<%BANNER%>

Surfactant Mediated Passivation to Achieve Chemical Mechanical Polishing Selectivity


PAGE 1

1 SURFACTANT MEDIATED PASSIVATION TO ACHIEVE CHEMICAL MECHANICAL POLISHING SELECTIVITY By KYOUNG-HO BU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Kyoung-Ho Bu

PAGE 3

3 To my beloved family, Mineok, Minji, and Seongah Byeon.

PAGE 4

4 ACKNOWLEDGMENTS It is a privilege to work with intelligent and committed individuals. Too many people to mention have influenced my work and provided inspiration and useful suggestions over many years, but I would especially like to express my appreciation to my advisor, Dr. Brij Moudgil, for his invaluable research guid ance and constructive support th rough intense discussions and productive feedback on this study. His sincere de dication to science, discipline in conducting research and considerate attention to details have always kept me moving forward and made significant contributions to this dissertation. I would also like to acknowledge the other me mbers of my advisory committee, Dr. Rajiv Singh, Dr. Stephen Pearton, Dr. Dinesh Sh ah, and Dr. Wolfgang Sigmund, for their indispensable support. I also wish to acknowle dge Dr. Susan Sinnott, Dr. Chang-Won Park, Dr. Yakov Rabinovich, Dr. Ivan Vakarelski, Dr. Pa rvesh Sharma, and Dr. Manoj Varshney who have informed and elaborated this work, with sp ecial appreciation to Dr. Ko Higashitani for his valuable insights. I am grateful to the National Science F oundations Engineering Research Center for Particle Science and Technology for financiall y supporting this resear ch (Grant EEC-94-02989). To Gary Schieffele, Gill Brubaker, and all other ERC staff, faculty, and administrators, I extend my hearty thanks for making my time there productive. Colleagues and friends who have contributed to this research through critical discussions as well as friendship include Scott Brown, V ijay Krishna, Madhavan Esayanur, Rhye Hamey, Marie Kissinger, Monica James, Dushyant Shek hawat, Suresh Yeruva, Kalyan Gokhale, Amit Singh, Debamitra Duta, Stephen Tedeschi, Se jin Kim, Takgeun Oh, Sangyup Kim, Won-Seop Choi, Seung-Mahn Lee, Kyo-Se Choi, Suho Jung, a nd Inkuk Jun. I also thank Bryce Devine and Bryan Opt Holt for training me how to use modeling tools.

PAGE 5

5 I have been blessed with Father Sangsun Park in Gainesville Korean Catholic Church who helps me have peace in mind, and blessed with my children, Minseok and Minji, who encourage me to overcome obstacles and motivate me to try my best in life. In addition, I owe particular debts to my parents and my parents-in-law for their strong confidence in my family. Finally, I am always grateful to my wife, Seongah, for her pa tience and support in spite of all ups and downs during my study. This work would not have been possible without her.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES................................................................................................................ .........9 ABSTRACT....................................................................................................................... ............13 CHAPTER 1 INTRODUCTION..................................................................................................................15 2 LITERATURE REVIEW.......................................................................................................21 Shallow Trench Isolation (STI) Stru cture and Selectivity of Slurry......................................21 Influence of Selectivity on Global Planarization in STI CMP Process..................................22 Nanotopography................................................................................................................. ....24 Surfactant Mediated Lubrication Effects................................................................................24 Surface Chemical Characteristics of SiO2 and Si3N4 Surfaces in Aqueous Solution.............25 Surfactants Adsorption on Silicon N itride and Lubrication Effect........................................26 Mixed Surfactants System......................................................................................................27 Research Approach.............................................................................................................. ...28 3 CMP CHARACTERISTICS OF SILI CA AND SILICON NITRIDE...................................37 Experimental................................................................................................................... ........38 Relationship between Material Rem oval Rate (MRR) and Youngs Modulus......................39 Role of Electrostatic Interactions on MRR.............................................................................41 Effect of pH ....................................................................................................................42 Effect of Salt Addition ...................................................................................................45 Parameters Affecting Surface Finish in STI CMP.................................................................48 Salt Mediated Lubrication......................................................................................................51 4 ROLE OF SURFACTANTS IN DEVLOP ING SELECTIVE PASSIVATION LAYER IN CMP......................................................................................................................... ..........72 High Selectivity Slurry Using Surfactants..............................................................................73 Surfactant Mediated Boundary Layer L ubrication for Selective Polishing............................75 Optimization of High Selectivity Slurry.................................................................................76 5 ADSORPTION STUDY OF SODIUM DODECYL SULFATE ON SILICA......................86 Adsorption Behavior of SDS on Silica...................................................................................87 Structure of Adsorbed SDS Molecules...................................................................................89

PAGE 7

7 6 APPLICATION OF DENSITY FUNTI ONAL THEORY BASED MODELING FOR SURFACTANT ADSORPTION STUDY...........................................................................100 Methodologies.................................................................................................................. ....101 Structures and Resources......................................................................................................104 Results and Discussion.........................................................................................................106 SDS Adsorption on Silica at, below, a nd above the Isoelectric Point (IEP).................107 SDS Adsorption on Silicon Nitride at IEP....................................................................108 TX-100 Adsorption on Silica at IEP.............................................................................109 7 CONCLUSIONS AND SUGGESTI ONS FOR FUTURE WORK......................................120 Conclusions.................................................................................................................... .......120 Suggestions for Future Work................................................................................................122 LIST OF REFERENCES.............................................................................................................125 BIOGRAPHICAL SKETCH.......................................................................................................133

PAGE 8

8 LIST OF TABLES Table page 1-1 A Product Generations and Chip Size Model Technology Trend TargetsNear-term Years.......................................................................................................................... ........17 3-1 Youngs modulus, hardness measured by nanoindentation method, material removal rate (MRR), ratio of MRR (CMP pressure of 7 psi), and ratio of Youngs modulus for silica and silicon nitride................................................................................................54 6-1 Adsorption energy (kcal/mol) calculated by density functional theory (DFT) based method (B3LYP) using 6-31G* basis set........................................................................111 6-2 Adsorption free energy (kcal/mol) of SDS on silica calculated from adsorption density data in Ch. 5 at different pH a nd two different added concentrations (1.6mM and 16mM). .................................................................................................................... .112

PAGE 9

9 LIST OF FIGURES Figure page 1-1. Schematic representation of chemi cal mechanical polishing (CMP) process. .................18 1-2. Moore's Law Means More Performance............................................................................19 1-3. Multilevel metallization, cross sect ion with silica diel ectric and aluminum metallization.................................................................................................................. .....20 2-1. Schematic shallow isolation structure................................................................................29 2-2. Nanotopography (a) Top view and (b ) cross-section graph of substrate nanotopography................................................................................................................. .30 2-3. In-situ friction force and material remova l rate responses of the baseline slurries (12 wt%, 0.2 mm primary particle size) and the slurries containing C12TAB, C10TAB and C8TAB surfactants at 32, 68 and 140 mM c oncentrations in th e presence of 0.6 M NaCl at pH 10.5................................................................................................................ .31 2-4. Lateral force as a func tion of loading force in the presence of surfactant [22].................32 2-5. Zeta potential behavior of silica, silicon nitride, ceriu m oxide (ceria), and polishing pad (polyurethane) with respect to the pH ........................................................................33 2-6. Maximum surface concentration of benzoic acid ( ) and pyridine ( ) obtained by fitting the adsorption data to a Langmuir-Freundlich equation.........................................34 2-7. Friction coefficient of si licon nitride ceramic as a functi on of load in pure water ( ) and silane aqueous solution ( ).........................................................................................35 2-8. The mechanism of high-ionic-strength sl urry stabilization by the synergistic mixture of anionic and nonionic surfactants...................................................................................36 3-1. Variations of mateiral removal rate (MRR) for silica a nd silicon nitride substrate as a function of applied pressure by using undilute d (30 wt%) colloidal silica slurry at pH 10.4........................................................................................................................... ..........55 3-2. Variations of MRR of silica and silicon nitride substrate and calculated electrostatic force between two abrasives as a function of pH of the dilute d (12 wt%) colloidal silica-based slurry (Klebosol 1501-50)..............................................................................56 3-3. Particle size distributions of colloidal silica slurry at two different pH conditions...........57 3-4. Zeta potential of colloidal silica slurry and electrost atic force between silica abrasive particles...................................................................................................................... ........58

PAGE 10

10 3-5. Variations of MRR and calculated electrost atic force between two abrasives as a function of slurry NaCl salt concentr ations in the slurry at pH 10.4.................................59 3-6. Particle size distributi ons of colloidal silica slurry (Klebosol 1501-50, 12 wt%) as a function of salt concen trations at pH 10.4.........................................................................60 3-7. Surface roughness of silica and silicon n itride substrate after CMP as a function of added salt (NaCl) concentration at pH 10.4.......................................................................61 3-8. Material removal rate of silica and silicon nitride as a function of repulsive electrostatic force betw een silica abrasives.......................................................................62 3-9. Surface roughness of silica and silicon nitr ide substrates after CMP as a function of slurry pH...................................................................................................................... ......63 3-10. Surface morphologies and profiles of s ubstrates from two pH conditions........................64 3-11. Material thickness change of silica and silicon nitride substrates as a function of immersed time in pH 13 NaOH solution...........................................................................65 3-12. Surface morphologies and profiles of subs trates before and after etching in pH 13 NaOH solutions................................................................................................................. .66 3-13. Etch pits formed on (a) silica and (b) silicon nitride substrate immersed in 0.1 M (pH 13) NaOH solution for 12 days..........................................................................................67 3-14. Lateral force of a 6.8 m silica particle interacting with a silica substrate in pure water and CsCl, NaCl, and LiCl solutions of 1 M.............................................................68 3-15. Schematic representation of th e hypothetical frictional mechanisms................................69 3-16. Particle size distributions of colloidal silica slurry (F uso PL-7) without salt and with 1 M LiCl and 1 M CsCl.....................................................................................................70 3-17. Material removal rate of silica substr ates by CMP using diluted (9.6 wt%) colloidal silica slurries (PL-7) without salt and with 1 M LiCl and 1 M CsCl as a function of applied polishing pressure..................................................................................................71 4-1. Influence of SDS addition on CMP performances............................................................79 4-2. Surface finish of silica and silicon nitr ide substrates processed with standard and high selectivity slurry........................................................................................................ .80 4-3. Variation of zeta potent ial of silica and silicon nitr ide substrate and adsorption density of 16mM SDS on silica and silic on nitride powder measured by total organic carbon (TOC)................................................................................................................... ..81

