Wettability aspects during silicon wafer cleaning in aqueous and organic systems.
Item Type text; Dissertation-Reproduction (electronic)
Authors Park, Jin-Goo.
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Download date 28/09/2021 21:19:39
Link to Item http://hdl.handle.net/10150/186200 INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed through, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
U·M·I University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. M148106·1346 USA 313/761·4700 800:521·0600
Order Number 9322700
Wettability aspects during silicon wafer cleaning in aqueous and organic systems
Park, Jin-Goo, Ph.D.
The University of Arizona, 1993
Copyright ©1993 by Park, Jin-Goo. All rights reserved.
U·M·I 300 N. Zeeb Rd. Ann Arbor, MI 48106
WETIABILITY ASPECfS DURING SIUCON WAFER CLEANING
IN AQUEOUS AND ORGANIC SYSTEMS
by
lin-Goo Park
Copyright CD lin-Goo Park 1993
A Dissertation Submitted to the Faculty of the
DEPARTMENT OF MATERIALS SCIENCE AND ENGINEERING
In Partial Fulfillment of the Requirements For the Degree of
DOcrOR OF PHILOSOPHY
In the Graduate College
THE UNIVERSITY OF ARIZONA
1993 2
THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE
As members of the Final Examination Committee, we certify that we have
read the dissertation prepared bY ___J~l~·n~-~G~o~o~P~a~r~k~ ______
entitled Wettability Aspects during Silicon Wafer Cleaning
in Aqueous and Organic Systems
and recommend that it be accepted as fulfilling the dissertation
requirement for Doctor of Philosophy
Date ; ,//;-r-II) D. B. Birnie III Date .[ Rra'i 2- /2. s/ 2.3 > Date I S. Raghavan
Final approval and acceptance of this dissertation is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.
Dissertation Director Date l S. Raghavan 3
STATEMENT BY AUTHOR
This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.
Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgement of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the copyright holder. 4
ACKNOWLEDGEMENTS
The author would like to express his great appreciation and respect to Professor Srini Raghavan, his advisor, for his guidance, encouragement and friendship during the course of his studies. He is also grateful to Professors 1. H. O'Hanlon, J. H. Enemark, D. P. Birnie III, and J. B. Hiskey for being on his committee and for their patience and encouragement. He would especially like to express his gratitude to Professor J. O'Hanlon, Director of the Center for Microcontamination Control, for his support during the course of this research.
He would like to thank Dr. D. Cooper of IBM-Thomas Watson Research Center, Dr. J. Carrejo and Mr. D. Tolliver of Motolora, Mesa, AZ, Mr. C. Schneider and Mr. J. Beauchamp of Intel, and J. Crozier of National Semiconductor for their assistance.
The author is also grateful to the Center for Microcontamination Control and the Sematech Center of Excellence at the University of Arizona for their financial support of this research.
The author wishes to thank Dr. Raghavan's research group, all his friends at the Arizona Materials Laboratory, especially Mr. P. K Sung, Mr. J. S. Jeon, Mr. D. Dunn, Mr. D. Jan, Ms. C. Jung, and Ms. K Sole for their help, friendship and encouragement. He would also like to thank Ms. P. Broyles for her help in preparing this dissertation.
Special thanks go to Reverend and Mrs. K H. Kim and the members of the Korean Presbyterian Church of Tucson for their love and prayers. They helped him mature and become closer to God with unforgettable memories during his study in Tucson.
The author would like to express his sincere gratitude to his family in Korea. He has to mention his adorable son, Jaeyeon, who gives him joy at the end of his study. Finally, he wants to thank his wife, Sukkyung, for her sacrifice and love. He dedicates this work to her. 5
TABLE OF CONTENTS
liST OF ILLUSTRATIONS ...... 8 liST OF TABLES ...... 17
ABSTRACf ...... 18
1. INTRODUCTION ...... 20
1.1. Introduction...... 20 1.2. Research Objectives ...... 22
2. BACKGROUND ...... 23
2.1. Wafer Cleaning ...... 23 2.1.1. Wet Cleaning Processes ...... 24 2.1.1.1. Alkaline Solutions in Semiconductor Processing. . . .. 31 2.1.1.2. Oxide Etch ...... 36 2.1.2. Drying of Silicon Wafers after Wet Processing ...... 42 2.1.2.1. Characterization of IPA ...... 44 2.2. Contact Angles ...... 46 2.2.1. Static Angles - Sessile Drop or Adhering Gas Bubble Method . 50 2.2.2. Dynamic Angles - Wilhelmy Plate Method ...... 53 2.2.3. Origin of Hysteresis ...... 59
3. EXPERIMENTAL MATERIALS AND PROCEDURES ...... 65
3.1. Materials ...... 65 3.1.1. Wafers ...... 65 3.1.2. Chemicals ...... 65 3.2. Experimental Procedures ...... 67 3.2.1. Preparation of Wafer Samples ...... 67 3.2.2. Wettability Measurements ...... 68 3.2.2.1. Dynamic Contact Angle Measurements ...... 68 3.2.2.2. Static Contact Angle Measurements ...... 71 3.2.3. Particle Deposition ...... 73 6
TABLE OF CONTENTS (continued)
3.3. Characterization Methods ...... 75 3.3.1. Scanning Electron Microscopy (SEM) ...... 75 3.3.2. Auger Electron Spectroscopy (AES) ...... 75 3.3.3. X-ray Photoelectron Spectroscopy (XPS, ESCA) ...... 75 3.3.4. Ellipsometry ...... 76 3.3.5. pH and Redox Potential Measurements and Optical Microscopy 76 3.3.6. Scanning Tunnelling (STM)/Atomic Force Microscopy (AFM) 77 3.3.7. Inductively Coupled Plasma Spectroscopy (ICP) ...... 78
4. RESULTS AND DISCUSSION...... 79
4.1. Characterization of Samples...... 79 4.2. Characterization of Chemicals ...... 79 4.2.1. Alkaline Cleaning Solutions ...... 79 4.2.2. Buffered Oxide Etch (BOE) ...... 91 4.3. Dynamic Contact Angle Measurements ...... 94 4.3.1. The Change of Contact Angles during the Storage of Samples 94 4.3.2. The Effect of Stage Speed of the Dynamic Contact Angle Analyzer on Wettability ...... 97 4.4. Dynamic Wetting of Wafers in Alkaline Solutions...... 100 4.4.1. Wettability of Wafers in Choline Containing Solutions ..... 100 4.4.1.1. Wettability of BOE-Etched p(100) Wafers ...... 100 4.4.1.2. Wettability of p(loo) Wafers with Thermally Grown Oxide ...... 104 4.4.1.3. Effects of Wafer Types and Orientations on Wettability 106 4.4.1.4. Effect of Hydrogen Peroxide on Wettability in Choline Solutions ...... 106 4.4.1.5. Effe~t of Surfactant on the Wettability of Silicon in Choline Solutions ...... 109 4.4.1.6. Effect of Temperature on Wettability in Choline Solutions ...... 115 4.4.2. Oxidation in Choline Solutions ...... 117 4.4.3. Wettability of Wafers in a Commercially Available Choline- Based Solution (SummaClean) ...... • ...... 133 4.4.4. Wettability of Wafers in TMAH (Tetramethyl Ammonium Hydroxide) Solutions ...... •...... •.••••...... 136 4.4.5. Wettability of Wafers in Ammonium Hydroxide Solutions ... 138 7
TABLE OF CONTENTS (continued)
4.5. Scanning Tunneling Microscopy (STM)/ Atomic Force Microscopy (AFM) Studies ...... 146 4.5.1. Effect of Addition of HzOz and Surfactant Roughness ...... 149 4.5.2. Effect of Addition of HzOz and Surfactant to TMAH Solutions ...... 153 4.6. Wettability of Wafers in IPA and Its Significance in Particle Removal ...... 154 4.6.1. Characterization of IPA Solutions ...... 154 4.6.2. Wettability of Wafers in IPA/Water Solutions ...... 158 4.6.3. Particle Removal on Wafers Using IPA-Based Solutions ..... 167 4.6.3.1. Particle Deposition ...... 167 4.6.3.2. Particle Removal on Hydrophilic Wafers ...... 171 4.6.3.3. Particle Removal from Hydrophobic Wafers ...... 182 4.6.4. Theoretical Analysis of Adhesion and Removal of Particles .. 188 4.6.4.1. Hydrophilic Particles on Silicon and SiOz-Coated Wafers ...... 192 4.6.4.2. Hydrophobic Particles on Silicon and SiOz-Coated Wafers ...... 195 4.6.4.3. Modification of Leenaars Equations...... 199
5. SUMMARY AND CONCLUSIONS ...... 203
5.1. Alkaline Solutions ...... 203 5.2. IPA/Water Solutions ...... 205
REFERENCES ...... 207 8
LIST OF ILLUSTRATIONS
2.1. Plot of calculated concentrations of HF and HFi for solutions of 15 M total fluoride as a function of pH ...... 38
2.2. Plot of calculated concentrations of possible species in the HF-Si02- H20 system as a function of pH ...... 40
2.3. Etch rates of thermal oxide at 25°C in various BOE compositions (from Ref. [2.44]) ...... 41
2.4. Equilibrium distribution diagram between liquid and vapor of IPA at their boiling temperatures. Replotted from Ref. [2.54] ...... 47
2.5. Change of activity coefficients of IP A and water as a function of IPA content in water. Replotted from Ref. [2.54] ...... 48
2.6. Schematic diagram of techniques measuring static angles: (a) sessile drops on a solid surface and (b) captive bubbles in a liquid underneath a solid surface ...... 51
2.7. Schematic diagram of a Wilhelmy plate apparatus ...... 54
2.8. Schematic diagram illustrating a Wilhelmy plate method (from Ref. [2.62]) ...... 55
2.9. Typical Wilhelmy plate hysteresis loops for (a) bare and (b) oxide- coated silicon wafers in DI water ...... 57
2.10. Change of contact angles as a function of roughness factor, 'r' (from Ref. [2.65]) ...... 61
2.11. Change of contact angles and contact angle hysteresis as a function of surface coverage of low-energy material (from Ref. [2.65]). Curves 1, 2, 3 and 4 indicate decreasing drop energies. Upper and lower sets of curves are advancing and receding angles, respectively 63
3.1. Schematic diagram of sample preparation procedures...... 69
3.2. Schematic diagram of a DCA contact angle analyzer ...... 70 9
LIST OF ILLUSTRATIONS (continued)
3.3. Schematic diagram of a static contact angle analyzer ...... 72
3.4. Schematic diagram of the experimental set-up for (a) electrophoretic deposition and (b) aerosol spray deposition ...... 74
4.1. SEM micrographs of (a) polished wafer (independent of wafer type and orientation), (b) p(100) unpolished side, (c) n(100) unpolished side, and (d) n(111) unpolished side of silicon wafers ...... 80
4.2. (a) pH, redox potential (vs. SHE) and (b) the surface tension (at 25°C and 35°C) of choline solutions ...... 82
4.3. Surface excess of choline solutions at air/solution interface as a function of concentration ...... 83
4.4. Changes of (a) pH and (b) redox potential (vs. SHE) as a function of H20 2 addition in choline solutions ...... 85
4.5. (a) pH and redox potential (vs. SHE) and (b) surface tension of TMAH solutions ...... 86
4.6. (a) pH and surface tension and (b) redox potentials of NH40H solutions with and without the addition of H20 2 • • • • • • • • • • • • • • • •• 87
4.7. Etch rates of p(100) silicon wafers in alkaline solutions at 25°C: (a) choline and TMAH and (b) 1:5 NH40H:H20 solutions as a function of their concentrations ...... 88
4.7. (continued). (c) Etch rates of p(100) silicon wafer in choline, TMAH, and 1:5 NH40H:H20 solutions as a function of pH at 25°C...... 89
4.8. Effects of surfactant addition in BOE and different vendors on contact angles in DI water and choline solution when silicon wafers were etched with BOEs ...... 93
4.9. ~ha!lge of (a) c?ntact ~ngles and (b) oxide thickness during storage m aIr as a functIon of tIme ...... 95
4.10. Effect of stage speed on contact angles in DI water ...... 98 10
LIST OF ILLUSTRATIONS (continued)
4.11. Changes of hysteresis loops at different stage speeds: (a) 264 p.m/sec, (b) 170 p.m/sec, (c) 64 p.m/sec and (d) 20 p.m/sec . .. 99
4.12. Advancing (OJ and receding (OJ angles on p(l00) bare wafers in choline solutions: (a) first cycle and (b) second cycle ...... 102
4.13. Contact angles in OI water on a sample immersed in choline (10000 ppm) for 10 min and on a sample subjected to a two-cycle test in choline ...... 103
4.14. Advancing (OA) and receding (OJ angles on oxide-coated p(l00) wafers in: (a) choline solutions and (b) in OJ water after conditioning in choline followed by a OJ water rinse for 5 min ...... 105
4.15. Wettability of BOE-etched p(100) and n(l00) wafers in: (a) the first cycle and (b) the second cycle in choline solutions ...... 107
4.16. Effect of wafer orientation on wettability in choline solutions: (a) the first cycle and (b) the second cycle ...... 108
4.17. Advancing (OJ and receding (OR) angles on BOE-etched OSP p(l00) wafers as a function of hydrogen peroxide concentration; choline content in solution: (a) 400 ppm and (b) 5000 ppm ...... 110
4.18. Contact angles in 1 wt% H20 2 solutions as a function of choline concentrations ...... 111
4.19. Change of surface tension as a function of surfactant (NCW 601A) concentration ...... 113
4.20. (a) Effect of surfactant addition (100 ppm) to choline solutions on wafer wettability and (b) contact angles in OJ water on samples subjected to a two-cycle immersion in a choline solution with surfactant (100 ppm) followed by a OJ water rinse for 5 min ...... 114
4.21. Contact angles on BOE-etched p(l00) wafers in: (a) choline solutions at 35°C and (b) OJ water after conditioning in choline at 35°C followed by a OJ water rinse for 5 min ...... 116 11
LIST OF ILLUSTRATIONS (continued)
4.22. Experimental matrix used to study the interaction of choline with silicon samples ...... 118
4.23. Changes of oxide thickness and wettability as a function of air exposure time after DI water rinse ...... 119
4.24. Effect of air exposure after immersion in choline solution on oxide thickness and contact angles in DI water (conditions: 10000 ppm choline, 10 min immersion, p(100) wafer) ...... 120
4.25. Changes of contact angle and oxide film thickness as a function of immersion time in 10000 ppm choline solutions ...... 122
4.26. Si (2p) XP spectra of: (a) BOE-etched p(lOO) Si, (b) Si treated in 1 wt% choline and rinsed immediately in DI water, and (c) Si "b" exposed to air for 10 min and then DI water rinsed ...... 123
4.27. Schematic diagram of a silicon wafer sample used to characterize the effect of repeated exposure to choline solution and air ...... 125
4.28. (a) Thickness changes of different areas as a function of stage speed and (b) wettability of areas 1 and 2 in DI water after cycling in choline solution ...... 126
4.29. XPS Si (2p) spectra obtained on area 2 of Si samples cycled at different stage speeds: (a) 264 p.m/sec, (b) 64 p.m/sec and (c) 20 p.m/sec ...... 128
4.30. Effect of cycling on: (a) oxide thickness and (b) contact angles of samples in DI water after treatment in 10000 ppm choline solution and DI water rinse .. ~ ...... 129
4.31. Change of oxide thicknesses at different choline temperatures as a function of choline concentrations ...... ;...... 130
4.32. Change of oxide thicknesses treated in choline/peroxide solutions followed by DI water rinse ...... • ...... 132 12
LIST OF ILLUSTRATIONS (continued)
4.33. (a) pH, redox potential (vs. SHE) and (b) surface tension of silicon in SummaClean as a function of choline concentration ...... 134
4.34. Contact angles of diluted SummaClean solutions on p(loo) silicon as a function of choline concentration ...... 135
4.35. (a) Advancing (8~ and receding (80 contact angles on bare SSP p(lOO) wafers in TMAH solutions and (b) the contact angles in D1 water on samples subjected to a two-cycle immersion in a TMAH solution followed by a 01 water rinse for 5 min ...... 137
4.36. Optical micrographs of p(lOO) silicon etched for 10 min in: (a) 400 ppm, (b) 1000 ppm, (c) 10000 ppm TMAH solutions, and (d) 10000 ppm choline solution. The bar scale is 100 I'm ...... 139
4.37. Surface profiles of samples treated in: (a) 10000 ppm, (b) 1000 ppm, (c) 400 ppm TMAH solutions, and (d) as-received polished silicon surface ...... 140
4.38. Wettability of silicon wafers in aqueous ammonium hydroxide solutions ...... 141
4.39. Optical micrographs of wafers subjected to 2 cycles in: (a) 0.1%, (b) 1%, (c) 10%, and (d) 50% NH40H:H20 solutions, and D1 rinsed for 5 min. The bar scale is 100 I'm ...... 143
4.40. Wettability of silicon wafers in: (a) alkaline H20 2 solution as a function of NH40H content and (b) D1 water after cycling in ammonia-peroxide solution twice followed by D1 water rinsing ...... 144
4.41. (a) STM micrograph on polished and (b) AFM micrograph on unpolished p(l00) wafer after 30 sec BOE etch ...... 147
4.42. STM micrographs of: (a) 0.1:1:5 and (b) 1:1:5 SC1-treated samples (500 X 500 run scanning area) ...... 148 13
LIST OF ILLUSTRATIONS (continued)
4.43. AFM micrographs of 10000 ppm choline-treated silicon samples followed by: (a) 10 min air exposure and DI water rinse and (b) DI water rinse only (no air exposure) ...... 150
4.44. AFM micrographs of 5000 ppm choline-treated silicon samples for 10 min: (a) without the addition of H20 2 and surfactants and (b) with the addition of 200 ppm H20 2 ••••••••••••••••••••••• 151
4.44. (continued) (c) with the addition of 1 wt% (10000 ppm) H20 2 and (d) with the addition of 200 ppm non-ionic surfactant ...... 152
4.45. AFM micrographs of silicon samples treated for 10 min in 600 ppm TMAH solutions: (a) in the absence of H20 2 and (b) with the addition of 200 ppm H20 2 •• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •• 155
4.45. (continued). (c) with the addition of 1 wt% (10000 ppm) H20 2 and (d) with the addition of 200 ppm non-ionic surfactant ...... 156
4.46. Surface tension of IPA-water solutions as a function of solution IPA content ...... 157
4.47. Surface excess of IPA at the air/IPA-water solution interface ...... 159
4.48. Contact angles on hydrophobic and hydrophilic wafers as a function of IPA content in water ...... 160
4.49. Nature of IPA/water interface formed by slowly placing IPA on the top of water in a glass optical cell as a function of time: (a) 0 hr and (b) 20 hrs later ...... 163
4.50. Surface tension of top and bottom layers formed by slowly placing IPA solutions of different concentrations on the top of water ...... 165
4.51. Contact angles on hydrophobic wafers as a sample is successively contacted with solutions of increasing IPA content .....•...... 166
4.52. Photomicrographs (at 5X) of droplets on hydrophobic wafers at different IPA contents: (a) less than 25 vol% IPA and more than 70 vol% IPA, (b) 25 vol% IPA, and (c) 50 vol% IPA ...... 168 14
LIST OF ILLUSTRATIONS (continued)
Figure
4.53. Optical micrographs of particles deposited by: (a) electrophoretic method, (b) aerosol spray method on hydrophilic wafers, and (c) aerosol spray method on hydrophobic wafers ...... 169
4.54. Change of the oxide thickness and contact angle as a function of deposl ·t· Ion t·lme ...... 170
4.55. Particle deposition methods applied to contaminate wafers ...... 172
4.56. Optical micrographs of samples rinsed once in: (a) DI water, (b) 2.4% IPA, (c) 9% IPA, and (d) 25% IPA. The bar scales are 100 I'm ...... •...... 173
4.56. (continued). (e) 50% IP A, and (f) 87 and 100% IP A The bar scales are 100 I'm ...... •...... 174
4.57. Appearance of samples successively rinsed in solutions increasing IPA content: (a) from DI water to pure IPA and (b) 50% IPA to pure IPA ...... , ...... 175
4.58. Optical micrographs of samples subjected to seven successive rinses in: (a) DI water, (b) 9% IPA, and (c) 25% IPA The bar scales are 100 I'm ...... ••...... •...... •... 177
4.58. (continued). (d) 50% IPA and (e) 87 and 100% IPA The bar scales are 100 I'm ...... 178
4.59. Schematic diagram illustrating the mechanism for enhanced particle removal from solutions of low IPA content: (a) particle on a substrate before contacting IP A-water solution, (b) particles at the air/IPA-water interface, (c) particles removed by surface pressure difference at the air/solution interface, and (d) substrate after rinsing ...... 179
4.60. Removal of silicon particles deposited on hydrophilic wafers: (a) as-deposited, (b) in DI water, and (c) in 9% IPA solution. The bar scales are 100 I'm ...... •...•.•...... •...... 181 15
LIST OF ILLUSTRATIONS (continued)
4.61. The removal of PSL particles which were electrophoretically deposited in IPA for 10 min: (a) as-deposited, (b) rinsed in 01 water, (c) rinsed in 9% IPA solution, and (d) in 100% IPA The bar scale is 100 ",m ...... •...... 183
4.62. Particle removal on silicon samples with different hydrophobicity: (a) bare, (b) anodized for 10 sec, (c) anodized for 2 min, and (d) anodized for 5 min and 10 min. The bar scale is 100 ",m ...... 185
4.63. Removal of silicon particles deposited on hydrophobic wafers: (a) as-deposited, (b) rinsed in DI water, and (c) rinsed in 9% IPA solution. The bar scales are 100 ",m ...... 187
4.64. Schematic diagram showing a spherical particle adhering to a substrate during the passage of a phase boundary: (a) ex < 90° and (b) CIt > 90° ...... 190
4.65. Comparison of IF.y •max l/2'l1"R'Ylg and IFal/2'l1"R'Ylg as a function of (J at different CIt values: (a) glass particles in 01 water and (b) glass particles in 9% IPA The curved lines are IF.y•max l/2'l1"R'Ylg at different CIt values and the straight lines are \Fal/2'l1"R'Ylg ...... 193
4.65. (continued). (c) alumina particles in 01 water, and (d) alumina particles in 9% IP A solutions ...... 194
4.66. Comparison of IF'Y.maxl/2'l1"R'Ylg and IFal/2'l1"R'Ylg as a function of (J at different CIt values: (a) PSL particles in 01 water and (b) PSL particles in 9% IPA The curved lines are IF'Y.maxl/2'1"R'Ylg at different CIt values and the straight lines are IFal/2'1"R'Ylg ...... 196
4.67. Comparison of IF'Y.maxI/2'1"R'Ylg and IFal/2'1"R'Ylg as a function of (J at different CIt values: (a) silicon particles in 01 water and (b) silicon particles in 9% IP A. The curved lines are IF'Y.maxl/2'1"R'Ylg at different CIt values and the straight lines are IFa I/2'l1"R'Ylg ...... 198 16
LIST OF ILLUSTRATIONS (continued)
B~ ~
4.68. Comparison of IFtotad/2TR'YIg and IFal/2TR'YIg as a function of T*: (a) PSL particles on hydrophilic wafers and silicon particles and (b) on hydropbiic and in 9% IFA solutions. The curved lines are IFtotad/2TR'YIg at different -,;* values and the straight lines are IFal/2TR'YIg .•..•....••••...•••.•..•••••••••.•••.•..... 201
4.68. (continued). (c) silicon particles on hydrophilic wafers and silicon particles and (b) on hydrophiic wafers and (d) alumina particles on hydrophilic wafers in 9% IPA solutions. The curved lines are IFtotai l/2-,;R'YIg at different -,;* values and the straight lines are IFal/2TR'YIg ••.••••.•...... ••••....•.••••.•...•...... 202 17
LIST OF TABLES
2.1. Wet Chemicals with Their Removals and Side Effects (from Ref. [2.11]) ...... 25
2.2. Physical Properties of Alcohols ...... 45
2.3. Advantages and Disadvantages of Contact Angle Methods (from Ref. [2.60]) ...... 49
2.4. Sources of Contact Angle Hysteresis (From Ref. [2.62]) ...... 64
3.1. Characteristics of Wafers Used in the Experiments...... 65
4.1. Etch Rates of Silicon Wafers in Alkaline Solutions at Higher Temperatures...... 91
4.2. Surface Tensions and Dynamic Contact Angles of Various Concentrations of BOB (400:1 and 200:1) on Bare p(1oo) Wafers from Two Different Vendors (30°C) ...... 92
4.3. STM/AFM Data on Roughness of Silicon Surfaces Treated with Various Alkaline Solutions ...... 146
4.4. Roughness of Samples Treated in 5000 ppm Choline for 10 min With and Without the Addition of H20 2 and Non-ionic Surfactant ... 153
4.5. Roughness of Samples Treated in 600 ppm TMAH for 10 min With and Without the Addition of H20 2 and Non-ionic Surfactant ... 154
4.6. Particle Removal in Hydrophobic and Hydrophilic Wafers ...... 185
4.7. Hamaker Constants for Various Materials Taken from Ref. [4.23] ...... 189
4.8. Calculated Values of A132 and IFal/2'1'R'Ylg Where 'Ylg of Water = 72.6 dynes/ern' nlg of 9% IPA Solution = 43 dynes/cm •...... 192
4.9 Contact Angles on Particles and Substrates in DI Water and 9% IP A Solutions ...... 197 18
ABSTRACI'
Alkaline solutions based on ammonium hydroxide and quaternary ammonium hydroxides such as choline (hydroxyethyl trimethyl ammonium hydroxide) and TMAH (tetramethyl ammonium hydroxide) are used widely in the wet processing of silicon wafers for the control of ionic and particulate impurities.
The Wilhelmy plate technique was used in characterizing the ability of alkaline solutions to alter the wettability of wafers. Choline improved the water wettability of wafers, and at concentrations greater than 1000 ppm, rendered the wafers very hydrophilic. Ellipsometric and XPS analyses showed that the exposure of choline treated surfaces to air resulted in the oxidation of Si to Si02• The increase of wettability of wafers in TMAH solutions was due to the roughness introduced by the high etch rate of TMAH solutions. Ammonia solutions without the addition of H20 2 did not increase wettability. The addition of H20 2 and a non-ionic surfactant to alkaline solutions significantly increased the wettability of wafers, decreased the etch rate, and resulted in smoother surfaces.
