only.

205

C.1 Peak removal rate versus the FOM-B calculated for the bulk material. Data for all nanodiamond fluids contain 0.01-vol% nanodiamonds.

209

C.2 Peak removal rate versus the FOM-S calculated 60nm into a sample in MR supernatant. Data for all nanodiamond fluids contain 0.01-vol% nanodiamonds.

209

C.3 Peak removal rate versus the FOM-D calculated 60nm into a dry sample. Data for all nanodiamond fluids contain 0.01-vol% nanodiamonds.

209

C.4 Peak removal rate versus the FOM-W calculated 60nm into a sample in DI water. Data for all nanodiamond fluids contain 0.01-vol% nanodiamonds.

209

C.5 Average areal p-v surface roughness versus inverse drag force with varying glass type for UK-Low nanodiamond MR fluids.

210

C.6 Average areal p-v surface roughness versus inverse drag force with varying glass type for UK-Medium A nanodiamond MR fluids.

210

C.7 Average areal p-v surface roughness versus inverse drag force with varying glass type for UK-High nanodiamond MR fluids.

210

C.8 Mechanical drawing for the force sensor assembly. 224

xxv

List of Symbols

Symbol or abbreviation Definition

A Contact area

Å Angstrom

Ac Abrasive concentration

b Coefficient (Model: Term 5)

BCI Coefficient (Model: Term 3)

Bnd Coefficient (Model: Term 3)

C Average water concentration

c Crack length

CB,2 Coefficient (Buijs et al.)

Cc Coefficient (Cook)

CCD Charge coupled device

CCI Carbonyl iron concentration

CI Carbonyl iron

CL Confidence level

CL-D Coefficient (Luo et al.)

CMP Chemical mechanical polishing

Cnd Nanodiamond concentration

Co Amount of material removed due to chemical etching

Cp Preston's coefficient

cP Centipoise

Cp,B1 Modified Preston’s coefficient (Buijs et al.)

Cp,M Modified Preston’s coefficient (Matsuo et al.)

Cp,M-D Modified Preston’s coefficient (Moon et al.)

Cp,S Modified Preston’s coefficient (Shorey et al.)

CSM Continuous stiffness measurement

xxvi


CVD Chemical vapor deposition

d Diffusion coefficient

D Length of diagonal

ddp Depth of deepest penetration

DI Deionized

ds Thickness of the substrate

Ds Percent weight loss

dt Thickness of substrate

E Young's modulus

EAEnergy required to activate bond between abrasive particle and work piece

EB Bulk Young’s modulus

EDX Energy dispersive X-ray

EG Ethylene glycol

Es Near surface Young’s modulus

ESA Electrokinetic Sonic Amplitude

Fd Drag force

Ff Frictional force

FJP Fluid jet polishing

FOM Figure of merit

FT-IR Fourier Transform Infrared

GANR Grazing angle neutron reflectometry

GPa Giga pascal

Hk Knoop microhardness

Hs Near surface Berkovitch nanohardness

HTSMD Heat treated synthetic monocrystalline diamond

Hv Vickers microhardness

I Transmitted intensity at the absorption peak

xxvii


I0 Absorption-free transmitted intensity

IEP Iso-electric point

IEPabr Iso-electric point of the abrasive particle

IPC Iron pentacarbonyl

ISE Indentation size effect

Kc Fracture toughness

L Normal load or force

Lm,c Mean size of the load bearing particle

Ls Length of the line scan

M∞ Saturated water uptake value

mm Millimeters

MR Magnetorheological

MRF Magnetorheological finishing

MRR Material removal rate

MRRpeak Peak material removal rate

MRRvol Volumetric material removal rate

Mt Experimental water uptake curve

mV Millivolt

n Number of indenting points or active abrasives

nd Refractive index

nm Nanometers

Nx or Ny Number of pixels in the x or y direction

p pressure

Pa Pascal

pHMRF Magnetorheological fluid pH

Pi Normal load per load bearing particle

PSD Power spectral density

xxviii


p-v Peak-to-valley

R Universal gas constant

R2 Relative predictive power of a model

rms Root mean square

Rp-v Peak-to-valley surface roughness

Rq rms average surface roughness calculated from a line scan

sbs Average single bond strength of the glass network formers

sbsabr Single bond strength of the abrasive particle

SEM Scanning electron microscope

SMD Synthetic monocrystalline diamond

SPD Polycrystalline diamond

Sq rms surface roughness calculated for an entire surface

STM Spot taking machine

t Time

T Temperature

TEM Transmission electron microscope

UDD Ultra dispersed diamond

v Velocity

vl Longitudinal wave velocity

Vr volume removed by one abrasive

vs Shear wave velocity

z Surface topography

µm Micrometers

ΒΕΤ Brunauer, Emmett and Teller surface area method

α Angle between opposite faces of Vickers indenter

ε Extinction coefficient

φ Abrasive diameter

xxix


φCI Average CI particle size

φnd Average nanodiamond particle size

µ Coefficient of friction

νd Abbe value

ρ Density

σ Abrasive size standard deviation

τ Shear stress

1

Chapter 1

Introduction

1.1 MRF Process

Magnetorheological finishing (MRF) is a deterministic sub-aperture polishing

process developed at the Center for Optics Manufacturing (COM) by a group of

international collaborators1 and commercialized by QED Technologies, Inc.2 MRF is

based on a magnetorheological (MR) fluid that consists of carbonyl iron (CI), non-

magnetic polishing abrasives, water, other carrier fluids and stabilizers. The MR

fluid in the absence of a magnetic field has the viscosity of approximately 0.04 to 0.1

Pa·s (at a shear rate of ~800 1/sec). Once introduced to a magnetic field (~2-3kG),

the viscosity increases four orders of magnitude.3

Figure 1.1 schematically shows an optic depressed into MR fluid flowing left

to right on a rotating wheel. The non-magnetic polishing abrasives are depicted as

red spheres. The CI particles are shown as gray spheres. Calculations have predicted

the existence of a thin layer of high shear in contact with the optic where polishing

occurs.4 Material removal for MRF has been shown to be dominated by shear stress,

τ, (equal to tangential drag force whose magnitude5 is on the order of 1 - 5N divided

by contact area). The normal force, L, of the abrasive acting on the surface during

MRF removal has been hypothesized to be very small and approximately equal to 1 x

10-4 mN.6 Particles involved in conventional pitch or pad polishing remove material

with normal forces in the range of 5 – 200mN.7 It is uncertain how the CI, the

abrasive, or both the CI and the abrasive contact the surface to cause removal using

conventional MR fluids.

The MRF removal function is characterized with MRF spots. An MRF spot is

created by lowering a non-rotating optic into the rotating MR fluid ribbon for a

known period of time. Material is removed in a characteristic D-shaped spot, as

shown in Figure 1.2. The peak removal rate is calculated by interferometrically

determining the depth of deepest penetration, ddp, and dividing by the spot time.

2

Material removal can also be reported as volumetric removal rate which is calculated

by dividing the volume of the material removed in the entire spot by the dwell time.

The volume is calculated interferometrically by finding the height of each data point

and multiplying it by the area of the pixel.8

Figure 1.1 Schematic diagram of the MRF contact zone.

Figure 1.2 Interferometric image of an MRF spot. Maximum removal in the region of deepest penetration is colored purple. Typical spotting time is 2-seconds to achieve approximately 0.2µm deep spots.

ddp

Flow Direction

3

1.2 Conventional optical material removal and smoothing models in grinding and polishing

Determining the removal mechanism for conventional optical polishing has

been studied for over 100 years. This section describes a few of the many material

removal models proposed for conventional optical polishing, ranging from purely

mechanical to purely chemical plus those that include both chemistry and mechanics.

F.W. Preston9 proposed that the amount of polishing was proportional to the

normal pressure, p, between the felt lap and the glass work piece multiplied by the

velocity v of the glass relative to the felt lap in time t. Equation 1.1 gives modern

version of Preston’s ideas, re-written for the material removal rate, MRR, shown in

equation 1.2 where Cp is Preston’s coefficient with units of Pa-1. This equation

applies equally well to both grinding and polishing. Preston also developed a model

for work, w, in the polishing process that is shown in equation 1.3. This model takes

into account the drag associated with polishing which can be calculated by

multiplying the coefficient of friction, µ, the contact area, A (between the felt lap and

the glass work piece) and the pressure. Preston believed that “the rate of polishing of

glass is proportional to the rate at which work is done on each unit area of glass,”

(p.217 in ref. 9).

Amt of polishing pvt∝ (1.1)

(1.2) pvCMRR p=

Apvtw µ= (1.3)

Sixty-six years after Preston published his theories on the material removal

process, Buijs et al. revised Preston’s work equation to incorporate the material

properties of Young’s modulus, hardness and fracture toughness in a new model for

lapping of optical glass. Their experiments were performed with Schott B270 glass,

copper and iron lapping plates and SiC abrasive lapping slurries of five different

sizes. Their model for glass removal rate, shown in equation 1.4, is based on the

assumption that material is being removed by rolling abrasive particles (i.e. three

body abrasion).10,11

4

pvHK

ELP

nCMRRvccm

iBp 2

45

,

43

1, ⋅= [µm/s] (1.4)

In equation 1.4, Cp,B1 is their unitless Preston coefficient, n is the number of

indenting points, Lm,c is the mean size of the load bearing particles, Pi is the normal

load per load bearing particle, Kc is the fracture toughness of the work piece, E is the

Young’s modulus of the work piece and Hv is the hardness of the work piece.

According to equation 1.4, the material removal rate appears to vary inversely with

the mean size of the load bearing particles, Lm,c. Buijs et al. go on to show that the

normal load per load bearing particle is inversely proportional to the number of load

bearing particles. The number of load bearing particles is proportional to Lm,c-2, which

in turn makes the material removal rate proportional to the square root of the mean

particle size of the load bearing particles.10

Buijs et al. found the average peak-to-valley surface roughness, Rp-v, to be

independent of pressure and velocity as shown in equation 1.5 for a given abrasive

size and lapping plate. The unitless constant CB2 in equation 1.5 is dependent on the

particle shape. They showed experimentally that the peak-to-valley surface

roughness increases with increasing mean abrasive particle size. The mean abrasive

size is not the same as the mean size of the load bearing particles. Buijs et al.

determined that an increase in abrasive size increased the load per particle.10 In

summary, Buijs et al. reported that when grinding glass, an increase in the average

abrasive size will increase both material removal rate and peak-to-valley surface

roughness.

2/12/1

2 iv

Bvp PHE

CR =− (1.5)

Izumitani summarized the different published mechanisms of removal for

conventional polishing of optical glass and classified them into three categories: wear

theory, flow theory and chemical theory. He proposed a theory for polishing

primarily with CeO2 on polyurethane pads that combined chemical and mechanical

interactions. Izumitani believed that a softer hydrated layer was developed at the

5

surface of the glass that occurred as a result of its chemical durability in water. This

layer was then mechanically removed with the polishing abrasives. The hardness of

the hydrated layer was measured using a Vickers micro-indenter after the glass had

been leached for 60 minutes in 0.1N nitric acid. A linear correlation between the

polishing rate and the Vickers micro hardness of the hydrated layer was found (see

Figure 1.3).12

Figure 1.3 Izumitani’s plot of polishing rate versus hardness of the hydrated layer [Ref. 12 with permission]

In Figure 1.3, Izumitani broke up the data into two categories: silicate glasses

and borate glasses. He found linear relationships between polishing rate and the

hardness of the hydrated layer for each glass type. The hydrated layer formed at

different rates for the two different glass types. Izumitani concluded that polishing

rate was inversely dependent on the hardness of the hydrated layer and directly

proportional to the rate at which the hydrated layer formed.12

Izumitani12 also compared the percent weight loss of glass in water to the

polishing rate. His data shown in the semi-log plot in Figure 1.4 indicates a non-

linear relationship between polishing rate and percent weight loss of glass in water.

The relationship appears to be independent of glass type, unlike his correlation

between polishing rate and the hardness of the glass hydrated layer.

6

Figure 1.4 Izumitani’s plot of polishing rate versus glass percent weight loss in water [Ref. 12 with permission]

Cornish and Watt13 believed that water hydrolyzed the glass surface to form a

less than 100Å hydrated surface layer. Their polishing experiments were performed

with felt laps, oxide polishing slurries and silicate glasses. They state that the

partially hydrated silica skeleton did not dissolve itself to give higher rates observed

with polishing oxides. They believed that, once hydrolysis occurred, the Na+ was

replaced with H+. Then once the polishing oxide (CeO2) came in close contact, a

bond was formed with the Si and the silica unit (Si(OH)4) was removed by adsorption

on the polishing grain. The residual surface was left hydrolyzed and ready for the

next polishing oxide particle. Cornish and Watt did not offer any quantitative model

for the proposed removal process.

Kaller14 proposed a theory for the mechanism of glass polishing that was

similar to the one proposed by Cornish and Watt. He focused on the importance of

extra lattice defects in the oxide polishing abrasives to aid in the gripping of the glass

material for removal. Work for his research was performed with quartz and quartz

glass (i.e. fused silica) on polishing pitch. He also believed that the polishing

abrasives should be softer than the material that is to be polished.

Kaller’s theory was that crystalline oxides, such as Fe2O3 and CeO2, contained

a large number of lattice defects due to the manufacturing process. These defects

reduced the strength of these oxides, causing them to mechanically rub off or break

7

when the polishing grain came in contact with the glass substrate. The newly

exposed defects on the polishing grain entered into a solid-state reaction with the

glass surface. The molecules on the glass surface bonded with the polishing abrasive

that was being rubbed across the surface. Once the new bond came in contact with

the aqueous fluid, the (glass-abrasive) molecule dissolved, creating a silicate

solution.15

Hoshino et al. 16 concentrated on polishing a thin film of silicon dioxide with

cerium oxide on a polyurethane pad. They agreed with Kaller that bonding occurred

between the abrasive particle and the glass surface. But unlike Kaller they also

believed that larger quantities of SiO2 were involved. Their theory was that when the

cerium oxide particles came in contact with the SiO2 film, Ce-O-Si bonds would be

formed and a lump of SiO2 would be mechanically removed from the surface.

Hoshino et al. defined a lump to be larger than a Si(OH)4 monomer. After removal

from the surface, the lump was released from the CeO2 particle into the aqueous

solution. They believed the rate of removal was affected not only by the formation

rate of the Ce-O-Si bonds, but also by the mechanical scraping of the lump off of the

surface. They verified their hypothesis by analyzing polishing residues and the

polished surface with the scanning electron microscope (SEM), X-ray micro-analysis

(XMA) and Fourier-transform-infrared-attenuated total reflectance (FT-IR-ATR).

They found surface layers on the polished SiO2 films using FT-IR-ATR which they

believe were modified by the CeO2 during polishing. They were able to restore the

surface layers to their original condition by washing the surfaces with an acidic

solution (HNO3/30% H2O2 = 1/1 by volume). Silica aggregates on the order of

600nm were found in the polishing waste. Hoshino et al. concluded that SiO2

particles of that size would leave visible damage on the surface of the polished film.

No surface damage was observed so they believe that the SiO2 particles aggregated

together after the SiO2 lumps were released by the CeO2 particles in the polishing

waste.

8

Cook17 proposed a similar hypothesis of the oxide polishing abrasives atomically

attaching to the silica in the optical glass being polished, and then being mechanically

ripped off. Cook did not perform his own experiments; he analyzed previously

published data for oxide polishing abrasives and optical glasses and developed an

equation for polishing rate which was dependent upon the (polishing abrasive) oxide

single bond strength (sbs), the polishing abrasive slurry pH and the isoelectric point

(IEP). [The IEP is the slurry pH where there is no net surface charge on an oxide

(substrate or abrasive).] His equation for removal rate can be found as equation 1.6,

where Cc is a proportionality constant.

)(log10 abrabr

c

IEPpHsbsCMRR

−⋅= (1.6)

Cook believed that the polishing abrasives should have IEP values similar to

the slurry pH, and that the slurry pH should have a higher pH than the glass surface

in order to maximize removal rates. The optimum pH for CeO2 and ZrO2 working

silicate glasses was hypothesized to be 9.8.17

Cumbo et al. 18 were more interested in achieving the smoothest surface, unlike

Cook who was interested in predicting the conditions for maximum removal. In his

thesis Cumbo developed the slurry-charge-control effect which states that, for silicate

glass types, the abrasive IEP should be lower than the pH of the fluid. This resulted

in the same sign for the surface charges on the glass and the polishing agent. Cumbo

found that this condition produced the smoothest surfaces, because surface chemistry

prevented particle agglomeration. He found that when the mean particle size (due to

agglomeration) was larger, the rms surface roughness was higher. A full factorial

experimental design was used to complete the polishing experiments. There were

three glass types: fused silica, borosilicate BK-7 and lead silicate SF-6; three pH

levels: 4, 7 and 10; and three types of polishing abrasives: CeO2, ZrO2 and Al2O3.

The experiment was performed using polyurethane foam on a continuous polishing

machine. Figure 1.5 is a plot of the rms surface roughness values as a function of

(pH-IEP), confirming Cumbo’s slurry-charge-control effect, which stated that the

9

charges on the glass surface and on the polishing abrasive should be the same sign

and the fluid pH should be larger than the abrasive particle IEP to achieve the

smoothest surfaces.

It is interesting to note that Cumbo’s18 slurry-charge-control effect yielded high

removal rates as well for certain abrasive/glass combinations, because minimal

agglomeration of slurry particles was thought to promote better fit to the lap and

closer contact for more abrasive particles to the part surface.

Figure 1.5 Cumbo’s surface roughness versus (pH – IEP). [Ref. 18 with permission]

1.3 CMP material removal and surface roughness models

Dunken19 stated that the chemical dissolution rate for glass was proportional

to Arrhenius’ equation shown in equation 1.7 where EA is the activation energy (the

additional energy required for a thermally activated reaction to take place), R is the

gas constant and T is the temperature. Sugimoto et. al.20 polished silicon dioxide with

fumed silica and polishing pads. They found that it was difficult to use the Arrhenius

equation to exclusively characterize removal because the Chemical Mechanical

Polishing (CMP) mechanism included both mechanics and chemistry.

RTEAeMRR −∝ (1.7)

Moon and Dornfeld21 modified Preston’s equation for their application of

CMP of silicon to develop a model for a polishing rate equation. They found that

Preston’s coefficient is proportional to the coefficient of friction measured in-situ

10

during the CMP process. Their proposed model can be found in Equation 1.8, where

Cp,M-D is their modified Preston coefficient with units of Pa-1.

pvCpvCMRR DMpp µ−== , (1.8)

Experimentally they determined the value of Cp,M-D to be 2.4 x 10-9 (Pa-1).

The experiment was performed on three bare silicon wafers using a 70-90 nm grain

size colloidal silica dispersion. The amount of abrasive in the slurry was 4.5wt%.

Three different types of polishing pads were used, and the coefficient of friction was

found to change with the type of pad. They found that the removal rate increased

linearly with the coefficient of friction, similar to Cumbo’s results.18, 21 Moon and

Dornfeld’s modified equation looks almost identical to Preston’s work equation 1.3,

if it were a rate equation. Equation 1.8 does not multiply a term for contact area by

the pressure as in equation 1.3.

Matsuo et al. 22 modified Preston’s equation using frictional force. A model

was developed that had the polishing rate proportional to the frictional force instead

of the polishing pressure, similar to Shorey’s6 model for MRF which will be

discussed later. The model for CMP of copper for ultra large-scale integrated circuits

can be found in equation 1.9, where Cp,M is their modified Preston coefficient with

units of N-1. In their experiments they polished copper using two 1wt% silica

(AEROSIL OX-50)-based slurries, where the concentration of glycine and water were

varied and the frictional force was measured during polishing.

(1.9) vFCMRR fMp,=

Frictional force, Ff, is equal to the normal force (or load), L, multiplied by the

coefficient of friction. Their experimental data supported the model by showing that

there is a better linear fit between the polishing rate and frictional force than the

polishing rate and polishing pressure.22

Luo and Dornfeld23 believed that the material removal rate was dependent on

more than just the parameters outlined by Preston which Moon and Dornfeld used for

their model.21 Luo and Dornfeld proposed a model for the material removal rate in

CMP applications that also included a term for chemical etching along with terms that

11

characterize mechanical removal. Their simplified model can be found in equation

1.10, where CL-D is their proportionality constant with units of kg(min/µm), ρ is the

density of the wafer, Vr is the volume of material removed by one abrasive particle

and Co is the material removed per unit time due to chemical etching. Active

abrasives are those that are in contact with the surface performing removal; abrasives

that are not in contact with the surface are considered inactive.

0CnVCMRR rDL += − ρ (1.10)

Luo and Dornfeld later proposed that the CMP material removal rate increased

linearly with abrasive weight concentration, breaking this dependence into three

regions (see Figure 1.6). The first region was where the material removal rate

increased due to dominating chemical effects. The material removal rate also

increased with abrasive concentration in the second region which Luo and Dornfeld

characterized as being the mechanically dominant region. Finally the third region

was referred to as the removal saturation region where the material removal rate no

longer increased with abrasive concentration. Removal saturation occurs at the

concentration where the addition of abrasive particles no longer increases the number

of active abrasives. Therefore Luo et al. concluded that the material removal rate in

the saturation region was still proportional to the number of active abrasives and was

mechanically dominant similar to the second region.24

Figure 1.6 Luo and Dornfeld’s graphical model of material removal rate for CMP (Ref. 24 with permission).

12

Bielmann et al.25 performed CMP experiments on tungsten using 0.1-10µm

alumina particles. They found that the surface roughness was insensitive to the size

of the particles. This did not agree with Buijs10 and Cumbo, although Cumbo was

referring to the agglomeration size, not the initial particle size, and also worked

exclusively with glass.18

Bielmann and Mahajan published another paper where they discussed results

for CMP polishing of oxide films with silica that showed that as the particle size of

the abrasive increased, there was a transition between the mechanisms of removal

from a surface area based model to an indentation-based model.26 The surface area

based model used when the abrasive particles are smaller than 0.5µm is shown in

equation 1.11. The idea behind this model was that the smaller particles had a larger

surface area and were able to remove more material than larger particles.26 In

equations 1.11, 1.12 and 1.13, Ac is the abrasive concentration and φ is the abrasive

particle diameter. 3131 −∝ φcAMRR (1.11)

The indentation based model is shown in Equation 1.12, which was used when

particles were larger than 0.5µm. The idea behind this model was that the larger

particles produced larger indents which gave higher removal rates; but as the amount

of particles increased, the force exerted on each individual particle was reduced and

the removal rate decreased.26 3431 φ−∝ cAMRR (1.12)

The transition between the two mechanisms of removal occurred at 0.5µm. They also

found that the surface became rougher when particles larger than 0.5µm were used.26

Luo and Dornfeld also did work trying to correlate the relationship between

the size of the abrasive and the material removal rate for CMP. They found an

inverse relationship between the material removal rate and the diameter of the

abrasive particle. A simplified version of their model is shown in Equation 1.13

where σ is the standard deviation of the particle size.27

13

3

2)3(φσφ +

∝MRR (1.13)

1.4 MRF material removal models

Lambropoulos et al. 28,29 were in agreement with Buijs that the lapping

removal rate scaled linearly with E/KcHk2 and that the surface roughness was

proportional to 1/Hk1/2 for optical glasses and crystals. They devised a form of

equation 1.4 that explicitly showed removal rates in loose abrasive grinding to vary

linearly with abrasive size. They attempted to fit conventional polishing data to this

model and found that there was not a linear correlation, but rather a power fit (MRR α

(E/Kc·Hk2)x) with an exponent of approximately 0.5. This along with Cook’s17

theories led to the implication that other chemo-mechanical factors played a role in

the removal process in the polishing of optical glasses.28

Although Buijs’ model did not correlate linearly with conventional polishing,

Lambropoulos et al. found a linear correlation between a similar ratio E7/6/(Kc Hk23/12)

and the volumetric material removal rate in MRF for glasses and crystals spotted with

a cerium oxide MR Fluid. These data implied that the MRF process was more

mechanical in nature compared to conventional polishing. The data that were

collected for the experiment are found in Figure 1.7.28 The R2 value for this linear fit

was 0.92.30

Figure 1.7 MRF removal rate plotted as a function of mechanical properties [Ref. 28 with permission]

14

Shorey6 also performed experiments with Magnetorheological Finishing

(MRF). He proposed an altered Preston’s equation for modeling the material removal

rate that included a term for the coefficient of friction. Equation 1.14 shows the

evolution from equation 1.2 to Shorey’s proposed model for material removal rate,

where Cp,S was a new coefficient, with units of Pa-1, that included information not

specifically called out with a quantitative term such as the chemistry of the carrier

fluid, abrasive type and glass type. The drag force, Fd, is equal to the coefficient of

friction, µ, multiplied by the normal force, L.

vCvAFCv

ALCv

ALCpvCMRR Sp

dSpSppp τµ

,,, ===== (1.14)

Shorey’s equation contained a term for the coefficient of friction, or drag

force similar to Preston’s wear equation. He justified this by stating that MRF

removal is controlled more by the shear stress rather than by the normal pressure, p.

His evidence is given in the two experimental plots found in Figures 1.8a and 1.8b.

Both plots have linear correlations between either the drag force or peak normal

pressure versus removal rate on fused silica, but only the drag force plot intercepts at

the origin. Therefore Shorey concluded that there must be drag force to have removal

in MRF, but normal pressure was not required.

15

Figure 1.8 Shorey’s experimental data for drag force and normal pressure versus MRF removal rate [Ref. 5 with permission].

Shorey also measured the effect of abrasive concentration on the drag force

for three types of abrasive at the same carbonyl iron concentration. He found that all

three abrasives reduced the drag force as more additions were made. The most

significant drop in drag force was observed for the nanodiamond abrasives. He also

found that as more abrasives were added, the removal rate increased. These two

findings seemed to be saying the exact opposite thing, but he hypothesized that both

were true, due to a transition from two-body abrasion to three-body abrasion. All of

Shorey’s drag force measurements were made on sapphire, which was not the

material being polished.6

16

Shorey’s work left many issues unresolved, providing topics for further study.

He brought up the question of how the CI particle size and shape influenced removal.

He also suggested future experiments to determine how the composition of the MR

fluid and the hardness of the CI particles affected the way material was removed.

Another area which he thought would benefit from more research was gaining a

better understanding of the hydrated layer and how MRF interacted with and removed

this layer. Finally, equation 1.14 contains information on chemistry, abrasive type

and glass type that are buried within Preston’s coefficient. He felt that it would be

very interesting to pull out more terms from this coefficient with the goal of better

understanding the MRF removal process.6

1.5 Overview of thesis

The motivation behind this thesis work was to gain a better understanding of the

MRF process, specifically the interaction between the nanodiamonds and glass. In

this thesis it will be shown how the addition of systematically varied nanodiamond

friability to the MR fluid affects the removal mechanism and surface texture for a set

of six optical glasses. The set of six optical glasses were chosen based on their

mechanical properties, shown by Lambropoulos28 to be proportional to the MRF

volumetric material removal rate.

In Chapter 2 our experimental approach is outlined. We also give descriptions of

all the instrumentation used for this thesis work.

The physical characteristics of the CI and nanodiamond particles will be

described in Chapter 3. Along with the descriptions is a discussion about the

interaction between these two particles and the glass surface during polishing.

Hydrated surface layers have been studied in conventional polishing. Work

reported here was done to measure the near surface nanohardness and Young’s

modulus of three optical glasses in dry and liquid environments. Chapter 4 will

discuss how water and MR fluid play a role in hydrating the surface layer of glasses.

17

The near surface nanohardness and Young’s modulus values will be included in our

MRF material removal rate model in Chapter 5.

Lambropoulos’ relationship28 between the mechanical figure of merit,

E7/6/KcHv23/12, and volumetric removal rate was modified and expanded in this thesis

by including additional terms for the near surface mechanical properties, shear stress,

polishing particle size and concentration, glass chemical durability and glass

composition. Each of these terms will be discussed individually and then the entire

model will be introduced and compared to experimental material removal rate data in

Chapter 5.

The presence of nanodiamonds plays a large role in the resultant surface texture

of the surface inside a MRF spot. Chapter 6 will explore the effects of nanodiamond

concentration and friability level on the surface roughness inside MRF spots.

The thesis will be summarized in Chapter 7 and suggestions for future work will

be given in Chapter 8. Appendix B is a comparison of the role of nanodiamond

abrasives in MRF and in fluid jet polishing (FJP), from experiments conducted by the

author in Switzerland in June 2005. In this work both CI and nanodiamonds were

used with the same glass set to help us better understand how the MRF material

removal process differs from polishing with an abrasive fluid jet.

References

1. S. D. Jacobs, D. Golini, Y. Hsu, B. E. Puchebner, D. Strafford, W. I. Kordonski, I. V. Prokhorov, E. Fess, D. Pietrowski, and V. W. Kordonski, "Magnetorheological Finishing: A Deterministic Process for Optics Manufacturing," SPIE 2576: Optical Fabrication and Testing, ed. T. Kasai, 372-383, (1995).

2. D. Golini, S. D. Jacobs, W. I. Kordonski, and P. Dumas, "Precision Optics Fabrication Using Magnetorheological Finishing," SPIE CR67: Advanced Materials for Optics and Precision Structures, ed. M. A. Ealey, 251-274, (1997).

18

3. S. D. Jacobs, S. A. Arrasmith, I. A. Kozhinova, L. L. Gregg, A. B. Shorey, H. J. Romanofsky, D. Golini, W. I. Kordonski, P. Dumas, and S. Hogan, "An Overview of Magnetorheological Finishing (MRF) for Precision Optics Manufacturing," ACERS 102: Ceramic Transactions, ed. R. Sabia, V. A. Greenhut, and C. Pantano, (1999).

4. W. I. Kordonski and S. D. Jacobs, "Model of Magnetorheological Finishing," Technomic Publishing Co. Sixth International Conference on Adaptive Structures, ed. C. A. Rogers, J. Tani, and E. J. Breitbach, 63-74, (1996).

5. A. B. Shorey, S. D. Jacobs, W. I. Kordonski, and R. F. Gans, "Experiments and observations regarding the mechanisms of glass removal in magnetorheological finishing," Applied Optics 40(1), 20-33 (2001).

6. A. B. Shorey, "Mechanisms of material removal in magnetorheological finishing (MRF) of glass," (University of Rochester, Rochester, NY, 2000).

7. V. H. Bulsara, Y. Ahn, S. Chandrasekar, and T. N. Farris, "Mechanics of polishing," Journal of Applied Mechanics-Transactions of the ASME 65(2), 410-416 (1998).

8. MetroPro Reference Guide, OMP-0347 (Zygo Corporation, Middlefield, CT).

9. F. W. Preston, "The theory and design of plate glass polishing machines," Journal of the Society of Glass Technology 11, 214 - 256 (1927).

10. M. Buijs and K. Korpel-van Houten, "A Model for Lapping of Glass," Journal of Materials Science 28(11), 3014-3020 (1993).

11. M. Buijs and K. Korpel-van Houten, "Three-Body Abrasion of Brittle Materials as Studied by Lapping," Wear 166(2), 237-245 (1993).

12. T. Izumitani and S. Harada, "Polishing mechanism of optical glasses," Glass Technology 12(5), 131 - 135 (1971).

19

13. D. C. Cornish and I. M. Watt, "The Mechanism of Glass Polishing," Report: R296 (SIRA Institute, Ltd., 1963).

14. A. Kaller, "Properties of polishing media for polishing optics," Glastechnische Berichte-Glass Science and Technology 71(6), 174 - 183 (1998).

15. A. Kaller, "The basic mechanism of glass polishing," Naturwissenschaften 87, 45 - 47 (2000).

16. T. Hoshino, Y. Kurata, Y. Terasaki, and K. Susa, "Mechanism of polishing of SiO2 films by CeO2 particles," Journal of Non-Crystalline Solids 283(1-3), 129-136 (2001).

17. L. M. Cook, "Chemical Processes in Glass Polishing," Journal of Non-Crystalline Solids 120(1-3), 152-171 (1990).

18. M. J. Cumbo, D. Fairhurst, S. D. Jacobs, and B. E. Puchebner, "Slurry Particle Size Evolution during the Polishing of Optical-Glass," Applied Optics 34(19), 3743-3755 (1995).

19. H. Dunken, "Surface chemistry of optical glasses," Journal of Non-Crystalline Solids 129, 64-75 (1991).

20. F. Sugimoto, Arimoto, Y., Ito, T., "Simultaneous Temperature Measurement of Wafers in Chemical Mechanical Polishing of Silicon Dioxide Layer," Japanese Journal of Applied Physics 34, 6314-6320 (1995).

21. Y. Moon and D. A. Dornfeld, "Investigation of the relationship between Preston's coefficient and friction coefficient in chemical mechanical polishing (CMP) of silicon," ASPE 17: Spring Topical Meeting, 78 - 82, (1998).

22. H. Matsuo, A. Ishikawa, and T. Kikkawa, "Role of frictional force on the polishing rate of Cu chemical mechanical polishing," Japanese Journal of Applied Physics 43(4B), 1813 - 1819 (2004).

20

23. J. Luo and D. Dornfeld, "Material removal mechanism in chemical mechanical polishing: Theory and Modeling," IEEE Transactions on Magnetics 14(2), 112 - 133 (2001).

24. J. Luo and D. Dornfeld, "Material removal regions in chemical mechanical planarization for submicron integrated circuit fabrication: coupling effects of slurry chemicals abrasive size distribution and wafer-pad contact area," IEEE Transactions on Magnetics 16(1), 45 - 56 (2003).

25. M. Bielmann, U. Mahajan, and R. K. Singh, "Effect of particle size during tungsten chemical mechanical polishing," Electrochemical and Solid State Letters 2(8), 401-403 (1999).

26. U. Mahajan, M. Bielmann, and M. Singh, "Abrasive Effects in Oxide Chemical Mechanical Polishing," Materials Research Society 566: Mat. Res. Soc. Symp., 27-32, (2000).

27. J. Luo and D. Dornfeld, "Effects of abrasive size distribution in chemical mechanical planarization: modeling and verification," IEEE Transactions on Magnetics 16(3), 469 - 476 (2003).

28. J. C. Lambropoulos, S. D. Jacobs, and J. Ruckmann, "Material removal from grinding to polishing," Ceramic Transactions 102, 113-128 (1999).

29. J. C. Lambropoulos, S. Xu, and T. Fang, "Loose abrasive lapping hardness of optical glasses and its interpretation," Applied Optics 36(7), 1501 - 1516 (1997).

30. J. C. Lambropoulos, F. Yang, and S. D. Jacobs, "Toward a Mechanical Mechanism for Material Removal in Magnetorheological Finishing," OSA 7: Optical Fabrication and Testing Workshop, 150-153, (1996).

21

Chapter 2

Experimental Approach

2.1 Spot taking machine (STM)

This work was carried out using an MRF research platform called the Spot

Taking Machine (STM). Unlike the commercial MRF machines, the STM has only

a z-axis which we use to make MRF spots; we are not able to polish out optics using

the STM. The STM has a peristaltic pump as opposed to a centrifugal pump used in

many commercial MRF machines. An MRF spot is made by partially submerging the

optic, without rotation, into the stiffened MR fluid ribbon, as shown in Figure 2.1.1

The main advantage of using the STM over a commercial machine is our ability to

easily take one spot while conserving space on the surface of a substrate. The

commercial MRF machines are programmed to take four spots in order to obtain

average removal rates for use in their optimization software. This must be manually

overridden to produce a single spot.

Another difference between the STM and the commercial Q22-Y MRF

machine is the magnetic field strength. Using a digital gaussmeter,2 we find that at

low currents the Q22-Y and STM have similar magnetic field strengths. As the

current is increased, the STM has a larger magnetic field than the Q22-Y. For

example, a spot taken at 15 amps on the STM is equivalent in depth to a spot taken at

21.6 amps on the Q22-Y. Magnetic current is shown as a direct read-out on the STM,

whereas this information is not provided to the user on the display screen of the Q22-

Y.

22

Spot time (sec)

Total time touching the ribbon (sec)

1 1.172 2.163 3.164 4.165 5.22

Figure 2.1 – Pictures of STM before and during spot formation.

We have calibrated the STM’s spotting time with a high speed digital

camera.3 The images of the spot formation were taken at 1000 frames per second.

Table 2.1 lists the data that was found by stepping through the frames that were

recorded. The total time is the time between the initial contact and the final contact.

These total time values are used in all peak removal rate calculations for this research.

Table 2.1 – Actual spot times measured using the high-speed camera.

The STM operating conditions are kept the same for all experiments

completed for this research. The wheel speed is 200rpm, the magnetic field is 10A

(1.8kG), the pump speeds are adjusted (range: 95 – 125rpm) to maintain a ribbon

height of 1.6mm and the distance between part surface and wheel surface, or gap is

23

held constant at 1.3mm. This results in a 0.3mm compression of the ribbon for all

spots taken in this work.

2.2 Optical glass substrates

Six different glass types are used for this research. Three of the six glasses

were chosen based on their economic importance: LHG-8, BK-7 and fused silica

(FS). LHG-8 is a phosphate laser glass that is extremely important to large, high peak

power laser systems throughout the world. BK-7 and FS are both popular optical

glasses used in many visible and ultraviolet applications. The remaining three

glasses, FCD-1, EFDS-1 and FD-60 were chosen in order to obtain a set of glasses

with a wide range of mechanical figure of merit4 (E7/6/(Kc Hk23/12)) values. The

glasses chosen, rank ordered by increasing mechanical figure of merit, can be found

in Table 2.1 along with their mechanical properties.

The Young’s modulus values are determined using the pulse-echo method.5

The Vickers microhardness values are measured in-house using 100g load and the

fracture toughness values are calculated using the Evans model6. The errors for the

measurements made in-house are all less than +/- 2.5%. The italicized value for FS

fracture toughness comes from reference 4.

Glass Glass Type

Source and Melt Number E [GPa] Kc [MPa m1/2] Hv [GPa] E7/6/(Kc*Hv

23/12) FS Fused Silica Corning 69 0.75 7.5 4.0

BK-7 Borosilicate Schott 81 0.80 6.0 7.0

FD-60 Titanium Silicate Hoya:20A-3419-39 93 0.69 6.3 8.6

EFDS-1 Titanium Phosphate Hoya:03A-4808-33 96 0.59 5.3 14.1

LHG-8 Phosphate Hoya 62 0.52 3.7 19.6 FCD-1 Fluorophosphate Hoya 73 0.47 4.0 22.1

Table 2.2 Optical glass mechanical properties rank

ordered by increasing mechanical figure of merit.

24

In addition to a large variation in mechanical properties, the optical glass set

has very different optical properties as well. The six glasses are indicated in Figure

2.2 on an optical glass diagram.

Figure 2.2 – Optical glass (nd/vd) diagram. The refractive

index (nd) at 590nm is given on the y-axis and the corresponding Abbe value (vd) is given on the x-axis. The six optical glasses used for this thesis are indicated on the diagram.

All of the substrates are disks that have been prepared in the LLE optical

fabrication shop. Side one was pitch polished to a p-v surface flatness less than 1µm,

p-v surface roughness less than 20nm and rms surface roughness less than 1nm (see

Section 2.3). Side two was left fine ground. The pitch-polished surfaces were

prepared using a Hastilite PO cerium oxide slurry and Gugolz 73 optical polishing

pitch. The sub-surface damage (ssd) is estimated to be less than 2x the p-v surface

roughness for ground surfaces.7 The substrates used for this work were pitch polished

to remove enough material to eliminate all ssd. The substrate diameters varied from

50mm to 62mm, and thicknesses varied from 10mm to 25mm. Only three samples of

25

each of the glass types are used in all of the experiments for this research. When

available the three samples are prepared from the same melt or rod of glass material.

2.3 Metrology

The surface quality of an optic is a defining factor for most optical systems.

In this research we examine the global shape (Zygo Mark IVxp), the surface

roughness (Taylor Hobson CCI 3000) and the surface texture using power spectral

density (PSD) analysis.

The peak removal rate measurements made for this work, described in

Chapter 1, were made using a Zygo Mark IVxp interferometer, shown in Figure 2.3.8

The Mark IVxp interferometer is a four inch HeNe Fizeau interferometer. In order to

get a good measurement the spot must be less than approximately 0.2µm deep. If a

spot is deeper than this, the slopes become too steep for the interferometer to

measure, dropout occurs and no data is collected for that region.

Figure 2.3 – Photograph Mark IVxp interferometer.

The surface microroughness is typically quantified by the root mean square

(rms) of the surface deviations of a surface profile. The rms surface roughness can be

calculated from either a line scan (Rq) or an entire surface (Sq). Equation 2.1 is the

formula for calculating the rms surface roughness from a line scan. In this equation

Ls is the length of the line scan. Equation 2.2 is the equation for determining the rms

26

value for a 3-D topographic image, and Nx x Ny corresponds to the number of pixels

in the image. In both equations the surface topography is represented by z.9

2/1

0

2 )(1⎥⎦

⎤⎢⎣

⎡= ∫

L

s

dxxzL

Rq (2.1)

2/1

1 1

21⎥⎥⎦

⎤

⎢⎢⎣

⎡= ∑∑

= =

x yN

i

N

jij

yx

zNN

Sq (2.2)

These quantitative values are obtained using commercial surface

characterization instruments and are typically the defining surface roughness

specification for an optical material. Although rms measurements are very popular,

the rms alone may not be giving the user the whole picture. PSD is another way to

characterize an optical surface. This measurement technique is not new, but it is

overshadowed by rms surface roughness characterization because the single rms

value is easier to interpret than a PSD plot for many process engineers who require

this information. PSD is defined as the surface height squared per spatial

frequency.10 The area under the PSD curve is equal to the rms value squared for a

given spatial frequency region. The benefit of using PSD to characterize a surface is

that specific spatial frequencies that correspond to features with the largest influence

on the surface topography can easily be identified.10

Surface roughness measurements are made using the Taylor Hobson Talysurf

CCI 3000 non-contact 3D surface profiler, shown in Figure 2.4.11 The CCI 3000 is a

white light interferometer with a digital CCD array. The digital (versus analog) CCD

array is crucial for orthogonal PSD analysis discussed below.12 Measurements are

taken with a 50x Mirau objective to provide a 350 x 350µm measurement area. Four

averages are automatically performed for each measurement made with the CCI

(multiple measurements setting); the purpose of this is to have four phase averages.

Phase averaging increases the signal to noise ratio which reduces system error. Four

27

phase averages are chosen based on personal communications with Maria Robinson

of Zygo Corporation13 and Paul Murphy of QED Technologies12. The instrument is

also calibrated daily with a certified CVD SiC reference surface (≤2Å rms).14 The

calibration procedure subtracts any system errors from each of the measurements to

improve measurement accuracy. The Talysurf CCI has a maximum resolution of

0.1Å in the z-axis and 0.47µm in the x-y axis (maximum optical resolution).

The average rms and p-v surface roughness for each of the spots are

determined by taking the average of five measurement sites within the depth of

deepest penetration (ddp) for each spot. Figure 2.5 indicates the ddp inside an MRF

spot as well as the measurement protocol.

Figure 2.4 – Photographs of the Talysurf CCI 3000.

28

Figure 2.5 – Diagram of an MRF spot and roughness measuring protocol

For this research PSD is calculated using an Excel program and the data are

collected from the areal scans.15 The one-dimensional average PSD plots are

averages of 20 line scans. The line scans are oriented in two separate directions. We

refer to the directions as horizontal (perpendicular to the MRF grooves) and vertical

(parallel to the MRF grooves) as depicted in Figure 2.3. The PSD calculated from the

horizontal line scans allow us to study the amplitude and periodicity of the residual

MRF grooving pattern, whereas the vertical line scans trace the path of the magnetic

CI and non-magnetic nanodiamond particles along the surface of the glass substrate.

2.4 Fluid analysis

The MR fluid properties are very important to the MRF polishing process.

Three of the main properties of interest to us are moisture content, fluid pH and

particle size.

The moisture content is measured using an Arizona Instrument Computrac

Max-1000 moisture analyzer.16 Moisture content measurements are taken before

each set of spots is made. Using the moisture content, the exact CI concentration is

calculated. This value will be discussed later in Chapter 5 as a term in the material

removal rate model.

Flow

0.35 x 0.352 mmareal measurement

Vertical Line Scans

ddp

Horizontal Line Scans

29

The MR fluid pH is also measured before each set of spots is taken. These

measurements are made using a Beckman pH meter17 and Sensorex pH probe18. In

Chapter 5 the MR fluid pH will be related to the chemical durability term used in the

material removal model.

The AcoustoSizer IIs from Collodial Dynamics19 is used to measure particle

size and the zeta potential of the particles in the MR fluid. The AcoustoSizer IIs can

be used with sample sizes as small as 20mL, and with solid concentrations from one

to forty volume percent. We can use it to measure particle size ranges from 0.02µm

to 10µm. The zeta potential of a particle is the charge on the surface of a particle in a

flowing carrier fluid, which changes as a function of pH. By titrating the sample with

acid or base, we can determine the iso-electric point (IEP) of the particle, or the pH

value of the suspension where there is no net charge on the particle. The

AcoustoSizer IIs is discussed in more detail in Appendix A.

The particle size and surface texture of the CI and nanodiamond particles are

examined using a field-emission scanning electron microscope (SEM)20. All samples

are prepared by adhering CI or nanodiamond particles to double sided carbon tape

attached to SEM sample stubs. Images are made at a variety of magnifications and

when possible the samples are not coated with a metal. The in-chamber and in-lens

secondary electron detectors are mixed to obtain all of the images. Images taken with

the SEM will be shown while discussing the MR fluid in the next chapter.

References

1. J. E. DeGroote, A. E. Marino, J. P. Wilson, K. E. Spencer, and S. D. Jacobs, "Effects of nanodiamond abrasive friability in experimental MR fluids with phosphate laser glass LHG-8 and other optical glasses," SPIE 5869: Optical Manufacturing and Testing VI, ed. H. P. Stahl, 121-132, (2005).

2. Digital Gaussmeter Model 9500, F.W. Bell, 6120 Hanging Moss Rd, Orlando FL, 32807.

30

3. MotionPro High-Speed CMOS PCI Camera, Redlake, 3440 E. Britannia Drive, Tucson, AZ 85706, www.redlake.com.


5. A. S. Birks and R. E. Green Jr., "Material Characterization Methods," in Nondestructive Testing Handbook, P. McIntyre, ed. (American Society for Nondestructive Testing), 386-402, (1991).

6. A. G. Evans, Fracture Toughness: The role of Indentation Techniques, Fracture Mechanics Applied to Brittle Materials (American Society for Testing and Materials, Philadelphia, 1979), Vol. 678, pp. 112-135.

7. J. C. Lambropoulos, "What mechanics and materials science can do for the modern optical workshop," APOMA Fall 1999 Workshop, (1999).

8. Mark IVxp Interferometer, Zygo Corporation, Middlefield, CT 06455.

9. T. Vorburger and J. Fu, "In the rough," SPIE's OEmagazine, March 2002, 2002.

10. H. Tobben, G. Ringel, F. Kratz, and D. R. Schmitt, "The use of power spectral density (PSD) to specify optical surfaces," SPIE 2775: Specification, Production, and Testing of Optical Components and Systems, 240-250, (1996).

11. Taylsurf CCI 3000 non-contact 3D surface profiler, Taylor Hobson Inc., Rolling Meadows, IL 60008.

12. P. Murphy, QED Technologies, 1040 University Ave., Rochester, NY (personal communication, October 2004).

13. M. Robinson, Zygo Corporation, Middlefield, Connecticut (personal communication, October 2004).

31

14. Certified CVD Silicon Carbide reference surface, General Optics, Inc. 5390 Kazuko Court, Moorpark, CA 93021.

15. Power spectral density Excel macro, Taylor Hobson Limited, England, 2005.

16. Computrac Max-1000 Moisture Analyzer, Arizona Instrument, 4114 E. Wood St., Phoenix, AZ 85040.

17. pH meter, Beckman Coulter, Inc., 4300 N. Harbor Boulevard, Fullerton, CA 92834.