PAGE 11

11 4-4. Variation of MRR and accompanying sele ctivity of Klebosol slurry (12 wt%) as a function of added SDS concentration at pH 2...................................................................82 4-5. Adsorption density of SDS on 12 wt% Kle bosol slurry with 16 mM SDS as a function of pH................................................................................................................. ...83 4-6. Effect of alkyl chain length of sodium alkyl sulfate on MRR and selectivity at pH 2......84 4-7. MRR and selectivity obtained by slurries with various surfactant and surfactant mixtures at pH 2............................................................................................................... ..85 5-1. Adsorption isotherm of SDS on colloid al silica (Klebosol 1501-50, 12 wt%) at pH 10.4........................................................................................................................... ..........93 5-2. Adsorption density of SDS on colloidal silica (12 wt% Klebosol 1501-50) at SDS concentration of 1.6 mM and 16 mM and zeta potential as a function of pH...................94 5-3. Zeta potential of Klebosol slurry as a function of SDS concentration at pH 10.4.............95 5-4. Pictorial depictions of the possible su rfactant aggregates film s at concentrations corresponding to I-IV in Figure 5-3...................................................................................96 5-5. Adsorption characteristics of SDS on Kl ebosol silica slurry a nd zeta potential as a function of concentration of SDS at pH 10.4 ....................................................................97 5-6. FTIR/ATR Spectra of SDS solution at 1, 2.5, 5 and 10 mM bulk concentration in the CH2 stretching region (2921, 2924) measured at pH 10.4 using Si ATR crystal..............98 5-7. Particle size dist ribution of Geltech SiO2 at pH 2 with and without 16 mM SDS 12 hours after pH change........................................................................................................99 6-1. Optimized (a) Si(OH)4, (b) Si(NH2)4, (c) Sodiumdodecyl sulfate (SDS), and (d) Triton X-100 (TX-100) stru cture using B3LYP method and 6-31G* basis set...............113 6-2. Optimized SiOH4 and DScomplex structure using B3 LYP method and 6-31G* basis set............................................................................................................................ .........114 6-3. Optimized SiOH5 + and DScomplex structure usin g B3LYP method and 6-31G* basis set...................................................................................................................... ......115 6-4. Sturcture of SiO4H3 and DScomplex. Optimization is not complete, since two molecules are being separated to decrease energy...........................................................116 6-5. Optimized SiO4H3 -, Na+, and DScomplex structure using B3LYP method and 631G* basis set................................................................................................................. .117 6-6. Optimized Si(NH2)4 and DScomplex structure using B3LYP method and 6-31G* basis set...................................................................................................................... ......118

PAGE 12

12 6-7. Optimized SiOH4 and TX-100 complex structure using B3LYP method and 6-31G* basis set...................................................................................................................... ......119

PAGE 13

13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SURFACTANT MEDIATED PASSIVATION TO ACHIEVE CHEMICAL MECHANICAL POLISHING SELECTIVITY By Kyoung-Ho Bu May 2007 Chair: Brij M. Moudgil Major: Materials Science and Engineering Chemical mechanical polishing (CMP) is an indispensable technique in the microelectronics industry to achie ve planarization and patterning of metal and dielectric layers. Device fabrication using high de nsity and small pattern size re quires precise control of CMP slurry properties. In this study, the performance of a colloidal silica CMP slurry for silica/silicon nitride, which consists of the shallow trench isolati on (STI) structures, was investigated. Factors determing material removal rate and surface finish were examined. It was found that electrostatic interactions can have significant effects on CM P performance. Emphasis was placed on selective removal of material. More than 10-fold increase in selectivity over conventional colloidal silica slurry was achieved with the a ddition of sodium dodecy l sulfate (SDS), an anionic surfactant. Adsorption characteristics of SDS on silica and silicon nitride we re measured as a function of slurry pH and surfactant concentr ation. It was determined that th e preferential adsorption of SDS on silicon nitride by electrostatic attraction results in the formation of a material-selective selfassembled passivation (boundary lubrication) laye r leading to selectiv e polishing. It was found that the adsorption density of surfactant plays a dominant role in determining selectivity.

PAGE 14

14 Accordingly, material-targeted boundary layer lubrication concept may be used to develop selective CMP polishing slurries. A theoretical approach based on density func tion theory was attempted to model various aspects of surfactant adsorption. Through this approach, it was possible to predict adsorption behavior and related thermodynamic properties to assist selection of passivating molecules.

PAGE 15

15 CHAPTER 1 INTRODUCTION Chemical mechanical polishing (CMP) is the planarization technique predominantly used for the fabrication of multilayer devices. Ma in components for CMP process include the substrate to be polished, the slu rry that provides the chemistr y and abrasives for mechanical removal-and the polishing pad. A schematic of CMP system is shown in Figure 1-1. Due to the demand for the faster and smaller devices, the nu mber of devices (dens ity) on a single wafer is expected to grow constantly as depicted in Figure 1-2. Accordingl y, the size of components of a device is expected to become sm aller as listed in Table 1-1. Hence the requirements for large scale integration are becoming more challenging. Current semiconductor devices are composed of multilayers as shown in Figure 1-3. Due to the planarity requirement for lithography proce sses, further processing is not possible if the required planarity is not achieve d. In addition, the standard for global planarization is becoming more demanding due to the high degree of device integration. Among the various structures requiring CMP, sh allow trench isolation (STI) is one of the most challenging, due to its larg e variation in pattern density. There are a number of possible approaches to accomplish global planarizati on in STI CMP process. Among these, the development of high selectivity slurries has been gaining more significance in order to accomplish a one-step CMP process for global planar ization. State of the art, high selectivity ceria based slurry has several drawbacks such as problems with coagulation and high defectivity, whereas conventional silica based slurries are kno wn to be free of those problems, but they exhibit low polishing selectivity be tween silicon nitride and silica s ubstrates. In this dissertation, silica based slurries were modified to achieve the targeted selectivit y of 15 or higher.

PAGE 16

16 The overall objective of the proposed investigati on is to improve the se lectivity (ratio of material removal rate of silica to silicon nitride) of the STI CMP slurry. Specific objective is to differentially modify surface states of silicon n itride and silica with surfactant or polymer adsorption, thereby selectively minimizing silicon n itride polishing, and thus leading to enhanced global planarization in STI CMP process. A synopsis of the various research tasks constituting this study is organized as follows. Chapter 2 reviews the literature on the STI CM P process and slurry selectivity. Different defects, hampering device performance will be addressed. The selectivity of the CMP slurry will be defined and its effect on global planarization w ill be discussed. Finally, strategies to increase the selectivity will be suggested. Chapter 3 covers the CMP characteristics of silica and silicon nitride substrates by colloidal silica slurry with respect to the material removal rate (MRR) and surface finish. Variables affecting the polishing pr ocess have been studied with special emphasis on electrostatic inte ractions. Chapter 4 presents the methodol ogies to increase the selectivity of the slurry. Specific mechanisms for observed resu lts will be discussed. Chapter 5 discusses the adsorption behavior of sodium dodecyl sulfate (S DS) on silica substrates, since it was found that SDS adsorption on silica abrasive particles dete rmines the necessary dosage of surfactant to fabricate high selectivity slurries. Chapter 6 describes the modeling efforts to develop methodologies based on density f unctional theory to predict optimal conditions for selective surfactant coating. Chapter 7 summarizes the conclu sions of this study and suggests future work.

PAGE 17

17 Table 1-1 A Product Generations and Chip Size Model Techno logy Trend TargetsNear-term Years [1]. Year of Production 2005 2006 2007 2008 2009 2010 2011 2012 2013 DRAM Pitch (nm) (contacted) 80 70 65 57 50 45 40 36 32 MPU/ASIC Metal 1 (M1) Pitch (nm) 90 78 68 59 52 45 40 36 32 MPU Printed Gate Length (nm) 54 48 42 38 34 30 27 24 21 MPU Physical Gate Length (nm) 32 28 25 23 20 18 16 14 13 ASIC/Low Operating Power Printed Gate Length (nm) 76 64 54 48 42 38 34 30 27 ASIC/Low Operating Power Physical Gate Length (nm) 45 38 32 28 25 23 20 18 16 Flash Pitch (nm) (un-contacted Poly)(f) 76 64 57 51 45 40 36 32 28

PAGE 18

18 Figure 1-1. Schematic representation of chemical mechanical polishing (C MP) process. (a) Side view; (b) Top view. (b) (a) Slurry Feed Holder Substrate Plate n Polishin g Pad Holder Slurry Feed

PAGE 19

19 Figure 1-2. Moore's Law Means More Performan ce. Processing power, measured in millions of instructions per second (MIPS), has steadily risen because of increased transistor counts [2].

PAGE 20

20 Figure 1-3. Multilevel metallization, cross se ction with silica diel ectric and aluminum metallization [3].

PAGE 21

21 CHAPTER 2 LITERATURE REVIEW Shallow Trench Isolation (STI) Stru cture and Selectivity of Slurry Chemical mechanical polishing or planarizat ion (CMP) is the key technology for shallow trench isolation (STI) process. STI process can reduce the required area for the device isolation and give better planarity relative to the local oxidation of silicon (LOCOS ) process. Therefore, existing sub-0.13 m technologies device isolation techni ques strongly depend on the STI CMP process [4-7]. There are several drawbacks such as dishing of silica, erosion of silicon nitride and failure to clear oxide that hamper global planarizati on in CMP process [8]. Typically the thickness uniformity across the substrate (usually called within-substra te non-uniformity, or WIWNU) must be below 3%, and dishing must typically be less than 20~50 nm. To minimize such defects, current STI CMP process is comprised of multi-steps [9] or raw structure modifications such as reverse mask, dummy active area, and additiona l active area [10]. For better productivity and process simplicity, a minimum number of proc ess steps are highly desired and accordingly, approaches for high selectivity single-step slur ry designs are being wide ly investigated [11-13]. Usually selectivity represents the ratio of material removal rate (MRR) of silica to silicon nitride: 1) (2 nitride Silion of rate removal Material Silica of rate removal Material y Selectivit In general, conventional silica abrasive based STI CMP slurry exhibits selectivity in the range of 3 to 4 [14]. According to the result re ported by J. Schlueter, erosion of silicon nitride could be minimized to less than 100 using ceria based high selectivity slurry in a multi-step STI CMP [15]. Besides the influence on planar ization, high selectivity provides enhanced endpoint detection capability. Genera lly, if oxide to nitride polishi ng selectivity is greater than

PAGE 22

22 15, monitoring substrate carrier motor current ca n be utilized for e ndpoint detection [11]. Therefore, research on improving selectivity and understanding the polishing mechanisms to achieve global planarization are needed. In this study, systematic approaches and strategies to improve selectivity of the STI CMP slurry for single-step CMP process were investigated. In the following sections, the detailed infl uences of selectivity on global planarization will be introduced, and issues of nanotopography th at justify a strong need for high selective STI CMP process will be outlined. Next, a brief re view of the polishing passivation/inhibition mechanism (i.e., surfactant mediat ed lubrication effects) will be provided. Surface chemical characteristics of silica and silicon nitridewill be reviewed, followed by examples of specifically adsorbing surfactants on silicon nitride surface. As an alternative to inhi bit polishing of silicon nitride by surfactants, silane additives to form passivation layer on silicon nitride will be introduced. Influence of Selectivity on Global Pl anarization in STI CMP Process As previously mentioned, several obstacles exist inhibiting global planarization in STI CMP. Figure 2-1 shows a schematic of a typical STI structure. It consists of a silicon device, a silicon nitride mask, and a silica in sulating layer inside of the tren ches. In the ideal CMP process, the oxide should be removed completely in all active regions, leaving it only in the trench regions (Figure 2-1 (b)) without eroding silicon n itride. In reality, there are three failure modes such as failure to clear oxide, excessive removal of nitride, and excessive removal of oxide [8]. The former is primarily an end-point detecti on issue, whereas the other two mechanisms are closely related to the pattern density of the devi ce, selectivity of slurr y, pad stiffness, imposing pressure, etc. [12]. To minimize these barriers, several approaches have been evaluated. One method is to use a stiffer pad and lower selectivity slurry [5], and the other is to use a softer pad and higher selectivity slurry [16]. When a stiffe r pad is used, which does not bend in the applied