The use of isopropyl alcohol (IPA) in the drying of wafers has been considered by the semiconductor industry. The addition of IP A to water resulted in a decrease in surface tension at the solution/vapor interface. The surface excess of IPA molecules at the solution/air interface was calculated to have a maximum value of 8.5 X 10-10 moles/em2 at a solution composition of 25% IPA and 75% 19 water. IPA solutions with less than 25% IPA were very effective in removing PSL particles on hydrophilic wafers.
Hydrophilic particles such as alumina and glass were difficult to remove from wafers in DI water and IP A solutions, however, hydrophobic particles such
as silicon were slightly removable in DI water and IPA solutions. The wettability
of particles (8) and substrate (ex) in solutions played important roles in removing
particles on substrates. Leenaars' equation for the calculation of magnitude of
surface tension force which balance adhesion force, FA' did not seem to hold for
solutions containing less than 25% IPA. Modification to this equation by adding a
surface pressure term, -r*, was considered in explaining the experimental results. 20
CHAPTER 1
INTRODUCI'ION
1.1. Introduction
With the increase in chip capacity and decrease in feature size, microcontamination control in semiconductor processing is becoming increasingly important. The DRAM (dynamic random access memory) chip area has increased about 40 to 50 percent every three years. At the same time, the defect density has increased significantly due to the decrease of minimum feature size and complexity of the device. The device yield decreases dramatically as chip
area and defect density increase [1.1]. Product yield can be improved by more
stringent contamination control throughout the manufacturing process because
more than 50% of the yield losses are attributed to micro contamination [1.2].
Organic films, metal ions and inorganic/organic particles are major sources
of wafer contamination during the silicon wafer fabrication sequence. Many wet
cleaning techniques reportedly remove contaminants from silicon wafers [1.2,1.3].
The commonly used RCA standard cleaning processes are based on acidic or
alkaline hydrogen peroxide solutions [1.4]. The alkaline RCA cleaning solution,
referred to as SC1, is based on ammonium hydroxide, hydrogen peroxide and
water at a ratio of 1:1:5. It is primarily designed to remove organic contaminants
from the wafer surface. 2t
The acidic RCA cleaning solution, referred to as SC2, is based on hydrochloric acid, hydrogen peroxide and water at the same ratio as SCt. It is known to be very effective in removing metal ions. As optional steps, piranha
(H2S04-H20 2 mixture) etch ants and HF etchants were combined with SCt and SC2 cleaning solutions. Piranha etch removed heavy organic contamination such as photoresists while HF removed native and/or chemically grown oxides on silicon.
Quaternary ammonium hydroxides (OAR) such as choline
([(CH3)3N+CH2CH20H]OH+) and tetramethyl ammonium hydroxide (TMAH) were tried in place of SCl for wafer cleaning [1.5-1.7]. Choline (pKt, = 0.1) and
TMAH (pKt, = 0) are much stronger bases than NH40H (pKt, = 4.7) and are typically free from alkali metals. Cleaning sequences involving choline showed better removal of particles [1.6] and heavy metals, and result in a lower number
of oxidation-induced stacking faults (OISF) after HF treatment [1.8].
Rinsing after wet cleaning is done with high-purity resistivity DI water.
Megasonic rinsing [1.9], centrifugal spray rinsing [1.10] and closed system rinsing
[1.11] were also tried. Wafer drying after rinsing must be done by physical
removal of the water rather than evaporation. Spin drying has been the most
widely used method. Drying of wet wafers with isopropyl alcohol (IPA) results in
a decrease in particulate contamination [1.11,1.12]. High-purity IPA vapor drying
[1.11] is effective in drying and particle removal. 22
Much of the above-mentioned work with cleaning solutions is concerned with the evaluation of the chemicals in terms of their performance such as contaminant removal. These reagents have the capacity to modify the wettability of a highly hydrophobic silicon surface produced by etching in HF or HF/NH 4F
(BOE) solutions. The hydrophilic/hydrophobic nature of a wafer surface after a cleaning step affects particulate contamination. Menon and Donovan [1.13] deposited glass beads, silicon dust and polystyrene latex (PSL) particles on both p and n-type wafers with contact angles ranging from 12 to 64 o. For all three particle types and both wafer conductivity types, low contact angle surfaces retained fewer particles after cleaning than the high contact angle surfaces.
1.2. Research Objectives
The hydrophobic silicon surface produced by HF or BOE (buffered oxide etch) increases contamination on wafers [1.14,1.15,1.16]. Fundamental information on the effect of cleaning solutions on the wettability of wafers is virtually nil. The objectives of this research were to: (1) investigate surface modification (wettability and microroughness) of silicon wafers in DI water, alkaline cleaning solutions, and IPA/water solutions and (2) relate the wettability to the cleaning effects on wafers in IPA/water solutions. 23
CHAPTER 2 BACKGROUND
2.1. Wafer Cleaning
Cleanliness during the processing of silicon wafers is crucial to the fabrication of devices and circuits with uniform electrical characteristics and high reliability. From a device fabrication viewpoint, a "clean surface" is a surface that does not contain a significant amount of harmful contaminants. It is well known that the yield of VLSI (very large-scale integration) fabrication lines can be
significantly improved by applying more stringent requirements for the purity of
chemicals, gases and deionized water used in processing [2.1,2.2]. While the
surfaces of wafers have traditionally been cleaned with wet chemicals, alternative
methods of obtaining impurity free surfaces are being investigated [2.1].
Surface cleaning procedures play a fundamental role in determining the
characteristics of semiconductor devices such as surface state type and density,
surface conductivity [2.3], electrical properties of oxides [2.4], and the kinetics of
oxide growth [2.5-2.7]. Atalla et a1. [2.3] studied the stabilization of silicon
surfaces by thermally grown oxides. They contended that from a device
technology viewpoint, truly atomically clean surfaces were not desirable. A device
with an atomically clean surface would not be operative since the silicon dangling
bonds would act as acceptor states and the resulting surface would be strongly
p-type, independent of bulk doping. From a practical standpoint, they argued that 24 it is more appropriate to refer to "surface cleaning" as "surface doping." The etches and rinses used in device preparation generally provide surface structures that are complex and unknown, compared to a truly atomically clean UHV cleaned surface. The surface ion state density and distributions may be affected by both the pre oxidation treatment and the concentration of certain impurities in silicon. They concluded that a thermally-grown oxide provided the most stable surface and found that impurities left on the surface were partially maintained through the oxidation process, affecting the interface characteristics and the device characteristics.
The fundamental role of any silicon cleaning procedure is to remove particles, organics, transition metals and alkali ions from the surface. For this purpose, the wafer cleaning methods practiced in the semiconductor industry can be divided into wet [2.1] and dry [2.8,2.9] processes even though they may be used in combination in certain situations.
2.1.1. Wet Cleaning Processes
In wet chemical cleaning, surface cleaning is commonly achieved by either immersing or spraying the wafer with liquid chemicals. Wet chemical cleaning using acidic and alkaline solutions has been practiced for many years [2.10]. This
cleaning procedure involves the immersion of wafers in various acidic or alkaline
solutions mixed with hydrogen peroxide and water, followed by a DI-water rinse. 25
The most commonly used wet chemicals are listed in Table 2.1 along with their contaminant removal characteristics.
Table 2.1. Wet Chemicals with Their Removals and Side Effects (from Ref. [2.11])
Contaminants Wet Chemicals to be Removed Side Effects Solvents Gross Organics Organic Contaminants Piranha etch Gross Organics Native Oxide H2S04:H20 2 (grease, wax, photoresist) SC1 Organic Residue Particles AI NH4OH:H20 2:H2O Group m. lIB metals Native Oxide Roughness SC2 Metallics Native Oxide HCl:H20 2:H2O (AI, Fe, Mg)
HF orBOE Si02 Reactive Surface (NH4F:HF)
Acid and acid-based hydrogen peroxide solutions are widely used in cleaning Si wafer surfaces. The sulfuric acid-hydrogen peroxide mixture, without additional water, is used to strip photoresist and remove trace organic impurities
[2.12]. Such a mixture is called "piranha" because its aggressive nature is designed
to leave a wafer surface free of organic materials. Solvents are also used for
removing gross organics on wafer surfaces. The hydrochloric acid-hydrogen
peroxide-water solution is designed to remove alkali ions and cations such as AI3+
and Fe3+ by forming chlorine complexes that are highly soluble. HF or BOE 26
(buffered oxide etch, a mixture of NH4F:HF) has also been used for oxide etch in semiconductor industries.
The basic ammonium hydroxide-hydrogen peroxide-water solution is
capable of removing organic contaminants by the solvating properties of the
ammonia and the oxidizing abilities of the peroxide. This solution can also
remove Cu, Ag, Ni, Co and Cd by complexation. However, Mg, AI and Fe form
insoluble hydroxide complexes in ammonium solutions and precipitate onto the
surface.
In a series of papers written in 1970, Kern [2.10,2.13,2.14] introduced a
cleaning procedure based on mixtures of hydrogen peroxide with acids or bases.
This cleaning procedure is now commonly known as the RCA clean. Kern
measured the adsorption of reagent components [2.13] and trace impurities [2.14]
by silicon and silica using a radioactive tracer method. He found that adsorption
of chloride ions from HCI solutions onto silicon was very weak and that the ions
could be easily removed by rinsing with deionized water at room temperature.
On the adsorption of fluoride ions from HF solutions, he reported that the major
portion of fluorine contamination resulted from the physisorption of fluoride ions,
but appreciable quantities of chemisorbed fluoride remained after 15 min of
rinsing. Kern noted that desorbed fluorine ions were effectively removed by
rinsing in acetone for a few minutes.
Deposition of trace impurity elements such as Au, Cu, Fe, Cr, Zn, Sb, Mn
and Mo was studied by radiochemical tracer techniques [2.14]. Severe metallic 27 contamination on silicon surfaces was observed after HF etching. This was attributed to the large difference between the oxidation potentials of the silicon and the metal ions in solution, leading to large quantities of electrochemically deposited Cu, Au and Cr. Adsorption of metal ions onto silica was very low. He found that complexing by acidic H20 2 is the most effective method of removing metal contaminants, including Au, Cu and Cr.
Based on adsorption and desorption studies, Kern et al. [2.10] introduced a new cleaning sequence using solutions based on hydrogen peroxide. Surface
contaminants were classified into several broad categories: molecular, ionic and
atomic. Typical molecular contaminants were natural and synthetic oils, waxes
and resins. The source of these organics were grease from fingers, grease from
films deposited during storage and residues from mechanical polishing of wafers.
These contaminants cause polarization and ionic drift on surface sensitive MOS
structures. Since these organics are not soluble in water, the presence of these
compounds would result in a hydrophobic surface and interfere with the effective
removal of any absorbed ionic or metallic impurities. Therefore, these
compounds must be removed as the first step in cleaning.
After removing gross organic residues in a hot trichloroethylene bath, Kern
[2.10] used a 1:2:5 solution of NH40H:H20 Z:HzO (SCI: standard clean 1) to remove the remaining organic contaminants by the solvating action of the
ammonium hydroxide as well as complexing some group I and group n metals
such as Cu, Ag, Ni and Co. The second process to remove the remaining heavy 28
metals was carried out in a 1:2:7 HCI:H20 2:H20 (SC2: standard clean 2)
solution. Over 30 different desorption treatments were done involving
combinations of complexing, chelating and oxidizing agents for removing copper
from silicon. The acidic-hydrogen peroxide was the most effective. Water and/or
hydrogen peroxide without acid were not effective. It is important to note that
the solutions did not attack either silicon or silicon oxide as long as sufficient
hydrogen peroxide was present. If the hydrogen peroxide became depleted or if
there was a large amount of fluoride ion contamination present, etching did occur.
It has become a common practice to modify the original RCA clean by
including a HF etch between the ammonium hydroxide-hydrogen peroxide and
hydrochloric acid-hydrogen peroxide steps, as suggested by Amick [2.15]. Amick
suggested a cleaning sequence with the following steps:
(1) removal of organics,
(2) removal of oxide film on silicon surface, and
(3) removal of metallic or ionic impurities.
He claimed that the sequence should be followed in this order since the presence
of an oxide or organic film on the surface could interfere with the removal of
metals and ions, while the presence of an organic film could interfere with the
removal of the underlying oxide.
Kern [2.16], in a later paper, suggested the addition of a brief etch in dilute
HF between SC1 and SC2. The HF etching in the cleaning process enhances the
carbon contamination, is a source of metal contamination, and introduces stacking 29 faults on (100) surfaces following oxidation [2.10,2.17]. Henderson [2.17] used high-energy electron diffraction (HEED) and auger electron spectroscopy (AES) to evaluate silicon cleaning by sequential immersions in a basic and acidic peroxide solution with HF etching. The HF etch as a final step in a cleaning procedure caused the formation of relatively large amounts of surface silicon carbide when a silicon wafer was heated to 800°C in ultrahigh vacuum. Also, a final HF etched wafer suffered extensive surface roughness with lloo°C heating in vacuum. However, SC1-HF-SC2 sequence provided a thin, nearly carbon-free, oxide layer that acted as a passivating and protective layer and did not introduce the formation of surface silicon carbide and roughness on the wafer when heated in vacuum.
Schwettmann et aI. [2.18] studied the dependence of oxidation kinetics on surface preparation for the thermally grown oxide. Wafers cleaned with a sulfuric-peroxide solution produced a considerably thicker thermal oxide when compared to the "no-clean" wafers and the ammonium hydroxide-hydrogen peroxide clean wafers.
Gould and Irene [2.6] varied the standard RCA clean to investigate the effect of surface clean procedures for thermally grown oxides thicker than 1200 A.
They found results similar to those of Schwettmann et al. [2.18]. The slowest oxidation rate was observed for the ammonium hydroxide-hydrogen peroxide cleaned sample. The HCI:H20 2 produced nearly the same retardation in 30
oxidation rate as the treatment in NH40H:H20 2 solution. An HF etched sample
showed the most rapid oxidation.
de Larios [2.11] reversed the cleaning sequence for the NH40H:H. 20 2 and
HCI:H20 2 solutions (reverse RCA) and added an HF:H20 2 step between the
original two steps of the RCA process. He found the retardation in oxidation
kinetics on surfaces cleaned with NH40H:H20 2 final step compared to the other
clean. This was attributed to the fact that less than 5 x 1013 AI atorns/cm2 were
deposited on the wafer surface during the wet cleaning process.
Ohsawa et al. [2.19] developed a method for preoxidation cleaning of
silicon trenches by slightly etching the surfaces using a HN03-HF-H20 solution
with an extremely low HF concentration. They claimed this method removed
near-surface damage associated with reactive ion etching (RIE) and reduced the
concentration of heavy metal surface impurities to one-tenth that of RCA
cleaning. Wafers were first cleaned by the modified RCA method (SCI-HF:H20-
SC2-SC1), then by a slight etch (SE) cleaning. SE cleaning (etch depth -30 nm)
reduced the surface concentration of heavy metals by a factor of ten and Ca, Mg
by a third when compared with RCA cleaning. They attributed this to the
removal of shallow contaminated layers on silicon wafers.
Peters and Deckert [2.20] investigated the ability of organic solvents,
plasmas and oxidizing solutions to strip positive and negative photoresists. The
organic solvents were adequate for stripping positive resists but left a significant 31 photoresists but left residue. A two-step procedure was recommended to clean surfaces: (1) removing the bulk of the photoresist film in organic solvents, and
(2) treating in hot (80°C) NH40H:H20 2 solutions. The SCl solution alone could not strip photoresists.
2.1.1.1. Alkaline Solutions in Semiconductor Processing
Alkaline solutions have traditionally been used to etch and clean silicon wafers. Cleaning wafers in straight (no hydrogen peroxide) alkaline solutions results in the etching of wafers, and the magnitude of the etching depends on the concentration of solutions. Strong alkaline chemicals such as NH40H [2.21], KOH, NaOH [2.22], ethyldiamine [2.23], hydrazine [2.24], and quaternary ammonium hydroxides [2.25-2.27] (e.g. tetraethyl ammonium hydroxide
«CH3)4NOH) and choline «CH3)3N(CH2CH20H)OH» are used for etching and cleaning processes. Inorganic strong bases such as NaOH and KOH are seldom used in the semiconductor industry because alkali elements are highly mobile
contaminants, thus harmful to electrical properties of devices.
During the etching of silicon with KOH and NaOH, the main reaction
species was found to be OH- [2.28,2.29]. The cation does not participate in the
etching and complexation of etched species. For most alkaline etching solutions,
a complexing agent such as iso-2-propyl alcohol and pyrocatechol (C6H4(OHh) is
add~d [2.23,2.24]. The complexation of the etched species in solutions is a very 32 important chemical reaction and prevents contaminant ions from redepositing on
the substrate.
Ammonium hydroxide is well known to etch and complex metals such as
Cu, Ag, Ni, Co, and Cd [2.30]. Ammonium hydroxide is a weak base and is
characterized by the following dissociation equilibrium:
NH40H t; NH4 + + OH-, pKt, = 4.75 at 25°C (2.1)
In the cleaning process, NH40H is seldom used alone, and additives such as H20 2
and surfactants are added to NH40H to control the etch rate and increase the
wettability of silicon. Without the presence of H20 2, ammonia solution etches
silicon and causes a rough surface [2.31]. Kern [2.16] studied the effect of H20 2
on the etch rate of silicon and silica in ammonia solutions. He found the etch
rate to be a function of the H20 2 concentration and a 1:1:5 NH40H:H20 2:H20
(SC1) solution at 85°C etched silica and silicon at 0.9 A/min and 0.5 A/min,
respectively. Without adding H20 2 to the SC1 solution, the etch rate of silicon
increased tremendously to -3.5 I'm/min at the same temperature but the etch
rate of silica was not affected.
Hydrogen peroxide is a powerful oxidizing and acidic agent. Dilute
solutions at 50°C are most stable at a pH near 4.5 to 5.0 and are least stable at
high pH's where they decompose to water and oxygen,
H20 2 ... H20 + 1/2 O2, (2.2)
Decomposition is catalyzed by traces of most heavy metals. Hydrogen peroxide is
a weak acid (PKa = 11.6). Thermally activated transfer of a proton from a 33 hydrogen peroxide molecule to a water molecule occurs to a small degree even at
25° [2.32]:
(2.3)
In SCl solutions, hydrogen peroxide gradually decomposes and this results in a loss of cleaning effectiveness. During cleaning, the temperature of the SCl solution is maintained in the range of 75-80°C because of the rapid decomposition of HzOz at higher temperatures and an increase in the volatilization of NH3 from the NH40H solution. The half-life of the SCI solution at 88-90°C was reported as approximately 5 min compared to 50 h at 23°C [2.16]. In acidic peroxide solutions, the decomposition of HzOz and volatilization of HCI occur at a much slower rate than from the alkaline peroxide solutions. There is no danger of silicon etching under any conditions.
Meerakker et al. [2.33] studied the etching n-type silicon of the (100) orientation in NH3/HzOz solutions at 7CfC. The addition of HzOz of less than 3 3 x 10- M to NH40H solutions enhanced the etch rate initially. When the HzOz concentration exceeded the value of 3 x 10-3 M (-100 ppm) at 70°C, the silicon
surface was passivated and etching stopped. They also investigated the influence
of the pH of NH3 solutions adjusted with NaOH. The pH showed a moderate
effect on the etch rate; increasing the OH- concentration by a factor of 30
enhanced the etch rate by a factor of 10. The effect of the NH3 concentration
was much stronger. The half-life of HzOz in SCI solutions operated at 80°C was calculated to be 9.3 min. 34
In SCI solutions, higher concentrations of ammonium hydroxide leave more particles and hazes as defects on silicon surfaces. Mishima et al. [2.34] reported that an NH40H-H20 2 solution of the standard ratio (1:1:5,
NH40H:H20 2:H20) increased the haze formation and decreased the particle
(polystyrene and silica latex) removal efficiency compared with lower NH40H content solutions. Proportions of NH40H ranging from 0.1 to 0.5 showed better cleaning efficiencies and formed much fewer hazes on the silicon than the standard ratio.
Quaternary ammonium hydroxides have recently been applied to clean wafers with RCA cleaning solutions. They showed better performance than any combinations of RCA cleaning solutions in removing particles [2.25,2.26], heavy metals and reducing oxidation induced stacking faults (OISF) [2.35] after treatment.
One such QAH as choline, a strong base, is a biologically, physiologically and pharmacologically important substance [2.36]. Choline can be produced by the reaction of trimethylamine with ethylene oxide.
Choline is free of alkali metal contaminants such as sodium or heavy metals, like
copper, whose inclusion should be avoided with utmost care. It is stable in dilute
solutions, but in concentrated solutions tends to decompose at 100°C, giving
ethylene glycol, polyethylene glycol and trimethylamine. Choline generally has the 35 capacity to remove organic matter such as oils and inorganic matter such as alkali metals and aluminum that may be deposited on a semiconductor substrate [2.27].
A choline-based cleaning solution, Summa-Clean, has gone through extensive testing in the semiconductor industry; it contains a proprietary surfactant and a
small amount of methyl alcohol.
Muraoka et al. [2.26,2.27] developed several techniques to clean silicon
wafers using choline. They reported that dilute aqueous solutions of choline with
selected nonionic surfactants can remove heavy metals from the silicon surface
and prevent replating of these metals from solutions. Even though choline etches
silicon, the etch rate can be controlled by adding nonionic surfactants. The
addition of hydrogen peroxide can be used to minimize the etch rate.
Hariri and Hockett [2.35] studied the influence of different wafer cleaning
methods on high temperature oxidation of silicon. They evaluated various
cleaning methods (nine different cleaning processes by combining RCA chemicals
and choline) by characterizing oxidation-induced stacking fault (OISF) formation,
oxide thickness and dielectric properties. Choline cleaning resulted in fewer
OISFs, the lowest levels of heavy metals, thicker oxides and no leakage in the
oxides.
Electronic-grade TMAH was introduced [2.37,2.38] in the 80's to clean
silicon wafers and develop resist films in the process for manufacturing
semiconductor devices. Dilute aqueous TMAH solutions, with and without
selected surfactants, can be used for removing organics, heavy metals and particles 36 from silicon wafers. At a temperature of near 70°C, TMAH shows better cleaning capability and etches silicon at rates ranging from 180 A/min at 0.077%
(1:30 dilution of 2.38% TMAH with surfactant). TMAH can also be added to a
SC1 mixture (1:1:3:20 TMAH:NH40H:H20 2:H20) for particle and metal ion removal.
2.1.1.2. Oxide Etch
HF in various dilutions in water and buffered with ammonium fluoride is the standard silicon dioxide wet etchant in the semiconductor industry. Etch rate of silicon in HF is negligible. Hot alkaline bases such as NaOH or KOH etch oxide at a useful rate, but they attack organic masking materials and also etch silicon.
HF-H20 System
In dilute water solutions, HF partially dissociates and generates the bifluoride ion HFi [2.39]:
3 KI = 1.3 X 10- (2.5)
K2 = 9.615 (2.6)
In the semiconductor industry, 49% or 50% HF is mixed with water ranging in weight ratios of HF:H20 from 5:1 to 400:1. The etch rate of silicon dioxide depends on the HF and HF2- concentrations but not on F in relatively dilute fluoride solutions. Figure 2.1 shows a plot of the calculated fractional 37 concentrations of HF and HF£ for solutions of 15 M total fluorine as a function
of pH. In the region of pH > 7 all fluoride exists as uncombined fluoride ions.
As the pH is lowered, the [HF£] increases and reaches a maximum in the region
of pH 3. Hydrofluoric acid becomes a significant species below a pH of 4 and
increases monotonically as the pH is further lowered.
Hydrofluoric acid reacts with Si02 to form hexafluosilicic acid and other
silicofluoride complexes according to the following equations [2.40-2.43]:
6 6 Si02 + 4 H+ + 6 F ... SiFl + 2 H2O K3 = 102 . (2.7)
2 H+ + SiFl ... H2SiF6 ~ = 35.82 (2.8)
2 Si02 + 4 H+ + 5 SiFl ... 6 SiFs- + 2 H 2O K3 = 10-1.05 (2.9)
7 15 2 Si02 + 8 H+ + 4 SiFl'" 6 SiF4 + 2 H 2O. K3 = 10- . (2.10)
The results of a calculation to identify the most predominant silicon fluoride
complexes in a 15 M HF solution saturated with Si02 are shown in Figure 2.2.
The most dominant species over the pH range 0 to 8 is shown as SiFl.
NH4F-HF-H20 System
Buffered oxide etch (BOE) is widely used as a chemical for processes such
as etching, patterning and cleaning of silicon wafer surfaces. It is a mixture of
40% NH4F and 50% HF, ranging in weight ratios of NH4F:HF from 5:1 to 400:1. 38
1.0
0.9
c -.2 0.6 ...u ~ -2 ~ .3 0.4
0.0 '-___---L.. ____ .::r:::= ___ L....-...... :::==- __ o z 4 6 8 pH
Figure 2.1. Plot of calculated concentrations of HF and HFi for solutions of 15 M total fluoride as a function of pH. 39
High concentrations of NH4F buffer the reaction rate of Si02 and to prevent the attack of photoresist by HF [2.44].
NH4F is a strong electrolyte, and a large amount of F ion is generated by the NH4F-HF-H20 system according to the following equilibrium: (2.11)
The fluoride ion combines with HF to form the bifluoride ion, HF2-, following reaction (2.9). The reaction of BOE with Si02 may be represented by the reactions:
Si02 + 3HF2- + H+ ... SiFl + 2H20 (2.12)
NH4 + + siFl ... (NH4hSiF6• (2.13)
Figure 2.3 shows the etch rates of thermal oxide at 25°C in various BOE compositions. The etch rates increase with an increase in HF concentration but it is almost independent of NH4F concentration above the equivalent mole ratio.