18. Combination pH Electrode, Sensorex, Garden Grove, CA 92841.

19. AcoustoSizer IIs, Colloidal Dynamics Inc., 11 Knight Street Building E18, Warwick RI 02886.

20. LEO 982 field emission scanning electron microscope, LEO Electron Microscopy is now Nano Technology Systems Division of Carl Zeiss NTS GmbH, One Zeiss Drive, Thornwood, NY

32

Chapter 3

Magnetorheological (MR) Fluid

3.1 Introduction

A typical MR fluid used in the MR finishing process consists of micron sized

magnetic carbonyl iron (CI) particles and non-magnetic polishing abrasives

suspended in an aqueous or non-aqueous carrier fluid.1 The aqueous MR fluid we use

for this research consists of CI particles, non-magnetic nanodiamond particles2, water

and stabilizers. As described in Chapter 1, in the presence of the magnetic field, most

of the CI particles consolidate to form a stiff layer against the rotating wheel. A thin

film of water, nanodiamond abrasives and some residual CI particles are believed to

move above this supporting layer against the part to remove material. Water

introduces chemistry into the removal process. Stabilizers are required for aqueous

fluids because they help reduce corrosion and sedimentation.3

In this chapter the two solid components of the MR fluid are discussed; CI and

nanodiamond particles. We use the AcoustoSizer IIs and scanning electron

microscope (SEM) images to learn more about the particle size, surface charge and

texture of CI particles, the main component in MR fluid. We also perform particle

size and zeta potential measurements on all nanodiamond abrasives. Since these

nanodiamond abrasives are used by us both as a dry powder and in specially prepared

aqueous suspensions, the issue of agglomeration arises and is addressed.

Nanodiamond friability is important to our research, and a discussion of this is given.

Finally, glass material removal is possible using only CI particles (no non-

magnetic nanodiamond particles) in an aqueous based MR fluid, but material removal

rate is significantly increased with the addition of non-magnetic nanodiamond

particles. Experimental removal rate data (with and without nanodiamonds) are

examined in the context of particle size, surface charge (zeta potential) and

nanodiamond friability to generate a need for the extended model.

33

3.2 Carbonyl Iron

Carbonyl iron is formed in a thermal decomposition cycle that begins with

iron pentacarbonyl (IPC) vapor.4 The cycle consists of the deposition of iron, then a

layer of carbon is deposited followed by heating. The carbon film inhibits the

formation of single iron crystals, because the iron cannot adopt the lattice structure

from the iron layer beneath the carbon layer.5 The resulting spherical, non-porous CI

particles have an internal onion-ring structure.4 Particle size is determined by the

length of time that the particles are subjected to the thermal decomposition cycle.5

The CI particles are sifted, milled and screened into various grades.4 Ulicny et al.6

observed with transmission electron microscopy that polycrystalline, mechanically

soft CI particles have grain sizes between 0.02 – 2 µm.

All of the work presented in this thesis was performed using a mechanically

hard grade of CI. Shorey et al. measured the Berkovitch nanohardness of these CI

particles to be 11.7 + 0.8 GPa.7 These CI particles are 2 - 4 microns in median size

and spherical. Figure 3.1 contains images taken with the SEM, as described in

Chapter 2.8 The patches on the surfaces of the CI particles are presumed to be the

silica that was used as a milling agent during the CI production cycle. Energy

dispersive x-ray spectrometry (EDX) performed on the CI particles with the SEM

shows that silicon is one of the elements detected along with iron and carbon. For

comparison, images taken of CI particles not milled with silica are shown in Figure

3.2. There are no patches on the surfaces of these non-silicated CI particles and we

are able to see the underlying CI particle surface texture. We hypothesize that both

the silica particles and the underlying CI particle texture play a role in the MRF

material removal process. It will be shown in Section 3.4 that an MR fluid without

non-magnetic polishing abrasives, consisting of only CI, water and stabilizers, is

capable of removing material on our glass set.

34

Figure 3.1 – SEM images of as received CI particles milled with silica.

Figure 3.2 – SEM images of as received non-

silicated CI particles.

3.2.1 CI particle size and zeta potential

Particle size and zeta potential were measured using the AcoustoSizer IIs.9

Samples are continuously circulated in the AcoustoSizer with a peristaltic pump to

prevent sedimentation. In addition the sample reservoir has a conical shape where

fluid is pulled out through the bottom and re-circulated through the top of the

container to reduce sedimentation. Agglomerates are broken up within the sample

reservoir by a stir bar.

35

The AcoustoSizer mixer was not aggressive enough to break the CI

agglomerates down to their primary particle size. In order to do this we made a 400g

batch of 10wt% CI in DI water with 0.25mL Darvan C10 (dispersing agent). This

batch was mixed with a high shear mixer11 at a setting of 2.5 for 5 minutes.

Following this procedure we found that the shape of the particle size distribution is

Gaussian, with a median particle size of 3.5µm +/- 0.02µm. The particle distribution

plot is shown in Figure 3.3. The particles range in size from approximately 2 – 5µm.

Fused particles are commonly observed in a typical powder sample (see the SEM

inset in Figure 3.3). These are a result of the manufacturing process. Such particles

make up the large end of the particle size distribution, and they do not break apart

with milling or high shear mixing. We do not consider them to be agglomerates.

Rather, they constitute a non-spherical, magnetic CI particle shape within our MR

fluid.

0

2

4

6

8

10

12

2.3 2.5 2.8

3.0 3.2 3.5 3.8 4.1 4.5 4.9 5.3

Particle diameter (µm)

Freq

uenc

y (%

)

Figure 3.3 – CI particle distribution plot. Fused particles (not agglomerates) make up the large end of the particle size distribution.

36

Using the AcoustoSizer we measured the zeta potential of CI as a function of

the pH of the host solution. The CI was placed in three different solutions; DI water,

10-2M KNO3, and MR carrier fluid* (see Figure 3.4 and Table 3.1). All three

suspensions were dispersed for 5 minutes using the high shear mixer prior to

measurement.

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

10

20

0 1 2 3 4 5 6 7 8 9 10 11 12 13

Host Solution pH

Zeta

Pot

entia

l (m

V)

DI Water

10-2M KNO3

MR Carrier Fluid

Figure 3.4 – Zeta potential versus host solution pH for CI.

DI Water 10-2M KNO3 MR carrier fluidpH Zeta Potential Particle Size pH Zeta Potential Particle Size pH Zeta Potential Particle Size

mV µm mV µm mV µm7.23 -82.70 3.60 7.04 -21.00 4.05 12.12 -26.40 5.717.11 -76.10 3.61 6.63 -17.70 3.92 12.06 -26.20 6.236.94 -73.70 3.61 6.38 -15.60 3.95 11.85 -27.00 6.526.97 -76.30 3.62 5.82 -8.50 4.04 11.67 -29.00 6.516.43 -75.10 3.62 5.48 -2.90 4.15 11.10 -27.60 5.986.13 -71.00 3.63 5.05 3.60 4.30 10.59 -27.40 6.195.86 -71.70 3.63 4.81 6.70 4.39 10.26 -28.10 6.155.54 -69.40 3.63 10.10 -27.80 6.005.25 -69.30 3.64 9.83 -27.60 5.744.92 -65.50 3.64 9.56 -27.30 5.804.56 -59.60 3.65 9.30 -27.60 5.504.37 -56.10 3.65 9.02 -27.40 5.734.09 -39.30 3.66

Table 3.1 – Zeta potential and particle size data for CI.

* MR carrier fluid is the liquid portion of the MR fluid (stabilizers dissolved in DI water) that has never been exposed to CI or nanodiamond particles. The pH is approximately 10. Later in Chapters 4 and 5 we discuss MR fluid supernatant. MR fluid supernatant is the liquid extracted from aged MR fluid (CI, nanodiamonds and stabilizers dissolved in DI water) through filtration so that is contains very few to no CI or nanodiamond particles. The pH of MR fluid supernatant is also approximately 10. We used MR carrier fluid as the host solution for all AcoustoSizer measurements described in Chapter 3 (even though the MR fluid supernatant is more similar to the environment during MR finishing), because we wanted to be 100% confident that the only particles circulating in the AcoustoSizer system were the CI or nanodiamond particles that we were attempting to measure. We also discuss in this Chapter AcoustoSizer measurements made on diluted MR fluid which is MR fluid (CI, nanodiamonds and stabilizers dissolved in DI water) that has been diluted with DI water to achieve a solids concentration of 20wt% (approximately 4 times less than the MR fluid solids concentration). It is necessary to dilute the MR fluid in order to measure it using the AcoustoSizer. The diluted MR fluid contains both CI and nanodiamond particles.

37

The natural pH level of neat CI particles mixed with DI water is

approximately neutral (7.2). A CI sample in DI water was subjected to an acid

titration to obtain the range of pH values shown in Figure 3.4, but the iso-electric

point (IEP) was never reached. The zeta potential values are extremely negative

throughout the range of pH values which implies that the CI particles are unlikely to

agglomerate in the DI water host solution.

The IEP of CI was determined while suspended in the 10-2M KNO3 host

solution, and its value is 5.4 (see Figure 3.4). Potassium nitrate (KNO3) is a common

electrolyte solution used for zeta potential measurements.12 Suspensions with zeta

potential values in the range |<30mV| are prone to particle agglomeration (see

Appendix A). Based on the data in Figure 3.4 for the CI particles in the KNO3 host

solution it can be inferred that if one were to work with this solution, in the pH range

shown above, that the particles would agglomerate, especially as the pH approached

5.

The data shown in Figure 3.4 indicate that the zeta potential for the CI

particles in the MRF carrier fluid is stable with respect to the pH range accessible to

titration. The zeta potential values are moderately negative, close to -30mV,

indicating that the CI particles are negatively charged and at a borderline magnitude

to prevent agglomeration.

The plot of CI particle size versus host solution pH in Figure 3.5 indicates that

the CI particles are not agglomerated in DI water, and may be slightly agglomerated

in 10-2M KNO3 especially as the pH approaches the IEP value. The data in the plot

also show that the CI particles in MR carrier fluid are almost twice the size of the

known CI median primary particle size shown in Figure 3.3. This is due to particle

agglomeration in spite of high shear mixing during sample preparation, presumably

because no dispersing agent was present.

38

0

1

2

3

4

5

6

7

0 1 2 3 4 5 6 7 8 9 10 11 12 13Host Solution pH

Part

icle

Siz

e ( µ

m)

DI Water

10-2M KNO3

MR Carrier Fluid

Figure 3.5 – CI particle size versus host solution pH.

3.3 Nanodiamonds

There are many advantages to using nanodiamonds in our study over other

non-magnetic abrasives. Nanodiamonds constitute the abrasive in one of the two

commercially available MR fluids, and they work well on a wide range of materials

such as optical glasses, single crystals (calcium fluoride, silicon) and ceramics (SiC,

WC).13 Nanodiamonds are fairly inert and, therefore, should exhibit fewer

confounding chemical interactions with the glass surface compared to polishing with

oxide abrasives such as cerium oxide that are known to have “chemical tooth”14. The

synthetic nanodiamond abrasives used for this research were supplied as a dry

powder2 and in an aqueous suspension to prevent agglomeration. The nanodiamonds

supplied in an aqueous suspension were from UK Abrasives Inc.15

The term synthetic nanodiamond encompasses many types of nanodiamond

particles such as synthetic monocrystalline diamond (SMD), polycrystalline diamond

(SPD), heat treated synthetic monocrystalline diamond (HTSMD) and ultra dispersed

diamond (UDD).16 In this work we used UDD diamonds exclusively. These UDD

nanodiamonds are also referred to as cluster diamonds. UDD nanodiamonds are

formed from carbon atoms that are contained within the explosive molecules.17 The

resulting nanodiamonds are composed mostly of carbon, oxygen, nitrogen and

hydrogen, but due to their fabrication method they contain other impurities. Many of

39

the impurities can be removed by washing the nanodiamonds with water (i.e. free

elctrolytes, salts) or acid (i.e. metals, oxides). The impurities that cannot easily be

removed are the impurities that are incorporated into the nanodiamond lattice

structure (i.e. silicon, calcium iron and sulfur). These admixture impurities constitute

less than 0.1-0.5 mass% of the UDD nanodiamonds.18

3.3.1 Nanodiamond friability

Friability (how easily a material is broken apart) is a variable that can be

adjusted during the production of UDD nanodiamonds. There are two ways to create

different friability nanodiamonds. The first is adjusting the amount of time that the

carbon is exposed to increased pressure. The longer the pressure is applied, the lower

the friability will be. The difference in exposure time between low and high friability

is on the order of 10-6 – 10-7 seconds. The second method varies the type of crushing

technique. This approach is typically used to make subtle changes in friability for

nanodiamonds produced with the same explosives and the same process time. Two

types of crushing techniques employed are ball milling (i.e. ceramic or iron milling

balls) and jet milling (air or water flow).19 In our work we used three different

friability levels of nanodiamond; low, medium and high. We chose to use different

friability nanodiamonds to see their effect on glass material removal rate and surface

roughness. The differences in behavior observed for the three nanodiamonds in the

MR fluids also allow us to learn more about the nanodiamond role in the MRF

material removal process.

The low friability nanodiamonds are very compact, difficult to break apart and

have rounded edges (see sketch in Figure 3.6). These low friability nanodiamonds

would have a difficult time in removing material, due to a lack of cutting edges. Low

40

friability nanodiamonds are typically used in CMP applications. High friability

nanodiamonds have many defects and jagged edges (see Figure 3.6) and are very

easily broken apart. It has been suggested that the high friability nanodiamonds

would probably break apart immediately when in contact with the glass, but might be

better suited for softer materials.19 Medium friability nanodiamonds have properties

in-between those of low and high friability nanodiamonds. They have enough defects

and jagged edges (see Figure 3.6) to cut off material from the glass surface, but they

offer resistance against immediate fracture and wear.19

Low Medium High

Figure 3.6 – Sketches of nanodiamond (primary particle) shapes.19 The average primary particle size of UDD nanodiamonds is approximately 4nm.17

Many unsuccessful attempts were made to image the individual nanodiamond

particles using the SEM. In order to image in the SEM, particles of differing

friability had to be removed from the aqueous suspension in which they were

supplied. This resulted in particle agglomeration. We washed the samples multiple

times with various solvents (i.e. chloroform) and DI water, but we were not able to

completely remove the residual surfactant on the surfaces. Due to the residual

surfactant and small particle sizes we were not able to obtain clear images of the

individual particles in order to observe changes in surface texture as a function of

friability. Examples of these SEM images are shown later in Figure 3.11 during the

discussion of the UK Abrasive nanodiamonds.

41

Surface area analysis was also considered as another technique for analyzing

nanodiamond friability. Surface area of abrasive particles is typically measured using

the Brunauer, Emmett and Teller or BET method.20-23 In order to have these

measurements made, the samples must be clean dry powders free from all

contaminants. Due to our inability to efficiently remove the nanodiamonds from

suspension and clean off all remaining surfactant, we were unable measure the

specific surface area for the nanodiamonds of varying friability.

3.3.2 Nanodiamond zeta potential and particle size

In this section we describe the zeta potentials and particle sizes of

nanodiamonds used in this work. In section 3.3.2.1 the nanodiamonds supplied in dry

powder form2 are discussed. In section 3.3.2.2 we discuss the nanodiamonds

manufactured by UK Abrasives Inc.15 These latter nanodiamonds were supplied in

aqueous suspensions, and they were provided to us in varying sizes and degrees of

friability.

All AcoustoSizer IIs measurements were made for samples of 1 wt%

nanodiamonds (solids concentration) in the host solution. [The aqueous suspending

solution† for the nanodiamonds in suspension was assumed to be water for calculating

dilution to the 1 wt% level. We did not attempt to separate the nanodiamonds from

the suspension, because when the nanodiamonds are used in the MR fluid, the

aqueous suspension is also added to the MR fluid.] Samples of nanodiamonds were

sonicated24 for 5 minutes instead of being subjected to high shear mixing, because

they are expensive and we had limited quantities. [The high shear mixer requires the

use of large volumes (>500mL).] Once in the AcoustoSizer IIs the samples were

continuously circulated through the system to minimize sedimentation and

agglomeration as described in Section 3.2.1.

† The supplier did not divulge the composition of the aqueous suspending solution, but they sent us samples containing no nanodiamonds and we were able to use this to create similar host solutions without nanodiamonds and perform a background subtraction. Background subtractions of the most similar host solution possible were performed for all samples, for both dry and suspended nanodiamonds.

42

3.3.2.1 Dry nanodiamond powder (NDP)

The synthetic dry nanodiamond powder will be referred to as NDP

nanodiamonds. An SEM image of the NDP nanodiamond material is shown in Figure

3.7. The sample was not coated with a metal; the particles were adhered to carbon

tape. The nanodiamonds are agglomerated and probably aggregated, since the

primary particle size is stated to be 4nm.2, 25 [A primary particle is a particle that is

homogeneously ordered, single domain or single crystal; an aggregate is two or more

primary particles strongly bonded together and difficult to separate; and an

agglomerate is a group of primary particles or aggregates, loosely bonded together

and easily separated.26] It is likely that the agglomerates break up once they are

milled with the CI in the MR fluid, but we have not been able to confirm this. We

attempted to extract NDP nanodiamonds from used MR fluid using a magnetic

separation technique for size and SEM analysis, but we were not able to recover

enough material to examine with either technique. Figure 3.8 gives a particle size

distribution plot for NDP nanodiamond powder (as received) dispersed in DI water to

a concentration of 1 wt%, the resulting pH was 2.8. The median aggregate particle

size (no agglomeration – no bimodal distribution) was determined to be 54nm +/-

3nm and the zeta potential was -21mV.

Figure 3.7 – SEM image of NDP nanodiamond agglomerates.

43

0

1

2

3

4

5

6

3 12 22 32 42 52 61 71 81 91 101 111 120 130 140 150

Particle Diameter (nm)

Freq

uenc

y (%

)

Figure 3.8 – Gaussian particle size distribution of NDP nanodiamonds in DI water at pH 2.8.

Zeta potential and particle size measurements were made for 1 wt% NDP

nanodiamonds in three different host solutions; DI water, 10-2M KNO3 and MR

carrier fluid. No dispersing agent was added to the host solutions and the samples

were all sonicated for 5 minutes prior to measurement with the AcoustoSizer IIs. We

titrated the DI water and KNO3 solutions with hydrochloric acid to attempt to

determine the IEP of the nanodiamonds. The KNO3 solution was also titrated with

sodium hydroxide to extend the pH range. The data is tabulated in Table 3.2.

DI Water 10-2M KNO3 MR carrier fluid

pH Zeta Potential Particle Size pH Zeta Potential Particle Size pH Zeta Potential Particle SizemV nm mV nm mV nm

2.81 -21.10 56 3.09 -13.90 34 11.64 6.40 622.70 -19.70 56 2.94 -13.50 34 11.71 3.20 662.61 -19.50 56 2.73 -13.60 34 11.75 2.20 682.53 -19.70 56 2.56 -12.60 32 11.77 0.80 692.43 -20.10 56 3.61 -14.30 322.34 -21.1 55 4.53 -15.30 33

5.55 -16.70 336.81 -21.40 348.63 -23.80 359.49 -27.20 36

10.49 -30.50 37Average 2.57 -20.20 56 5.49 -18.44 34 11.72 3.15 66St. Dev. 0.17 0.72 0 2.92 6.26 2 0.06 2.38 3

Table 3.2 – Zeta potential and particle size measurements for NDP nanodiamonds in three different host solutions.

44

The results in Figure 3.9 show that the NDP nanodiamonds are moderately

negatively charged in the DI water (-21 mV) and KNO3 host solutions (-12 mV to -15

mV) at low pH. We were not able to determine the IEP of these NDP nanodiamonds

in either of these solutions, since the zeta potentials remained negative throughout the

pH testing range accessible through titration. The zeta potential of the NDP

nanodiamonds in the MR carrier fluid was slightly positive at pH 11.7. This is very

interesting because the CI particles were found to be negatively charged in the MR

carrier fluid at pH 11. From electro-static considerations27, 28 it is possible that an

attractive force exists between these particles when combined in the MR carrier fluid.

-35

-30

-25

-20

-15

-10

-5

0

5

0 2 4 6 8 10 12

Host Solution pH

Zeta

Pot

entia

l (m

V)

10-2 KNO3

DI Water

MR Carrier Fluid

14

Figure 3.9 – Zeta potential versus host solution pH for NDP nanodiamonds.

The median values for NDP nanodiamond particle size as a function of pH are

shown in Figure 3.10 for the three different host solutions. The value in MR carrier

fluid (60nm at pH 11.7) is a close match to the result shown earlier in Figure 3.8 for

DI water (55nm at pH 2.8). The median particle sizes for the NDP nanodiamonds in

KNO3 are 65% lower (33-35 nm for the pH range 2-11). At the high pH end, this

might be ascribed to the moderately negative zeta potentials for these small particles,

but this does not explain the small particle size measured at the low pH end.

45

0

10

20

30

40

50

60

70

0 2 4 6 8 10 12

Host Solution pH

Med

ian

Par

ticle

Siz

e (n

m)

10-2 KNO3

DI WaterMR Carrier Fluid

14

Figure 3.10 – Particle size versus host solution pH for NDP nanodiamonds.

3.3.2.2 UK Abrasive nanodiamonds

The nanodiamonds received from UK abrasives were supplied in suspension

to prevent agglomerates (1.2 – 1.3 wt%). Table 3.1 lists their properties. The particle

sizes in Table 3.1 were measured at LLE with the AcoustoSizer IIs. DI water was

added to the samples to achieve 1 wt% suspensions. UK – Low, UK – Medium A

and UK – High were all intended to be the same size, that being the lowest particle

size offered by UK Abrasives. These fall over a 15nm range from 19 - 34nm. The

UK – Medium B and UK – Medium C were prepared at our request with the intention

of having larger particle sizes compared to UK – Medium A. This was accomplished

with a modest degree of success, falling over a 15nm range from 29 – 44nm. The

right-most column in Table 3.1 lists the largest aggregate particle size observed for

the six UK abrasive nanodiamonds (all data were Gaussian). We find that the order

of increasing largest particle size does not correspond with order of increasing median

aggregate particle size. The smallest nanodiamonds are the UK-Low friability

46

nanodiamonds and the largest nanodiamonds are the UK-Medium C friability

nanodiamonds.

The friability index data given in Table 3.3 were measured by UK

Abrasives15. They would not divulge their technique but told us that the friability

index is based on the percentage change in size after crushing.

Nanodiamond Designation Friability index† Median Particle size 99% particles all smaller than% nm nm

UK - Low 20 19 +/- 2 60UK - Medium A 50 29 +/- 1 82

UK - High 90 34 +/- 1 114UK - Medium B 50 35 +/- 1 98UK - Medium C 50 44 +/- 2 121

†Supplied by vendor Table 3.3 – UK Abrasive nanodiamond properties.

Sample SEM images are shown in Figure 3.11 for the six nanodiamond

products. These images were difficult to obtain, because the nanodiamonds as

received were dispersed with the aid of a surfactant (suspending agent) that created

charging problems. Most of the surfactant was removed by washing the samples with

chloroform, but the procedure was not successful in removing all of the surfactant. It

is difficult to see variations between the different friability samples due to the very

small particle size and agglomeration. All of the samples agglomerated after they

were taken out of suspension to prepare samples for SEM work.

47

(c) UK - High (d) UK – Medium B

(b) UK – Medium A (a) UK - Low

(e) UK – Medium C

Figure 3.11 – SEM images of (a) UK-Low, (b) UK-Medium A, (c) UK-High, (d) UK-Medium B and (e) UK-Medium C nanodiamond agglomerates after removal from suspension, and washing.

48

The particle size distributions of the as-received UK nanodiamonds are shown

in Figure 3.12. The measurements were made with the AcoustoSizer IIs. In general

the median particle size and maximum particle size in the distribution increase with

friability for UK – Low, UK – Medium A and UK - High. The lack of a bimodal

distribution suggests well-adhered aggregates and no agglomerates.

0

1

2

3

4

5

6

0 20 40 60 80 100 120 140

Particle diameter (nm)

Freq

uenc

y (%

)

UK - Low UK - Medium A UK - High UK - Medium B UK - Medium C

Figure 3.12 – Gaussian particle size distribution plot of UK nanodiamonds.

The zeta potential and particle size data for neat UK Abrasive nanodiamonds

(1 wt%) in DI water, 10-2KNO3, and MR carrier fluid are given in Figures 3.13 – 3.17.

We titrated the nanodiamonds in 10-2M KNO3 with hydrochloric acid in an attempt to

determine the IEP value, based on earlier success with the NDP nanodiamonds.

Multiple measurements were made in the DI water and MR carrier fluid host

solutions over a period of 30 minutes circulating in the AcoustoSizer IIs. In what

follows we discuss results for each of the nanodiamonds separately, and then we give

a comparison of results for NDP and UK Abrasives nanodiamonds in section 3.3.2.3.

49

UK-Low nanodiamonds (Figure 3.13 and Table 3.4) exhibit negative charges

in all three host solutions. The natural pH of the UK-Low nanodiamonds in DI water

is 8.2. The surface charge on the UK-low nanodiamond particles is -46mV which

implies that the particles will not be attracted to each other. This is confirmed by the

corresponding particle size measurements that show a median particle size of 22nm,

which is similar to the results shown in Table 3.1. The zeta potential of the UK-Low

nanodiamonds is more negative in the KNO3 solution compared to DI water. With

hydrochloric acid titration the zeta potential becomes less negative for a pH range

between 6 and 2. The nanodiamonds particles agglomerated as expected, as the

environment became more acidic and the zeta potential approached zero. Finally if

we examine the zeta potential and particle size data for the UK-low nanodiamonds in

the MR carrier fluid, we find that the zeta potential is less negative in MR carrier

fluid at pH 10.8 compared to DI water at pH 8.2. Zeta potential values closer to zero

promote particle agglomeration. The particle size data shown in Figure 3.13 support

this by showing that the measured particle size is 50% higher in the MR carrier fluid

host solution compared to that found in DI water.

-90

-80

-70

-60

-50

-40

-30

-20

-10

00 1 2 3 4 5 6 7 8 9 10 11

Host Solution pH

Zeta

Pot

entia

l (m

V)

DI Water

10-2M KNO3

MR Carrier Fl

12

uid

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10 11 12Host Solution pH

Part

icle

Siz

e (n

m)

DI Water

10-2M KNO3

MR Carrier Fluid

Figure 3.13 – Zeta potential and particle size measurements for UK – Low nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.

50


mV nm mV nm mV nm8.28 -47.30 25 6.18 -60.10 27 10.77 -20.20 418.17 -46.60 22 5.91 -59.20 29 10.76 -21.00 438.14 -46.40 21 5.51 -58.70 30 10.77 -22.20 458.13 -46.20 21 5.00 -57.70 32 10.78 -22.70 46

4.51 -55.40 354.03 -40.20 432.05 -19.20 51

Average 8.18 -46.63 22 4.74 -50.07 35 10.77 -21.53 44St. Dev. 0.07 0.48 2 1.41 15.25 9 0.01 1.14 2

Table 3.4 - Zeta potential and particle size measurement data for UK – Low nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions.

The data presented in Figure 3.14 and Table 3.5 give the UK-Medium A

nanodiamond zeta potential and particle size measured as a function of host solution

pH in the three different host solutions. The smallest particle size is measured when

the nanodiamonds are in the DI water host solution (29nm). This result is surprising

based on the zeta potential values of the nanodiamonds in the DI water host solution

(-50mV) compared to the nanodiamonds in the KNO3 host solution (-60mV) where

the particles are approximately 30% larger (44nm). The surface charge become less

negative in the KNO3 host solution with acid titration, but the particle size shows

little sensitivity to these changes. The zeta potential is less negative in the MR carrier

fluid compared to water and the KNO3 host solutions. The zeta potential value of

nanodiamonds in the MR carrier fluid is in a region (|<30 mV|) where the particles are

predicted to agglomerate. The largest measured particle size is obtained when the

nanodiamonds are in the MR carrier fluid host solution (48nm), approximately 40%

larger than the median aggregate particle size.

51

-90

-80

-70

-60

-50

-40

-30

-20

-10

00 1 2 3 4 5 6 7 8 9 10 11 12 13

Host Solution pH

r Fluid

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10 11 12 13Host Solution pH

Part

icle

Siz

e (n

m)

DI Water

10-2M KNO3

MR Carrier Fluid

Figure 3.14 – Zeta potential and particle size measurements for UK – Medium A nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.



6.38 -51.10 31 6.13 -60.60 44 11.79 -10.40 456.43 -50.90 28 6.03 -60.00 45 11.79 -11.50 486.43 -50.60 28 5.51 -59.00 45 11.81 -12.90 50

5.00 -58.40 454.52 -58.20 454.03 -57.00 453.56 -55.10 463.04 -51.60 482.54 -35.20 48

Average 6.41 -50.87 29 4.48 -55.01 46 11.80 -11.60 48St. Dev. 0.03 0.25 2 1.29 7.92 1 0.01 1.25 3

Table 3.5 - Zeta potential and particle size measurement data for UK – Medium A nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions.

The data presented in Figure 3.15 and Table 3.6 give the UK-High

nanodiamond zeta potential and particle size measured as a function of host solution

pH in the three different host solutions. The UK-High nanodiamonds exhibit a

negative charge in the DI water host solution at pH 6.4. The particles are not

expected to agglomerate with themselves in this environment which is shown in the

median particle size value of 29nm which is actually slightly smaller than the

measured median aggregate size for these nanodiamonds as reported in Table 3.1

(34nm). The zeta potential was also more negative for this measurement (-51 mV)

Zeta

Pl (

DI Water

10-2M KNO3

MR Carriem

V)

oten

tia

52

compared to the one indicated in Table 3.1 (-42 mV) which may account for the size

difference. The UK-High nanodiamonds exhibit an approximately 50% decrease in

surface charge as the KNO3 host solution becomes more acidic with titration, but the

value in the most acidic environment is still more negative than -30mV. This implies

a stable solution. The median particle sizes are all larger than the smallest aggregate

size measured. This shows that there was some agglomeration, and it increased

slightly as more acid was added to the solution. The least negative zeta potential

value was measured for the UK-High nanodiamonds in the MR carrier fluid host

solution. The median agglomerate particle size in the MR carrier fluid (45nm) was

approximately 25% higher than the median aggregate particle size (34nm) for the

UK-High nanodiamonds reported in Table 3.1.

-90

-80

-70

-60

-50

-40

-30

-20

-10

00 1 2 3 4 5 6 7 8 9 10 11 12 13

Host Solution pH

uid

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10 11 12 13Host Solution pH

Part

icle

Siz

e (n

m)

DI Water

10-2M KNO3

MR Carrier Fluid

Figure 3.15 – Zeta potential and particle size measurements for UK – High nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.

Zeta

Pl (

DI Water

10-2M KNO3

MR Carrier Fl

mV

)ot

entia

53


mV nm mV nm mV nm6.38 -51.10 31 6.17 -52.90 45 12.06 -17.10 426.43 -50.90 28 5.95 -52.80 47 12.11 -20.20 456.43 -50.60 28 5.55 -52.30 48 12.15 -22.10 47

5.05 -51.30 494.56 -49.90 504.01 -46.90 523.50 -36.40 513.03 -31.00 512.54 -31.80 52

Average 6.41 -50.87 29 4.48 -45.03 49 12.11 -19.80 45St. Dev. 0.03 0.25 2 1.30 9.27 2 0.05 2.52 3

Table 3.6 - Zeta potential and particle size measurement data for UK – High nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions.

The zeta potential data in Figure 3.16 and Table 3.7 show that the UK-

Medium B nanodiamonds have an extremely negative surface (-66mV) charge in DI

water host solution. The corresponding particle size shows no agglomeration with a

median particle size of 35nm which is the same as the median aggregate size of

35nm. The UK-Medium B nanodiamonds were the only UK Abrasive nanodiamonds

for which we were able to determine the IEP in the KNO3 host solution. The IEP

occurred at pH 3.6 (see Figure 3.17 and Table 3.7). The measured particle size for

the nanodiamonds in the KNO3 solution did not vary as much as one might expect

due to the huge variations in surface charge. There was only about a 20% increase in

particle size as the KNO3 solution became more acidic. The zeta potential values for

the UK-Medium B nanodiamonds in MR carrier fluid were in the region where

agglomeration is more likely to occur. We observe an approximately 25% increase in

particle size if we compare the median aggregate nanodiamond particle size to the

UK-Medium B nanodiamonds in the MR carrier fluid host solution.

54

-90-80-70-60-50-40-30-20-10

0102030405060708090

0 1 2 3 4 5 6 7 8 9 10 11 12 13

Host Solution pH

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10 11 12 13Host Solution pH

Part

icle

Siz

e (n

m)

DI Water

10-2M KNO3

MR Carrier Fluid

Figure 3.16 – Zeta potential and particle size measurements

for UK – Medium B nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.



7.81 -65.80 34 7.99 -50.70 41 11.87 -12.70 457.84 -65.90 35 7.54 -49.10 43 11.90 -14.90 487.85 -65.70 36 7.04 -47.40 45 11.92 -15.20 49

6.53 -43.10 465.94 -49.80 475.29 -63.50 474.93 -72.10 484.44 -57.50 494.02 -31.60 493.52 4.90 503.04 38.80 502.54 75.90 50

Average 7.83 -65.80 35 5.24 -28.77 47 11.90 -14.27 47St. Dev. 0.02 0.10 1 1.80 45.19 3 0.03 1.37 2

Table 3.7 - Zeta potential and particle size measurement data for UK – Medium B nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions.

The zeta potential and particle size data for the UK-Medium C nanodiamonds

in the three host solutions are shown in Figure 3.17 and Table 3.8. The zeta potential

was moderately negative for the UK-Medium C nanodiamonds in the DI water host

solution. The corresponding particle size in DI water was the smallest (45nm)

compared to the two other host solutions. It was almost the same as the median

Zeta

Pl (

DI Water

10-2M KNO3MR Carrier Flu

IEP = 3.6

mV

)ot

entia

id

55

aggregate size (44nm) shown in Table 3.3. The surface charge and particle size of the

UK-Medium C nanodiamonds were not as sensitive to an acid titration compared to

other nanodiamonds discussed earlier, but there was very little agglomeration

observed in the KNO3 host solution. The surface charge of the UK-Medium C

nanodiamonds in the MR carrier fluid was less negative compared to the two other

solutions, but the values were still more negative than -30mV. In general the UK-

High nanodiamonds were not prone to agglomeration in any of the three host

solutions.

-90

-80

-70

-60

-50

-40

-30

-20

-10

00 1 2 3 4 5 6 7 8 9 10 11 12 13

Host Solution pH

Zeta

Pot

entia

l (m

V)

DI Water

10-2M KNO3

MR Carrier Fluid

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10 11 12 13Host Solution pH

Part

icle

Siz

e (n

m) DI Water

10-2M KNO3

MR Carrier Fluid

Figure 3.17 – Zeta potential and particle size measurements for UK – Medium C nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions plotted as a function of host solution pH.


mV µm mV µm mV µm7.79 -46.50 44 8.01 -49.60 49 12.11 -23.90 507.81 -46.40 45 7.54 -49.40 50 12.12 -24.90 517.81 -46.30 45 7.00 -49.40 51 12.12 -26.10 51

6.55 -49.10 515.98 -49.40 515.54 -48.80 515.04 -48.20 524.53 -42.60 524.03 -40.30 49

Average 7.80 -46.40 45 6.02 -47.42 51 12.12 -24.97 51St. Dev. 0.01 0.10 1 1.36 3.46 1 0.01 1.10 1

Table 3.8 - Zeta potential and particle size measurement data for UK – Medium B nanodiamonds in DI water, 10-2M KNO3 and MR carrier fluid host solutions.

56

3.3.2.3 Nanodiamond properties in MR carrier fluid environments

In this section we compare the average zeta potential values and average

median particle sizes of the six nanodiamonds suspended in MR carrier fluid. The

data presented in this section have been presented separately in sections 3.3.2.1 and

3.3.2.2.

The data plotted in Figure 3.18 are the average zeta potential values for 1wt%

nanodiamonds suspended in MR carrier fluid. The UK-Low nanodiamond

suspension was the only nanodiamond suspension that did not have a pH value

between 11.6 and 12.1. The lower pH value may be due to different nanodiamond

manufacturing or washing techniques proprietary to the vendor.

All of the zeta potential values in Figure 3.18 are |<30mV|. The NDP

nanodiamonds are the only nanodiamonds to exhibit a positive charge. All of the UK

Abrasive nanodiamonds are negatively charged in MR carrier fluid. With the

exception of the UK-Low nanodiamond suspension, there is a trend that higher pH

values create more negative surface charges on the nanodiamonds independent of

manufacturer, size or friability level when the nanodiamonds are suspended in MR

carrier fluid.

57

-30.0

-25.0

-20.0

-15.0

-10.0

-5.0

0.0

5.0

10.0

10.6 10.8 11.0 11.2 11.4 11.6 11.8 12.0 12.2

Host Solution pH

Zeta

Pot

entia

l (m

V)

UK-Low

Nanodiam

UK-Medium A

UK-High

UK-Medium B

UK-Medium C

NDP

Figure 3.18 – Average zeta potential versus host solution pH for nanodiamonds suspended in MR carrier fluid.

The average median particle size data for the six nanodiamonds suspended in

MR carrier fluid are plotted versus the host solution pH in Figure 3.19. The NDP

nanodiamonds are the largest particle sized nanodiamonds measured in the MR

carrier fluid. The UK-Low, UK-Medium A, UK-Medium B and UK-High have

similar particle sizes within error. The UK-Medium C nanodiamonds exhibit the

largest particle size of the UK Abrasive nanodiamonds in the MR carrier fluid host

solution. It is likely that, as these nanodiamond agglomerates are milled by the CI

particles in the STM operating system, they break down to their original aggregate

size.

58

35

40

45

50

55

60

65

70

10.6 10.8 11.0 11.2 11.4 11.6 11.8 12.0 12.2Host Solution pH

Par

ticle

Siz

e (n

m)

NDP UK - Low UK - Medium A UK - High UK - Medium B UK - Medium C

UK-Low

NDP

UK-Medium A

UK-High

UK-Medium B

UK-Medium C

Figure 3.19 – Average particle size versus host solution pH for nanodiamonds suspended in MR carrier fluid.

3.4 Magnetorheological (MR) fluid behavior in polishing

MR fluids used for this thesis work consisted of CI particles, DI water,

stabilizers and non-magnetic nanodiamond abrasives. MR fluid containing only CI

particles, DI water and stabilizers is referred to as abrasive-free fluid. In order to

maintain consistency, all of our experimental MR fluids contained fixed

concentrations of CI particles, DI water and stabilizers. Our main MR fluid variables

were the types, sizes and amounts of nanodiamonds.

In this section we introduce a small subset of our experimental data to show

how spherical magnetic CI particles alone in the abrasive free MR fluid are capable of

removing glass material, in some instances with good surface smoothing efficiency.

Once irregularly-shaped nanodiamonds are added to the abrasive free MR fluid, the

59

removal efficiency is greatly improved, as is the surface smoothing for some glasses.

We show results and discuss our hypothesis as to why nanodiamond friability is a

very important property for determining the performance of the nanodiamond MR

fluid. We also report zeta potential and particle size results for diluted MR fluids

measured with the AcoustoSizer IIs. (We wanted to measure the properties of the

MR fluid as it is used for polishing, but it is necessary to dilute the MR fluid with DI

water for measurements with the AcoustoSizer.) These data are used to interpret

removal rate data collected on our glass set using the same (undiluted) MR fluids.

3.4.1 Abrasive free MR fluid

In the abrasive free MR fluid the magnetic CI particles act both as the lap and

as the polishing abrasive to remove material. Figure 3.20 (and Table 3.9) gives the

peak removal rate data and corresponding surface roughness values inside MRF spots

made with an abrasive-free MR fluid, compared to results for the initial pitch polished

surface. With the exception of LHG-8, the surfaces inside the spots are comparable

in roughness to the initial pitch polished surface roughness values. In some cases the

surfaces inside the spots are smoother than those of the pitch polished surfaces. The

high rms surface roughness value inside the LHG-8 spot is due to pitting. Surface

interactions between the MR fluid and LHG-8 will be discussed in more detail in

Chapter 6. Additional discussion and data that support the primarily mechanical

nature of MRF with an abrasive free MR fluid are given in Appendix C (section C.1).

60

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

FS LHG-8 BK-7 FD-60 EFDS-1 FCD-1

Ave

rage

are

al rm

s su

rface

roug

hnes

s (n

m)

0

1

2

3

4

5

6

7

8

Pea

k R

emov

al R

ate

( µm

/min

)

Surface roughness inside spot Initial pitch polished surface roughness Peak removal rate

Figure 3.20 – Average areal rms surface roughness (left axis) and peak removal rate (right axis) experimental data for the abrasive free MR fluid spots. Surface roughness values for the initial pitch polished surfaces have been included for comparison.

Glass Peak Removal Rate st. dev. Avg. p-v st. dev. Avg. rms st. dev.µm/min nm nm

FS 1.80 0.05 11.49 3.50 0.87 0.03LHG-8 3.25 0.13 44.86 3.13 3.74 0.59BK-7 3.28 0.06 12.06 2.24 0.83 0.01

FD-60 4.58 0.12 9.56 1.80 0.75 0.02EFDS-1 4.94 0.05 9.09 0.96 0.74 0.04FCD-1 6.46 0.45 11.47 2.22 0.86 0.02

Table 3.9 – Abrasive free MR fluid data.

Throughout this thesis we will use the results from the abrasive free MR fluid

experiments as a basis for comparison with various nanodiamond MR fluids. It is

important to consider the CI particles interaction without nanodiamonds in order to

better understand their continued interaction with the glass surface as nanodiamonds

are added.

61

3.4.2. Nanodiamond effectiveness and friability

Adding nanodiamonds to the abrasive free MR fluid significantly affects the

removal rate. Very small quantities (30mg) of nanodiamonds were added to abrasive

free MR fluids to test our hypothesis that, in the presence of the magnetic field, the

nanodiamonds are trapped between the layer of CI particles and glass surface. There

are three circumstances that would cause the effectiveness of the nanodiamonds to

diminish: if the nanodiamonds continued to be evenly distributed throughout the

fluid, if they were “attached” (i.e. strongly embedded) to each of the CI particles, or if

they settled out in crevices or stagnant zones in the fluid delivery system. In addition

to examining nanodiamonds for their performance as nonmagnetic abrasives in the

MR fluid, we also studied the importance of nanodiamond friability to the removal

process.

Experimental results are shown in a semi-log plot in Figure 3.21 for spots

made with the abrasive free, 0.001-vol% UK-Low, 0.001-vol% High and 0.001-vol%

UK-Medium A friability nanodiamond MR fluids. Results for the abrasive free MR

fluid were discussed in the previous section. We observe an increase of 5 – 250% for

all of the peak removal rates after the addition of nanodiamonds, compared to the

abrasive MR fluid results. In order for such a small amount of nanodiamonds (24 –

138 nanodiamond particles to one CI particle) to have such a significant impact on the

removal process, the nanodiamonds must be transported to the surface of the rotating,

stiffened ribbon in the magnetic field.

With the exception of LHG-8 and FCD-1 (the two softest glasses) the high

friability nanodiamonds show the smallest increase in removal compared to the

abrasive free data. These nanodiamonds break apart very easy. If the nanodiamond

breaks on contact with glass, then it would be more difficult for this particle to

efficiently remove material, especially on harder glasses. The opposite is true for the

low friability nanodiamonds where they are very difficult to break apart and edges are

rounded with abrasion instead of breaking off and exposing fresh, sharp edges. The

low friability nanodiamonds exhibit higher removal rates compared to the high

62

friability nanodiamonds for the four hardest glasses (FS, BK-7, FD-60 and EFDS-1).

In general, the medium friability nanodiamonds exhibit one of the highest removal

rates compared to the two other nanodiamond types. Medium friability

nanodiamonds have properties that are in-between the low and high friability

nanodiamonds. It is assumed that, as they mill and interact with the glass, their edges

break off and expose fresh surfaces that can easily cut through and remove material

for optimum results.

1

10

FS LHG-8 BK-7 FD-60 EFDS-1 FCD-1

Peak

Rem

oval

Rat

e ( µ

m/m

in)

Abrasive free MR fluid UK-Low Friability Nanodiamond MR fluidUK-Medium A Friability Nanodiamond MR fluid UK-High Friability Nanodiamond MR fluid

(11.6)

Figure 3.21 – Peak removal rate values [semi-log] for the six glasses in four MR fluids: abrasive free, 0.001-vol% UK-Low, 0.001-vol% High and 0.001-vol% UK-Medium A. [Peak removal rate data are tabulated in Tables C.1 – C.3 in Appendix C].

Based on these data alone we see that glass properties (i.e. hardness) play a

role in determining peak removal rates. In Chapter 4 we introduce nanohardness

measurements of the near surface hydrated layer of each of the six glasses. Then in

Chapter 5 we take the nanohardness values and other near surface glass properties

and incorporate them into an MRF material removal rate model.

63

3.4.3 Nanodiamond surface charge and glass removal rate

In section 3.3.2 the surface charges of the nanodiamond aggregates were

determined for particles suspended in the MR carrier fluid. These data and the zeta

potential for CI (-30mV) suspended in MR carrier fluid are plotted in Figure 3.22

with experimental removal rate data for five MR fluids: abrasive free, 0.01-vol% UK-

High, 0.01-vol% UK Low, 0.01-vol% UK-Medium A and 0.01-vol% NDP

nanodiamond MR fluid. These results suggest a linear relationship between the zeta

potential of the nanodiamond in MR carrier fluid host solution and peak removal rate.

Linear trend lines have been drawn for each of the glass types and in general they

have good correlations, with the exception of FD-60 and FS. Three ways in which

the zeta potential of the nanodiamond particles might affect the removal process are

1) the particle agglomerate size 2) the zeta potential of the entire MR fluid and/or 3)

the interaction between the CI and nanodiamond particles.

FSR2 = 0.24CL: <75%

BK-7R2 = 0.77CL: 95%

FD-60R2 = 0.11CL: <75%

FCD-1

R2 = 0.54CL: 80%

EFDS-1R2 = 0.53CL: 80%

LHG-8R2 = 0.56

Confidence Level (CL): 80%

0

5

10

15

20

25

30

35 -30 -25 -20 -15 -10 -5 0 5

eta Potential of nanodiamonds suspended in MR carrier fluid host solution (mV)

Pea

k R

emov

al R

ate

( µm

/min

)

CI UK-Low UK-Medium AUK-High NDP

-

Z

Figure 3.22 – Peak removal rate data for abrasive free and 0.01-vol% UK-Low, 0.01-vol% UK-High, 0.01-vol% UK Medium A and 0.01-vol% NDP nanodiamond MR fluids versus surface charge measured for CI and nanodiamond particles in MR carrier fluid. [Data tabulated in Tables C.7 – C.9 in Appendix C]

64

When zeta potential values approach zero, particles are more likely to

agglomerate. The resulting larger nanodiamond particles may create higher removal

rates. In Figure 3.24 we test this hypothesis by plotting the median agglomerate

particle size of the nanodiamond particles measured in the MR carrier fluid versus

peak removal rate. Linear trend lines have been drawn for each glass type and all

have confidence levels lower than 80%. Therefore the high correlation between peak

removal rate and the nanodiamond zeta potential is not due its effect on the

nanodiamond agglomerate size.

