research on surface finish improvement using abrasive jet polishing, annealing, and polish air driving fluid jet polishing processes doctor of philosophy major mechanical engineering

國立台灣科技大學 機械工程系 博士學位論文 學號：D9803807 Research on Surface Finish Improvement Using Abrasive Jet Polishing, Annealing, and Air-Driving Fluid Jet Polishing Processes 研 究 生 ：Pham Huu Loc 指導教授：修芳仲 博士 中華民國 101 年 12 月 14 日 Research on Surface Finish Improvement Using Abrasive Jet Polishing, Annealing, and Air-Driving Fluid Jet Polishing Processes by Pham Huu Loc Department of Mechanical Engineering National Taiwan University of Science and Technology ABSTRACT Surface finish plays an important role in product quality due to its direct effects on product appearance Hence, improvement of the surface finish is an essential requirement in industrial products In an attempt to improve the surface finish of bulk metallic glass (BMG) material, some common methods have been used, such as milling, grinding, and lapping However, the BMG surface finish has not yet been significantly improved by using these methods Therefore, this thesis proposes sequential abrasive jet polishing (AJP) and annealing processes that can considerably improve the BMG surface finish In addition, this thesis also takes into account optimal parameters for both the AJP and annealing processes based on the Taguchi’s L18 and L8 orthogonal array experimental results, respectively The experimental results show that using the optimal AJP parameters, the surface roughness (Ra) of the ground BMG can be significantly improved from 0.675 to 0.016 µm After the AJP process, the surface roughness (Ra) of the polished BMG can be further improved from 5.7 to nm within an area of 5×5 µm by using the optimal annealing parameters i Furthermore, this thesis also proposes both air-driving fluid jet polishing (FJP) and AJP process that can improve the surface roughness of N-BK7 optical glass In addition, this thesis also investigates optimal parameters for air-driving FJP and AJP processes based on the Taguchi’s L18 orthogonal array experimental results The surface roughness (Ra) of ground N-BK7 optical glass can be improved from 0.350 to 0.032 µm by using the optimal air-driving FJP parameters and improved from 0.350 to 0.018 µm by using the optimal AJP parameters Finally, the determined optimal plane parameters for the airdriving FJP and the AJP are applied to the freeform surface finish of an N-BK7 spherical lens, and the surface roughness (Ra) of the spherical lens can be improved from 0.421 to 0.202 àm within an area of 283.6ì200 àm by using the optimal plane air-driving FJP parameters and improved from 0.421 to 0.232 àm within an area of 283.6ì200 µm by using the optimal plane AJP parameters Keywords: Abrasive jet polishing; Air-driving fluid jet polishing; Bulk metallic glass; Optical glass; Taguchi’s method; Annealing; Surface roughness ii ACKNOWLEDGEMENTS First, I would like to express my sincerest gratitude to my advisor, Fang-Jung Shiou, Professor, Mechanical Engineering Department Chair, Director of Opto-Mechatronics Technology Center in National Taiwan University of Science and Technology, who has supported me throughout my thesis with his patience and knowledge Besides my advisor, I would also like to thank the rest of my thesis committee for the valuable comments In addition, I would also like to thank Professor Jason S C Jang of National Central University and Professor Jinn P Chu of National Taiwan University of Science and Technology for providing samples of BMG material used in this study Furthermore, I am also grateful to Dr Arif and teacher Sun-Peng Lin, who guided me in operating the CNC machining center in the workshop In addition, I would also like to express my appreciation and thank Mr Son, Mr Quang, Mr Nguyen, Mr Andy, Dr Dat, Mr Duc, Mr Assefa and Mr Dang who helped me in all the time of research and writing of this thesis Their assistance and guidance have been of great value in this study I would also like to express my appreciation to the Instrument Technology Research Center (ITRC), National Applied Research Laboratories, Hsinchu Science Park for providing the financially support and thank Dr Wei-Yao Hsu, Mr Zong-Ru Yu