The equivalent mole ratio line is determined from the fact that the molecular weight of NH4F is 37 and that the HF is 20. Kikuyama et al. [2.44] found that the excess NH4F concentration in the ordinary composition of BOE did not enhance the ionic reaction with silicon oxide during the etching and seriously degraded the etching uniformity and linearity due to the lack of solubility of etching product in
BOE. They improved the BOE by using an NH4F concentration of about 15% instead of 40%. The increase of NH4F concentration in BOE increased the solubility of etching species and provided the solid phase segregation during the
storage of BOE. 40
0 F-
HF;
-5 HF
-c: 0 :.;:; u u:e -10 -!! ~
-15
~o~--~~----~--~~--~~----~--~ -4 -2 o 2 4 6 8 pH
Figure 22. Plot of calculated concentrations of possible species in HF-Si02-H20 system as a function of pH. 41
HF' Cone. ( moVl )
-~ 30 -. c -~ () -e , - c: , c( ~. 0 - 01( 0 ~ - -t- 0 0 &! - '0 -• 10' c( 0 0 a; 20 10 I " .~- , ~ -0 - u u. 0 6--tt.'" ' 5 N ~ ~ ... ,,~\'P ... $ ~ .... ~. 4 i ~'(\. 10 3 2 1 0 0 10 HF WeIght Cone. (")
Figure 2.3. Etch rates of thermal oxide at 25°C in various BOE compositions (from Ref. [2.44]). 42
Kikuyama et al. [2.45] controlled the wettability of silicon in BOE through the addition of surfactants. Surfactants to be used in BOE solutions must have
the following characteristics: (1) good solubility in BOE, (2) hydrophilic property
at the wafer surface, (3) stability in BOE, (4) non-reaction with BOE,
(5) sufficient lowering of contact angle at the critical micelle concentration
(CMC). They found a binary surfactant system, consisting of a combination of
aliphatic amine and aliphatic alcohol or aliphatic acid, to meet all these
requirements.
2.1.2. Drying of Silicon Wafers after Wet Processing
In the aqueous processing of silicon wafers, drying after rinsing is a critical
step and must be accomplished by the physical removal of water. Spin
rinse/drying is an accepted and widely practiced procedure in industry to remove
water. However, it has been argued that this method [2.46] recontaminates the
wafer surface with particles and aerosols caused by the mechanical motion of the
dryer itself.
A method known as capillary drying removes water from the surface of
wafers by capillary action. In this method, wafers are first submerged in hot DI
water and then pulled out. Apparently, less than 1% of the water remains on the
surface to evaporate, leaving a particle-free surface [2.46].
In solvent vapor (usually isopropyl alcohol) drying, wafers wet with DI
water are dried by placing them in a solvent vapor, which displaces the water on 43 the wafer surface. Mishima et al. [2.47] found that the water content in IPA, the temperature distribution of IPA around the wafers and the velocity of IPA vapors affect the cleanliness of surfaces after IPA drying. They determined optimum conditions for IPA vapor drying such as the water content in IPA of less than
1000 pp~ IPA temperature around the wafers of 82°C and an IPA velocity of
5.0 em/sec. These conditions resulted in the cleanest surfaces during IPA drying.
Another modified IPA vapor drying process, known as direct displacement
IPA drying, eliminates the transport of rinsed wafers from another processing bath
[2.48,2.49]. In this technique, the wafers are first cleaned and then rinsed in DI water in a closed chamber. Following the rinse step, the wafers, still immersed in water, are exposed to IPA vapor. Bare, oxide-coated and silicon nitride-coated
silicon wafers were tested for particle removal in direct displacement IPA drying
[2.49]. Best particle removal was measured on bare wafers, whereas no particle
removal was observed on oxide and nitride coated wafers. It was hypothesized
that a condensate of IPA forms on the DI water and that the IPA spread readily
at the water/hydrophobic wafer interface displacing the aqueous phase. DI water
on hydrophilic wafers was believed to compete with IFA, thereby reducing the
effectiveness of IFA
The success of direct displacement IFA vapor drying may be explained by
the Marangoni effect [2.50,2.51], which is the flow along a liquid surface induced
by local variations in surface tension. When a flow of water-soluble IPA vapor
reaches a wafer/water/vapor interface, more vapor is adsorbed at the 44 water/vapor interface close to the wafer than at the water/vapor interface far from the wafer. This leads to a concentration gradient along the meniscus causing a surface pressure gradient. The surface pressure gradient induces a Marangoni flow of water directed away from the wafer.
2.1.2.1. Characterization of IPA
Physical Properties of Alcohols
Table 2.2 shows a comparison of the physical properties of alcohols with isopropyl alcohol. Alcohols can be divided into hydrophilic and hydrophobic alcohols. Alcohols with HLB (hydrophilic-lipophilic balance) numbers between 3 and 6 give water-in-alcohol type (i.e., hydrophobic); those with numbers between 8 and 18 give the alcohol-in-water type (i.e., hydrophilic). The empirical HLB number can be related to the chemical nature of alcohols by assigning numbers to structural groups and adding the pertinent ones together [2.52]. The relation is:
HLB = 1; hydrophilic group numbers
- 1; lipophilic group numbers + 7. (2.14)
Octyl, butyl, hexyl and benzyl alcohols have HLB values less than 7, low
solubilities in water, low vapor pressures and high boiling points. They are
hydrophobic and both advancing and receding contact angles reach zero, i.e., show
complete wetting on hydrophobic silicon wafers. Isopropyl, ethyl, methyl and allyl
alcohols, however, have opposite properties: HLB values larger than 7, complete
solubilities in water, high vapor pressures and low boiling points with Table 2.2. Physical Properties of Alcohols
P (1),(Il) 'Y, dynes/em, vapor , T~~~f Solubility in Solubility of Alcohols at 25°C mmHg water(l),(Il), wt% water(l), wt% HLB °A OR Octyl 19.7 0.14(2) 196 0.059(2) - 5.1 0 0 Isopropyl 21.2 33 82.4 Complete Complete 7.48 0 14 Butyl 21.7 5.5 117.7 7.7 20 7 0 0 Hexyl 21.7 1(2) 156.5 0.58 7.2 6.05 0 0 Ethyl 22.1 44 78.32 Complete Complete 7.95 0 14 Methyl 22.8 99 64.7 Complete Complete 8.43 0 15 Allyl 25.5 23.8(2) 96-98 Complete Complete 7.48 0 13 Benzyl 35.7 3.75(3) 230-205 4 5.2 5.58 0 0 ------(I)at 200C (2)at 25°C (3)at 77°C (Il)from Ref. [2.53]
~ 46 hydrophilicity. It should be noted that the receding contact angles did not reach zero in these hydrophilic alcohols.
IPA-Water System
Wilson and Simons [2.54) studied the vapor-liquid equilibria of the isopropyl alcohol-water system over a range of pressures from 95 mm Hg to
4 atmospheres. They reported the azeotropic composition and boiling point as
85.27 vol% IPA and 80.1O°C at 760 mm Hg, respectively. Figure 2.4 shows an equilibrium distribution diagram between liquid and vapor of IP A at their boiling temperatures. At an IPA content in water lower than 25%, more IPA evaporates from IP A/water solutions than water, and then remains constant at 80% IP A in water. Figure 2.5 shows the change of activity coefficients of IP A and water as a function of IPA content in water. Below 25%, IPA shows much higher activity coefficients than water. It demonstrates that, at a given pressure, IP A has a greater tendency to escape from IP A/water mixtures than water.
2.2. Contact Angles
Hydrophobicity /hydrophilicity of a silicon wafer surface is one of the important parameters in understanding the source of micro contamination. Oxide coated and RCA-cleaned wafer surfaces are hydrophilic, and HF-etched surfaces are hydrophobic due to the passivation by hydrogen termination [2.55].
Hydrophobic surfaces are characterized by large contact angles of water, often in 47
100 Cd 90· [[pcP ~ 80 COCO >" CCll CC CIO .5 70 . C ~ .... 60 . C ..,0 c 50 Q) I::! 40 C ~ 0 ~ 30 ~ C5 > 20· 10·
0 0 10 20 30 40 50 60 70 80 90 100
Volume Percent of IPA in Liquid
Figure 2.4. Equilibrium distn'bution diagram between liquid and vapor of IP A at their boiling temperatures. Replotted from Ref. [2.54). 48
1.0 Ol 0.9' D D D Log Activity Coef. of IPA 0.8 • Log Activity Coef. of Water .. 0.7 D I 0.6 • D O.S D 1Q 0.4 .- f c _. •• ! 0.3 I:b 0.2' 0 •• •0 0.1 • .. Do roo 0.0 -... •• .- - . . Ol:ln.n. 0 10 20 30 40 SO 60 70 80 90 100
Volume Percent of IPA in Water
Figure 25. Change of activity coefficients of IP A and water as a function of IP A content in water. Replotted from Ref. [2.54]. 49 the range of 40 to 110 degrees and low heats of immersion into water, 6 to
90 dynes/em [2.56]. Static contact angles have been measured on HF-etched silicon wafers [2.57,2.58]. Using methanol/water solutions, Gould and Irene [2.57] determined a value of critical surface tension ('Yc), of 27 dynes/em for Si from static contact angle measurements. The low value of 'Yc was attributed to the presence of hydrogen and fluorine species on the HF-etched Si surface. William and Goodman [2.58] measured the contact angle of water on silicon surfaces covered with oxide layers of different thicknesses. They found HF-etched silicon to be hydrophobic (8 :I:: 90°) and oxides thicker than 30 A to be hydrophilic
(8 cOO). Contact angles were a function of the oxide thickness. Contact angle values are very sensitive to surface contamination [2.59]. The contact angle measurement method is inexpensive and rapid in characterizing the solid surface even though the surface chemistry information is ambiguous. Table 2.3 shows the
Table 2.3. Advantages and Disadvantages of Contact Angle Methods (from Ref. [2.60])
Advantages Disadvantages Inexpensive to use Rapid to measure angles Artifact-prone method such as --=------.....:;------1 roughness, substrate contamination _S_im~p_le_to_o...:p:...e_r_at_e______I or modification and operator bias Theory is well developed Provides information related to /surface Provides ambiguous surface energetics chemistry information 50 advantages and disadvantages of the contact angle method in characterizing the solid surface.
The choice of a method to measure contact angles depends on the gross geometry of the system. For example, the most convenient method for a flat plate is not usable for the inner surface of a capillary tube, and fine textile fibers and powders create specific problems of measurements that are not present in other systems. Static and dynamic contact angle methods can be applied in measuring contact angles. The dynamic method is more accurate and objective and provides more information on surface energetics than static methods.
2.2.1. Static Angles - Sessile Drop or Adhering Gas Bubble Method
The method of sessile drop or, alternatively, of the adhering gas bubble is the most widely used static technique and is shown in Figure 2.6. It is, in general, the most convenient method and it has the two advantages: (1) requires only very small quantities of liquid and (2) is able to work with small samples. The equipment for the sessile drop method typically consists of a horizontal stage on which the flat specimen is mounted, a micrometer pipette with a small tip or other device for forming the drop, a source of illumination of the
drop from behind, a light filter to minimize heating by the light source and a
telescope with a protractor eyepiece. The contact angle may be determined by
either using a telescope equipped with a goniometer eyepiece or taking a
photograph and measuring the angle. 51
(a)
------::I--~ ---9 f/; 'It
(b)
Figure 26. Schematic diagram of techniques measuring static angles: (a) sessile drops on a solid surface, and (b) captive bubbles in a liquid underneath a solid surface. 52
For the captive bubble method, the specimen is immersed in the liquid and attached to the top of the chamber. The gas for the bubble must be supplied via a small-tipped tube, and the volume of the gas must be controlled.
On a smooth, homogeneous, rigid, isotropic solid surface, the equilibrium contact angle of a pure liquid is given by Young's equation,
'Ysv - "(sl = 'Ylv cos8c (2.15) where "(SV is the solid-vapor interfacial free energy, "(&1 the solid-liquid interfacial free energy, "(Iv the liquid-vapor interfacial tension, and 8e the equilibrium contact angle. For most systems, however, the equilibrium contact angle state is impossible to obtain.
In the sessile drop method, the advance and retreat of a drop on substrate make it possible to measure both advancing and receding angles. It is generally possible, by use of a motor driven syringe to control the rate of the addition of a liquid or its removal and to measure the rate of motion of the drop front. An appropriate rate for a sessile drop is about 0.01-0.1 mm/min linear advance or
retreat. Advancing angles are typically larger than receding angles (both depend upon the rate of advance and retreat). This introduces hysteresis in contact angle
measurements. To fully characterize materials, both advancing and receding
contact angles must be measured [2.61]. 53
2.2.2. Dynamic Angles - Wilhelmy Plate Method
The Wilhelmy plate method is widely used in the measurement of dynamic contact angles. The basic apparatus is shown schematically in Figure 2.7. The specimen, in the form of a plate, is suspended by a thin rod from an
electrobalance. The balance may be interfaced with a computer. The stage
carrying a beaker containing the liquid of interest may be moved up and down by
an electric motor to establish the advancing and receding angles at variable
speeds. Advancing and receding angles can be measured by immersing the plate
to some depth (advancing angle) and subsequently withdrawing the plate
(receding angle). In the Wilhelmy method, the perimeter must be constant along
the plate.
n a smooth, vertical plate of thickness 't' and width 'w' is brought into
contact with a liquid, as indicated in Figure 2.8, the liquid will exert a downward
force on the plate, which is given by
F p = P 'Ylv cosO (2.16)
where 'p' is the perimeter of the plate, 'Ylv is the surface tension of liquid, and 0 is
the contact angle. As the sample continues to penetrate the liquid, the measured
mass decreases due to buoyancy forces,
Fb = PI Vi g = PI ghtw (2.17) 54
Micro-balance
Sample Plate
'---I--Contact Angle Wetting Medium
t ~ Moving Stage
Figure 2.7. Schematic diagram of a Wilhelmy plate apparatus. 55
t F F P = 2(t+w) 1
2 F
mg
Figure 2.S. Schematic diagram illustrating a Wilhelmy plate method (from Ref. [2.62]). 56 where PI is density of the liquid, g is the acceleration due to gravity, Vi is the volume of plate immersed in the liquid and h is immersion depth. The total force on the plate during the immersion and emersion can be given by
F = mg + Fp - Fb
= mg + P 'Ylv cosO - Fb or
(2.18) where 'm' is mass of the plate. To evaluate the contact angle, it is necessary to know the surface tension of liquid medium. The surface tension of liquid can be measured using a highly clean, completely wettable solid (e.g. radio frequency glow discharged mica, prehydrated in water or precleaned glass with acetone and flame).
When a sample plate is slowly immersed in and withdrawn from solutions, force changes can be measured as a function of the depth of immersion of the plate. The result of such a measurement is generally a hysteresis loop. Figure 2.9 shows typical Wilhelmy plate hysteresis loops for (a) bare and (b) oxide-coated p(100) silicon wafer samples in DI water. The bare wafer sample was taken through two successive cycles. Prior to the contact between the plate and the
liquid, the force remains constant (line AB in Figure 2.9(a)). At zero net depth
of immersion (i.e. when the sample just touches the liquid) the force decreases
due to the capillary depression of the liquid at the plate if 0 > 90° as shown in
Figure 2.9(a). The force increases if 8 < 90° due to the capillary rise of the liquid 57
10000
SO.OO
·SO.OO
'100.00
·ISO.00
.m·~.~OO------I-.00~-----~-00------5.00~----~~~00------~9~.00~
Sb~ POllllon, mm
(a)
400.00
350.00
300.00 F 0 250.00 r c 200.00 e (119) 150.00 100.00
50.00
0.00
-50.00
-IOO.O~I.OO 1.00 3.00 5·00 7.00' 9.00 11·00 13.00 15.00 Stage Position (1It1ll)
(b)
Figure 2.9. Typical Wilhelmy plate hysteresis loops for (a) bare and (b) oxide coated silicon wafers in DI water. 58 as shown for the oxide-coated wafer (Figure 2.9(b». Further immersion decreases the force due to the buoyancy effect. In the emersion (withdrawal) step, the buoyancy force decreases and the net force increases as shown in Figure 2.9.
These buoyancy slopes in both immersion/emersion can be extrapolated to zero
depth of immersion to find advancing (F.J and receding (FJJ forces. In
Figure 2.9(a), F A,lst and FR,lst are used to compute the first cycle advancing (O.J
and receding (OR) angles on a bare wafer, respectively. If the average contact
angle along the line of contact does not remain constant during the immersion or
emersion or if the surface of the liquid is vibrating, the loop line will not be
straight but contorted.
The Wilhelmy plate method relies on weight change, which can be
measured with much higher accuracy and objectivity than the direct reading of an
angle with a goniometer. The reproducibility is therefore only limited by the
reproducibility of the solid surface and the reliability of determining the perimeter
'p' of the plate. However, there are several drawbacks that are not immediately
obvious and can restrict the usefulness of this method [2.61]. The plate must have
the same composition and morphology at all surfaces--front, back and both edges.
If one wants to investigate films deposited in vacuum, polished surfaces or
anisotropic systems, this condition may be difficult to meet. In measurements that
extend over appreciable time intervals, swelling of the solid (e.g. polymer) may
become a problem. Also, adsorption of the vapor of the liquid at various parts of
the gravimetric system other than the plate may affect the measurement of 59 contact angles. This is particularly serious for measurements of the temperature dependence of contact angles.
2.2.3. Origin of Hysteresis
On most surfaces, liquids that form nonzero contact angles usually exhibit hysteresis because of the heterogenous nature of surfaces, such as different
chemical composition, different crystallographic faces and the existence of defects
and roughness on surfaces [2.63]. Hysteresis is defined as the difference between
the advancing and receding angles
1:&0 = OA - OR (2.19)
Contact angles in dynamic system depend upon the rate of advance and
retreat of a liquid over a solid. The measurement of contact angles in a dynamic
system was made a function of rate for organic liquids on Teflon TFE [2.63]. It
was found that (JA (dynamic system) > (JA (static system), OR (dynamic system) <
OR (static system), and 1:&0 (dynamic system) > I:&(J (static system) for liquids such
as decane, over the entire range of speeds employed, which were as slow as 0.01
em/min. Attempts were made to extrapolate dynamic angles to zero rate. No
satisfactory extrapolation was found such that the limiting hysteresis, as rate
approached zero, was equal to the quasi-static hysteresis that was observed after
motion had been stopped. 60
Roughness
The macroscopic contact angles measured can be very different from the actual microscopic or true angles. Wenzel [2.64] attempted to relate contact angles to the true or total area, 'A', and to the apparent or geometric area A-.
He defined roughness, 'r', as
r = A/A (2.20)
Wenzel further claimed that
cosO = r cosOo (2.21)
where 0 is the observed macroscopic angle and 00 the true microscopic angle. For
a given 00, it is clear that as 'r' increases 0 may increase or decrease. Figure 2.10
gives 0 as a function of 'r' for various 00• For 00 > 90°, 0 increases as 'r' increases,
but for 00 < 90°,0 decreases as 'r' increases. If 0 is 90°, the contact angle is
independent of the surface roughness.
Surface Heterogeneity
Surface heterogeneity can also cause hysteresis. If a surface consists of two
materials having contact angles, 01 and O2, the equilibrium contact angle of a
heterogeneous surface, 0 is the weighted average of the angles on each of the two
phases:
(2.22) 61
1801r------~------r_------~
15
. B.· 90·
°1~--~--~------~3~------~4--~ SURFACE ROUGHNESS RATIO. r
Figure 2.10. Change of contact angles as a function of roughness factor, 'r' (from Ref. [2.65]). 62
where fl and f2 are the area fractions (fl + f2 = 1) with contact angles 91 and 82,J respectively. Advancing angles are associated with the intrinsic angle of the high contact-angle regions of surfaces. Similarly, receding angles are associated with low-contact-angle areas. Figure 2.11 gives a family of curves showing how contact angles and contact angle hysteresis vary with the percentage of the surface covered with the low-contact-angle material. The center line is calculated from. equation (2.22). The curves above the straight line represent possible advancing angles and those below represent corresponding receding angles [2.65]. It can also be seen from Figure 2.11 that, if the solid surface consists mainly of low energy (high-contact angle) regions, any change in the surface concentration of high-energy (low-contact angle) regions should produce large changes in the receding contact angles without significantly affecting the advancing angle.
Deformation, swelling and penetration of liquid into the surface were
observed in polymer materials and regarded as a major sources of contact angle hysteresis. Surface reorientation and motions are also possible in polymeric
surface. Table 2.4 summarizes the sources of contact angle hysteresis. For an
inorganic solid surface such as a silicon wafer, sources of contact angle hysteresis
are mainly caused by the surface roughness and heterogeneity due to the
contamination during the wafer process. 63
150 140 I~O
: lIoUI _120~~======~==~=CI 0.100 CI ~ 90 ~ eo ~ 70 ...~ 60 ~ 50 540 u 3020
~~~~-===:J~:S:::S:::5~~~~~~~~~60 70 eo 90 100
per cent 82 AREA
Figure 2.11. Change of contact angles and contact angle hysteresis as a function of surface coverage of low-energy material (from Ref. [2.65]). Curves 1,2, 3, and 4 indicate decreasing drop energies and upper and lower sets of curves are advancing and receding angles, respectively. Table 2.4. Sources of Contact Angle Hysteresis (From Ref. [2.62])
General Assumption Specific Assumption Effect on Hystersis Time Dependent
Surface is smooth Surface must be smooth /j.(J increase with increasing No at the 0.1 to 0.5 I'm roughness «(JA increases and level (JR decreases with increasing roughness if (J A > 90°)
Surface is heterogeneous Surface must be (J A dependent on low-energy No homogeneous at the phase; (JR dependent on 0.1 I'm level and above high-energy phase Surface is nondeformable Modulus of elasticity Yes > 3 x lOS dyne/cm2 - Wetting liquid does not Liquid molecular Increased liquid penetration Yes--due to penetrate surface volume> leads to increased /j.(J diffusion 60-70 cc/g-mole Surface does not reorient Reorientation time > > Increased tendency to Yes time of measurement orientation leads to increased /j.(J
~ 65
CHAPTER 3
EXPERIMENTAL MATERIALS AND PROCEDURES
3.1. Materials
3.1.1. Wafers
One hundred mm and 150 mm single-side polished (SSP) and double-side polished (DSP) silicon wafers, with and without thermally grown oxide coatings
(570 A), were used. Thermally grown oxides were prepared on p(100) wafers in a quartz tube furnace at 850°C for 30 min in a steam ambient. Table 3.1 shows the characteristics of wafers used in the experiments.
Table 3.1. Characteristics of Wafers Used in the Experiments
Type Direction Oxide Thickness, A Resistivity, O' cm p 100 570 10-80 P 100 No grown oxide 10-80 n 111 No grown oxide 1-15 n 100 No grown oxide 30-100
3.1.2. Chemicals
Choline was purchased from Sigma Chemical CO.(l) as a 50% solution.
Aliquots of this solution were appropriately diluted with 18 MO-cm DI water to
prepare choline solutions ranging in concentration from 6.5 to 10000 ppm
(l)Sigma Chemical Co., P.O. Box. 14508, St. Louis, MO 63178. 66
(1 wt%). An electronic-grade choline solution called "SummaClean" mixed with methanol and a proprietary surfactant was provided by the Mallinkrodt Chemical
CoP)
Reagent-grade TMAH was purchased from Aldrich Chemical CoP) as a
10% solution. Aliquots of this solution were diluted with Dl water to prepare
TMAH solutions. Electronic-grade TMAH (PFC-1) was provided by Moses
Chemical CO.(4) as a 1.2% solution. Electronic-grade NH40H (28% as NH3) and H20 2 (30%) reagents were used to make appropriate compositions of SCI solutions. Semiconductor-grade H2S04 (98%) and 400:1 BOE (etch rate, 28 A/min) and 5:1 BOE (etch rate, 1000 A/min) were also used in the experiments. High-purity non-ionic surfactant (NCW 601A) was provided by
WAKO Chemical Co. (5) as a 30% solution.
Reagent grade isopropyl alcohol (CH3CHOHCH3, 99.5%) was purchased from EM Science.(6) IPA-water solutions were prepared with 18 MO-cm Dl water. Red fluorescence polystyrene latex (PSL) microspheres (1.8 I'm in diameter), soda lime glass microspheres (1.6 .±. 0.3 I'm) and alumina particles
(2)Mallinkrodt, Inc., P.O. Box. M, Paris, KY 40361. (3)Adrich Chemical Co., P.O. Box 355, Milwaukee, WI 53201. (4)Moses Lake Industries, 760 Randolph Rd., Moses Lake, WA 98837. (S)WAKO Chemical Co., Tokyo, Japan. (6)EM Science, 480 Democrat Rd., Gibbstown, NJ 08027. 67
(1.2-1.4 I'm) were purchased from Duke Scientific CoP> In the as-received condition, PSL particles were suspended in water containing a surfactant with a particle concentration of 2.9 x 109/ml. Glass and alumina particles were in powder.
Silicon particles were purchased from Johnson and Mathey(8) as sieved -
325 mesh powder. These particles were filtered through an 8 I'm filter and the minus 8 I'm fraction was etched using 100:1 HF, DI water rinsed thoroughly, and then used for particle deposition experiments.
3.2. Experimental Procedures
3.2.1. Preparation of Wafer Samples
Wafers were cut with a diamond saw into 13 mm x 19 mm size samples in a class 100 cleanroom. Immediately after cutting, samples were spin-rinse cleaned with DI water to remove silicon particles produced during cutting. Analysis of cut samples by SEM/EDXA and Auger techniques showed no contamination except oxygen and carbon from the atmosphere during cutting. Samples were transferred to a class 10 cleanroom for further chemical cleaning. Piranha (8:2 ratio of
H2S04:H20 2) etch was performed for 10 min followed by a DI water rinse for 5 min. Bare wafers were cleaned by etching in 5:1 BOE for 30 sec and then rinsing in DI water for 30 sec. Oxide-coated wafers were etched in 400:1 BOE for
(7)Duke Scientific Co., P.O. Box 50005, Palo Alto, CA 94303. (8)Johnson and Mathey, 152 Andover St., Danvers, MA 01923. 68
1 min and DI water rinsed for 30 sec to obtain a fresh hydrophilic surface. The samples were then dried using dry nitrogen gas. Figure 3.1 shows the schematic diagram of the sequence of the cleaning procedures.
3.2.2. Wettability Measurements
3.2.2.1. Dynamic Contact Angle Measurements
A Cahn DCA-312 Dynamic Contact Angle Analyzer(9) was used in measuring the force on sample plates as they were immersed in and emersed from
aqueous solutions at speeds ranging from 2 to 264 p.m/sec at room temperature
and 35°C. The analyzer (Figure 3.2) measures dynamic contact angles based on
the Wilhelmy plate technique [3.1]. The instrument was interfaced with an
mM-PC for data acquisition.