PAGE 23

23 pressure range, the highest portion of the surface will start to be polishe d ultimately resulting in global planarization. However, there is also a possibility of poor surface finish and wafer breakage. When a softer pad is used, which ha s a greater flexibility, all the structures on substrate will be in contact to pad, and hence a high selectivity slurry wi ll be required not to preferentially polish the unwanted structure. In this case, the risk of poor surface finish and substrate breakage will be reduced. Current high selectivity slurries in STI CMP usually contain ce ria abrasives showing higher material removal rate for silica than silicon nitride [16]. In general, for higher pattern densities of which the area of sili ca isolation layer is not large, di shing effect decreases, since the pad bending is limited. For lower pattern densitie s, dishing effect increases because the pad bending is high [17]. Therefore in each case, the pad materials and operating pressure should be chosen appropriately. Kim et al. investigated the influence of slu rry selectivity of the slurry on erosion and planarity by modeling. It was pred icted that above 30% active pattern density, high selectivity slurries show good planarity [18]. In these cases, planarity is defined as the difference of height between the highest region and the lowest region on a substrate. Considering that higher pattern densities of the device will be re quired with decreasing device size in the future, a systematic research for a high selectivity slurry will be essential to meet these goals. In general, current STI CMP processes use sili ca abrasives that show low selectivity (about 3 4) [14]. W. G. America investigated the infl uence of selectivity on material removal rate of silica and silicon nitride using silic a and ceria abrasives. In this cas e, the material removal rate of silica and silicon nitride was determined to be about 2700 /min and 700 /min, respectively [19]. Recently, ceria abrasives have shown higher selectivity (more than 5), and are being

PAGE 24

24 investigated for further enhancement. According to W. G. America, mate rial removal rate of silica using ceria abrasives was more than 5700 /min as compared to 800 /min for silicon nitride [19]. However in ceria CMP, the pH at which maximum polishing rate and maximum selectivity are achieved is about 8, which also is the isoelectric point (IEP) of ceria. This results in coagulation of ceria yielding poor surface morphologies with scratches and higher roughness [13]. Recently, it has been reported that by decreasi ng the size of ceria abrasive particles, the number of scratches can be decreased significantly [20]. Nanotopography An emerging issue impairing global planariz ation in STI CMP is nanotopography. This phenomenon is becoming a strong driving force for developing high selectivity slurries. Nanotopography is a term used to describe relatively gentle (10-100 nm) surface height variations occurring over lateral distances of 1-10 mm on unpa tterned silicon substrates (Figure 2-2). Boning et al. have inves tigated this issue by modeling and verified it by experiments, and have suggested that due to the height variati on of blanket wafer, several defect mechanisms come into play such as failure to clear oxide and excess nitride thinning (erosion). It has been commonly believed that stiffer pa d would yield acceptable planari zation [5]. On the contrary, it has been shown that softer pad and lower pressure is more effective in minimizing such defects [5]. With respect to this phenomen a, if the selectivity of the slurry is not high enough and endpoint detection is not accurate accompanying erosion will be unavoidable. Silicon nitride erosion can be minimized if only additional prot ective layers exist on silicon nitride surface. Surfactant Mediated Lubrication Effects As a protective mechanism from polishing for si licon nitride, one of the approaches is to incorporate surfactant mediated l ubrication effects. Basim et al. have shown that the addition of

PAGE 25

25 long chain cationic surfactant (e.g. C12TAB) produces an enhanced defect-free surface morphology but the polishing rate was extremely sma ll due to the lubrication effect of surfactant (Figure 2-3.) [21]. Although this research was fo cused on dispersion of abrasive particles, it implied that long chain surfactant can act as an anti-polishing agent. Vakarelski et al. showed that the primary mechanism of lubric ation is the formation of an intervening surfactant aggregate fi lm on solid-liquid interface largel y by electrostatic interactions [22]. In addition, the decrease in frictional force depends on the concentration of surfactant. After the concentration reaches critical micelle concentrati on (CMC), there was no further decrease in lateral (frictional) force. The effect of surfactant c oncentration on the lateral force is illustrated in Figure 2-4. Surface Chemical Characteristics of SiO2 and Si3N4 Surfaces in Aqueous Solution Understanding the surface chemistry of substrates is the first step to implement the above approach to create a selective passivation/lubrica tion layer. It is well known that silicon nitride forms the same type of surface hydroxyl layer as silica in an aqueous solution. However, there is a difference in the surface group compositions. Figur e 2-5 illustrates the ze ta potential variation with respect to pH. Unlike silica (IEP of 2.2), silicon nitride exhib its an IEP of about 5.8. This difference is explained on the basis of relati ve number of silanol (Si-OH) and amine (Si2-NH) groups on the silicon nitride surface as compared to only silanol groups on silica surface [23]. The silanol groups are acidic in nature and thus result in a lowe r IEP, while the presence of amine groups results in a higher IEP. In the case of silicon nitride powder with an IEP of pH 6, the ratio of nitrogen to oxygen was calculated to be appr oximately 0.2, and it was nearly 1 for powders with an IEP of pH 7.9 [23]. Sonnefeld et al. reported, based on potentiometric titration measurements, that the surface site densities of amine group (Si2NH) and that of silanol group (SiOH) are 0.56 /nm2 and 1.83

PAGE 26

26 /nm2, respectively on the silicon nitride surface [24] Density of silanol groups on silica surface was estimated to be 0.74 /nm2 [25]. From these values, conv erted area of amine and silanol groups for molecular adsorption is 1.79 nm2 and 0.546 nm2, respectively on the silicon nitride surface, and area of silanol gropup on silica was 1.35 nm2. Surfactants Adsorption on Silicon Nitride and Lubrication Effect There have been many reports on stabilization of silicon nitride powders using polymeric dispersants [26-30]. Malghan et al. investigated the dispersion behavior of silicon nitride powder using both cationic Betz 1190 (quaternized polyamine epoxychlorohydrin) and anionic Darvan C (ammonium poymethacrylate) polymers [29]. In the case of cationic polyelectrolyte (CPE), there was strong electros tatic attraction between CPE and silicon nitride powder at pH 9 leading to stable dispersion, while in the case of anionic polyel ectrolyte (APE), the adsorption was very restricted due to the similar surface charge, consequently, small adsorption occurred possibly due to the hydrogen bondi ng. According to Hackley et al., anionic poly acrylic acid (PAA) adsorption on silicon nitride surface decrea sed from 100% at pH 3 to around 25% at pH 10, however, stable dispersion was achieved due to de pletion forces in the pr esence of PAA [26]. Besides the sign of surface charge, hydrogen bon ding plays an important role in adsorption of organic molecules on silicon ni tride. Bergstrm et al. investig ated the adsorption behavior of various organic probe molecules in cyclohexane [31]. They showed that benzoic acid and benzyl amine prefer to adsorb on the basic amine (Si2NH) groups via hydrogen bonding (N-H) (Figure 2-6). To accomplish selective ad sorption of surfactants or polym ers on silicon nitride surface, anionic surfactants should be i nvestigated first considering th at nitride shows higher negative zeta potential at pH 10.5 for current CMP conditions. Philipossi an et al. showed that by applying anionic poly-carboxylate, th e selectivity increased from 5 to 100 [32]. They used ceria abrasives for silica polishing at pH 8. According to their results, most of anionic surfactant adsorbed on

PAGE 27

27 silicon nitride with some amount of polymer adsorption on silica and ceria abrasive particles, resulting in overall decreased MRR from 500 to 100 (a.u.). Hibi et al. investigated th e lubrication effect of s ilane coupling agents (3-(2aminoethylaminopropyl) dimethoxymethylsilane) on silicon nitride and alumina ceramics (Figure 2-7) [33]. They reported that amino-containing silane c oupling agents formed the crosslinked polysiloxane by hydrolysis and dehydrative condensation, which was effective in reducing both friction and wear of silicon nitride. In othe r words, the additives reduced the wear of silicon nitride as a result of inhibition of silicon nitride reaction with wa ter. In this case, the silane agents reacted with the oxide (silanol group) on silicon nitride surface. As mentioned above, since the density of silanol groups on si lica and silicon n itride surface was estimated to be 0.74 /nm2 and 1.83 /nm2, respectively [24, 25], the extent of the passivation on silicon nitride and silica is expected to be different. Mixed Surfactants System Palla et al. investigated the use of mixed surfactants to di sperse the alumina abrasive particles in CMP. They reported that by applyi ng anionic surfactant, sodium dodecyl sulfate (SDS), mixed with various noni onic surfactants, the dispersion stability was highly improved [34]. The schematic of the slurry stabilization of alumina abrasives is shown in Figure 2-8. In this scheme, adsorption was attributed to str ong adsorption of ionic surfactants on abrasive particles, and associa tion of nonionic surfactants with ionic surfactants via hydrocarbon chain interactions (attractive hydrophobi c forces). Alumina is known to have Lewis active site similar to silicon nitride, hence, it can be envisioned that mixed surfac tants concept can be applied to silicon nitride-silica sy stem. However, under the normal CMP pH conditions, zeta potential of silicon nitride is negative, indi cating the greater significance of electrostatic interaction.

PAGE 28

28 Research Approach Commercial ceria abrasive STI CMP slurries with selectivity of about 5 are known to result in high defectivity and post-CMP cleaning problems, while colloidal silica slurries has a lower selectivity of 3 to 4, although they e xhibit acceptable defectivity. Therefore in the proposed research, surfacta nts that selectively adsorb on silico n nitride will be investigated and methods to inhibit polishing and the mechanisms will be studied to improve global planarity. One of the major challenges is the fact th at both materials have silanol group on their surfaces in water and show negative zeta potent ial at the conventiona l CMP pH of 10.4. The ideal solution is to find a surfactant, which has selective affinity only to silicon nitride. To achieve this goal, several anionic surfactants and mixed surfactant systems will be investigated in terms of adsorption with respect to pH and added surfactant concentration. In using silica abrasives under current CMP conditions, anionic surf actants adsorption on abrasive particles will be largely opposed due to the similar (negative) charge of the adsorbate and adsorbent. Therefore, to increase the amount of surfactant adsorptio n on silicon nitride, readjustment of CMP process pH to a lower va lue may be required. Since pH plays a dominant role in determining surfactant ad sorption through electros tatic interactions, de tailed investigation of the adsorption behavior of the anionic surfactant as a function of pH will be required to achieve optimal surfactant adsorption.