FSR2 = 0.003

BK-7R2 = 0.57

FD-60R2 = 0.31

EFDS-1R2 = 0.15

FCD-1R2 = 0.22

LHG-8R2 = 0.17

0

5

10

15

20

25

30

40 45 50 55 60 65

Median Agglomerate Size measured in MR carrier fluid (nm)

Pea

k R

emov

al R

ate

( µm

/min

)

UK-Low UK-Medium AUK-High NDP

Figure 3.23 – Peak removal rate data for abrasive free and 0.01-vol% UK-Low, 0.01-vol% UK-High, 0.01-vol% UK Medium A and 0.01-vol% NDP nanodiamond MR fluids versus nanodiamond agglomerate size measured nanodiamond particles in MR carrier fluid. [Data tabulated in Tables C.7 – C.9 in Appendix C]

Our second hypothesis for why there was a strong correlation between the

charge on the nanodiamonds and peak removal rate is that the nanodiamond surface

charge affected the overall zeta potential of the MR fluid, and consequently affecting

agglomeration and the interaction with the glass surfaces. Using the AcoustoSizer IIs

we measured the zeta potential of diluted nanodiamond MR fluids (CI, 0.01-vol%

65

nanodiamonds, DI water and stabilizers). Sample preparation consisted of diluting

the MR fluid with DI water to a 20 wt% solids concentration, then mixing the fluid by

pouring the solution between two glass beakers continuously until it was placed into

the AcoustoSizer reservoir. The constant agitation by pouring back and forth was

done to avoid sedimentation. No additional steps (high shear mixing or sonification)

were taken to break up any agglomerates. We wanted conditions to be as similar as

possible to those of the MR fluid in the STM delivery system. (Dilution was

necessary due to AcoustoSizer measurement constraints). Figure 3.24 shows the

same peak removal rate data from Figures 3.22 and 3.23 plotted against the zeta

potential of the diluted 0.01-vol% nanodiamond MR fluids. Each of the MR fluids is

labeled with the corresponding nanodiamond near the x-axis. Linear trend lines are

drawn for each glass type and have confidence levels less than 80%. The zeta

potential of the MR fluid does not explain the results we saw in Figure 3.22. The

rank order of zeta potential values are the same for the individual nanodiamonds

compared to the nanodiamond MR fluids, shown in Table 3.10.

FSR2 = 0.0001

BK-7R2 = 0.48

FD-60R2 = 0.49

EFDS-1R2 = 0.20

FCD-1R2 = 0.21

LHG-8R2 = 0.12

0

5

10

15

20

25

30

-56 -55 -54 -53 -52 -51 -50 -49

Zeta Potential (mV)

Pea

k R

emov

al R

ate

( µm

/min

)

UK-Low UK-Medium AUK-High NDP

Figure 3.24 – Peak removal rate versus zeta potential of diluted MR fluid for 0.01-vol% UK-Low, 0.01-vol% UK-High, 0.01-vol% UK-Medium A, 0.01-vol% NDP and abrasive free MR fluid. [Data tabulated in Tables C.7 – C.9 in Appendix C]

66

Nanodiamond

Zeta potential of nanodiamond in MR carrier fluid

Zeta potential of diluted MR fluid

mV mVUK-Low -21.53 -54.80UK-High -19.80 -53.54

UK-Medium A -11.60 -53.44NDP 3.20 -49.62

Table 3.10 – Zeta potential values of UK-Low, UK-High, UK-Medium A and NDP nanodiamond and 0.01-vol% nanodiamond MR fluids (diluted).

Our third hypothesis is that removal is dependent on the attraction/repulsion

of the nanodiamonds to the CI particles. For example if the nanodiamonds have a

slight attraction to the CI particles (i.e. NDP) then the nanodiamonds might coat the

CI particle surface in the removal zone. The CI particle will then act as a miniature

lap pressing the nanodiamonds onto the surface of the glass. The nanodiamonds will

act as a lubricating layer, coating the CI particles and efficiently removing material.

Alternatively, if the CI and nanodiamond charges are similar, the nanodiamonds

could tend to migrate away from the CI particles and congregate in areas in-between

the CI particles in the polishing zone (see sketch of extreme cases in Figure 3.25).

The resulting removal might not be as efficient. Therefore, for the most efficient

removal it is desirable to have the largest surface charge difference possible between

the zeta potentials of the CI and nanodiamonds measured in the MR carrier fluid host

suspension. This last hypothesis seems to explain the results shown in Figure 3.22.

In either scenario, once the MR fluid flows past the removal zone, and out of

the magnetic field, the nanodiamonds will dislodge, if they are weakly adhered to the

CI particle, and homogenously mix back into the MR fluid. It is important that the

nanodiamonds are free particles in order to be available to return to the surface of the

MR fluid ribbon on their next pass through the polishing zone.

67

(a) Glass substrate

CI particle

(b)

nanodiamond

Figure 3.25 – Sketch of nanodiamond/CI interaction hypothesis in polishing zone: (a) Attraction – more efficient removal, (b) Repulsion – less efficient removal. [Not to scale]

In this chapter we have shown that the nanodiamonds used in this work differ

in friability, particle size and surface charge. We have shown some experimental

removal rate data and found correlations between nanodiamond friability,

nanodiamond surface charge and peak removal rate. In the next two chapters we

discuss the near surface mechanical properties of glass and then introduce a material

removal model. This model contains five terms, including terms for the mechanical

property of glass, nanodiamond particle size and the pH of the MR fluid.

References

1. S. D. Jacobs, S. A. Arrasmith, I. A. Kozhinova, L. L. Gregg, A. B. Shorey, H. J. Romanofsky, D. Golini, W. I. Kordonski, P. Dumas, and S. Hogan, "An Overview of Magnetorheological Finishing (MRF) for Precision Optics Manufacturing," ACERS 102: Ceramic Transactions, ed. R. Sabia, V. A. Greenhut, and C. Pantano, (1999).

68

2. S. D. Jacobs, "Nanodiamonds Enhance Removal in Magnetorheological Finishing," Finer Points 7, 47-54 (1995).

3. S. D. Jacobs, W. I. Kordonski, I. V. Prokhorov, D. Golini, G. R. Gorodkin, and T. D. Strafford, "Magnetorheological Fluid Composition," Patent No. 5,804,095, September 8, 1998.

4. J. E. Japka, "Microstructure and Properties of Carbonyl Iron Powder," Journal of Metals 40(8), 18-21 (1988).

5. L. B. Pfeil, "The Nature, Properties, and Applications of Carbonyl-Iron Powder," The Iron and Steel Institute Symposium on Powder Metallurgy: Section C - Magnetic Powders and Products, 47-51, (1947).

6. J. C. Ulicny, M. P. Balogh, N. M. Potter, and R. A. Waldo, "Magnetorheological fluid durability test - Iron analysis," Materials Science and Engineering A 443, 16 - 24 (2007).

7. A. B. Shorey, K. M. Kwong, K. M. Johnson, and S. D. Jacobs, "Nanoindentation hardness of particles used in magnetorheological finishing (MRF)," Applied Optics 39(28), 5194-5204 (2000).


9. AcoustoSizer IIs, Colloidal Dynamics Inc., 11 Knight Street Building E18, Warwick RI 02886.

10. Darvan C, Ammonium Polymethacrylate and water, R.T. Vanterbuilt Company, 30 Winfield St. Norwalk, CT 06855.

11. T50 basic ULTRA-TURRAX, IKA Works, Inc., 2635 North Chase Parkway SE, Wilimington, NC 28405.

69

12. R. J. Hunter, Foundations of Colloidal Science, 2nd ed. (Oxford University Press, Oxford, 2001).

13. QED Technologies, 1040 University Avenue, Rochester, NY 14607 USA. www.qedmrf.com.


15. UK Abrasives Inc. 3045 Mac Arthur Blvd. Northbrook, IL 60062.

16. N. Tumavitch, "Diamond dynamics," R&D Magazine 47(11), 35-35 (2005).

17. O. A. Shenderova, V. V. Zhirnov, and D. W. Brenner, "Carbon nanostructures," Critical Reviews in Solid State and Materials Sciences 27(3-4), 227-356 (2002).

18. V. Y. Dolmatov, "Detonation synthesis ultradispersed diamonds: properties and applications," Russian Chemical Review 70(7), 607-626 (2001).

19. V. Slobodsky, UK Abrasives Inc. 3045 Mac Arthur Blvd. Northbrook, IL 60062 (personal communication, September 5, 2006).

20. R. Domesle, U. Gobel, L. Mussmann, E. Lox, and T. Kreuzer, "Catalyst Material," Patent No. US 6,475,951 B1, November 5, 2002.

21. H. L. Zhu, D. E. Niesz, V. A. Greenhut, and R. Sabia, "The effect of abrasive hardness on the chemical-assisted polishing of (0001) plane sapphire," Journal of Materials Research 20(2), 504-520 (2005).

22. Quantachrome Instruments, Nova Series Rapid BET Surface Area Analyzer and Pore Size Analyzer System, 1900 Corporate Drive, Boynton Beach, FL 33426.

23. B. Palosz, E. Grzanka, C. Pantea, T. W. Zerda, Y. Wang, J. Gubicza, and T. Ungar, "Microstructure of nanocrystalline diamond powders studied by powder diffractometry," Journal of Applied Physics 97(064316), 1-6 (2005).

70

24. Branson Ultrasonics Corporation Model 250 Sonifier, 41 Eagle Road, Danbury CT 06810.

25. J. E. DeGroote, H. J. Romanofsky, I. A. Kozhinova, J. M. Schoen, and S. D. Jacobs, "Polishing PMMA and other optical polymers with magnetorheological finishing," SPIE 5180: Optical Manufacturing and Testing V, ed. H. P. Stahl, (2003).

26. Definitions taken from “Fundamentals of Particle Sizing”, Tech Note (7/94) from Nanophase Technologies Corp., 453 Commerce St., Burr Ridge, Il 60521.

27. J. C. Lambropoulos, Rochester, NY (personal communication, March 27, 2007).

28. J. F. Hughes, Electrostatic Powder Coating, Electrostatics and Electrostatic Application Series (Research Studies Press, Ltd., Letchworth, Hertfordshire, England, 1984).

71

Chapter 4

Glass modified surface layer

4.1 Introduction

Glass is a three dimensional matrix consisting of silicon and oxygen atoms,

and depending on the glass composition additional modifier ions. In silicate glasses,

and in the presence of water (liquid or vapor), the H2O molecules interact with the Si-

O-Si chains. This reaction is summarized in Equation 4.1.1

Si-O-Si + H2O → 2Si(OH) (4.1)

This modified surface layer is commonly referred to as the glass hydrated

surface layer. The thickness of this layer has been predicted to be between 10Å and

several thousand angstroms depending on the glass composition and environmental

conditions. Its rate of formation is due to the chemical durability of the glass.

Cornish and Watt2 did not believe this layer was a distinct film, but rather a gradual

change in composition from the surface of the substrate into the bulk material. They

also believed that the formation of the hydrated layer advances as quickly as the

material is being removed by conventional polishing. A rate of 130 Å/s was

proposed.

Izumitani3 concluded that the conventional polishing rate of glass was

determined by the rate which the hydrated layer formed and by the hardness of the

hydrated layer. The optical glass set that Izumitani worked with consisted of mostly

silicate and borate glasses over a large range of Vickers microhardness values.

Izumitani4 found an inverse relationship between the conventional polishing

rate in aqueous ceria slurry and a Vickers microhardness measured on a hydrated

layer. This layer was formed by leaching each of the substrates with 0.1N

hydrochloric acid for 60 minutes. This experiment proved to him that the softer the

72

hydrated layer the easier it was for the polishing abrasive particles to remove

material, resulting in a higher polishing rate.

Many people have tried to learn more about the thickness, hardness and

formation rate of the glass hydrated layer.5-18 Doremus et al.5 used a nuclear reaction

technique (Dynamitron linear accelerator) to examine the hydrogen and alkali profile

of the hydrated layer in several silicate glasses. The samples were 40mm x 10mm x

1mm and they were polished on both sides with an aqueous 1µm alumina slurry. The

samples were etched with 5% HF for 3-4 minutes after polishing to ensure that any

residual hydrated layer from the polishing process had been removed. Three of the

glasses were placed in 600mL of pH 5.5 buffer solutions at 50oC or 90oC. The

remaining five glasses were placed in several liters of distilled water at 90oC. The

amount of time that the samples were hydrated varied for each experiment. They

found that the hydrated layer was formed on the less durable alkali silicate glasses,

but that no layer was found on the more durable glasses. The hydrogen depth results

varied from 0.05µm – 1.0µm for the glasses where the hydrated layer was observed.

Tomozawa et al.6 used an IR spectrometer to examine water diffusion in

commercial borosilicate glass. The sample sizes were 20mm x 20mm by 30µm to

50µm thick. The samples were subjected to constant water vapor pressure (355 mm

Hg) at elevated temperatures (200oC – 500oC). After the surface treatment the

transmission of each sample was measured with an IR spectrometer over a range of

4000 – 2000cm-1. They found two absorption peaks at 3600cm-1 and 2750cm-1 which

were used to evaluate the water concentration using Beer’s Law as shown in Equation

4.2, where ε [liters/(mol⋅cm)] is the extinction coefficient, I is the transmitted

intensity at the absorption peak, I0 is the absorption-free transmitted intensity, C

[moles/liter] is the average water concentration in the substrate and dt [cm] is the

thickness of the substrate.

tdCII ε−=100 (4.2)

73

The diffusion coefficient was determined by comparing the experimental water

uptake curve, Mt, shown in Equation 4.3, with the theoretical water uptake curve to

give the relationship shown in Equation 4.4.

(4.3) dCCdM t 0−=

21

))((4 πDtdMMt ∞= (4.4)

C0 is the initial water concentration in the sample, M∞ is the saturated water uptake

value, t is the effective diffusion time and D is the diffusion coefficient. They

experimentally determined a water diffusion coefficient of 2.8 x 10-7 e[-52(kJ/mol)/RT]

(cm2/s) into silica at high temperatures.

Tomozawa et al.7 also used IR spectroscopy to examine a Knoop hardness

indent that had been made on silica (Surprasil-W1) that was submerged in water. A

small IR beam diameter of ~10µm was used and the amount of water absorbed was

observed along the main diagonal of the indent. They found that the amount of water

diffused into the silica was proportional to the deformed volume. The depth of

diffusion was estimated to be on the order of a few microns. The Knoop hardness of

silica was found to decrease with increasing loading time for the measurement made

in water. The hardness was independent of loading time for measurements made in

benzene, acetone, acetonitrile, formamide and hydrazine. They also made an indent

on silica in a dry environment and then exposed it to water; their infared results did

not show a water absorption peak for this indent. They believe that water diffuses

into the glass only during indentation and the diffusion causes the hardness of the

surface layer to decrease. They hypothesize that this is why polishing glass in

aqueous environments is more efficient than polishing in other non-aqueous

environments.

Nakashima et al.10 used micro Fourier Transform Infrared (FT-IR)

spectroscopy to examine the water content in geological samples taken from the

Yanazawa-Kamimura area near the Median Tectonic Line in Japan. They found that

they needed to have the sample measurement housing purged with dried air in order

to avoid moisture fluctuations.

74

The same group analyzed the hydrated layer of silica glass under extreme

conditions using the same micro FT-IR technique a few years later. The silica

samples were subjected to pure water at temperatures of 400-500oC and at pressures

of 40-96 MPa for extended periods of time. They found that the hydration layer

increased during initial hydration but reached a steady state thickness of 0.1mm after

tens of hours.11

Maaza et al.12 observed a 60Å thick layer on borosilicate glass designed for

neutron optics applications using grazing angle neutron reflectometry (GANR). The

samples were polished with a felt polisher for two hours with cerium oxide slurry and

then cleaned with 50oC RBS® detergent solution, 25oC water and ethyl-alcohol. The

surface layer was less dense than the bulk sample and it was hypothesized to be

caused by both mechanical and chemical effects which include an interaction with

water to form a hydrated layer, creep at the surface due to rubbing of the polishing

tool with the glass surface, mechanical molecular erosion of the surface and ionic

exchanges with the polishing slurry. In addition they also thought that increased local

pressure exerted by the polishing abrasives would promote the growth of a hydrated

surface layer. Since this reaction would cause densification19, it would not cause the

surface layer to become less dense, which is contrary to what they found.

SEM13 and transmission electron microscopy (TEM)14 have also been used to

observe a hydrated layer in glass or geological materials. Ballif et al.13 were

interested in determining how borosilicate glass fibers reacted to saline solution.

They built a device that allowed them to leach fibers at a slightly elevated

temperature (37oC) in saline solution. The fibers were sliced and the cross-sections

were examined using the backscattered electron detector on an SEM. They were able

to see an unaltered core surrounded by the hydrated layer. With increasing time

exposed to the saline solution they found that the diameter of the unaltered core

decreased as the hydrated layer increased. The observed hydrated layer thicknesses

varied from 0 to 2µm.

75

Kawano et al.14 observed a hydrated layer on volcanic material from the

Sakurajima volcano in Japan. They found that the naturally weathered surfaces had

<0.1µm thick layers with no structure. Energy dispersive X-ray (EDX) analysis was

also performed on these layers and it was found that they had less silicon and more

aluminum than the bulk material.

Ellipsometry is another method researchers have used to study glass hydrated

layers.15-17 The measurements made by Yokota et al.15 were performed in desiccated

air. This was done because prior results showed that the measurements were sensitive

to any humidity in the ambient environment. They found that the thicknesses of the

hydrated layer on a variety of glasses (flints, crowns, silicas and Pyrex) after pitch

polishing with a ceria suspension for 90 minutes were on the order of 10 - 50nm. The

thickest layers were on silica glass and Pyrex and the thinnest layers were on the flint

glasses. They also reported that the index of the hydrated layer was lower than that of

the bulk material for the less chemically durable glasses and higher than the bulk

material for the more chemically durable glasses. They believe that two factors

control the difference in index between the surface layer and the bulk material: high

local pressures exerted by the polishing abrasive particles which will increase the

index of the surface layer through densification of the glass, and the leaching action

of water which will decrease the index by extracting metal ions out of the surface

layer. The change in index will depend on the balance of these two factors.

Ji et al.18 were interested in the mechanical behavior of borosilicate glass after

exposure to an aqueous environment. Polished borosilicate samples were exposed to

an aqueous solution that was enriched with silicon, boron and sodium. Vickers

microhardness indents were made on the surface before and after the samples were

subjected to the aqueous environment. It was found that the fracture toughness

decreased approximately 20% for the surface exposed to the aqueous environment for

56 days compared to the pristine surface.

Cumbo20 measured the Vickers microhardness of a variety of optical glass

samples immersed in aqueous fluids of pH 4 and 10, deionized (DI) water (pH 7) and

76

methanol environments. He found that the pH had no measurable effect on the

hardness and fracture toughness of the glasses measured. He also observed that the

hardness and fracture toughness were higher in the methanol environment compared

to the DI water for most of the glasses. His experimental results suggest that the near

surface hardness decreased with exposure to DI water during microindentation for

silica-rich glasses (FS and BK-7), but that it increased for lead-rich glasses (F7 and

SF6).

Park et al.21 examined the surface of silicon using nanoindentation with a

Berkovitch indenter. They used potassium hydroxide (KOH) to hydrate the silicon

surfaces. Four different concentrations of KOH were used: 0wt%, 5wt%, 10wt% and

15wt%. The results showed that the hardness decreased as a function of KOH

concentration.

It has been confirmed through various analytical techniques1-21 that the

introduction of water creates a hydrated or modified surface layer on glass. This

surface layer has been found to have lower hardness values than the bulk layer2, 3, 7, 12,

18, 20, 21, depending on glass composition and the properties of the exposure medium5,

6, 15. Many literature sources3, 6, 7, 12, 16-18, 20, 21 believe that it is the creation of the

softened hydrated layer that promotes easier material removal (i.e. without

scratching) using aqueous polishing techniques. We also believe that glass surfaces

hydrate and soften with exposure to the aqueous MR fluids, and that each glass

responds differently to the MR fluid environment.

We chose to use the nanoindentation technique to study the hydrated layers of

our glass set. Nanoindentation measurements allow us to measure the nanohardness

and Young’s modulus of dry surfaces, and to compare these results with those taken

on glass surfaces exposed to MR fluid supernatant. From this comparison we can

determine the depth of the hydrated layer for our glasses. Hardness and Young’s

modulus are two terms in the existing MRF material removal rate model.22 Our work

is unique in that we will measure these properties for our glasses in the environment

to which they are subjected during MR finishing.

77

4.2 Nanoindentation in fluid environments

Our goal is to simulate the conditions during MRF polishing on the surface of

glass and measure the nanohardness as a function of depth for these conditions. For

this work we use the Nano Indenter XP and Testworks 4 software23, shown in Figure

6.1. A magnetic coil is used to apply the load with the Nano Indenter XP and a

capacitance gauge is used to measure the displacement. The load resolution is 50nN

and the displacement resolution for the instrument is <0.01nm. We perform all

measurements using a Berkovitch pyramidal diamond indenter. The Testworks 4

software calculates the nanohardness and Young’s modulus using the Oliver and

Pharr model.24

Figure 4.1 – Photograph of the Nano Indenter XP.

We use the continuous stiffness measurement (CSM) technique for all of our

nanoindentation work. Stiffness is continuously measured as a function of

indentation depth into a sample during CSM. The nanohardness (material’s

resistance to plastic flow) and Young’s modulus (material’s resistance to elastic

deformation) values are calculated from the stiffness values which allow the user to

acquire both values as a function of depth into the surface. Due to Berkovitch tip

blunting (or imperfection), the data at shallower depths must be ignored. Each tip is

unique, and the cutoff value is determined by plotting load divided by stiffness

squared versus displacement into the surface. The point at which the load divided by

78

stiffness squared values level off are considered the depth at which the data can be

considered valid.25 For our work we will consider all values valid for depths greater

than 40nm; all measurements made at less than 40nm are discarded.

The indentation size effect (ISE) can be seen in some of the data presented in

the nanohardness versus indentation depth plots, where the hardness of material

decreases as the depth into the surface increases.26 This effect typically occurs for

hardness measurements made at depths less than 10µm. ISE has been seen in both

low load microindentation and nanoindentation.27,28,29,30,31 We do not believe that the

ISE modifies any of the conclusions we draw based on the data we report here,

because a) we compare the near surface hardness and Young’s modulus values

measured in different environments for a given glass type and the same depths and b)

we make glass to glass comparisons.

Sample preparation consists of placing the glass samples in an oven in air at

115oC for 30 minutes to bake out the surfaces, allowing them to cool in the oven and

then immediately placing them in sealed boxes containing desiccant. This procedure

is performed on each sample before each nanoindentation test to ensure that we start

with dry surface layers. In order to measure the nanohardness in a liquid

environment, sample dimensions must not exceed 21mm in diameter and 5mm in

thickness. The areal surface roughness values of all samples are less than 5 nm rms

and 50nm p-v.

CSM nanoindentation is made for each of the glasses in three different

environments: dry, DI water and MR fluid supernatant (pH 10.5). Figure 4.2 shows a

drop of supernatant being placed on the surface of BK-7. It is important to have a

thin layer of liquid and not a droplet. In order for this to happen the surfaces are

cleaned with acetone prior to drying and then if needed a corner of a Kimwipe is used

to evenly distribute the droplet across the surface. The liquid is introduced

approximately one minute before the measurement is made. The liquid does not need

to be replenished because the measurement process only takes approximately ten

minutes and the liquid does not evaporate in that short amount of time. We clean the

79

liquid off with Kimwipes in-between measurements in order to view the surface with

a microscope objective and select the next measurement site. A fresh drop of liquid is

applied after the site has been located and the procedure is repeated. The time

between measurements is less than five minutes. All of the experiments are

performed in an ambient environment (23 +/- 1oC).

We repeat the measurement four times on each surface. The data we report in

nanohardness and Young’s modulus versus depth plots are averages of four CSM

measurements. Comparisons of the nanohardness values for the different hydrated

layers are made at a depth of 60nm. This value is chosen because it is larger than the

largest p-v surface roughness value, it is in the valid data region as defined by Oliver

ad Pharr25 and it is in the modified layer region according to various literature

sources.2, 5, 10, 12, 13

Figure 4.2 – Photograph of a drop of supernatant applied to the surface of BK-7.

80

4.2.1 Nanohardness of LHG-8 in an Ethylene Glycol environment

In 1977 Brown et al.32 determined the absolute weight loss of phosphate laser

glass LHG-7 (and three additional glasses) in ethylene glycol (EG) solutions varying

from 100% EG to 100% DI water (see Figure 4.3). They found that increasing

amounts of water in the EG solution created higher weight loss. We expect that if we

measure the nanohardness of the phosphate glass, LHG-8, in EG solutions the

nanohardness will decrease with increasing water concentration.

Figure 4.3 – Absolute weight loss in EG and DI water solutions for three phosphates and one silicate glass composition. [Ref. 32 with permission].

LHG-7 and LHG-8 are high-power phosphate laser glasses produced by

HOYA Corporation. Based on product literature33 the glasses have similar

compositions and their chemical durability is the same in water. We measure and

compare the nanohardness of LHG-8 in four different environments: dry, 100% EG,

50% EG/50% DI water and 100% DI water (see Figure 4.4).

81

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

10 100 1000Displacement (nm)

Nan

ohar

dnes

s (G

Pa)

100% DI Water

100% Ethylene Glycol (EG)

50% EG 50% DI water

Dry

Environment Average (GPa) st. dev.Dry 6.5 0.3

100% EG 6.2 0.550% EG 50% DI Water 6.0 0.1

100% DI Water 5.7 0.1

Nanohardness at 60nm

Figure 4.4 – Average CSM nanohardness data for LHG-8 in EG and DI water solutions plotted on a semi-log plot. [Inset: average nanohardness value at 60nm depth]

We observe a softened surface layer on LHG-8 (see Figure 4.4). The

measurements made in DI water exhibit the lowest nanohardness values to a depth of

approximately 700nm. The values at 60nm show a distinct softening of the surface

layer as a function of water concentration. This agrees with our hypothesis based on

Brown’s data.32 The error for the measurement made in 100% EG is higher than that

for the other measurements. This is most likely due to the higher viscosity of

ethylene glycol (19.9cP at 20oC)34 compared to water (1.0cP at 20oC)34, creating more

resistance between the indenter and the surface. We conclude from this data that a

softened hydrated layer exists on LHG-8 when it is exposed to water, compared to the

original dry surface.

82

4.2.2 Young’s Modulus in fluid environments

Young’s modulus test results for the six glass types from the CSM technique

in dry, DI water and MR fluid supernatant environments are listed in Table 4.1.

Included in the tables are measurements that were made on the bulk material using

the pulse-echo method35. The bulk material measurements are discussed in more

detail in Chapter 5. The values are given here as a reference for qualitative

comparison.

We observe a 9% increase in the near surface Young’s modulus value for FD-

60 after exposure to DI water compared to the dry surface layer. DI water exposure

does not appear to affect the near surface elastic properties of EFDS-1 and FCD-1.

FS, BK-7 and LHG-8 all exhibit reduced near surface Young’s modulus values (by 3

– 12%) after exposure to DI water during nanoindentation.

Young's ModulusGlass Bulk Dry DI Water MR Supernatant

[GPa] [GPa] [GPa] [GPa]FS 69 +/- 1 81 +/- 1 74 +/- 1 78 +/- 1

BK-7 81 +/- 2 97 +/- 2 94 +/- 6 85 +/- 2FD-60 93 +/- 2 100 +/- 2 109 +/- 1 99 +/- 2

EFDS-1 96 +/- 2 105 +/- 4 108 +/- 5 108 +/- 4LHG-8 62 +/- 1 67 +/- 2 59 +/- 1 55 +/- 2FCD-1 73 +/- 1 85 +/- 3 87 +/- 4 81 +/- 5

Nanoindentation

Table 4.1 – Young’s modulus data measured on the bulk glass (pulse-echo method) and at a depth of 60nm into the surface (nanoindentation) in three different fluid environments.

The data in Table 4.1 show that a modified surface layer is created by the MR

fluid supernatant on most of the glass surfaces, making them less resistant to elastic

deformation than dry surfaces or surfaces exposed to DI water. The reduction is from

1 - 18%. The only exception is EFDS-1, which also has the highest Young’s modulus

value in our glass set. For EFDS-1 the Young’s modulus does not change with

exposure to different fluid environments.

83

4.2.3 Nanohardness in fluid environments

Nanohardness test results for the six glass types from the CSM technique in

dry, DI water and MR fluid supernatant environments are listed in Table 4.2.

Included in the table are measurements that were made on the bulk material using the

Vicker’s microindentation36 (microhardness) technique. The bulk material

measurements will be discussed in more detail in Chapter 5. The values are given

here as a reference for qualitative comparison.

We observe an increase in nanohardness for FD-60 by ~9% after exposure to

DI water for one minute compared to the dry surface layer. We are not the first to

observe this result. Cumbo20 found that SF6 had near surface hardening (increase of

3%) when exposed to DI water during microindentation. SF6 is the Schott37

equivalent to Hoya’s33 FD-60.

BK-7, EFDS-1 and FCD-1 nanohardness values measured in DI water are

similar to the dry surface test results. We are not able to distinguish differences

between the test results for the dry surface and the surface immersed in DI water for

these three glasses.

FS and LHG-8 exhibit reduced nanohardness values from 12 – 34% with

exposure to DI water. It is interesting that these are the hardest and the softest glasses

in our set, and yet they are the most sensitive to water. We believe this occurs

because pure silica networks (i.e. FS) are easily attacked by water1, and the phosphate

glass former in LHG-8 is known to have very poor chemical durability in aqueous

environments32.

84

Table 4.2 – Microhardness data measured on the bulk glass (Vickers) and nanohardness measured at a depth of 60nm into the surface (nanoindentation) in three different fluid environments.

HardnessGlass Bulk Dry DI Water MR Supernatant

[GPa] [GPa] [GPa] [GPa]FS 7.5 +/- 0.4 11.4 +/- 0.0 9.1 +/- 0.5 10.0 +/- 0.2

BK-7 6.0 +/- 0.3 8.6 +/- 0.3 8.9 +/- 0.1 8.5 +/- 0.2FD-60 6.3 +/- 0.3 9.1 +/- 0.4 9.9 +/- 0.4 8.3 +/- 0.4

EFDS-1 5.3 +/- 0.3 8.4 +/- 0.4 8.8 +/- 0.4 8.8 +/- 0.1LHG-8 3.7 +/- 0.2 6.5 +/- 0.3 5.7 +/- 0.1 4.4 +/- 0.1FCD-1 4.0 +/- 0.2 6.4 +/- 0.2 6.7 +/- 0.2 5.5 +/- 0.5

Nanoindentation

The data in Table 4.2 show that the modified surface layers produced by

introducing the MR fluid to most of the glass surfaces are softer than the dry surfaces.

[The only exception is EFDS-1. For this glass the nanohardness does not change with

exposure to different fluid environments. We do not find evidence of a softened

(modified) layer on EFDS-1 at depths greater than 40nm. If the MR fluid supernatant

does produce a softened layer it occurs at depths smaller than 40nm. This is

discussed more in section 5.5.]

The nanohardness increases gradually as the depth increases. A well-defined

boundary does not exist between the modified layer and the bulk material. The

nanoindentation data for the six glasses are shown in Figure 4.5 as a function of depth

into the material. LHG-8 is the most sensitive of all of our glasses; the nanohardness

never rises to that of the dry surface over the tested depth range of 1µm. We

hypothesize that, because the nanoindent is made with the surface being covered with

MR fluid supernatant, the hydrated layer continues to re-establish itself as the

nanoindenter penetrates deeper into the glass. We observe this for BK-7 as well.

Based on the data in Figure 4.5 we estimate that the initial modified layer depth for

both LHG-8 and BK-7 is approximately 200nm by observing where the dry (shaded)

and MR fluid supernatant (solid) lines first touch. We observe a softened modified

85

layer of approximately 80nm for FS. We observe similar results for the remaining

three glasses, all of which have softened modified layers of 100nm or less. These

results agree well with Cornish and Watt’s2 hypothesis on the depth (1-1000nm) and

formation rate (13 nm/sec) of the modified surface layer. This information regarding

the nanohardness and Young’s modulus of the modified surface layers is applied in

Chapters 5 and 6 where our MRF material removal rate model is introduced and we

discuss the surface texture produced inside an MRF spot.

3

4

5

6

7

8

9

10

11

12

10 100 1000Depth (nm)

Nan

ohar

dnes

s (G

Pa)

LHG-8

FS

FD-60

BK-7

Figure 4.5 – Average CSM nanoindentation data for all six glasses. The measurements were made on the dry surfaces (shaded lines) and then again with the glass surfaces covered with a layer of MR supernatant (solid lines).

EFDS-1

FCD-1

Dry (D)

Supernatant (S)

DS

D

D

S

S D S

D

S

86

References

1. W. A. Lanford, "Glass Hydration - Method of Dating Glass Objects," Science 196(4293), 975-976 (1977).


3. T. Izumitani, "Polishing of Optical Glass: Mainly Investigated from the View Point of Material Science," Report: HGW-0-3E (Hoya Glass Works, Tokyo, 1971).


5. R. H. Doremus, Y. Mehrotra, W. A. Lanford, and C. Burman, "Reaction of Water with Glass - Influence of a Transformed Surface-Layer," Journal of Materials Science 18(2), 612-622 (1983).

6. H. Tomozawa and M. Tomozawa, "Diffusion of Water into a Borosilicate Glass," Journal of Non-Crystalline Solids 109(2-3), 311-317 (1989).

7. M. Tomozawa and K. Hirao, "Diffusion of Water into Oxides during Microhardness Indentation," Journal of Materials Science Letters 6(7), 867-868 (1987).

8. K. Hirao and M. Tomozawa, "Microhardness of Sio2 Glass in Various Environments," Journal of the American Ceramic Society 70(7), 497-502 (1987).

9. M. Nogami and M. Tomozawa, "Effect of Stress on Water Diffusion in Silica Glass," Journal of the American Ceramic Society 67(2), 151-154 (1984).

10. S. Nakashima, H. Matayoshi, T. Yuko, K. Michibayashi, T. Masuda, N. Kuroki, H. Yamagishi, Y. Ito, and A. Nakamura, "Infrared Microspectroscopy Analysis of Water Distribution in Deformed and Metamorphosed Rocks," Tectonophysics 245(3-4), 263-276 (1995).

87

11. N. Yanagisawa, K. Fujimoto, S. Nakashima, Y. Kurata, and N. Sanada, "Micro FT-IR study of the hydration-layer during dissolution of silica glass," Geochimica et Cosmochimica Acta 61(6), 1165-1170 (1997).

12. M. Maaza, B. Farnoux, F. Samuel, C. Sella, and P. Trocellier, "Effect of Mechanical Polishing on the Surface-Structure of Glasses Studied by Grazing Angle Neutron Reflectometry," Optics Communications 100(1-4), 220-230 (1993).

13. P. Baillif, B. Chouikhi, L. Barbanson, and J. C. Touray, "Dissolution kinetics of glass fibers in saline solution: in vitro persistence of a sparingly soluble aluminum-rich leached layer," Journal of Materials Science 30, 5691-5699 (1995).

14. M. Kawano and K. Tomita, "TEM-EDX, study of weathered layers on the surface of volcanic glass, bytownite, and hypersthene in volcanic ash from Sakurajima volcano, Japan," American Mineralogist 86(3), 284-292 (2001).

15. J. Yokota, H. Sakata, M. Nishibori, and K. Kinosita, "Ellipsometric study of polished glass surfaces," Surface Science 16, 265 - 274 (1969).

16. G. M. Mansurov, R. K. Mamedov, A. S. Sudarushkin, V. K. Sidorin, K. K. Sidorin, V. I. Pshenitsyn, and V. M. Zolotarev, "Study of the Nature of Silica-Glass Polished Surface by Ellipsometry and Spectroscopy Methods," Optika I Spektroskopiya 52(5), 852-857 (1982).

17. M. Malin and K. Vedam, "Ellipsometric Studies of Environment-Sensitive Polish Layers of Glass," Journal of Applied Physics 48(3), 1155-1157 (1977).

18. H. Ji, T. Rouxel, A. Abdelouas, B. Grambow, and P. Jollivet, "Mechanical behavior of a borosilicate glass under aqueous corrosion," Journal of the American Ceramic Society 88(11), 3256-3259 (2005).

19. J. D. Mackenzie, "High-Pressure Effects on Oxide Glasses: I. Densification in Rigid State," Journal of the American Ceramic Society 46(10), 461-470 (1963).

88

20. M. J. Cumbo, "Chemo-mechanical interactions in optical polishing," (University of Rochester, Rochester, NY, 1993).

21. J. M. Park and H. D. Jeong, "A study on the micro machining of Si wafer using surface chemical reaction," Advances in Abrasive Technology Vi 257-258, 459-464 (2004).


23. Nano Indenter XP and Testworks 4 software, MTS Nano Instruments, 1001 Larson Drive, Oak Ridge, TN 37830.

24. W. C. Oliver and G. M. Pharr, "An Improved Technique for Determining Hardness and Elastic-Modulus Using Load and Displacement Sensing Indentation Experiments," Journal of Materials Research 7(6), 1564-1583 (1992).

25. W. C. Oliver and G. M. Pharr, "Measurement of hardness and elastic modulus by instrumented indentation: Advances in understanding and refinements to methodology," Journal of Materials Research 19(1), 3-20 (2004).

26. A. C. Fischer-Cripps, Nanoindentation, Mechanical Engineering Series (Springer-Verlag New York, Inc., New York, 2002).

27. Y. G. Wei, X. Z. Wang, and M. H. Zhao, "Size effect measurement and characterization in nanoindentation test," Journal of Materials Research 19(1), 208-217 (2004).

28. J. G. Swadener, A. Misra, R. G. Hoagland, and A. Nastasi, "A mechanistic description of combined hardening and size effects," Scripta Materialia 47(5), 343-348 (2002).

29. R. Rodriguez and I. Gutierrez, "Correlation between nanoindentation and tensile properties Influence of the indentation size effect," Materials Science and

89

Engineering a-Structural Materials Properties Microstructure and Processing 361(1-2), 377-384 (2003).

30. W. D. Nix and H. J. Gao, "Indentation size effects in crystalline materials: A law for strain gradient plasticity," Journal of the Mechanics and Physics of Solids 46(3), 411-425 (1998).

31. J. W. Hutchinson, "Plasticity at the micron scale," International Journal of Solids and Structures 37, 225-238 (2000).

32. D. C. Brown, S. D. Jacobs, J. A. Abate, O. Lewis, and J. Rinefierd, "Figures of Merit and Correlations of Physical and Optical Properties in Laser Glasses," National Bureau of Standards (U.S.), Spec. Publ. 509 9th Annual Symposium on Optical Materials for High Power Lasers, in Laser Induced Damage in Optical Materials, ed. A. J. Glass and A. H. Guenther, 416-433, (1977).

33. HOYA Corporation, 572 Miyazawa-cho, Akishima-shi, Tokyo, Japan. (1998 version).

34. R. C. Weast, ed., CRC Handbook of Chemistry and Physics (CRC Press Inc., Boca Raton, FL, 1986), pp. F169 - F172.


36. "Standard Test Method for Vickers Indentation Hardness of Advanced Ceramics," ASTM International C 1327-03, 462-469 (2003).

37. SCHOTT North America, Inc., 555 Taxter Road, Elmsford, NY 10523. (2001 version).

90

Chapter 5

MRF Material Removal Rate Model

5.1 Introduction Preston1 was the first to model optical polishing by claiming proportionality

of removal rate to pressure and velocity. His equation has been modified for

conventional, CMP and MRF polishing processes. Buijs et al.2 made many

modifications to Preston’s equation to show that the material removal rate was also

dependent on the mechanical properties of the glass and the abrasive size. Matsuo et

al.3 found that the CMP material removal rate was dependent on the frictional force

rather than the polishing pressure. Luo and Dornfeld4 proposed that the CMP

material removal rate depended on more parameters than Preston’s equation. They

included a term for chemical etching in their model. Bielmann et al.5 demonstrated

that the material removal rate was inversely proportional to the abrasive size for

particles less than 0.5µm, and proportional to the abrasive size particles larger than

0.5µm. Lambropoulos et al.6,7 extended Buijs work to MRF and volumetric removal

rate. Shorey8 proposed a modified Preston’s equation for his work with MRF to

incorporate shear stress. Dunken9 incorporated quantitative values for chemistry by

stating that glass material removal rate was proportional to an Arrhenius-like

equation, and Cook10 showed that the material removal rate was related to the single

bond strength of the polishing abrasive.

For this thesis research we incorporate many of these ideas into an improved

model for the MRF material removal rate. Equation 5.1 is our material removal rate

equation where ES [units: GPa] and HS [units: GPa] are the Young’s modulus and

nanohardness of the near surface layer of the substrate immersed in MR fluid

supernatant, Kc [units: MPa m1/2] is the fracture toughness of the bulk substrate, Fd

[units: N] is the drag force between the MR fluid ribbon and substrate, A [units: mm2]

is the contact area between the MR fluid ribbon and the substrate, v [units: rpm] is the

velocity of the wheel and hence the polishing abrasive, φnd [units: nm] and Cnd [units:

91

vol. %] are the average particle size and concentration of the nanodiamond particles,

φCI [units: µm] and CCI [units: vol. %] are the average particle size and concentration

of the CI, Bnd [units: µm4/3] and BCI [units: µm-1/3] are coefficients (both empirically

equal to 1), Ds(pHMRF) is the percent weight loss of the glass in the MR fluid

supernatant, as a function of MR fluid pH, as determined by our chemical durability

test, sbs [units: J/mol] is the single bond strength of the glass network formers, b is a

unitless coefficient, R [8.314 J/mol K] is the universal gas constant and T is the room

temperature in Kelvin. For this research we hold the following terms constant: A

(28mm2), v (200rpm), φCI (3.5µm), CCI (45 vol. %) and T (296K).

1 3 1 3 4 3 3 10

2 v (− −⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡∝ ⋅ ⋅ ⋅ ⋅φ ⋅ + ⋅φ ⋅ ⋅ ⋅⎢ ⎥ ⎣ ⎦ ⎣⎢ ⎥⋅ ⎣ ⎦⎣ ⎦sbs bRTS d

peak nd nd nd CI CI CI s MRFc S

E FMRR B C B C D pH eK H A

) ⎤ ⎡ ⎤⎦ ⎣ ⎦ (5.1)

Term 1 Term 5 Term 2 Term 4 Term 3 In this chapter we discuss each of the five terms separately and then combine

them in the model to examine how mechanics, fluid properties and chemistry all work

together in the MRF material removal process. All of the experimental data discussed

in this section are tabulated in Appendix C.

5.2 Term 1: Mechanical Figure of Merit In 1996, Lambropoulos et al.7 established a positive linear correlation between

a set of bulk material mechanical properties (specifically bulk Young’s modulus, EB,

fracture toughness, Kc, and Knoop microhardness, Hk) and the volumetric rate of

removal, MRRvol, as follows:

1223

67

kc

Bvol HK

EMRR⋅

∝ (5.2)

The correlation coefficient was not given, but the plot is reproduced in Figure 1.7

(Chapter 1).

We modify Lambropoulos’ model in two ways. We use the bulk mechanical

properties with exponents that fit better for our results to obtain

92

2Vc

Bpeak HK

EMRR⋅

∝ (5.3)

and we substitute the near surface Young’s modulus, ES, and near surface Berkovitch

nanohardness, HS, for EB and Hk, to obtain

2Sc

Speak HK

EMRR⋅

∝ (5.4)

where our MRRpeak are peak MRR. A discussion of these near surface properties,

their measurement, and the results for our six glass set was given in Chapter 4.

In what follows we describe our measurements of the bulk mechanical

properties EB and Vickers microhardness, Hv, for our six glass set, and the calculation

of Kc from the Hv data. We compare our bulk data to literature values. We then

construct and examine several mechanical figure of merit (FOM) values based on the

measurement of ES and HS in dry, DI water and MR fluid supernatant environments.

5.2.1 Young’s modulus

Young’s modulus or elastic modulus (E) is the resistance of a material to

elastic deformation.11 We measured the bulk Young’s modulus values of our glass

set using ultrasonic wave propagation, or the pulse-echo method. The measurements

were made on the same pitch polished glass samples used for our MRF experiments.

The measurements were accomplished by introducing shear and longitudinal waves

with ultrasonic transducers (hand held against the glass with a small amount of

coupling gel in the interface) and measuring the wave velocities, vs and vl

respectively.12 Knowing the density of the glass sample, ρ, along with the wave

velocities, we determined EB using Equation 5.4.13 Measurements were performed

once on each glass sample. We estimate approximately 2% error in these

measurements due to a combination of errors in the measurements of wave speed and

density.

93

( )( )2

s2l

2s

2l

2

vvv4v3

−−

= sB

vE ρ (5.4)

Near surface values for Young’s modulus were measured with the Nano Indenter XP

as described in Chapter 4, using the Oliver and Pharr method.14-16

Table 5.1 gives our bulk measurements compared to bulk data from he

literature, along with our near-surface data.17-19 The near surface data are for a depth

of 60nm with nanoindentation in dry, DI water and MR supernatant environments.

In general we find good agreement between the measured and literature values for the

bulk Young's modulus. The near-surface Young's modulus values were consistently

higher than both the measured and literature bulk values. There is a 2x variation in ES

for our glass set.

Young's Modulus

Glass Pulse-echo, E B Literature, E B Dry DI Water MR Supernatant[GPa] [GPa] [GPa] [GPa] [GPa]

FS 69 +/- 1 73 81 +/- 1 74 +/- 1 78 +/- 1BK-7 81 +/- 2 82 97 +/- 2 94 +/- 6 85 +/- 2FD-60 93 +/- 2 92 100 +/- 2 109 +/- 1 99 +/- 2

EFDS-1 96 +/- 2 99 105 +/- 4 108 +/- 5 108 +/- 4LHG-8 62 +/- 1 50 67 +/- 2 59 +/- 1 55 +/- 2FCD-1 73 +/- 1 80 85 +/- 3 87 +/- 4 81 +/- 5

Nanoindentation, E S

Table 5.1 – Bulk and near surface Young's modulus data

for our glass set, with bulk data from the literature17-19 for comparison.

5.2.2 Hardness

Hardness is the measure of resistance of a material to plastic deformation. We

measured the bulk Vicker’s microhardness20 of our glasses using a 100g load under

ASTM standard conditions.21 Five indents were made on each of the plane parallel

pitch polished samples used in our MRF experiments. Equation 5.5 was used to

determine the Vickers microhardness, where L is the applied load, D is the length of

the indentation diagonal and α is the angle between the opposite faces of the Vickers

indenter (136o).22

94

( )2

2sin2D

LHvα

= (5.5)

The resulting micro-indent depths are between 2-3µm. Table 5.2 lists the bulk Knoop

microhardness data from the product literature14-16 and our measured values using

Vickers microindentation21, along with nanohardness data that was discussed in

Chapter 4. Vickers and Knoop microhardness data have a linear relationship, where

the Knoop values are lower based on the indenter shape.23 If we compare the

measured Vickers microhardness data to the Knoop microhardness values taken from

the literature we find that all of the Knoop values are lower than the Vickers values

with the exception of BK-7, where there the value is the same. The Knoop values for

FS and FD-60 are lower than the value for BK-7 which does not agree with the

Vickers microhardness values that we measured for our materials.

Hardness

Glass Vickers, H v Knoop, H k : Literature Dry DI Water MR Supernatant[GPa] [GPa] [GPa] [GPa] [GPa]

FS 7.5 +/- 0.4 5.5 11.4 +/- 0.0 9.1 +/- 0.5 10.0 +/- 0.2BK-7 6.0 +/- 0.3 6.0 8.6 +/- 0.3 8.9 +/- 0.1 8.5 +/- 0.2FD-60 6.3 +/- 0.3 5.4 9.1 +/- 0.4 9.9 +/- 0.4 8.3 +/- 0.4

EFDS-1 5.3 +/- 0.3 4.5 8.4 +/- 0.4 8.8 +/- 0.4 8.8 +/- 0.1LHG-8 3.7 +/- 0.2 3.1 6.5 +/- 0.3 5.7 +/- 0.1 4.4 +/- 0.1FCD-1 4.0 +/- 0.2 3.4 6.4 +/- 0.2 6.7 +/- 0.2 5.5 +/- 0.5

Nanoindentation, H S

Table 5.2 – Measured and literature17-19 micro- and

nanohardness values for our glass set. (Literature values were all made with 100gf load but according to different testing standards).