Contact angle analysis was performed from the generated force vs. depth of
immersion curves. Contact angles were calculated from the formula [3.2,3.3]
aF = P'YLvCos8 - pgax (3.1)
where aF is the change in force as the sample is immersed or emersed, 'g' is the
gravitational constant, 'p' is the perimeter of the sample, 'YLV is the surface tension
of the liquid/vapor interface (dynes/em), 8 is the contact angle, p is the density of
wetting medium, 'a' is the cross-sectional area of the sample and 'x' is the depth
of immersion. The second term in equation (3.1) is a correction for the sample
buoyancy and is eliminated by extrapolation of the data to zero depth of
(9)Cahn Instruments, Inc., 16207 S. Carmenita Rd., Cerritos, CA 90701. 69
Wafers cut into ,13 mm x 19 mm
Piranha etch for 10 min
, DI water rinse for 5 min
BOE etch 400:1 BOE, 1 min for oxide-coated wafers 5:1 BOE, 30 sec for bare wafers
, DI water rinse for 30 sec
Dry in N2 gas u
Measurements of dynamic contact angles within 48 hrs
Figure 3.1. Schematic diagram of sample preparation procedures. 70
PC Micro-Balance
Inside Chamber ----, Temperature Controlled Circulating Water Bath printer
Water ++ Mo\Ang Stage
Figure 3.2. Schematic diagram of a DCA contact angle analyzer. 71 immersion. The contact angle is then calculated from the force at zero depth of immersion, 'F, by
F = 'YLVPcos8 (3.2) 0.981
In the above equation, the factor 0.981 in the denominator is a conversion factor from gm to dynes.
All contact angle and surface tension measurements were made in a laminar flow hood with an atmosphere of class 100 quality. For most experiments, a stage speed of 64 p.m/sec was used. The surface tension of aqueous solutions was measured using a 24 x 30 mm cleaned glass slide as the solid sample and assuming a contact angle of zero. For the measurements of the surface tension in BOE solutions, a clean Pt plate was used instead of the glass slide.
3.2.2.2. Static Contact Angle Measurements
Static contact angles were measured using an optical goniometer(IO) on
polished and unpolished sides of the wafers. Figure 3.3 shows a schematic
diagram of a static contact angle analyzer. Droplets of 10 p.l volume were used in
the measurements.
(IO)Rame-hart, Inc., NRL C. A. Gonometer, Model 100-00115,43 Bloomfield Ave., Mountain Lakes, NJ. 72
Liquid Illuminator
Power
Figure 3.3. Schematic diagram of a static contact angle analyzer. 73
3.2.3. Particle Deposition
Glass, alumina, silicon and fluorescent PSL particles were deposited on silicon using electrophoretic and aerosol spray methods. Figure 3.4 shows a schematic diagram of the experimental setups used for electrophoretic and aerosol spray deposition. In the electrophoretic technique, a wafer sample was held as
the anode in a cell containing an aqueous dispersion of PSL (106 particles/ml)
and glass particles (0.05 mg/IOO ml). A potential of 6 volts was applied between
the silicon sample and a titanium counter electrode was placed at a distance of
1 em from the sample. Particle deposition was typically carried out for 10 min. It
should be mentioned that the bare silicon sample in the electrophoretic deposition
became progressively hydrophilic by anodic oxidation during deposition.
An aerosol method was also used to deposit glass, PSL, alumina and silicon
particles on samples. This involved the use of an air brush.(ll) Polystyrene
latex particles were dispersed at a concentration of 108 particles/ml (for PSL);
glass, alumina and silicon particles were dispersed at a concentration of 0.05
mg/lOO ml in DI water. The deposition was carried out at a pressure of 20 psi
for approximately 30 sec. Both hydrophobic and hydrophilic samples were
coatable by this technique.
(ll)Richper Air Brush, Model 024C, Tokyo, Japan. 74
Ti Si
...... :::::::::::':::::::::::::::::: ... :::::::::::::::::::::::::::::::::::: ......
PSL Particles in DI Water
(a)
Silicon j:fttir~~~:1~::~~~:1======3 ,.. ., :.. : ...... -:,,:':;~"" :," Particles To air com pressor at 20 psi
(b)
Figure 3.4. Schematic diagram of the experimental set-up for (a) electrophoretic deposition and (b) aerosol spray deposition. 75
3.3. Characterization Methods
3.3.1. Scanning Electron Microscopy (SEM)
After cleaning the wafers, they were inspected using a JEOL SEM(12) with a Tracor(l3) x-ray analysis system. Samples were Au/Pd-coated to observe
the morphology of surfaces. For energy dispersive x-ray analysis, samples were
coated with carbon.
3.3.2. Auger Electron Spectroscopy (AES)
A Perkin Elmer Scanning Auger Microprobe(14) was used to characterize
the surface before and after cleaning wafers. A beam voltage of 5 kV and a beam
current of 10 nA were used at zero sample tilt. A sputtering rate of 10 A/sec was
used to sputter the surface.
3.3.3. X-ray Photoelectron Spectroscopy (XPS, ESCA)
A VG ESCALAB MkII(lS) photoelectron spectrometer with an
aluminum (Ka1 ,2 = 1486.6 eV) anode as the x-ray source was employed in the XPS analysis. Samples were left in a sample vacuum chamber at least a day
(12)JEOL, JSM-840A
(13)Tracor Norton TN 55023.3. Characterization Methods. (14)PHI 590 Scanning Auger Microprobe, Perkin Elmer Corporation, 1161C San Antonio Rd., Mountain View, CA 94043.
(lS)VG Instruments Inc., ESCALAB MkII XPS, 300 Broad Street, Stamford, CT 06901. 76 before obtaining spectra. Spectral data were obtained at a pressure of < 10- 11 torr. All spectral data were referenced to the C Is peak at 284.6 eV.
3.3.4. Ellipsometry
The thickness of films formed on the silicon substrate was determined using a Gaertner ellipsometer(16) at a wavelength of 6328 A and an incident angle of 70°. In the calculation of the thickness of silicon oxides, the index of refraction of the oxide film was assumed to be 1.46.
3.3.5. pH and Redox Potential Measurements and Optical Microscopy
Solution pH and oxidation and reduction potentials (ORP) were measured using a Cole-Parmer(17) epoxy body combination Ag/ AgCI electrode and an
Orion(l8) platinum redox electrode.
A Nikon Epiphot Metallurgical Optical Microscope(19) with differential interference contrast (DIC) was used to photograph deposited particles. An
Olympus Epi-Fluorescent Microscope(20) with long working distance objective
(16)Gaertner Scientific Co., Auto Gain Ellipsometer L116B, Chicago, IL. (I7)Cole-Parmer,7425 N. Oak Park Ave., Chicago, IL 60648. (18)OriOD, 529 Main St., Boston, MA 02129.
(19)NikoD, 623 Stewart Ave., Garden City, NY 11530. (20)Olympus, 4 Nevada Dr., Lake Success, NY 11042. 77 lenses was used to observe the movement of fluorescent microspheres
(excitation/emission, 541/611 nm) at the IPA/water/wafer interface.
3.3.6. Scanning Tunnelling (STM)/Atomic Force Microscopy (AFM)
The roughness of silicon surfaces was measured with a Digital Nanoscope
n.(21) Approximately 1 cm2 pieces were cleaved from the center of each of the
samples and degreased in acetone and methanol followed by nitrogen blow-dry.
For obtaining STM images, the native oxide on the sample surfaces was removed
by dipping samples in 49% HF for 10 sec or by leaving samples under 0.05 wt%
HF solution during imaging. AFM images were obtained by scanning the surface
with a Si3N4 tip without removing the native oxide. The root-mean-square surface
roughness (Rrms) and peak-valley roughness (~-v) of samples were measured
within the regions investigated. Rrms is defined as below:
(3.3)
where Zj = height at point (x,y), 'i' is the number of measurement and 'N' is the
number of data points. l\-v is the height difference between the highest and lowest peaks.
(21)Nanoscope II, Digital Instruments, Santa Barbara, CA 78
3.3.7. Inductively Coupled Plasma Spectroscopy (ICP)
The etch rate of silicon in various alkaline solutions was measured by analyzing the silicon concentrations in alkaline solutions using a Jarrell-Ash 800
Inductively Coupled Plasma Spectrometer (ICP).(22) SSP and DSP p(loo) wafers were immersed in 20 ml of ammonium hydroxide, choline and TMAH solutions for 3 hrs. The concentration of silicon in solutions was measured in ppm. The concentration of silicon was converted into total dissolved weight of
silicon, and then density of silicon was used to calculate the etched thickness of
silicon using the following equations which assume Psolution is 1:
m, grams = (C, ppm/l06) x V, ml (3.4)
and
t, em = m/(PSi x A) (3.5)
where, 'm' is the amount dissolved in solutions in grams, 'e' is the silicon
concentration measured in ppm, 'V' is the volume of solution, Psi is the density of
3 silicon (2.33 g/cm ), 'A' is the total area of silicon sample (5.2642 cm2) and 't' is
the thickness of silicon etched in solutions.
(22)Jarrell-Ash, 590 Rincoln St., Walpham, MA 02254. 79
CHAPTER 4
RESULTS AND DISCUSSION
4.1. Characterization of Samples
Figure 4.1 shows the SEM micrographs of polished (front) and unpolished
(back) sides of p(100), n(100), and n(111) silicon wafers used in the experiments.
It is interesting to note that the morphology of the back sides of wafers are quite different. This is due to the different etch chemicals (NaOH for (100) and nitric/HF for (111» used in the chemical thinning step prior to the final polishing of one side of the wafer. However, it is not clear yet why SEM morphologies of the unpolished side of p(100) and n(lOO) are different.
Cut samples were analyzed by Energy Dispersive Analysis by X-rays
(EDAX) and Scanning Auger Spectroscopy techniques before and after cleaning.
No contamination except oxygen and carbon from the atmosphere was detected
during cutting and cleaning.
4.2. Characterization of Chemicals
4.2.1. Alkaline Cleaning Solutions
The pH, redox potentials (reported on a standard hydrogen electrode
scale), and surface tensions of choline, TMAH and ammonium hydroxide
solutions were measured as a function of the solution concentration at 25°C. The 80
(a) (b)
(c) (d)
Figure 4.1. SEM micrographs of (a) polished wafer (independent of wafer type and orientation), (b) p(lOO) unpolished side, (c) n(lOO) unpolished side, and (d) n(1ll) unpolished side of silicon wafers. 81 effect of the addition of hydrogen peroxide to these solutions was also characterized.
Rapid changes in the pH were observed when the concentration of choline in the solution was raised from 0 to 6.3 ppm as shown in Figure 4.2(a). The solution redox potential decreased with increasing concentration in choline solutions. Above these concentrations, the pH continued to increase but less rapidly. Choline appears to become surface active at the air-solution interface at a solution concentration of approximately 400 ppm as evidenced by the onset of lowering of surface tension. A 10000 ppm solution of choline is characterized by a surface tension of 47.6 dynes/em. As expected, the surface tension of solutions at 35°C is slightly lower than at 25°C as shown in Figure 4.2(b). From the measured surface tension values of choline solutions, the Gibbs surface excess (r) of choline at the solution/air interface was calculated as a function of the solution composition using the following equation [4.1]:
r - _ C dl' (4.1) Choline RT de
These surface excess values are plotted in Figure 4.3. It may be seen that the surface excess of choline increases at solution concentrations above 1000 ppm and
has a maximum value of 2.89 x 10-10 moles/cm2 at 5000 ppm choline. A slight
decrease of surface excess was calculated above 5000 ppm of choline in solution.
The surface tension of choline solutions was not affected significantly by
the addition of hydrogen peroxide up to concentrations of 1000 ppm. The redox 82
600 14 1\ 1111111 I I ~~~ ~+H~-+~# \ ~ Ell 12 r\ 500 >e r\ ,..( 400 ~ 10 ! 300 I 8 II 1I 200 It I
4+-~~~-4~~*-~+# o o 10 100 1000 10000 Choline Concentration, ppm (a)
80
'l- """1j'-. f\ ~ at 3S·C -0- at 2S·C I\~ ,~
40 o 10 100 1000 10000 Choline Concentrltlon, ppm (b)
Figure 4.2. (a) pH, redox potential (vs. SHE) and (b) the surface tension (at 25°C and 35°C) of choline solutions. 83
3 ~ t\[\ ~... v M. ~> N ....E 2 V Kl 15 E / Ii 1/ 1/1 Q) U M LI.I , / "u t ~ / II
o 100 1000 10000
Choline Concentration, ppm
Figure 4.3. Surface excess of choline solutions at air/solution interface as a function of concentration. 84 potential and pH were slightly affected by adding hydrogen peroxide. The change of redox potential and pH is shown in Figure 4.4 as a function of hydrogen peroxide concentration in different choline solutions. The redox potential increased as the concentration of hydrogen peroxide in the choline solutions increased. The pH of choline solutions, however, decreased with the addition of hydrogen peroxide into solutions. These changes indicate that hydrogen peroxide is both the oxidizing and acidic agent (p~ of H20 2 is 11.7 at 25°C). TMAH showed similar changes in redox potential and pH as choline. The rapid change in pH, as shown in Figure 4.5, was observed when the concentration of TMAH in solution was raised from 0 to 10 ppm. Above these concentrations, pH continued to increase but less rapidly. The solution redox potential decreased with increasing concentration. The surface tension of TMAH solutions decreased only slightly as the concentration of TMAH was increased. This is in sharp contrast to the behavior of choline solutions.
As shown in Figure 4.6, small additions of NH40H to DI water rapidly changed both the pH and redox potential of the solutions. The surface tension of
ammonia solutions decreased slightly as the concentration of NH40H was
increased. Also, the presence of H20 2 in aqueous ammonium hydroxide solutions at a concentration typically used in a conventional SC1 formulation lowered the
pH from 11.7 to 10.3. However, slight decrease increase of redox potential was
observed in the presence of H20 2 in NH40H solutions. 85
450 I I T 1 200 ppm Choline 400 pp!" Chol.lne .~ 1000 ppm Choline >e - 350 5000 ppm Choline - - ~ - IU iii .....~ 250 ..... ---t J-- V ~ ...... - ~ ~ I l../ It--~ l/ ....- ~ 150 It--+-" J V V 50 o 200 400 600 BOO 1000 H202 Addition in Choline, ppm (a)
14 I I I I r-o-- 200 ppm Choline f-+- ~oo ~ Choline 13 on: 1000 ppm ChoHne _ 5000 ppm Choline I- '"""1 I"------~ ~ 12 """1 i-- -...... J-- r--- """--1 ---t ~ ~ I--t ]... 11 ..., ~ -...., >-- ~ ~ r--- - 10
9 o 200 400 600 800 1000 H202 Adcition in Choline Solutions, ppm (b)
Figure 4.4. Changes of (a) pH and (b) redox potential (vs. SHE) as a function of HzOz addition in choline solutions 86
,.. 600
12 500 ~ .-I:
10 = "-i i "00 8 I 300 6 Ia:
200 "0 10 100 1000 10000 TMAH Concentration. ppm
(a)
80.------,
75 - I 70 i 65
I 60 -- Surface Tension ... 55
I 50 "5
o 10 100 1000 10000 TMNI Concentration. ppm (b)
Figure 4.5. (a) pH and redox potential (vs. SHE) and (b) surface tension of TMAH solutions. 87
80 14
75 12 70
I 65 i 10 :: 60 a. 8 J 55 NH40H-HZO 8 --0- Surface Tension pH 50 • i 6 NH40H In 1:~ Hz0Z-HZO 45 --a- Surface Tension -II- pH 40 4 0 .1 10 100
Volume Percent of NH40H In Water (a)
700
600
>e --0- NH40t+H20 ,.... SOO --0-- NH40H In 1:5 H2O-H2O ~ ! ~OO "i i ! 300 I 200
100 0 .1 10 100
Vohne Percent of MI40H In Water (b)
Figure 4.6. (a) pH and surface tension, and (b) redox potentials of NH40H solutions with and without the addition of H20 2• 88
The alkaline solutions used in this research are capable of etching silicon.
In order to characterize the etching capability of these solutions, the etch rate of p(100) samples was calculated by measuring the amount of dissolved silicon in solutions using Inductively Coupled Plasma (ICP) atomic emission spectroscopy.
Figure 4.7 shows the etch rate of p(100) bare silicon in alkaline solutions at room temperature. The etch rate of silicon was measured to be 18 A/min and
234 A/min in 10000 ppm choline and TMAH solutions, respectively. The etch rate of silicon in TMAH solutions depended on the concentration of TMAH.
When the concentration of TMAH was decreased to 100 ppm, an etch rate of
8.8 A/min was measured. Interestingly, SumrnaClean, which is a comrnercially available choline solution, etched silicon at a rate of 7 A/min. A 1:5
NH40H:H20 (17 vol. %) solution was found to etch silicon at a rate of 380 A/min. When the ammonia volume ratio was decreased to 0.01 in water, the etch
rate was still high (73 A/min). The etch rate of the SiOrcoated wafer was
measured to be <0.1 A/min in these alkaline solutions.
As shown in Figure 4.6(c), at a given pH, the highest etch rate was
measured in ammonia solutions. This shows that OH- is not the only factor
controlling the etch rate, perhaps the cations of the different alkalis play an
important role in etching. The cation of the surface active choline may adsorb on
the silicon surface, thereby reducing the etch rate significantly.
Meerakker and Staraaten [4.2] studied the influence of pH in NH40H and NaOH solutions by varying pH from 11.5 to 13. They found that pH had a 89
250 I I -<>- Etch rate of Choline at 2S'C ~~ 200 1---0- Etch rate of TMAH at 2S'C I VI pH- 12. 9 ,.V , ! 150 I7pH-12.6- f 100 / pH-11.9- V ~I-" pH- 12.4 pH- 11.0 V L'!. so , pH- 12.7 / k pH-11.2 pH-11.7 I' ~pH-l0.3~ II JJt 1""11' o 100 1000 10000 Solution Concentration, ppm (a)
500
1 ~ 1 1 I. 1 I --e- Etch rate of anmonia solution at 2S'C 400 LI; ~ ~ / pH-12.4 ~ 300 / pH- 12.2 t"""-"- J / ... 200 pH-11.6 ~ rf I 100 ~ pH-11.2
o o 10 20 30 40 so
Volume " of tet40H in Water (b)
Figure 4.7. Etch rates of p(100) silicon wafer in alkaline solutions at 25°C, (a) choline and TMAH and (b) 1:5 NH.OH:HzO solutions as a function of their concentrations. 90
.. 00 :1 -.- tfl40H II --0-- TMAH 300 -....- Choline J j 200 .! I 1 II 100 ~) !--" j .iO o .. 6 ----8 10 12 14 pH
(c)
Figure 4.7. (Continued) (c) Etch rates of p(lOO) silicon wafer in choline, TMAH, and 1:5 NH40H:H20 solutions as a function of pH at 25°C. 91 moderate effect on the etch rate; increasing the OH- concentration by a factor of
30 enhanced the etch rate by a factor of 10. The etch rate of the NH40H solution was much higher than that of the NaOH at the same pH values. Table
4.1 gives the etch rate of silicon in various alkaline solutions at elevated temperatures as reported in literature. When etching temperature increases from room temperature to 90°C, choline significantly increases etch rate from 14 A/min to 2300 A/min at a concentration of 5000 ppm (0.5 wt%) while 1:5 NH40H:H20 solution increases the etch rate from 370 A/min to 1100 A/min. It might be due
to the fast evaporation of NH40H in a 1:5 NH40H:H20 solution at elevated
temperature.
Table 4.1. Etch Rates of Silicon Wafers in Alkaline Solutions at Higher Temperatures
Etch Rate, Chemicals A/min T,oC Substrates Reference Choline 0.5 wt% 2300 90 p(l00) [4.3] Choline 1 wt% with 10 90 (100) [4.4] 3 wt% H20 2 TMAH 0.076 wt% with 180 70 n(l00) [4.5] surfactant
1:1:5 NH4OH:H20 2:H2O 0.5 85 p(100) [4.6]
1:5 NH4OH:H2O 1100 85-92 (100) [4.7]
4.2.2. Buffered Oxide Etch (BOE)
Buffered oxide etch solutions, with and without surfactant, from two
different vendors were characterized in terms of their ability to wet silicon. 92
Table 4.2 shows the surface tension and contact angles of BOE (400:1 and 200:1) solution on p(1oo) bare silicon wafers. For the measurement of surface tension of
BOE solutions, a platinum plate was used instead of a glass slide. BOE without
surfactant obtained from both vendors showed the same surface tension and
contact angles. The addition of surfactant significantly changes the surface
tension and contact angles. The difference in the surface tension of surfactant-
containing BOE solutions from the two different vendors is most likely due to the
difference in the type of proprietary surfactants used. Poly-oxyalkylene alkyl and
alkyl phenol ether are the general type of surfactants used in BOE.
Table 4.2. Surface Tensions and Dynamic Contact Angles of Various Concentrations of BOE (400:1 and 200:1) on Bare p(1oo) Wafers from Two Different Vendors (30°C)
BOE 'Y, dynesl cm OA OR Vendor A wlo surfactant 90.8 93.6 82.2 wi surfactant 22.0 0 0 VendorB wlo surfactant 90.4 93.5 82.5 wI surfactant 33.0 28.2 20.2
Figure 4.8 shows the effects of cleaning in BOE solutions on contact angles
of DI water and 400 ppm choline solutions. In these experiments, p(100) wafer
samples were treated with two different BOEs for 1 min and then rinsed in DI
water for 1 min. For samples cleaned in surfactant-free BOE, larger angles were
measured in DI water. Significant differences were noticed in the water 93
100~------=~--~ • MY. Angle. in 01 Water IZI Rec. Angles in DI water o MY. Angles in 400 ppm Choline 80 B Ret. Angles in 400 ppm Choline
I 60 ~
I 40
20
o Vendor A Vendor B Vendor A Vendor B wlo Surfactant wI Surfactant
Figure 4.8. Effects of surfactant addition in BOE and different vendors on contact angles in DI water and choline solution when silicon wafers were etched with BOEs. 94 wettability of samples treated in surfactant-containing BOE from the two different sources. Both advancing and receding angles of water on sample A were higher than that on sample B. Ideally, DI rinsing of sample showed remove the surfactant and return the wafer to the highly hydrophobic state. But, as may be noticed from the data in Figure 4.8, the sample treated with BOE from vendor B is not very hydrophobic even after a DI water rinse. This means that the surfactant in the BOE sample (B) adsorbs rather strongly on the silicon. This is an undesirable effect.
The treatment of wafers in surfactant-containing BOE did not provide the same reproducible surface in terms of contact angles. This might be due to a variation in the amount of surfactants remaining on the silicon surfaces. The use of BOE without surfactant provided reproducible bare silicon surfaces.
4.3. Dynamic Contact Angle Measurements
4.3.1. The Change of Contact Angles during the Storage of Samples
Precleaned silicon wafers were transferred from the cleanroom to the laboratory and were stored under the laminar flow hood until experiments were
performed. The change in the wettability of BOE etched p(lOO) samples was
characterized as a function of storage time. Figure 4.9 shows the contact angles
of DI water and oxide thickness on wafers as a function of the storage time. The
oxide thickness almost tripled over a storage time of 30 days. Both advancing and
receding contact angles decreased steadily with storage time. These results with 95
100 I I --0- Adv. Anglel Rec. Ang~1 80 i%-c r--o- ~ ~ 60 f ~ ~ h 40 I " 'l.J
20
o o 10 20 30
Tee, days (a)
30
....A Oxide Thickess. It.
20 ~ ,---~t"'" ~ r--- ~ 10 /: ~~
o o 10 20 30
Tine, days (b)
Figure 4.9. Cbange of (a) contact angles and (b) oxide thickness during storage in air as a function of time. 96
BOE-cleaned wafers are quite similar to the observation of Philipossian [4.8] who measured the contact angles and oxide thickness of HF-treated hydrophobic surfaces as a function of air exposure time. He etched wafers with vapor and liquid-HF /H20 and exposed samples to air up to 10000 min. A decrease of contact angles (62° to 35°) and an increase of oxide thickness (8 A to 18 A) were observed as a function of air exposure time.
It is also known that the hydrophilic wafer becomes more hydrophobic at atmosphere. Suzuki et al. [4.9] measured the contact angles of water on oxide coated wafers as a function of storage time in a cleanroom. They observed the increase in contact angles and attributed this to the adsorption of gases in cleanroom. Kissinger et al. [4.10] also reported that the aging of RCA-cleaned wafers strongly decreased the hydrophilicity after 24 hrs and thereafter much more slowly at room temperature. He etched wafers with vapor and liquid-
HF/H 20 and exposed samples to air up to 10000 min. The decrease of contact angles and increase of oxide thickness were observed as a function of air exposure time.
It is pertinent to point out that the hydrophobic wafers become less hydrophobic [4.8] and the hydrophilic wafers become less hydrophilic [4.9,4.10]
during storage of wafers in a cleanroom. It is known that the hydrophilicity of a
silicon surface is caused by surface hydroxyls or adsorbed OH- groups and
hydrophobicity by H- and CHx- groups [4.11]. A freshly BOE-etched surface is
terminated by hydrogen atoms. This surface can slowly adsorb H20 or O2 from 97 the atmosphere, which can explain the slow reduction in hydrophobicity over a period 30 days. Therefore, it is necessary to process wafers within 24 hrs after etching to maintain the same surface wettability as a freshly etched sample.
4.3.2. The Effect of Stage Speed of the Dynamic Contact Angle Analyzer on Wettability
In measuring dynamic contact angles by the Wilhelmy plate method, the stage speed is known to affect the measured contact angles. Figure 4.10 shows the dynamic contact angles of p(100) bare SSP wafers in DI water as a function of
the stage speed. The advancing angles were independent of the stage speed in
the range 2 to 264 p.m/sec. The receding angles, however, were a strong function
of stage speed for speeds less than 20 p.m/sec. A receding angle of 32° was
measured at a stage speed of 2 p.m/sec. At high speeds (> 170 p.m/sec), it was
difficult to obtain a linear extrapolation of the buoyancy line for the calculation of
contact angles. This introduces some inaccuracy in the measured contact angles.