and ITRC staffs for helping me during my studying time In my daily work, I am indebted to many of my other lab mates to provide me a happy and peaceful environment I would also like to thank the Library staffs who helped me in gathering a lot of information for this study I am also very appreciative of the NTUST for providing the financial support from Sept., 2009 to August, 2012 during iii Ph.D program I offer my regards to all of those who supported me in any respect during my studying time Finally, I would also like to thank my family for everything they have done for me Without their love, this thesis would not be finished iv TABLE OF CONTENTS ABSTRACT i ACKNOWLEDGEMENTS iii TABLE OF CONTENTS v LIST OF FIGURES ix LIST OF TABLES xiii CHAPTER INTRODUCTION 1.1 Background about the development and application of BMG 1.2 Literature review 1.2.1 The properties and machining ability of BMG 1.2.2 The properties and machining ability of optical glass 1.2.3 Abrasive jet polishing and air-driving fluid jet polishing process 1.2.4 Annealing process 1.3 Research motivation and thesis objectives 1.4 Outline of thesis CHAPTER BASIC PRINCIPLE 2.1 Production of bulk metallic glass 2.2 Grinding process 12 2.3 Lapping process 13 2.4 Abrasive jet polishing process 15 2.4.1 Polishing mechanisms 18 2.4.2 AJP process Parameters 20 2.5 Air-driving fluid jet polishing process 22 2.6 Annealing process 23 2.7 Taguchi method 24 v 2.7.1 Introduction 24 2.7.2 Control factors and noise factors 24 2.7.3 Orthogonal array 24 2.7.4 Analysis of variance (ANOVA) and S/N ratio analysis 26 2.7.5 Confirmation experiment 30 CHAPTER DEVELOPMENT OF AN AJP AND AN AIR-DRIVING FJP SYSTEM 31 3.1 Abrasive jet polishing system 31 3.2 Air-driving FJP polishing system 37 3.3 Velocity measurement system in AJP process 40 3.4 Velocity measurement system in air-driving FJP process 44 CHAPTER EXPERIMENTAL WORK 49 4.1 Material of the test specimens 50 4.2 Experimental setup 53 4.2.1 Experimental setup of the AJP process 53 4.2.2 Experimental setup of the air-driving FJP process 54 4.3 Configuration of Taguchi’s orthogonal array 57 4.3.1 AJP process on BMG material 57 4.3.2 Annealing process on BMG material 61 4.3.3 Air-driving FJP process on N-BK7 optical glass 63 4.3.4 AJP process on N-BK7 optical glass 68 CHAPTER EXPERIMENTAL RESULTS AND DISCUSSION 69 5.1 Experimental results for the AJP process on ground BMG 69 5.1.1 Combination of the optimal level for each factor 69 5.1.2 Confirmation experiments 70 vi 5.1.3 Analysis of variance 76 5.1.4 Influence of pressure on the surface roughness 77 5.1.5 Influence of impact angle on the surface roughness 77 5.1.6 Influence of particle size on the surface roughness 78 5.1.7 Influence of polishing time on the surface roughness 79 5.1.8 Influence of abrasive material type on the surface roughness 79 5.1.9 Influence of standoff distance on the surface roughness 80 5.2 Experimental results of the annealing process on polished BMG 81 5.2.1 Combination of the optimal level for each factor 81 5.2.2 Confirmation experiments 82 5.3 Experimental results for the air-driving FJP process on ground N-BK7 optical glass 83 5.3.1 Combination of the optimal level for each factor 83 5.3.2 Confirmation experiments 84 5.3.3 Analysis of variance 87 5.3.4 Influence of air pressure on the surface roughness 88 5.3.5 Influence of polishing time on the surface roughness 89 5.3.6 Influence of impact angle on the surface roughness 90 5.4 Experimental results for the AJP process on ground N-BK7 92 5.4.1 Combination of the optimal level for each factor 92 5.4.2 Confirmation experiments 94 5.4.3 Analysis of variance 97 5.4.4 Influence of abrasive concentration on the surface roughness 98 5.4.5 Influence of impact angle on the surface roughness 99 5.4.6 Influence of pressure on the surface roughness 99 vii CHAPTER APPLICATION AND COMPARISON OF AIR-DRIVING FJP AND AJP FOR THE SURFACE FINISH OF N-BK7 OPTICAL GLASS 101 6.1 Comparison of the air-driving FJP and the AJP for the surface finish of N-BK7 optical glass 101 6.2 Application of the air-driving FJP and AJP 102 6.2.1 Application of the air-driving FJP 102 6.2.1 Application of AJP 106 CHAPTER CONCLUSIONS AND FUTURE WORK 109 7.1 Conclusions 109 7.2 Future