The literature values, although measured with the same load, were taken from the

product literature of the respective manufacturers who follow different standard test

methodologies.18, 19 The discrepancy we see here in the rank order of microhardness

is most likely due to differences in measurement techniques among the

manufacturers. Qualitatively, we find that the rank order for nanohardness of the dry

glass surface is almost the same as the Vickers microhardness data with the exception

of switching LHG-8 and FCD-1.

95

It is expected that the rank order of near surface nanohardness will change

from measurements made on the dry surface to those measured in DI water or MR

fluid supernatant due to chemistry. Variations in nanohardness for glasses tested in

fluid environments are influenced by chemical effects, as discussed in Chapter 4.

There is a 2x variation in HS for our glass set.

5.2.3 Fracture Toughness

Fracture toughness is a measure of the resistance of a material to fracture or

extension of a pre-existing crack.11 We used 100g load Vickers indents from

hardness tests and the Evans’ model24 (see Equations 5.4 – 5.6) to calculate the

fracture toughness, Kc, of our optical glasses. In Equation 5.5, c is the half the crack

length shown in Figure 5.1. Table 5.3 gives the bulk fracture toughness data for our

six glasses. There is nearly a 2x variation in Kc for our glass set. We were not able

to measure the fracture toughness of FS due to densification25; therefore the value

given in Table 5.3 is a literature value.6 We know of no method for generating near

surface fracture toughness data because cracks/fracture do not occur at low

loads/depths.

)(4.0

102

xf

vvc H

EDHK ⎟⎟⎠

⎞⎜⎜⎝

⎛= (5.4)

⎟⎟⎠

⎞⎜⎜⎝

⎛=

2log10 D

cx (5.5)

5432 32.1697.2423.1102.234.059.1)( xxxxxxf +−+−−−= (5.6)

96

2c D

Vickers Indentation

Radial Cracks

Figure 5.1 – Drawing of the top view of a Vickers indent. Glass Bulk Fracture Toughness Literature

[MPa m1/2] [MPa m1/2]FS 0.75 0.75

BK-7 0.80 +/- 0.04 0.85FD-60 0.69 +/- 0.03

EFDS-1 0.59 +/- 0.03LHG-8 0.52 +/- 0.03 0.50FCD-1 0.47 +/- 0.02

Table 5.3 – Bulk fracture toughness data for our six optical glasses with data from literature6 (when available) for comparison.

5.2.4 Mechanical FOM term

The FOM term (Term 1 in Equation 5.1) may be calculated in four ways from

the bulk and near surface mechanical properties of the glasses. In Table 5.4 these

FOM terms are listed along with Lambropoulos’ FOM. Lambropoulos FOM is the

FOM calculated from the bulk sample measurements (pulse-echo EB and Hv) using

the original Lambropoulos model (equation 5.2).7 FOM – B is the FOM (equation

5.3) calculated from the bulk sample measurements (pulse-echo EB and Hv), FOM – D

(equation 5.4) is calculated from the dry nanoindentation measurements, FOM – W

(equation 5.4) is calculated from the DI water nanoindentation measurements and

FOM – S (equation 5.4) is calculated from the MR fluid supernatant nanoindentation

measurements (ES and HS). The Kc values calculated from the Vickers indents are

used for all of the FOM terms. To see if the use of near surface mechanical properties

improves the linear positive correlation between this FOM element and peak MRF

removal rate, we look at a select subset of data. Additional comparisons of the FOM

terms are included in section C.1 in Appendix C of this thesis.

97

MR fluid supernatant DI water Dry Bulk Bulk

Glass Lambropoulos FOM FOM - B FOM - D FOM - W FOM - SFS 4.0 1.6 0.8 1.2 1.0

BK-7 6.8 2.8 1.6 1.5 1.5FD-60 8.5 3.4 1.8 1.6 2.1

EFDS-1 14.3 5.8 2.5 2.4 2.4LHG-8 19.3 8.7 3.0 3.5 5.5FCD-1 22.1 9.6 4.4 4.1 5.7

Table 5.4 – Mechanical FOM values calculated from Young’s modulus and hardness values measured for the bulk (B), the dry near surface (D), the near surface in DI water (W) and the near surface in MR fluid supernatant (S).

We find that the near surface mechanical properties measured in MR fluid

supernatant describe MRF material removal rate better than the mechanical properties

of the bulk material. The near surface and bulk figure of merit values are plotted

versus peak removal rate data for the six optical glasses and 0.01-vol% UK-Medium

A MR fluid in Figure 5.2. Linear fits have been drawn for each set of FOM values.

The FOM-S values provide a better linear fit than the Lambropoulos FOM and FOM-

B values. In comparison, the near surface FOM-D values provide the lowest

confidence level of all the FOM values plotted in Figure 5.2. The FOM-D values

would best describe a mechanically dominant removal that was not dependent on

water or chemistry. This is clearly not the case in our work. The goodness of the

linear fit increases significantly when we add the effect of water by plotting the FOM-

W values with the same peak removal rate data. Finally the best of the three near

surface mechanical FOM linear fits is found for FOM-S. These data support our

hypothesis that the MR fluid supernatant softens the near surface layer of glass, and

then the layer is mechanically sheared off with polishing abrasives at a rate

proportional to the mechanical properties of the softened hydrated layer. This

98

interpretation follows the conventional pad polishing process description of

Izumitani.26 There is approximately an 80% variation in Term 1 for our glass set.

Lambropoulos' modelR2 = 0.52


FOM-DR2 = 0.31CL: 75%

FOM-WR2 = 0.50CL: 85%

FOM-SR2 = 0.64CL: 90%

FOM-BR2 = 0.57CL: 90%

0

5

10

15

20

25

30

0 5 10 15 20 25

FOM

Peak

Rem

oval

Rat

e ( µ

m/m

in)

FS

BK-7FD-60

EFDS-1

LHG-8

FCD-1

Figure 5.2 – Bulk and near surface mechanical figure of merit

(FOM) values plotted with peak removal rate data for MRF spots taken with 0.01-vol% UK-Medium A nanodiamond MR fluid. [Data tabulated in Table C.8 in Appendix C].

5.3 Term 2: Modified Preston’s equation

The second term in our model contains terms for drag force, Fd, spot contact

area, A, and wheel speed, v, (velocity of the abrasive particle). The term Fd/A is also

referred to as shear stress. This section will focus primarily on drag force

measurements. For most of our work the spot contact area and wheel speed were held

constant to reduce confounding effects. Experiments were performed to determine

99

the influence on peak removal rate from variations in both contact area and wheel

speed, and these results are discussed first.

5.3.1 Wheel speed and contact area

Our material removal model contains terms for spot contact area and wheel

speed (150mm diameter wheel). These two terms are difficult to vary without

changing other parameters in the MRF process. For example, if we change the wheel

speed, the ribbon height will change which will affect the spot contact area.

Therefore it is very difficult to change only one parameter. The data in Figures 5.3

and 5.4 are spot contact area and wheel speed versus peak removal rate for BK-7. We

chose to vary the pump speed to maintain the desired ribbon height. The data in

Figures 5.3 and 5.4 show that there is a good, inverse linear relationship between spot

contact area and peak removal rate and a strong, positive linear relationship between

wheel speed and peak removal rate as predicted by Preston1. The magnetic field was

held constant (10A) throughout these experiments.

R2 = 0.69Confidence Level: 90%

0

2

4

6

8

10

12

14

16

18

0 5 10 15 20 25 30 35

Spot Contact area, A (mm2)

Peak

Rem

oval

Rat

e ( µ

m/m

in)

Contact Area

(mm2)

Removal Rate

(µm/min)22.0 16.3124.0 14.4625.0 15.9028.0 11.2830.0 12.72

Figure 5.3 – Spot contact area versus peak removal rate for BK-7. Wheel speed was held constant; pump speed was varied to maintain a constant ribbon height.

100

R2 = 0.99Confidence Level: >99%

0

2

4

6

8

10

12

14

16

0 50 100 150 200 250 300Wheel Speed, v (rpm)

Peak

Rem

oval

Rat

e ( µ

m/m

in)

Wheel Speed (rpm)

Removal Rate

(µm/min)150 7.79175 9.49200 11.38200 11.13225 12.26250 13.85

[2.0 m/s] [1.2 m/s] [1.6 m/s]

Figure 5.4 – Wheel speed versus peak removal rate for BK-7. Spot contact area was held constant and the pump speed was varied to maintain a constant ribbon height. The corresponding linear velocities with units of meters/second are given in brackets.

The contact area and wheel speed terms were held constant (A = 28mm2 and v

= 200rpm) in the development of the MRF material removal rate model discussed in

this thesis. Their effects are linear and well-understood as seen in this section.

5.3.2 Drag force

Shorey expressed the importance of the coefficient of friction in properly

determining the material removal rate model for MRF. He used a strain gauge to

determine the drag force induced by the rotating MR fluid ribbon. All of his drag

force measurements were made on sapphire.8 In this research we continue his

approach to further determine how different nanodiamond abrasive properties and

their concentration affect the coefficient of friction and simultaneously the removal

rate. Our measurements are unique in that we measure the drag force on the same

material that we use to measure peak removal rates.

101

5.3.2.1 Force sensor experimental set up

Drag force measurements were made using a piezo-electric force sensor (see

Figure 5.5).27 This sensor’s high sensitivity (error less than +/- 0.2N) allows us to

detect the drag forces corresponding to subtle changes in non-magnetic abrasive

concentration and/or glass type. The three second data acquisition time allows for

multiple measurements on our six glass set in times as short as 30 minutes. The

LabVIEW28 software program to control the force sensor was created internally.29

Data was analyzed in Microsoft Excel. All measurements were made with a 1.6mm

ribbon height and the part depth into the ribbon of 0.3mm. Six blocking bodies were

fabricated for the force sensor setup (drawing located in Appendix C).30 The use of

these blocking bodies allowed us to mount and measure the drag force for our entire

six glasses in the same experiment.

Force sensor

Part

Ribbon

Wheel rotating

clockwise

Figure 5.5 - Photograph of the piezo-electric force sensor used to measure drag force, mounted on the STM.

The output signal from the force sensor is proportional to the drag force. An

example of the output is shown in Figure 5.6. The data acquired during the MRF spot

is averaged and the standard deviation is calculated for each measurement. Only one

spot is taken for every glass/fluid combination.

102

∝ 1.4N

Figure 5.6 – An example of the force sensor output for LHG-8 pressed against an NDP nanodiamond MR fluid ribbon, plotted with Microsoft Excel.

5.3.2.2 Drag force results

We determined how an increase in nanodiamond concentration would affect

the drag force for each of our glasses. We chose to use the NDP nanodiamonds

because they are supplied in dry form, and we can add increasing amounts to the

same MR fluid without the complexity of additional water and surfactants associated

with nanodiamonds supplied in suspension. We observe from the data in Figure 5.7

that, as the concentration of nanodiamonds is increased, there is a positive linear

correlation with drag force for each of our six glasses. The drag force data in Figure

5.7 are plotted against the corresponding peak removal rate data in Figure 5.8. Our

model predicts that there is a positive linear relationship between drag force and peak

removal rate, which we see in Figure 5.8. The confidence levels are all higher than

95%.31 For LHG-8, a very modest increase in drag force caused a 5x increase in

removal rate. The increase in removal rate was 2x for FCD-1, EFDS-1 and BK-7 and

50% for FS and FD-60. There is approximately 60% variation in Term 2 for our

glass set. The range of drag force values is comparable to what was observed in

Figure 1.8a on page 15.8, 32

103

LHG-8R2 = 0.90CL: >99%

FSR2 = 0.93CK: >99%

EFDS-1 R2 = 0.83CL: 98%

FCD-1R2 = 0.66CL: 95%

BK-7R2 = 0.67CL: 95%

FD-60R2 = 0.65


0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0 0.002 0.004 0.006 0.008 0.01Nanodiamond Concentration (vol%)

Dra

g Fo

rce

(N)

Figure 5.7 – Drag force plotted versus NDP nanodiamond concentration for our six optical glasses. [Data tabulated in Table C.12 in Appendix C].

LHG-8R2 = 0.95

Confidence Level (CL): >99%

FCD-1

R2 = 0.68CL: 95%

EFDS-1R2 = 0.79CL: 98%

FD-60R2 = 0.67CL: 95%

FSR2 = 0.94CL: >99%

BK-7R2 = 0.72CL: 95%

0.0

5.0

10.0

15.0

20.0

25.0

0.0 1.0 2.0 3.0 4.0 5.0Drag Force (N)

Pea

k R

emov

al R

ate

( µm

/min

)

Figure 5.8 – Peak removal rate plotted versus drag force for our six optical glasses in MR fluids with increasing concentrations of NDP nanodiamonds. [Data tabulated in Table C.12 in Appendix C].

104

Figure 5.9 gives a plot of peak removal rate versus drag force for four

nanodiamond MR fluids and all glasses. The nanodiamonds in the fluids are NDP,

UK-Low, UK-Medium A and UK-High at a concentration of 0.01-vol%. The

variable in this plot is glass type. We observe that drag force and peak removal rate

do not show the same linear correlation across all six glasses. The phosphate glasses

(LHG-8, FCD-1 and EFDS-1) show a negative linear correlation between drag force

and peak removal rate and the silicates (FS, BK-7 and FD-60) have a positive linear

correlation. This is true regardless of the nanodiamond manufacturer or friability

level. Chemistry and glass composition play a significant role in the MRF material

removal process, and removal rate cannot be characterized by drag force alone.

0

5

10

15

20

25

30

1.0 1.5 2.0 2.5 3.0 3.5 4.0Drag Force (N)

Peak

Rem

oval

Rat

e ( µ

m/m

in)

NDP

UK Low

UK Medium A

UK High(Silicates indicated with filled

shapes, Phosphates indicated with hollow shapes)

LHG-8

FCD-1 EFDS-1

BK-7 FD-60

FS

Figure 5.9 – Peak removal rate versus drag force for four 0.01-vol% nanodiamond MR fluids. Spots were taken on all six glasses with each fluid. The glass types are identified for the NDP points only to make the figure easier to read. The glass order is the same for the other nanodiamond fluids. [Data tabulated in Tables C.7, C.8, C.9 and C.12 in Appendix C].

105

The phosphate glasses require lower drag force for the MRF process to

remove material compared to the silicate glasses. If we recall the near surface

nanohardness of the glasses immersed in MR fluid supernatant, Hs, we find that there

is an inverse relationship between Hs and drag force for the phosphate glasses, and

there is a positive relationship for the silicate glasses. In Figure 5.10 we plot peak

removal rate data versus drag force divided by Hs and we find linear correlations with

all confidence levels higher than 70%. Therefore we find that our model already

accounts for the difference between the phosphate and silicate glasses.

R2 = 0.79; CL: 98%

R2 = 0.86; CL: 99%UK High

R2 = 0.57; CL: 90%

R2 = 0.27; CL: ~70%

0

5

10

15

20

25

30

0.0 0.1 0.2 0.3 0.4 0.5 0.6

Fd/Hs (N/GPa)

Pea

k R

emov

al R

ate

( µm

/min

)

NDP

UK Low

UK Medium A

(Silicates indicated with filled shapes, Phosphates indicated

with hollow shapes)

FS

BK-7 FD-60

FCD-1

LHG-8

EFDS-1

Figure 5.10 – Peak removal rate versus Fd/Hs for four 0.01-vol% nanodiamond MR fluids. Spots were taken on all six glasses with each fluid. The glass types are identified for the NDP points only to make the figure easier to read. The glass order is the same for the other nanodiamond fluids. [Data tabulated in Tables C.7, C.8, C.9 and C.12 in Appendix C].

106

5.4 Term 3: Abrasive size and concentration

The third term in our model [Bndφnd-1/3Cnd

1/3+BCIφCI4/3CCI] incorporates fluid

properties into the removal process. This term is based on work by Mahajan et al.33

for CMP polishing of oxide films with silica. They found relationships for material

removal rate based the size of the abrasive particles which they referred to as the

surface area based mechanism and the indentation-based mechanism. If the particles

were smaller than 0.5µm, smaller particles and higher concentrations produced higher

CMP material removal rates, but if the particles were larger than 0.5µm the inverse

occurred.

MRF is unique in that it contains two abrasive particles; CI particles that are

larger than 0.5µm and nanodiamonds which are smaller than 0.5µm. The two models

presented by Mahajan et al. are modified and combined into our third term which

allows us to incorporate size and concentration effects from both particles. We chose

to add the two terms together so that when there are no non-magnetic abrasives

present in the MR fluid our model is not forced to zero.34 The two coefficients, Bnd

[units: µm4/3] and BCI [units: µm-1/3] were included to balance the units for Term 3.

These two coefficients are empirically equal to one. We have substantial data on

polishing with only CI particles (see section 3.4.1).

5.4.1 Nanodiamond size and concentration

As our model predicts, smaller nanodiamond particles produce higher material

removal rates as seen for each glass type in the plots in Figures 5.11 (phosphate

glasses) and 5.12 (silicate glasses). These two Figures contain our experimental data

for the 29nm (UK-Medium A) and 54nm (NDP) nanodiamonds for various

concentrations. The difference is largest for LHG-8 and greater than 50% at low

concentrations. For the other glasses it is approximately 2 – 15%.

107

0

5

10

15

20

25

30

0 0.005 0.01 0.015 0.02 0.025 0.03Nanodiamond Concentration (vol%)

Pea

k R

emov

al R

ate

( µm

/min

)

LHG-8 - 54nm FCD-1 - 54nm EFDS-1 - 54nm

LHG-8 - 29nm FCD-1 - 29nm EFDS-1 - 29nm

Figure 5.11 – Peak removal rate versus nanodiamond concentration for the 29nm (UK-Medium A) and 54nm (NDP) nanodiamonds for phosphate glasses LHG-8, FCD-1 and EFDS-1. [Data tabulated in Tables C.2, C.5, C.8 and C.12 in Appendix C].

108

0

1

2

3

4

5

6

7

8

9

10

0 0.005 0.01 0.015 0.02 0.025 0.03Nanodiamond Concentration (vol%)

Pea

k R

emov

al R

ate

( µm

/min

)

FS - 54nm BK-7 - 54nm FD-60 - 54nmFS - 29nm BK-7 - 29nm FD-60 - 29nm

Figure 5.12 – Peak removal rate versus nanodiamond concentration for the 29nm (UK-Medium A) and 54nm (NDP) nanodiamonds for silicate glasses FS, BK-7 and FD-60. [Data tabulated in Tables C.2, C.5, C.8 and C.12 in Appendix C].

Material removal rate is sensitive to very small additions of nanodiamonds.

For example the first addition (3mg) of NDP nanodiamonds increases the peak

removal rate 7 - 37% based on the glass type. We calculate that the ratio of 54nm

nanodiamond particles to 3.5µm CI particles is approximately 5:1 after the first

addition of NDP nanodiamonds. We were able to reach higher nanodiamond

concentrations with the 54nm NDP nanodiamonds compared to the 29nm UK-

Medium A, because the NDP nanodiamonds were supplied in dry form. We can add

dry nanodiamonds to an MR fluid circulating in the STM without altering the water

content and other fluid properties. The nanodiamond concentration range of 0.01 to

0.03 vol. % is also where the peak removal rates start to show asymptotic behavior

109

with further increases in nanodiamond concentration for most glasses. In Figure 5.13

we plot the data from Figure 5.12 with the third term in our material removal model.

The two main variables are nanodiamond size and concentration. The CI size

(3.5µm) and concentration (45-vol%) are held constant. The coefficients Bnd and BCI

were both empirically equal 1. The coefficients were included to balance the units for

Term 3 and they are not included in Figures 5.13 and 5.14. The positive linear

correlation between the peak removal rate and term 3 is very high for all six of our

glasses.

LHG-8R2 = 0.90

FSR2 = 0.81

BK-7R2 = 0.94

FCD-1

R2 = 0.90

FD-60R2 = 0.91

0

5

10

15

20

25

30

239.0 239.5 240.0 240.5

φnd-1/3Cnd

1/3+φCI4/3CCI

Pea

k R

emov

al R

ate

( µm

/min

)

EFDS-1

R2 = 0.90

Figure 5.13 – Peak removal rate data versus our third term of the MRF material removal model. The 29nm UK-Medium A and 54nm NDP nanodiamond concentration are varied. CI size and concentration are constant as described in the text. All of the linear trend lines have confidence levels greater than 99%. [Data is tabulated in Tables C.2, C.5, C.8 and C.12 in Appendix C].

110

The plot in Figure 5.14 is similar to the one in Figure 5.13, but this plot

contains the data from nanodiamond concentration experiments using the 35nm (UK-

Medium B) and 44nm (UK-Medium C) medium friability UK nanodiamonds in

addition to the UK-Medium A and NDP nanodiamond results. The R2 values are

slightly lower in this plot compared to Figure 5.13, but due to the increased number

of data points, the confidence levels are still much greater than 99%. There is

approximately a 40% variation in Term 3 for the MR fluid conditions we tested.

LHG-8R2 = 0.80

FSR2 = 0.72

BK-7R2 = 0.89

FCD-1

R2 = 0.84

FD-60R2 = 0.87

0

5

15

20

25

30

239.0 239.2 239.4 239.6 239.8 240.0 240.2 240.4 240.6

φnd-1/3Cnd

1/3+φCI4/3CCI

Peak

Rem

oval

Rat

e ( µ

m/m

in)

EFDS-1R2 = 0.88

10

Figure 5.14 – Peak removal rate data versus the third term of the MRF material removal rate model. The experimental data for this plot includes the increasing nanodiamond concentration for the 29nm, 35nm, 44nm and 54nm nanodiamond experiments. CI size and concentration are constant as described in the text. The confidence levels are all greater than 99%. [Data is tabulated in Tables C.2, C.5, C.8, C.10, C.11 and C.12 in Appendix C].

111

5.4.2 CI size and concentration

Most of our work involved varying the nanodiamond size and concentration.

We did not do any work with varying the size distribution of CI particles. We

hypothesize that using larger CI particles in the MR fluid would result in higher

removal rates. At this time we do not have any experimental data to determine the

specific dependence, therefore we decided to use the term φCI4/3 based on the CMP

model33.

We performed one experiment where we varied the concentration of the CI

particles in an MR fluid that contained only CI particles. The results from this

experiment are shown in Figure 5.15. In order to increase CI concentration it was

necessary to decrease the water content. We accomplished this in reverse by starting

with a very low water content and gradually adding water to the fluid. We observe

from these results that there is a strong positive linear correlation between peak

removal rate and increasing CI concentration. Confidence levels are all greater than

99% for each glass. The material removal rate increases due to the increasing CI

concentration which results in increased in-field MR fluid viscosity8 and increased

ribbon stiffness. [Currently in our laboratory we can only measure out-of-field

viscosity (see Figure 5.15).] It is very difficult to separate the effect of these two

factors. The linear CCI dependence shown in this section was reflected in Term 3 of

the model plotted in Figures 5.13 and 5.14.

112

FSR2 = 0.98

BK-7R2 = 0.98

LHG-8R2 = 0.95

FCD-1R2 = 0.97

0

2

4

6

8

10

12

14

16

40 41 42 43 44 45 46 47 48 49CI Concentration (vol%)

Peak

Rem

oval

Rat

e ( µ

m/m

in)

0

10

20

30

40

50

60

70

80

90

100

Out

-of-f

ield

MR

flui

d vi

scos

ity (c

P)

Out-of-fieldMR Fluid Viscosity

Figure 5.15 – Peak removal rate and out-of-field MR fluid viscosity versus CI concentration for four glasses. CI concentration was varied with the addition of DI water. [Data tabulated in Table C.13 in Appendix C].

5.5 Term 4: Glass chemical durability

There are very few quantitative models presented in the literature1, 8, 10, 32, 35-40

that include terms relating to the chemical durability of the glass being polished. One

of our goals for this research is to improve the existing mechanical models by

including a quantitative term for glass chemical durability.

In Chapter 1, Figure 1.4, Izumitani’s36 results are shown for polishing rate

versus the percent weight loss of glass in water. The relationship is not linear, but it

is a power relationship (MRR ∝ [Percent weight loss]1/10). We calculate an R2 value

of 0.63 (confidence level >99%) with Izumitani’s data. We also find a power

relationship between chemical durability and the MRF material removal rate.

113

In this section we discuss how we modified the O’HARA water resistance test

to determine the chemical durability for each of the six glasses. We used these data

to define the term Ds(pHMRF) specific to each glass type. Chemical durability, Ds, is a

function of the MR fluid pH, pHMRF. The pHMRF is measured before we take each set

of spots; therefore the Ds(pHMRF) term in our model takes into account the varying pH

of the MR fluid. Finally we discuss how we experimentally determined the power

dependence in the relationship between the chemical durability term and the MRF

peak removal rate.

5.5.1 Chemical durability testing protocol and removal rate correlation

We modified the O'HARA water resistance test41 for determining chemical

durability of our glasses in the MR fluid supernatant and in tap water. MR fluid

supernatant is the liquid portion of the MR fluid which typically has a pH value of

between 9 and 11 (see pg. 36).

The supernatant used for this work was collected from 20-vol. % CI MR fluid

(375g of CI) that was aged on a roll mill for 2 days prior to filtration. The MRF

supernatant was extracted from the MR fluid by allowing the fluid to sit for one hour

in which time the CI settled to the bottom of the container. The supernatant was

siphoned off the top and filtered with filter paper. No CI particles were present in the

supernatant. A batch of 250mL of 20-vol% CI MR fluid produced enough

supernatant (100mL) to run one durability test. The supernatant was used the same

day it was filtered. Testing was completed at pH 7.0 (tap water), 9.0, 9.8, 10.5, 11.0

and 12.0. The supernatant pH level is naturally around 10.5 and was adjusted using

either boric acid or NaOH.

The glass samples were ground using mortar and pestle and sized using a set

of sieves to obtain a size range between 425 – 600µm. A test was run by filling a

platinum wire basket with the ground glass (0.3 – 0.5g depending on glass density)

and weighing it (to 5 decimal places) with an analytical balance. Testing solution

114

(tap water or MR fluid supernatant) was heated in a Pyrex beaker to 100oC and the

platinum basket filled with ground glass was submerged for one hour as shown in

Figure 5.16.

Figure 5.16 – Photograph of the glass chemical durability

testing set up.

After one hour the platinum basket and sample was removed from the

supernatant and placed in an oven preheated to 120oC for one hour to dry. The

sample and glass were weighed again and the difference from the initial weight was

recorded. This procedure was followed 6 times for each glass/fluid pH combination.

Figure 5.17 gives a plot of Ds versus testing solution pH for all six glasses on

the semi-log scale. Power trend lines [Ds ∝ (pHMRF)B] are drawn, and the equations

are included in Figure 5.17. All of the confidence levels are higher than 99%. These

equations will be used in our MRF material removal model.

115

Figure 5.17 – Percent weight loss, Ds, versus tes

solution pH for all six optical glasses. Ds relationships that are used in our Mmaterial removal model are located onright hand side of the figure.

The power relationship between Ds and removal rate w

experimentally. Figure 5.18 shows an example of 0.01-vol% NDP nan

plotted on a semi-log scale with the corresponding Ds(pHMRF) value

equation is also shown in Figure 5.18 for an exponent of 0.3 with a c

of 90%. For comparison, Izumitani’s data and power fit with an expo

also shown in Figure 5.18. The good fit is further illustrated by con

for the other nanodiamond abrasives and aging MR fluid in the next s

not perform aging experiments with NDP nanodiamond MR fluids.

LHG-8Ds α (pHMRF)8.71

BK-7Ds α (pHMRF)8.22

FSDs α (pHMRF)10.20

FD-60Ds α (pHMRF)11.70

EFDS-1Ds α (pHMRF)9.70

FCD-1Ds α (pHMRF)6.56

0.001

0.01

0.1

1

10

100

6 7 8 9 10 11 12 13 1

Testing solution pH

Perc

ent w

eigh

t los

s, D

s

MR fluid supernatant Tap Water

B = 8.71

B = 8.22

B = 11.70
B = 9.70
B = 6.56

B = 10.20

ting The RF

the

as determined

odiamond data

s. The best fit

onfidence level

nent of 0.1 are

sidering results

ection. We did

4

116

FS

FD-60

EFDS-1

FCD-1

BK-7

LHG-8

SK16

LaLK3

LaK12

LaLF2BK7

KF2

SK2

SF6LaK10

NbSF3

NbF1

TaF2

0.01-vol% NDP Nanodiamond MR Fluid, pHMRF = 10PRR = 8.3 Ds(pHMRF)0.3


Conventional CeO2 Pad Polish, pH 7PRR = 0.4 Ds

0.1


0.1

1

10

100

0.01 0.1 1 10Ds

Peak

Rem

oval

Rat

e ( µ

m/m

in)

Figure 5.18 – Peak removal rate versus chemical durability, Ds, for 0.01-vol% NDP nanodiamond MR fluid experimental data and Izumitani’s36 data for conventional CeO2 pad polishing. [Data tabulated in Table C.12 in Appendix C].

5.5.2 Chemical durability and aging MR fluid

The pH value of an MR fluid, pHMRF, that is allowed to naturally age in an

MRF machine will gradually decrease if no corrective action is taken.42 The MR

fluids used for our experiments exhibited this behavior and we wanted to account for

this in our model.

The 0.01-vol% UK-Low, UK-Medium A and UK-High friability

nanodiamond MR fluids were circulated in the STM for 9 days, during which time 10

sets of spots were taken under the same operating conditions. DI water was added

when needed to maintain a constant viscosity. We monitored the pHMRF and the

results are shown in Figure 5.19 as a function of elapsed time that the fluids were in

117

the STM. As previously indicated, a time-appropriate pHMRF value is used to

determine the specific quantitative chemical durability value for each spot. Term 4 of

our model is plotted with the 0.01-vol% UK-Low friability nanodiamond MR fluid in

Figure 5.20. The data from the 0.01-vol% UK-Medium friability nanodiamond MR

fluid are shown in Figure 5.21 and the 0.01-vol% UK-High friability nanodiamond

MR fluid results are plotted in Figure 5.22. There is an approximate 60-70%

variation in Term 4 for our glass set.

9.0

9.2

9.4

9.6

9.8

10.0

10.2

10.4

0 2000 4000 6000 8000 10000 12000 14000Elapsed Time (minutes)

pHM

RF

UK-Low UK-Medium A UK-High

Day 1 Day 9

Figure 5.19 – Measured MR fluid pH values as a function of time for 0.01-vol% UK-Low, UK-Medium A and UK-High nanodiamond MR fluids. The fluids were allowed to naturally age in the STM for 9 days. [Data tabulated in Tables C.7 – C.9].

118

LHG-8R2 = 0.99

FSR2 = 0.83

BK-7R2 = 0.97

FCD-1R2 = 0.84

FD-60R2 = 0.83

0

2

4

6

8

10

12

14

0.0 0.5 1.0 1.5 2.0

Ds(pHMRF)0.3

Pea

k R

emov

al R

ate

( µm

/min

)

EFDS-1

R2 = 0.90

Figure 5.20 – Peak removal rate versus Term 4 of our MRF material removal model. The 0.01-vol% UK-Low nanodiamond fluid was allowed to naturally age for 9 days in the STM. All spots were made using the same operating conditions. The confidence levels for all of the linear trend lines drawn in the figure are greater than 99%. [Data tabulated in Table C.7 in Appendix C].

119

LHG-8R2 = 1

BK-7R2 = 0.89FS

R2 = 0.79FD-60

R2 = 0.92

EFDS-1R2 = 0.88

FCD-1R2 = 0.89

0

5

10

15

20

25

30

0.0 0.5 1.0 1.5 2.0

Ds(pHMRF)0.3

Pea

k R

emov

al R

ate

( µm

/min

)

Figure 5.21 – Peak removal rate versus Term 4 of our MRF material removal model. The 0.01-vol% UK-Medium A nanodiamond fluid was allowed to naturally age for 9 days in the STM. All spots were made using the same operating conditions. The confidence levels for all of the linear trend lines drawn in the figure are greater than 99%. [Data tabulated in Table C.8 in Appendix C].

120

LHG-8R2 = 0.88CL: >99%

FSR2 = 0.60CL: >99%

BK-7R2 = 0.82CL: >99%

FD-60R2 = 0.31CL: 90%

EFDS-1

R2 = 0.76CL: >99%

FCD-1R2 = 0.12

Confidence Level (CL): <75%

0

2

4

6

8

10

12

14

16

0.0 0.5 1.0 1.5 2.0

Ds(pHMRF)0.3

Pea

k R

emov

al R

ate

( µm

/min

)

Figure 5.22 – Peak removal rate versus Term 4 of our MRF material removal model. The 0.01-vol% UK-High nanodiamond fluid was allowed to naturally age for 9 days in the STM. All spots were made using the same operating conditions. [Data tabulated in Table C.9 in Appendix C].

Figure 5.22 gives a plot of removal rate versus Term 4 for the 0.01-vol% UK-

High friability nanodiamond MR fluid. FS and LHG-8 show strong positive linear

dependencies (>99%) with Term 4. The confidence level for the positive linear

correlation between FD-60 and Term 4 is lower than that for FS and LHG-8, but it is

still 90%. There is a negative linear relationship between removal rate and Term 4

for FCD-1, EFDS-1 and BK-7. MRF is not a purely chemical process; therefore we

can not expect to perfectly describe every polishing situation using only chemical

terms. This is apparent for these three glasses using the UK-High nanodiamonds

only, and we do not have an explanation.

121

5.6 Term 5: Glass average single bond strength

The fifth term [e-sbs/bRT] that we incorporate into our material removal model

for MRF is based on work by Dunken who stated that the chemical dissolution rate

for glass was proportional to Arrhenius’ equation.9 Sugimoto et al. found that it was

difficult to use the Arrhenius equation to characterize Chemical Mechanical Polishing

(CMP) removal exclusively, because the CMP mechanism included both mechanics

and chemistry.43 We also believe this to be true of the MRF removal mechanism; it

involves both mechanics and chemistry. Term 5 is an Arrhenius-like equation where

we use the average single bond strength, sbs, (energy required to break bonds) of each

glass as opposed to the activation energy (energy required to activate chemical

reaction) used in the traditional Arrhenius equation and b is a unitless coefficient

(empirically equal to 1000). This substitution creates a term that accounts for both

the glass chemical composition and the mechanical energy required to break the

bonds.

5.6.1 Determination of glass average single bond strength

In order to determine the average single bond strength of our glasses we

needed to know the exact composition. Ground samples of LHG-8, BK-7, FCD-1,

FD-60 and EFDS-1 were sent to NSL Analytical44 for quantitative elemental analysis.

The composition data resulting from this analysis are listed in Table 5.5. Fused silica

was not analyzed. We assume it to be 100% SiO2.

SiO2 B2O3 Na2O K2O MgO BaO As2O3 CaO Zr2O TiO2 Sb2O3 P2O5 Al2O3 Li2O AlF3 BaF2 SrF2 CaF2 MgF2 Nd2O3 Nb2O5

FS 100BK-7 69 11 10 8 1 1FD-60 27 9 4 15 1 2 30 12EFDS-1 3 2 4 19 6 1 19 46LHG-8 1 12 1 12 57 8 1 3 1FCD-1 1 30 15 10 1 1 22 14 6

Table 5.5 – Composition data for our glass set. Values listed are weight percentages.

Based on work by Sun 45,46, and knowing the exact compositions, we calculate

the average single bond strength for each of our six glasses.47 An example of this

122

calculation for SiO2 is shown in Equation 5.7. The calculated average single bond

strength values are given in Table 5.6 along with the values for the resulting

Arrhenius-like equation (Term 5). Temperature was constant (296K) for these

calculations. There is approximately a 12% variation in Term 5 for our glass set.

(5.7) %__2 Silicon+⋅=

%___

___%_2 WtWtSilicon

SiliconofmassAtomic

OxygenofmassAtomicWtSiO

Glass sbs (kJ/mol) e -sbs/RT e -sbs/bRT

FS 799.6 5.86E-145 0.723BK-7 695.5 3.51E-126 0.754

FD-60 665.6 8.69E-121 0.763EFDS-1 679.7 2.49E-123 0.759LHG-8 499.1 9.38E-91 0.816FCD-1 513.9 2.01E-93 0.812

Table 5.6 – Calculated average single bond strength (sbs) [units: kJ/mol] and Term 5 (R is the gas constant [units: kJ/mol K], T is temperature = 296K) and b is a unitless coefficient empirically equal to 1000.

5.6.2 Glass average single bond strength results

Figure 5.23 contains our experimental peak removal rate data for Abrasive

free and four 0.01-vol% nanodiamond MR fluids versus an Arrhenius-like equation,

the fifth term in the MRF material removal model. We observe from the data in the

figure that the material removal rates are higher for glasses with lower single bond

strengths. Linear trend lines are drawn for each MR fluid. The confidence levels for

the nanodiamond fluids are all higher than 95%. The abrasive free MR fluid has only

an 80% confidence level, and it shows the weakest dependence. We have previously

hypothesized that the nanodiamond particles cut through material where as the CI

particles skip and/or roll across the surface during removal. This hypothesis would

also explain the results shown in Figure 5.23 where the nanodiamonds continuously

break bonds as they travel across the glass surface and have strong correlations with

123

Term 5 where as the CI particles can be thought of as skipping across the surface in a

discontinuous manner.

Abrasive-freeR2 = 0.49CL: 80%

UK-LowR2 = 0.92CL: 99%

UK-HighR2 = 0.89CL: 99%

UK-Medium AR2 = 0.66


NDPR2 = 0.80CL: 98%

0

5

10

15

20

25

30

0.70 0.72 0.74 0.76 0.78 0.80 0.82 0.84 0.86

e(-sbs/bRT)

Pea

k R

emov

al R

ate

( µm

/min

)

Figure 5.23 – Peak removal rate versus Term 5 with the average single bond strength for five MR fluids. [Data tabulated in Tables C.7, C.8, C.9 and C.12 in Appendix C].

5.7 MRF material removal rate model

In this section we revisit the five terms we discussed in this chapter and

combine these to form our MRF material removal rate model (see Equation 5.1). The

experimental data in Figure 5.24 came from 650 spots taken on six glasses using 0 –

0.01-vol% concentrations of UK-Low, UK-Medium A, UK-High, UK-Medium B,

UK-Medium C and NDP nanodiamond MR fluids. [Data for Figures 5.24 and 5.25

are tabulated in Tables C.1 – C.12 in Appendix C.]

124


0

5

10

15

20

25

30

0 500 1000 1500 2000 2500 3000 3500 4000

Peak

Rem

oval

Rat

e ( µ

m/m

in)

1 3 1 3 4 3 3 10

2 v (− −⎡ ⎤ ⎡ ⎤ )⎡ ⎤ ⎡ ⎤ ⎡ ⎤⋅ ⋅ ⋅ ⋅ φ ⋅ + ⋅ φ ⋅ ⋅ ⋅⎢ ⎥ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦⎢ ⎥⋅ ⎣ ⎦⎣ ⎦sbs bRTS d

nd nd nd CI CI CI s MRFc S

E F B C B C D pH eK H A

Figure 5.24 – Experimental peak removal rate data for six glasses with various MR fluids versus our MRF material removal rate model, incorporating mechanics, polishing particle properties and chemistry. The terms A, v, φCI, CCI, BCI, Bnd, b, R and T are all constant.

Our new model allows us to build a better understanding about glass material

removal using the MRF process with nanodiamond MR fluids. In this chapter we

have shown the importance of the near surface mechanical properties to the MRF

material removal rate. Our Term 1 (eqn. 5.4) provides a better fit compared to the

figure of merit of the bulk glass properties (FOM-B, eqn. 5.3) for MRF peak removal

rate data. Term 1 also spans the largest range of values for our glass set compared to

all other terms in our model.

We found that increasing the drag force or shear stress (Fd/A) in the modified

Preston’s equation [Term 2: (Fd/A)⋅v] produced higher material removal for each

125

glass type with increasing nanodiamond concentration. Glasses with phosphate

network formers require less drag force to remove material compared to glasses with

silica network formers. In comparison between glass types for a given MR fluid, the

phosphate glasses have a negative linear correlation and the silicate glasses have a

positive linear relationship. We concluded that this is due to the hardness of the

hydrated layer, Hs, and when we combine Fd from Term 2 and Hs-1 from Term 1 we

find a positive linear relationship for all glasses.

Polishing particle properties [Term 3: Bndφnd-1/3Cnd

1/3+BCIφCI4/3CCI] play a

significant role in the removal process. The addition of nanodiamonds increases the

material removal efficiency, but eventually the nanodiamonds reach a saturation point

where removal no longer increases. Smaller nanodiamond particles increase MRF

peak removal rate due to increased contact or surface area with the glass substrate.

The CCI term in Term 3 was found to have a positive linear relationship with MRF

material removal rate, and it is included to the power 1 for our model, even though it

was kept constant.

We found that even though the MRF material removal process is dominated

by mechanics, chemical durability of glass and hence the pH of the MR fluid play a

role in the process. Similar to Izumitani’s work with conventional pad polishing we

found that the nanodiamond MR fluid peak removal rates increase with Term 4

[Ds(pHMRF)3/10] for all six of our glasses. Term 4 influences the material removal rate,

but should not be used alone, because the MRF material removal mechanism is not

purely chemical.

Finally, the average strength of the glass network bonds, or Term 5 [e-sbs/bRT],

is very important to the process even though it spans the smallest range compared to

the four other terms in our model. Experimental data show that glasses with low

average single bond strength values have higher material removal rates compared to

glasses with higher average single bond strength values.

Most importantly we found that MRF material removal is best described not

with one single term, but with a combination of terms that involve mechanics,

126

polishing particle properties and chemistry. Our new model provides us with a better

understanding of the interaction between the MR fluid and the glass surface during

material removal. This model also brings us one step further to predicting material

removal.

In Chapter 6 we conclude the major body of research in this thesis. We will

discuss the surface texture that is left inside the MRF spots that were made with the

nanodiamond MR fluids discussed in this chapter.

References

1. F. W. Preston, "The theory and design of plate glass polishing machines,"

Journal of the Society of Glass Technology 11, 214 - 256 (1927).

2. M. Buijs and K. Korpel-van Houten, "A Model for Lapping of Glass," Journal of Materials Science 28(11), 3014-3020 (1993).

3. H. Matsuo, A. Ishikawa, and T. Kikkawa, "Role of frictional force on the polishing rate of Cu chemical mechanical polishing," Japanese Journal of Applied Physics 43(4B), 1813 - 1819 (2004).

4. J. Luo and D. Dornfeld, "Material removal mechanism in chemical mechanical polishing: Theory and Modeling," IEEE Transactions on Magnetics 14(2), 112 - 133 (2001).

5. M. Bielmann, U. Mahajan, and R. K. Singh, "Effect of particle size during tungsten chemical mechanical polishing," Electrochemical and Solid State Letters 2(8), 401-403 (1999).


127

7. J. C. Lambropoulos, S. D. Jacobs, and J. Ruckmann, "Material removal mechanism from grinding to polishing," Ceramic Transactions 102, 113-128 (1999).


9. H. Dunken, "Surface chemistry of optical glasses," Journal of Non-Crystalline Solids 129, 64-75 (1991).


11. W. F. Smith, Foundations of Materials Science and Engineering, 3rd ed. (The McGraw-Hill Companies, Inc., New York, 2004).


13. M. J. Bamber, K. E. Cooke, A. B. Mann, and B. Derby, "Accurate determination of Young's modulus and Poisson's ratio of thin films by a combination of acoustic microscopy and nanoindentation," Thin Solid Films 398-399, 299-305 (2001).

14. Nano Indenter XP and Testworks 4 software, MTS Nano Instruments, 1001 Larson Drive, Oak Ridge, TN 37830.

15. W. C. Oliver and G. M. Pharr, "An Improved Technique for Determining Hardness and Elastic-Modulus Using Load and Displacement Sensing Indentation Experiments," Journal of Materials Research 7(6), 1564-1583 (1992).

16. W. C. Oliver and G. M. Pharr, "Measurement of hardness and elastic modulus by instrumented indentation: Advances in understanding and refinements to methodology," Journal of Materials Research 19(1), 3-20 (2004).

128

17. Corning, One Riverfront Plaza, Corning, NY. (1993 version).



20. Tukon Model 300, Wilson Instruments, Binghamton, NY. The instrument operates with a 50x microscope objective. Measurements were made using a 15s loading time with the Vickers micro-indenter.

21. "Standard Test Method for Vickers Indentation Hardness of Advanced Ceramics," ASTM International C 1327-03, 462-469 (2003).

22. B. W. Mott, Micro-indentation hardness testing (Butterworths Scientific Publications, London, 1956).

23. J. C. Lambropoulos, T. Fang, P. D. Funkenbusch, S. D. Jacobs, M. J. Cumbo, and D. Golini, "Surface microroughness of optical glasses under deterministic microgrinding," Applied Optics 35(22), 4448-4462 (1996).

24. A. G. Evans, Fracture Toughness: The role of Indentation Techniques, Fracture Mechanics Applied to Brittle Materials (American Society for Testing and Materials, Philadelphia, 1979), Vol. 678, pp. 112-135.

25. J. C. Lambropoulos, T. Fang, S. Xu, and S. M. Gracewski, "Constitutive law for the densification of fused silica with applications in polishing and microgrinding," SPIE 2536: Optical Manufacturing and Testing, ed. V. J. Doherty and H. P. Stahl, 275-286, (1995).

26. T. S. Izumitani, Optical Glass (American Institute of Physics, New York, 1986).

27. Kistler single axis slim line shear load cell measuring system, model number 9143B21 (www.kistler.com). Maxwell Bennett Associates, P.O. Box 401, W. Henrietta, NY 14586.

129

28. National Instruments Corporation, 11500 N Mopac Expwy, Austin, TX 78759-3504.

29. MRF force sensor program was written while Jason Keck was employed at The Laboratory for Laser Energetics. Jason is now employed at Semrock, Inc. 3625 Buffalo Road, Suite 6, Rochester, NY 14624.

30. The force sensor set up was designed by Andrew Dillenbeck of The Laboratory for Laser Energetics, 250 E. River Road, Rochester NY 14623.

31. A. Wheeler and A. Ganji, Introduction to Engineering Experimentation (Prentice-Hall Inc., Upper Saddle River, NJ, 1996).

32. A. B. Shorey, S. D. Jacobs, W. I. Kordonski, and R. F. Gans, "Experiments and observations regarding the mechanisms of glass removal in magnetorheological finishing," Applied Optics 40(1), 20-33 (2001).

33. U. Mahajan, M. Bielmann, and M. Singh, "Abrasive Effects in Oxide Chemical Mechanical Polishing," Materials Research Society 566: Mat. Res. Soc. Symp., 27-32, (2000).

34. J. C. Lambropoulos, Rochester, NY (personal communication, October 28, 2005).

35. V. H. Bulsara, Y. Ahn, S. Chandrasekar, and T. N. Farris, "Mechanics of polishing," Journal of Applied Mechanics-Transactions of the ASME 65(2), 410-416 (1998).



38. A. Kaller, "Properties of polishing media for polishing optics," Glastechnische Berichte-Glass Science and Technology 71(6), 174 - 183 (1998).

130

39. T. Hoshino, Y. Kurata, Y. Terasaki, and K. Susa, "Mechanism of polishing of SiO2 films by CeO2 particles," Journal of Non-Crystalline Solids 283(1-3), 129-136 (2001).

40. M. J. Cumbo, D. Fairhurst, S. D. Jacobs, and B. E. Puchebner, "Slurry Particle Size Evolution during the Polishing of Optical-Glass," Applied Optics 34(19), 3743-3755 (1995).

41. "OHARA Optical Glass Catalogue," (OHARA Corporation, Somerville, NJ), 7,

42. M. Schinhaerl, E. Pitschke, R. Rascher, P. Sperber, R. Stamp, L. Smith, and G. Smith, "Temporal stability and performance of MR polishing fluid," SPIE 5523: Current Developments in Lens Design and Optical Engineering, ed. P. Z. Mouroulis, W. Smith, and R. B. Johnson, 273-280, (2004).

43. F. Sugimoto, Arimoto, Y., Ito, T., "Simultaneous Temperature Measurement of Wafers in Chemical Mechanical Polishing of Silicon Dioxide Layer," Japanese Journal of Applied Physics 34, 6314-6320 (1995).