Figure 4.11 shows the hysteresis loops at different stage speeds. The presence of
less noisy buoyancy slopes at lower stage speeds can easily be noticed from their
figures. 98
100
.h. T _ ~o.. ~ l-----<> iii v - ---0- Adv. Angles 80 -0-- Rec. Angles
ell ..-0 G) .:r... __ C, .... c ';t' < 60 ~ --- ~ U ...,t'O C / 0 U 40 ~ l
20 a 100 200 300
Stage Speed. pm/sec
Figure 4.10. Effect of stage speed on contact angles in DI water. 99
DI E
Stage Position, nun
Figure 4.11. Changes of hysteresis loops at different stage speeds: (a) 264 p.m/sec, (b) 170 p.m/sec, (c) 64 p.m/sec and (d) 20 p.m/sec. 100
4.4. Dynamic Wetting of Wafers in Alkaline Solutions
For most of the measurements of advancing and receding angles, an
immersion/emersion rate of 64 p.m/sec was used. Multiple immersion/emersion
cycles were performed to observe the changes in advancing and receding angles, if
any, due to cycling.
4.4.1. Wettability of Wafers in Choline Containing Solutions
4.4.1.1. Wettability of BOE-Etched p(loo) Wafers
Most of the wafers used in measuring contact angles were single-side
polished (SSP) wafers. Some measurements were done on double-side polished
(DSP) p(100) wafers to determine the dependence of contact angles on surface
finish. Multiple immersion/ emersion cycles were performed to observe the
change in advancing and receding angles on Si wafers in choline solutions.
Figure 4.12(a) shows advancing (fJA) and receding (fJ0 contact angles on
bare p(100) wafers (single- and double-side polished) during the first
immersion/ emersion cycle as a function of choline concentration. Double-side
polishing did not significantly alter advancing «(JA) or receding (fJ0 contact angles.
Advancing angles remained almost constant until a choline concentration of
400 ppm was reached, and then began to decrease. The receding angles
decreased gradually and were around 20° for choline concentrations in excess of
5000 ppm. 101
The second-cycle behavior of wafers in choline solutions is shown in
Figure 4.12(b). On both SSP and DSP wafers, OA decreased significantly when choline concentrations above 200 ppm increased. A comparison of Figure 4.12(b) to 4.12(a) clearly reveals that at high choline concentrations the second cycle OA
and OR values are almost equal. A DI water rinse after the choline treatment did
not return the wafers to their original hydrophobic states.
It is well known that if the contact angle is less than 90°, it is decreased,
and if larger than 90°, increased, by roughness [4.12]. Equation (4.2) relates
change of contact angles to the roughness factor, r:
(4.2)
where Or is the contact angle on a rough (apparent) surface, 0& is the contact angle
on a smooth (true) surface, and r the roughness factor (ratio of area of rough
surface to that of smooth surface). In the first and second cycles, both advancing
and receding angles on both SSP and DSP wafers did not show much difference.
This lack of difference in contact angles between SSP and DSP wafers was
initially very puzzling. This may be due to the fact that the roughness scale on
the unpolished side was not large enough to measure macroscopic contact angles.
When DSP wafer surfaces were roughened using sandpaper (100 grit) to generate
very rough surfaces, dynamic contact angles, (J A and OR' were reduced to 85° and
200 from 92° and 65°, respectively.
Figure 4.13 displays contact angles of DI water on samples immersed in
choline solutions for 10 min and on samples that were subjected to a two-cycle 102
100
~ eo ~t-. -0- SSP 111 Adv. .!.!!Q- 'sSp'1iiReC. -.- OSP 111 Adv. ~ __ OSP 111 AIle. eo r--. t""" h:' ~ ..... , '1~!\ ...
o o 10 100 1000 10000 Choline Concentration, ppm (a)
100 ..... I eo ~ ~j-.. h -0- SSPrt~ -0- SSP ald AIle. ~~ ~ -.- OSP ald Afti. ~. DsP2ndAec. ~~ ~ ~ ~, i'~ ,~ ,...... " :'-.I' - o o 10 100 1000 10000 Choline Concentratlon,ppm (b)
Figure 4.12. Advancing (8,J and receding (8.J angles on p(100) bare wafers in choline solutions, (a) first cycle and (b) second cycle. 103
100 -a- 10 min dipping, Adv. • 10 min Dipping, Rec • a After 2 cycles, Adv. 80 "" After 2 cycles, Rec. ~ C. 60 .5 ] ~ < 40 tl
u~ 20
O~J~~mm~~~m-~~~~~~ o 10 100 1000 10000 Otoline Concentration. ppm. before DI Water Rinse
Figure 4.13. Contact angles in OI water on a sample immersed in choline (10000 ppm) for 10 min and on a sample that was subjected to a two cycle test in choline. 104 test in choline solutions. Both these samples were rinsed in DI water for 5 min before making measurements. Rinsing in DI water after conditioning in choline did not return the wafer to the original highly hydrophobic surface state. A sample that went through two cycles in 1 wt% choline exhibited much lower advancing angles than a sample subjected to a 10 min dipping in choline. The reason for this will be discussed later.
4.4.1.2. Wettability of p(100) Wafers with Thermally Grown Oxide
The wetting of oxide-coated p(lOO) wafers was investigated next.
Figure 4.14(a) shows advancing and receding angles of choline solutions on oxide
coated (570 A) wafers. Advancing angles in the first cycle decreased as choline
concentration increased. Zero advancing angles (complete wetting) were
measured in the second cycle in solutions at a choline concentration greater than
100 ppm. Receding angles decreased slowly until 200 ppm and increased to 20° at
10000 ppm. Cycling did not affect the receding angles over the entire choline
concentration. Figure 4.14(b) shows contact angles of DI water on samples
subjected to a two-cycle immersion in choline followed by a DI water rinse for 5
min. Zero advancing and receding angles were measured for samples cycled in
choline solutions with a concentration of 5000 ppm or higher. 105
--0- 1st Ai:Jv. I -0-- 1st Rec. -.- 2nd ADv. ___ 2nd Rec. ~ ! 30 \ t-- I r\ .... I 1 \ I I \ 'j I I I I ~ ~ ~ I 10 - r\ i/ l?-t- IJ' o ...-r ~ o 10 100 1000 10000 o.oIine Concent., ppm, before DI Water Rinse (a)
~'1~!~~sl --0-- Rec. An9~s ;
,..j..( V I I I I I i I
J-. I 10 !:--:--- !J 1\! I ~ ~ ~ f\ ,,~~ o o 10 100 1000 10000 o.oIlne Concent., ppm, before DI Water Rinse (b)
Figure 4.14. Advancing (8.J and receding (8.J angles on oxide coated p(l00) wafers in (a) choline solutions and (b) in DI water after conditioning in choline followed by a DI water rinse for 5 min. 106
4.4.1.3. Effects of Wafer Types and Orientations on Wettability
The effects of choline on the wettability of p(100), n(100) and n(111) wafers are shown in Figures 4.15 and 4.16. Figure 4.15 shows advancing and receding angles on BOE-etched p(100) and n(100) wafers as a function of the choline concentration. Similar behavior was observed on both types of wafers.
This reveals that dopants do not affect the wettability of wafers with the same crystal orientation.
Contact angles on n-type wafers of (100) and (111) orientation are presented in Figures 4.16(a) and (b). The surface energies of (111) and (100) planes of silicon have been reported to be 1230 dynes/em and 2130 dynes/cm, respectively [4.13]. Higher solid surface energy ('Ysv) should ideally lead to better wetting. The measurements of both advancing and receding angles on BOE etched silicon wafers, however, showed no effects of crystal orientation on wettability. This might be due to the passivation of the BOE-etched silicon surface by species such as hydrogen [4.14-4.16]. A low value of critical surface tension of wetting of 27 dynes/em has been reported for Si cleaned in BOE, independent of crystal orientation.
4.4.1.4. Effect of Hydrogen Peroxide on Wettability in Choline Solutions
As mentioned in the background section, choline/peroxide solutions have
shown promise as cleaning agents for silicon wafers. The addition of H20 2 to choline solutions increased the redox potentials but decreased the pH. The effect 107
100 I 11 ...... eo II "'l'" --0- P(100) 1st MY. --0- P(100) 11t Rec. --e- n(100) 1st MY. ~ n(100) 2nd Rec:. r" ---- ~ r'\i' ~ I',~ 1'\ .... I
o o 10 100 1000 10000 Choline Concentration, ppm
(a)
100 r--?- P(100)2nd~1 ~ P(100) 2nd Rec. If:::= ~ n(100) 2nd MY. eo F"'I" ~. n(100) 2nd Rec. i""'" 1-:1== t- 1\ I~ ~ ~~ .~ l\ ~\ ~ ~ ~I' " Y o o 10 100 1000 10000 Choline Concentrations, ppm
(b)
Figure 4.15. Wettability of BOE etched p(l00) and n(l00) wafers in (a) the first cycle and (b) the second cycle in choline solutions. 108
100
...... eo ~~ I n(100) 1st Adv. If'\ n(100) 1st Rec. ~ r---- n(111) 1st Adv. "f'o -- n(lll) 1st Rec. ~ ~~ t'-. .. 'i yo
o o 10 100 1000 10000
Choline Concentration, ppm (a)
100 JJl"~~00)'~d'1WI r--...f=:::: --:a--:. °° 0(100) 2nd °Rec.° ° eo ,...... !e!-:.. I I ~(111) 2n'd Ac"/:' ""'I"- --+-:-" ii(111)'2Iic! 'Rae:' ~~ ~ 'I " ~ N> ~ i\. ~ 1\ '\ ,~ ~ "'-.j ~~
o o 10 100 1000 10000 Choline Concentration, ppm (b)
Figure 4.16. Effect of wafer orientation on wettability in choline solutions, (a) the first cycle and (b) the second cycle. 109
of the addition of H20 2 to choline solutions on wafer wettability is shown in
Figures 4.17(a) and (b). In a 400 ppm choline solution, H20 2 significantly lowered advancing angles in the second cycle. Receding angles remained the same in both the first and second cycles throughout the entire hydrogen peroxide concentration investigated. At a choline concentration of 5000 ppm, the second cycle advancing and receding angles were low and identical. The decrease of advancing angles in the second cycle in both 400 and 5000 ppm choline solutions indicates that hydrogen peroxide helps to increase the wettability of silicon.
Contact angles of bare wafers were also measured in H 20 2 (1 wt%) solutions containing choline at different levels. In the absence of choline, the solution with hydrogen peroxide reduced the advancing angle to 37° in the second cycle as shown in Figure 4.18. Adding choline to 1 wt% H20 2 solutions further decreased the advancing angles both in the first and second cycles. As mentioned earlier, choline/peroxide solutions are more effective in reducing contact angles than peroxide solutions.
4.4.1.5. Effect of Surfactant on the Wettability of Silicon in Choline Solutions
In aqueous solutions, surfactants, depending on their structure, may form organized aggregates called micelles, in which the lipophilic parts of the surfactants associate in the interior of the aggregate and leave hydrophilic parts to face the aqueous medium. The surfactant concentration at which micelle formation becomes significant is called the critical micelle concentration (CMC) 110
100
I eo ~ 1st /ldv. -- 1st Ret. I 2nd /Idv. eo - 2nd Ret._ I "'~
40 I ':-., ,...... - ...... :'" o o 10 100 1000 10000
Hydrogen Peroxide Concentration, ppm (a)
100
~
~ 1st IvJv. -0- 1st Ret., eo -e-- 2nd /Idv. -- 2nd Ret.
~
o o 10 100 1000 10000
Hydrogen Peroxide Concentration, ppm (b)
Figure 4.17. Advancing (SA) and receding (S.J angles on BOB etched DSP p(100) wafers as a function of hydrogen peroxide concentration; (a) choline content in solution 400 ppm and (b) 5000 ppm. 111
100
--i't--.N 80 1st Adv. ... 1st Rec. - 2nd Adv • • ") ~ • 2nd Rec • ~ 60 < t ~ 40 ...
..... ~ C1 ~ 20 ~ ~ Ii'"
o o 10 100 1000 10000
Choline Concentration, ppm, with 1 wtC)i H202
Figure 4.18. Contact angles in 1 wt% H20 2 solutions as a function of choline concentrations. 112 and can be determined from a plot of surface tension versus concentration [4.17].
The CMC of the non-ionic surfactant, NCW 601A (polyoxyalkylene alkylphenol ether), used was determined by measuring the surface tension of solutions as a function of surfactant concentration. Figure 4.19 shows the change of surface tensions of DI water, 400 ppm and 5000 ppm choline solutions as a function of surfactant addition. No changes in surface tension were measured above a surfactant concentration of 100 ppm; this was determined to be the CMC.
Figure 4.20(a) shows the effects of surfactant addition to choline solutions on wafer wettability. Even when added to 01 water, the surfactant significantly lowered the advancing angle in the first cycle and complete wetting was observed
during the second cycle immersion step. Receding angles remained the same both
in the first and second cycles across the entire choline concentration range
investigated.
Figure 4.20(b) shows the contact angles of 01 water on samples subjected
to a two-cycle immersion in choline solutions containing the surfactant at a
concentration of 100 ppm, followed by a 01 water rinse for 5 min. Based on
receding angles, it may be said that samples exposed to surfactant containing
choline solutions of concentrations larger than 1000 ppm retained their
hydrophilicity. For example, for the sample conditioned twice in a 1000 ppm
choline solution, the advancing contact angle decreased from 75° to about 400.
This indicates the surface modification in choline-surfactant solutions. Adding
choline to DI water continued to modify the silicon surface up to a choline 113
80 ~ 1111111 In 01 Water :~ • e 70 ~ ClIoIine 400 ppm - ClIoIine 5000 ppm i ~ - -6' 60 ~~ 1\1' co I' '! I~ ... ~ 50 ,....~ 8 ~ ~ ~ I'~~ ~ r-.. en 40 .-
30 o 10 100 1000 10000
Surfactant Concentration, ppm
Figure 4.19. Change of surface tension as a function of surfactant (NCW 601A) concentration. 114
100 11]1l! 11J~ -""" 1st Rec. 80 - 2ndMi - 2nd Rec
I 60
I 40
20
10 100 1000 10000 Choline Concentration, ppm (a)
100 ~"'Adv.k~ - Rec. Angles
t---I'--o l"'-r-- r-,r. r--- :>-- I"'- ~ ..... r"'"' ~ o o 10 100 1000 10000 Choine Concentration, .... befan DI Water Rinse (b)
Figure 4.20. (a) Effect of surfactant addition (100 ppm) to choline solutions on wafer wettability and (b) contact angles in DI water on samples subjected to a two-cycle immersion in a choline solution with surfactant (100 ppm) followed by a DI water rinse for 5 min. 115 concentration of 1000 ppm and reached constant angles in both advancing and receding angles. At choline concentrations higher than 1000 ppm, choline modified silicon surfaces much more dominantly than the surfactant did.
4.4.1.6. Effect of Temperature on Wettability in Choline Solutions
The wettability of silicon wafers in choline solutions was also measured at
35°C. Since the condensation of the wetting medium on the sample plate and sample holder may mislead the measurement of contact angles at high temperatures, a slightly elevated temperature of 35°C was chosen.
Figure 4.21(a) shows the change of wettability as a function of choline
concentration at 35°C. When compared to the data at 25°C shown in Figure 4.12,
it may be seen that the advancing angles in the second cycle were much lower at
35°C than at 25°C. At 35°C, the second cycle advancing angle in 200 ppm choline
solution was measured to be only 30°, which is about 40° lower than that
measured at 25°C. Figure 4.21(b) shows contact angles measured in DI water on
samples that were subjected to a two-cycle test in choline. These samples were
rinsed in DI water for 5 min before making measurements. The advancing angles
were lower at 35°C than that at 25°C. The improved wettability at higher
temperature may be due to the lower surface tension of choline solutions as well
as due to the thicker oxide formation on silicon, which will be discussed in the
next section. 116
100
j~ ...... : 80 t--." --0- lit Aljy. I "to. ~ lit Ree. i~ --e- 2nd Adv. ) 2nd Rec. I ~~ ,,~ I' --- ~ ~~ ~ "'" I ~ 20 't== ~h!. o I • o 10 100 1000 10000
Choline Concentration, ppm (a)
100 I i': 80 I -... AlJy. " ~ Rec. I'... I "" t..;; f'..
I 1.• ~ ~ I 1'-." ~'" I ~ to... 20 "'" ;J.. [il o II o 10 100 1000 10000
Choine Concentration, ppm, before III Water RInse (b)
Figure 421. Contact angles on BOB etched p(l00) wafers in (a) choline solutions at 35°C and (b) in DI water after conditioning in choline at 35°C followed by a DI water rinse for 5 min. 117
4.4.2. Oxidation in Choline Solutions
The change of the surface state from hydrophobic to hydrophilic was
observed on samples treated with solutions containing greater than 5000 ppm
choline. It was initially speculated that the reasons for improved wettability may
be the adsorption of choline on the surface, surface modification such as oxidation
and roughness, and/or lower surface tension of choline solutions. Figure 4.22
shows the experimental matrix used to decipher the mechanism by which choline
improved the wettabaility of silicon. The first set of experiments was carried out
on samples treated with 10000 ppm (1 wt%) choline for 10 min, DI rinsed for
5 min, dried by nitrogen gas, and exposed to air as a function of time. The oxide
thickness and wettability of these samples were measured. As shown in Figure
4.23, oxide thickness and wettability did not change significantly with air exposure
time after DI water rinse.
Figure 4.24 shows contact angles and oxide film thickness on choline
treated samples that were exposed to air before a DI water rinse. In these tests,
samples were immersed in 1 wt% choline solutions for 10 min, removed, and then
exposed to air for different periods of time before rinsing in DI water. On
samples that were rinsed in DI water immediately after a choline dip, advancing
and receding angles were -64° and 20°, respectively. The oxide thickness on
these samples was 27. ± 4 A. On samples exposed to air for 10 min before DI
rinsing, the advancing angle dropped to 30° and the receding angle decreased to
15 0. The average oxide film thickness on these samples were 33 ± 3 A, 118
01 rinsed Exposed to DI rinsed 01 rinsed 01 rinsed for 5 min air as a fn for 5 min for 5 min for 5 min N2 dried of time N2 dried N2 dried N2 dried
Measured oxide thick. & contact angles in 01 water
Measured oxide thick. & contact angles in 01 water
Figure 4.22. Experimental matrix used to study the interaction of choline with silicon samples. 119
80 40
... 4~ ~:t S 60 35 . r ~ ~ --0- Adv. Angles -< Q '" --0- Rec.. ~gles - ~f .5 Oxide Thickness, A Q) III [/ J2 ..S! 40 t:s:s 30 (,J C) ...... c j rl ~ < Q) ~ I)W" 'C t1S ·sc ~ 0 20 ~~ 25 8 -l~ ~ Y 'r'
o 20 o 30 60 90 120
Air Exposure Time after 01 Water Rinse, min
Figure 4.23. Changes of oxide thickness and wettability as a function of air exposure time after DI water rinse. 120
-0- Oxide Thickness, A Adv. Angles • Rec. Angles 40 80
70 35
~ 60 ~ -< C'CI 30 ~ ~ 50 CI.I a ~u .5 25 40 III CI.I ~ C, Q) C "C 30 < d 20 ~u 5 20 c 0 15 u 10
10 0 0 10 20 30 40 50 60
Air Exposure Time before DI Rinse, min
Figure 4.24. Effect of air exposure after immersion in choline solution on oxide thickness and contact angles in DI water (conditions: 10000 ppm choline, 10 min immersion, p(lOO) wafer). 121 indicating that exposure to air after a choline dip resulted in the thickening of the oxide layer on silicon.
Figure 4.25 shows the effect of dipping time in 10000 ppm (1 wt%) choline solutions on the wettability of and oxide fllm growth on Si wafers. In these tests, samples were DI water rinsed immediately after removal from choline solutions.
After DI water rinsing, advancing angles in water were still quite large, ca. 60°. A comparison of Figure 4.25 with Figure 4.23 shows that air exposure after dipping in choline solution resulted in improved wettability. Interestingly, there are no significant differences in the trends of ellipsometric data for the two cases. X-ray photoelectron spectroscopy was used to probe the surface of choline- treated samples. Figure 4.26 shows Si(2p) peaks in the XP spectra for silicon wafers that were: (a) BOB etched, (b) treated with 10000 ppm (1 wt%) choline for 10 min and rinsed immediately in DI water, and (c) treated with choline and exposed to air for 10 min before a DI water rinse. The spectrum for sample (a) shows only a
Si(2p) peak characteristic of Si at a binding energy of 98.2 eV. Sample (c), exposed to air, is characterized by Si(2p) peaks for Si and Si02• In sample (b), the Si(2p) peak for Si02 appeared to be in the initial stage of development. Interestingly, a nitrogen Is peak was not detected on the samples.
Since there was evidence that air exposure after a choline rinse resulted in
enhanced oxidation, a method was developed to form areas of different degrees of
exposure to air on a sample used in wettability studies. This was achieved by
varying the extent of immersion of the sample in the choline solution in the first 122
SO 100 --0- Oxide Thickness, A • Adv. Angles • Rec. Angles 40 80 ~
< ~ 60 c 6 30 .5 .eu In c,Go) ~ Co ., 20 40 < )( "0 ~
~0 10 20 (,)
O~~~~~~~~~~~II~~~~~TO o 20· 40 60 80 100 120 Choline Treatment Time, min
Figure 4.25. Changes of contact angle and oxide film thickness changes as a function of immersion time in 10000 ppm choline solutions. 123
Si2p
Si2p(SiOZ)
c
b
a
107 104 101 98 95 Binding Energy, eV
Figure 4.26. Si (2p) XP spectra of (a) BOB etched p(lOO) Si, (b) Si treated in 1 wt% choline and rinsed immediately in DI water, and (c) Si "b" exposed to air for 10 min and then DI water rinsed. 124
and second cycle. Figure 4.27 shows the different areas on the sample. Area 1 was immersed in choline and was not exposed to air during multiple cyclings.
Areas 2 and 3 were alternately exposed to choline and air twice and once,
respectively. Area 4 did not contact choline but was always exposed to air during
two successive cycles. In these experiments, different stage speeds were used to
observe the effect of air exposure time during cycling (faster stage speeds resulted
in less exposure time to choline and air).
Figure 4.28(a) shows oxide thicknesses in different areas as a function of
stage speed. At lower stage speeds «64 p.m/sec), area 2 developed a much
thicker oxide coating than the other areas. At higher stage speeds
(>200 p.m/sec), there was no detectable difference in oxide thickness in areas 1, 2
and 3. A comparison of the oxide thickness values for areas 2 and 3 revealed that
repeated alternate exposure to choline and air resulted in significant oxidation of
silicon.
Dynamic contact angles were measured in DI water on repeatedly cycled
samples that were rinsed. Figure 4.28(b} shows contact angles on areas 1 and 2 as
a function of the stage speed during choline treatment. The advancing and
receding angles on area 1 were not a function of the stage speed. However, the
values of the advancing angle on area 2 decreased to nearly those of the receding
angles as the stage speed decreased. The receding angles in both areas 1 and 2
did not change with the stage speed. 125
.. Area 4 ~ Unexposed to Solution r------
Area 3 ~OneCycle
r------
Area 2 ~ Two Cycles
r------
Area 1 ~ Immersed in Choline Unexposed to Air
Figure 4.27. Schematic diagram of a silicon wafer sample used to characterize the effect of repeated exposure to choline solution and air. 126
140 "'!""T'--"-'-----., 80~------_,
lZ0 --0- Area 4 Area 3 --0- Area 2 100 Area 1 -< 6 f 80 .5 Area 1. Adv. ~ 40 Area 1. Rec. C, Area 2. Adv. c ~ 60 < Area 2. Rec. ~ tl o 40 ~ 20 (3 20
a +-r..,.-..,..,..,..-,-...... ,----t o +-Ir-.J...-.-.-.-..,...,.-.-,r-o-I a so 100 1 SO ZOO 250300 o 50 1001 SO 200 250300 Stage Speed, wn/sec Stags Speed, wn/sec
(a) (b)
Figure 4.28. (a) Thickness changes at different areas as a function of the stage speed, and (b) wettability of areas 1 and 2 in DI water after cycling in choline solution. 127
Figure 4.29 shows Si(2p) peaks of area 2 at different speeds of immersion into choline. The intensity of the Si(2p) peak for Si02 increased as the stage speed decreased. In addition, the difference in the binding energies of Si(2p) peaks for oxide and bare silicon increased as the stage speed decreased. These two features reveal that progressive oxidation of Si to Si02 occurred by repeated exposure to choline and air. It is important to point out that a change in the separation of Si(2p) peak for Si and Si02 with thickness has been measured for very thin oxide «20 A) films [4.18].
The effect of cycling on contact angles and oxide thickness was investigated by performing multiple cycle experiments on samples at a stage speed of
64 p.m/sec. Samples were cycled in a 1 wt% choline solution, rinsed in DI water for 5 min, and then the oxide thickness and contact angles were measured in DI water. As can be seen from Figure 4.30, the oxide thickness and contact angles
did not change after the second cycle.
The effect of temperature on wettability in choline solutions was discussed
earlier (Section 4.4.1.6.). Figure 4.31 compares the oxide thickness at 25°C and
35°C as a function of choline concentration. An increase in solution temperature
increased the oxide thickness and resulted in better wettability.