work 111 REFERENCES 112 APPENDIX A 116 APPENDIX B [30] 119 PUBLICATION 124 CURRICULUM VITAE 125 viii LIST OF FIGURES Fig 2.1.a Schematic illustration of the melt spinning process [4] 10 Fig 2.1.b Schematic diagram of the high-pressure die casting equipment designed and used by Inoue [4] 12 Fig 2.2.a Surface grinding machine [25] 13 Fig 2.2.b Principle of the grinding process [26] 13 Fig 2.3 Single Sided Lapping Set-Up [28] 14 Fig 2.4 Variables in AJP technology [12] 15 Fig 2.5 Schematic diagram of AJP technology 17 Fig 2.7 The process of material removal in brittle material mode [29] 19 Fig 2.8 The process of material removal in ductile material mode [30] 20 Fig 2.9 Schematic diagram of the air-driving FJP 23 Fig 2.10 Steps of Taguchi’s method to determinate the optimal parameters 25 Fig 2.11 The illustration of the control and noise factors 25 Fig 3.1 Schematic illustration of AJP system 31 Fig 3.2 The tank, stirring device in (a) rest state (b) in process 33 Fig 3.3 The inverter control pump 33 Fig 3.4 Photo of the developed components of the AJP tool head 34 Fig 3.5 The container with some of outlets 35 Fig 3.6 Slurry within the tank (a) without mixed hydraulic oil and (b) with mixed hydraulic oil after hours of the AJP process 36 Fig 3.7 Photo of the used 3-axis machining center, type MV-3A 36 Fig 3.8 Schematic illustration of the air-driving FJP system [20] 37 Fig 3.9 Photo of the tank 39 Fig 3.10 Photo of the air pressure regulator 39 ix Fig 4.6 Experimental setup for AJP 55 Machining center MV-3A Clamping shank MP 700 probe Tool holder NC controller NC codes for polishing path RS232 PC Pressure gauge NC Table Nozzle Specimen Atomizer Container Hose Stirrer Compressed air source Air pressure regulator Slurry Tank Fig 4.7 Experimental setup for air-driving FJP 56 4.3 Configuration of Taguchi’s orthogonal array 4.3.1 AJP process on BMG material In the AJP process, the effect of polishing parameters was determined by conducting matrix experiments using Taguchi’s orthogonal array [38] The fixed AJP parameters are summarized in Tables 4.4 Some parameters of the AJP process such as standoff distance, impact angle, abrasive material, abrasive concentration, pressure, and polishing time have significant effects on the surface roughness of BMG, and these parameters were selected as six control factors to investigate by Taguchi’s method and ANOVA Two levels of abrasive material parameter and three levels of the other parameters are shown in Table 4.5 Accordingly, the L18 orthogonal array was selected to perform the matrix experiments to identify the optimal AJP parameters Eighteen trials were then carried out separately based on the configured L18 orthogonal array The level value of parameters in Table 4.5 was estimated as following: Table 4.4 Fixed AJP factors in Taguchi design experiments Fixed factors Value Aperture of nozzle (material: Aluminum) 1.5 mm ANSI mesh of abrasive particles #2000 SiC (diameter: 6.7 µm) #4000 Al203 (diameter: 2.7 µm) Additives Water and machining oil Workpiece BMG Stirring velocity 100 rpm Revolution of the workpiece rpm Polishing path 2-axis path Size of each trial 3×5 mm 57 Table 4.5 The control factors and their levels in AJP experiments Level Control factors A Abrasive material SiC Al2O3 - B Abrasive concentration (abrasive: water) 1:5 1:8 1:10 C Impact angle (o) 30 40 50 D Standoff distance (mm) 10 15 E Pressure (kg/cm2) F Polishing time (min) 30 60 90  Abrasive material estimation Two types of abrasive material were determined based on the availability in market, and abrasive of SiC and Al2O3 were selected because these types were generally used in polishing industries  Abrasive concentration estimation According to Booji’s investigation [10], if the concentration is higher, particle-particle interaction will occur The particles will lose their kinetic energy due to these collisions, and the lost energy will not be used for material removal In order to avoid particle interactions in during a polishing process, the abrasive concentration was not allowed over 50wt% She suggested that range of abrasive concentration was about 5-20 wt% So in this thesis, the abrasive concentration of 1:5 (20 wt%), 1:8, 1:10 (10 wt%) were selected for the