44. NSL Analytical, 7650 Hub Parkway, Cleveland, Ohio 44125, Phone: (216) 447-1550, E-mail: [email protected], http://www.nslanalytical.com.

45. K.-H. Sun, "Fundamental Condition of Glass Formation," Journal of the American Ceramic Society 30, 277 - 281 (1947).

46. K.-H. Sun and M. L. Huggins, "Energy Additivity in Oxygen-Containing Crystals and Glasses, II," Journal of Physics and Colloidal Chemistry 51(2), 438-443 (1947).

47. R. C. Weast, ed., CRC Handbook of Chemistry and Physics (CRC Press Inc., Boca Raton, FL, 1986), pp. F169 - F172.

131

Chapter 6

The role of nanodiamonds in the polishing zone

6.1 Introduction

There is a surface texture inside MRF spots that consists of grooves that

follow the direction of the MR fluid flow. These grooves are especially apparent in

spots where there is no substrate rotation. It is hypothesized that the grooving is a

result of an interaction between abrasive particles in the MR fluid and the part

surface. Analysis of this unique surface texture allows us to learn more about the

surface/particle interaction during the polishing process.

When abrasives, specifically nanodiamonds, are added to the MR fluid they

are homogeneously mixed with the CI particles. When the MR fluid passes into the

magnetic field, the CI particles are pulled toward the surface of the wheel because of

a magnetic field gradient, leaving the non-magnetic abrasive particles in a thin film at

the surface of the stiffened ribbon.

If we take this hypothesis one step further, we can consider which particles are

involved in removing material. We have previously discussed the role of CI particles

themselves in removing material without any nonmagnetic abrasives present in the

MR fluid. In this chapter we specifically examine the role of nanodiamonds which

we purposefully add in very small quantities (i.e. 30mg to 3.9kg of MR fluid). As the

optic is depressed into the ribbon, the nonmagnetic nanodiamonds are trapped

between the CI particles and the glass part, where they cut through and remove

material from the surface.

In this chapter we attempt to better understand how the nanodiamonds and CI

interact with the glass surface by studying the surface texture. To do this we use a

Taylor Hobson CCI 30001 to measure areal surface roughness. In addition to

studying the surface roughness we use the data collected from the CCI 3000 to

generate and analyze orthogonal power spectral densities (PSD). PSD allows a

132

breakdown of the surface roughness over a specific spatial frequency range. [All

experimental data that support this chapter are tabulated in Appendix C.]

6.2 Surface texture and the MRF material removal process

Lambropoulos et al.2, 3 and Shorey4 proposed that the material removal

mechanism for MRF is micro-lateral-fracture. We speculate that the mechanism for

material removal with the abrasive free MR fluid is “micro-gouging” where the CI

particles do not cut and fracture the surface but roll and stick across the surface

causing pits that are polished into a comet-like form by the MR fluid. Once

nanodiamonds are introduced into the MR fluid we agree with Lambropoulos and

Shorey that the micro-lateral-fracture mechanism is dominant. The following

abrasive free and nanodiamond MR fluid surface texture data support our hypothesis.

LHG-8 is the most sensitive glass in our set to changes in the MR fluid

properties. The images in Figure 6.1 are false color surface maps of an LHG-8

surface after a pitch polish†, together with maps that were taken inside MRF spots

made with the abrasive-free and with three nanodiamond fluids ten minutes after the

nanodiamonds were added. Spot times for the abrasive free, low and medium

friability nanodiamond MR fluids were 2 seconds in duration. The spot time for the

high friability nanodiamond MR fluid was one second. Spot times were adjusted in

order to achieve desirable spot depths (0.15 – .2µm) for measurements on the Mark

IVxp (see Chapter 2).

There are major differences in the surface textures inside the four MRF spots.

The abrasive-free spot contains severe pitting or micro-gouges and the average areal

surface roughness values (both p-v and rms) are approximately four times larger than

those of the pitch polished surface. Ten minutes after the addition of very small

amounts of nanodiamonds (30mg) into the abrasive-free fluid, the pitting is either

greatly reduced or eliminated. This is rather remarkable. Simple calculations show

that there are approximately 15, 25 or 87 nanodiamond particles (assumes mono-sized † All glass surfaces used for spotting experiments were initially pitch polished by Alex Malstev in the LLE fabrication shop using Gugolz 73 optical polishing pitch and a Hastilite PO cerium oxide slurry.

133

19, 29 or 34nm diameter spheres) for each CI particle (assumes mono-sized 3µm

diameter spheres). These results strengthen our hypothesis that the nanodiamonds are

being efficiently transported to the surface of the stiffened ribbon for polishing,

thereby causing the transition from micro-gouging to micro-lateral-fracture.

0 0.05 0 .1 0.15 0.2 0 .2 5 0 .3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

05101520253035404550556065

0 0.05

rms: 3.74 +/- 0.59 p-v: 44.9 +/- 3.1

rms: 0.96 +/- 0.09 p-v: 13.6 +/- 4.3

0.1 0 .15 0.2 0.25 0.3 0 .35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

5

10

15

20

25

30

35

40

0 0.05 0.1 0 .15 0.2 0.25 0.3 0 .35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

012345678910111213

0 0.05 0 .1 0.15 0.2 0 .2

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Pitch polished Abrasive-free

Low friabi Medium friability High friabil

(2 second spot)

rms: 1.31 +/- 0.04 p-v: 14.0+/- 2.9

rms: 1.20 +/- 0p-v: 20.9 +/-

rms: 2.01 +/- 0.04 p-v: 41.3+/- .4 8

lity (2 second spot)

(2 second spot) (1 second spo

Figure 6.1 - False color surface maps and areal roughness values (all in nm) of the LHG-8 surfaces. Nogrooving is observed on the pitch polished surface (upper left). The paths taken byabrasives across the part surface within spotsare seen as grooves extending from top tobottom in all other images. The field of viewis 0.35mm x 0.35mm. Nanodiamondconcentration is 0.001-vol%. [MR fluid flow direction indicated].

Figure 6.2 examines the evolution of areal rms surface roughness

over time as the MR fluid aged, and Figure 6.3 shows a plot of the cor

peak removal rate data. The results are given in Tables 6.1 - 6.4. The ab

MR fluid rms surface roughness remained around 4nm for the entire exper

Flow direction

5 0 .3 0.35 m m nm

02468101214161820222426

ity

.07 5.9

t)

for LHG-8

responding

rasive-free

iment, and

134

pitting continued throughout. The low friability nanodiamond MR fluid initially

showed reduced pitting and reduced rms surface roughness, but as time elapsed the

pitting returned and the rms surface roughness increased to values similar to those

found for the abrasive-free spots. This suggests that the edges of the low friability

nanodiamonds became even more rounded or dull as the fluid was circulated in the

STM, resulting in a shorter performance lifetime. They were less likely to create

micro-lateral fractures compared to that for the medium and high friability

nanodiamonds. The implication from the data in Figure 6.2 is that the medium and

high friability nanodiamonds continued to break during milling to expose fresh sharp

edges. The results shown in Figure 6.2 help support our hypothesis that the lower

friability nanodiamonds have a shorter period of activity compared to the higher

friability nanodiamonds. The medium and high friability nanodiamonds had rms

surface roughness values slightly higher than 1nm for the entire experiment. There

was hardly any pitting within any of these spots.

The peak removal rate data shown in Figure 6.3 indicate that all three

nanodiamonds increased the removal rate and progressively declined in efficiency as

the MR fluid aged. Although the low friability nanodiamonds were no longer active

in producing smooth surfaces, they still continued to polish more efficiently than the

abrasive free MR fluid. Therefore the nanodiamonds, although dulled, were still

present in the polishing zone after three days of use in the STM. The medium and

high friability nanodiamond removal rates were very dissimilar, unlike their surface

roughness results. The medium friability nanodiamond MR fluid removal rate was

significantly higher than that of the high friability nanodiamond MR fluid. Based on

these results it appears that it is possible to engineer a nanodiamond with a specific

friability level that will give both high peak removal rate values and produce smooth

surfaces. The ideal nanodiamond would be dependent on the nanohardness of the

glass hydrated layer. For example LHG-8 (4 GPa) would require a higher friability

nanodiamond compared to FS (10 GPa).

135

0

1

2

3

4

5

6

7

1 10 100 1000 10000

Elapsed Time (min)

Ave

rage

Are

al R

MS

(nm

)

Low Friability Medium Friability High Friability Pitch Polished Abrasive-Free

Pitch Polished

Abrasive-Free

Low Friability

High FriabilityMedium Friability

Abrasive Free

Figure 6.2 - Average areal rms surface roughness versus elapsed time for LHG-8.

0

2

4

6

8

10

12

14

1 10 100 1000 10000

Elapsed Time (minutes)

Pea

k R

emov

al R

ate

( µm

/min

)

Abrasive Free Low Friability Medium Friability High Friability

Figure 6.3 – Peak removal rate versus elapsed time for LHG-8 corresponding to surface roughness data shown in Figure 6.2.

136

Time (min)

Spot Time

Removal Rate

(um/min)Average

RMS (nm) St. Dev.Average PV (nm) St. Dev.

initial 0.90 0.03 13.40 3.741 2 3.25 3.74 0.59 44.86 3.13

106 2 3.28 3.66 0.10 47.40 2.23257 2 3.14 3.94 0.27 59.16 7.50382 2 3.06 5.21 0.31 62.30 15.19

1537 2 2.58 3.72 0.45 48.94 2.381624 2 2.44 4.21 0.62 47.44 3.542866 2 2.03 3.17 0.35 42.90 4.72

Table 6.1 – Abrasive free MR fluid peak removal rate and surface roughness data for LHG-8.

Time (min)

Spot Time (sec)

Removal Rate

(um/min)Average


initial 0.96 0.09 13.61 4.32abrasive free 2 3.47 7.30 0.56 83.82 9.15

15 2 5.25 2.01 0.37 41.28 8.3776 2 4.64 3.69 1.14 57.94 13.42140 2 4.61 4.52 0.86 66.06 7.34203 2 4.44 4.65 0.70 69.80 16.62295 2 3.58 5.36 1.12 65.98 11.34

1391 2 3.39 5.27 0.81 64.58 8.341636 2 3.39 4.41 0.88 57.78 7.332794 2 3.11 3.10 1.22 65.26 19.51

Table 6.2 – 0.001-vol% low friability nanodiamond MR fluid peak removal rate and surface roughness data for LHG-8.

Time (min)

Spot Time (sec)

Removal Rate

(um/min)Average



7 2 11.61 1.31 0.04 13.98 2.8856 1 8.97 1.11 0.07 11.13 1.56166 1 7.74 1.16 0.25 25.86 26.73232 1 7.49 1.12 0.04 14.34 5.66312 1 7.44 1.21 0.16 22.04 16.59

1379 1 4.82 1.62 0.70 34.30 19.571593 2 4.56 1.36 0.28 34.00 22.692816 2 3.81 1.39 0.19 31.06 9.61

Table 6.3 – 0.001-vol% medium friability nanodiamond MR fluid peak removal rate and surface roughness data for LHG-8.

Time (min)

Spot Time (sec)

Removal Rate

(um/min)Average



7 1 6.15 1.20 0.07 20.88 5.9444 1 5.23 1.37 0.14 35.08 21.51106 1 5.23 1.81 0.07 26.00 7.40171 1 4.92 1.16 0.03 13.88 2.22228 2 5.06 1.38 0.11 20.56 7.73

1341 2 3.94 1.42 0.16 33.88 30.671494 2 4.17 1.30 0.07 14.60 2.982838 2 3.28 1.73 0.52 27.78 10.11

Table 6.4 – 0.001-vol% high friability nanodiamond MR fluid peak removal rate and surface roughness data for LHG-8.

137

6.3 Surface roughness and power spectral density

In this section we look at the effects of nanodiamond concentration, glass

mechanical properties and drag force on the p-v and rms surface roughness inside

MRF spots made with select nanodiamond MR fluids. Spot times varied from one to

five seconds depending on glass type (see Appendix C). Where applicable we will

include orthogonal power spectral density (PSD) analyses inside the MRF spots. The

p-v surface roughness inside MRF spots made with nanodiamond MR fluid is also

compared to a relationship determined by Buijs and Korpel-van Houten for lapping

glass.

6.3.1 Varying nanodiamond concentration

The average areal p-v surface roughness inside the MRF spots on six optical

glasses made with varying concentrations of NDP nanodiamonds* is plotted in Figure

6.4. It was mentioned in section 6.3 that LHG-8 is the most sensitive glass to changes

in the MR fluid. This is observed again in the data shown in Figure 6.4. The surface

inside the LHG-8 spot made with abrasive free MR fluid is covered with pits. Once

nanodiamonds are added to the MR fluid, pitting is almost completely eliminated and

the p-v surface roughness decreases almost by a factor of 4. The average areal p-v

surface roughness does not vary with the next four additions of nanodiamonds, but

the surface roughness is always higher than the initial pitch polished surface. With

the last addition of nanodiamonds (to give a total of 0.03-vol%) the p-v surface

increased due to deep grooves on the surface (shown in Figure 6.3). The average

areal p-v surface roughness values for remaining six glasses (FS, BK-7, FCD-1, FD-

60 and EFDS-1) are not sensitive to changes in nanodiamond concentration. All of

the values were similar to the initial pitch polished surface independent of the

nanodiamond concentration.

* Spots were taken approximately 10 – 15 minutes after each addition of nanodiamonds to the MR fluid.

138

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

0.0001 0.001 0.01 0.1 1

Nanodiamond Concentration (vol%)

Aver

age

area

l p-v

sur

face

roug

hnes

s (n

m)

LHG-8FSBK-7FCD-1FD-60EFDS-1Series7Series8Series9Series10Series11Series12

0Abrasive free

initial surface

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

10

20

30

40

50

60

70

80

90

0

LHG-8: Pitting (2 second spot)

0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

5

10

15

20

25

30

35

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

02468101214161820222426

LHG-8: Pitting almost eliminated

(1 second spot)

LHG-8: Deep MRF grooves (1 second spot)

Figure 6.4 – Average p-v surface roughness inside MRF spots made with NDP nanodiamond MR fluid at various nanodiamond concentrations. [Data tabulated in Table C.12 in Appendix C].

From a research standpoint, the high sensitivity of LHG-8 to small changes in

the MR fluid is very interesting; it allows us to learn more about the MRF process. In

Figures 6.5 and 6.6 we compare the orthogonal PSD analysis inside spots made on

LHG-8 with the abrasive free, 0.001-vol% and 0.03-vol% NDP nanodiamond MR

fluid in the horizontal (x) and vertical (y) directions respectively (see Chapter 2 for

detailed measurement description). In both figures the initial pitch-polished surface is

included for comparison.

We first examine the PSD in the horizontal direction (perpendicular to the

MRF grooving) (see Figure 6.5). The PSD analyzed in the horizontal direction

allows us to examine the amplitude and periodicity of the residual MRF grooving

pattern. The abrasive-free data has the highest amplitude for almost the entire spatial

frequency range in Figure 6.5. The higher amplitude is due to surface pitting which

139

we deem is caused by the spherical CI particles rolling along, sticking to, and gouging

out the softened modified surface layer. The horizontal PSD analysis supports this

hypothesis. We find that the abrasive-free amplitude becomes significantly higher

than the pitch polished and nanodiamond MR fluid spot surfaces within the spatial

frequency range that corresponds to the 2 – 5µm size of the CI particles. The addition

of nanodiamonds decreases the PSD amplitude over the entire spatial frequency

range. We have mentioned throughout this thesis that the non-magnetic

nanodiamonds are transported to the surface of the MRF ribbon in the presence of a

magnetic field. Even with the small amount of nanodiamonds present in the MR fluid

we assume that the nanodiamonds form a very thin layer between the CI particles and

the glass surface. This aids in cutting and efficiently removing the softened modified

layer, leaving behind a pit-free, smoother surface. The thickness of this layer of

nanodiamonds increases for the 0.03-vol% NDP nanodiamond MR fluid compared to

the 0.001-vol% MR fluid. One hypothesis for the increase in PSD amplitude for this

higher nanodiamond concentration is MRF groove decoration. Once a groove is

produced, the nanodiamonds remove material inside the groove increasing its depth

and width. This produces very deep grooves of various widths and the resulting PSD

analysis will look jagged specifically at spatial frequencies lower than the CI particle

size range as shown in Figure 6.5. Another hypothesis is that the NDP nanodiamonds

agglomerate to form larger particles at higher concentrations in the polishing zone

which will create larger residual grooves. The initial pitch-polished surface exhibits

lower PSD amplitude compared to all of the surfaces inside the MRF spots.

140

0.1

1

10

100

1000

10000

0.001 0.01 0.1 1

Spatial Frequency (1/µm)

Ave

rage

Pow

er D

ensi

ty (n

m2 µ

m)

Initial pitch polished surface Abrasive free 0.001-vol% 0.03-vol%

[1 µm][10 µm][100 µm]

[Periodicity]

[1000 µm]CI particle size range [5µm] [2µm]

Figure 6.5 – 1-D PSD analyzed in the horizontal direction (perpendicular to the MRF grooves) inside LHG-8 MRF spots made with 0, 0.001 and 0.03-vol% NDP nanodiamond MR fluids and compared to the initial pitch polished surface.

The PSD in the vertical direction (parallel to the MRF grooves) is shown in

Figure 6.6. PSD analyzed in the vertical direction allows us to trace the path of the

magnetic CI and non-magnetic nanodiamond particles along the surface of the glass

substrate. Similar to the horizontal PSD analysis, the LHG-8 abrasive free spot

surface has the highest PSD amplitude throughout the entire spatial frequency region

shown in Figure 6.6. This is again due to pitting, which is visible in both the vertical

and horizontal directions. The 0.001-vol% and 0.03-vol% NDP nanodiamond

horizontal PSD results are very similar. We think that the nanodiamonds are the

particles that are touching the glass surface removing material and the CI particles act

as a lap pressing the nanodiamonds against the surface. According to this hypothesis

the vertical PSD results will be similar because the nanodiamond paths are almost

141

identical, as is shown in Figure 6.6. The amplitudes are low for the two nanodiamond

MR fluids in the 0.2 – 0.5 (1/µm) spatial frequency region that corresponds to the size

of the CI particles. Therefore we infer that the CI particles are no longer gouging the

surface during the removal process. We also observe that the 0.001-vol% and 0.03-

vol% NDP nanodiamond MR fluid spots have lower PSD amplitudes compared to the

pitch polished surface for spatial frequencies 0.03 (1/µm) and higher. This shows us

that the nanodiamond particles are creating very smooth surfaces, smoother than a

pitch polished surface over a very large spatial frequency region.

0.1

1

10

100

1000

10000

0.001 0.01 0.1 1

Spatial Frequency (1/µm)A

vera

ge P

ower

Den

sity

(nm

2 µm

)

Initial pitch polished surface Abrasive free

[1 µm][10 µm][100 µm]

[Periodicity]

[1000 µm]

Figure 6.6 – 1-D PSD analyzed in th(parallel to the MRF groMRF spots made with vol% NDP nanodiamocompared to the initial pi

The average areal rms surface roughness ins

various concentrations of NDP nanodiamonds is sho

CI particle size range [5µm] [2µm]

0.001-vo

e verticoves) ins0, 0.001

nd MR tch polish

ide the s

wn in Fig

l% 0.03-vol%

al direction ide LHG-8 and 0.03-fluids and ed surface.

ame spots made with

ure 6.7. The average

142

areal rms surface roughness values inside the LHG-8 spots are similar to average

areal p-v surface roughness values shown in Figure 6.4. There is a small gradual

increase in areal rms surface roughness inside FCD-1 spots, but the silicate glasses

(FS, BK-7 and FD-60) which are harder than most phosphates do not appear to be

affected by changes in nanodiamond concentration.

The rms surface roughness results for EFDS-1 in Figure 6.6 are very

interesting. The average areal rms surface roughness inside the abrasive free spot

(0.72 +/- 0.03nm) is smoother than the initial pitch polished surface (0.89 +/-

0.03nm). Gradually as nanodiamonds are added, the surface roughness increases and

reaches its maximum (2.62 +/- 0.62nm) at 0.007-vol% NDP nanodiamond

concentration. The surface roughness inside the spots becomes smoother as the

nanodiamond concentration continues to increase, and at a 0.1-vol% NDP

nanodiamond concentration, the rms surface roughness is 1.34 +/- 0.13nm. False

color surface images taken with the CCI 3000 white light interferometer are shown in

Figure 6.7 for the 0, 0.007, 0.1-vol% NDP nanodiamond MR fluid spots and the

initial EFDS-1 initial pitch polished surface. The abrasive free surface appears to

have disjointed MRF grooves. These are most likely caused by the CI particles

rolling or skidding across the surface while removing material. The 0.007-vol% and

0.1-vol% NDP nanodiamond concentration spots have continuous MRF grooves. The

0.007-vol% spot also appears to have wider grooves compared to the 0.1-vol% spot

which we assume accounts for the higher average areal rms surface roughness for the

0.007-vol% spot. We take a closer look at the EFDS-1 spots using orthogonal PSD

analysis in Figures 6.8 and 6.9.

143

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

0.0001 0.001 0.01 0.1 1Nanodiamond Concentration (vol%)

Aver

age

area

l rm

s su

rface

roug

hnes

s (n

m)

LHG-8FSBK-7FCD-1FD-60EFDS-1Series7Series8Series9Series10Series11Series12

0(Abrasive free)

initial surface

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

1

2

3

4

5

6

7

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

2

4

6

8

10

12

14

16

18

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

1

2

3

4

5

6

7

8

9

10

EFDS-1: 0.1-vol%

(1 second spot)

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

1

2

3

4

5

6

7

8

9

10

11

EFDS-1: Initial pitch

polished surface

EFDS-1: 0.007-vol%

(1 second spot)

EFDS-1: Abrasive Free (2 second spot)

Figure 6.7 – Average rms surface roughness inside MRF spots made with NDP nanodiamond MR fluid at various nanodiamond concentrations. [Data tabulated in Table C.12 in Appendix C].

The PSD analysis in the horizontal (perpendicular to the MRF grooves)

direction is given in Figure 6.8. The 0.007-vol% NDP nanodiamond MR fluid spot

has the highest amplitude for the lower spatial frequency region (>0.03 µm-1), which

we are able to observe in wider MRF grooves in false color surface images in Figure

6.7. This spatial frequency region contributes to the higher average areal rms surface

roughness value seen in Figure 6.7. It is not obvious why the low spatial frequency

features are present in the 0.007-vol% NDP nanodiamond fluid and not at lower or

higher nanodiamond concentrations. We found through our nanoindentation work

(Chapter 4) that EFDS-1 does not create a softened hydrated layer in the presence of

144

MR fluid, therefore this phenomenon is most likely caused by a mechanical

mechanism.

0.1

1

10

100

1000

10000

0.001 0.01 0.1 1


Ave

rage

Pow

er D

ensi

ty (n

m2 µ

m)


[1 µm][10 µm][100 µm]

[Periodicity]

[1000 µm]


Figure 6.8 – 1-D PSD analyzed in the horizontal direction (perpendicular to the MRF grooves) inside EFDS-1 MRF spots made with 0, 0.007 and 0.1-vol% NDP nanodiamond MR fluids and compared to the initial pitch polished surface.

The vertical (parallel to the MRF grooves) PSD analysis shown in Figure 6.9

indicates that both the CI particles and the nanodiamonds create smoother surfaces as

they remove material compared to the initial pitch polished surface. The vertical PSD

analysis for the abrasive free spot indicates that the track that the CI particles take is

as smooth as the tracks made with the nanodiamonds. This is not obvious from visual

inspection of the false colored surface images in Figure 6.7.

145

0.1

1

10

100

1000

0.001 0.01 0.1 1


Ave

rage

Pow

er D

ensi

ty (n

m2 µ

m)


[1 µm][10 µm][100 µm]

[Periodicity]

[1000 µm]


Figure 6.9 – 1-D PSD analyzed in the vertical direction (parallel to the MRF grooves) inside EFDS-1 MRF spots made with 0, 0.007 and 0.1-vol% NDP nanodiamond MR fluids and compared to the initial pitch polished surface.

146

6.3.2 Surface roughness and drag force

Figures 6.10 and 6.11 give the average areal p-v and rms surface roughness

inside MRF spots for all glasses made with 0 – 0.01-vol% NDP nanodiamond MR

fluids and the corresponding drag forces. We find that the linear trend lines drawn in

both figures have negative slopes with very high confidence levels.


0

10

20

30

40

50

60

70

80

90

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0Drag Force (N)

Aver

age

area

l p-v

sur

face

roug

hnes

s (n

m)

Figure 6.10 – Average areal p-v surface roughness inside MRF spots made with 0 – 0.01 vol% NDP nanodiamond MR fluid versus the corresponding drag force. [Data tabulated in Table C.12 in Appendix C].

147


0

1

2

3

4

5

6

7

8

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0Drag Force (N)

Ave

rage

are

al rm

s su

rface

roug

hnes

s (n

m)

Figure 6.11 – Average areal rms surface roughness inside MRF spots made with 0 – 0.01 vol% NDP nanodiamond MR fluid versus the corresponding drag force. [Data tabulated in Table C.12 in Appendix C].

The same data shown in the previous two figures are plotted again in Figures

6.12 and 6.13, but this time linear trend lines are drawn for the individual glass types.

It is very interesting to observe in Figure 6.12 that, with the exception of LHG-8,

there is a positive linear correlation between the p-v surface roughnesses and drag

force, although the confidence levels are all very low (less than 80%). All of the

linear trend lines have positive slopes in Figure 6.13 for the rms surface roughness

data and drag force. The confidence level for FD-60 is 90%, but it is less than 80%

for the remaining five glasses. In general it appears that making a prediction of the

resulting surface roughness inside MRF spots with knowledge of the drag force is

difficult for one glass type and varying nanodiamond concentration. Based on the

data shown in Figures 6.14 and 6.15 we find that drag force is a much better predictor

of resulting surface roughness inside MRF spots if we use constant nanodiamond

concentration and vary the glass type.

148

LHG-8R2 = 0.09

FSR2 = 0.08

BK-7R2 = 0.17

FCD-1R2 = 0.30

FD-60R2 = 0.38

0

10

30

40

50

60

70

80

90

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

Drag Force (N)

Av

area

l p-v

sur

face

roug

hnes

s (n

m)

EFDS-1R2 = 0.10

20erag

e

Figure 6.12 – Average areal p-v surface roughness inside MRF spots made with 0 – 0.01 vol% NDP nanodiamond MR fluid versus the corresponding drag force. Linear trend lines are drawn for the individual glass types. [Data tabulated in Table C.12 in Appendix C].

149

LHG-8R2 = 0.03

FSR2 = 0.09

BK-7R2 = 0.39

FCD-1R2 = 0.59

FD-60R2 = 0.56

0

1

2

3

4

5

6

7

8

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0Drag Force (N)

Ave

rage

are

al rm

s su

rface

roug

hnes

s (n

m)

EFDS-1R2 = 0.04

Figure 6.13 – Average areal rms surface roughness inside

MRF spots made with varying NDP nanodiamond concentration (0 – 0.01 vol%) MR fluid versus the corresponding drag force. Linear trend lines are drawn for the individual glass types. [Data tabulated in Table C.12 in Appendix C].

The best fit we found for comparing p-v surface roughness to drag force is

[1/Fd] across glass types for each specific nanodiamond concentration. The power

trend lines in Figure 6.14 all have confidence levels higher than 90%. The abrasive

free MR fluid fit best to [1/Fd3]. Additional p-v surface roughness values for spots

made with UK-Low, UK-Medium A and UK-High nanodiamonds are compared to

drag force in section C.4 in Appendix C.

Similar curve fits were found for the rms surface roughness data and drag

force across glass types in Figure 6.15. The dependencies varied from glass type to

glass type but in general are all proportional to [1/Fd].

150

Abrasive free

p-v α 1/Fd3

R2 = 0.90Confidence Level (CL): 99%

0.001-vol% p-v α 1/Fd

R2 = 0.80CL: 98%

0.003-vol%p-v α 1/Fd

R2 = 0.65CL: 95%

Figure 6.14 – Average areal p-v surface roughness inside MRF spots made with varying glass type versus the corresponding drag force. Linear trend lines are drawn for the different NDP nanodiamond concentrations (0–0.01-vol%) in the MR fluid. [Data tabulated in Table C.12 in Appendix C].


R2 = 0.87CL: 99%


R2 = 0.56CL: 90%

0

10

20

30

40

50

60

70

80

90

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0Drag Force (N)

Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)


R2 = 0.86CL: 99%

151

Abrasive Freerms α 1/Fd

3

R2 = 0.90CL: >99%

0.001-vol%rms a 1/Fd

R2 = 0.69CL: 95%

0.003-vol%rms α 1/Fd

R2 = 0.66CL: 95%

Figure 6.15 – Average areal rms surface roughness inside MRF spots made with varying glass type versus the corresponding drag force. Linear trend lines are drawn for the different NDP nanodiamond concentrations (0-0.01-vol%) in the MR fluid. [Data tabulated in Table C.12 in Appendix C].

6.3.3 Surface roughness and glass mechanical properties

Buijs and Korpel-van Houten5 determined a relationship between the

mechanical properties of glass and the p-v surface roughness after lapping, shown in

Equation 6.1. We plotted the same relationship with our data in Figure 6.16. The

relationship is valid for our data; the confidence level in the linear fit is 80%. Similar

to the mechanical figure of merit and material removal rate, we assume that the

surface roughness is not wholly dependent on the glass mechanical properties. We

can improve the fit slightly by using the near surface mechanical properties in place

of the bulk properties, Es and Hs. Using this substitution in Figure 6.17 we find that

the confidence level increases to 85%.


R2 = 0.52CL: 85%


R2 = 0.48CL: 80%

0

1

2

3

4

5

6

7

8

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

Drag Force (N)

Ave

rage

are

al rm

s su

rface

roug

hnes

s (n

m)


2

R2 = 0.51CL: 85%

152

vHEvp

21

∝− (6.1)


0

5

10

15

20

25

30

0 0.5 1 1.5 2 2.5

E1/2/Hv

Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)

FSBK-7

LHG-8

FCD-1

FD-60

EFDS-1

Figure 6.16 – Average areal p-v surface roughness values

versus E1/2/Hv. The spots were taken using 0.01-vol% NDP MR fluid. The mechanical properties were measured on the bulk glass. [Data tabulated in Table C.12 in Appendix C].

153


0

5

10

15

20

25

30

0 0.25 0.5 0.75 1 1.25 1.5 1.75 2

Es1/2/Hs

Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)

FSBK-7

LHG-8

FCD-1

FD-60

EFDS-1

Figure 6.17 – Average areal p-v surface roughness values versus Es

1/2/Hs. The spots were taken using 0.01-vol% NDP MR fluid. The near surface mechanical properties were measured in MR fluid supernatant environment. [Data tabulated in Table C.12 in Appendix C].

If we incorporate a drag force term into this model, we can significantly

improve the linear fit (see Figure 6.18). We found in the previous section that drag

force is inversely proportional to the p-v surface roughness when comparing spots

taken on different six glasses under the same conditions. We find that with this

relationship the softer glasses that polish faster also have higher surface roughness

values and at the same time the glasses that require lower drag forces to remove

material also have higher removal rates and higher surface roughness values. This

relationship is valid only when comparing the surface roughness inside MRF spots

taken with the same fluid. It is not valid when comparing removal rates and surface

roughness values on the same glass with changing fluid properties.

154


0

5

10

15

20

25

30

0 0.25 0.5 0.75

(Es1/2/Hs) (1/Fd)

Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)

FS

BK-7

LHG-8

FCD-1

FD-60

EFDS-1

●

1

Figure 6.18 – Average areal p-v surface roughness versus (Es

1/2/Hs)⋅(1/Fd). Spots were taken using the 0.01-vol% NDP fluid. [Data tabulated in Table C.12 in Appendix C].

In this chapter we have used surface roughness and PSD analysis as tools to

help us better understand the MRF process. We have verified that in the presence of a

magnetic field the magnetic CI particles are pulled toward the rotating wheel leaving

the non-magnetic nanodiamond particles in a thin film at the surface of the ribbon.

This allows even very small amounts of nanodiamonds to be involved in the removal

process. These nanodiamonds act as a lubricant to aid in increased efficiency of

removal and in some cases lead to smoother surfaces. We also hypothesize that as we

increase the nanodiamond concentration the thickness of the nanodiamond layer

increases and the CI particles are less likely to touch the surface of the glass and act

more as a polishing lap as opposed to actually removing material. Finally the average

areal p-v surface roughness was found to be proportional to the near surface glass

mechanical properties using a relationship used for the lapping of glass. We

improved upon this relationship by adding a term for inverse drag force between the

MR fluid ribbon and the optic being polished.

155

References


2. J. C. Lambropoulos, S. D. Jacobs, and J. Ruckmann, "Material removal mechanism from grinding to polishing," Ceramic Transactions 102, 113-128 (1999).

3. J. C. Lambropoulos, F. Yang, and S. D. Jacobs, "Toward a Mechanical Mechanism for Material Removal in Magnetorheological Finishing," presented at the Optical Fabrication and Testing Workshop, 1996.


5. M. Buijs and K. Korpel-van Houten, "Three-Body Abrasion of Brittle Materials as Studied by Lapping," Wear 166(2), 237-245 (1993).

156

Chapter 7

Summary

Throughout this thesis, the interaction between the MR fluid and the glass

surface has been discussed. In Chapter 3 we examined the particle size, surface

charge and surface texture of the CI and nanodiamond particles that make up MR

fluid. Magnetic CI particles have a median particle size of 3.5µm and range in size

from 2 – 5 µm. Through SEM imaging we showed that the larger end of the CI

particle size distribution was due to fused, non-spherical CI particles. We measured

the zeta potential of the CI particles in various environments and found that the

surface charge in MR carrier fluid at pH 10 was -30mV. The non-magnetic

nanodiamond particles we used for this thesis work have median aggregate sizes

between 19 and 54nm in diameter. The nanodiamonds that were supplied in

suspension had negative surface charge in the MR carrier fluid environment. The dry

nanodiamond powder, NDP, was positively charged in the same environment. We

hypothesized that the difference in charge between the CI and nanodiamond particles

influences MRF material removal. We showed that increasing peak removal rates

were linearly related to the increasing (less negative) surface charge of the

nanodiamond particle. We introduced an attraction/repulsion hypothesis for the CI

and nanodiamond particles in the MR fluid. In this hypothesis we suggest that

positively (or neutral) charged nanodiamond particles will coat (weakly adhere)

themselves to the negatively charged CI particles at the top of the MR fluid ribbon in

the polishing zone. These nanodiamonds will create a lubricating layer on the CI

particles and increase the efficiency of MRF material removal. Alternatively the

negatively charged nanodiamonds will have a weak repulsion to the negatively

charged CI particles and will not coat the CI particles and therefore the material

removal will not be as efficient. In both scenarios, after the CI and nanodiamonds

leave the polishing zone, the nanodiamonds will detach from the CI particles due to

mechanical milling as the fluid circulates through the delivery system.

157

In Chapter 4 we focused on learning more about the near surface mechanical

properties (Young’s modulus and nanohardness) of the glass modified surface layer

after exposure to MR fluid during polishing. We implemented the Nano Indenter XP

nanohardness tester with a Berkovitch diamond tip to make nanoindentations on all

six of our glasses. We made measurements on the dry surface and compared these to

measurements made on surfaces that were immersed in DI water and MR fluid

supernatant. We found that the MR fluid supernatant created a softened modified

layer for five out of the six glasses. The Young’s modulus and nanohardness of the

titanium phosphate glass, EFDS-1, did not change with exposure to MR fluid

supernatant. The near surface Young’s modulus and nanohardness values we

obtained through these measurements were implemented in our MRF material

removal rate model in Chapter 5.

In Chapter 5 we introduced our new glass MRF material removal rate model.

This model incorporates terms for mechanics, polishing particle properties and

chemistry. Through experiments done to determine the terms in our new model, we

gained an improved understanding of many facets of MRF with nanodiamonds.

The first term in our model is a mechanical figure of merit term using the near

surface mechanical properties of the glass measured in MR fluid supernatant, together

with the bulk fracture toughness (Es/(Kc⋅Hs2)). Term 1 exhibits a strong positive

linear correlation with peak removal rates for our nanodiamond MR fluid. We

observed an approximate 6% improvement in the R2 value for the fit of Term 1 in our

model compared to the same relationship using the bulk mechanical properties.

The second term in our model is a modified Preston’s equation ((Fd/A)⋅ v).

We verified that peak removal rate increases linearly with wheel speed and has an

inverse relationship with the spot contact area. Most of our work with this term

focused on the drag force. We found that peak removal rate increases linearly with

drag force as the concentration of nanodiamonds increases in the MR fluid. This

relationship however is specific to glass type. By measuring Fd on each glass type,

we observed that the silicate glasses have a positive linear correlation between

158

MRRpeak and Fd, but that the phosphate glasses have a negative linear correlation for a

given nanodiamond MR fluid. This relationship is true regardless of the

nanodiamond manufacturer, size or friability level. This obviously affects the quality

of the overall model, but if the drag force term is multiplied by 1/Hs, then we observe

strong positive linear correlations for all six glasses. We did not have to adjust our

model because inverse near surface nanohardness is in Term 1.

Polishing particle properties make up the third term of our model

(Bndφnd-1/3Cnd

1/3 + BCIφCI4/3CCI). We validated the power 1 dependence for CCI, but

otherwise we chose to keep the size and concentration of the CI particles constant for

our experiments. By varying nanodiamond size from 19 – 54nm and concentration

from 0 – 0.01-vol% we observed a very strong positive linear correlation between

Term 3 and peak removal rate for each of the glass types.

The fourth term in our model is a term for glass chemical durability which is a

function of the MR fluid pH (Ds3/10(pHMRF)). The chemical durability of each of our

glasses was determined experimentally by measuring the percent weight loss of

ground glass subjected to MR fluid supernatant at an elevated temperature for one

hour. We determined the dependence on pH by varying the pH of the MR fluid

supernatant from 9 -12 and measuring the percent weight loss in tap water at pH 7.

Therefore each glass has its own relationship with the varying pH of MR fluid. The

power relationship was determined experimentally. The MRF material removal

mechanism is not purely chemical, and therefore Term 4 cannot be expected to

accurately describe removal by itself, but combined with mechanical terms, it can

improve the process understanding.

The final term that we introduced is the glass average single bond strength

incorporated into an Arrhenius-like equation (e-(sbs/bRT)). Term 5 quantitatively

describes how material is sheared off faster/easier for glasses with weaker bond

strengths compared to glasses with stronger bonds.

In Chapter 6 we discussed the role of nanodiamonds in the polishing zone.

We showed evidence through surface texture and roughness results that the abrasive

159

free MR fluid has a micro-gouging mechanism of removal. Once nanodiamonds are

added to the MR fluid, this mechanism changes to a micro-lateral-fracture

mechanism, which is more efficient at removing material. We also used the resulting

surface roughness inside MRF spots made with varying friability nanodiamonds to

support our hypothesis that the nanodiamonds are transported to the surface of the

MR fluid ribbon in the polishing zone based on the known properties of the

nanodiamonds. Finally we modified a conventional lapping model by substituting the

near surface mechanical properties and incorporating 1/Fd to find strong correlations

with the p-v surface roughness inside MRF spots.

The MRF material removal mechanism is very complex. After completion of

this thesis work, many questions have been answered, but many still remain.

160

Chapter 8

Future Work

There are many areas of this work that can be extended in the future. Several

of these topics are summarized here.

In our work we evaluated nanodiamonds with three levels of friability. Our

results, shown in Figure 8.1, indicate a parabolic dependence of removal rate on

friability level. It would be very interesting to incorporate a term for nanodiamond

friability in the MRF material removal model. This would require additional work

using additional nanodiamond friability levels, making a connection to near surface

mechanical properties. In turn this information along with surface roughness data

could provide additional insight for selecting the “optimal” nanodiamond for use in

MRF polishing of specific glass.

0

5

10

15

20

25

30

0 20 40 60 80

Friability Index

Pea

k R

emov

al R

ate

( µm

/min

)

Low Medium HighFS

BK-7FD-60

EFDS-1FCD-1

LHG-8

100

Figure 8.1 – Peak removal rate for FS and BK-7 versus nanodiamond friability. The three MR fluids contained 0.01-vol% nanodiamonds.

161

The nanodiamond and CI attraction/repulsion hypothesis should also be

explored more in the future. Experiments could be conducted to systematically vary

the surface charge on the nanodiamond particles that have the same size and friability

level. These results would help to prove or disprove the hypothesis suggested in this

thesis that the highest removal rates occur when the CI and nanodiamonds are

oppositely charged and are attracted to each other. At the same time these

experiments would enable us to learn more about producing a nanodiamond that

works the best with CI particles in MR fluids.

Examination of the CI particles themselves could help to determine how the

surface texture or different surface treatments affect MRF material removal rates. It

may also be possible to bond non-magnetic polishing abrasives to the CI particles. In

this case the non-magnetic polishing abrasives would no longer be free particles in

the MR fluid, but would be permanently attached to the CI particles.

Throughout this thesis we kept certain variables in our model constant,

specifically the size and concentration of CI and the temperature. We did not vary the

size of the CI particles in this work due to the high cost of obtaining suitably sieved

batches from the broad distribution in the commercial product, but this information

would be very valuable to further understanding the MRF process. We briefly looked

at the effects of varying the CI concentration of the MR fluid in Chapter 5. Unlike

traditional optical finishing processes the magnetic field and fluid viscosity play very

important roles. Additional work should be done to gain a better understanding of

how to model the combined effects of CI concentration, the magnetic field and

resulting viscosity in the model. Temperature was not varied in our work. We

predict that increasing temperature will also increase peak removal rate as indicated

in the model. Increasing temperature can potentially be very tricky. The viscosity of

the MR fluid is strongly temperature dependent. Glass and many other materials

change shape with exposure to increased temperature, which may influence the

metrology measurements. In addition care should be made to make sure that, if a

162

material has a high expansion coefficient, the temperature gradient from the air to the

“warm” MR fluid is not high enough to shock and fracture the material.

All of the work done with this model is based on aqueous nanodiamond MR

fluids, but in the future it could also be extended to other non-magnetic polishing

abrasives such as cerium oxide and non-aqueous MR fluids. In addition the MRF

material removal rate model can be extended to other materials such as the hard

ceramics or soft polymers.

A model for residual p-v surface roughness inside MRF spots was introduced.

This p-v surface roughness model contains terms for the near surface mechanical

properties and drag force. Additional work can be done to determine additional terms

for the surface roughness model, for example chemical durability or nanodiamond

friability. More knowledge about the factors that affect peak removal rate and the

residual surface roughness can only improve how fast and how smooth MRF is

capable of polishing optical surfaces.

163

Appendix A

Particle size and zeta potential All of the particle size and zeta potential measurements for this research are

made using the Colloidal Dynamics AcoustoSizer IIs.1 The AcoustoSizer measures

particle size and zeta potential using an electroacoustic method.

The electroacoustic method involves applying an alternating electric field to a

colloid suspension, forcing the charged particles to move back and forth between the

electrodes. The ultrasound wave (frequency >20,000 Hz) generated from the

particles by an electric field is called the Electrokinetic Sonic Amplitude (ESA)

effect. The ESA sound wave is detected by a piezoelectric transducer. Using the

magnitude and phase of the resulting ultrasound wave, the charge and size of the

particle are calculated by the AcoustoSizer software.1

In addition to the ESA measurements, the AcoustoSizer also can be used to

perform ultrasonic attenuation measurements. In this method the ultrasound is

generated by the external transducer, not the particles. A large sinusoidal voltage

pulse is applied, causing the transducer to expand and contract. The result is the

generation of an ultrasound wave which passes through the colloid to another

transducer. The ultrasound wave causes the particles to move back and forth if the

density of the particles is different from that the suspending medium. The motion of

the particles causes some attenuation of the ultrasound wave. Using the AcoustoSizer

software, one can determine the particle size from the change in frequency of the

ultrasonic wave.1

One of the main advantages of AcoustoSizer over light scattering

measurements is the ability to measure opaque dispersions at higher concentrations.

Particle sizes from 0.01µm to 1000µm can be determined with the electroacoustic

method.2

Zeta potential is the charge a particle acquires under flow within in a

particular medium. The zeta potential of the particle is the potential measured at the

164

slipping plane (see Figure A.1) in a suspension. A negatively charged particle will

attract positively charged ions that are present in a suspension. Ions that are strongly

bound to the particle form the Stern layer. Ions that are less firmly attached, but still

move or travel with the negatively charged particle, form a diffuse region. The

boundary of this diffuse region is referred to as the slipping plane.3

Particle with negative surface charge

Slipping plane

Figure A.1 – Drawing of a negatively charged particle in suspension. [Adapted from Ref. 3].

The magnitude of the zeta potential is a strong indication of the stability of the

colloid suspension. If the zeta potential is >|30| mV the particles will remain stable in

the suspension and not form agglomerates due to their strong similar charges. For

lower values of zeta potential, agglomeration may occur, causing the suspension to

become unstable.3

The pH of the host suspension has a large influence on the zeta potential of a

particle. The zeta potential of a particle tends to go more negative when it is placed

in basic solutions compared to more acidic solutions, where the charge can be

neutralized and eventually become positive. The point at which there is no net charge

on a particle is called the iso-electric point (IEP).3

165

References

1. AcoustoSizer IIs, Colloidal Dynamics Inc., 11 Knight Street Building E18,

Warwick RI 02886.

2. I. D. Morrison and S. Ross, Colloidal Dispersions: Suspensions, Emulsions, and Foams (John Wiley and Sons, Inc., New York, 2002).

3. "Chapter 16: Zeta Potential Theory " in ZetaSizer Nano Series User Manual (Malvern Instruments Ltd., Malvern, Worcestershire WR14 1XZ, UK, 2004).

166

Appendix B

Fluid Jet Polishing Fluid Jet Polishing (FJP) is a novel deterministic sub-aperture polishing

process developed at Delft University of Technology.1 The FJP process consists of

forcing slurry containing abrasive particles through a nozzle at a fixed distance and

angle at the substrate to be polished. Removal is characterized with FJP footprints

which are created by introducing the abrasive slurry to the substrate for a given

amount of time while keeping the distance, angle and substrate stationary. Using the

depth of the footprint, the removal rate can be determined for these conditions and

used to deterministically polish a similar substrate with suitable CNC controls.2

The goal of our work with FJP is to learn more about the unique behavior of

the CI and nanodiamond particles used in the MRF process by working with them in

another polishing process. To be meaningful, we use the same glass set as that in our

MRF experiments. The polishing particles that we use in our FJP experiments are CI

particles, silicon carbide (SiC) and 20nm low, medium and high friability

nanodiamonds, denoted again as UK-Low, UK-Medium and UK-High. These are the

same particles (except for SiC) that are studied throughout this thesis.

In section B.1 we discuss the FJP experimental set up, then in section B.2 we

review our glass set. We use five different slurries for this work. The slurry

containing only CI particles is discussed in section B.3, the slurry containing SiC

particles is discussed in section B.4 and the three varying friability nanodiamond/CI

slurries are discussed in section B.5. In section B.6 we compare the performance of

all slurries using FJP, and we relate these observations to MRF.

B.1. Fluid Jet Polishing (FJP) experimental set-up Planning for the FJP experiments was begun in October 2004, and they were

carried out by the author in June 2005 at Fisba Optik3 located in St. Gallen

167

Switzerland using their experimental set up (Figure B.1). Slurry is loaded into the

cylindrical tank with the system in bypass. The fluid is continuously circulated in

bypass which keeps the abrasive particles from settling. The only time the system is

not in bypass is to take footprints. The path the slurry takes while generating

footprints is through the pump, past the pressure sensor, into a delivery system and

then through the rotating nozzle.4 The nozzle rotates to ensure uniform slurry flow

and to inhibit preferential wear inside the nozzle. A software program is used to

move the chuck-mounted sample into the slurry jet for a specified amount of time and

then move it back when the footprint is complete.5 The slurry is collected through a

drain running back out of the machine into the cylindrical tank. Once the footprint is

taken the system is placed back into bypass.