It is should be noted from Figure 4.12(b) that at choline concentrations
above 5000 ppm, receding angles in the second cycle were slightly larger than the
advancing angles. This behavior of 0A < OR was observed on SSP wafers at all 128
SiZp(SiOZ) SiZp
b
a
107 104 101 98 9S Binding Energy, eV
Figure 4.29. XPS Si (2p) spectra obtained on area 2 of Si samples cycled at different stage speeds. (a) 264 11m/sec, (b) 64 11m/sec and (c) 20 11m/sec. 129
100 ~
80 \
\ --<>-- Mv. Angln- \ --0-- Rec:. Angle. r\ \ \ \ \ ~ \ \. 20 , \. /. ~ V o o 2 4 6
Number of Cycles (a)
100
Oxide !ThIckness 80 'r--ts- / V / I " / ~ 40 / ? 20 / ~ o o 2 4 6
Number of Cycles (b)
Figure 4.30. Effect of cycling on: (a) oxide thickness and (b) contact angles of samples in DI water after treatment in 10000 ppm choline solution and DI water rinse. 130
200
150 --0-- at 35°C -< --0- at 2S'C H .6 u 100 ~ 1-"1-' -8 V " ~ S ~ V 50 V ...4~ ~ t-" -~ l- ~ ~ 1== I=i" o r' o 10 100 1000 10000
Choline Concentration, ppm
Figure 4.31. Change of oxide thicknesses at different choline temperatures as a function of choline concentrations. 131 immersion/ emersion speeds investigated. The reason for this apparent autophobic(23) effect can only be speculated. As discussed earlier, at the end of the first cycle, the part of the sample that was immersed in choline is exposed to
air. The maximum exposure time was 3 min at a stage speed of 64 /lm/sec. This
exposure was shown to result in the oxidation to Si02• Perhaps the critical
surface tension of the oxide layer with adsorbed choline is less than that of the
choline solution, resulting in (J A < (JR' In choline solutions, the receding angles
never reached zero but often remained at 10-20° even though advancing angles
reached zero. It appears that the receding contact angles are affected by a
Marangoni effect which arises due to surface tension gradients in a liquid
meniscus. This effect may either contract or expand the meniscus (Le., affect the
apparent contact angle), depending on the direction/sign of the surface tension
gradient. It is known that a Marangoni effect usually occurs when the dissolved
surface active solute has a different volatility than that of the solvent.
In order to investigate the effect of hydrogen peroxide addition on oxide
growth, hydrogen peroxide (1 wt%) was added to choline solutions of different
concentrations. Figure 4.32 shows the oxide thickness measured on samples
subjected to two successive cycles followed by a DI water rinse. There is no
change in oxide thicknesses with increasing choline concentration which indicates
(23%e films formed on the surface by withdrawing the plate from solution or a melt of a long chain RX-type compound would often not be wet by the bulk parent compound. Such films are called auto-phobic. 132
so
40
-< iii en cu 30 c --0-- Oxide Thickness ~ :cu ~ cu 20 '). ';( 1'.... "0 -- 10
o o 10 100 1000 10000
Choline Concentration, ppm, with 1 ~ H202
Figure 4,32. Change of oxide thicknesses treated in choline/peroxide solutions followed by DI water rinse. 133 that hydrogen peroxide passivates the surface initially and prevent oxide growth as a sample is exposed to air.
It appears that the exposure of choline-contacted silicon to air resulted in oxide formation and hence lower contact angles. Thus, if a quick dump rinser with multiple cycles is used for choline rinsing, the wafer wettability can be improved significantly.
4.4.3. Wettability of Wafers in a Commercially Available Choline-Based Solution (SummaClean)
A choline-based solution of electronic grade is marketed by Mallinkrodt
Chemical Co. under the trade name "SummaClean." It is claimed to contain 5000
ppm choline, methyl alcohol and a proprietary surfactant. The pH, redox
potential (vs. SHE) and surface tension of SummaClean solutions were measured
by diluting the as-received chemical to various levels as shown in Figure 4.33.
The pH and redox potential of the as-received solution were 12.5 and 95 mV,
respectively. The surface tension of SummaClean was measured to be
40 dynes/em.
Contact angles of SummaClean solutions on p(lOO) silicon are shown in
Figure 4.34. Mueh lower contact angles were measured in SummaClean than in
pure choline. This is due to the presence of a surfactant in SummaClean. As
shown earlier in Figure 4.20(a), which depicted the effect of surfactant on 134
14 600 ['\ -0- Iii 500 1Z \ ~ >e ,/ ..:: \ 11.1 I--'" 400 6i 10 r\. Y .....~ :z: X 300 a. "II j v ~ 1:1 8 i I- ZOO 15 / f'c.. ~ -0- '~I 6 / ~ "0 100 i=
4 o o 10 100 1000 10000
SUmma-Clean Concentration, ppm (a)
80
70 r.....
"' t'-... 60 " '\ so '" r\'\.
40 \
30 o 10 100 1000 10000
SUmma-CIean Concentration, ppm (b)
Figure 4.33. (a) pH, redox potential (vs. SHE) and (b) surface tension of silicon in Summa Clean as a function of choline concentration. 135
100 1st Adv. 1st Ree. 2nd Adv. 80 2nd Ree.
Xl C) c 60 <.., u ~ <3 40
20+------~----~~~------~------~
04-~~~~~ __~~~ __~~~a_ __~~~ o 10 100 1000 10000
Summa Clean Concentration, ppm
Figure 4.34. Contact angles of diluted Summa-Clean solutions on p(l00) silicon as a function of choline concentration. 136 wettability, Summa-Clean also showed advancing angles and non-zero receding angles in the second cycle.
4.4.4. Wettability of Wafers in TMAH (Tetramethyl Ammonium Hydroxide) Solutions
Figure 4.35(a) shows advancing (OJJ and receding (OJJ contact angles on bare SSP p(l00) wafers in TMAH solutions. In the first cycle, advancing angles remained almost constant up to a TMAH concentration of 1000 ppm and then decreased slightly. The receding angles decreased gradually and were around 20° at TMAH concentrations in excess of 50 ppm. In the second cycle, 0A decreased far more rapidly above 200 ppm TMAH than in the first cycle and reached a value of 10°. Samples that went through two cycles in TMAH solutions showed
similar advancing and receding angles above 400 ppm of TMAH, indicating
modification of the silicon surface by TMAH. Similar receding angles were
measured in both cycles. Figure 4.35(b) displays contact angles measured in DI
water on samples that were subjected to a two-cycle test in TMAH solutions.
These samples were rinsed in DI water for 5 min before measurements were
made. Above 200 ppm of TMAH, the difference between 0A and OR was very
small. As discussed previously (Section 4.2.1), the etch rate of silicon became
significant even in low concentrations of TMAH. The surface roughness values, l\-v and Rnna, measured by AFM were 832.2 and 123.3 A for silicon treated for 137
80 r-.... I-.. ~ htAAiY. ±--t-t+++fl1k::...... "19"n.tt1It--++,~~- 1st Ree.' I. I~ ~ r-.- 2nd AAJy I 60 \1\ = 2nd Rec.
I 40 +-+-''-Hd-+t'I.II:-~--I-+H-H+II--+1 \-\H-Ift+Ht-+-++H+Iti
20 t1-m-rnn-~Ir'\~~im~lW»F11f!Om 14 o+-~~~~~##~~~~~~~ o 10 100 1000 10000 TMAH Concentration. ppm (a)
100 c
eo "'r--.
1[\ ~. DlAAiy. 1\ 1--0-- DI Rec. \ , 1\ I\. I'
20 ~ r\ i' 'f..,
o o 10 100 1000 10000 TMAH Concent., ppm, before DI water Rinsing (b)
Figure 4.35. (a) Advancing (OA) and receding (O.J contact angles on bare SSP p(l00) wafers in TMAH solutions, and (b) the contact angles in DJ water on samples subjected to a two cycle immersion in a TMAH solution followed by a DI water rinse for 5 min. 138
10 min in 600 ppm TMAH. The low contact angles in TMAH solutions above
200 ppm are most likely due to the roughness induced by TMAH solutions.
Silicon surfaces treated with TMAH were observed using an optical microscope. Bare silicon wafers were immersed in various concentrations of
TMAH solutions from 400 ppm to 1 wt% for 10 min and OI rinsed for 5 min.
The conditioned surfaces are shown in Figure 4.36. An island-like rough surface was observed even on samples conditioned in 400 ppm TMAH solutions. A silicon surface etched in 1 wt% choline solution was much smoother than the sample etched in 400 ppm TMAH solution, as may be seen by comparing
Figures 4.36(a) and (d). Figure 4.37 shows the surface roughness of the etched sample scanned by a OEKTAK IIA surface profiler. Above 1000 ppm TMAH, the surface roughness became significant.
4.4.5. Wettability of Wafers in Ammonium Hydroxide Solutions
Advancing and receding angles of aqueous ammonium hydroxide solutions
of different composition on p(100) silicon are shown in Figure 4.38. These
measurements were made at a stage speed of 64 p.m/sec. Up to an ammonium
hydroxide (58%) to water ratio of 0.1 (-9 vol%), the advancing angles in the first
cycle remained high. Receding angles decreased as the NH40H content in solution increased. Advancing angles in the second cycle decreased more rapidly
in solutions containing more than 10% by volume of NH40H. This decrease might be due to the increased surface roughness caused by etching at higher 139
I', '~ '4." I'", .\ " "'"
(a) (b)
',1"' .. 1~1 i::!,<
" . ;';~"'~'~~~';~:' . ~.: " -
(c) (d)
Figure 4.36. Optical micrographs of p(100) silicon etched for 10 min in (a) 400 ppm, (b) 1000 ppm, (c) 10000 ppm TMAH solutions, and (d) 10000 ppm choline solution. The bar scale is 100 IJ.ffi. 140
3000.00
-< 1000.00
b ~ c..~ ~ -1000.00
c
d
-3000.00 iii iii iii ii' ii' iii iii iii iii iii I I I • , i , I • i i 0.00 1000.00 2000.00 3000.00 4000.00
Distance, pm
Figure 4.37. Surface profiles of samples treated in: (a) 10000 ppm, (b) 1000 ppm, (c) 400 ppm TMAH solutions, and (d) as-received polished silicon surface. 141
100~~------,
80 -e-- 1st Adv. 1st Rec. III Q) 60 a 2nd Adv. -=e 2nd Rec. < ... U ell ...e 40 0 U
20
O~)~--~~--~~~~~~~~--~~~ o .1 10 100
Volume Percent of NH40H in Water
Figure 4.38. Wettability of silicon wafers in aqueous ammonium bydroxide solutions. 142
concentrations of NH40H. Figure 4.39 shows the DIC micrographs of wafers subjected to 2 cycles in different NH40H:H20 solutions and DI rinsed for 5 min.
Even for the 0.1% NH40H solution, roughening of the surface (etch rate,
202 A/min) was observed as shown in Figure 4.39(a). An etch rate of 380 A/min was measured for a 17% ammonia solution at room temperature as discussed earlier (Section 4.2.1).
The wettability of silicon by ammonia-peroxide solutions (SC1) was measured as a function of ammonium hydroxide content in a H20 2-H20 solution of 1:5 volume ratio. The addition of ammonium hydroxide to hydrogen peroxide solutions increased the wafer wettability, as shown in Figure 4.40. In the first cycle, the advancing contact angle on silicon was 93 ° in the absence of NH40H, but this angle started to decrease when the volume percent of NH40H exceeded
0.17% (0.01:1:5 SC1). The advancing angle decreased to 33° for a NH40H volume percent of 14.3 (1:1:5 SC1). The decrease of advancing angles at higher volume fractions of NH40H may be attributed to the etching of the silicon surface
on a micro-scale with passivation of the surface by H20 2• Roughening of surfaces
by etching would indeed result in lower contact angles [4.12]. The receding angles
in the first cycle were quite low compared to advancing angles. Even in the
absence of NH40H, the first cycle receding angle was only 30°. This pronounced
hysteresis is most likely due to the oxidation of silicon by H20 2 on immersion into
the solution. It should be mentioned that the presence of H20 2 in NH40H
solutions increases the redox potential and then would promote the passivation of 143
(a) (b)
(c) (d)
Figure 4.39. Optical micrographs of wafers subjected to 2 cycles in (a) 0.1 %, (b) 1 %, (c) 10 % and 50 % NH40H:H20 solutions and DI rinsed for 5 min. The bar scale is 100 tLm. 144
100~~------____~
80
60 i ---0- 1st Adv. ~ 1st Rec. g 2nd Adv• 40 • 8 2nd Rec.
20
.1 1 10 100
VoIwne " of NH40H in 1:5 H202:H20 (a)
40 ---0- DI Adv. .. • Of Rec. ..." 30 ~ is .5 II 20 ., ~ < t J! ~ 10 ~ -~ ~. c§ r J.
o r' ., o .1 10 100 Vaune " of NH40H in 1:5 H202:H20
(b)
Figure 4.40. Wettability of silicon wafers in: (a) alkaline H20 2 solution as a function of NH40H content, and (b) DI water after cycling in ammonia-peroxide solution twice followed by DI water rinsing, 145 surface (Section 4.2.1). In the second cycle, advancing angles became 0°
(complete wetting) when the ammonia content exceeded 0.017% (0.001:1:5 SCI).
Receding angles were also low ( - 10°) but interestingly never reached 0°. It is
noted that Mishima et al. [4.19] reported better cleaning efficiencies and the
formation of much fewer hazes on the silicon when proportions of NH40H ranging from 0.1 to 0.5 were used in SC1 solutions.
Samples treated with an alkaline hydrogen peroxide solution were rinsed in
DI water, nitrogen dried, and then characterized for their wettability. Advancing
and receding angles remained low (~200) and constant on samples pretreated
with solutions containing ammonium hydroxide up to a volume fraction 7.7%
(0.5:1:5 SCI). Above this volume fraction, advancing angles decreased to values
less than the receding angles, but receding angles did not change. The decrease
of advancing angles from 20 to 10° on samples pretreated in high ammonia
containing solutions indicates the presence of surface roughness on these samples.
The oxide thickness on samples immersed in solutions with different
NH40H concentrations was measured to be 13 ± 1 A, and was independent of
NH40H content and air exposure time. This finding is similar to the result
observed with choline/peroxide solutions. Perhaps the presence of H20 2 passivates the surface and limits the oxide growth. 146
4.5. Scanning Tunneling Microscopy (STM)/Atomic Force Microscopy (AFM) Studies
STM/AFM techniques were used to characterize the effect of alkaline
solutions on surface roughness. Figures 4.41(a) and (b) show STM and AFM
micrographs of polished and unpolished sides of p(tOO} wafers after a 30 sec BOB
etch. The unpolished surface was 500-600 times as rough as the polished side
{l\,-v of 1250 Avs. 2 A}.
Figures 4.42(a) and (b) show STM micrographs of surfaces treated with
0.1:1:5 and 1:1:5 SC1 solutions for 10 min. To make the surface conductive, these
samples were HF (49% }-etched for 10 sec prior to measurement. Within the area
investigated, Rms increased with NH40H concentration for the two SC1-treated-
samples. A flat and featureless surface was observed for the 0.1:1:5 sample. In
SC1 solutions, the reduction of the ammonium hydroxide proportion from 1 to 0.1
decreased the surface roughness (Rnns) from 6.4 to 0.8 A. Table 4.3 summarizes
the roughnesses measured by STM/ AFM.
Table 4.3. STM/ AFM Data on Roughness of Silicon Surfaces Treated with Various Alkaline Solutions
p(100) Silicon Wafers ~-v (A) Rnns (A) Polished side, BOB etched for 30 sec 2.0 0.2 Unpolished (back) side, BOB etched 1250 - for 30 sec
SCt 0.1:1:5, NH4OH:H20 2:H2O 7.9 0.8
SC1 1:1:5, NH4OH:H20 2:H2O 52.3 6.4 Choline, t wt% (10000 ppm) 200 20 147
(a)
(b)
Figure 4.41. (a) STM micrograph on polished and (b) AFM micrograph on unpolished p(100) wafer after 30 sec BOE etch. 148
-...... ~. • • t- ,_. -'~~t~~~;~;..:::
"- ••' '4~4 • - .~ ;~; ,:~':--:----.- ....,..,."'!
(a)
(b)
Figure 4.42. STM micrographs of: (a) 0.1:1:5 and (b) 1:1:5 SCI-treated samples (500 x 500 nm scanning area). 149
As discussed previously, the exposure of choline-treated samples to air increased the wettability. In order to check whether the improved wettability is related to surface roughness, two samples with different degrees of exposure to air after choline treatment were studied using AFM. The first sample was treated with 1 wt% choline for 10 min, exposed to air for 10 min and DI water rinsed.
The second sample was treated with choline and rinsed in DI water immediately before exposing to air. Then samples were observed using an AFM in air as shown in Figure 4.43. The peak-valley and root mean square roughnesses of both samples were in the neighborhood of 200 and 20 A, respectively. This indicates that roughness is not the factor which increases the wettability of silicon samples exposed to air after choline treatment. Chemical oxidation of samples in choline is the main factor producing better wettability as discussed previously.
4.5.1. Effect of Addition of H20 2 and Surfactant Roughness Hydrogen peroxide and a non-ionic surfactant (NCW 601A) were added to
5000 ppm choline solutions to observe the effect of these chemicals on surface
roughness. Figure 4.44 shows samples treated with 5000 ppm choline with and
without the addition of H20 2 and surfactant. In plain choline solutions, the
silicon surface was rendered rough with ~-v and Rrms values of 180.9 A and 22.4 A, respectively. 150
(a)
(b)
Figure 4.43. AFM micrographs of 10000 ppm choline-treated silicon samples followed by: (a) 10 min air exposure and OI water rinse and (b) DI water rinse only (no air exposure). 151
(a)
(b)
Figure 4.44. AFM micrographs of 5000 ppm choline-treated silicon samples for 10 min: (a) without the addition of H:P2 and surfactants and (b) with the addition of 200 ppm H20 2• 152
(c)
(d)
Figure 4.44 (continued). (c) with the addition of 1 wt% (10000 ppm) H20 2 and (d) with the addition of 200 ppm non-ionic surfactant. 153
The addition of 200 ppm H20 2 to choline reduced the values of l\,-v and
Rnna to 48.6 and 6.4 A, respectively. Increasing the H20 2 addition to 1 wt% further reduced the surface roughness. Also, when 200 ppm of the non-ionic
surfactant was added to the choline solution, the surface roughness was similar to
that of a sample treated with a 200 ppm H20 2 solution. Table 4.4 summarizes the roughness values of these samples.
Table 4.4. Roughness of Samples Treated in 5000 ppm Choline for 10 min With and Without the Addition of H20 2 and Non-ionic Surfactant
Samples treated in 5000 ppm choline 180.9 22.4
5000 ppm choline with 200 ppm H20 2 48.6 6.4 5000 ppm choline with 32.8 4.2 1 wt% (10000 ppm) H20 2 5000 ppm choline with 40.1 5.6 200 ppm non-ionic surfactant
4.5.2. Effect of Addition of H20 2 and Surfactant to TMAH Solutions The etch rate of TMAH solutions was much greater than that of choline at
equivalent concentrations. Even in the 600 ppm TMAH solution, the l\-v and Rnna values were measured to be 832.2 and 123.3 A, respectively, which were
much higher than values measured in the 1 wt% (10000 ppm) choline solution.
Adding H20 2 to TMAH significantly decreased the surface roughness. The
addition of 200 ppm H20 2 to TMAH provided a smoother surface than the 154 addition of 200 ppm non-ionic surfactant. Figure 4.45 shows samples treated with
600 ppm TMAH solutions with and without the addition of H20 2 and surfactant. Table 4.5 summarizes the roughness (AFM) of these TMAH-treated samples.
Table 4.5. Roughness of Samples Treated in 600 ppm TMAH for 10 min With and Without the Addition of H20 2 and Non-ionic Surfactant
Samples treated in 600 ppm TMAH 832.2 123.3
600 ppm TMAH with 200 ppm H20 2 14.3 1.8 600 ppm TMAH with 13.5 1.8 1 wt% (10000 ppm) H20 2 600 ppm TMAH with 20.1 2.8 200 ppm non-ionic surfactant
4.6. Wettability of Wafers in IPA and Its Significance in Particle Removal
4.6.1. Characterization of IP A Solutions
The surface tension of IP A/water solutions is plotted in Figure 4.46 as a function of the solution composition. The surface tension of water decreases
rapidly with IPA additions up to 25% by volume. Beyond this IPA concentration,
the surface tension decreases slowly and approaches that of pure IP A
Because of the IPA molecular structure, the interface between a water-IPA
solution and air will be enriched in IPA molecules. From the measured surface
tension values of water-IP A solutions, the Gibbs surface excess (r) of IP A at the
solution/air interface was calculated as a function of the solution composition 155
(a)
(b)
Figure 4.45. AFM micrographs of silicon samples treated for 10 min in 600 ppm TMAH solutions: (a) in the absence of H20 2 and (b) with the addition of 200 ppm H20 2• 156
(c)
(d)
Figure 4.45 (continued). (c) with the addition of 1 wt% (10000 ppm) H20 2 and (d) with the addition of 200 ppm non-ionic surfactant. 157
80
I 70
c 60 .2 en c ~ q Q) SO u ~ 5t :I en 40 ~ ~ l~ 30 ~ ~ - 20 o 20 40 60 80 100
Volume Percent IPA in Water
Figure 4.46. Surface tension of IP A-water solutions as a function of solution IP A content. 158 using equation (4.1). These surface excess values are plotted in Figure 4.47. It may be seen that the surface excess of IP A is higher at lower solution IP A contents and has a maximum value of 8.5 x 10-10 moles/cm2 around 25 volume percent IP A As will be shown later, the higher surface activity of more dilute solutions plays a key role in particle removal. It should be mentioned that surface excess values never reach zero.
4.6.2. Wettability of Wafers in IPA/Water Solutions
The wetting characteristics of hydrophobic and hydrophilic wafer samples in IPA/water solutions of different compositions are shown in Figure 4.48. These data were taken for an immersion/emersion speed of 64 p.m/sec. For hydrophobic wafers, the contact angle values, similar to surface tension data, decreased steadily up to a solution containing 25% IP A by volume. Interestingly, at the stage speed used, the advancing angles became smaller than the receding angles for IPA volume fractions greater than 0.25. The reasons for this apparent anomalous behavior may be attributed to the expansion or contraction of the meniscus due to the Marangoni effect as explained earlier with choline solutions
(Section 4.4.2). In the case of hydrophilic wafers, the contact angles were quite low in all solutions. Complete wetting was observed during immersion as well as emersion in solutions with 9 to 67% IPA. 159
10
~ 0 ..0 8 >C ~ N E ) \ ...... 7 III 6 G) \ ~ 1 vi 4 ?f \ III G) I U 0 >C \ U.I J G) \ u 2 ,U ~:;, f'\ en ~ o :--0- " o 20 40 60 80 100
Volume Percent IPA in Water
Figure 4.47. Surface excess of IPA at the air/IPA-water solution interface. 160
l00T------r------r-----,,-----,------,
---0-- Hydrophilic Adv. Angles 80+-~---T----~ Hydr~philic Rec. ":ngles • Hydrophobic Adv. Angles Hydrophobic Rec. Angles III G) ~ 60~r---~------_r------+_----~----__; c < ~ 40+---~-+--~--+_-----+------T_----_; ~
o+-~~~~~~==~~_+~~~ o 20 40 60 80 100
Volume Percent IPA in Water
Figure 4.48. Contact angles on hydrophobic and hydrophilic wafers as a function of IPA content in water. 161
In a commercially-developed IP A process mentioned earlier (Section
2.1.1.3), an IPA layer, under slight pressure, is used to replace the wafer/water interface with a wafer /IPA interface. Because of the complete miscibility of IP A
and water, it is likely that the IP A layer develops a concentration gradient and the wafer is actually exposed to a series of solutions with different IP A contents as
IP A is pushed past it. In order to study the nature of interface between IP A and
water, IP A and IPA/water solutions of different compositions were gently
contacted with water. A dye (ethidium bromide) was dissolved in IPA to make
the interface more visible.
The addition of various concentrations of IP A solutions (9% to 100%) to
water initially introduced a rapid flow at the interface as well as inside the
aqueous solution. Since the addition of IF A to water lowers the interfacial
tension, an inhomogeneous distribution of IP A within the interface may result in
the creation of local tension gradients. The highest tension occurs in those
regions where the IPA density is lowest; thus, the tensile restoring force acts in
the opposite direction to the IF A density gradient. The ensuing fluid motion
arising from the interfacial tension gradient is known as the Marangoni effect
[4.20).
When certain compositions of IPA solutions were added to water, a sharp
interface started to form between the IPA solution and water. The top layer was
stable for several hours with a clear boundary between the water and IP A 162 solution. To photograph the formation of a distinct top layer, a glass optical cell was used. First, 50 ml of water was added into the cell, and then 20 ml of
different concentrations of IPA solutions were slowly added on top of the water using the wall of the optical cell. Among the IP A solutions prepared for this
experiment, those solutions containing less than 9% IP A did not form a distinct
layer on the top of water. Other solutions, 25%, 50%, 70%, and 100% IP A,
initially formed a very stable IP A layer. The lower the IPA content of the
solutions, the greater the amount of mixing between the water and IP A solutions.
Figure 4.49(a) shows a photograph of the IPA layer formed on water using 100%
IP A in a cell. After 20 hrs, the boundary between the water and IP A solutions
became diffuse as shown in Figure 49(b). After 72 hrs, complete mixing between
the water and IPA solutions was observed. This indicates that IP A can form a
distinct layer initially, but given sufficient time, IPA in the top layer will continue
to diffuse into the water-rich bottom layer.
Attempts were made to measure the surface tensions of the top IP A layer
and bottom water layer immediately after the formation of a distinct top layer.
The surface tension of the top layer was measured by immersion/withdrawal of a
pre-cleaned glass slide without interrupting the IPA/water interface. Next, the
top layer was removed very carefully using a syringe, and then the surface tension
of the bottom layer was measured. Figure 4.50 shows the surface tensions of the
top and bottom layers of solutions when various IP A solutions were added to the 163
(a) (b)
Figure 4.49. The change of IP A layer formed by slowly placing IP A on the top of water in a glass optical cell as a function of time: (a) 0 hr and (b) 20 hrs later. 164 water. The surface tension of the top layer did not change significantly after placing IP A solutions on top of the water, however, the surface tension of the bottom layer decreased to around 45 dynes/cm over the range of IPA contents investigated. From the measured surface tension values shown in Figure 4.50, the
composition of the top and bottom layers was calculated. The IP A compositions
of the bottom layer were in the range of 9 to 14%. The IPA content of the top
layer was 2 to 20% less than the volume percent of IPA in the solution added
onto water.
To understand the impact of the mixing at the solution interface, the
wettability of the same hydrophobic wafer sample was measured in solutions of
successively increasing IPA content; the results are displayed in Figure 4.51. It
may be noted that the methodology used in this test was different from that used
to obtain the data in Figure 4.48, where a different sample was used for each test.
While the trend is the same in both types of tests, the successive immersion tests
gave lower advancing and receding angles due to the presence of IP A films at the
solution/air interface.
During the successive immersion/ emersion tests with hydrophobic surfaces,
several interesting observations were made. When a sample was immersed and
then emersed from solutions of less than 25 volume percent or more than
70 volume percent IP A, it dried without leaving any droplets on the surface.