AJP process  Impact angle (α) estimation Total amount of removed material (W) is estimated by flowing equation [10]: W  l hw (4.1) 58 where l is length of the impact, h is depth of the impact, and w is width of the impact l, h, and w are calculated by flowing equation [10]: l = C1dvcos(α) (4.2) h = C2dvsin(α) (4.3) √ (4.4) The material removal (W) is a function of the impact angle (α), as shown in Fig 4.8 The highest material removal is obtainable at the impact angle of 45° The maximum length of a single impact occurs for particles approaching the surface parallel to the surface (α = 0), and the material removal (W) will be zero there, because the depth of the impact will be zero The maximum depth of one impact is reached for the impact angle of 90°, but the material removal (W) is zero again, because the length of the impact is zero Based on above analysis and Fig.4.8, range of the impact angle should was determinate around 30°-50°, and three levels of the impact angle of 30°, 40°, and 50° were selected for the AJP process 59 Fig 4.8 Relationship between the material removal and the impact angle [10]  Standoff distance estimation According to Booji’s investigation [10], range of standoff distance should be selected suitably for AJP system as long as the slurry does not diverge during a polishing process Based on preliminary tests for the AJP system, to avoid the slurry divergence, the range of standoff distance was determinate about 3-20 mm Hence, three levels of the standoff distance of mm, 10 mm, and 15 mm were selected for the AJP process  Pressure estimate Higher pressure results in higher surface roughness, due to the higher kinetic energy per impacting particle, which could cause deeper depressions in the surface [10] Range of the pressure is selected so that cutting force is sufficient for peak destruction but not sufficient for grooves formation and the smallest surface roughness is attained Based on research result of Booij [10] and Yan et al [15], range of the pressure of 2-5 kg/cm2 was applied for AJP process Accordingly, the three levels of the pressure of kg/cm2, kg/cm2, and kg/cm2 were selected for AJP process 60  Polishing time estimate Polishing time for each trial depends on some parameters such as size of each trial, feed rate, nozzle diameter and step over In this thesis, each trial size is 3×5 mm, as shown in Fig 4.9 Step over is 1.5mm using nozzle diameter of 1.5 Hence, the polishing time is calculated as follows: t (4.5) l 9mm   f f where l is length of tool path (mm), f is feed rate (mm/min), and t is polishing time (min) Based on preliminary tests for AJP system, range of feed rate of 0.1-0.3 mm/min was applied, the surface roughness of BMG could significantly improve By substitute f = 0.10.3 mm/min into Eq.4.5, the range of polishing time was determined about 30-90 Tool path Polishing area 3.0 5.0 Fig 4.9 Tool path for polishing area of each trial Hence, the three levels of the polishing time of 30 min, 60 min, and 90 were selected for the AJP process 4.3.2 Annealing process on BMG material After the AJP process, the improvement of the polished BMG surface finish was carried out by an annealing process in the super-cooled liquid region [22] In the annealing process, the effect of annealing parameters was determined by conducting matrix 61 experiments using Taguchi’s orthogonal array [38] Some parameters of the annealing process such as annealing temperature and annealing time have significant effects on the surface roughness of polished BMG, and these parameters were selected as two control factors to investigate by Taguchi’s method The control factors and their levels are summarized in Table 4.6 As shown in Table 4.6, the two control factors are assigned as follows: one 4-level factor and one 2-level factor Accordingly, the L8 orthogonal array was selected to perform the matrix experiments to identify the optimal annealing process parameters The level value of parameters in Table 4.6 was estimated as following: Table 4.6 The control factors and their levels in annealing experiments Level Control factors  A Annealing temperature ( o C) 450 470 490 510 B Annealing time (min) - - Annealing temperature estimation Based on the DSC analysis with