PAGE 29

29 Figure 2-1. Schematic shallow isolation structure: (a ) Initial structure before CMP: typical trench isolation structure used to isolate activ e regions on a substrate where devices will be built. The nitride layer has been pattern ed and a shallow trench etched into the silicon. An oxide has then been deposited into the trench, which also results in overburden oxide above the nitride active area s. (b) ideal result af ter CMP: the oxide is removed completely in all active regions leaving oxide only in the trench regions. Three key failure mechanisms may arise: (c) excessive removal (ero sion) of nitride in active areas, (d) excess removal of oxide (dis hing) within the trench, and (e) failure to clear oxide from nitride active areas [8]. SiO2 (d) Dishing (c) Erosion (e) Failure to clear oxide ( a ) Before CMP (b) Ideal result after CMP Si3N4 Si

PAGE 30

30 Figure 2-2. Nanotopography (a) Top view and (b) cross-section graph of substrate nanotopography. Dotted line in (a) shows path of scan. The x axis in (b) indicates the distance along the scan path in (a), moving from left to right [8]. WAFER HEIGHT (nm)100 80 60 40 20 0 -20 -40 -60 -80 -100 Nanotopography Length 100 nm -100 nm 100 nm -100 nm 100 nm -100 nm (a) (b)

PAGE 31

31 Figure 2-3. In-situ friction force a nd material removal rate responses of the baseline slurries (12 wt%, 0.2 mm primary particle size) and the slurries containing C12TAB, C10TAB and C8TAB surfactants at 32, 68 and 140 mM c oncentrations in th e presence of 0.6 M NaCl at pH 10.5. (Striped bars represent the Friction Force responses and the solid bars represent the Removal Rate responses) [21].

PAGE 32

32 Figure 2-4. Lateral force as a f unction of loading force in the presence of surfactant [22]. 0 50 100 150 200 250 300 050010001500Loading Force (nN)Lateral Force (nN) Pure Water 1mM C12TAB 8mM C12TAB 16mM C12TAB 32mM C12TAB

PAGE 33

33 Figure 2-5. Zeta potential behavior of silica, silicon nitride, ceriu m oxide (ceria), and polishing pad (polyurethane) with respect to the pH [32].

PAGE 34

34 Figure 2-6. Maximum surface concentration of benzoic acid ( ) and pyridine ( ) obtained by fitting the adsorption data to a Langmuir-Freundlich equation [31]. 0 10 20 30 40 50 Max. surface concentration ( mol/m2) 3.5 Amount amino groups (%)3.0 2.5 2.0 1.5 0 10 20 30 40 50 Max. surface concentration ( mol/m2) 3.5 Amount amino groups (%)3.0 2.5 2.0 1.5 3.5 Amount amino groups (%)3.0 2.5 2.0 1.5

PAGE 35

35 Figure 2-7. Friction coefficient of silicon nitride ceramic as a func tion of load in pure water ( ) and silane aqueous solution ( ) [33].

PAGE 36

36 Figure 2-8. The mechanism of highionic-strength slurry stabilizat ion by the synergistic mixture of anionic and nonioni c surfactants [34].

PAGE 37

37 CHAPTER 3 CMP CHARACTERISTICS OF SILI CA AND SILICON NITRIDE The Shallow trench isolation (STI) chemical mechanical pol ishing (CMP) process involves polishing of silica and silicon nitr ide layer. Therefore, the charact eristics of the both materials are very important for process optimization and overall STI CMP process performance. Besides, silicon nitride is widely used for various appl ications such as giant magnetoresistance (GMR) and ceramic ball bearings making the research on the CMP characteristics of silicon nitride more significant [35, 36]. There are several abrasives used in STI CMP slurries according to its specific purposes [19, 36, 37]. Among them, colloidal silica is the traditional material, wh ich has long been used for various applications, and its disper sion stability towards various el ectrolytes is well documented [38-41]. A unique property is th at it shows high dispersion stabili ty around its isoelectric point (IEP, pH 2 ~ 4), unlike other materials. It has long been believed that hydration force due to modified water structure at the silica surface or silanol (SiOH) gr oups give rise to a repulsive forces, which is responsible for the observed phenomena [39, 42]. Another explanation is that the formation of a surface gel layer or short polymerlike hairs protruding from the silica surface can give rise to steric repulsion [43, 44]. In intermediate pH ra nge, silicic acid chains (-Si(OH)2-OSi(OH)2-OH) or siloxane bonds (Si-O-Si) are report ed to form silica ge l relatively easily by reaction between acidic ionized silanol (SiO-) and neutral silanol (SiOH). At a higher than pH 10, colloidal silica shows stable dispersion again through electrostatic repu lsion between almost completely ionized silanol groups. As a result, co lloidal silica suspensions are stored and used usually under high pH conditions. Wh en a lower pH application is required, the pH transition is performed in a very short time period to avoid gelation.

PAGE 38

38 Silica is a promising candidate for the STI CMP due to its high surface quality as compared to other materials. However, the basic CM P characteristics for silica and silicon nitride, which consist of the STI structure, are not completely understood. In this chapter, CMP characteristics of silicon and sili con nitride by colloidal silica abra sives will be discussed with an emphasis on the electrostatic interact ions encountered in the system. Experimental The CMP slurry used in this study was Kl ebosol 1501-50 from Rodel Co. The original slurry of 30 wt% colloidal silic a abrasives was diluted with nano-pure water to 12 wt%. The slurry pH was measured to be 10.4 after dilution. HCl and KOH solutions were used for further adjustment of the slurry pH. The study of lubric ation by hydrated cations utilized PL-7 supplied by Fuso Chemical Co., which is originally at 20 wt% colloidal silica abrasives. It was diluted with nano pure water to 9.6 wt%, with a final sl urry pH of 7.3. Salt concentration was controlled to 1 M by adding the proper amount of 5 M salt so lution to the slurry. Concentrated 5 M solution was prepared with analytical grade LiCl and Cs Cl purchased from Fisher Scientific Co. Silica and silicon nitride wafers were purchased from Silicon Quest International. Two m thickness of silica thin film was deposited on (100) Si subs trate by plasma enhanced chemical vapor deposition (PECVD) method using Tetra Ethyl Ort ho Silicate (TEOS) as a source on (111) Si. For the silicon niride wafers, 3000 thickness silicon nitride film was deposited on the 3000 silica, which was used as a diffusion barrier on (100) Si by low-pre ssure chemical vapor deposition (LPCVD) method us ing dichlorosilane (SiCl2) and ammonia (NH4) as source materials. IC 1000/Suba IV stacked pads s upplied by Rodel Inc. and TegraPol-35 with TegraForce-5 from Struers Co. tabletop polisher were utilized for CMP purposes. The rotation speed was controlled to 150 rpm both for the pad and the wafer. Material removal rate (MRR) was measured using ellipsometry (Woollam EC 110 Ellipsometer) by dividing the decrease in

PAGE 39

39 thickness by polishing time. In the present st udy, MRR reproducibility was within 5 %. Prior to each polishing step, the pad underwent 30 sec onds of conditioning with diamond conditioner. The actual time for polishing was controlled to 30 seconds. Youngs modulus and hardness were measured by Nanoindentation method using Hysitr on Triboindenter purchased from Hysitron Co. Digital Instruments Nanoscope III atomic force microscope was used for the measurement of surface roughness of substrates after CMP. Zeta potential of the slurry was measured by Acoustosizer purchased from Colloidal Dynamics Co. A variation in the zeta potenti al values (20 mV) at pH 10.4 was observed for different batches purchased from slurry supplie r. A decrease in zeta pot ential was also observed with aging time (10 mV upon 1 y ear aging). Accordingly, zeta poten tial values at the same pH were found to be different depending on the batc h and aging time. However for a given sample, the reproducibility of measurement was found to be within 3 mV over a month period. Particle size distribution wa s measured by Coulter partic le size analyzer (Coulter LS13320). After dissolution, the pictures of th e substrate surface were taken by optical microscopy (Olympus BX60). Relationship between Material Re moval Rate and Youngs Modulus The MRR of silica and silicon nitride wafers as a function of polishing pressure is plotted in Figure 3-1. In this experiment, original slurry (30wt% solids lo ading) was used without further dilution. The MRR showed a linear relationship with polishing pressure, as predicted by the empirical Preston equation [45]: 1) (3 t s P K MRRp where, Kp is Preston coefficient, P is polishing pressure, and s is the relative travel between glass surface and lap over in which the wear occurs (platen speed) during time interval t [45].

PAGE 40

40 The MRR of silicon nitride was determined to be lower than silica. In CMP of Si-based materials such as silica and sili con nitride, it is we ll known that water play s a significant role, because no material removal occurs in non aqueous medium. It is commonly believed that water attacks and breaks the siloxane bonds by the following reaction: 2) (3 SiOH SiOH O H Si O Si2 It has been reported that the hard ness of silica decreases to aro und 50% of the original value in aqueous systems [46, 47]. The above reaction is believed to be controlled by the diffusion of water in silica, which in tu rn affects surface hardness. There have been several attempts to explai n MRR theoretically [45, 48]. One of them is Cooks model, assuming Hert zian penetration [45]: 3) (3 t s P E 2 1 MRR where, E is the Youngs modulus of the material. Cons idering that the modulus is the resistance of the material to tensile or compressive deform ation, above equation indicat es that material with high modulus should be harder to polish. A more elaborate model incorporating chemical effects was proposed by Chi-Wen and co-workers [48]: 4) (3 t s P ) E 1 E 1 ( C MRRw a where, C is the coefficient accounting for chemical e ffect of a slurry and other properties of CMP consumables, Ea and Ew are the Youngs modulus of abra sive particle and substrate, respectively. Trends in experimental results with substrates of di fferent moduli were in aggrement with those predicted by Equation (3-4). To evaluate the correlation between MRR and m echanical properties of substrate materials, Youngs modulus and hardness of both substrat es were measured by the nano-indentation

PAGE 41

41 method and are summarized in Table 3-1. The MRR and Youngs modulus ratio indicated a correlation between MRR and mechanical properties of the material. However, according to this explanation, silicon nitride cannot be polished by silica abrasive particles, since silicon nitride has a higher hardness than silica, in contrast to experimental ev idence. In reality, the formation of a thin silica layer (around 1 nm) on the silicon nitride surface by spontaneous oxidation represented by the equation below and is expected to influence the polishing characteristics of silicon nitride [23, 31, 49] 5) (33 2 2 4 3NH 4 SiO 3 O H 6 N Si It has been reported that the ra te-limiting step for the above re action is the breakage of Si-N bonds [50], with relatively faster breakage of Si-O bonds due to diffusion of water. In other words, the reaction of water with silicon nitrid e for breaking the Si-N bond is slower than water diffusion. As a result, the thickness of the newl y formed silica layer on silicon nitride will be very thin compared to that of the silica substr ate, thereby resulting in different MRR of the two substrates. Theoretically, Youngs modulus reflec ts the bond strength of the material on an atomic scale [51]. In other words, a higher modul us means stronger bonds, which will be harder to break. Role of Electrostatic Interactions on MRR It has long been observed that MRR is depe ndent on the pH of the slurry in various polishing processes including CMP. As was disc ussed by Choi and co-workers, electrostatic interactions can influence the CMP performa nce. However, systematic approaches and quantitative analysis to explain the effect and modulation have not been attempted until now.