IN

OUT

Sample chuck

nozzle

drain

Bypass valve

Pressure sensor

Cylindrical tank

Peristaltic pump

Figure B.1 – FJP experimental set up. Typically 8L of slurry is continuously circulated through this system. The delivery system and nozzle are part of a prototype Zeeko4 FJP machine.

B.2. Experimental glass set The seven glasses used in these FJP experiments are chosen based on their

mechanical properties. Six of the seven glasses are the same glasses used in the MRF

experiments at the University of Rochester and taken to Switzerland from Rochester;

the samples of NSF-6 are supplied by Fisba. NSF-6 is the Schott6 equivalent of

Hoya’s7 FD-60. The seven optical glasses and their mechanical properties as

168

measured at the University of Rochester are in Table B.1, rank ordered according to

increasing mechanical figure of merit (FOM).

Glass Glass Type

Source and Melt Number E [GPa] Kc [MPa m1/2] Hv [GPa]

Mechanical FOM

E7/6/(Kc*Hv23/12)

FS Fused Silica Corning 69 0.75 7.5 4.0 BK-7 Borosilicate Schott 81 0.80 6.0 7.0

FD-60 Titanium Silicate Hoya:20A-3419-39 93 0.69 6.3 8.6

NSF-6 Niobium Titanium

Alkali Silicate

Schott 52 0.56 3.8 14.0

EFDS-1 Titanium Phosphate Hoya:03A-4808-33 96 0.59 5.3 14.1

LHG-8 Phosphate Hoya 62 0.52 3.7 19.6 FCD-1 Fluorophosphate Hoya 73 0.47 4.0 22.1

Table B.1 Optical glass mechanical properties rank

ordered by increasing mechanical figure of merit.

B.3. Carbonyl iron (CI) slurry

Eight liters of fluid are loaded into the machine at 2pm on day 1 of the

experiment. The fluid consists of 4L DI water, 4L blue buffer solution and 875g of

CI (10 wt% or 1.5 vol% CI). There are no polishing abrasives other than the CI

particles themselves. The pH of the fluid is measured periodically throughout this

experiment. We determined later through SEM images (see section B.5) of this slurry

that there was some contamination with cerium oxide particles used in the previous

experiment on the FJP machine.*

Two different programs are used to allow for two different spotting times.5

The softer glasses (LHG-8, FD-60, FCD-1 and NSF-6) require 40 second dwell times

and the harder glasses (FS, BK-7 and EFDS-1) require 80 second dwell times. Four

* We believe that even with the contamination the CI particles are the primary particles involved in the removal process, based on a comparison between the removal rate results in this section and the initial footprints taken before the nanodiamonds were added in the nanodiamond slurry experiments.

169

sets of footprints are taken over a period of 24 hours, two on the first day and two on

the second day.

The operating conditions are: 2mm inner diameter nozzle, 20mm standoff

distance, 45 degree impact angle and 10% pump speed. [The pressure fluctuates

between 8-9 barr using a 10% pump speed.] Booij2 determined that the velocity of

the slurry out of the nozzle scaled linearly with the pressure. According to her results

the slurry velocity range for our experiments is 33-37 m/s. In this experiment all

parameters are kept as constant as possible. Nothing is added to the fluid during this

experiment, and the CI that remains in the polishing chamber at the end of the day is

collected with a magnet and put back into circulation. The water level was seen to

decrease over the experiment, but no more water is added to the slurry, the pressure is

not affected by the evaporation.

The footprint depths are measured using a Zygo interferometer with a Fisba

transmission flat and Fisba software. The technique for determining the ddp is to

measure only the area directly around the footprint and to use the areal p-v

measurement as the ddp.

The footprints are characterized for surface roughness using the Taylor

Hobson CCI 30008 at the Laboratory for Laser Energetics at the University of

Rochester. We observe all the footprints in the same orientation where the slurry

flow is oriented from the bottom to the top of the viewing screen. An example of a

footprint and the measurement procedure is shown in Figure B.2.

Figure B.3 is a plot of the peak removal rates versus elapsed time for all seven

glass types. There is a significant drop between the first and second set of spots, a

plateau after the evening hours and then a second drop on the second day. The slurry

pH values are recorded throughout the experiment and their values are included in

Figure B.3. It is hypothesized that the drop in removal rate is strongly correlated with

the drop in pH.

In order to determine which glass type has the largest percent change, we

normalize the data to the first footprint for each glass type (Figure B.4). LHG-8 is the

170

most sensitive glass to whatever is causing this drop (approx. 60%) in removal. NSF-

6 appears to be the least sensitive to the change (approx. 25%).

vertical

horizontal

CCI 3000 Areal scan: 350x350µm50mm

70mm Flow

Figure B.2 – FJP footprint shape and orientation for surface roughness measurements using the CCI 3000.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0 200 400 600 800 1000 1200 1400 1600 1800Elapsed Time (minutes)

Peak

Rem

oval

Rat

e ( µ

m/m

in)

9.00

9.05

9.10

9.15

9.20

9.25

Slur

ry p

H

LHG8 FD60 EFDS1 FS BK7 FCD1 NSF6 Slurry pH

Day 1 Day 2Slurry left milling in bypass overnight

Figure B.3 – Peak removal rate versus elapsed time for the CI slurry.

171

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0 200 400 600 800 1000 1200 1400 1600Elapsed Time (minutes)

Nor

mal

ized

Pea

k R

emov

al R

ate

9.00

9.05

9.10

9.15

9.20

9.25

Slur

ry p

H

LHG8 FD60 EFDS1 FS BK7 FCD1 NSF6 Slurry pH

Day 1 Day 2Slurry left milling in bypass overnight

Figure B.4 – Normalized peak removal rate versus elapsed time for the CI slurry.

The average areal p-v surface roughness data for the footprints taken on the

seven glasses are plotted in Figure B.5. The initial pitch polished surface roughness

values are included for comparison; their values are listed in Table B.2. Selected

roughness maps are shown in Figure B.6. The error for EFDS-1 is larger compared to

the other six glasses. This error is most likely due to the presence of large pits on the

surface inside the footprints (see Figure B.6b). There is typically only one of these

pits in a single field of view, but pitting is frequent enough that we are not able to

avoid them. Fused silica and BK-7 have the lowest p-v surface roughness values,

comparable to their pitch polished values throughout the entire experiment. Peak

removal rate and surface roughness values for this experiment are listed in Table B.2.

172

Table B.2 – Removal rate and surface roughness data for FJP footprints made with the CI slurry.

Figure B.5 – Average areal p-v surface roughness for footprints taken with the CI slurry.

-6 Initial 14.00 1.83 1.08 0.209.23 90 0.28 1.00 42.80 2.04 1.85 0.329.13 210 0.20 0.729.11 1315 0.18 0.63 25.54 5.02 1.70 0.069.04 1550 0.11 0.41 30.60 3.97 1.46 0.09

EFDS-1 Initial 12.01 3.96 0.87 0.019.23 50 0.26 1.00 84.80 38.74 2.92 1.129.13 175 0.15 0.57 104.90 24.79 2.76 0.409.11 1230 0.12 0.45 93.78 23.47 2.55 0.629.04 1518 0.07 0.29 75.06 52.45 2.26 1.01

LHG-8 Initial 14.96 0.47 1.07 0.019.23 30 1.06 1.00 69.58 24.00 3.16 0.219.13 165 0.42 0.40 48.44 1.76 2.77 0.309.11 1220 0.39 0.36 55.92 9.83 2.77 0.149.04 1508 0.21 0.19 44.18 13.41 2.54 0.16

FCD-1 Initial 14.06 2.76 1.09 0.019.23 70 0.44 1.00 31.40 1.26 2.75 0.149.13 197 0.29 0.65 33.88 5.76 2.23 0.239.11 1250 0.27 0.62 42.06 13.50 2.38 0.139.04 1541 0.18 0.41 34.92 11.41 2.14 0.06

0

20

40

60

80

100

120

140

0 200 400 600 800 1000 1200 1400 1600


Ave

rage

are

al rm

s su

rface

roug

hnes

s (n

m)

LHG8 FD60 EFDS1 FS BK7 FCD1 SF6

initial surface

LHG8 FD60 EFDS1 FS BK7 FCD1 NSF6

Glass Slurry pH Elapsed time Peak removal rateNormalized peak

removal rate Average areal p-v p-v st. dev. Average areal rms rms st. dev.(min) (µm/min) (nm) (nm)

FS Initial 17.88 6.46 0.80 0.079.23 55 0.03 1.00 14.90 3.50 0.90 0.019.13 183 0.02 0.52 15.28 2.62 0.87 0.019.11 1238 0.02 0.46 16.84 4.34 0.88 0.029.04 1525 0.01 0.30 17.50 5.43 0.86 0.00

BK-7 Initial 13.18 3.92 0.81 0.029.23 62 0.06 1.00 14.50 4.76 0.90 0.019.13 190 0.04 0.63 18.06 6.48 0.93 0.029.11 1245 0.04 0.64 17.14 4.23 0.89 0.019.04 1535 0.02 0.28 15.03 5.38 0.85 0.01

FD-60 Initial 15.66 8.69 0.84 0.039.23 40 0.18 1.00 48.06 12.93 1.16 0.059.13 170 0.08 0.47 37.70 6.97 1.08 0.129.11 1225 0.07 0.39 40.72 13.08 1.04 0.059.04 1512 0.04 0.21 43.90 4.60 1.12 0.09

SFN

173

a.) LHG-8 (phosphate)

c.) FD-60 (titanium silicate)

d.) FCD-1 (fluorophosphate)

b.) EFDS-1 (titanium phosphate)

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

2

4

6

8

10

12

14

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

5

10

15

20

25

30

35

40

45

flow

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

1

2

3

4

5

6

7

8

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

5

10

15

20

25

30

35

40

45

50

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

1

2

3

4

5

6

7

8

9

10

11

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

5

10

15

20

25

30

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0. 51

0.2

0. 52

0.3

0. 5

nm

30

1

2

3

4

5

6

7

8

9

10

11

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

5

10

15

20

25

30

35

40

45

50

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

10

20

30

40

50

60

70

80

second footprint

initial surface first footprint




RMS: 1.07 + 0.01 nmPV: 14.96 + 0.47 nm

RMS: 3.16 + 0.21 nm PV: 69.58 + 24.0 nm

RMS: 0.87 + 0.01 nm PV: 12.01 + 3.96 nm

RMS: 2.92 + 1.12 nmPV: 84.80 + 38.7 nm

RMS: 2.76 + 0.40 nmPV: 104.9 + 24.8 nm

RMS: 0.84 + 0.03 nmPV: 15.66 + 8.69 nm

RMS: 1.16 + 0.05 nm PV: 48.06 + 12.9 nm

RMS: 1.09 + 0.01 nmPV: 14.06 + 2.76 nm

RMS: 2.75 + 0.14 nm PV: 31.40 + 1.26 nm

174

e.) FS (fused silica)

f.) BK-7 (borosilicate)

g.) NSF-6 (Niobium titanium alkali silicate)

Figure B.6 – False color images of the initial pitch polished

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

2

4

6

8

10

12

14

16

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

01234567891011121314


RMS: 0.80 + 0.07 nmPV: 17.88 + 6.46 nm

RMS: 0.90 + 0.01 nm PV: 14.90 + 3.50 nm

surfaces and the surfaces inside the FJP footprints (1st set of footprints).

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

2

4

6

8

10

12

14

16

18

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

1

2

3

4

5

6

7

8

9


RMS: 0.81 + 0.02 nmPV: 13.18 + 3.92 nm

RMS: 0.90 + 0.01 nm PV: 14.50 + 4.76 nm

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 m m

m m

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

5

10

15

20

25

30

35

40

nm

0

2

4

6

8

10

12

14

16


RMS: 1.08 + 0.20 nmPV: 14.00 + 1.83 nm

RMS: 1.85 + 0.20 nm PV: 42.80 + 2.04 nm

175

0.5

1

1.5

2

2.5

3

3.5

4

0 200 400 600 800 1000 1200 1400 1600


Ave

rage

are

al rm

s su

rface

roug

hnes

s (n

m)

LHG8 FD60 EFDS1 FS BK7 FCD1 SF6

initial surface

silicates

phosphates

NSF6

Figure B.7 – Average areal rms surface roughness for footprints taken with CI slurry.

The average areal rms surface roughness versus elapsed time is shown in

Figure B.7. Similar to the p-v surface roughness data, the EFDS-1 has larger error

compared to the other six glass types, and the FS and BK-7 are the smoothest,

comparable to their pitch polished surfaces. One very interesting trend is that the

silicates consistently have lower rms surface roughness values compared to the

phosphates glasses. Typically silicate glasses are harder than phosphate glasses

which suggests plotting the rms surface roughness values against 1/Hv1/2 (see Figure

B.8). Lambropoulos et al.9 found a very high correlation when lapping optical glass.

The pitch polished data have a very high R2 value compared to the first FJP footprint

data. According to Wheeler et al.10 the R2 value for the FJP data have greater than

90% correlation.

176

y = 11.836x - 3.3658R2 = 0.5539

y = 2.1044x - 0.0077R2 = 0.9336

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

0.35 0.37 0.39 0.41 0.43 0.45 0.47 0.49 0.53 0.55

1/Hv1/2 (1/GPa1/2)

Ave

rage

are

al rm

s su

rface

roug

hnes

s (n

m)

0.51

FJP CI slurry Pitch polished Linear (FJP CI slurry) Linear (Pitch polished)

FS FD-60 BK-7 EFDS-1 FCD-1 SF-6 LHG-8N-SF6

Figure B.8 – Average areal rms surface roughness for the pitch polished surfaces and the first FJP footprints versus 1/Hv

1/2.

The orthogonal PSD plots comparing the first footprint for LHG-8 and FS are

shown in Figure B.9. The horizontal and vertical curves overlap for the entire spatial

frequency range for all seven glasses. Due to the 45 degree angle of incidence we

hypothesize that there should be directionality to the surface texture inside the

footprints. This is not the case; no periodic features are found inside the footprints.

177

ig

.4. Silicon carbide (SiC) slurry on all seven glasses (not measured over time)

re

e rem

0.1

1

10

100

1000

0.001 0.01 0.1 1


Ave

rage

Pow

er D

ensi

ty (n

m2 µ

m)

CI alone footprint - Horizontal Line Scans CI alone footprint - Vertical Line ScansPitch Polished

FS10000

0.1

1

10

100

1000

0.001 0.01 0.1 1


Aver

age

Pow

er D

ensi

ty (n

m2 µ

m)

CI alone footprint - Horizontal Line Scans CI alone footprint - Vertical Line ScansPitch Polished

LHG-810000

F ure B.9 – PSD plots for LHG-8 and FS measured inside

the CI slurry footprint and the adjacent pitch polished surface.

B

We chose to use SiC in our FJP experiments because we wanted to compa

th oval of the round CI particles (~3µm) to irregularly shaped particles of similar

size. The SiC particles (~5µm) that we use for this experiment are slightly larger than

the CI particles.

178

One footprint is taken on each of the seven glasses. The slurry pH is 9.22

throughout the experiment. The SiC slurry consists of 4L DI water, 4L blue buffer

solution and 859g 5µm SiC (10 wt% or 3.36 vol% SiC). Results in this section

compare the performance of ~5µm irregularly shaped SiC particles to the ~3µm

spherical CI particles (see SEM images included in Figure B.10). The peak removal

rates measured for each of the glasses are plotted versus the approximated glass

mechanical figure of merit (E/HvKc2) and compared to the results of the CI slurry in

Figure B.10. The peak removal rates and linear fit are significantly higher and better

for the SiC slurry compared to the CI slurry. We infer from the data that removal

with the SiC slurry is dominated by mechanics, unlike that for the CI slurry which

may be more heavily influenced by chemical factors in the removal process (i.e. pH).

The difference in particle size and the rounder shape of the CI particles must also play

a role in producing lower removal rates.

SiC

y = 0.2421x - 0.0753R2 = 0.8121

y = 0.0282x - 0.0203R2 = 0.2965

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

0 5 10 15 20 25 30

Mechanical Figure of Merit (E/HvKc2)

Peak

Rem

oval

Rat

e ( µ

m/m

in)

SiC slurry CI slurry Linear (SiC slurry) Linear (CI slurry)

FS BK-7 FD-60 SF-6 EFDS-1 LHG-8 FCD-1NSF-6

5µm

CI

Figure B.10 – Peak removal rate versus glass mechanical figure of merit for SiC and CI slurries. SEM images of SiC and CI are included to the right of the plot. (scale indicated)

179

The p-v and rms surface roughness values are higher inside the SiC slurry

footprints compared to the surfaces inside the CI slurry footprints. Figure B.11 gives

the data and those for the initial pitch polished surface. The surface textures inside

the SiC footprints (Figure B.12) all looked similar to the surfaces inside the CI slurry

footprints shown in Figure B.6 with the exception of the EFDS-1 sample. For this

part the surface inside the CI slurry footprint has many deep holes, whereas the

surface inside the SiC footprint has fewer and shallower holes. The peak removal

rate and surface roughness data for the SiC experiment are listed in Table B.3.

0

20

40

60

80

100

120

140

160

FS BK7 FD60 SF6 FCD1 LHG8 EFDS1

Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)

SiC slurry CI slurry Pitch polished

NSF-6

Figure B.11 – Average areal p-v surface roughness measured inside SiC footprints compared to CI footprints and the pitch polished surface.

Glass Average areal rms Peak Removal Rate Average areal p-v p-v st. dev. rms st. dev.

Table B.3 – Peak removal rate and surface roughness data

for FJP footprints made with SiC slurry.†

† The values for the pitch polished data are listed in Table B.2, the same glass samples were used for both experiments.

(µ m/min) (nm) (nm)FS 0.51 1.80 24.04 3.11 0.06

BK7 1.16 2.37 27.56 2.40 0.14FD60 2.24 2.23 44.42 5.27 0.47NSF6 3.82 3.92 48.26 15.21 0.45FCD1 5.33 4.76 73.64 12.23 0.22LHG8 4.86 4.73 84.88 10.10 0.70

EFDS1 2.57 4.64 118.08 23.35 0.68

180

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

10

20

30

40

50

60

70

80

90

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0102030405060708090100110120130140

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

5

10

15

20

25

30

35

40

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

10

20

30

40

50

60

70

EFDS-1 FD-60 FCD-1 LHG-8

Figure B.12 – False color surface images inside SiC footprints.

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

2

4

6

8

10

12

14

16

18

20

22

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

02

468

101214

161820

222426

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 mm

mm

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

nm

0

5

10

15

20

25

30

35

40

45

FS BK-7 NSF-6

The plot in Figure B.13 indicates that the SiC slurry gives higher rms surface

roughness values compared to the CI slurry. The rms surface roughness values are

plotted versus 1/Hv1/2 along with the data plotted in Figure B.8. Similar to the peak

removal rate data, the rms surface roughness values inside the SiC footprints seem to

be strongly connected to the mechanical properties of the glass. SiC particles of this

size and shape are typically associated with grinding processes.

SiC

CI

y = 2.1044x - 0.0077R2 = 0.9336

y = 11.836x - 3.3658R2 = 0.5539

y = 17.717x - 4.4618R2 = 0.7051

0

1

2

3

4

5

6

0.35 0.37 0.39 0.41 0.43 0.45 0.47 0.49 0.51

1/Hv1/2 (GPa-1/2)

Ave

rage

are

al rm

s su

rface

roug

hnes

s (n

m)

SiC slurry CI slurry Pitch Polished Linear (Pitch Polished) Linear (CI slurry) Li

FS FD-60 BK-7 EFDS-1 FCD-1 SF-6 LNSF-6

Figure B.13 – Average areal rms surface roughnes1/Hv

1/2.

Pitch

0.53 0.55

near (SiC slurry)

HG-8

s versus

181

B.5. Addition of UK Abrasive nanodiamonds to CI slurry Three one-day experiments are conducted to examine the performance of an

MRF-type polishing fluid under FJP conditions. Different nanodiamond abrasives

are studied each day. The sequence of events is as follows: fresh CI slurry (10wt%

CI) is loaded into the FJP machine, one footprint is taken on all seven glasses with the

CI slurry and then 5g of nanodiamond (500mL nanodiamond suspension) are added

to the fluid. Three footprints are taken on each of the seven glasses throughout the

day with the nanodiamond/CI slurry. The data for all of the peak removal rate and

surface roughness measurements for the FJP footprints made with the three

nanodiamond/CI slurries are given in Tables B.4 – B.6 located in section B.5.8.

Because of the strong sensitivity to slurry pH found earlier (see Figure B.4),

the pH of the slurries is monitored throughout each experiment. Figure B.14 is a plot

of the slurry pH as a function of elapsed time. The slurry pH does not drop at the

same rate for the nanodiamond/CI slurries compared to the neat CI slurry. Diamonds

are typically considered to be inert, but these nanodiamonds are synthetic and

dispersed in a solution containing suspending agents that may influence the chemistry

of this dual particle slurry.

Samples of the used CI slurry and the three nanodiamond/CI slurries have

been analyzed using the scanning electron microscope (SEM)11 located at the

University of Rochester.‡ Images of each of the four slurries are given in Figure

B.15. There are no visible differences between the CI particles before and after use in

FJP (see Chapter 3). Based on the energy dispersive spectrum (EDS) information

there is a cerium oxide contaminant in the CI slurry (see discussion at the beginning

of section B.3). There is no ceria detected in the three nanodiamond/CI slurries.

‡ The particles were adhered to carbon tape and then placed on SEM sample stubs. The samples were not coated with gold. The SEM settings for imaging were 2kV accelerating voltage, 5mm working distance, aperture 2 and Gamma 3.

182

9.05

9.1

9.15

9.2

9.25

9.3

0 50 100 150 200 250 300 350 400 450 500 550 600


Slu

rry p

H

Low Friability Medium Friabiliy High Friability CI Slurry (Sect. B.3)

Nanodiamond - Addition TimesLow Friability - 68 minutes

Medium Friability - 85 minutesHigh Friability - 90 minutes

1200

Figure B.14 – Slurry pH as a function of time.

CI slurry Low friability nanodiamond/CI slurry

Medium friability nanodiamond/CI slurry High friability nanodiamond/CI slurry

Figure B.15 – SEM images of FJP slurries.

183

B.5.1. BK-7 Figure B.16 contains a plot comparing the peak removal rate data for the four

slurries on BK-7. The first data points at t ≈ 20 minutes are for the CI slurry without

nanodiamonds. Subsequent data points are for footprints generated after the addition

of nanodiamonds to the slurry. It is difficult to compare the different slurries when

the initial CI slurry footprints have different removal rates, most likely due to

different pH values. The data are normalized to their respective initial CI slurry

footprints and plotted in Figure B.17. BK-7

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0 100 200 300 400 500 600


Pea

k R

emov

al R

ate

( µm

/min

)

Low Friability Nanodiamonds Medium Friability Nanodiamonds High Friability Nanodiamonds

CI Slurry

Nanodiamonds/CI Slurry

Figure B.16 – Peak removal rate data on BK-7 with 3 slurries. BK-7

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500 600


Nor

mal

ized

Pea

k R

emov

al R

ate

Low Friability Nanodiamonds Medium Friability Nanodiamonds High Friability Nanodiamonds CI Alone

1200

CI only

Nanodiamonds added

Figure B.17 – Normalized peak removal rate for BK-7 with 3 slurries.

184

Initial inspection of the data plotted in Figure B.16 suggests that the

nanodiamonds help increase removal rates, but normalizing the data (Figure B.17)

shows that this is not the case. [The addition of medium friability nanodiamonds

appears to improve the efficiency for one data point (circled), but within a few hours

the normalized removal rate is similar to that for all other slurries.]

Silvia Booij2 found that the material removal scales quadratically with the

abrasive particle radius. She found that the contact area (radius2) was important in

removal, and not the mass of the abrasive (radius3). If we extend the quadratic

dependence to the abrasives used in this experiment, the 3µm CI particles should be

10,000x more efficient than the 30nm nanodiamond particles. Another hypothesis for

why the nanodiamonds do not increase removal is that there is no mechanism for the

nanodiamonds to be efficiently transported to the polishing zone. All of the particles

are being forced out of the nozzle at the same pressure, and if we assume that the

nanodiamonds and CI particles are free particles in the slurry, the more massive CI

particles will dominate the removal process.

Within error, the addition of nanodiamonds does not affect the p-v surface

roughness inside of the footprints (Figure B.18), which are only slightly higher than

that for the initial pitch polished surface. The average areal rms surface roughness

values inside CI slurry footprints at the beginning of the nanodiamond/CI

experiments are higher than those for the CI slurry experiment (Figure B.19). The

rms surface roughness values do not vary much for the medium and high friability

nanodiamond/CI slurries as a function of time. The values continue to increase with

time for the low friability nanodiamond/CI slurry.

185

BK-7

0

5

10

15

20

25

30

35

0 100 200 300 400 500 600


Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)

Low Friability Nanodiamonds Medium Friability NanodiamondsHigh Friability Nanodiamonds Pitch Polished

Nanodiamonds added

Figure B.18 – Average areal p-v surface roughness for BK-7 with 3 nanodiamond/CI slurries.

BK7

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

1.1

1.15

1.2

0 100 200 300 400 500 600


Ave

rage

are

al rm

s su

rfac

e ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.19 – Average areal rms surface roughness for BK-7 with 3 nanodiamond/CI slurries.

186

B.5.2. FS The peak removal rate data for footprints taken on FS with the three

nanodiamond/CI slurries are included in Figure B.20. All three nanodiamond/CI

slurries have similar peak removal rate values after the nanodiamonds were added to

the CI slurries. There was very little variation in peak removal rate for all three

nanodiamond/CI slurries throughout the experiment.

The peak removal rate data normalized to the respective initial CI slurry data

point is plotted in Figure B.21. The medium friability nanodiamond/CI slurry has the

higher normalized removal rate data compared to the low and high friability

nanodiamond/CI slurries. We believe that the medium friability nanodiamonds

continue to play a small role in increasing the material removal throughout the

experiment compared to the other two nanodiamond/CI slurries.

The p-v surface roughness inside all of the footprints taken with all four

slurries is similar to the pitch polished surface (Figure B.22). The rms surface

roughness values (Figure B.23) are higher inside the CI slurry footprints compare to

the pitch polished surface. The same is true for the medium and low friability

nanodiamond/CI slurry footprints. Two out of three high friability nanodiamond/CI

slurry footprints have the rms surface roughness values similar to the pitch polished

surface. FS

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0 100 200 300 400 500 600


Pea

k R

emov

al R

ate

( µm

/min

)


CI Slurry

Nanodiamonds added

Figure B.20 – Peak removal rate for FS with 3 nanodiamond/CI slurries.

187

FS

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500 600


Nor

mal

ized

Pea

k R

emov

al R

ate


CI only

Nanodiamonds added

Figure B.21 – Normalized peak removal rate for FS with 3 nanodiamond/CI slurries.

FS

0

5

10

15

20

25

30

0 100 200 300 400 500 600


Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)


1200

Nanodiamonds added

Figure B.22 – Average areal p-v surface roughness for FS

with 3 nanodiamond/CI slurries.

188

FS

0.7

0.75

0.8

0.85

0.9

0.95

0 100 200 300 400 500 600


Ave

rage

are

al rm

s su

rfac

e ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.23 – Average areal rms surface roughness for FS for 3 nanodiamond/CI slurries.

B.5.3. LHG-8 The peak removal rate data for footprints taken with the three friability

nanodiamond/CI slurries is shown in Figure B.24. The low and high friability

nanodiamond/CI slurries have a higher initial peak removal rates before the

nanodiamonds are added compared to the medium friability nanodiamond/CI slurry,

but the opposite is true after the nanodiamonds are added. When all of the data is

normalized to the first footprint of each experiment (Figure B.25), the medium

friability data has the smallest percent change in removal rate from the CI slurry

compared to the low and high friability nanodiamond/CI slurries.

The average areal p-v surface roughness values (Figure B.26) are higher

inside all of the nanodiamond/CI slurry footprints compared to the initial pitch

polished surface. Within error it is difficult to make comparisons between the

slurries, but as time elapses the high friability nanodiamond/CI slurry appears to

produce the surface with the lowest p-v surface roughness values. The average rms

surface roughness values (Figure B.27) converged to a similar value as time elapsed.

189

The values measured inside the nanodiamond/CI footprints are all larger than the

initial pitch polished surface roughness.

LHG-8

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

0 100 200 300 400 500 600


Pea

k R

emov

al R

ate

( µm

/min

)


CI Slurry

Nanodiamonds added

Figure B.24 – Peak removal rate for LHG-8 with 3 nanodiamond/CI slurries.

LHG-8

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500 600


Nor

mal

ized

Pea

k R

emov

al R

ate


1200

CI Slurry

Nanodiamonds added

Figure B.25 – Normalized peak removal rate for LHG-8 with 3 nanodiamond/CI slurries.

190

LHG-8

0

20

40

60

80

100

120

0 100 200 300 400 500 600


Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.26 – Average areal p-v surface roughness for

LHG-8 with 3 nanodiamond/CI slurry footprints.

LHG-8

0.8

1.3

1.8

2.3

2.8

3.3

3.8

4.3

0 100 200 300 400 500 600


Ave

rage

are

al rm

s su

rfac

e ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.27 – Average area rms surface roughness for LHG-8 with 3 nanodiamond/CI slurry footprints.

191

B.5.4. FCD-1 Similar to BK-7, FS and LHG-8 the addition of nanodiamonds to the CI slurry

do not increase the efficiency of removal compared to the CI slurry as shown in

Figures B.28 and B.29. Unlike the previous three glasses, the low friability

nanodiamond/CI slurry produces the highest peak and normalized removal rates in

comparison to the medium and high friability nanodiamond/CI slurries.

The p-v surface roughness inside the nanodiamond/CI footprints on FCD-1

(Figure B.30) did not change within error as time elapsed, although all of the surfaces

inside the footprints have higher p-v surface roughness values compared to the initial

pitch polished surface. The rms surface roughness values inside the FCD-1 footprints

(Figure B.31) similar to the p-v surface roughness values are higher than the initial

pitch polished surface. The rms surface roughness values do not change much as

time elapsed with the exception of the low friability nanodiamond slurry, the rms

value inside the last footprint taken was lower than the previous footprints.

FCD-1

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0 100 200 300 400 500 600


Pea

k R

emov

al R

ate

( µm

/min

)


CI Slurry

Nanodiamonds added

Figure B.28 – Peak removal rate data for FCD-1 footprints made with 3 nanodiamond/CI slurries.

192

FCD-1

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500 600


Nor

mal

ized

Pea

k R

emov

al R

ate


CI Slurry

Nanodiamonds added

Figure B.29 – Normalized peak removal rate data for FCD-1 footprints made with 3 nanodiamond/CI slurries.

FCD-1

0

10

20

30

40

50

60

70

80

90

100

0 100 200 300 400 500 600


Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.30 – Average areal p-v surface roughness data inside FCD-1 footprints made with 3 nanodiamond/CI slurries.

193

FCD-1

0.7

1.2

1.7

2.2

2.7

3.2

3.7

4.2

0 100 200 300 400 500 600


Ave

rage

are

al rm

s su

rfac

e ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.31 – Average areal rms surface roughness data inside FCD-1 footprints made with 3 nanodiamond/CI slurries.

B.5.5. FD-60 Figures B.32 and B.33 are plots of the peak removal rate and normalized

removal rate for the footprints taken on FD-60 with the three nanodiamond/CI

slurries. The data given in these plots indicate that the nanodiamonds do not play a

roll in improving the efficiency of the removal process of FD-60 compared to the CI

slurry.

The average areal p-v surface roughness (Figure B.34) and average areal rms

surface roughness (Figure B.35) inside the FD-60 footprints are plotted as a function

of time. The addition of the low and medium friability nanodiamonds do not affect

the p-v or rms surface roughness values compared to the surface roughness inside the

CI slurry footprints. Both the p-v and rms surface roughness values decreased inside

the FD-60 footprints after the addition of high friability nanodiamonds. The surfaces

continued to become smoother as the slurry aged. The surface roughness inside the

194

last high friability nanodiamond/CI slurry footprint was almost as smooth as the

initial pitch polished surface.

FD-60

0.00

0.05

0.10

0.15

0.20

0.25

0 100 200 300 400 500 600


Pea

k R

emov

al R

ate

( µm

/min

)


CI Slurry

Nanodiamonds added

Figure B.32 – Peak removal rate data for FD-60 footprints made with 3 nanodiamond/CI slurries.

FD-60

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500 600


Nor

mal

ized

Pea

k R

emov

al R

ate


CI Slurry

Nanodiamonds added

Figure B.33 – Normalized peak removal rate data for FD-60 footprints made with 3 nanodiamond/CI slurries.

195

FD-60

0

10

20

30

40

50

60

70

80

0 100 200 300 400 500 600


Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.34 – Average areal p-v surface roughness data inside FD-60 footprints made with 3 nanodiamond/CI slurries.

FD-60

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

0 100 200 300 400 500 600


Ave

rage

are

al rm

s su

rfac

e ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.35 – Average areal rms surface roughness data inside FD-60 footprints made with 3 nanodiamond/CI slurries.

196

B.5.6. EFDS-1 All three of the slurries have higher peak removal rates before the

nanodiamonds were added compared to the CI slurry footprints shown in Figure B.36.

The normalized peak removal rates (Figure B.37) indicate that the medium friability

nanodiamond/CI slurry is more efficient compared to the low and high friability

nanodiamond/CI slurries.

We observed in section B.3 that the EFDS-1 surface had pits inside footprints

made with CI slurry. We also observe pits after the low and medium friability

nanodiamonds were added to the CI slurry. We did not see any pits inside the high

friability nanodiamond/CI slurry. Pitting on the surface of a part causes higher p-v

and rms surface roughness values.

The average areal p-v and rms surface roughness values are plotted in Figures

B.38 and B.39. The CI slurry footprint taken before the high friability nanodiamonds

were added only had very shallow pits. We believe that the pits are sub-surface

damage (SSD) uncovered in the FJP polishing process. All of the experiments except

the high friability experiment are performed on the same piece of EFDS-1. The high

friability experiment is performed on a fresh piece of EFDS-1. The two pieces are

from the same melt, and were both polished by the same optician and had the same p-

v and rms surface roughness values, but it is still possible that the first piece of

EFDS-1 had SSD. It is unlikely that the high friability nanodiamonds inhibit pitting

because the high friability nanodiamonds had very low, if any involvement in the

material removal process which is seen in the removal rate data.

197

EFDS-1

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0 100 200 300 400 500 600


Pea

k R

emov

al R

ate

( µm

/min

)


CI Slurry

Nanodiamonds added

Figure B.36 – Peak removal rate data for EFDS-1 footprints made with 3 nanodiamond/CI slurries.

EFDS-1

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500 600


Nor

mal

ized

Pea

k R

emov

al R

ate


CI Slurry

Nanodiamonds added

Figure B.37 – Normalized peak removal rate data for EFDS-1 footprints made with 3 nanodiamond/CI slurries.

198

EFDS-1

0

20

40

60

80

100

120

140

160

180

200

0 100 200 300 400 500 600


Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.38 – Average areal p-v surface roughness data

inside EFDS-1 footprints made with 3 nanodiamond/CI slurries.

EFDS-1

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

0 100 200 300 400 500 600


Ave

rage

are

al rm

s su

rfac

e ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.39 – Average areal rms surface roughness data inside EFDS-1 footprints made with 3 nanodiamond/CI slurries.

199

B.5.7. NSF-6 The peak removal rate results for NSF-6 with the nanodiamond/CI slurries

were unlike the other glasses in our set in two ways. The peak removal rate (Figure

B.40) did not decrease for the first footprint taken after the medium friability

nanodiamonds were added to the CI slurry. Subsequent medium friability

nanodiamond/CI slurry footprints were all lower than the CI slurry. The peak

removal rate for the last spot taken with the low friability nanodiamond/CI slurry was

equal to the CI slurry peak removal rate. These two observations are accentuated in

the normalized peak removal rate data shown in Figure B.41.

The additions of nanodiamonds have very little effect on the surface texture of

NSF-6. All of the surfaces inside the footprints had higher average areal p-v (Figure

B.42) and rms (Figure B.43) surface roughness values compared to the initial pitch

polished surface, but within error the values inside the spots are all the same.

SF-6

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0 100 200 300 400 500 600


Pea

k R

emov

al R

ate

( µm

/min

)


CI Slurry

Nanodiamonds added

Figure B.40 – Peak removal rate data for NSF-6 footprints made with 3 nanodiamond/CI slurries.

200

SF-6

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500 600


Nor

mal

ized

Pea

k R

emov

al R

ate


1200

CI Slurry

Nanodiamonds added

Figure B.41 – Normalized peak removal rate data for NSF-6 footprints made with 3 nanodiamond/CI slurries.

SF-6

0

10

20

30

40

50

60

70

0 100 200 300 400 500 600


Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.42 – Average areal p-v surface roughness data inside NSF-6 footprints made with 3 nanodiamond/CI slurries.

201

SF-6

0.7

0.9

1.1

1.3

1.5

1.7

1.9

2.1

2.3

2.5

0 100 200 300 400 500 600


Ave

rage

are

al rm

s su

rfac

e ro

ughn

ess

(nm

)


Nanodiamonds added

Figure B.43 – Average areal rms surface roughness data inside NSF-6 footprints made with 3 nanodiamond/CI slurries.

B.5.8. Experimental data for three nanodiamond slurries



FS 9.26 38 0.05 1.00 21.14 4.18 0.89 0.019.28 95 0.03 0.52 18.82 4.90 0.91 0.029.18 227 0.02 0.47 20.00 2.42 0.89 0.029.09 412 0.02 0.41 17.14 3.04 0.89 0.02

BK-7 9.26 48 0.10 1.00 20.80 4.55 0.97 0.039.28 100 0.07 0.72 23.92 4.89 1.01 0.019.18 232 0.06 0.60 22.60 5.22 1.02 0.049.09 417 0.07 0.69 18.72 5.16 1.12 0.01

FD-60 9.26 25 0.23 1.00 38.28 10.99 1.35 0.069.28 83 0.09 0.40 53.00 14.99 1.29 0.089.18 216 0.09 0.39 27.48 7.07 1.17 0.069.09 403 0.09 0.38 31.40 7.31 1.28 0.03

SF-6 9.26 66 0.34 1.00 39.56 8.96 1.97 0.049.28 115 0.30 0.91 27.34 7.53 1.92 0.119.18 252 0.27 0.80 35.18 9.81 2.00 0.069.09 434 0.33 0.98 48.66 10.43 2.08 0.21

EFDS-1 9.26 31 0.37 1.00 67.02 20.77 2.52 0.229.28 90 0.19 0.52 148.40 13.61 3.10 0.649.18 221 0.18 0.48 118.64 52.99 3.63 1.069.09 408 0.16 0.45 103.46 20.87 2.79 0.34

LHG-8 9.26 20 1.42 1.00 51.48 8.14 3.14 0.169.28 75 0.61 0.43 74.70 25.46 3.61 0.309.18 212 0.49 0.34 83.26 20.05 3.10 0.349.09 398 0.38 0.27 55.88 4.53 2.58 0.25

FCD-1 9.26 52 0.53 1.00 52.94 7.86 3.18 0.059.28 106 0.39 0.74 64.10 13.57 2.88 0.119.18 237 0.39 0.75 54.44 9.81 2.90 0.079.09 422 0.41 0.78 51.88 10.83 1.93 0.11

NSF-6

Table B.4 – Peak removal rate and surface roughness data for the UK-Low friability nanodiamond experiment. The first row of data for each glass is the data for the footprint made with the slurry containing only CI particles; the nanodiamonds were added after the footprints were taken.

202

Glass Slurry pH Elapsed time Peak removal rate

Normalized peak removal rate Average areal p-v p-v st. dev. Average areal rms rms st. dev.

(min) (µm/min) (nm) (nm)FS 9.24 46 0.04 1.00 15.16 2.53 0.90 0.01

9.24 113 0.03 0.65 19.32 5.40 0.92 0.019.12 316 0.02 0.63 17.22 3.11 0.92 0.019.16 505 0.02 0.60 20.02 8.33 0.91 0.01

BK-7 9.24 50 0.09 1.00 16.58 5.00 0.97 0.029.14 265 0.07 0.81 25.92 5.52 0.99 0.029.12 320 0.06 0.63 19.70 3.99 0.96 0.019.16 508 0.06 0.64 20.70 5.90 0.95 0.02

FD-60 9.24 37 0.19 1.00 46.78 10.41 1.25 0.159.24 106 0.11 0.58 38.42 5.58 1.27 0.079.12 307 0.09 0.46 40.06 7.85 1.17 0.049.14 492 0.09 0.46 43.94 10.71 1.11 0.07

SF-6 9.24 62 0.42 1.00 45.48 6.37 1.88 0.209.14 276 0.41 0.98 41.00 13.50 1.86 0.219.14 337 0.29 0.71 39.32 7.42 1.95 0.179.16 525 0.28 0.68 38.22 7.58 1.83 0.21

EFDS-1 9.24 41 0.35 1.00 112.44 40.50 3.02 0.619.24 109 0.19 0.55 97.86 30.90 2.74 0.369.12 307 0.18 0.53 74.02 50.93 1.85 0.579.16 497 0.19 0.53 86.82 32.33 3.05 0.65

LHG-8 9.24 33 1.11 1.00 58.20 13.18 2.63 0.339.24 102 0.64 0.58 74.52 23.68 2.50 0.249.12 292 0.48 0.43 71.92 13.01 2.76 0.329.14 486 0.47 0.43 63.80 5.37 2.79 0.25

FCD-1 9.24 55 0.55 1.00 54.54 10.77 3.46 0.169.14 270 0.42 0.75 64.96 3.75 3.28 0.059.14 324 0.35 0.64 55.40 6.74 3.20 0.039.16 515 0.41 0.73 63.18 23.86 3.21 0.36

NSF-6

Table B.5 – Peak removal rate and surface roughness data for the UK-Medium friability nanodiamond experiment. The first row of data for each glass is the data for the footprint made with the slurry containing only CI particles; the nanodiamonds were added after the footprints were taken.



FS 9.27 27 0.05 1.00 19.24 4.95 0.87 0.019.23 108 0.03 0.56 19.12 2.30 0.81 0.009.18 257 0.02 0.48 14.70 2.99 0.85 0.019.08 440 0.02 0.42 21.30 6.08 0.82 0.01

BK-7 9.27 36 0.10 1.00 25.34 6.49 0.99 0.029.23 117 0.07 0.65 18.42 4.65 0.99 0.029.18 265 0.05 0.45 18.92 4.49 0.91 0.029.08 447 0.05 0.50 24.08 8.51 0.98 0.02

FD-60 9.27 18 0.23 1.00 38.40 10.29 1.22 0.079.23 100 0.09 0.37 18.22 1.05 1.03 0.029.18 249 0.06 0.26 21.36 8.30 0.93 0.079.08 430 0.05 0.22 20.02 6.77 0.90 0.03

SF-6 9.27 42 0.40 1.00 41.58 15.76 1.97 0.029.23 123 0.29 0.72 30.30 5.20 1.83 0.069.18 251 0.27 0.68 30.64 8.19 1.76 0.129.08 457 0.22 0.54 31.48 4.57 1.95 0.08

EFDS-1 9.27 23 0.41 1.00 77.58 12.47 2.53 0.329.23 104 0.18 0.45 51.04 13.03 2.12 0.039.18 252 0.14 0.33 40.96 13.98 1.91 0.199.08 436 0.12 0.30 33.20 7.49 1.90 0.05

LHG-8 9.27 15 1.39 1.00 55.40 10.65 3.49 0.139.23 96 0.44 0.32 52.78 14.14 3.30 0.189.18 245 0.34 0.24 44.96 11.76 2.50 0.219.08 427 0.30 0.21 38.74 4.93 2.79 0.46

FCD-1 9.27 33 0.65 1.00 46.48 5.46 3.16 0.059.23 112 0.36 0.55 42.22 6.53 2.55 0.159.18 261 0.31 0.47 38.02 6.66 2.88 0.029.08 444 0.27 0.42 46.96 19.10 2.70 0.15

NSF-6

Table B.6 – Peak removal rate and surface roughness data for the UK-High friability nanodiamond experiment. The first row of data for each glass is the data for the footprint made with the slurry containing only CI particles; the nanodiamonds were added after the footprints were taken.

203

B.6. Comparison of polishing abrasives The peak removal rate data for the five experiments discussed in sections B.3

– B.5 are plotted in Figure B.44; each glass has its own plot. All seven plots are

plotted on the same semi-log scale. In this section we compare the peak removal rate

range for all of the slurries.

We discussed in Chapter 3 that an MR fluid containing only spherical CI

particles removes material off glass substrates in the MRF platform. We learned from

our first FJP experiment with CI slurry, that the CI particles are also capable of

removing material from glass surfaces in the FJP platform, which does not introduce

a magnetic field.

The FJP platform provided us with a unique way of comparing non-magnetic

similarly sized abrasive particles to CI particles. We chose to use SiC particles in

order to determine how the shape of the abrasive particle plays a role in the FJP

removal process. The data presented indicates that the irregularly shaped SiC

abrasive remove material much more efficiently compared to all of the other slurries.

We determined that the more massive, irregularly shaped SiC particles produce much

higher removal rates and higher surface roughness compared to the spherical CI

particles. The MRF platform is not capable of performing this experiment because

the SiC particles are non-magnetic.

The peak removal rate ranges for the footprints made with the

nanodiamond/CI slurries are also included in Figure B.44. According to our

experimental data the nanodiamonds appear to play a role increasing material

removal efficiency for NSF-6. The nanodiamonds do not appear to increase material

removal efficiency for the remaining six glasses that we use in our MRF experiments.

Combining what we learned from our MRF and FJP experimental results we

hypothesize that the increase in removal efficiency with the addition of

nanodiamonds to the abrasive free MR fluid in the MRF platform is due to the

magnetic field pulling the CI particles down towards the rotating wheel leaving the

204

nanodiamond particles at the top of the MR fluid ribbon where they play an active

role in the MRF material removal process. At the same time we hypothesize that the

nanodiamonds played very little role in the FJP removal process when added to the CI

slurry because there was no force transporting them to the polishing zone.

205

BK-7 FS

0.0

0.2

0.4

0.6

0.8

1.0

1.2

CI Alone SiC UK-Low UK-Medium UK-High

m

in)

µm/

al R

ate

(

Rem

ov

Peak

Figure B.44 – Peak removal rate data ranges for the FJP footprints taken on the set of seven glasses with the five different slurries. The nanodiamond/CI slurries are indicated by the nanodiamond only.

0.0

0.2

0.4

0.6

0.8

1.0

1.2


Peak

Rem

oval

Rat

e ( µ

m/m

in)

3.82

FD-60 NSF-6

EFDS-1 LHG-8

0.0

0.2

0.4

0.6

0.8

1.0

1.2


Pea

k R

emov

al R

ate

( µm

/min

)

0.0

0.2

0.4

0.6

0.8

1.0

1.2


Peak

Rem

oval

Rat

e ( µ

m/m

in)

2.24

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Alon iC -Low UK-Medium UK-High

Peak

Rem

oval

Rat

e ( µ

m/m

in)

CI e S UK

FCD-1

2.57

0.0

0.2

0.4

0.6

0.8

1.0

1.2


Peak

Rem

oval

Rat

e ( µ

m/m

in)

4.86

0.0

0.2

0.4

0.6

0.8

1.0

1.2


Peak

Rem

oval

Rat

e ( µ

m/m

in)

5.33

206

B.7. Summary

The majority of the FJP experimental slurries we used implemented carbonyl

iron (CI) as the main abrasive. It was discovered that the pH of the slurry and the

peak removal rates dropped significantly during the course of a three-day experiment.

The polishing efficiency of the CI slurry was compared to slurry containing

SiC abrasive particles whose mean particle size is slightly larger than the CI particles.

The SiC abrasive is irregularly shaped where as the CI particles are spherical.