heating rate of 40°C, the glass transition temperature and the crystallization temperature for BMG are 442.9°C and 529.4°C, respectively, as shown in Fig 4.3 The annealing temperature must be chosen to be above the glass transition temperature and below the crystallization temperature [22] Hence, range of the annealing temperature was determined about 450-525°C Accordingly, the fourth levels of the annealing time of 450°C, 470°C, 490°C, and 510°C were selected for the annealing process (Table 4.6)  Annealing time estimation Based on research result of Kumar et al [6] and [22], range of annealing time for BMG material was about 1-5 Therefore, the two levels of the annealing time for the BMG of min, were selected for the annealing process (Table 4.6) 62 4.3.3 Air-driving FJP process on N-BK7 optical glass The effect of polishing parameters was determined by conducting matrix experiments using Taguchi’s orthogonal array [38] Some parameters of the air-driving FJP process having significant effects on the surface roughness are abrasive material, abrasive concentration, impact angle, standoff distance, air pressure, and polishing time These parameters were investigated by Taguchi’s method and ANOVA The fixed polishing parameters are summarized in Tables 4.7 Two levels of abrasive material parameter and three levels of the other parameters are shown in Table 4.8 Accordingly, the L18 orthogonal array was selected to perform the matrix experiments to identify the optimal air-driving FJP parameters Eighteen experiments were then carried out separately based on the configured L18 orthogonal array The level value of parameters in Table 4.8 was estimated as following: Table 4.7 Fixed air-driving FJP factors in Taguchi design experiments Fixed factors Value Aperture of nozzle (material: Copper) mm Abrasive particle size CeO2 (diameter: µm) Al2O3 (diameter: 2.7 µm) Additives Water Workpiece N-BK7 optical glass Stirring velocity 100 rpm Revolution of the workpiece rpm Polishing path 2-axis path Size of each trial 6×6 mm 63 Table 4.8 The control factors and their levels in air-driving FJP experiments Level Control factors A Abrasive material CeO2 Al2O3 - B Abrasive concentration (wt%) 10 15 20 C Impact angle (o) 40 50 60 D Standoff distance (mm) 12 E Air pressure (kg/cm2) F Polishing time (min) 15 30 45  Abrasive material estimation Two types of abrasive material were determined based on the availability in market, and abrasive of CeO2 and Al2O3 was selected to polish N-BK7 optical glass because these types were generally used in polishing industries  Abrasive concentration estimation Abrasive concentration affects material removal of polishing process The higher concentration doesn’t produce higher material removal rate because abrasive particles collides with each other easily in high concentration slurry [10], and the collision phenomenon reduces the kinetic energy of each abrasive particle Hence, less kinetic energy can be used to remove material The material removal ability will be reduced with over slurry concentration in the polishing process Based on experimental results as shown in Fig 4.10, Yu et al [21] have used range of the abrasive concentration from 5-20 wt% for N-BK7 polishing using air-driving FJP process So in this thesis, the abrasive concentration of 10 wt%, 15 wt%, and 20 wt% were selected for the air-driving FJP process 64  Impact angle estimation According to investigation of Yu et al [21], range of impact angle from 30°-90° have been applied using air-driving FJP process However, that study showed that the smallest surface roughness can be obtained with range of impact angle from 30°-60° Hence, in this thesis, three levels of the impact angle of 40°, 50°, and 60° were selected for the Depth of material removal (mm) air-driving FJP process Processing time (sec) Fig 4.10 Investigation of slurry concentration on N-BK7 [21]  Standoff distance estimation Based on research results of Yu et al [21], range of impact angle from 3-12 mm have been applied using air-driving FJP process That study revealed that the material removal decreases as stand-off distance increases, as shown in Fig 4.11 Hence, in this thesis, three levels of the standoff distance of mm, mm, and 