When samples were withdrawn from solutions containing 25 to 70 volume percent 165
80
I) 70 ~ Top layer E u 1\ ~ Bottom layer- III "II) 60 \ ~ \ C 0 \ en 50 ...... n..: c ~ 1--0.. ~ - CI) \ -- -~ u 40 1", ~ \ en~ 30 li ~ r--.o- --..n..- 20 . ~ o 20 40 60 80 100
Volume " of IPA in top layer
Figure 450. The surface tension of IPA solutions and water layer formed by slowly placing the IP A solutions on the top of water. 166
100
)
80 1\
~ 0- Adv. Angles III • • Q) ~~ .~ --a- Rec. Angles c;, 60 c < \"\ ~ U ell ~~ ~ 40 c u0 \l \ 20 /J \ -.T_ ~y -f. o 5---...., ~...., -rI...., ..; o 20 40 60 80 100
Volume Percent IPA in Water
Figure 4.51. Contact angles on hydrophobic wafers as a sample is successively contacted with solutions of increasing IP A content. 167
IPA, droplets were initially seen on the surface but they eventually dried up.
Also, the size of the droplets decreased with increasing IPA content in this range.
This indicates that the IPA solutions larger than 25% and less than 70% may leave particles and hazes as the droplets dry during rinsing. Figure 4.52 shows the micrographs of droplets on hydrophobic wafers at different IPA/water contents.
4.6.3. Particle Removal on Wafers Using IPA-Based Solutions
4.6.3.1. Particle Deposition
Particles were deposited onto silicon samples using two different techniques. The first technique was based on electrophoretic deposition.
Polystyrene (PSL) and glass particles were deposited on silicon using this method.
As shown in Figure 4.53(a), a rather uniform deposition of particles was achieved using this technique. However, it was found that the wafer surface was oxidized during the deposition process. The water contact angle (advancing) on the sample was 10° at the end of the deposition, indicating that the sample became hydrophilic during the deposition process. Figure 4.54 shows the change of
contact angle and oxide thickness as a function of deposition time.
An aerosol method was also used to deposit PSL, glass, alumina and silicon
particles on samples. Both hydrophilic and hydrophobic samples were coatable by
this technique. Representative photographs of hydrophilic and hydrophobic
wafers coated by the aerosol technique are shown in Figures 4.53(b) and (c), 168
(a) (b)
(c)
Figure 4.52. Photomicrographs (at 5X) of droplets on hydrophobic wafers at different IPA content, at (a) less than (» 25 vol.% IPA and more than (.$.) 70 vol.% IPA, (b) 25 vol.% IPA, and (c) 50 vol.% IPA. 169
(a) (b)
(c)
Figure 4.53. Optical micrographs of particles deposited by (a) electrophoretic method, (b) aerosol spray method on hydrophilic wafers and (c) aerosol spray method on hydrophobic wafers. 170
100 50 :,..0- t'1~ C) I--'" 1"':" 1/ ~ 80 l,...-' 40 I--'" ltf" I-"" I/" -< I· ~ Adv. Angles Ii 60 V· • ,.!. • I I I I , I 30 cu I" Rec. Angles 1I I -I I I I .... .5u Jt. Oxide Thickness, A - ~ ~ ~ 40 ~ ~ 20 cu ['\ :g II I ?--......
Anodization Time, min
Figure 454. Change of the oxide thickness with contact angle as a function of deposition time. 171 respectively. Much fewer particles were deposited on the hydrophobic surface.
Figure 4.55 summarizes various deposition methods applied to contaminate hydrophilic as well hydrophobic wafers.
4.6.3.2. Particle Removal on Hydrophilic Wafers
As mentioned earlier, electrophoretic deposition of PSL particles onto wafer samples resulted in the hydrophilization of the samples. A photomicrograph of a sample in the as-deposited condition is shown in Figure
4.53(a). Coated samples were then separately rinsed (immersed and then emersed) in DI water, 2.4% IPA solution, 9% IPA solution, 25% IPA solution,
50% IPA solution, 87% IPA solution or 100% IPA at a stage speed of 64 p.m/sec.
The appearances of the various samples after rinsing are shown in Figure 4.56.
During DI water rinsing (Fig. 4.56(a», particles were removed from many areas but the removed particles formed agglomerate strings (lines) in the direction of the sample movement. Rinsing in pure IPA (Fig. 4.56(f) was not at all effective in removing particles. Of the six IP A/water solutions investigated, solutions with
. less than 25% IPA appeared to be effective in removing particles.
The results obtained from rinsing a contaminated hydrophilic wafer successively in seven solutions of different IP A contents (DI water ... 9% IPA ...
25% IPA ... 50% IPA ... 67% IPA ... 87% IPA ... 100% IPA) are displayed in
Figure 4.57(a). This rinsing scheme appears to be very effective in removing PSL 172
Hydrophobic Wafers Hydrophilic Wafers I I II II II Electrophoretic Electrophoretic Air-Gun Air-Gun Deposition Deposition Spray Spray in DI Water inIPA
Hydrophilic Hydrophobic Hydrophobic Hydrophilic Wafers Wafers Wafers Wafers w/ Particles w/ Particles w/ Particles w/ Particles
:Can Control Hydrophilicity with Deposition Time
Figure 4.55. Particle deposition methods applied to contaminate wafers. 173
(a) (b)
(c) (d)
Figure 4.56. Optical micrographs of samples rinsed once in (a) DI water, (b) 2.4% IPA, (c) 9% IPA and (d) 25% IPA. The bar scales are 100 J.Lm. 174
(e) (f)
Figure 4.56. (Continued) Optical micrographs of samples rinsed once in (e) 50% IPA, and (f) 87 and 100% IPA. The bar scales are 100 J-Lm. 175
(a)
(b)
Figure 4.57. Appearance of samples successively rinsed in solutions increasing IPA content: (a) from DI water to pure IPA and (b) 50% IPA to pure IPA 176 particles. A rinsing scheme starting with 50% or more IPA was not very effective in removing particles as shown in Figure 4.57(b). This may be related to the surface excess of IPA at the solution/air interface which was shown to be larger at lower «25%) IPA contents than at higher IPA contents.
Repeated cycling for seven times in the same solution was carried out to
ascertain that the solution sequence discussed earlier is indeed important. The
results obtained are shown in Figure 4.58. In solutions containing 50% or more
IP A, repeated rinsing had minimal effect on particle removal. Rinsing in a 9%
IPA solution or 25% IPA solution seven times removed a substantial amount of
particles. However, it must be mentioned that the cleanliness of samples
subjected to successive cycling in a series of solutions with increasing IP A content
was remarkably better than that samples subjected to repeated cycling in anyone
solution.
From the experimental data obtained, a mechanism for the enhanced
particle removal from solutions of low IPA content will now be proposed.
Schematic sketches of the immersion and emersion stages of a contaminated
hydrophilic wafer are shown in Figure 4.59. It may be noted that the solution-air
(vapor) interface is populated by IPA molecules and the surface excess (or crudely
speaking, surface concentration) of IPA molecules at the solution/air interface is
higher in solutions of low IPA content than in solutions of high IP A content. As 177
(a) (b)
(c) (d)
Figure 4.58. Optical micrographs of samples subjected to seven successive rinse in: (a) DI water, (b) 9% IPA, and (c) 25% IPA. The bar scales are 100 /.Lm. 178
(e)
Figure 4.58. (Continued) Optical micrographs of samples subjected to seven successive rinse in: (d) 50% IP A and (e) 87 and 100% IP A. The bar scales are 100 jLm. 179
Hydrophilic I b Particle Wafer U Vapor
(a) (b) ~.
(c) (d)
Figure 459. A schematic diagram illustrating the mechanism for enhanced particle removal from solutions of low IPA content: (a) particle on a substrate before contacting IP A-water solution, (b) particles at the air lIP A-water interface, (c) particles removed by surface pressure difference at the air/solution interface and (d) substrate after rinsing. 180 the particle laden wafer penetrates the solution/air interface, the distribution of surface IPA molecules may be altered.
A scenario shown in Figure 4.59(b) is likely that the surface excess of IP A becomes higher at the part of the interface that is near the wafer (region I) than at the part far away (region TI). If this happens, the surface tension of region I would decrease slightly and that of region TI would increase slightly. This would cause a difference in surface pressure, '1',(24) between regions I and TI, with the gradient in the direction of I to n. The force on the particle due to this surface pressure gradient may be responsible for particle removal. It should be mentioned that the immersion/ emersion speed should be slow enough for the particles to experience this force.
The removal of glass and alumina particles deposited onto an oxidized wafer was extremely difficult. However, as shown in Figure 4.60, hydrophobic silicon particles deposited on a hydrophilic wafer were removed in DI water, and particles in 9% IP A solutions were removed even more efficiently. For PSL particles deposited onto an oxidized wafer, some particle removal was noticed in
DI water, but the extent of removal was substantially improved by the addition of small amounts of IP A to water.
(24)Surface pressure, '1', is defined as the difference between the surface tension of a pure solvent/vapor interface and that of a solution/vapor interface; in this case 'Ywater - 'YwaurIlPA' 181
(a) (b)
(c)
Figure 4.60. The removal of silicon particles deposited on hydrophilic wafers (a) as deposited, in (b) DI water and (c) 9 % IPA solution. The bar scales are 100 f..£m. 182
4.6.3.3. Particle Removal from Hydrophobic Wafers
Electrophoretic deposition of particles on wafers in water resulted in substantial oxidation of the surface. In order to prevent oxidation during deposition, pure IP A was used instead of DI water as the liquid medium for the electrophoretic deposition. Uniform deposition of PSL particles was observed on bare silicon wafers after 10 min. Deposited particles in pure IP A were removable by rinsing in DI water or 9% IPA solution as shown in Figure 4.61. However, in both these media, the extent of particle removal was much less than that observed for the samples contaminated from DI water. The advancing and receding angles of water on silicon samples anodically held in IPA for 10 min were 83° and 51°,
respectively. These values indicate that wafer hydrophobicity was only marginally
decreased during the deposition in IP A. Thus, it appears that PSL particle
removal from hydrophobic wafers is inefficient.
The effects of the nature of surface charge on particles and particle type on
particle removal from hydrophobic wafers were studied by depositing alumina and
glass onto hydrophobic wafers. Glass is negatively charged, while alumina is
positively charged in DI water. Alumina (1.4-4.6 #Lm) and silica (1.9 #Lm) were
dispersed in DI water at a concentration of 0.01 mg/100 m1 of water with a
tOO ppm nonionic surfactant (NCW601A). Particle depositions were performed
using an air gun on hydrophobic silicon samples. The contaminated samples were 183
(a) (b)
(c) (d)
Figure 4.61. The removal of PSL particles which were electrophoretically deposited in IPA for 10 min: (a) as deposited, (b) rinsed in DI water, (c) rinsed in 9 % IPA solution, and (d) in 100% IPA. The bar scale is 100 JLm. 184 rinsed in DI water and a 9% IPA solution at a speed of 64 p.m/sec. Neither repeated nor sequential rinsing resulted in removal of alumina and glass particles.
Two deposition methods were developed to observe the effect of surface hydrophobicity on particle removal. In the first method, PSL particles were first deposited using an air gun spray, and then the anodization of particle-deposited samples was performed for 10 sec, 20 sec, 30 sec, 2 min, 5 min, and 10 min at a potential of 6 V / cm. Water contact angles and oxide thickness on these samples were identical to those shown in Figure 4.54. When these samples were rinsed once in 9% IPA solutions at a speed of 64 p.m/sec, no particle removal was observed. Multiple cyclings and stage speed in the range 20 p.m/sec to
264 p.m/sec were not effective in removing particles from the surfaces. This indicates strong bonding between PSL particles and the hydrophobic surface. The anodization of the wafer surface after air gun deposition did not modify the strong bonding between the substrate and particles.
In the second method, wafers were anodized and then particles were deposited using the air gun spray. Even for a sample anodized for 10 sec (contact angle = 67°), slight particle removal was noticed in the 9% IPA solution.
Hydrophilic samples prepared by prolonged anodization were more amenable to cleaning in the same solution, as shown in Figure 4.62. It is clear that the increase of wafer wettability increased the removal force acting on the particles and resulted in better removal of particles. 185
(a) (b)
(c) (d)
Figure 4.62. Particle removal on changing the hydrophobicity of silicon wafer surfaces (a) as deposited, (b) anodized for 10 sec, (c) anodized for 2 min, and (d) anodized for 5 min and 10 min. The bar scale is 100 J.Lm. 186
The effect of particle hydrophobicity on the efficiency of removal was investigated by depositing hydrophobic silicon particles. As shown in Figure 4.63, silicon particles deposited on hydrophobic wafer samples were slightly removable in both DI water and 9% IP A solutions. This indicates that hydrophobic particles adhere more weakly on silicon wafers than hydrophilic particles and are removable during rinsing in water and 9% IP A. Table 4.6 summarizes particle removal on both hydrophobic and hydrophilic wafers. It is likely that for silicon
Si02-coated substrates, the contact angle between liquid and particle plays an important role in removing particles from substrates. The best combination for
Table 4.6. Particle Removal in Hydrophobic and Hydrophilic Wafers
Particle Type Hydrophobic Wafer Hydrophilic Wafer Hydrophilic Particles No significant removal No significant (Glass, Alumina) removal Hydrophobic Particles PSL-no significant Significant removals (PSL, Silicon) removal Si-slight removal Positively Charged Particles No significant removal No significant (Alumina in DI Water) removal Negatively Charged PSL, Glass-no significant PSL, Si-significant Particles removal removal (PSL,Silicon, Glass in DI Si-slight removal Glass-no significant Water) removal 187
(a) (b)
(c)
Figure 4.63. The removal of silicon particles deposited on hydrophobic wafers; (a) as deposited, (b) rinsed in DI water, and (c) rinsed in 9% IPA solution. The bar scales are 100 J.l.m. 188 significant particle removal was the hydrophobic particles (PSL, silicon) and hydrophilic wafer.
Leenaars' work [4.21] with oxide contaminants deposited onto hydrophilic wafers stressed the importance of low immersion speeds (Le. 3 p.m/s) and the relative wettabilities of the substrate and particles in DI water. He found that
hydrophilic oxide (Si02) particles deposited onto a hydrophilic wafer were
extremely difficult to remove in DI water. However, hydrophobic particles (Ti02,
a-Fe203 and silylated Si02) deposited onto a hydrophilic wafer were removable in DI water.
4.6.4. Theoretical Analysis of Adhesion and Removal of Particles
For particles smaller than about 10 ",m, adhesion of particles to a surface is
primarily caused by van der Waals forces. For particles deposited on silicon
wafers, the following equation can be used [4.22]:
Fa = - AR/6H2 (4.3)
where Fa is the London-van der Waals force, A is the Hamaker constant, 'R' is
the particle radius, and 'H' the distance of the closest separation between the
particle and the substrate.
The Hamaker constant must be calculated to obtain the adhesion force, FB)
of particles to a substrate. The effective Hamaker constant, Al32 for the
interaction of materials 1 and 2 embedded in a medium 3, can be calculated from 189 the Hamaker constants of the individual materials using the following equation
[4.23]:
(4.4)
Table 4.7 summarizes the Hamaker constants, ~i' for each material used in the experiment.
Table 4.7. Hamaker Constants for Various Materials Taken from Ref. [4.23]
Materials All' x 102° J Si 25.6
Si02 50 PSL 6.3
Al20 3 15.5 Glass 8.5 Water 4.38
Figure 4.64 shows a spherical particle adhering to a substrate during the passage of a phase boundary between liquid and air. For a particle adhering to a substrate, the interface close to the particle is no longer horizontal but has a contact angle, a, between the liquid and substrate. The contact angle, fJ, between the liquid and the particle depends on the hydrophobicity /hydrophilicity of the 190
Substrate
(a)
Substrate
(b)
Figure 4.64. Schematic diagram showing a spherical particle adhering to a substrate during the passage of a phase boundary: (a) a < goo and (b) a > 9 2 0 F'Y.max = 2'11"R'Ylg sin (9/2) COsa for a < 90 (4.5) and for a > 900 (4.6) When F'Y.max is divided by 2'11"R'Ylg' equations (4.5) and (4.6) become a function of (J and a only: for a < 900 (4.7) and for a > 900 (4.8) During the immersion and withdrawal of samples in solutions, the particle removal is possible only when F 'Y,max is greater than IFa I. Table 4.8 shows the calculated values of Am and IFa l/2'11"R'Ylg for different particle-substrate combinations for a "H" value to be 4 A (4 X 10-8 cm) [4.21]. 192 Table 4.8. Calculated Values of AI32 and IFal/211'R'Ylg Where 'Ylg of Water = 72.6 dynes/cm''YI,. of 9% IPA Solution = 43 dynes/cm Materials, Am IFal /211'R'Ylg' IF a I/211'R 'Ylg' Particles/Medium/Substrate Water 9%IPA Si/water /Si02 14.76 0.33 0.56 Si/waterlSi 8.82 0.20 0.33 PSL/water /Si02 2.09 0.047 0.076 PSL/water/Si 1.25 0.028 0.048 Al20 3/water/Si02 9.18 0.2 0.34 Al20 3/water lSi 5.46 0.12 0.20 4.6.4.1. Hydrophilic Particles on Silicon and Si02-Coated Wafers In the case of hydrophilic particles such as glass and alumina on a 0 0 hydrophilic wafer, Cl < 20 and 0 < 20 in DI water; in the 9% IPA solution, Cl and 0 are less than 100. For hydrophilic particles on a hydrophobic wafer, 65° < 0 0 Cl < 90 and 0 < 20 in DI water, depending on the advancing or receding interface. In the 9% IPA solution, 500 < ex < 650 and 8 < 100. Figure 4.65 shows IFoy,maxl/211'R'Ylg and IFal/211'R'Ylg in DI water and the 9% IPA solution for hydrophilic particles on bare and Si02-coated wafers as a function of 8. Curved lines are IF'V,maxl/211'R'Ylg at different Cl values. The straight lines are IFal/211'R'Ylg for particles on hydrophilic and hydrophobic wafers. When 8 is less than 200, IFal is much larger than IF 'V,max I for glass and alumina particles on silicon and 193 01 Water/Glass 0.4 0.3 DI) ~ D: tll 0.2 -~ 0.1 GIassISIOl GlassfSI a= 890 0 0 20 40 60 80 100 e (a) 9 % IPAlGlass 0.6 0.5 0.4 DI) ~ ! 0.3 -~ 0.2 0.1 0 0 20 40 60 80 100 8 (b) Figure 4.65. Comparison of IF,,maxl/21rRy,,-and IF.I/2",RYIg as a function of 8 at different a values, (a) glass particles in DI water, and (b) glass particles in 9% IPA The curved lines are IF,,max1/21rRYlg at different a values and the straight lines are IF. 1/2". R Ytr 194 01 Water/Alumina 0.6 O.S 0.4 III:t ....t!I 0.3 ~ 0.2 0.1 0 0 20 40 60 80 100 (c) 9 % IPAIAlumina 0.6 O.S 0.4 I III: ....t!I 0.3 ~ 0.2 0.1 0 0 20 60 80 100 e (d) Fig. 4.65 (continued). (c) alumina particles in DI water, and (d) alumina particles in 9% IPA solutions. 195 SiOz-coated wafer surfaces in DI water and 9% IP A solutions. This is the reason glass and alumina particles on wafer surfaces cannot be removed. Figure 4.66 shows IF or.max I/2",R'Ylg and IFa I/2",R'Ylg for PSL particles on both hydrophilic and hydrophobic wafers at different values of a and 6. Since a is less than 25° for a hydrophilic wafer, it is not possible to remove PSL particles using DI water if 6 is less than 25°. For a hydrophobic wafer, 600 < a < 89°, PSL removal will not occur in DI water if 6 is less than 30°. It might be recalled that PSL particles on a hydrophilic wafer were removable slightly in DI water, whereas PSL particles on a hydrophobic wafer were not removable in DI water. For measured a values, the contact angle, 6, should be 25° < 6 < 30° in order to satisfy the experimental observation in DI water. In 9% IPA solutions, a and 6 are smaller than in DI water. Smaller 6 would make it difficult to remove PSL particles from hydrophilic and hydrophobic wafers. However, experimental results showed better removal of particles in 9% IPA than in DI water. This indicates that surface tension effects, as given by equations (4.5) and (4.6), are not sufficient for the removal of deposited particles. 4.6.4.2. Hydrophobic Particles on Silicon and Si0z-Coated Wafers For hydrophobic particles such as silicon on hydrophilic wafers, a is less than 25° but 6 is near 90° in DI water. In 9% IPA solutions, a is near zero and 6 is less than 65°. For silicon particles on a hydrophobic wafer, CIt has values 196 01 Water /PSL 0.4 0.3 a:~ ....!I 0.2 cl" 0.1 0 0 6 (a) 9 % IPA/PSL 0.6 0.5 0.4 a:~ ....!I 0.3 cl" 0.2 0.1 PSUSI02 PSUSI 0 0 20 40 60 80 100 e (b) Figure 4.66. Comparison of IF,.maxI/211"RYI and IF.1/211"RYIg as a function of e at different a values, (a) PSL particles in Dr water, and (b) PSL particles in 9% IPA The curved lines are IF"maxI/21rRYlg at different a values and the straight lines are IF.1/2wRYlg" 197 between 65° and 89° and 0 is near 90° in DI water. The advancing and receding contact angles of particles and substrates in DI water and 9% IPA solutions are tabulated in Table 4.9. Table 4.9. Contact Angles on Particles and Substrates in DI Water and 9% IPA Solutions In DI Water In 9% IPA Particle /Substrate OtAO OtRO eo OtAO OtRO eo Glass/Si 89 60 <20 65 50 <10 Glass/Si02 <25 <20 <20 0 0 <10 Alumina/Si 89 60 <20 65 50 <10 Alumina/Si02 <25 <20 <20 0 0 <10 Si/Si 89 60 89 65 50 65 Si/Si02 <25 <20 89 0 0 65 PSL/Si 89 60 25<, <30 65 50 <25 PSL/Si02 <25 <20 25<, <30 0 0 <25 Figure 4.67 shows IF 'Y,max I/ 2TR'Ylg and IFa I/2TR'Ylg in DI water and 9% IP A solutions for silicon particles on both types of wafers as a function of 0 at different Ot values. For silicon particles on a hydrophilic wafer, IF'Y,max I /2TR'Ylg is larger than IFa I/2'1'R'Ylg because Ot is less than 25° and 0 is close to 90° in DI water as shown in Figure 4.67(a}. This implies that particles on a bydrophilic wafer are removable in DI water and in fact experimental results shows this. 198 01 WaterlSi 0.6 0.5 0.4 It D' .!I 0.3 .r'" 0.2 0.1 0 0 20 40 60 80 100 9 (a) 9 % IPA/Si 0.6 0.5 ------~r__ SIISIOZ I a= 450 0.4 s:III i pC a:= SOO c!I 0.3 ------~-~~---SI/SI -rl" I 0.2 a= 6SO 0.1 0 0 20 60 80 100 9 (b) Figure 4.67. Comparison of IFy,muI/2'1rRy and IFal/2".RYIg as a function of e at different a values, (a) silicon particles in\)I water, and (b) silicon particles in 9% IPA The curved lines are IFy,max I/2".RYIg at different Q values and the straight lines are IFal /2". RYI&' 199 However, IF.y,max1/2 ...R'Ylg is slightly smaller than IFal/2... R'Ylg in 9% IPA solutions, thus particle removal should not be possible. This prediction contradicts the experimental observations of partial particle removal in 9% IP A For silicon particles on a hydrophobic wafer, particle removal is possible in DI water because IF'Y,maxl/2 ...R'Ylg is slightly larger than IFal/2... R'Ylg (aR is around 60°) as shown in Figure 4.67(a). However, as shown in Table 4.10, in 9% IP A solutions it is not possible to remove silicon particles on hydrophobic wafers. The calculated IF'Y,max I is always smaller than IFa I, which is opposite to the experimental observations in 9% IP A solutions. 