Polishing efficiency was higher for the SiC slurry compared to the CI slurry. We

found very high correlation between the peak removal rate values and the glass

mechanical properties for the SiC slurry footprints.

Three nanodiamond suspensions were added to the CI slurry, and the peak

removal rate data suggests that they were not involved in the removal process.

Previous work by Booij suggests that the CI particles are 10,000x more efficient

compared to the nanodiamond particles based on their size and therefore would play a

very small roll in the removal process.

In comparison of our results using the same polishing abrasives and glass set

using two polishing platforms with very different mechanisms of removal we realize

to a larger extent that it is a combination of the magnetic field and the addition of

nanodiamonds to the MR fluid containing only CI particles that provides the increase

in removal efficiency. We learned that a polishing process without a driving force

transporting the nanodiamonds to the polishing zone, the larger particles dominate the

removal process.

207

References 1. O. W. Fahnle, H. van Brug, and H. J. Frankena, "Fluid jet polishing of optical

surfaces," Applied Optics 37(28), 6771-6773 (1998).

2. S. M. Booij, "Fluid Jet Polishing: Possibilities and limitations of a new fabrication technique," (Delft University of Technology, 2003).

3. FISBA OPTIK AG, Rorschacher Strasse 268, CH-9016 St.Gallen

4. Zeeko Limited, 4 Vulcan Way, Vulcan Court, Hermitage Industrial Estate Coalville, Leicestershire, LE67 3FW, UK.

5. St. Gallen, Switzerland., 2005.




9. J. C. Lambropoulos, S. Xu, and T. Fang, "Loose abrasive lapping hardness of optical glasses and its interpretation," Applied Optics 36(7), 1501 - 1516 (1997).

10. A. Wheeler and A. Ganji, Introduction to Engineering Experimentation (Prentice-Hall Inc., Upper Saddle River, NJ, 1996).


208

Appendix C

The data provided in Appendix C is intended to supplement data presented in

Chapters 5 and 6 of this thesis. Individual topics are divided into subsections.

C.1 – Comparison of near surface and bulk mechanical FOM values

In this section we show additional experimental data relating to the four

different bulk and near surface mechanical figure of merit values. These data are

supplemental to the data discussed in Section 5.2.4 in the main thesis.

The bulk and near surface mechanical figure of merit values are plotted with

our experimental results for five MR fluids (Figures C.1 – C.4). The five fluids are

four 0.01-vol% nanodiamond fluids and one MR fluid containing only CI particles,

which we will refer to as abrasive-free fluid. The relationship plotted in Figure C.1 is

FOM-B calculated from the mechanical properties measured for the bulk substrate

versus peak removal rate. The confidence levels for the linear relationships between

the peak removal rates versus the FOM-B are above 90%.

The FOM-S values (Figure C.2) have improved linear fits for the UK Medium

A, NDP and UK low nanodiamond MR fluids compared to FOM-B. The UK High

fluid R2 value for the FOM-S values is slightly lower compared to the FOM-B values.

The confidence level for the abrasive free MR fluid linear fit with FOM-S was the

lowest for all of the FOM values.

The worst fits are associated with the FOM-D values (Figure C.3) for all four

nanodiamond MR fluids. The FOM-D does however appear to describe the removal

for the abrasive-free fluid better than all of the FOM values. We expect the

confidence levels between FOM-D and MRF peak removal rates to be lower because

the dry surface layer does not resemble the glass as MRF is removing material. We

find it surprising that the confidence level for the abrasive-free MR fluid linear fit is

as high as 98%. The high correlation implies that a mechanical mechanism

dominates glass material removal with abrasive-free MR fluid.

209

Figure C.4 has been included for comparison. The linear fits all have

confidence levels better than 80% for the FOM measurements made on the near

surfaces of the glasses immersed in DI water.

Abrasive FreeR2 = 0.62CL: 90%

UK-LowR2 = 0.73CL: 95%

UK-MediumR2 = 0.57

Confidence Level (CL): 90%NDP

R2 = 0.70CL: 95%

0

5

10

15

20

25

30

0 2 4 6 8 10 12 14FOM - B

Peak

Rem

oval

Rat

e ( µ

m/m

in)

UK-HighR2 = 0.88CL: 99%


UK-LowR2 = 0.81CL: 98%

UK-MediumR2 = 0.64CL: 90%

NDPR2 = 0.74CL: 95%

0

5

10

15

20

25

30

0 1 2 3 4 5 6 7 8FOM - S

Pea

k R

emov

al R

ate

( µm

/min

)

UK-HighR2 = 0.85CL: 99%

F

.

0

5

10

15

20

25

30

0

Re

l Rm

/

F

min

)at

e ( µ

mov

aP

eak

igure C.1 – Peak removal rateversus the FOM-B calculatedfor the bulk material. Data forall nanodiamond fluids contain0.01-vol% nanodiamonds.


UK-LowR2 = 0.60CL: 90%


NDPR2 = 0.47CL: 80%

1 2 3 4 5 6 7FOM - D

UK-HighR2 = 0.70CL: 95%

0

5

10

15

20

25

30

Peak

Rem

oval

Rat

e ( µ

m/m

in)

Figure C.2 – Peak removal rateversus the FOM-S calculated 60nminto a sample in MR supernatantData for all nanodiamond fluidscontain 0.01-vol% nanodiamonds.

Abrasive freeR2 = 0.60CL: 90%

UK-LowR2 = 0.68CL: 95%


NDPR2 = 0.64CL: 90%

0 1 2 3 4 5 6FOM - W

UK-HighR2 = 0.81CL: 98%

igure C.3 – Peak removal rate versus the FOM-D calculated 60nm into a dry sample. Data for all nanodiamond fluids contain 0.01-vol% nanodiamonds.

Figure C.4 – Peak removal rateversus the FOM-W calculated60nm into a sample in DI water.Data for all nanodiamond fluidscontain 0.01-vol% nanodiamonds.

210

C.2 – p-v surface roughness and drag force In section 6.3.2 in Chapter 6 of this thesis we found an inverse relationship

between p-v surface roughnesses and drag force for various concentrations of NDP

nanodiamond MR fluid. Figures C.5 – C.7 plot the average areal p-v surface

roughness versus inverse drag force for various glass types taken with 0.001, 0.005

and 0.01-vol% UK-Low, UK-Medium A and UK-High nanodiamond MR fluids. In

general we find that the data for these MR fluids agrees with the relationship we

discussed earlier in this thesis. The agreement is very low for higher concentrations

of UK-Medium A nanodiamond MR fluid. The UK-Medium A p-v surface

roughness values are consistently low and there is little variation between glass types,

which results in low correlations.

Figure C.5 – Average areal p-v surface roughness versus inverse drag force with varying glass type for UK-Low nanodiamond MR fluids. [Data tabulated in tables C.1, C.4 and C.7.]

0.001-vol%R2 = 0.74


0.01-vol%R2 = 0.09CL: <70%

0.005-vol%R2 = 0.003CL: <70%

0

5

10

15

20

25

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1/F

0.001-vol%R2 = 0.97

Confidence Level (CL): >99%

0.005-vol%R2 = 0.63CL: 90%

0.01-vol%R2 = 0.004CL: <70%

0

10

20

30

40

50

60

70

80

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

1/Fd (N-1)

Figure C.6 – Average areal p-v surface roughness versus inverse drag force with varying glass type for UK-Medium A nanodiamond MR fluids. [Data tabulated in tables C.2, C.5 and C.8.]

Figure C.7 – Average areal p-v surface roughness versus inverse drag force with varying glass type for UK-High nanodiamond MR fluids. [Data tabulated in tables C.3, C.6 and C.9.]

d (N-1)

Aver

age

area

l p-v

sur

face

roug

hnes

s (n

m)

0.001-vol%R2 = 0.84


0.01-vol%R2 = 0.63CL: 90%

0.005-vol%R2 = 0.21CL: ~75%

0

10

20

30

40

50

60

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1/Fd (N-1)

Ave

rage

are

al p

-v s

urfa

ce ro

ughn

ess

(nm

)

s (n

m)

e ro

ughn

esag

e ar

eal p

-v s

urfa

cA

ver

211

C.3 – Experimental peak removal rate and surface roughness data tables The following Tables C.1 – C.12 contain experimental data used in this thesis. The glass type is indicated above each table and the fluid name is indicated in the table description.

LHG-8

Spot #Time (min)

Viscosity (cP)

Moisture Content (wt. %) Cnd Fluid pH

Pump Speed (rpm)

Drag Force (N) st dev

Spot Time (sec) ddp (um)

Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.65 13.20 0 10.27 118 1.03 0.05 2.16 0.16 4.50 68.84 5.03 8.07 0.272 10 43.73 13.11 0.001 10.13 123 1.12 0.04 2.16 0.22 6.06 60.30 6.95 5.78 0.993 156 44.87 12.97 0.001 10.03 127 1.08 0.04 2.16 0.20 5.42 69.10 18.98 6.70 0.654 261 44.99 12.87 0.001 9.95 127 0.99 0.04 2.16 0.19 5.33 85.12 33.87 6.18 0.635 1490 44.86 12.79 0.001 9.72 128 1.09 0.07 2.16 0.16 4.31 82.72 23.72 6.99 0.306 1696 44.80 12.77 0.001 9.70 126 1.11 0.04 2.16 0.16 4.47 73.16 18.01 7.09 0.417 2799 45.02 12.71 0.001 9.58 123 1.09 0.04 2.16 0.15 4.17 107.32 47.88 6.81 0.42

FS

Spot #Time (min)

Viscosity (cP)

Moisture Content Cnd Fluid pH

Pump Speed (rpm)

Drag Force (N) st. dev.


Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.60 13.20 0 10.27 118 1.84 0.07 4.16 0.12 1.77 13.62 4.08 0.86 0.022 19 43.74 13.11 0.001 10.13 123 1.90 0.07 4.16 0.13 1.85 18.86 7.78 0.84 0.053 159 45.11 12.97 0.001 10.03 127 2.03 0.06 4.16 0.12 1.76 13.54 6.67 0.88 0.054 264 45.03 12.87 0.001 9.95 127 1.81 0.07 4.16 0.11 1.59 11.69 2.66 0.88 0.025 1493 44.93 12.79 0.001 9.72 128 1.81 0.08 4.16 0.08 1.20 25.40 7.26 0.94 0.046 1700 44.85 12.77 0.001 9.70 126 1.80 0.06 5.22 0.07 0.78 21.76 7.80 1.01 0.147 2803 44.88 12.71 0.001 9.58 123 1.67 0.06 7.00 0.06 0.50 67.34 22.94 1.92 0.72

BK-7

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 45.06 13.20 0 10.27 118 2.14 0.07 3.16 0.16 2.96 8.88 1.05 0.77 0.012 22 43.93 13.11 0.001 10.13 123 2.29 0.07 3.16 0.20 3.87 13.17 3.48 0.82 0.023 163 44.99 12.97 0.001 10.03 127 2.34 0.06 3.16 0.20 3.84 9.46 1.60 0.87 0.024 255 44.89 12.87 0.001 9.95 127 2.19 0.06 3.16 0.19 3.66 11.22 1.82 0.85 0.025 1495 44.95 12.79 0.001 9.72 128 2.20 0.08 3.16 0.16 2.98 11.23 3.20 0.88 0.066 1702 44.83 12.77 0.001 9.70 126 2.22 0.07 3.16 0.12 2.20 10.45 1.98 0.86 0.027 2806 45.05 12.71 0.001 9.58 123 2.06 0.07 4.16 0.13 1.89 10.80 1.69 0.93 0.02

FD-60

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.48 13.20 0 10.27 118 2.39 0.07 2.16 0.16 4.44 11.93 2.26 0.73 0.022 30 43.98 13.11 0.001 10.13 123 2.58 0.08 2.16 0.18 5.06 12.52 5.40 0.74 0.033 168 45.08 12.97 0.001 10.03 127 2.77 0.07 2.16 0.18 4.94 11.55 4.45 0.75 0.044 273 45.04 12.87 0.001 9.95 127 2.54 0.07 2.16 0.18 4.86 8.32 2.40 0.72 0.015 1501 45.02 12.79 0.001 9.72 128 2.71 0.10 2.16 0.16 4.56 8.83 1.13 0.76 0.026 1709 45.00 12.77 0.001 9.70 126 2.73 0.08 2.16 0.16 4.36 28.30 19.71 0.83 0.167 2812 45.01 12.71 0.001 9.58 123 2.64 0.10 2.16 0.15 4.22 17.41 17.62 0.77 0.10

EFDS-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.65 13.20 0 10.27 118 2.24 0.08 2.16 0.19 5.19 13.83 6.65 0.70 0.032 33 43.95 13.11 0.001 10.13 123 2.40 0.07 2.16 0.22 6.14 6.79 0.58 0.67 0.013 170 44.99 12.97 0.001 10.03 127 2.54 0.06 2.16 0.21 5.94 7.98 1.39 0.68 0.024 275 44.91 12.87 0.001 9.95 127 2.49 0.07 2.16 0.21 5.83 14.70 2.42 1.31 0.285 1503 44.88 12.79 0.001 9.72 128 2.40 0.09 2.16 0.20 5.44 8.90 1.40 0.72 0.056 1710 45.07 12.77 0.001 9.70 126 2.52 0.07 2.16 0.20 5.50 8.00 1.91 0.68 0.037 2815 44.84 12.71 0.001 9.58 123 2.59 0.08 2.16 0.20 5.42 7.00 0.60 0.70 0.03

FCD-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.71 13.20 0 10.27 118.00 2.05 0.07 1.17 0.14 7.28 11.42 1.94 0.89 0.042 28 43.88 13.11 0.001 10.13 123.00 2.20 0.06 1.17 0.16 8.21 13.48 1.75 0.87 0.023 165 45.01 12.97 0.001 10.03 127.00 2.34 0.06 1.17 0.15 7.79 9.41 0.85 0.90 0.024 270 45.02 12.87 0.001 9.95 127.00 2.15 0.07 1.17 0.17 8.46 9.29 0.31 0.89 0.025 1498 44.92 12.79 0.001 9.72 128.00 2.31 0.08 1.17 0.15 7.69 9.34 0.94 0.92 0.026 1705 45.00 12.77 0.001 9.70 126.00 2.41 0.07 1.17 0.16 8.05 12.17 2.08 0.92 0.017 2809 44.86 12.71 0.001 9.58 123.00 2.28 0.07 1.17 0.15 7.79 10.67 1.95 0.87 0.02

Table C.1 - 0.001-vol% UK-Low friability MR fluid data.

212

LHG-8

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.51 13.15 0 10.25 113 1.04 0.04 2.16 0.16 4.36 98.38 16.45 7.89 0.252 15 43.96 13.13 0.001 10.24 118 1.22 0.05 2.16 0.39 10.78 17.40 2.28 1.16 0.083 168 44.92 12.90 0.001 10.07 120 1.24 0.05 1.17 0.18 9.44 26.94 5.26 1.21 0.284 323 44.94 12.89 0.001 10.02 120 1.22 0.06 1.17 0.15 7.59 30.56 4.73 1.23 0.155 1359 44.59 12.90 0.001 9.92 120 1.16 0.04 1.17 0.12 5.95 38.22 7.92 1.71 0.276 1685 15.15 12.73 0.001 9.83 118 1.06 0.06 2.16 0.20 5.47 60.82 4.85 5.25 0.847 2800 45.17 12.74 0.001 9.70 122 1.09 0.06 2.16 0.18 4.92 74.00 13.60 5.67 0.94

FS

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.75 13.15 0 10.25 113 2.04 0.06 4.16 0.12 1.73 11.47 3.65 0.84 0.032 18 44.05 13.13 0.001 10.24 118 2.21 0.07 4.16 0.15 2.19 12.98 3.84 0.84 0.013 165 45.09 12.90 0.001 10.07 120 2.21 0.08 4.16 0.15 2.11 13.22 3.28 0.82 0.024 326 45.02 12.89 0.001 10.02 120 2.18 0.08 4.16 0.14 1.96 11.44 3.49 0.82 0.025 1362 44.60 12.90 0.001 9.92 120 2.05 0.07 4.16 0.13 1.90 9.44 1.74 0.82 0.016 1689 45.19 12.73 0.001 9.83 118 1.82 0.09 4.16 0.09 1.31 12.50 3.60 0.88 0.037 2803 45.08 12.74 0.001 9.70 122 1.79 0.08 5.22 0.09 1.07 10.14 1.31 0.87 0.02

BK-7

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.86 13.15 0 10.25 113 2.25 0.06 3.16 0.17 3.17 9.13 1.03 0.82 0.002 21 44.19 13.13 0.001 10.24 118 2.51 0.07 3.16 0.23 4.37 12.07 3.20 0.84 0.033 167 45.12 12.90 0.001 10.07 120 2.45 0.08 3.16 0.21 3.99 8.48 0.51 0.83 0.034 328 44.91 12.89 0.001 10.02 120 2.49 0.08 3.16 0.22 4.08 9.73 1.27 0.84 0.035 1365 44.72 12.90 0.001 9.92 120 2.36 0.07 3.16 0.19 3.65 10.13 1.53 0.82 0.026 1687 45.09 12.73 0.001 9.83 118 2.26 0.10 3.16 0.18 3.49 9.54 0.87 0.90 0.047 2806 45.39 12.74 0.001 9.70 122 2.21 0.10 3.16 0.16 3.02 11.92 2.55 0.89 0.03

FD-60

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.72 13.15 0 10.25 113 2.45 0.07 2.16 0.16 4.36 10.70 5.89 0.74 0.032 27 44.51 13.13 0.001 10.24 118 2.77 0.07 2.16 0.20 5.47 7.60 0.38 0.76 0.033 173 45.29 12.90 0.001 10.07 120 2.81 0.09 2.16 0.19 5.28 10.01 2.03 0.81 0.034 333 45.04 12.89 0.001 10.02 120 2.73 0.09 2.16 0.18 4.97 15.11 13.90 0.78 0.045 1370 44.81 12.90 0.001 9.92 120 2.70 0.07 2.16 0.15 4.03 13.39 4.06 0.76 0.026 1693 44.87 12.73 0.001 9.83 118 2.64 0.10 2.16 0.16 4.56 17.27 10.74 0.81 0.057 2812 44.91 12.74 0.001 9.70 122 2.73 0.10 2.16 0.15 4.22 8.38 2.46 0.71 0.01

EFDS-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.89 13.15 0 10.25 113 2.27 0.07 2.16 0.19 5.19 7.53 1.11 0.68 0.022 29 44.43 13.13 0.001 10.24 118 2.48 0.08 2.16 0.24 6.56 7.32 0.27 0.71 0.013 176 45.23 12.90 0.001 10.07 120 2.35 0.07 2.16 0.22 6.22 6.98 0.39 0.71 0.024 336 44.88 12.89 0.001 10.02 120 2.39 0.08 2.16 0.23 6.28 7.92 1.50 0.73 0.025 1372 44.59 12.90 0.001 9.92 120 2.33 0.07 2.16 0.21 5.78 7.14 0.32 0.76 0.026 1696 44.88 12.73 0.001 9.83 118 2.31 0.09 2.16 0.21 5.69 7.47 0.78 0.74 0.037 2814 45.01 12.74 0.001 9.70 122 2.42 0.11 2.16 0.20 5.58 7.17 0.30 0.73 0.01

FCD-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.56 13.15 0 10.25 113.00 2.04 0.06 1.17 0.14 7.38 11.96 2.45 0.85 0.022 24 44.18 13.13 0.001 10.24 118.00 2.27 0.07 1.17 0.18 9.03 10.29 1.62 0.92 0.023 170 45.05 12.90 0.001 10.07 120.00 2.39 0.08 1.17 0.17 8.87 14.08 1.03 1.00 0.034 331 45.06 12.89 0.001 10.02 120.00 2.38 0.08 1.17 0.18 9.03 11.21 2.33 0.91 0.035 1368 44.77 12.90 0.001 9.92 120.00 2.33 0.06 1.17 0.17 8.67 9.75 1.07 0.91 0.046 1690 44.92 12.73 0.001 9.83 118.00 2.20 0.08 1.17 0.16 8.00 12.72 2.69 0.96 0.017 2809 45.18 12.74 0.001 9.70 122.00 2.18 0.08 1.17 0.15 7.69 13.52 3.04 0.95 0.02

Table C.2- 0.001-vol% UK-Medium A friability MR fluid data.

213

LHG-8

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.72 13.18 0 10.22 110 1.22 0.06 2.16 0.16 4.39 81.14 13.16 8.07 0.422 17 43.90 13.04 0.001 10.18 114 1.27 0.06 2.16 0.23 6.33 45.28 5.18 2.95 1.253 165 45.06 12.89 0.001 10.06 117 1.34 0.06 2.16 0.22 6.08 54.84 13.19 3.65 1.104 259 44.91 12.88 0.001 9.98 117 1.27 0.06 2.16 0.22 6.08 77.98 20.26 6.57 0.725 1356 44.88 12.86 0.001 9.78 121 1.36 0.06 2.16 0.18 5.03 62.08 3.22 6.22 0.736 1655 45.11 12.72 0.001 9.78 121 1.28 0.07 2.16 0.17 4.83 62.38 6.25 5.67 0.577 2792 45.00 12.70 0.001 9.61 124 1.25 0.12 2.16 0.16 4.53 52.96 2.93 4.85 0.48

FS

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.57 13.18 0 10.22 110 1.80 0.06 4.16 0.12 1.76 14.98 2.21 0.84 0.032 21 44.30 13.04 0.001 10.18 114 1.92 0.08 4.16 0.14 1.96 8.73 0.93 0.81 0.013 167 45.01 12.89 0.001 10.06 117 1.94 0.09 4.16 0.13 1.86 10.36 2.44 0.82 0.024 264 45.05 12.88 0.001 9.98 117 1.88 0.09 4.16 0.13 1.80 11.56 3.20 0.82 0.015 1359 44.83 12.86 0.001 9.78 121 1.92 0.08 4.16 0.12 1.69 16.80 6.38 0.91 0.026 1658 45.10 12.72 0.001 9.78 121 1.78 0.11 5.22 0.13 1.46 16.18 3.52 0.85 0.057 2795 44.94 12.70 0.001 9.61 124 1.79 0.08 5.22 0.11 1.28 14.58 2.62 0.86 0.02

BK-7

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.90 13.18 0 10.22 110 2.05 0.06 3.16 0.16 3.09 10.85 2.26 0.81 0.032 24 44.04 13.04 0.001 10.18 114 2.20 0.07 3.16 0.21 3.89 9.39 1.84 0.82 0.013 170 45.04 12.89 0.001 10.06 117 2.23 0.09 3.16 0.19 3.65 8.43 0.50 0.82 0.024 266 45.00 12.88 0.001 9.98 117 2.17 0.09 3.16 0.19 3.68 9.44 0.85 0.86 0.025 1362 44.91 12.86 0.001 9.78 121 2.26 0.08 3.16 0.19 3.65 9.45 1.36 0.89 0.046 1660 44.91 12.72 0.001 9.78 121 2.14 0.10 3.16 0.18 3.36 10.79 1.66 0.88 0.017 2798 45.12 12.70 0.001 9.61 124 2.18 0.09 3.16 0.17 3.19 9.48 0.95 0.87 0.03

FD-60

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.50 13.18 0 10.22 110 2.42 0.07 2.16 0.16 4.42 32.39 35.79 0.87 0.232 29 44.11 13.04 0.001 10.18 114 2.51 0.09 2.16 0.19 5.17 14.88 10.58 0.77 0.043 175 44.92 12.89 0.001 10.06 117 2.67 0.09 2.16 0.19 5.14 8.24 0.56 0.75 0.024 273 45.13 12.88 0.001 9.98 117 2.71 0.10 2.16 0.19 5.28 15.14 11.34 0.77 0.055 1368 44.93 12.86 0.001 9.78 121 2.73 0.09 2.16 0.17 4.64 7.60 0.44 0.75 0.036 1666 44.85 12.72 0.001 9.78 121 2.58 0.11 2.16 0.17 4.69 10.50 4.05 0.73 0.027 2803 44.85 12.70 0.001 9.61 124 2.83 0.10 2.16 0.17 4.61 7.84 0.54 0.77 0.02

EFDS-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.62 13.18 0 10.22 110 2.28 0.08 2.16 0.19 5.25 6.69 0.51 0.69 0.032 31 44.16 13.04 0.001 10.18 114 2.44 0.08 2.16 0.22 6.17 7.23 0.89 0.72 0.033 177 44.98 12.89 0.001 10.06 117 2.44 0.09 2.16 0.22 6.03 7.06 0.36 0.75 0.034 275 44.97 12.88 0.001 9.98 117 2.40 0.09 2.16 0.23 6.31 7.87 2.15 0.71 0.035 1371 44.92 12.86 0.001 9.78 121 2.61 0.09 2.16 0.22 6.00 7.68 0.36 0.83 0.016 1668 44.93 12.72 0.001 9.78 121 2.59 0.10 2.16 0.22 6.08 9.19 2.06 0.85 0.037 2805 44.83 12.70 0.001 9.61 124 2.63 0.09 2.16 0.21 5.86 7.46 0.23 0.80 0.02

FCD-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 1 44.75 13.18 0 10.22 110.00 2.01 0.07 1.17 0.14 7.33 9.60 0.43 0.94 0.032 26 43.94 13.04 0.001 10.18 114.00 2.15 0.07 1.17 0.17 8.72 11.51 1.82 1.18 0.203 172 45.10 12.89 0.001 10.06 117.00 2.22 0.09 1.17 0.17 8.56 10.77 2.00 0.92 0.024 270 45.15 12.88 0.001 9.98 117.00 2.27 0.09 1.17 0.16 8.31 10.13 1.22 0.92 0.045 1365 44.90 12.86 0.001 9.78 121.00 2.35 0.08 1.17 0.16 8.10 10.75 1.66 0.92 0.016 1663 44.84 12.72 0.001 9.78 121.00 2.21 0.10 1.17 0.16 8.21 9.27 0.65 0.89 0.027 2800 45.01 12.70 0.001 9.61 124.00 2.28 0.07 1.17 0.16 8.00 10.15 2.27 0.89 0.02

Table C.3 - 0.001-vol% UK-High friability MR fluid data.

214

LHG-8

Spot #Time (min)

Viscosity (cP)

Moisture Content (wt. %) Fluid pH

Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 145 45.60 13.20 10.40 103 1.60 0.07 1.17 0.20 10.21 33.40 6.07 2.18 0.032 256 45.02 13.03 10.30 107 1.69 0.07 1.17 0.18 9.23 23.10 7.00 1.12 0.193 380 45.30 12.94 10.00 112 1.63 0.06 1.17 0.15 7.444 486 45.12 12.92 9.97 112 1.62 0.06 1.17 0.14 7.18 45.50 17.80 1.52 0.285 1619 45.03 12.88 9.98 112 1.63 0.06 1.17 0.13 6.876 1946 44.90 12.76 9.89 112 1.62 0.06 1.17 0.12 5.95 38.70 6.10 2.27 0.617 3035 44.99 13.07 9.70 109 1.45 0.07 1.17 0.11 5.38 39.90 6.85 2.67 0.688 4804 44.99 12.79 9.56 113 1.53 0.04 2.16 0.17 4.839 9156 45.09 12.79 9.46 111 1.25 0.05 2.16 0.16 4.31 48.60 12.00 2.88 0.87

10 10255 45.06 12.87 9.39 113 1.34 0.05 2.16 0.15 4.2511 13207 45.17 12.82 9.38 109 1.32 0.07 2.16 0.15 4.06 29.70 6.97 1.39 0.24

FS

Spot #Time (min)

Viscosity (cP)

Moisture Content Fluid pH

Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 149 45.45 13.20 10.40 103 2.13 0.07 3.16 0.12 2.26 10.82 0.74 0.91 0.022 260 45.41 13.03 10.30 107 2.18 0.07 3.16 0.12 2.24 9.87 1.34 0.86 0.023 382 45.20 12.94 10.00 112 2.14 0.08 3.16 0.11 2.054 488 45.09 12.92 9.97 112 2.10 0.06 3.16 0.11 2.09 9.23 0.55 0.95 0.115 1621 44.95 12.88 9.98 112 2.24 0.07 4.16 0.14 1.986 2008 44.95 12.76 9.89 112 2.23 0.08 4.16 0.14 2.06 9.26 0.89 0.92 0.027 3040 45.03 13.07 9.70 109 2.19 0.12 4.16 0.16 2.31 10.18 3.43 0.84 0.028 4808 44.95 12.79 9.56 113 2.15 0.08 4.16 0.10 1.469 9160 45.19 12.79 9.46 111 1.59 0.08 4.16 0.07 1.07 34.40 11.91 1.96 0.46

10 10559 45.11 12.87 9.39 113 1.60 0.06 5.22 0.04 0.4411 13215 44.88 12.82 9.38 109 1.55 0.08 15.00 0.07 0.30 78.60 38.40 3.59 1.37

BK-7

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 152 45.70 13.20 10.40 103 2.11 0.07 2.16 0.17 4.58 9.42 1.14 0.86 0.042 263 45.17 13.03 10.30 107 2.22 0.08 2.16 0.16 4.533 384 45.15 12.94 10.00 112 2.25 0.07 2.16 0.14 3.97 17.50 3.60 0.87 0.024 492 45.23 12.92 9.97 112 2.32 0.07 2.16 0.15 4.085 1623 45.03 12.88 9.98 112 2.40 0.07 2.16 0.15 4.25 10.00 1.20 0.81 0.026 2011 45.11 12.76 9.89 112 2.36 0.08 2.16 0.15 4.14 11.90 2.60 0.79 0.037 3043 45.13 13.07 9.70 109 2.41 0.11 2.16 0.13 3.69 8.90 0.80 0.79 0.038 4812 44.94 12.79 9.56 113 2.47 0.09 2.16 0.12 3.449 9162 45.24 12.79 9.46 111 2.08 0.14 2.16 0.10 2.67 10.70 1.90 0.84 0.03

10 10562 45.05 12.87 9.39 113 2.14 0.13 3.16 0.14 2.7011 13220 44.98 12.82 9.38 109 1.87 0.15 3.16 0.11 2.05 8.30 0.20 0.82 0.04

FD-60

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 159 45.74 13.20 10.40 103 2.33 0.08 2.16 0.21 5.94 11.36 0.70 1.27 0.102 268 45.18 13.03 10.30 107 2.43 0.08 2.16 0.22 6.00 12.90 2.09 1.14 0.133 393 45.85 12.94 10.00 112 2.55 0.09 2.16 0.18 5.084 502 45.21 12.92 9.97 112 2.47 0.08 2.16 0.20 5.58 9.08 1.45 0.91 0.045 1628 44.98 12.88 9.98 112 2.66 0.09 2.16 0.20 5.506 2017 45.16 12.76 9.89 112 2.62 0.08 2.16 0.19 5.39 8.62 1.24 0.81 0.077 3048 45.17 13.07 9.70 109 2.67 0.13 2.16 0.16 4.56 8.06 0.63 0.90 0.078 4818 44.85 12.79 9.56 113 2.77 0.13 2.16 0.16 4.429 9167 44.98 12.79 9.46 111 2.74 0.16 2.16 0.14 3.86 9.82 2.06 0.96 0.05

10 10568 44.98 12.87 9.39 113 2.91 0.16 2.16 0.13 3.6111 13225 45.04 12.82 9.38 109 2.65 0.16 2.16 0.14 3.86 9.96 0.59 1.02 0.09

EFDS-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 163 45.70 13.20 10.40 103 2.33 0.08 1.17 0.14 7.13 11.20 0.76 1.25 0.102 271 45.06 13.03 10.30 107 2.40 0.08 1.17 0.14 6.97 11.23 1.62 1.23 0.143 395 44.73 12.94 10.00 112 2.34 0.07 1.17 0.13 6.414 504 45.16 12.92 9.97 112 2.48 0.07 1.17 0.13 6.87 11.84 0.89 1.39 0.145 1630 44.95 12.88 9.98 112 2.47 0.09 1.17 0.12 6.366 2017 44.87 12.76 9.89 112 2.66 0.08 1.17 0.13 6.62 13.32 1.40 1.47 0.047 3050 45.03 13.07 9.70 109 2.68 0.10 1.17 0.13 6.62 10.71 0.87 1.22 0.058 4820 44.94 12.79 9.56 113 2.87 0.11 1.17 0.12 6.219 9170 45.07 12.79 9.46 111 2.98 0.22 1.17 0.12 6.10 12.72 2.60 1.29 0.13

10 10568 45.03 12.87 9.39 113 2.88 0.15 1.17 0.10 5.3311 13228 45.16 12.82 9.38 109 2.59 0.14 2.16 0.21 5.89 11.31 1.33 1.20 0.21

FCD-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 155 45.76 13.20 10.40 103.00 2.07 0.07 1.17 0.18 9.38 9.24 0.46 0.93 0.072 265 44.73 13.03 10.30 107.00 2.19 0.07 1.17 0.18 9.183 389 45.02 12.94 10.00 112.00 2.27 0.06 1.17 0.16 8.41 10.33 0.50 1.03 0.054 500 45.15 12.92 9.97 112.00 2.10 0.06 1.17 0.18 9.035 1326 44.91 12.80 9.98 112.00 2.38 0.07 1.17 0.17 8.87 10.48 1.40 0.90 0.046 2014 45.18 12.76 9.89 112.00 2.44 0.07 1.17 0.18 9.23 9.79 2.23 0.89 0.047 3045 45.04 13.07 9.70 109.00 2.24 0.08 1.17 0.16 8.36 10.26 0.64 0.96 0.048 4814 44.81 12.79 9.56 113.00 2.45 0.07 1.17 0.17 8.729 9165 45.07 12.79 9.46 111.00 2.23 0.09 1.17 0.16 8.36 11.99 2.84 0.99 0.11

10 10565 45.04 12.87 9.39 113.00 2.32 0.06 1.17 0.15 7.4911 13222 44.91 12.82 9.38 109.00 2.36 0.11 1.17 0.15 7.64 12.22 0.86 1.03 0.11

Table C.4 - 0.005-vol% UK-Low friability MR fluid data.

215

LHG-8

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 128 44.91 13.20 10.08 110 1.45 0.08 1.17 0.36 18.21 16.66 4.17 1.14 0.042 252 45.85 13.09 10.02 115 1.59 0.05 1.17 0.31 15.85 12.89 3.06 1.06 0.063 353 45.11 12.95 9.83 115 1.36 0.04 1.17 0.27 13.954 460 45.40 12.94 9.79 115 1.45 0.05 1.17 0.26 13.28 11.02 2.21 1.12 0.265 1608 44.97 12.87 9.78 122 1.44 0.05 1.17 0.21 10.566 1891 45.06 12.81 9.76 122 1.47 0.04 1.17 0.21 10.67 11.48 2.14 1.02 0.057 3100 45.05 12.76 9.62 122 1.49 0.04 1.17 0.19 9.69 14.88 1.92 1.26 0.128 4771 44.97 12.70 9.43 123 1.32 0.04 1.17 0.18 9.039 9096 44.87 12.72 9.37 106 1.28 0.05 1.17 0.15 7.54 17.42 4.77 1.14 0.12

10 9566 45.15 13.08 9.31 112 1.28 0.05 1.17 0.14 7.3311 10996 44.99 12.75 9.27 120 1.30 0.04 1.17 0.14 6.97 20.74 3.74 1.22 0.08

FS

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 132 44.93 13.20 10.08 110 2.26 0.06 3.16 0.15 2.89 8.89 0.84 0.82 0.012 255 45.35 13.09 10.02 115 2.46 0.06 3.16 0.14 2.58 12.18 4.01 0.80 0.003 356 45.02 12.95 9.83 115 2.33 0.05 3.16 0.12 2.304 463 45.33 12.94 9.79 115 2.28 0.05 3.16 0.13 2.39 10.62 2.49 0.81 0.015 1610 45.00 12.87 9.78 122 2.20 0.06 4.16 0.14 2.056 1889 44.98 12.81 9.76 122 2.33 0.05 4.16 0.13 1.89 13.08 1.64 0.81 0.067 3104 45.13 12.76 9.62 122 2.22 0.06 4.16 0.12 1.72 12.67 2.44 0.84 0.048 4783 45.06 12.70 9.43 123 2.08 0.05 5.22 0.12 1.349 9094 44.92 12.72 9.37 106 1.86 0.07 5.22 0.10 1.10 25.42 7.74 1.17 0.15

10 9563 45.03 13.08 9.31 112 1.97 0.08 10.00 0.19 1.1211 10993 45.01 12.75 9.27 120 2.26 0.12 8.00 0.16 1.21 13.36 5.14 0.82 0.05

BK-7

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 135 45.23 13.20 10.08 110 2.15 0.06 2.16 0.20 5.58 9.89 2.14 0.90 0.042 258 45.27 1309.00 10.02 115 2.23 0.06 2.16 0.20 5.47 8.64 0.70 0.85 0.023 358 44.56 12.95 9.83 115 2.07 0.05 2.16 0.18 5.084 465 45.33 12.94 9.79 115 2.02 0.05 2.16 0.18 5.11 8.54 0.35 0.87 0.035 1613 45.04 12.87 9.78 122 2.11 0.05 2.16 0.15 4.066 1888 44.95 12.81 9.76 122 2.10 0.05 2.16 0.17 4.81 13.48 2.96 0.93 0.127 3108 45.21 12.76 9.62 122 2.13 0.05 2.16 0.16 4.53 9.56 0.83 0.92 0.068 4777 45.03 12.70 9.43 123 2.12 0.06 2.16 0.15 4.039 9091 45.06 12.72 9.37 106 2.17 0.10 2.16 0.16 4.44 9.77 1.48 0.85 0.03

10 9565 45.05 13.08 9.31 112 2.29 0.11 2.16 0.15 4.0311 10996 45.14 12.75 9.27 120 2.60 0.16 2.16 0.14 3.89 9.87 1.13 0.87 0.05

FD-60

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 143 45.44 13.20 10.08 110 2.10 0.06 2.16 0.24 6.69 9.15 0.96 0.97 0.062 263 45.10 13.09 10.02 115 2.20 0.07 2.16 0.23 6.33 11.49 2.97 1.04 0.043 365 43.91 12.95 9.83 115 2.00 0.06 2.16 0.23 6.504 469 45.25 12.94 9.79 115 2.13 0.06 2.16 0.23 6.47 12.00 2.16 1.07 0.075 1618 44.90 12.87 9.78 122 2.14 0.06 2.16 0.21 5.866 1886 44.98 12.81 9.76 122 2.18 0.05 2.16 0.19 5.33 10.88 2.02 0.92 0.057 3115 45.05 12.76 9.62 122 2.16 0.05 2.16 0.20 5.56 9.69 0.70 0.99 0.048 4785 44.91 12.70 9.43 123 2.14 0.08 2.16 0.19 5.369 9099 44.87 12.72 9.37 106 2.23 0.11 2.16 0.19 5.14 9.81 1.34 0.96 0.02

10 9570 45.08 13.08 9.31 112 2.46 0.12 2.16 0.18 4.9211 11001 45.06 12.75 9.27 120 2.72 0.15 2.16 0.17 4.81 10.63 1.82 1.09 0.05

EFDS-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 146 44.58 13.20 10.08 110 2.70 0.06 1.17 0.16 8.15 13.98 1.59 1.61 0.102 266 44.99 13.09 10.02 115 2.76 0.07 1.17 0.16 8.05 12.90 1.44 1.51 0.103 369 45.76 12.95 9.83 115 2.82 0.06 1.17 0.15 7.904 471 45.14 12.94 9.79 115 2.73 0.06 1.17 0.15 7.59 16.42 1.16 2.07 0.145 1621 44.92 12.87 9.78 122 2.85 0.06 1.17 0.14 7.136 1883 44.88 12.81 9.76 122 2.99 0.06 1.17 0.14 7.13 12.72 1.43 1.44 0.167 3117 44.97 12.76 9.62 122 2.83 0.06 1.17 0.14 7.33 11.38 0.63 1.36 0.088 4788 44.78 12.70 9.43 123 2.79 0.06 1.17 0.14 7.089 9102 44.79 12.72 9.37 106 2.70 0.09 1.17 0.14 6.97 14.16 3.43 1.53 0.09

10 9574 45.08 13.08 9.31 112 2.97 0.14 1.17 0.14 7.1811 11004 45.19 12.75 9.27 120 3.01 0.12 1.17 0.14 7.13 13.74 1.63 1.52 0.19

FCD-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 139 45.05 13.20 10.08 110.00 2.08 0.05 1.17 0.22 11.23 11.37 1.02 1.03 0.012 261 45.18 13.09 10.02 115.00 2.24 0.06 1.17 0.21 10.62 15.04 2.33 0.98 0.043 263 44.49 12.95 9.83 115.00 2.11 0.05 1.17 0.20 10.154 467 45.21 12.94 9.79 115.00 2.19 0.06 1.17 0.21 10.92 9.79 1.32 0.91 0.015 1616 44.93 12.87 9.78 122.00 2.22 0.05 1.17 0.20 10.006 1876 44.95 12.81 9.76 122.00 2.15 0.06 1.17 0.19 9.85 11.94 2.40 0.91 0.057 3111 45.20 12.76 9.62 122.00 2.28 0.05 1.17 0.19 9.59 9.54 0.88 0.94 0.048 4791 44.96 12.70 9.43 123.00 2.16 0.05 1.17 0.19 9.749 9104 44.82 12.72 9.37 106.00 2.11 0.07 1.17 0.18 9.28 8.40 0.24 0.86 0.02

10 9568 45.16 13.08 9.31 112.00 2.28 0.09 1.17 0.19 9.5911 10998 45.04 12.75 9.27 120.00 2.43 0.08 1.17 0.18 9.28 9.57 0.79 0.89 0.01

Table C.5 - 0.005-vol% UK-Medium A friability MR fluid data.