PAGE 42

42 Effect of pH One of the best ways to modulat e the electrostatic inte raction is to change pH of the slurry. Colloidal silica slurry is the best candidate fo r this purpose, since it shows stable dispersion throughout a wide pH range, if only the pH was adju sted just before polishing. To investigate the effect of electrostatic intera ction on CMP performance, the M RR for both substrates as a function of slurry pH was measured and plotted in Figure 3-2. Particle si ze distribution at pH 2 and 10.4 in Figure 3-3 confirmed th at there was no measurable coag ulation of silica particles at pH 2. MRR as a function of pH reached a maximum as slurry pH is reduced. At high pH beyond 11, MRR steeply increased for silica and remained constant for silicon nitride. The CMP results of silicon and silica as a function of pH were re ported by several authors [52-54]. Choi et al. attributed the increase in MRR at lower pH to the electros tatic attraction between the oppositely charged silica substrate and silica abrasive particles, and a higher MRR at higher pH to increased softening of silica induced by its high solubility at higher pH. According to their report, the electrostatic force between si lica particles and substrate s howed a maximum around 0.4 mN/m (force/radius of partic le) at pH 10.4. The contact area of the CMP pad and the substrate was reported to be around 1% due to the asperity ch aracteristics of the pad materials employed in their study at the same pH [55]. Assuming that ha lf of the individual abra sive particle will be embedded in the substrate surface and the other ha lf of the particle will be captured by pad asperities during the CMP process, the contact area will yield the number of particles in contact with the substrate. If 1% of a 1 1 inch wafer is in contact with the abrasive particles, then there will be approximately 109 particles of diameter of 90 nm in the system. The total electrostatic force is calculated to be 18 mN According to experiments in the present study, if one assumes that there is no electrostatic force contribution at pH 3 (due to its near ly zero value of zeta

PAGE 43

43 potential), a pressure caused by repulsive force of 6.85 N on 1 1 inch wafer, is required to make a difference in MRR. This is more than tw o orders of magnitude difference in electrostatic force contribution between the ab rasive and the wafer. It is, however, possible that induced repulsion by electrostatic interact ions may contribute to lubrica tion effects. According to Choi, there was approximately 25% decrease of frictiona l force between colloidal silica abrasives and the wafer when the slurry pH was increased from 2 to 10.4. Mahajan also reported that the frictional force between pad and the wafer decrease d at higher pH due to increased electrostatic repulsion between them [56]. It is well known that in the case of boundary lubrication, fr iction follows the equation for interfacial sliding, as pro posed by Tabor et al.[57]. 6) (3 A S Fc friction where, Ffriction is a frictional force, Sc is a critical shear stress that depends on the details of the interfacial region, and A is the contact area. It is not clear which term is affected by the electrostatic interaction for the current system. Ho wever, it seems reasonable that if electrostatic repulsion between the abrasive and subs trate is high, critical shear stress ( Sc) will be reduced, resulting in overall reduction in the fric tional force. On the other hand, surface layer characteristics can also change upon a shift in pH, resulting in change s in contact area (A) between the pad and the substrate. Yeruva reported that there was no consistent evidence that the Youngs modulus of the pad, which is directly rela ted to the contact area, changes with solution pH. Recently, Taran et al. have reported that a lubr ication effect between silica particles and the substrate resulted in reduced lateral force at high pH above 9.6, using lateral force microscopy [58]. Below pH 9.6, there was no noticeable change They correlated their observations with

PAGE 44

44 solubility of silica and formation of surface gel layer, which is beli eved to form at high pH due to high solubility [58]. It seems likel y that the lubrication phenomena may play a role in explaining low MRR at high pH, but it is not possible at present to e xplain high MRR below pH 8.6. Another possibility is that the electrostatic forces between pa rticles can change the number of abrasive particles par ticipating in the polishing process, depending upon their dispersion/coagulation characterist ics. It has been generally know n that MRR is almost linearly proportional to solids loading of the slurry [59, 60]. Zeta potential of the ab rasive particle will produce electrostatic repulsive forces that will re sist the particles to come within a certain distance of the substrate resulti ng in limited number of particle s participating in polishing at a certain pH. The repulsive force can be calcula ted using simplified Poisson-Boltzman equation [61] 7) (3D 2 o oe 2 R / F where, F/R is the electrostatic fo rce per particle radius, is the Debye-Huckel parameter, o is surface potential, and D is the distance between particles which is assumed to be 1 nm. The absolute force value can change as a function of distance, but th e trend should be similar. Zeta potential was assumed to be the same as the surface potential, since th ere were no specific adsorbing ions in the slurry. Figure 3-4 shows the measured ze ta potential of silica and the corresponding electros tatic force between abrasive partic les calculated from the potential as a function of pH (also plotted in Figure 3-2). At pH around 3 (IEP of silica), the electrostatic force leveled off and approached zero and MRR for sili ca also reached a maximum value at pH 3. In the intermediate pH range (3 ~ 10), MRR and th e electrostatic force were inversely proportional to each other.

PAGE 45

45 At pH above 11, the MRR of silica showed a sudden increase, probably related to the solubility of silica. However, the MRR of silicon n itride, which has a lower solubility than silica, showed the same trend as electrostatic force. Ov erall, it appears that there exists an inverse correlation between the MRR and repul sive electrostatic forces betw een the abrasive particles. The zeta potential of the subs trate and colloidal silica s hould be similar, since both materials are amorphous silica, therefore it may be sa fe to assume that the calculated electrostatic force also represents the trend in the force between abrasive partic les and substrate. It is clear that the electrostatic forc es induced by zeta potential of variou s materials has a significant effect on MRR in terms of (i) opposing force against polis hing pressure or (ii) number of particles participating in the CMP process. Effect of Salt Addition It is well known that various salts reduce the surface charge of the particles in a colloidal system, decreasing the electrostatic repulsion and thereby promoting their coagulation by attractive van der Waals inter actions [38, 39]. The minimum c oncentration of salt causing coagulation of particles is cal led the critical coagulation c oncentration (CCC). This phenomenon can be utilized to modulate the electrostatic force in the CMP process. Among various salts, monovalent ions are most suitable for this purpo se in terms of controlla bility, since multivalent ions have far lower CCC than monovalent ions. A llen and co-workers have reported that CCC of NaCl for colloidal silica wa s around 0.4 M and that of CaCl2 was around 1 mM, at pH 9. CMP was conducted as a function of NaCl concentr ation added to the slurry. The MRR for both substrates and calculated electrostatic force between abrasive particles from zeta potential values are plotted in Figure 3-5. The first thing to be monitored is the coagula tion of particles whenever salt is added into slurry. Figure 3-6 shows the partic le size distribution as a functi on of NaCl concentration. Below

PAGE 46

46 0.5 M NaCl, the particle size maintained a narrow mono size distribution. When the concentration reached 0.5 M, gelation occurred and particle size distribution showed multiple peaks. It is not clear from Figure 3-6 if there is coagulation, since the additional peak(s) from coagulation are not noticeable due to the multiple peaks from gelation. It is very likely that there is some degree of coagulation at that concentrat ion. Gelation usually occurs at intermediate pH and high salt concentration in a colloidal sili ca system, and it is different from coagulation. Gelation is reversible, i.e. the dispersion stability can be restor ed simply by stirring or dilution, but if the coagulation oc currs, it is not usually reversible. In gelation, silica particles form a network by siloxane (Si-O-Si) bonds. In coagulat ion, they do not form any network, but they simply collide with each other by Brownian moti on leading to very strong attractive van der Waals interactions. It is not known how gelation of abrasive particles affects the CMP performance. A colloidal silica slurry adjusted to neutral pH and kept for some time to promote geltation without any salt can be a good candidate to isolate such effects. Below a salt concentration of 0.5 M, NaCl a ddition to the polishing slurry showed the same trend in MRR change as the pH change. Th ere was a steep decrease in the MRR after the salt concentration exceeded the CCC (0.5 M NaCl) for silica, however. The silicon nitride substrate did not show such drama tic change. It has been reported that at fixed solids loading, the MRR decreases as a function of particle size afte r reaching a critical size of particles [62, 63]. This leads to the explanation of how the coagula tion might affect MRR. At a fixed solids loading, coagulation leads to two possible effects, (i) reduction in the number of abrasive particles participating in the polishing process thereby de creasing the contact area between particles and substrate, (ii) increased penetration depth due to size enlargement resulting in higher MRR. As was discussed by Yeruva, optimal indentation de pth is determined by the thickness of the

PAGE 47

47 modified surface layer of silica caused by reaction with water, which is believed to be on the order of nm in thickness [55]. Besides, the optimum mean part icle size resulting in maximum MRR was reported to be around 75 nm experi mentally [63]. In the present study, the agglomerated particle size is larger than 100 nm, hence a decrease in MRR and poor surface finish are expected and experimental results confirmed these predictions. Choi reported that at intermedia te salt concentrations, Stber silica slurry showed a broader distribution with a larger part icle size accompanying the MRR increase, and was attributed to reduced electrostatic forces and increased particle size due to coagulation [64]. At a higher salt concentration, they reported low MRR and high r oughness values attributed to coagulation of the silica abrasive particles. In th e present study with a colloidal s ilica slurry, the increase in MRR can solely be attributed to reduced electrost atic repulsion, since ther e was no particle size increase. Measured surface roughness values indicated that up to 0.3 M NaCl, there was not much difference in surface roughness (Figure 3-7). However at 0.5 M, a rough surface with low MRR on silica but not on silicon nitride was observed. On the silicon nitride subs trate, the coagulation of abrasive particles does not seem to have as high an effect as on the silica probably due to the higher hardness of silicon nitride substrate as compared to the silica abrasive particles. Salt addition has been reported to increase frictional force between the pad and substrate as also observed by Mahajan [56]. This suggests that coagulation of abrasive particles is a major factor in determining frictional forces, which in turn impact MRR. In order to further establish a correlation between the MRR and elec trostatic forces, the MRR for both materials is plotte d in Figure 3-8 as a function of electrostatic repulsive force between colloidal silica abrasive particles at di fferent levels of pH and salt concentrations.