12 mm were selected for the air-driving FJP process 65 Depth of material removal (mm) Processing time (sec) Fig 4.11 Investigation of standoff distance on N-BK7 [21]  Air pressure estimation In order to determine range of pressure, three preliminary tests of polished N-BK7 were carried out in polishing time of 20 sec The removal shape was measured by a Zygo interferometer (VerifireTM AT+), and the peak to valley (P-V) of removal shape represents the material removal in this study The result indicated that the material removal increases when pressure increases, as shown in Fig 4.12 However, when air pressure was provided about kg/cm2, flow rate of slurry was too small (0.09 l/min) to remove material, as shown in Table 3.5 When range of the air pressure was about 2-5 kg/cm2, the flow rate of slurry was enough to remove material Hence, the three levels of pressure of kg/cm2, kg/cm2, kg/cm2 was selected for the air-driving FJP process 66 Polishing time: 20 sec Sample: Polished N-BK7 Abrasive: CeO2 Concentration: 10 wt% Standoff distance: mm Nozzle diameter: mm Pressure (kg/cm2) Fig 4.12 The relationship between material removal and pressure  Polishing time estimation Polishing time for each trial depends on parameters such as size of each trial, feed rate, and step over In this thesis, each trial size for the air-driving FJP is 6×6 mm, as shown in Fig 4.13 and step over is 1.5 mm using nozzle diameter of 1.5 mm, the polishing time is calculated as follows: t l 12mm   f f (4.6) where l is length of tool path, f is feed rate, and t is polishing time Based on research result of Yu et al [21], range of feed rate of 0.25-1 mm/min was applied on N-BK7 optical glass By substitute f = 0.25-0.8 mm/min into Eq.4.6, the range of polishing time was determined about 12-48 Hence, the three levels of the 67 polishing time of 15 min, 30 min, and 45 were selected for the air-driving FJP process Polishing area 6.0 1.5 6.0 Tool path 6.0 Fig 4.13 Tool path for polishing area of each trial 4.3.4 AJP process on N-BK7 optical glass In order to compare polishing performance between the AJP and the air-driving on NBK7 optical glass, hence, the L18 orthogonal array was selected to perform the matrix experiments to identify the optimal AJP parameters Eighteen experiments were then carried out separately based on the configured L18 orthogonal array The fixed parameters and control factors of the AJP was similar to the fixed parameters and control factors of the air-driving FJP, as shown in Table 4.7 and 4.8, respectively 68 CHAPTER EXPERIMENTAL RESULTS AND DISCUSSION This chapter describes experimental results when the AJP, the annealing, and the air-driving FJP processes applied to the surface roughness of BMG, N-BK7 optical glass In addition, the effect of dominant polishing parameters on the surface roughness is also discussed using ANOVA Finally, optimal parameters for the AJP, the annealing, and the air-driving FJP processes are determined based on S/N ratio value In this chapter, the experimental results include the following items: the experimental results for the AJP process on ground BMG, the annealing process on polished BMG, the air-driving FJP process on ground N-BK7 optical glass, and the AJP process on ground N-BK7 optical glass 5.1 Experimental results for the AJP process on ground BMG 5.1.1 Combination of the optimal level for each factor Signal-to-noise (S/N) ratio is used as an objective function, and the surface roughness value of polished surface should be smaller than that of original surface Thus, the AJP process is an example of the smaller-the-better type problem The optimization target of the smaller-the-better type problem is to maximize η defined by Eq 2.2 Levels for each factor that have maximum the S/N ratio value are selected as optimal levels According to experimental results for the AJP process, Table 5.1 indicates the measured polished surface roughness value (Ra) and the S/N ratio value of each L18 orthogonal array that was calculated using Eq 2.2 Table 5.2 reveals the average S/N ratio for each level of the six factors The average S/N ratio for each level of the six factors is shown graphically in Fig 5.1 Based on the S/N ratio value, the optimal level for each 69