216

LHG-8

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 117 44.81 13.27 10.46 109 1.48 0.06 1.17 0.21 10.62 14.28 3.20 1.15 0.052 208 45.09 13.09 10.35 115 1.44 0.06 1.17 0.19 9.90 15.90 4.86 1.26 0.143 327 45.18 12.90 10.11 118 1.39 0.06 1.17 0.18 9.384 445 45.09 12.88 10.02 121 1.47 0.07 1.17 0.19 9.54 21.38 3.44 1.43 0.075 1591 44.99 12.84 9.78 123 1.52 0.07 1.17 0.17 8.776 1878 45.06 12.63 9.70 122 1.44 0.09 1.17 0.19 9.85 55.06 27.01 3.10 1.607 3050 45.04 12.68 9.54 121 1.45 0.05 1.17 0.16 8.41 24.78 3.36 1.61 0.068 3362 45.07 12.70 9.56 121 1.31 0.04 1.17 0.17 8.929 7687 44.86 12.63 9.43 123 1.36 0.05 1.17 0.15 7.64 16.71 5.66 1.03 0.0910 8826 45.01 12.63 9.33 119 1.34 0.05 1.17 0.14 7.0311 9803 45.18 12.65 9.30 120 1.17 0.13 6.82 22.00 3.56 1.91 0.3512 10997 44.97 12.62 9.26 123 1.31 0.04 1.17 0.13 6.82

FS

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 120 44.84 13.27 10.46 109 2.20 0.08 3.16 0.13 2.39 11.52 2.08 0.81 0.022 211 45.00 13.09 10.35 115 2.33 0.08 4.16 0.16 2.34 11.25 1.07 0.82 0.013 330 45.25 12.90 10.11 118 2.08 0.08 4.16 0.16 2.294 447 44.94 12.88 10.02 121 2.42 0.12 4.16 0.16 2.28 8.03 0.43 0.81 0.015 1594 44.91 12.84 9.78 123 2.41 0.10 5.22 0.15 1.716 1880 44.95 12.63 9.70 122 2.33 0.12 5.22 0.18 2.02 12.50 1.45 0.87 0.017 3052 45.08 12.68 9.54 121 2.27 0.06 5.22 0.16 1.86 10.39 2.79 0.84 0.018 3365 44.97 12.70 9.56 121 2.00 0.07 5.22 0.14 1.629 7689 44.82 12.63 9.43 123 2.28 0.11 5.22 0.11 1.26 9.82 1.01 0.84 0.0310 8828 44.94 12.63 9.33 119 2.29 0.11 5.22 0.13 1.4611 9809 45.26 12.65 9.30 120 5.22 0.12 1.39 9.51 1.78 0.79 0.0112 10999 44.94 12.62 9.26 123 2.15 0.12 5.22 0.11 1.28 9.95 3.60 0.77 0.02

BK-7

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 122 44.93 13.27 10.46 109 2.11 0.07 2.16 0.17 4.67 9.10 0.52 0.88 0.012 213 44.97 13.09 10.35 115 2.19 0.08 2.16 0.17 4.58 11.25 2.86 0.89 0.033 332 45.22 12.90 10.11 118 2.11 0.08 2.16 0.16 4.534 450 45.07 12.88 10.02 121 2.25 0.09 2.16 0.17 4.58 15.28 2.56 0.83 0.095 1596 44.98 12.84 9.78 123 2.33 0.09 2.16 0.16 4.566 1883 44.93 12.63 9.70 122 2.39 0.12 2.16 0.18 5.06 10.12 2.34 0.84 0.027 3055 44.91 12.68 9.54 121 2.23 0.06 2.16 0.18 4.86 8.83 0.73 0.87 0.038 3368 45.13 12.70 9.56 121 2.24 0.07 2.16 0.18 4.869 7692 45.08 12.63 9.43 123 2.52 0.09 2.16 0.17 4.72 10.09 2.21 0.84 0.0310 8831 44.98 12.63 9.33 119 2.67 0.13 2.16 0.18 4.9711 9811 45.27 12.65 9.30 120 2.16 0.18 5.03 12.35 4.78 0.84 0.0412 11002 44.98 12.62 9.26 123 2.64 0.15 2.16 0.17 4.78 9.78 2.56 0.83 0.02

FD-60

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 128 45.63 13.27 10.46 109 2.02 0.07 2.16 0.22 6.22 7.48 0.40 0.75 0.012 218 44.96 13.09 10.35 115 2.13 0.08 2.16 0.22 6.17 7.41 0.28 0.74 0.023 337 45.08 12.90 10.11 118 2.08 0.08 2.16 0.23 6.394 456 45.06 12.88 10.02 121 2.23 0.09 2.16 0.23 6.42 7.35 0.52 0.73 0.015 1602 44.93 12.84 9.78 123 2.37 0.10 2.16 0.23 6.396 1888 44.99 12.63 9.70 122 2.39 0.13 2.16 0.23 6.47 7.00 0.16 0.72 0.027 3060 44.95 12.68 9.54 121 2.31 0.15 2.16 0.22 6.00 7.51 0.16 0.75 0.028 3357 45.01 12.70 9.56 121 2.22 0.08 2.16 0.21 5.949 7682 44.94 12.63 9.43 123 2.82 0.17 2.16 0.21 5.89 8.84 1.32 0.76 0.0310 8837 44.83 12.63 9.33 119 2.70 0.20 2.16 0.20 5.6111 9826 45.22 12.65 9.30 120 2.16 0.21 5.75 9.06 3.00 0.78 0.0312 11007 44.75 12.62 9.26 123 2.69 0.14 2.16 0.19 5.36 7.37 0.25 0.76 0.02

EFDS-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 130 45.04 13.27 10.46 109 2.39 0.09 1.17 0.14 7.38 12.92 1.14 1.41 0.042 221 44.96 13.09 10.35 115 2.53 0.09 1.17 0.14 7.33 14.82 4.04 1.98 0.803 339 45.04 12.90 10.11 118 2.66 0.09 1.17 0.14 7.284 459 45.13 12.88 10.02 121 2.72 0.10 1.17 0.14 7.23 13.20 1.63 1.57 0.225 1604 44.93 12.84 9.78 123 2.73 0.10 1.17 0.16 8.056 1891 44.96 12.63 9.70 122 2.86 0.13 1.17 0.17 8.51 13.54 1.92 1.46 0.117 3062 45.02 12.68 9.54 121 2.71 0.07 1.17 0.15 7.64 9.95 0.57 1.07 0.098 3360 44.96 12.70 9.56 121 2.64 0.07 1.17 0.15 7.649 7684 44.79 12.63 9.43 123 2.98 0.12 1.17 0.17 8.72 11.92 2.16 1.20 0.2110 8839 44.83 12.63 9.33 119 2.99 0.13 1.17 0.17 8.4611 9836 45.24 12.65 9.30 120 2.95 0.18 1.17 0.18 9.08 11.36 0.80 1.34 0.1612 11009 44.94 12.62 9.26 123 2.88 0.09 1.17 0.16 8.31 11.11 2.45 1.20 0.07

FCD-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 125 44.64 13.27 10.46 109.00 2.17 0.08 1.17 0.20 10.05 11.43 2.22 0.95 0.032 216 44.90 13.09 10.35 115.00 2.26 0.08 1.17 0.19 9.85 10.24 1.89 0.96 0.013 334 45.16 12.90 10.11 118.00 2.27 0.08 1.17 0.20 10.314 454 45.05 12.88 10.02 121.00 2.28 0.09 1.17 0.20 10.26 14.34 2.88 1.08 0.375 1600 44.86 12.84 9.78 123.00 2.51 0.09 1.17 0.20 10.106 1886 44.96 12.63 9.70 122.00 2.52 0.13 1.17 0.21 10.56 11.96 2.83 1.02 0.187 3058 44.97 12.68 9.54 121.00 2.37 0.07 1.17 0.20 10.46 9.17 0.85 0.87 0.078 3355 44.99 12.70 9.56 121.00 2.19 0.07 1.17 0.20 10.419 7694 45.02 12.63 9.43 123.00 2.42 0.10 1.17 0.22 11.44 13.36 2.80 0.90 0.0810 8834 44.94 12.63 9.33 119.00 2.35 0.07 1.17 0.21 10.7211 9814 45.24 12.65 9.30 120.00 1.17 0.21 10.87 14.06 2.54 1.01 0.1612 11005 44.71 12.62 9.26 123.00 2.47 0.06 1.17 0.21 10.51 10.56 1.44 0.91 0.05


217

LHG-8

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 105 44.64 13.17 10.31 101 1.41 0.07 1.17 0.26 13.08 10.66 3.85 1.00 0.132 215 44.96 13.05 10.22 105 1.40 0.10 1.17 0.24 12.31 9.45 0.31 0.94 0.013 346 44.77 12.96 10.16 107 1.17 0.09 1.17 0.23 11.85 9.53 0.43 0.93 0.014 490 44.99 12.90 10.05 108 1.41 0.08 1.17 0.22 11.08 12.12 1.94 0.92 0.015 1554 44.88 12.83 9.96 107 1.30 0.05 1.17 0.19 9.64 10.24 0.93 0.92 0.016 1980 44.83 12.75 9.82 108 1.32 0.06 1.17 0.17 8.92 11.37 3.19 0.87 0.027 3296 44.86 12.75 9.71 107 1.39 0.06 1.17 0.17 8.51 11.35 4.34 0.88 0.028 4467 45.14 12.64 9.63 108 1.28 0.06 1.17 0.14 7.23 15.92 2.99 0.90 0.039 11740 45.13 12.86 9.45 111 1.20 0.05 2.16 0.22 6.06 16.44 5.79 1.07 0.14

10 13095 44.91 12.80 9.38 112 1.08 0.04 2.16 0.21 5.92 9.93 1.48 0.96 0.03

FS

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 108 44.56 13.17 10.31 101 1.88 0.11 3.16 0.15 2.75 12.70 1.26 0.84 0.022 218 44.91 13.05 10.22 105 1.99 0.08 3.16 0.14 2.68 11.60 3.07 0.81 0.013 349 45.02 12.96 10.16 107 2.04 0.11 3.16 0.14 2.62 11.48 3.12 0.80 0.034 493 44.93 12.90 10.05 108 2.06 0.11 4.16 0.18 2.54 11.47 3.12 0.79 0.015 1558 44.83 12.83 9.96 107 2.08 0.07 4.16 0.17 2.47 9.02 1.71 0.79 0.016 1911 44.93 12.75 9.82 108 2.08 0.09 4.16 0.16 2.34 8.72 0.93 0.80 0.017 3300 44.84 12.75 9.71 107 2.10 0.08 4.16 0.15 2.15 11.72 2.17 0.78 0.018 4470 45.10 12.64 9.63 108 1.89 0.10 4.16 0.12 1.79 9.49 1.54 0.81 0.019 11745 45.19 12.86 9.45 111 1.60 0.07 5.22 0.06 0.74 10.82 1.43 0.80 0.01

10 13097 44.94 12.80 9.38 112 1.46 0.06 10.00 0.10 0.61 9.60 1.98 0.81 0.02

BK-7

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 111 44.53 13.17 10.31 101 2.03 0.10 2.16 0.20 5.44 13.88 5.32 0.86 0.022 220 44.96 13.05 10.22 105 2.15 0.11 2.16 0.20 5.42 11.28 3.20 0.85 0.043 353 45.06 12.96 10.16 107 2.14 0.11 2.16 0.19 5.33 15.16 2.16 0.87 0.024 496 44.66 12.90 10.05 108 2.19 0.10 2.16 0.19 5.22 8.80 0.71 0.84 0.015 1562 45.01 12.82 9.96 107 2.17 0.07 2.16 0.18 4.92 9.25 2.18 0.83 0.016 1914 44.91 12.75 9.82 108 2.20 0.09 2.16 0.17 4.78 8.29 0.41 0.82 0.027 3302 44.75 12.75 9.71 107 2.42 0.10 2.16 0.16 4.56 9.64 2.06 0.83 0.018 4473 45.14 12.64 9.63 108 1.77 0.11 2.16 0.15 4.22 8.12 0.24 0.82 0.029 11747 45.15 12.96 9.45 111 2.11 0.10 2.16 0.15 4.06 9.68 1.72 0.80 0.02

10 13100 44.96 12.80 9.38 112 2.17 0.11 2.16 0.13 3.69 11.90 4.00 0.81 0.01

FD-60

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 120 44.50 13.17 10.31 101 2.10 0.09 1.17 0.17 8.82 8.73 2.16 0.82 0.042 228 44.96 13.05 10.22 105 2.27 0.10 1.17 0.17 8.56 7.84 0.82 0.82 0.033 361 45.02 12.96 10.16 107 2.35 0.10 1.17 0.16 8.41 7.80 0.44 0.78 0.024 503 44.88 12.90 10.05 108 2.44 0.07 1.17 0.16 8.10 8.25 0.54 0.82 0.045 1569 44.87 12.82 9.96 107 2.32 0.08 1.17 0.16 8.05 11.36 4.38 0.80 0.016 1923 44.98 12.75 9.82 108 2.49 0.10 1.17 0.16 8.36 11.40 3.04 0.83 0.067 3310 44.86 12.75 9.71 107 2.52 0.10 1.17 0.16 8.10 9.74 1.93 0.84 0.028 4482 45.09 12.64 9.63 108 2.41 0.10 1.17 0.15 7.74 10.71 2.64 0.86 0.029 11757 45.01 12.96 9.45 111 2.56 0.12 1.17 0.14 7.18 1.23 1.76 1.02 0.12

10 13110 45.02 12.80 9.38 112 2.45 0.13 1.17 0.13 6.82 9.36 1.14 0.95 0.05

EFDS-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 117 44.61 13.17 10.31 101 2.13 0.11 1.17 0.13 6.51 8.04 1.13 0.74 0.032 226 44.82 13.05 10.22 105 2.27 0.13 1.17 0.13 6.46 8.05 0.94 0.74 0.023 359 45.40 12.96 10.16 107 2.37 0.10 1.17 0.13 6.56 8.91 1.31 0.75 0.024 501 44.93 12.90 10.05 108 2.35 0.10 1.17 0.12 6.31 7.32 0.18 0.73 0.025 1567 44.93 12.82 9.96 107 2.35 0.09 1.17 0.12 6.10 7.71 0.29 0.77 0.026 1921 45.04 12.75 9.82 108 2.46 0.09 1.17 0.12 6.00 7.92 1.66 0.73 0.017 3308 44.70 12.75 9.71 107 2.60 0.13 1.17 0.11 5.79 4.46 1.21 0.71 0.018 4480 45.10 12.64 9.63 108 2.43 0.11 2.16 0.19 5.25 7.97 1.23 0.74 0.029 11754 45.09 12.96 9.45 111 2.57 0.12 2.16 0.17 4.67 8.37 2.46 0.73 0.01

10 13107 44.90 12.80 9.38 112 2.62 0.19 2.16 0.16 4.44 7.38 0.92 0.72 0.03

FCD-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 114 44.78 13.17 10.31 101.00 1.99 0.09 1.17 0.21 10.72 11.90 1.57 0.92 0.052 223 44.85 13.05 10.22 105.00 2.07 0.12 1.17 0.21 10.51 11.56 0.80 0.93 0.023 356 45.04 12.96 10.16 107.00 2.17 0.09 1.17 0.21 10.56 10.08 2.99 0.91 0.044 498 44.90 12.90 10.05 108.00 2.13 0.09 1.17 0.21 10.62 12.71 2.23 0.88 0.035 1564 44.84 12.82 9.96 107.00 2.04 0.07 1.17 0.21 10.56 12.41 5.05 0.90 0.036 1917 44.94 12.75 9.82 108.00 2.13 0.09 1.17 0.21 10.51 9.06 0.29 0.91 0.027 3305 44.93 12.75 9.71 107.00 2.12 0.08 1.17 0.19 9.90 10.67 1.74 0.90 0.038 4476 45.15 12.64 9.63 108.00 2.06 0.08 1.17 0.19 9.85 10.28 0.68 0.90 0.049 11751 45.04 12.96 9.45 111.00 2.12 0.07 1.17 0.18 9.18 9.39 1.09 0.90 0.03

10 13103 44.90 12.80 9.38 112.00 2.03 0.06 1.17 0.18 9.13 9.44 0.64 0.89 0.02 Table C.7 - 0.01-vol% UK-Low friability MR fluid data.

218

LHG-8

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 143 44.99 13.02 10.18 100 1.65 0.08 1.17 0.54 27.49 10.89 1.20 1.08 0.042345678 7299 44.91 12.71 9.39 111 1.45 0.06 1.17 0.25 12.72 12.18 1.64 1.12 0.049 8627 45.17 12.69 9.36 111 1.42 0.06 1.17 0.24 12.3610 10068 45.05 12.70 9.32 112 1.41 0.06 1.17 0.22 11.1811 11654 45.12 12.70 9.31 109 1.37 0.06 1.17 0.22 11.38 15.26 3.85 1.19 0.0512 12984 45.04 12.71 9.23 110 1.43 0.06 1.17 0.21 10.56 11.20 0.92 1.07 0.07

FS

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 147 45.03 13.02 10.18 100 2.10 0.09 3.16 0.18 5.66 10.48 3.38 0.85 0.062 226 44.89 12.93 10.10 103 2.14 0.10 3.16 0.18 5.60 9.51 1.27 0.80 0.023 376 45.08 12.82 9.98 104 2.21 0.10 3.16 0.19 5.924 494 45.00 12.71 9.95 104 2.19 0.11 3.16 0.17 5.22 11.29 1.31 0.76 0.045 1626 45.10 12.65 9.78 105 2.22 0.09 3.16 0.17 5.256 1934 45.26 12.64 9.70 105 2.09 0.11 3.16 0.17 5.22 1.02 3.10 0.83 0.037 3036 45.07 12.68 9.66 109 2.18 0.10 3.16 0.15 4.72 12.36 0.61 0.86 0.038 7221 44.94 12.66 9.38 113 2.34 0.14 3.16 0.14 4.53 12.05 2.31 0.89 0.059 8629 45.20 12.69 9.36 111 2.39 0.19 3.16 0.13 3.9910 10071 45.07 12.70 9.32 112 2.57 0.17 4.16 0.13 3.13 10.64 2.56 0.84 0.0211 11657 45.06 12.70 9.31 109 2.28 0.20 4.16 0.12 2.8412 12987 45.15 12.71 9.23 110 2.25 0.19 5.22 0.12 2.34 10.27 1.00 0.87 0.03

BK-7

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 151 45.10 13.02 10.18 100 2.30 0.09 2.16 0.25 7.03 12.14 0.72 1.08 0.082 229 44.81 12.93 10.10 103 2.44 0.10 2.16 0.25 7.00 10.48 0.26 1.07 0.063 380 45.30 12.82 9.98 104 2.50 0.11 2.16 0.25 7.00 10.91 1.29 1.01 0.054 496 45.10 12.71 9.95 104 2.50 0.11 2.16 0.26 7.08 10.32 0.33 1.00 0.055 1629 45.09 12.65 9.78 105 2.27 0.09 2.16 0.24 6.536 1936 45.27 12.64 9.70 105 2.42 0.09 2.16 0.25 6.92 10.66 1.24 1.05 0.057 3.38 45.09 12.68 9.66 109 2.61 0.11 2.16 0.22 6.00 12.20 1.35 1.02 0.048 7225 45.09 12.66 9.38 113 2.99 0.20 2.16 0.20 5.61 12.92 1.63 1.00 0.059 8633 45.13 12.69 9.36 111 2.93 0.17 2.16 0.20 5.5310 10074 45.31 12.70 9.32 112 2.97 0.18 2.16 0.19 5.28 12.47 3.55 0.98 0.0311 11659 45.08 12.70 9.31 109 2.71 0.14 2.16 0.19 5.3112 12990 45.30 12.71 9.23 110 2.75 0.16 2.16 0.18 4.94 11.98 1.62 0.96 0.08

FD-60

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 160 44.87 13.02 10.18 100 2.28 0.10 1.17 0.16 8.00 9.86 1.80 0.92 0.072 236 44.80 12.93 10.10 103 2.47 0.11 1.17 0.16 7.95 9.21 0.55 0.94 0.063 387 45.15 12.82 9.98 104 2.46 0.12 1.17 0.14 7.28 14.30 2.14 1.35 0.134 502 44.97 12.71 9.95 104 2.47 0.12 1.17 0.14 7.33 10.60 1.24 1.12 0.125 1638 45.01 12.65 9.78 105 2.61 0.12 1.17 0.13 6.77 11.20 0.64 1.00 0.096 1941 45.18 12.64 9.70 105 2.63 0.12 1.17 0.14 6.92 10.80 1.29 0.90 0.027 3044 45.17 12.68 9.66 109 2.64 0.16 1.17 0.12 6.00 10.40 1.67 0.86 0.048 7231 45.12 12.66 9.38 113 2.97 0.15 1.17 0.12 6.00 10.40 2.08 0.97 0.069 8641 45.01 12.69 9.36 111 2.95 0.19 1.17 0.12 6.21 11.00 0.71 0.90 0.0210 10079 45.11 12.70 9.32 112 2.95 0.17 1.17 0.12 5.90 9.44 0.96 0.89 0.0611 11666 44.90 12.70 9.31 109 2.96 0.21 2.16 0.21 5.92 10.10 1.52 0.90 0.0512 12994 45.18 12.71 9.23 110 2.70 0.13 2.16 0.20 5.67 10.70 1.44 0.92 0.03

EFDS-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 163 44.95 13.02 10.18 100 2.22 0.09 1.17 0.21 10.72 10.86 1.69 1.10 0.142 238 45.08 12.93 10.10 103 2.43 0.10 1.17 0.22 11.03 11.18 1.70 1.32 0.273 390 45.23 12.82 9.98 104 2.33 0.12 1.17 0.21 10.77 16.34 3.40 1.24 0.094 505 44.96 12.71 9.95 104 2.42 0.13 1.17 0.20 10.31 13.92 2.19 1.20 0.045 1641 44.97 12.65 9.78 105 2.56 0.12 1.17 0.20 10.31 10.50 1.17 1.14 0.126 1944 45.23 12.64 9.70 105 2.67 0.15 1.17 0.20 10.05 9.20 1.34 0.93 0.067 3047 45.10 12.68 9.66 109 2.65 0.16 1.17 0.19 9.548 7236 45.03 12.66 9.38 113 2.69 0.14 1.17 0.19 9.49 8.30 0.57 0.93 0.089 8644 44.92 12.69 9.36 111 2.97 0.15 1.17 0.19 9.74 9.06 1.74 0.87 0.0910 10082 45.14 12.70 9.32 112 2.90 0.16 1.17 0.18 9.38 10.42 1.21 1.02 0.0511 11668 44.94 12.70 9.31 109 2.84 0.19 1.17 0.18 9.08 9.16 0.98 0.97 0.0512 12995 45.27 12.71 9.23 110 2.92 0.17 1.17 0.17 8.82 8.51 0.23 0.93 0.04

FCD-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 156 44.97 13.02 10.18 100.00 2.11 0.09 1.17 0.25 12.97 9.72 0.48 0.93 0.022 233 45.07 12.93 10.10 103.00 2.24 0.09 1.17 0.26 13.33 9.21 1.00 0.89 0.033 383 45.28 12.82 9.98 104.00 2.26 0.10 1.17 0.25 12.82 11.23 3.05 0.89 0.034 499 44.89 12.71 9.95 104.00 2.28 0.10 1.17 2.56 131.28 20.34 10.24 0.98 0.065 1635 44.96 12.65 9.78 105.00 2.27 0.07 1.17 0.24 12.21 8.83 0.24 0.89 0.046 1939 45.17 12.64 9.70 105.00 2.21 0.10 1.17 0.23 11.85 12.18 1.71 0.91 0.037 3042 45.23 12.68 9.66 109.00 2.32 0.09 1.17 0.22 11.08 9.32 0.34 0.95 0.028 7228 44.96 12.66 9.38 113.00 2.46 0.08 1.17 2.17 111.28 9.17 0.38 0.92 0.039 8638 45.08 12.69 9.36 111.00 2.60 0.14 1.17 0.22 11.44 11.28 2.60 0.96 0.0810 10077 45.17 12.70 9.32 112.00 2.75 0.15 1.17 0.21 10.87 11.12 2.71 0.99 0.0211 11662 45.95 12.70 9.31 109.00 2.66 0.18 1.17 0.21 10.56 10.34 0.29 0.99 0.0212 12993 45.20 12.71 9.23 110.00 2.52 0.10 1.17 0.21 10.56 15.84 3.40 1.09 0.05

Table C.8 - 0.01-vol% UK-Medium A friability MR fluid data.

219

LHG-8

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 147 44.67 13.24 10.31 106 1.60 0.08 1.17 0.28 14.31 27.80 10.01 3.13 1.652 236 44.87 13.06 10.23 110 1.58 0.08 1.17 0.27 13.74 17.14 3.86 1.35 0.353 362 44.93 13.04 10.15 110 1.59 0.08 1.17 0.28 14.364 490 45.32 12.94 10.13 112 1.57 0.09 1.17 0.28 14.10 15.50 2.25 1.23 0.135 1585 44.65 12.92 9.94 115 1.59 0.05 1.17 0.24 12.316 1777 45.09 12.72 9.82 115 1.57 0.06 1.17 0.26 13.13 14.05 2.85 1.09 0.067 3005 44.84 12.74 9.76 115 1.59 0.04 1.17 0.25 12.87 11.92 1.44 1.05 0.048 7630 44.96 12.70 9.54 115 1.46 0.05 1.17 0.24 12.05 13.04 1.82 1.06 0.099 8811 45.12 12.71 9.53 116 1.46 0.05 1.17 0.24 12.3110 10528 44.89 12.77 9.39 117 1.47 0.04 1.17 0.22 11.03 11.18 1.52 1.10 0.0411 11999 45.04 12.75 9.40 119 1.46 0.04 1.17 0.22 11.0812 12474 45.13 12.73 9.34 120 1.52 0.05 1.17 0.21 10.97 32.92 8.70 4.02 1.37

FS

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 150 44.61 13.24 10.31 106 1.99 0.09 3.16 0.15 2.83 14.50 1.51 0.91 0.052 240 45.01 13.06 10.23 110 2.14 0.10 3.16 0.15 2.85 11.60 1.86 0.97 0.053 365 44.81 13.04 10.15 110 2.11 0.10 3.16 0.14 2.604 493 45.45 12.94 10.13 112 2.13 0.10 3.16 0.14 2.62 11.36 1.61 0.94 0.065 1587 44.84 12.92 9.94 115 2.19 0.07 3.16 0.14 2.626 1780 45.02 12.72 9.82 115 2.17 0.07 3.16 0.15 2.77 11.78 0.91 0.86 0.057 3008 45.00 12.74 9.76 115 2.17 0.07 3.16 0.14 2.60 12.22 1.14 0.94 0.068 7630 45.01 12.70 9.54 115 2.24 0.09 3.16 0.13 2.53 13.98 2.07 0.86 0.039 8813 44.91 12.71 9.53 116 2.42 0.15 3.16 0.14 2.6610 10531 44.80 12.77 9.39 117 2.26 0.12 3.16 0.11 2.05 11.79 2.56 0.81 0.0511 12001 45.13 12.75 9.40 119 2.33 0.12 3.16 0.11 2.0512 12476 45.05 12.73 9.34 120 2.23 0.17 4.16 0.12 1.72 11.92 1.91 0.88 0.07

BK-7

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 153 44.93 13.24 10.31 106 2.06 0.10 2.16 0.19 5.39 11.70 1.47 0.91 0.032 242 44.97 13.06 10.23 110 2.17 0.09 2.16 0.20 5.42 13.86 1.43 0.93 0.083 367 44.89 13.04 10.15 110 2.25 0.10 2.16 0.21 5.754 495 45.02 12.94 10.13 112 2.22 0.11 2.16 0.20 5.56 10.60 2.07 0.92 0.115 1589 44.85 12.92 9.94 115 2.29 0.07 2.16 0.20 5.676 1782 45.01 12.72 9.82 115 2.29 0.08 2.16 0.22 5.97 12.62 2.15 0.94 0.027 3010 44.88 12.74 9.76 115 2.41 0.07 2.16 0.21 5.83 12.40 2.00 0.94 0.058 7632 44.96 12.70 9.54 115 2.73 0.19 2.16 0.22 6.22 10.05 1.37 0.94 0.059 8816 45.04 12.71 9.53 116 2.75 0.19 2.16 0.22 6.2210 10533 44.85 12.77 9.39 117 2.70 0.15 2.16 0.22 6.11 10.38 0.69 0.91 0.0311 12004 45.05 12.75 9.40 119 2.80 0.15 2.16 0.23 6.3312 12478 44.94 12.73 9.34 120 2.63 0.14 2.16 0.22 5.97 10.38 1.76 0.89 0.02

FD-60

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 159 44.80 13.24 10.31 106 2.24 0.10 1.17 0.14 6.97 8.31 1.00 0.78 0.012 247 44.88 13.06 10.23 110 2.32 0.10 1.17 0.13 6.72 8.28 0.35 0.79 0.043 373 44.96 13.04 10.15 110 2.36 0.11 1.17 0.13 6.724 501 45.02 12.94 10.13 112 2.35 0.11 1.17 0.14 7.38 10.21 2.49 0.83 0.035 1595 44.87 12.92 9.94 115 2.50 0.08 1.17 0.14 6.926 1787 45.24 12.72 9.82 115 2.50 0.10 1.17 0.14 7.18 13.16 1.14 1.24 0.067 3015 44.89 12.74 9.76 115 2.34 0.11 1.17 0.14 7.08 10.04 1.14 1.06 0.138 7636 44.76 12.70 9.54 115 2.77 0.17 1.17 0.14 6.92 10.80 1.17 1.07 0.079 8822 44.96 12.71 9.53 116 3.07 0.21 1.17 0.13 6.7710 10538 44.92 12.77 9.39 117 2.75 0.15 1.17 0.13 6.46 13.76 3.58 1.35 0.3411 12010 44.91 12.75 9.40 119 2.83 0.16 1.17 0.13 6.4612 12483 45.08 12.73 9.34 120 2.77 0.15 1.17 0.12 6.31 10.16 0.22 1.13 0.06

EFDS-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 162 44.65 13.24 10.31 106 2.24 0.10 1.17 0.18 9.28 10.97 0.94 1.23 0.082 249 44.93 13.06 10.23 110 2.35 0.10 1.17 0.18 9.33 12.86 1.49 1.26 0.113 375 45.00 13.04 10.15 110 2.31 0.11 1.17 0.19 9.544 503 44.94 12.94 10.13 112 2.41 0.11 1.17 0.18 9.38 10.30 1.04 1.15 0.045 1597 44.81 12.92 9.94 115 2.46 0.07 1.17 0.20 10.006 1790 45.05 12.72 9.82 115 2.51 0.08 1.17 0.20 10.05 10.82 1.67 1.14 0.097 3017 44.78 12.74 9.76 115 2.65 0.08 1.17 0.19 9.49 11.27 1.49 1.23 0.168 7638 44.81 12.70 9.54 115 2.82 0.19 1.17 0.20 10.26 14.36 2.68 1.42 0.039 8824 45.01 12.71 9.53 116 2.98 0.18 1.17 0.20 10.3110 10540 45.03 12.77 9.39 117 2.84 0.18 1.17 0.20 10.21 14.22 2.14 1.56 0.1511 12012 45.08 12.75 9.40 119 2.83 0.15 1.17 0.20 10.1012 12485 44.99 12.73 9.34 120 2.68 0.14 1.17 0.20 10.15 13.42 2.85 1.42 0.20

FCD-1

Spot #Time (min)

Viscosity (cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 156 44.61 13.24 10.31 106.00 2.02 0.09 1.17 0.22 11.49 9.64 0.63 0.97 0.022 245 44.93 13.06 10.23 110.00 2.10 0.09 1.17 0.23 11.85 12.22 1.84 0.94 0.023 370 45.06 13.04 10.15 110.00 2.12 0.10 1.17 0.23 11.90 13.61 4.81 0.95 0.024 499 45.02 12.94 10.13 112.00 2.19 0.11 1.17 0.23 11.59 10.34 1.07 0.98 0.025 1591 44.65 12.92 9.94 115.00 2.23 0.07 1.17 0.23 11.79 10.18 1.22 0.96 0.026 1785 45.31 12.72 9.82 115.00 2.31 0.07 1.17 0.23 11.59 10.37 1.22 0.97 0.037 3012 44.93 12.74 9.76 115.00 2.27 0.06 1.17 0.23 12.00 12.30 2.77 0.99 0.048 7634 45.04 12.70 9.54 115.00 2.25 0.08 1.17 0.24 12.21 12.51 3.37 97.00 0.029 8819 45.01 12.71 9.53 116.00 2.47 0.10 1.17 0.25 12.67 12.76 2.95 0.96 0.0110 10535 44.95 12.77 9.39 117.00 2.34 0.10 1.17 0.23 12.00 13.68 1.49 1.03 0.0211 12008 45.08 12.75 9.40 119.00 2.63 0.17 1.17 0.24 12.26 57.06 21.45 8.33 2.7712 12481 45.16 12.73 9.34 120.00 2.36 0.06 1.17 0.22 11.23


220

LHG-8

Spot # CndViscosity

(cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 44.55 13.37 10.42 126 1.17 0.04 2.16 0.12 3.25 72.00 17.00 5.84 0.812 0.001 44.92 131 1.22 0.03 2.16 0.20 5.44 74.00 13.00 4.80 0.463 0.003 45.32 13.20 10.18 135 1.46 0.04 1.17 0.16 8.05 33.60 9.04 1.78 0.274 0.005 44.97 137 1.23 0.03 1.17 0.22 11.33 32.40 6.10 3.58 1.375 0.007 45.14 13.10 10.08 121 1.50 0.05 1.17 0.28 14.51 81.00 18.00 2.13 0.596 0.01 45.17 135 1.50 0.08 1.17 0.35 17.74 17.62 3.93 1.35 0.09

FS

Spot # CndViscosity

(cP)Moisture Content Fluid pH

Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 44.40 13.37 10.42 126 1.81 0.05 4.16 0.11 1.56 11.72 2.49 0.84 0.022 0.001 44.96 131 1.85 0.05 4.16 0.12 1.69 12.18 2.50 0.79 0.043 0.003 45.44 13.20 10.18 135 1.84 0.04 4.16 0.14 2.00 14.11 6.18 0.76 0.034 0.005 44.96 137 1.89 0.05 4.16 0.16 2.24 20.12 6.58 0.82 0.025 0.007 44.99 13.10 10.08 121 1.77 0.05 4.16 0.17 2.42 13.53 6.60 0.77 0.036 0.01 45.12 135 1.84 0.08 4.16 0.19 2.71 18.18 2.91 0.80 0.047 0.02 45.01 13.04 10.00 118 3.16 0.17 3.28 25.84 6.73 0.88 0.02

BK-7

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 44.50 13.37 10.42 126 1.59 0.03 3.16 0.16 3.04 15.58 6.46 0.85 0.022 0.001 44.83 131 1.78 0.04 3.16 0.19 3.59 10.20 1.79 0.86 0.013 0.003 44.93 13.20 10.18 135 1.89 0.05 2.16 0.16 4.44 15.64 3.00 0.83 0.014 0.005 45.11 137 1.69 0.05 2.16 0.19 5.19 10.44 3.04 0.82 0.025 0.007 45.31 13.10 10.08 121 1.78 0.05 1.17 0.11 5.44 15.40 2.90 0.85 0.026 0.01 45.10 135 1.82 0.08 1.17 0.12 6.15 8.28 0.59 0.82 0.037 0.02 45.09 13.04 10.00 118 1.88 0.06 1.17 0.14 7.18 9.42 1.12 0.86 0.03

FD-60

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 44.24 13.37 10.42 126 1.79 0.05 2.16 0.17 4.64 8.76 1.94 0.73 0.022 0.001 44.89 131 2.04 0.05 2.16 0.19 5.33 13.77 2.95 0.76 0.033 0.003 45.12 13.20 10.18 135 2.10 0.04 1.17 0.12 5.90 12.21 4.70 0.73 0.054 0.005 44.93 137 1.72 0.05 1.17 0.13 6.46 10.17 1.79 0.70 0.015 0.007 45.25 13.10 10.08 121 1.87 0.05 1.17 0.14 7.08 10.06 3.01 0.73 0.026 0.01 45.17 135 2.03 0.09 1.17 0.15 7.44 11.38 2.22 0.73 0.027 0.02 45.18 13.04 10.00 118 2.10 0.06 1.17 0.16 8.41 13.95 4.04 0.78 0.01

EFDS-1

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 45.22 13.37 10.42 126 1.88 0.04 2.16 0.19 5.19 11.64 1.47 1.30 0.062 0.001 44.97 131 2.13 0.05 1.17 0.12 6.00 14.50 5.77 1.60 0.583 0.003 45.00 13.20 10.18 135 2.12 0.05 1.17 0.15 7.49 13.94 1.17 1.57 0.174 0.005 44.90 137 1.82 0.04 1.17 0.17 8.51 11.96 0.82 1.34 0.075 0.007 45.19 13.10 10.08 121 1.99 0.06 1.17 0.18 9.33 13.84 1.36 1.61 0.266 0.01 45.16 135 2.17 0.08 1.17 0.21 10.77 14.43 3.28 1.38 0.187 0.02 45.28 13.04 10.00 118 2.19 0.06 1.17 0.25 12.97 13.24 1.23 1.53 0.14

FCD-1

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 44.83 13.37 10.42 126.00 1.82 0.05 1.17 0.15 7.79 14.00 2.96 0.89 0.012 0.001 45.02 131.00 1.91 0.05 1.17 0.16 8.21 34.20 9.59 4.53 1.673 0.003 45.28 13.20 10.18 135.00 2.00 0.19 1.17 0.19 9.69 15.70 4.89 1.30 0.554 0.005 44.69 137.00 1.74 0.04 1.17 0.21 10.67 11.00 3.80 0.92 0.055 0.007 45.16 13.10 10.08 121.00 1.84 0.06 1.17 0.23 11.59 10.24 1.02 0.98 0.026 0.01 45.28 135.00 1.99 0.08 1.17 0.25 12.62 10.47 1.54 0.98 0.027 0.02 45.08 13.04 10.00 118.00 2.05 0.06 1.17 0.30 15.18 10.32 1.00 1.02 0.04

Table C.10 - UK-Medium B MR fluid data.

221

LHG-8

Spot # CndViscosity

(cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 46.06 13.25 10.38 133 1.82 0.06 3.16 0.32 6.152 0.001 44.62 109 1.42 0.07 1.17 0.13 6.77 80.01 17.90 7.94 0.993 0.003 44.88 13.18 10.21 112 1.36 0.08 1.17 0.20 10.26 14.60 2.37 1.14 0.044 0.005 44.84 114 1.57 0.09 1.17 0.26 13.23 10.60 0.86 1.13 0.075 0.007 44.79 13.07 10.07 117 1.53 0.10 1.17 0.29 14.826 0.01 44.91 114 1.71 0.10 1.17 0.37 18.82 14.30 4.10 1.37 0.28

FS

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 45.84 13.25 10.38 133 2.64 0.06 5.22 0.15 1.762 0.001 44.70 109 1.95 0.09 4.16 0.13 1.88 10.50 2.20 0.78 0.023 0.003 44.93 13.18 10.21 112 1.93 0.11 4.16 0.14 2.06 12.10 3.00 0.77 0.014 0.005 44.92 114 1.94 0.11 4.16 0.17 2.41 11.58 2.01 0.79 0.035 0.007 44.65 13.07 10.07 117 2.10 0.12 4.16 0.18 2.586 0.01 4.78 114 2.11 0.12 4.16 0.19 2.77 9.16 1.59 0.78 0.027 0.02 44.79 12.98 9.99 113 1.93 0.13 4.16 0.24 3.43 11.03 2.18 0.81 0.02

BK-7

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 45.33 13.25 10.38 133 2.45 0.06 4.16 0.23 3.332 0.001 44.66 109 1.91 0.09 3.16 0.20 3.76 9.33 2.00 0.85 0.023 0.003 44.94 13.18 10.21 112 1.90 0.11 3.16 0.23 4.35 16.60 2.60 1.00 0.014 0.005 45.00 114 1.87 0.11 2.16 0.18 4.89 11.65 4.06 0.96 0.015 0.007 44.85 13.07 10.07 117 2.01 0.13 2.16 0.20 5.446 0.01 45.16 114 2.14 0.12 2.16 0.22 6.08 11.87 2.51 0.98 0.027 0.02 44.87 12.98 9.99 113 2.03 0.14 1.17 0.14 7.18 10.04 1.30 0.98 0.04

FD-60

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 46.26 13.25 10.38 133 2.87 0.10 2.16 0.18 5.062 0.001 45.18 109 2.04 0.10 1.17 0.10 5.13 9.64 1.88 0.73 0.023 0.003 44.99 13.18 10.21 112 2.14 0.12 1.17 0.12 6.26 21.34 3.23 0.74 0.014 0.005 44.99 114 2.20 0.13 1.17 0.13 6.51 8.45 1.83 0.71 0.015 0.007 45.06 13.07 10.07 117 2.20 0.15 1.17 0.14 6.976 0.01 44.88 114 2.85 0.14 1.17 0.15 7.74 10.35 3.78 0.75 0.027 0.02 45.08 12.98 9.99 113 2.18 0.16 1.17 0.17 8.67 9.69 3.12 0.77 0.01

EFDS-1

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 45.41 13.25 10.38 133 2.80 0.08 2.16 0.21 5.942 0.001 44.77 109 2.14 0.10 1.17 0.12 6.10 14.28 4.65 1.40 0.113 0.003 44.87 13.18 10.21 112 2.29 0.12 1.17 0.14 7.28 13.50 0.91 1.71 0.134 0.005 45.12 114 2.32 0.13 1.17 0.17 8.72 14.60 7.05 1.41 0.185 0.007 44.93 13.07 10.07 117 2.34 0.15 1.17 0.19 9.496 0.01 45.12 114 2.58 0.15 1.17 0.21 10.72 13.00 2.03 1.40 0.147 0.02 44.83 12.98 9.99 113 2.45 0.17 1.17 0.25 13.03 12.20 1.84 1.39 0.11

FCD-1

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 44.80 13.25 10.38 133.00 2.64 0.07 2.16 0.30 8.192 0.001 44.84 109.00 1.99 0.09 1.17 0.19 9.49 12.64 1.48 0.94 0.013 0.003 45.00 13.18 10.21 112.00 2.17 0.12 1.17 0.20 10.15 9.78 0.72 0.98 0.054 0.005 45.00 114.00 2.11 0.12 1.17 0.22 11.23 10.03 0.51 0.98 0.046 0.01 45.08 114.00 2.57 0.15 1.17 0.29 14.62 12.46 0.87 1.32 0.067 0.02 45.08 12.98 9.99 113.00 2.28 0.16 1.17 0.31 15.85 12.42 1.02 1.06 0.02

Table C.11 - UK-Medium C MR fluid data.

222

LHG-8

Spot # CndViscosity

(cP)


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 53.29 13.12 10.37 93 1.36 0.06 2.16 0.16 4.33 71.00 13.00 6.60 0.872 0.001 53.06 13.25 10.24 94 1.43 0.04 1.17 0.14 6.92 26.00 4.70 2.40 0.423 0.003 52.90 13.28 10.20 95 1.59 0.05 1.17 0.21 10.82 23.00 2.10 3.20 0.374 0.005 52.40 13.12 10.16 95 1.67 0.05 1.17 0.27 13.90 21.00 4.10 2.90 0.215 0.007 52.15 13.07 10.08 94 1.63 0.05 1.17 0.31 16.00 25.00 1.70 3.50 0.286 0.01 52.53 13.02 10.07 96 1.81 0.05 1.17 0.41 20.77 24.00 4.10 3.70 0.677 0.03 52.39 12.96 9.91 96 2.09 0.06 1.17 0.47 24.05 39.00 8.50 5.80 0.87

FS

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 53.08 13.12 10.37 93 2.27 0.06 5.22 0.16 1.86 11.50 3.31 0.89 0.012 0.001 53.17 13.25 10.24 94 2.29 0.07 5.22 0.17 2.00 11.41 2.64 0.97 0.023 0.003 52.76 13.28 10.20 95 2.45 0.08 4.16 0.16 2.24 11.66 2.30 0.94 0.044 0.005 52.27 13.12 10.16 95 2.61 0.10 4.16 0.18 2.60 12.66 1.42 0.93 0.015 0.007 52.43 13.07 10.08 94 2.62 0.11 4.16 0.19 2.78 9.32 1.45 0.89 0.046 0.01 52.41 13.02 10.07 96 2.72 0.09 3.16 0.17 3.27 13.38 2.75 0.92 0.027 0.03 52.61 12.96 9.91 96 2.58 0.09 3.16 0.23 4.27 13.72 4.77 0.90 0.018 0.07 52.68 12.92 9.95 97 2.58 0.08 2.16 0.19 5.33 11.49 3.27 0.91 0.049 0.1 52.98 13.05 9.92 99 2.48 0.07 2.16 0.19 5.39 11.11 1.60 0.94 0.02

BK-7

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 53.41 13.12 10.37 93 2.49 0.07 3.16 0.17 3.13 9.96 3.53 0.83 0.032 0.001 53.12 13.25 10.24 94 2.72 0.07 3.16 0.19 3.68 10.16 0.51 1.04 0.053 0.003 52.89 13.28 10.20 95 3.03 0.19 2.16 0.16 4.39 14.90 3.68 1.21 0.054 0.005 52.34 13.12 10.16 95 3.07 0.15 2.16 0.20 5.53 12.10 2.46 1.16 0.045 0.007 52.59 13.07 10.08 94 3.09 0.11 1.17 0.12 6.00 11.30 2.73 1.01 0.076 0.01 52.66 13.02 10.07 96 3.09 0.13 1.17 0.14 7.08 11.20 1.61 1.19 0.187 0.03 52.66 12.96 9.91 96 3.03 0.10 1.17 0.19 9.64 14.30 4.64 1.01 0.058 0.07 53.16 12.92 9.95 97 3.19 0.15 1.17 0.20 10.26 10.42 0.64 1.05 0.039 0.1 53.03 13.05 9.92 99 3.15 0.15 1.17 0.22 11.03 12.30 2.62 1.03 0.02

FD-60

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 53.15 13.12 10.37 93 2.99 0.11 2.16 0.18 4.89 8.19 0.79 0.77 0.012 0.001 53.01 13.25 10.24 94 3.36 0.15 2.16 0.19 5.33 10.71 1.93 1.10 0.113 0.003 52.81 13.28 10.20 95 3.20 0.19 1.17 0.11 5.49 11.22 2.15 1.09 0.094 0.005 52.52 13.12 10.16 95 3.49 0.20 1.17 0.11 5.74 9.96 0.45 1.02 0.075 0.007 53.05 13.07 10.08 94 3.35 0.16 1.17 0.12 6.36 10.76 0.80 1.23 0.096 0.01 52.78 13.02 10.07 96 3.63 0.14 1.17 0.14 6.97 10.04 0.89 1.09 0.097 0.03 53.18 12.96 9.91 96 3.50 0.19 1.17 0.18 9.33 10.89 0.77 1.24 0.078 0.07 51.86 12.92 9.95 97 3.69 0.15 1.17 0.24 12.51 12.68 0.98 1.32 0.089 0.1 52.90 13.05 9.92 99 3.58 0.16 1.17 0.27 13.74 14.28 0.89 1.36 0.07

EFDS-1

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 53.07 13.12 10.37 93 2.44 0.07 2.16 0.20 5.56 6.97 0.28 0.72 0.032 0.001 53.01 13.25 10.24 94 2.46 0.07 1.17 0.11 5.85 10.72 0.65 1.31 0.103 0.003 52.88 13.28 10.20 95 2.74 0.14 1.17 0.12 6.36 13.04 0.90 1.67 0.134 0.005 52.62 13.12 10.16 95 2.95 0.18 1.17 0.14 7.38 14.66 1.14 2.08 0.445 0.007 53.05 13.07 10.08 94 2.81 0.15 1.17 0.16 8.15 18.78 3.41 2.62 0.626 0.01 52.71 13.02 10.07 96 3.04 0.14 1.17 0.19 9.54 15.68 0.98 2.42 0.157 0.03 53.15 12.96 9.91 96 3.15 0.21 1.17 0.28 14.36 12.74 1.41 1.65 0.148 0.07 52.56 12.92 9.95 97 3.33 0.21 1.17 0.35 18.15 12.52 1.10 1.35 0.109 0.1 52.68 13.05 9.92 99 3.35 0.22 1.17 0.38 19.59 11.46 1.02 1.34 0.13

FCD-1

Spot # CndViscosity


Pump Speed (rpm)



Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 0 53.41 13.12 10.37 93.00 2.32 0.08 1.17 0.14 7.33 11.65 2.33 0.97 0.052 0.001 52.99 13.25 10.24 94.00 2.38 0.07 1.17 0.16 8.00 14.48 1.80 1.42 0.133 0.003 52.22 13.28 10.20 95.00 2.56 0.06 1.17 0.17 8.56 16.66 2.10 1.38 0.044 0.005 53.06 13.12 10.16 95.00 2.62 0.07 1.17 0.20 10.31 14.20 1.83 1.60 0.135 0.007 53.29 13.07 10.08 94.00 2.49 0.06 1.17 0.21 10.72 14.78 1.54 1.65 0.166 0.01 53.17 13.02 10.07 96.00 2.66 0.06 1.17 0.24 12.21 14.38 1.01 1.65 0.167 0.03 53.47 12.96 9.91 96.00 2.96 0.06 1.17 0.30 15.18

Table C.12 - NDP nanodiamond MR fluid data.

223

LHG-8

Spot # CCI

Viscosity (cP)

Moisture Content (wt. %)

Pump Speed (rpm)


Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 47.82 95.14 11.68 115 2.16 0.46 13.68 28.60 12.80 3.21 1.402 46.79 85.26 12.15 116 2.16 0.38 11.52 17.50 5.50 1.50 0.933 46.46 78.26 12.30 118 2.16 0.34 10.05 19.70 7.10 2.55 1.044 46.10 70.40 12.47 115 2.16 0.29 8.67 20.30 1.30 2.27 0.675 45.58 59.90 12.72 115 2.16 0.24 7.26 28.30 10.70 1.56 0.456 44.45 48.80 13.27 120 2.16 0.16 4.80 29.60 6.10 1.87 1.057 43.45 41.70 13.78 124 2.16 0.14 4.11 37.00 5.90 3.29 1.248 42.63 36.90 14.21 127 2.16 0.11 3.24 29.50 4.80 1.94 0.289 41.00 31.02 15.10 127 2.16 0.04 1.08 31.60 4.60 1.31 0.10

FS

Spot # CCI

Viscosity (cP)

Moisture Content

Pump Speed (rpm)


Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 47.82 95.14 11.68 115 5.22 0.28 3.31 12.30 2.80 0.80 0.062 46.79 85.26 12.15 116 5.22 0.23 2.80 6.80 1.40 0.72 0.023 46.46 78.26 12.30 118 5.22 0.21 2.54 7.50 1.40 0.68 0.024 46.10 70.40 12.47 115 5.22 0.19 2.33 6.50 0.50 0.64 0.025 45.58 59.90 12.72 115 5.22 0.18 2.18 7.20 1.90 0.58 0.016 44.45 48.80 13.27 120 5.22 0.15 1.85 4.50 0.10 0.49 0.017 43.45 41.70 13.78 124 5.22 0.13 1.54 5.30 1.30 0.50 0.018 42.63 36.90 14.21 127 5.22 0.11 1.26 5.60 0.80 0.48 0.029 41.00 31.02 15.10 127 5.22 0.07 0.83 4.30 0.20 0.45 0.01

BK-7

Spot # CCI

Viscosity (cP)

Moisture Content

Pump Speed (rpm)


Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 47.82 95.14 11.68 115 3.16 0.27 5.42 12.10 3.30 0.90 0.032 46.79 85.26 12.15 116 3.16 0.23 4.64 10.10 1.60 0.76 0.013 46.46 78.26 12.30 118 3.16 0.22 4.38 6.70 0.60 0.72 0.014 46.10 70.40 12.47 115 3.16 0.22 4.34 6.30 0.50 0.70 0.015 45.58 59.90 12.72 115 3.16 0.19 3.72 5.50 0.40 0.62 0.026 44.45 48.80 13.27 120 3.16 0.16 3.24 5.60 1.50 0.55 0.037 43.45 41.70 13.78 124 3.16 0.13 2.54 6.30 1.40 0.53 0.018 42.63 36.90 14.21 127 3.16 0.11 2.10 5.10 0.50 0.51 0.019 41.00 31.02 15.10 127 3.16 0.08 1.54 6.60 0.40 0.50 0.02

FCD-1

Spot # CCI

Viscosity (cP)

Moisture Content

Pump Speed (rpm)


Removal Rate

(um/min)Avg PV

(nm) st devAvg RMS

(nm) st dev1 47.82 95.14 11.68 115.00 2.16 0.36 10.68 6.80 1.00 0.72 0.042 46.79 85.26 12.15 116.00 2.16 0.36 10.80 6.10 0.40 0.69 0.033 46.46 78.26 12.30 118.00 2.16 0.35 10.53 7.80 2.10 0.66 0.024 46.10 70.40 12.47 115.00 2.16 0.33 9.75 8.20 3.20 0.65 0.035 45.58 59.90 12.72 115.00 2.16 0.31 9.24 14.80 4.30 0.62 0.026 44.45 48.80 13.27 120.00 2.16 0.26 7.86 6.00 1.30 0.60 0.037 43.45 41.70 13.78 124.00 2.16 0.23 6.90 8.90 6.10 0.65 0.038 42.63 36.90 14.21 127.00 2.16 0.19 5.79 14.30 5.80 0.61 0.019 41.00 31.02 15.10 127.00 2.16 0.15 4.56 8.70 2.30 0.63 0.02

Table C.13 – Abrasive free ramping CI data.

224

C.4 – Mechanical drawing for force sensor assembly Figure C.8 contains the mechanical drawing for the Kistler force sensor assembly. This assembly along with the force sensor was constructed and used to measure all of the drag force data in this thesis. Drag force is specifically discussed in Chapter 5. Figure C.8 – Mechanical drawing for the force sensor assembly.

Documents

Surface Interactions Between Nanodiamonds and Glass in