PAGE 48

48 Except under the extreme conditions such as pH 2, 11.5 and NaCl concentration of 10 mM, where calculated electrostatic force was not sensitiv e to experimental variab les, an inverse linear relationship was observed between MRR and electrostatic forces. Parameters Affecting Surface Finish in STI CMP Figure 3-9 shows the surface roughn ess of silica and silicon nitr ide substrates as a function of slurry pH. Selected surface morphologies and roughness profiles of the silica and silicon nitride after CMP at pH 10.4 and 11.5 for both ma terials are plotted in Figure 3-10. CMP by colloidal silica slurry impr oved the roughness of both materials below pH 11. 5 and silica showed higher roughness values than silicon nitrid e over the entire pH range examined in this study. At pH 11.5, CMP resulted in poor surface fi nish for both materials but the increase of roughness was higher for silica. Scratches from th e CMP process were not observed on either substrate. This variation of roughness follows exactly the sa me trend as the silica solubility results by Iler [65]. It is known that the so lubility of silica shows a steep in crease in the basic pH condition. Iler reported about a th ree orders of magnitude increase in si lica dissolution rate as the pH value changed from 2 to 11 [65]. The increase in solubili ty is believed to be due to the hydroxyl ion (OH-) acting as a catalyst for attack by water on the siloxane (Si-O-Si) network. Specifically, hydroxyl ions create an excess of electrons resulting in a higher negative surface potential and consequently more attacks by H3O+ [21]. Therefore, it has been widely believed that the high dissolution rate of silica at hi gh pH is responsible for the hi gh MRR [53, 54]. The effect of solubility on surface roughness has not been well unde rstood. It should be noted that solubility of silica is known to depend on the curvature of the s ilica surface [66]. Hulett et al. reported that the convex surface of colloidal silica shows higher so lubility than the concave one, and a smaller radius of curvature exhibits hi gher solubility [66]. This imp lies that surface convex impurities

PAGE 49

49 will dissolve faster than flat substrates. However, this prediction is contrary to our experimental observation of the effect of solubility on MRR and surface roughness, and requires further investigation. To evaluate the effect of solubility of silica and silicon nitride on CMP performance, dissolution rate was determined by measuring the thickness of bot h substrates immersed in a 0.1M (pH 13) NaOH solution for 12 days without stirring (Figure 3-11). Surface roughness of the substrates before and after dissolution is pr esented in Figure 3-12. The dissolution rate of silica was three orders of magnit ude higher than that of silico n nitride most probably due to higher bond strength of the latter. Even though the experiment was conducted at pH 13, the magnitude of dissolution of both substrates was re latively low. However, in a real CMP process, dissolution can be increased by th e imposed pressure re sulting in higher tens ile stress created by the abrasive particles as they abrade silica surface. Nogami and co-workers reported a 50% increase in solubility when 30 MPa compressive stress was applie d compared with the stress-free condition [47]. Additionally, when abrasive partic les abrade the surface, the temperature can be higher due to heat generated by friction. It has been reported by Iler that so lubility of colloidal silica increased by more than ten times at 200 oC than at room temper ature [65]. However, incorporation of all thos e factors still gives a far less dissolu tion rate than the MRR increase at pH 11.4 for silica. Regarding this apparent discre pancy, it should be noted that the attack of hydroxyl ions will be higher at higher pH resulting in a softer layer, which can be removed easily, and is prone to damage by abrasion. Consequently, the attack of hydroxyl ions increases the solubility and promotes formation of a softer layer on the subs trate at high pH. The disso lution of silica itself does not seem to play a bigger role in determinin g MRR. The extent of hy droxyl ion attack will

PAGE 50

50 also be dependent on the bond strength, and according to Youngs moduli of the materials, this may explain the reason for low MRR and dissolution rate of silicon nitride. Figure 3-12 illustrates the surface morphologies of the two substrates after dissolution at pH 13 for 12 days. There was very small increas e in surface roughness for both of the substrates. The inverse pyramidal-type etch pits observed in Figure 3-13 are common phenomena when highly concentrated alkaline solution is used fo r etching silicon in the micromachining of silicon substrates [67-69]. The reason for the anisotropic etching is differe nt reactivities of certain crystal planes of silicon. In other words, anisotropic etchants etch much faster in one dir ection than in another, which is usually (111) planes of silicon. Ther efore, anisotropic etch ing of (100) silicon by alkaline solution results in the inverse pyramidal -type etch pits, as was observed experimentally. Since the thin film used in this research was deposited on (100) silicon, the silica film will have a similar atomic arrangement as the underlying silicon. It has been well observed in silicon anisotropic etching that when a dilute alkalin e solution (20 wt%) is us ed, the etching produces high surface roughness. Palik et al reported that the high surface roughness is attributed to the formation of hydrogen bubbles acting as a pseudoma sk, thus inhibiting uniform etching [69]. In silica, the overall reaction of the diss olution can be described as follows: 9) (32 3 2H 2 1 ) OH ( Si OH O H SiOH Gas bubbles were observed during dissolution experi ments and are believed to be hydrogen gas. For the reaction shown in Equation (3-9) to occur, the nucleation of hydrogen bubbles is a dominant step and it is much eas ier to nucleate them on high ener gy sites giving rise to surface defects.

PAGE 51

51 Salt Mediated Lubrication Donose and co-workers have reported that various cations adsorbing on silica from electrolyte solutions can induce lubrication throu gh the formation of a hyd rated cation layer [70]. Due to the difference in hydration enthalpy of different cations, resultant lubrication was different for each added salt. Similar phenomen a have been reported by Raviv and Klein by a modified surface force apparatus [71]. They meas ured the shear force between mica surfaces and concluded that hydration layers of adsorbed cations act as a hi ghly efficient boundary lubricant. Their research was mostly done by lateral for ce microscopy and the macroscopic effect on CMP was not investigated. Figure 3-14 shows the lateral force as a function of loading force in the presence of various salts such as LiCl, NaCl and CsCl reported by Donose and co-workers [70]. According to their results, every salt showed higher lubrication e ffect than pure water. The thickness of the adsorbed cation layer increases with increas ing electrolyte concen tration. Highly hydrated cations such as Li+ can form a thick and soft layer resu lting in higher lubrication than poorly hydrated cation such as Cs+. It was observed that the degree of lubrication followed their order of hydration, which is Li+ > Na+ > Cs+. Schematics shown in Figure 315 illustrate this concept. For pure water, at least one layer of water molecu les are bound to the silica surface, but this layer is relatively thin and firmly adsorbed to the silica surface resulting in rigid interface. In the presence of an electrolyte solution, there is a thicker hydration la yer than pure water. The model suggested by Raviv et al. states that the cations surrounded with water molecules are very hard to remove and remain fluid like in a lateral directio n and promote lubrication. It is well known that smaller Li+ ion has the highest hydration enthal py and hydrated radius among various

PAGE 52

52 cations[72]. Accordingly, Li+ ions have a thicker and more e ffective lubricating layer on silica surface, while Cs+ ions have a thinner and less effective lubrication layer. To investigate how the variations in lubric ation affects the real CMP performance, CMP was conducted as a function of applied polishing pr essure using three slurries with no salt, 1 M LiCl, and 1 M CsCl. The salt c oncentration was selected corres ponding to the results reported by Donose and co-workers [70]. The particle size was measured to assess if the selected salt addition causes any coagulation of the abrasive particles (Figure 316). When appropriate amounts of 5 M LiCl and 5 M CsCl were added to change the salt concentration, there was no particle size increase initiall y up to about 10 minutes after mi xing. As time passed, gelation took place slowly and the peak height of the particle size decreased and the size distribution became broader. While it is not well understood how th e gelation affects the CMP performance, CMP was performed 5 minutes after the mixing of 5 M salt solution to avoid the possible effect of gelation and ensure uniform mixing of added salt. Surface roughness measurement showed that RMS surface roughness was around 0.15 nm for all the conditions at the same polishing pressure and there was no increase from salt addition. Theref ore, it appears that th e variation in MRR is due to the effect of salt on mate rial properties and not necessarily from gelation and coagulation. Figure 3-17 shows the variation in the MRR w ith and without 1 M LiCl and CsCl as a function of polishing pressure. Increase in the M RR with added salt suggests that electrostatic interactions play a dominant role in polishing. The MRR of the silica substrate using a slurry with 1 M LiCl is lower than that of 1 M CsCl showing results in agreement with those from lateral force microscopy measurements. Consideri ng the same electrolyte concentration in both experiments, the electrostatic forces should be similar. The lubrication effect of individual

PAGE 53

53 particles should reduce the MRR, but increase in the number of abrasive pa rticles due to reduced electrostatic repulsive forces between them seem s to have resulted in overall higher MRR. In summary, CMP performance using colloidal silica slurry in a silica and silicon nitride system revealed that Youngs modulus of the substr ate material is more likely the reason for the differences in their MRR, with electrostatic repul sive force imposed by pH change in the slurry playing a dominant role. The el ectrostatic interaction was valid ated by monovalent salt addition to the slurry. A linear relationship between the M RR and electrostatic forces implied that such repulsive interactions probably resulted in gove rning the number of particles engaged in the polishing process. Dissolution rates were meas ured by immersing substrates into 0.1 M NaOH solution for 12 days and the resu lts showed that dissolution of silica was much higher than silicon nitride, however, the rate of dissolution was too low to make any significant difference in the MRR. It seems that the attack of hydroxyl ions at higher pH is res ponsible for poor surface finish and higher MRR due to the formation of a softer top layer. Dissolution in alkaline solutions produced a poor surface finish due to nucleation of hydrogen gas bubbles. The effect of the nature of added ions on CMP performance was also investigated. The Lubrication effect of hydrated cations was determined not to be a dominant factor in MRR. However, a slurry with LiCl showed lower M RR than one with CsCl, which suggests that the lubrication of the hydrated cations is playi ng a limited role in determining the MRR.

PAGE 54

54 Table 3-1. Youngs modulus, hardness measured by nanoindentation method, material removal rate (MRR), ratio of MRR (CMP pressure of 7 psi), and ratio of Youngs modulus for silica and silicon nitride. E (GPa) H (GPa) MRR (/min)MRRSiO2/MRRSi3N4 ESi3N4 /ESiO2 SiO2 84.6 3.0 8.5 0.3 4696 Si3N4 176.6 2.5 23.5 1.0 1382 3.4 2.1

PAGE 55

55 Figure 3-1. Variations of mateiral removal rate (MRR) for silica a nd silicon nitride substrate as a function of applied pressure by using undilute d (30 wt%) colloidal silica slurry at pH 10.4. Material Removal Rate (/min) 024681012141618 -1000 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 pH 10.4 SiO2 Si3N4 Pressure (psi)

PAGE 56

56 Figure 3-2. Variations of MRR of silica and silicon nitride substrat e and calculated electrostatic force between two abrasives as a function of pH of the dilute d (12 wt%) colloidal silica-based slurry (Klebosol 1501-50). Material Removal Rate ( /min ) 24681012 400 600 800 1000 1200 1400 1600 1800 2000 2200 23456789101112 400 600 800 1000 1200 1400 1600 1800 2000 2200 24681012 0.00 0.02 0.04 0.06 0.08 0.10 MRR SiO2 MRR Si3N4Electrostatic Force/R (mN/m) Force pH

PAGE 57

57 Figure 3-3. Particle size di stributions of colloidal silica slurry at two different pH conditions. 0.11 -2 0 2 4 6 8 10 12 14 16 Differential Volume (%) pH 10.4 pH 2Particle Size (m)

PAGE 58

58 Figure 3-4. Zeta potential of colloidal silica slurry and electrost atic force between silica abrasive particles. Force was calculated from th e zeta potential values by constant surface charge model. The distance between abrasives was assumed to be 1 nm. 24681012 -90 -75 -60 -45 -30 -15 0 123456789101112 0.00 0.02 0.04 0.06 0.08 0.10 Electrostatic Force/R (mN/m) Zeta potential (mV) p H