4.6.4.3. Modification of Leenaars Equations In IPA-containing solutions, the force term due to the surface tension gradient at the solution/air interface should be considered with the force caused by surface tension. From the surface tension data, it was concluded that a surface excess of IPA was present at the IPA solution/air interface. As the wafer with particles passes through this layer, the distribution of surface IP A molecules could be altered to cause a surface pressure difference (.... ). The total force including the surface pressure term, .... , can be written as: Ftotal = 2... R hlg + .... ) sin2(8/2)cosa or (4.9) 200 In the above equation, it is assumed that the removal force due to the surface pressure acts on the particle in the same direction as surface tension force. Figure 4.68 shows IFtotall acting against IFal in 9% IPA solutions as a function of surface pressure, T -. For these calculations, the values of a and (J in Tahle 4.9 were used for PSL particles on hydrophilic wafers and silicon particles on hydrophilic and hydrophobic wafers. From Fig. 4.68(a), the minimum surface pressure, T -, necessary for the removal of PSL particles may be calculated to be 26. The minimum surface pressures for the removal silicon particles on both types of wafers may be similarly determined to be 40 and 33 dynes/cm, respectively, as shown in Figures 4.68(b) and (c). When surface pressures larger than 40 dynes/em are generated due to the differential enrichment of IPA molecules at the IPA solution/air interface, PSL particles on hydrophilic and silicon particles on both types of wafers are removable. For the removal of alumina particles on a hydrophilic wafer, however, the surface pressure of 260 dynes/cm should be generated in 9% IPA solutions. This is impossible in aqueous solutions and this may explain the inability of IPA-based solutions to remove deposited Al20 3• The surface pressure model developed assumes that the surface tension and surface pressure forces are addictive and act in the same direction. This is obviously a crude assumption and a more rigorous and accurate approach is required to confirm the extent of impact of the surface pressure forces. 201 0.1 0.09 cc ,::: Ill: 0.08 ~ ..... PSUSI02 ! 0.Q7 0 r.... 0.06 0.05 0 10 20 30 40 so x 'It, dynes/cm (a) 0.4 0.35 cc ,::: Ill: ~ ..... 0.3 Ci ...0 r.... 0.25 0.2 0 10 20 30 40 so K *, dynes/cm (b) Figur~ 4.68. Comparison of IFtotad/211'RYlg and 1F.1/211'RY14 as a function of".o (a) PSL particles on hydrophilic wafer and silicon particles \b) on hydrophilic and 0 in 9% IPA solution. The curved lines are IFlotall /2". R YIg at different ". values and the straight lines are IF.1/211'RYIg" 202 0.6 0.55 DC $: CI: 0.5 ~ .... 0.45 "5 0 OJ r:&:. 0.4 0.35 0.3 0 10 20 30 40 so X,* dynes/cm (c) 0.2 Alumlna/SiOz DC 0.15 $: CI: )i .... 0.1 r-~ 0.05 0 0 SO 100 ISO 200 2SO 300 X.. , dynes/cm (d) Fig. 4.68 (continued). (c) silicon particles on hydrophilic wafers and silicon particles and (b) on hydrophiic wafers and (d) alumina particles on hydrophilic wafers in 9% IPA solutions. The curved lines are IFtotad/2~R'Y18 at different .,:* values and the straight lines are IF.I/2..-R'Yla' 203 CHAPTER 5 SUMMARY AND CONCLUSIONS 5.1. Alkaline Solutions (1) C~oline appears to become surface active at the liquid/air interface at a solution concentration of approximately 400 ppm. The addition of 10,000 ppm of choline decreased the surface tension of water to 47.6 dynes/em. In contrast, the surface tension of ammonia solutions decreased only slightly as the concentration of NH40H was increased from 0 to 50 volume percent. (2) Single and double side polished p(100) bare wafers showed the same advancing (O.J and receding (OR) angles in the first cycle in choline. Single side polished wafers showed the reversal of angles (0 A < (J~ in the second cycle, but this was not observed on the double side polished wafers. (3) On oxide-coated p(lOO) wafers, advancing angles decreased in the first cycle as the choline concentration increased and complete wetting was observed above 100 ppm of choline in the second cycle. (4) Types and orientations of silicon wafers did not affect the wettability of BOE-etched silicon surface in alkaline solutions investigated. This might be due to the passivation of silicon surface by hydrogen. (5) When multiple cycles were performed in choline solutions, advancing angles decreased much more rapidly in the second cycle than in the first cycle. At 204 a choline concentration of O.S wt%, advancing angles reached a value of 200 in the second cycle indicating improved hydrophilicity. (6) Air exposure of choline-rinsed silicon wafers resulted in enhanced oxidation. Oxidation (as determined by ellipsometry and XPS) and wettability of samples increased when samples, after a 10-min choline treatment, were exposed to air before being rinsed in DI water. Lower stage speeds « 124 p.m/sec, longer air exposure time) resulted in better wettability than higher stage speeds (> 124 p.m/sec, shorter air exposure time). (7) TMAH was not surface active at concentrations lower than 1000 ppm. A rapid decrease of advancing angles in the second cycle occurred at TMAH concentrations above 200 ppm. The decrease of advancing angles was due to the etching of silicon surface by TMAH. (8) In a 600 ppm TMAH solution, the ~-v and Rrms values were measured to be 832.2 and 123.3 A by STM/AFM, respectively, which were much higher than values measured in a 10,000 ppm choline solution. The addition of 200 ppm of hydrogen peroxide and a non-ionic surfactant was effective in increasing the wettability, even at low TMAH and choline concentrations, and resulted in much smoother surfaces after treatments when observed by STM/ AFM. (9) The wettability of silicon in ammoniacal solutions was significantly improved by the addition of hydrogen peroxide. Advancing angles became 00 (complete wetting) in the second cycle when ammonia volume fraction exceeded 205 0.017%. Receding angles were low (_tOO) but never reached 0° in either cycle. The oxide thickness of wafers immersed in solutions with different NH40H concentrations was measured to be 13 ± 1 Aand was independent of NH40H content and air exposure time. S.2. IPA/Water Solutions (1) The addition of IPA to water resulted in a decrease of the surface tension of solution/vapor interface. The surface excess of IP A molecules at the solution/air interface was calculated to have a maximum value of 8.5 x to- IO moles/cm2 in a solution composition of 25% IPA and 75% water. (2) IP A solutions with less than 25% IPA were very effective in removing hydrophobic particles such as PSL and silicon on hydrophilic wafers. This particle removal from hydrophilic wafer samples may be attributed to the surface pressure differences generated along the solution/air interface during the entry of the sample into the solution. (3) The wettability of particles (6) and substrates (a) in solutions played important roles in particle removal. The large values of 8 and small a were preferred to obtain the maximum removing force due to surface tension against the adhesion force. The lowest removal force was calculated for alumina and glass particles on hydrophobic wafers. The highest removal force was calculated for silicon particles on hydrophilic wafers. 206 (4) Leenaars equation for the calculation of magnitude of surface tension force which balances FA did not seem to hold good for the solutions containing IPA less than 25%. This equation was modified by adding a surface pressure term, 7·. The modified form of Leenaars equation was able to explain experimental results in IPA-containing solutions. 207 REFERENCES References for Chapter 1 1.1. T. Hattori, "Contamination Control: Problems and Prospects," Solid State Technolo~, ~(7), SI (1990). 1.2. W. Kern, "The Evolution of Silicon Wafer Cleaning Technology," J.. Electrochem. Soc., 137(6), 1887 (1990). 1.3. V. B. Menon, L. D. Michaels, R. P. Donovan and D. S. Ensor, "Effects of Particulate Size, Composition, and Medium on Silicon Wafer Cleaning," Solid State Technology/Contamination Control Supplement, J2(3), S7, (1989). 1.4. W. Kern and D. A Puotinen, "Cleaning Solutions Based on Hydrogen Peroxide for Use in Silicon Semiconductor Technology," RCA Rev., ~, 187 (1970). 1.5. M. Asano, T. Cho and H. Muraoka, "Application of Choline in Semiconductor Technology," Abstracts 354, The Electrochemical Society Extended Abstracts, Vol. 76-2, Las Vegas, Nevada, October 17-22, 1976, pp. 911-913. 1.6. H. Muraoka, K. 1. Kuosawa, H. Hiratsuka and T. Usami, "Cleaning Solutions Based on Choline," Abstracts 238, The Electrochemical Society 208 Extended Abstracts, Vol. 81-2, Denver, Colorado, October 11-16, 1981, pp. 570-573. 1.7. Anonymous, Moses Lake TMAH Seminar. 1.8. A Hariri and R. S. Hockett, "Evaluate Wafer Cleaning Effectiveness," Semicond. International., 12(9}, 74 (1986). 1.9. V. B. Menon, A C. Clayton and R. P. Donovan, "A Critical Evaluation of Megasonic Cleaning for Removal of Particulate Contaminants from Silicon Wafers," Microcontamination, 1(6}, 31 and 107 (1989). 1.10. D. Burkman, "Optimizing the Cleaning Procedure for Silicon Wafer Prior to High Temperature Operations,"Semicond. International., !(7), 103 (1981). 1.11. A E. Walter and C. F. McConnell, "Stationary In Situ Wet Chemical Processing: The Impact on Wafer Surface Phenomena," Semicond. International., 12(1}, 36 (1989). 1.12. H. Mishima, T. Yasui, T. Mizuniwa, M. Abe and T. Ohmi, "Particle Free Wafer Cleaning and Drying Technology," IEEE Trans. Semi. Manuf., 2(3} (1989). 1.13. V. B. Menon and R. P. Donovan, "Review of Particle Control Methods During Wet Chemical Cleaning of Silicon Wafers," in Semiconductor Cleanin~ Technolo~11989, PV. 90-9, J. Ruzyllo and R. E. Novak (Eds.), 209 pp. 167-187, The Electrochemical Society Softbound Proceeding Series, Pennington, New Jersey (1989). 1.14. Anonymous, Mallinckrodt HF Particle Seminar. 1.15. O. J. Antilla, M. V. Tilli, M. Schaelars and C. L Claeys, "Effect of Chemical on Metal Contamination on Silicon Wafers," J. Electrochem. k .132(4), 1180 (1992). 1.16. C. F. McConnell, "Investigating the Impact of Wafer History on the Removal of Particles from Silicon Surfaces," Microcontamination. R(10), 45 (1990). References for Chapter 2 2.1. W. Kern, IThe Evolution of Silicon Wafer Cleaning Technology," 1.. Electrochem. Soc" 131(6), 1887 (1990). 2.2. K Sugiyama and T. Ohmi, "Ultraclean Gas Delivery Systems Part I. VLSI Fab Must Begin with Ultraclean Nitrogen Systems," Microcontamination, 2(11), 49 (1988). 2.3. M. M. Atalla, E. Tannenbaum and E. J. Scheibner, "Stabilization of Silicon Surfaces by Thermally Grown Oxides," The Bell System Technical Journal, ~(3), 749 (1959). 2.4. S. Raider, R. Flitsch and M. Palmer, "Oxide Growth on Etched Silicon in Air at Room Temperature," J. Electrochem. Soc., 112(3),413 (1975). 210 2.5. F.N. Schwettmann, K.. L. Chiang and W. A Brown, "Variation of Silicon Dioxide Growth Rate with Pre-Oxidation Clean," Abstraact No. 276. Extended Abstracts, Vol. 78-1, The Electrochemical Society Spring Meeting, Seattle, Washington, May 21-26, 1978. 2.6. G. Gould and E. A Irene, "The Influence of Silicon Surface Cleaning Procedures on Silicon Oxidation," J. Electrochem Soc., 134(4), 1031 (1987). 2.7. J. M. de Larios, C. R. Helms, P. B. Kao and B. E. Deal, "Effect of Silicon Surface Cleaning Procedures on Oxidation Kinetics and Surface Chemistry," Appl. Sur. Sci., 3Q, 17 (1987). 2.8. A R. Reinberg, "Dry Processing for Fabrication of VLSI Devices," in VLSI Electronics: Microstructure Science. Vol. 2, N. G. Einspruch (Ed.), Ch. 1, Academic Press, New York, 1981. 2.9. S. J. Fonash, "An Overview of Dry Etching Damage and Contamination Effects," J. Electrochem. Soc., ill(12), 3885 (1990). 2.10. W. Kern and D. A Puotinen, "Cleaning Solutions Based on Hydrogen Peroxide for Use in Silicon Semiconductor Technology," RCA Rev., J.1, 187 (1970). 2.11. J. M. de Larios, Ph.D. Thesis, Stanford University, 1989. 2.12. J. Davison, "Acid Reprocessors - Part 1: Piranha or Sulfuric Acid Reprocessor," Solid State Technol., JS,(3), Sl-S5 (1992). 211 2.13. W. Kern, "Radiochemical Study of Semiconductor Surface Contamination I. Adsorption of Reagent Components," RCA Review, 11, 207 (1970). 2.14. W. Kern, "Radiochemical Study of Semiconductor Surface Contamination II. Desorption of Trace Impurities on Silicon and Silica," RCA Review, ll, 234 (1970). 2.15. J. A Amick, "Cleanliness and the Cleaning of Silicon Wafer," Solid State Technol., 12(11),47 (1976). 2.16. W. Kern, "Purifying Si and Si02 Surfaces with Hydrogen Peroxide," Semicond. International., 7.(4), 94 (1984). 2.17. R. C. Henderson, "Silicon Cleaning with Hydrogen Peroxide Solutions: A High Energy Electron Diffraction and Auger Electron Spectroscopy Study," J. Electrochem, Soc" 112(6), 772, (1972). 2.18. F. N. Schwettmann, K L. Chiang and W. A Brown, "Variation of Silicon Dioxide Growth Rate with Preoxidation Clean," Abstract No. 276, Extended Abstracts, Vol. 78-1, The Electrochemical Society Spring Meeting, Seattle, Washington, May 21-26, 1978. 2.19. A Ohsawa, R. Takizawa, K. Honda, T. Matsutani and K. Imaoka, "Improvements of Trench Capacitor Characteristics by Pre-Oxidation Surface Clean with HN03-HF-H20 Solution," IEDM 88, 726 (1988). 2.20. D.A Peters and C. A Deckert, "Removal of Photoresist Film," L Electrochem. Soc., 126(5), 883 (1979). 212 2.21. W. Kern, "Chemical Etching of Silicon, Germanium, Gallium Arsenide, and Gallium Phosphide," RCA Review. ~ 278 (1978). 2.22. P. J. Holmes, in The Electrochemistll of Semiconductors. P. J. Holmes (Ed.), Academic Press, Ltd., London, 1962, p. 329. 2.23. R. M. Finne and D. L. Klein, "A Water-Amine-Complexing Agent System for Etching Silicon," J. Electrochem. Soc., 114(9), 965 (1969). 2.24. D. B. Lee, "Anisotropic Etching of Silicon," J. Appl. Phys., 40(11), 4569 (1969). 2.25. M. Asano, T. Cho and H. Muraoka, "Application of Choline in Semiconductor Technology," Abstracts 354, The Electrochemical Society Extended Abstracts, Vol. 76-2, Las Vegas, Nevada, October 17-22, 1976, pp. 911-913. 2.26. H. Muraoka, K J. Kurosawa, H. Hiratsuka and T. Usami, "Cleaning Solutions Based on Choline," Abstracts 238, The Electrochemical Society Extended Abstracts, Vol. 81-2, Denver, Colorado, October 11-16, 1981, pp. 570-573. 2.27. H. Muraoka, M. Asano and T. Ohashi, "Method of Etching a Semiconductor Substrate," U.S. Patent 4172005. 2.28. E. D. Palik, V. M. Bermudez and O. J. Glembock, "Ellipsometric Study of Orientation-Dependent Etching of Silicon in Aqueous KOH," L Electrochem. Soc., 132(4), 871 (1985). 213 2.29. E. D. Palik, H. F. Gray and P. B. Klein, "A Raman Study of Etching Silicon in Aqueous KOH," J, Electrochem, Soc"' 130(4),956 (1983). 2.30. B. Meddings and V. N. Mackiw, ''The Gaseous Reduction of Metals from Aqueous Solutions," in A~plication of Fundamental Thermodynamics to Metallur~cal Processes. G. R. Fillerer (Ed.), Gordon and Breach Sci. Publ., New York, 1967, p. 105. 2.31. G. Gould and E. A. Irene, "An In Situ Ellipsometric Study of Aqueous NH40H Treatment of Silicon," J, Electrochem, Soc" 136(4), 1108 (1989). 2.32. W. C. Schumb, C. N. Satterfield and R. L. Wentworth, "Hydrogen Peroxide," American Chemical Society Monograph Series No, 128, Ch. 7, p. 354, Reinhold Publishing Co., New York, NY (1955). 2,33. J. E. A M. van den Meerakker and M, H. V. van del' Straaten, itA Mechanistic Study of Silicon Etching in NH3/H20 2 Cleaning Solutions," L Electrochem. Soc., 137(4), 1235 (1990). 2.34. H. Mishima, T. Yasui, T. Mizuniwa, M. Abe and T. Ohmi, "Particle-Free Wafer Cleaning and Drying Technology," IEEE Trans. Semi, Manuf., ~(3), 69 (1989). 2.35. A Hariri and R. S. Hockett, "Evaluate Wafer Cleaning Effectiveness," Semicond, International" 12(9), 74 (1986). 2.36. Y. Suzuki, "Choline," in Ullmann's Encyclopedia of Industrial Chemistty. Vol. A7, 1985, pp. 39-41. 214 2.37. Anonymous, Moses Lake TMAH Chemicals Seminar. 2.38. U.S. Patent # 4634509, "Method for Production of Aqueous Quaternary Ammonia Hydroxide Solution." 2.39. J. S. Judge, "A Study of the Dissolution of Si02 in Acidic Fluoride Solutions," J. Electrochem. Soc., ill(l1), 1772 (1971). 2.40. P. Kubelka and V. Pristoupil, "Die Hydrolyse des Silikofluoridions," Z. Anorg. Chern., 121, 391 (1931). 2.41. R. C. Weast (Ed), Handbook of Chemistry and Physics, 73th ed., CRC Press (1991). 2.42. K Kleboth, Monntshfte fur Chemie., "Fluorkomplexe des Siliciums in wabriger losung," ,22, 1177 (1968). 2.43. M. K. Karepetiants, Thermodynamic Constants of Inorganic and Organic Compounds, Ann-Arb or-Humphrey Science Publish (1970). 2.44. H. Kikuyama, N. Miki, K. Saka, J. Takano, 1. Kawanabe, M. Miyashita and T. Ohrni, "Principles of Wet Chemical Processing in ULSI Microfabrication," IEEE Trans. Semicond. Manufact., ~(1), 26 (1991). 2.45. H. Kikuyama, N. Miki, K. Saka, J. Takano, 1. Kawanabe, M. Miyashita and T. Ohrni, "Surface Active Buffered Hydrogen Fluoride Having Excellent Wettability for ULSI Processing," IEEE Trans. Semincond. Manufact., ~(3), 99 (1990). 215 2.46. K. Skidmore, "Dry Wafers Cleanly: Without Spinning," Semicond. International, 12(8), 80 (1989). 2.47. H. Mishima, T. Yasui, T. Mizuniwa, M. Abe and T. Ohmi, "Particle-Free Wafer Cleaning and Drying Technology," IEEE Trans. Semi. Manuf., 2(3) (1989). 2.48. A E. Walter and C. F. McConnell, "Direct Displacement Wet Processing: How It Affects Wafer Surface Phenomena," Microcontamination. ~(1), 35 (1990). 2.49. C. F. McConnell, "Examining the Effects of Wafer Surface Chemistry on Particle Removal Using Direct-Displacement Isopropyl Alcohol Drying," Microcontamination, 2(2), 35 (1991). 2.50. S. Raghavan, "Dynamics and Stability of IPAjWaterjWafer Interfaces with Special Relevance to Particle Removal from Wafer Surfaces," Presentation given at the Semi-Annual Meeting of the Center for Microcontamination Control, University of Arizona, Tucson, Arizona, April 4-6, 1991. 2.51. P. H. Singer, "New Ultraclean Drying Method," Semicond. International, ll(I), 24 (1992). 2.52. R. D. VoId and M. J. VoId, Colloid and Interface Chemistry, Ch, 12, p. 395, Addison-Wesley Publishing Co. (1983). 2.53. E. W. Flick, Industrial Solvents Handbook, 4th ed., Noyes Data Co., Park Ridge, New Jersey (1991). 216 2.54. A Wilson and E. L. Simons, "Vapor-Uquid Equilibria, 2 propanol and Water System," Ind. En~. Chern., 44{9}, 2214 {1952} 2.55. V. A Burrows, Y. J. Chabal, G. S. Higashi, K. Raghavachari and S. B. Christman, "Infrared Spectroscopy of Si{11l} Surfaces after HF Treatment: Hydrogen Termination and Surface Morphology," Annl. Phys. Lett, ~(11), 998 (1988). 2.56. A C. Zettlemoyer, "Hydrophobic Surfaces," J. Colloid Interface Sci.. 2R(3/4), 343 (1968). 2.57. G. Gould and E. I. Irene, "An In Situ Study of Aqueous HF Treatment of Silicon by Contact Angle Measurement and Ellisometry," J. Electrochem. ~, ill, 1535 (1988). 2.58. R. Williams and A M. Goodman, "Wetting of Thin Layers of Si02 by Water," Annl. Phys. Lett., 25(10), 531 (1974). 2.59. M. L. White, ''The Detection and Control of Organic Contaminants on Surfaces," in Clean Surfaces. G. Goldfinger (Ed.), p. 361, Marcel Dekker, New York (1970). 2.60. B. D. Ratner, S. C. Yoon and N. B. Mateo, "Surface Studies by ESCA of Polymers for Biomedical Applications," in Polymer Surfaces and Interfaces. W. J. Feast and H. S. Manto (Eds.), p. 231, John Wiley and Sons, New York (1987). 217 2.61. A W. Neumann and R. J. Good, 'Techniques of Measuring Contact Angles," in Surface and Colloid Science. Vol. 11, R. J. Good and R. R. Stromberg (Eds.), p. 31, Plenum Press, New York (1979). 2.62. J. D. Andrade, L M. Smith and D. E. Gregonis, "The Contact Angle and Interface Energetics," in Surface and Interfacial Aspects of Biomedical Polymers. Vol. 1, J. D. Andrade {Ed.}, Plenum Press, New York (1985). 2.63. R. J. Good, "Contact Angles and the Surface Free Energy of Solids," in Surface and Colloid Science, Vol. 11, R. J. Good and R. R. Stromberg (Eds.), p. 1, Plenum Press, New York (1979). 2.64. R. N. Wenzel, "Resistance of Solid Surfaces to Wetting by Water," Ind. Eng. Chern., 2.8(8), 988 (1936). 2.65. R. E. Johnson, Jr. and R. H. Dettre, "Wettability and Contact Angle," in Surface and Colloid Science. Vol. 2, E. Matijevic (Ed.), Wiley-Interscience, New York (1969). References for Chapter 3 3.1. A W. Neumann and R. J. Good, 'Techniques of Measuring Contact Angles," in Surface and Colloid Science. Vol. 11, R. J. Good and R. R. 'Stromberg (Eds.), p. 31, Plenum Press, New York (1979). 3.2. J. D. Andrade, L. M. Smith and D. E. Gregonis, "The Contact Angle and Interface Energetics," in Surface and Interfacial Aspects of 218 Biomedical Polymers. Vol. 1, p. 249, J. D. Andrade (Ed.), Plenum Press, New York (1985). 3.3. Cahn Instrument Co., DCA 312 Instruction Manual. References for Chapter 4 4.1. A W. Adamson, Physical Chemistry of Surface Chemistry of Surfaces, 4th ed., Ch. 3, p. 49, John Wiley & Sons, New York (1982). 4.2. J. E. A M. van den Meerakker and M. H. V. van der Straaten, "A Mechanistic Study of Silicon Etching in NH3/H20 2 Cleaning Solutions," L. Electrochem. Soc., 137(4), 1235 (1990). 4.3. M. Asano, T. Cho and H. Muraoka, "Application of Choline in Semiconductor Technology," Abstracts 354, The Electrochemical Society Extended Abstracts, Vol. 76-2, Las Vegas, Nevada, October 17-22, 1976, pp. 911-913. 4.4. H. Muraoka, K. J. Kuosawa, H. Hiratsuka and T. Usami, "Cleaning Solutions Based on Choline," Abstracts 238, The Electrochemical Society Extended Abstracts, Vol. 81-2, Denver, Colorado, October 11-16, 1981, pp. 570-573. 4.5. Anonymous, Moses Lake TMAH Seminar. 4.6. W. Kern, "Purifying Si and Si02 Surfaces with Hydrogen Peroxide," Semicond. Internationa1., 1(4), 94 (1984). 219 4.7. W. Kern, "Chemical Etching of Silicon, Germanium, Gallium Arsenide, and Gallium Phosphide," RCA Review, J2, 278 (1978). 4.8. A Philipossian, II The Activity of HF /H20 Treated Silicon Surfaces in Ambient Air Before and After Gate Oxidation," J. Electrochem. ~ .u2(10), 2956 (1992). 4.9. Y. Suzuki, T. Kajima, S. Nishiate, T. Tsokiguchi, C. Watanabe, K. Kiuchi and C. Hashimoto," Gas Adsorption on Wafers," Microelectronics Manufacturing Technology, 14(12),38 (1991). 4.10. G. Kissinger, W. Kissinger, H. Hofmann and J. Kruger, "SOl by Silicon Wafer Direct Bonding-Problems of Wafer Warpage and Surface Chemistry," in Gettering and Defect Engineering in Semiconductor Technology, GADEST'91, M. Kittler and H. Richter (Eds.), Solid State Phenomena, Vols. 19 and 20, pp. 625-630, Elsevier, Netherlands (1991). 4.11. M. Grunder, "Einsatz von Oberflachenmethoden zur Charakerisierung von Si-Wafers nach Unterschieldlicher Vorbehandlung," Z. Anal. Chern .. 319, 853 (1984). 4.12. R. N. Wenzel, "Resistance of Solid Surfaces to Wetting by Water," I.ruh Eng. Chern" ~(8), 988 (1936). 4.13. R. J. Jaccodine, "Surface Energy of Germanium and Silicon," L Electrochem. Soc. 11.Q(6), 524 (1963). 220 4.14. G. Gould and E. I. Irene, "An in situ Study of Aqueous HF Treatment of Silicon by Contact Angle Measurement and Ellipsometry," J, Electrochem. ~ lli(6), 1533 (1988). 4.15. V. A Burrows, Y. J. Chabal, G. S. Higashi, K. Raghavachari and S. B. Christman, "Infrared Spectroscopy of Si (111) surfaces After HF Treatment: Hydrogen Termination and Surface Morphology," Alllll. Phys. !&ll., ~(11), 998 (1988), 4.16. D. B. Fenner, D. K. Biegelsen and R. D. Bringans, "Silicon Surface Passivation by Hydrogen Termination: A Comparative Study of Preparation Methods," J. Alllll. Phys., 6Q(1), 419 (1989). 4.17. L. L. Schram, "Petroleum Emulsions," in Emulsions, L. L. Schramm (Ed.), Advances in Chemistry Series, Vol. 231, p. 1, American Chemical Society, Washington, DC (1992). 4.18. R. Flitsch and S. I. Raider, "Electron Mean Escape Depths from X-ray Photoelectron Spectra of Thermally Oxidized Silicon Dioxide Films on Silicon," J. Vac. Sci. Tech., 12(1), 305 (1975). 4.19. H. Mishima, T. Yasui, T. Mizuniwa, M. Abe and T. Ohmi, "Particle-Free Wafer Cleaning and Drying Technology," IEEE Trans. Semi. Manuf., ~(3), 69 (1989). 221 4.20. D. A Edwards, H. Brenner and D. T. Wasan, Interfacial Transport Processes and Rheology, Ch. 5, Butterworth-Heinemann, Boston, Massachusetts (1991). 4.21. A F. M. Leenaars, "A New Approach to the Removal of Sub-Micron Particles from Solid (Silicon) Substrates," Particles on Surface 1. Detection, Adhesion, and Removal. K L. Mittal (Ed.), p. 361, Plenum Press, New York (1988). 4.22. G. Kar, S. Chander and T. S. Mika, ''The Potential Energy of Interaction Between Dissimilar Electrical Double Layers," J. Colloid and Interface Sei., 44(2), 347 (1973). 4.23. J. Visser, "On Hamaker Constants: A Comparison between Hamaker Constants and Lifshitz-Van der Waals Constants," Adv. Colloid Interface S£b ~, 331 (1972).