PAGE 59

59 Figure 3-5. Variations of MRR and calculated electrost atic force between two abrasives as a function of slurry NaCl salt concentr ations in the slurry at pH 10.4. Material Removal Rate (/min) 101102400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 101102 0 5 10 15 20 25 MRR SiO2 MRR Si3N4NaCl Concentration (mM) pH 10.4 ForceElectrostatic Force/R (mN/m)

PAGE 60

60 Figure 3-6. Particle size distribut ions of colloidal silica slurry (Klebosol 1501-50, 12 wt%) as a function of salt concen trations at pH 10.4. 0.11 0 2 4 6 8 10 12 14 16 No Salt 0.1 M NaCl 0.3 M NaCl 0.5 M NaClDifferential Volume (%)Particle Size (m)

PAGE 61

61 Figure 3-7. Surface roughness of silica and silicon nitride substrate after CMP as a function of added salt (NaCl) concentration at pH 10.4. 00.30.5 0.0 0.1 0.2 0.3 0.4 0.5 RMS Roughness (nm)NaCl Concentration (M) SiO2 Si3N4

PAGE 62

62 Figure 3-8. Material removal ra te of silica and silicon nitrid e as a function of repulsive electrostatic force between silica abrasives: (a) pH effect and (b) Salt (NaCl) addition at pH 10.4. Material Removal Rate (/min) 300 600 900 1200 1500 1800 2100 2400 0.000.020.040.060.080.10 0.000.06 0.000.06 (a) SiO2Electrostatic force/R ( mN/m ) Si3N4 0510152025 600 900 1200 1500 1800 2100 2400 pH 10.4(b) () Electrostatic Force/R (mN/m) SiO2 Si3N4Material Removal Rate (/min)

PAGE 63

63 Figure 3-9. Surface roughness of silica and silicon n itride substrates after CMP as a function of slurry pH. 2345678910111213 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 RMS Roughness (nm)pH SiO2 Si3N4

PAGE 64

64 Figure 3-10. Surface morphologies and profiles of substr ates from two pH conditions; (a) silica at pH 10.4, (b) silica at pH 11.5, (c) silicon nitride at pH 10.4, and (d) silicon nitride at pH 11.4 Si3N4, pH 11.5, RMS Roughness: 0.22 nm (d) Si3N4, pH 10.4, RMS Roughness: 0.14 nm (c) (a) SiO2, pH 10.4, RMS Roughness: 0.24 nm SiO2, pH 11.5, RMS Roughness: 0.48 nm (b)

PAGE 65

65 Figure 3-11. Material thickness change of silica and silicon nitride substrates as a function of immersed time in pH 13 NaOH solution. Elli psometer was used to measure thickness change. Calculated dissoluti on rates were also shown. Removed Thickness ( ) 024681012 0 200 400 600 800 1000 1200 1400 1600 1800 2000 SiO2 Si3N4Si3N4 Dissolution rate: 0.1 M NaOH (pH 13) SiO2 Dissolution rate: Days 0.109 /min 0.966 x 10-3/min

PAGE 66

66 Figure 3-12. Surface morphologies and profiles of s ubstrates before and after etching in pH 13 NaOH solutions; (a) bare silica (b) silica after 12 days, (c) bare silicon nitride, and (d) silicon nitride after 12 days. SiO2, RMSRoughness: 0.338 n m (a) Si3N4, RMS: 0.204 nm, R max: 3.073 n m (c) Si3N4 at pH 13 for 12 days, RMS Roughness: 0.241 nm (d) (b) Si3N4, RMS Roughness: 0.204 nm SiO2at pH 13 for 12 days, RMSRoughness: 0.434 n m

PAGE 67

67 Figure 3-13. Etch pits formed on (a) silica and (b) silicon nitride substrate immersed in 0.1 M (pH 13) NaOH solution for 12 days. (a) (b)

PAGE 68

68 Figure 3-14. Lateral force of a 6.8 m silica particle interacting with a silica substrate in pure water and CsCl, NaCl, and LiCl solutions of 1 M: dependence of friction on the applied load at a fixed scan rate of 2 m/s [70].

PAGE 69

69 Figure 3-15. Schematic representation of the hypothetical frictional mechanisms [70].

PAGE 70

70 Figure 3-16. Particle size distributions of colloid al silica slurry (Fuso PL-7) without salt and with 1 M LiCl and 1 M CsCl. 0.11 -2 0 2 4 6 8 10 12 14 16 18 PL-7 9.6 wt%, pH 7.3 No Salt 1 M CsCl 1 M LiCl Particle Size (m)Differential Volume (%)

PAGE 71

71 Figure 3-17. Material removal rate of silica substrates by CMP us ing diluted (9.6 wt%) colloidal silica slurries (PL-7) without salt and with 1 M LiCl and 1 M CsCl as a function of applied polishing pressure. 024681012 0 30 60 90 120 150 180 210 240 pH 7.3 Fuso PL-7 1 M CsCl 1 M LiCl No SaltMaterial Removal Rate (nm/min)Pressure (psi)

PAGE 72

72 CHAPTER 4 ROLE OF SURFACTANTS IN DEVLOPING S ELECTIVE PASSIVATION LAYER IN CMP In this chapter, a surfactant mediated passiva tion approach to increase STI CMP selectivity is discussed. Selective adsorption of a surfactant is necessary to develop selective passivation in CMP. It can be achieved if there is an adequate diffe rence in the surface charge characteristics of the substrates. This concept has been successfully used to achieve selective coating of surfactants in mineral flotation [73, 74]. Interactions between a solid surface and charged polar head of the surfactant molecule determine the adsorption st rength and the resultant adsorption density. In CMP, surfactants have been used not only to di sperse abrasive particles but also to create lubricating layers, yielding passivation agains t polishing. In the present study, an anionic surfactant, SDS, was used to create a selectiv e passivating layer only on the silicon nitride and not on the silica substrate. It is well known that isoelectric point (IEP ) of silicon nitr ide is higher than silica resulting in less negative potentia l for silicon nitride a bove the IEP [23]. The concentration of SDS was adjust ed to 16 mM, twice the critical micelle concentration (CMC), which has been shown previously to yield stable dispersion of silica abra sives in a CMP slurry [75]. For adsorption studies, silica fr om Geltech Co. and silicon ni tride from Ube Co. (SN-E10) were used to simulate silica and silicon nitr ide substrates. The particle size of silica was measured to be around 0.53 m by Coulter, and th at of silicon nitride, which was measured by centrifugal sedimentation was reported to be around 0.5 m by the manufact urer. Their specific surface areas were measured to be 8.1 m2/g and 10.4 m2/g, respectively by Quantachrome Nova 1200, BET surface area measurement technique. The specific surface area of abrasive silica particles was measured to be 34 m2/g by Quantachrome Autosorb 1C-MS. The Phoenix 8000

PAGE 73

73 UV-Persulfate TOC Analyzer was used to measure the SDS concentration. 99% sodium dodecyl sulfate (SDS) surfactants from Acros Organics Co. and Fisher Scientif ic Co. were used as received. 98% dodecyl alcohol from Eastman Koda k Co., 95% Sodium tetradecyl sulfate from Acros Organics Co. and Tween 80 from Fischer Sc ientific were also used, as received. High Selectivity Slurry Using Surfactants The addition of SDS to the slurry was found to result in a lower value of MRR of silica and silicon nitride in the entire pH range investigated in the pres ent study (Figure 4-1). However, significant increase in selective polishing of s ilica was measured below pH 3, yielding a selectivity of 25 as compared to state-of-the-art ceria abrasives of 5. Th e silicon nitride surface appeared to be fully passivated with the surfact ant layer at a pH below its IEP of pH 4.5, with minimal effect on silica CMP. The surface quality of substrates plotted in Figure 4-2 indicated that surfactant addition did not cause any additional defects m easured as root mean square (RMS) roughness. To understand the reasons for the observed selectivity, zeta potential and adsorption density measurements were conducted as a function of slurry pH (Figure 4-3). The IEP of silicon nitride and silica substrates we re measured to be about pH 4.5 and pH 2.2, respectively. The difference in the IEP results from the different surface groups constituting each material. As mentioned earlier, acidic silanol (SiOH) are th e major surface groups on silica, while the silicon nitride surface consists of basic amine (Si2NH) and acidic silanol (SiOH) groups [23]. These surface groups can acquire charge in aqueous solution according to following reactions: 1) (4 H SiO SiOH 2) (4 2 2 2NH Si H NH Si

PAGE 74

74 Consequently, zeta potential of silicon nitride is more positive due to the positively charged amine groups on its surface. The adsorption density of SDS was measured to be higher on silicon nitride than silica at a pH below their IEP. This is attributed to the resultant electrostatic in teractions between the substrate and surfactant molecules. At pH 2, the zet a potential of silicon n itride was measured to be +40 mV, whereas, that of silica was around +3 mV. Accordingly, the adsorption density on silicon nitride was determined to be six times higher than on silica resulting in complete passivation of the former. At pH values above the IEPs for both materials, there was still measurable adsorption on both materials, however the adsorption density on silicon nitride was higher probably due to more positive sites on silicon nitride from su rface amine groups. There have been several reports of SDS adsorption on the negatively charged silica surface. Hydrogen bonding and sodium ion mediated surfactant bonding are proposed as plausible mechanisms [76, 77]. In order to measure the eff ect of surfactant concentrati on on selectivity, polishing was conducted as a function of added surfactant conc entration (Figure 4-4). The MRR for both silica and silicon nitride started to decrease upon SD S addition and reached a minimum above 16mM. The maximal decrease in the MRR for silica was ar ound 20% from its original value, and that for silicon nitride was more than 90% resulting in 10 times higher po lishing selectivity than without surfactant addition. No further ch ange in the MRR or selectivity was observed once the added surfactant concentration exceeded 16mM. It ha s been reported that once the equilibrium concentration reaches CMC, no more adsorption changes are observed due to electrostatic repulsion between adsorbed micellar aggregat es and free micelles in solution [78].

PAGE 75

75 Surfactant Mediated Boundary Layer L ubrication for Selective Polishing Vakarelski et al. have shown that be yond the CMC of the cationic surfactant (dodecyltrimethylammonium bromide, C12TAB), there was no further decrease in the lateral force on silica substrate [22]. C onsequently, it is hypothesized that the maximum decrease in the MRR will occur when the bulk concentration reaches the CMC of SDS (around 8mM) [79]. However, in the present study, two times higher c oncentration of surfactant than the CMC was required to achieve maximum selectivity. Th e measurement of SDS adsorption on the CMP slurry as a function of pH showed that a bout 91% of added (16mM) SDS adsorbed on the abrasive particles at pH 2 as shown in Figur e 4-5. The area per molecule using the Gibbs adsorption equation, was cal culated to be around 70 2/molecule, which is higher area per molecule than the literature value of 53 2/molecule [79] at the liquid/gas interface. The possible reasons for the higher dosage of surfactant than expected are that the surfactant adsorption does not reach true equilibrium conditi ons due to process conditions encountered in CMP. This phenomenon may also be related to the dynamic aspects of surfactant. The reported for SDS is around 2.32 10-3 s [80]. However, accord ing to Patist and co workers, when 15 mM SDS was used for foaming experiments, the dynamic surface tension decreased as a function of bubble life time until it reached the saturation after about two seconds [81]. Recently, Philipossian et al. have reported the mean residence time of colloidal silica slurry between pad and substrate to be of the sa me order of a few seconds under the present experimental conditions [82]. A ssuming that other conditions are similar, the mean residence time in our study is expected to be 2 3 s econds. Considering that these two numbers are comparable, migration of surfactant to the newly formed substrate surface may be limited due to the high speed rotation of pad and wafer in CMP.

PAGE 76

76 The adsorption free energy is the driving force for surfactant adsorption and is the sum of various molecular interactions [ 78]. In the current study, it can be categorized into two categories, (i) interactions be tween the polar head of SDS and the surface through electrostatic and hydrogen bonding, and (ii) hydrophobic interacti ons between alkyl chains of adsorbed SDS molecules. By using the measured adsorption density plotted in Figure 4-5, and the radius of the SDS micelle (20 ) [79], the adsorption free energy of SDS on silica abrasive s was calculated to be -3.58 kcal/mol at pH 2 using modifi ed Stern-Graham equation [78]. 3) (4 kT G exp rC 2 o ads o where, is the adsorption density, r is the e ffective radius of the adsorbed ion, k is the Boltzman constant, Co is the bulk concentration, T is 298 K, and o adsG is the adsorption free energy. The electrostatic component of the adsorption free en ergy was calculated to be -0.76 kcal/mol using ze where, z is the valency of the adsorbate speci es, e is the charge of the electron, and the is the potential at the plane (assumed to be the zeta pote ntial). These calculations indicate that significant adsorption of the surfactant on th e abrasive particles is more favorable and may act as an additional energy barrier. Optimization of High Selectivity Slurry It is clear from the above disc ussion that the adsorption dens ity of surfactant molecules on the substrate is an important fact or in determining the slurry se lectivity. In order to reduce the required dosage of the surfactant, longer alkyl ch ain length surfactants were examined, since it was expected to exhibit better lu brication effects at a lower amount of added concentration. This is attributed to the formation of more comp act surfactant layers [75]. The MRR and polishing selectivity as a function of alkyl chain length of the sodium alkyl surfactant are plotted in Figure

PAGE 77

77 4-6. The surfactant concentration was selected to be twice the CMC value to compensate for the loss of surfactant due to adsorpti on on silica abrasive particles. As expected, SDS with longer alkyl chain length (C12) resulted in higher MRR decrease for silicon nitride with almost negligible effect on silica, thus yielding hi gher selectivity than sodium decyl sulfate (C10). However, when sodium cetyl sulfate (C14) was examined, there was a smaller decrease in MRR of silicon nitride resulting in lower selectivity. Considering that the Krafft point of C14 sodim sulfate (30 oC) [80] is higher than room temperat ure and higher than that of SDS (16 oC) [80], the surfactant was not completely solubilized and therefore failed to form a functional passivation layer. Another approach to decrease the dosage of the surfactant required to achieve desired selectivity involved using mixe d surfactant system (Tween 80/ SDS and dodecyl alcohol/SDS) at pH 2. The MRR and selectivity for the selected sy stems are plotted in Figure 4-7. In the case of dodecanol and SDS, selectivity was lower for th e mixed surfactant system than for 16mM SDS alone. It is possible that the a ddition of a small amount of dodeca nol promotes adsorption of SDS both on silica and silicon nitrid e. Although there was no appreci able change in the MRR on silicon nitride, the higher adsorption of SDS on silica also passivated its surface. It has been reported by Pala and co-workers that surfactant mixture of SDS and various nonionic surfactants can produce sy nergistic effects for dispersi on of slurry under high ionic strength conditions [34, 37]. Wh en 8mM Tween 80 was added to 16mM SDS, the MRR of silica was highly suppressed, whereas that of silicon nitride remained almost unchanged, thereby resulting in poor selectivity. It is well known that nonionic surf actant such as Tween 80, which has ethylene oxide groups (OC2H4), can adsorb on silanol gr oups (SiOH) on silica through hydrogen bonding [83]. These observa tions strongly suggest that a surfactant or surfactant

PAGE 78

78 system that exhibits strong preference only fo r silicon nitride is essential for developing surfactant-based high selectivity slurries. In summary, colloidal silica, which shows high dispersion stability in the range of pH 2 to 11, was utilized to develop a hi gh selectivity slurry. The additi on of SDS at pH 2 resulted in more than ten times higher selectivity than the conventional slurry. Additionally, AFM roughness measurement showed an acceptable surf ace finish. Adsorption density measurements revealed that there is a preferential higher adso rption of SDS on silicon ni tride, possibly due to electrostatic attraction, as comp ared to silica. The SDS adsorption results in differential passivation/lubrication and hence lower polishing efficiency of silicon nitride as compared to silica. The CMP characteristics ex amined as a function of added SDS showed that decrease in MRR and increase in selectivity leveled off at about twice th e surfactant CMC and remained unchanged, thereafter. The surfactant requirements appear to be driven by their adsorption primarily on silica abrasive part icles. To reduce the surfactant dos age, longer alkyl chain length surfactants were tested, which yi elded higher selectivity at lowe r dosage. However, the addition of a long chain length alcohol to substitute fo r the surfactant resulted in lower selectivity, probably due to higher adsorption of the surfactant on silica. Mi xed ionic and nonionic surfactant systems, on the other hand, resulted in poor sele ctivity due to passivati on of both silica and silicon nitride, although to a different degree.

PAGE 79

79 Figure 4-1. Influence of SDS addition on CMP performances: (a) Variation of material removal rate (MRR) as a function of slurry pH with and without 16mM sodium dodecyl sulfate (SDS), (b) Accompanyi ng selectivity of the slurry. 1234567891011 0 5 10 15 20 25 30 246810 1234567891011 Selectivity without SDS(b) SelectivitypH Selectivity with SDS-400 0 400 800 1200 1600 2000 2400 2800 3200 1234567891011 pH MRR with SDS SiO2 Si3N4 MRR Without SDS SiO2 Si3N4(a) Material Removal Rate ( /min )

PAGE 80

80 Figure 4-2. Surface finish of silica and silicon nitride substrates processed with standard and high selectivity slurry. Standard (pH 10.4)pH 216mM SDS at pH20.00 0.07 0.14 0.21 0.28 0.35 Roughness (RMS, nm) SiO2 Si3N4

PAGE 81

81 Figure 4-3. Variation of zeta pote ntial of silica and silicon nitride substrate and adsorption density of 16mM SDS on silica and silic on nitride powder measured by total organic carbon (TOC). 1234567891011 -160 -140 -120 -100 -80 -60 -40 -20 0 20 40 60 246810 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 SiO2 Si3N4Zeta Potential (mV) Adsorption density (mol/m2)pH SiO2 Si3N4

PAGE 82

82 Figure 4-4. Variation of MRR a nd accompanying selectivity of Klebosol slurry (12 wt%) as a function of added SDS concentration at pH 2. -30369121518212427 0 70 140 210 280 900 1200 1500 1800 2100 2400 2700 -30369121518212427 0 10 20 30 40 50 Selectivity MRR SiO2 MRR Si3N4 () SDS Concentration (mM)pH 2 Selectivity Material Removal Rate ( /min )

PAGE 83

83 Figure 4-5. Adsorption density of SDS on 12 wt% Kle bosol slurry with 16 mM SDS as a function of pH. 24681012 2.20 2.22 2.24 2.26 2.28 2.30 2.32 2.34 2.36 2.38 pH Adsorption density (mol/m2)

PAGE 84

84 Figure 4-6. Effect of alkyl chai n length of sodium alkyl sulfat e on MRR and selectivity at pH 2. The concentration was adjusted to 2 times the CMC to compensate the loss during CMP process. C10C12C140 30 60 2000 2100 2200 2300 2400 2500 MRR SiO2 MRR Si3N4 Selectivity C10C12C14 0 10 20 30 40 50 pH 2Selectivity Material Removal Rate (A/min) 66 mM C10 Sodium Sulfate 16 mM C12 Sodium Sulfate 4.2 mM C14 Sodium Sulfate Material Removal Rate ( /min)

PAGE 85

85 Figure 4-7. MRR and selectivity obtained by slurries with vari ous surfactant and surfactant mixtures at pH 2. Slurry A (16mM SDS) was included for comparison purpose. 0 30 60 1000 1500 2000 2500 0 10 20 30 40 50 pH 2Selectivity MRR SiO2 MRR Si3N4 Selectivity Material Removal Rate (A/min) A B C A = 16 mM SDS B = 0.8 mM Dodecanol/15.2 mM SDS C = 8 mM Tween 80/16 mM SDS Material Removal Rate ( /min )

PAGE 86

86 CHAPTER 5 ADSORPTION STUDY OF SODIUM DODECYL SULFATE ON SILICA Considering the relevance of surfactants in developing selective CMP slurries, it is important to understand their adso rption mechanisms on different substrates to optimize their performance as passivating agents. Accordingly, adsorption behavior of sodium dodecyl sulfate (SDS) on silica was studied. Special emphasis was placed on SDS adsorpti on on colloidal silica particles at high pH where both consti tuents exhibit negative charges. To measure adsorption density of SDS on collo idal silica particles, diluted Klebosol colloidal silica slurry (12 wt%) was prepared. Afte r dilution, the slurry pH was measured to be around 10.4. Suspension pH was adjusted with HCl and KOH solutions prepared with analytical grade substances purchased from Fisher Sc ientific Co. A proper amount of 100 mM SDS solution was added to obtain 1 to 5 mM SDS c oncentrations. Higher SDS concentrations were achieved by adding dry SDS powder. Adsorpti on density measurement incorporated the following steps: 1) add surfactan t to silica suspension, 2) magnetically stirring for 10 min, 3) centrifuge at 1500 rpm and 4) separate appropriate amount of supernat ant and dilute with nanopure water to yield a concentration with in calibration range (a round 50 ppm). Finally, 40 ml vials were loaded to total organic carbon (T OC) analyzer and measure the residual (bulk) SDS concentration in the supernatant. Concentr ation of adsorbed surf actant on particle was calculated from the difference in input concentr ation and residual bulk co ncentration. Specific surface area of colloidal sili ca was measured to be 34 m2/g by Quantachrome Autosorb 1C-MS. In order to gain insight into specific bi nding mechanisms, Fourier transform infrared spectroscopy (FTIR) measurements were c onducted. A nitrogen-purged Nicolet Magna 760 spectrometer equipped with a DTGS detect or was used to conduct FTIR analysis. FTIR/attenuated total re flection (ATR) method is well establ ished for its sensitivity to the

PAGE 87

87 surface property change [84-86]. Si nce it is well known that the surf ace of silicon is covered with silica by spontaneous oxidation, Si ATR crystal a nd surfactant solution were used to investigate the adsorption behavior. All the sp ectra were the results of 512 co -added scans at a resolution of 4 cm-1. Surfactant solutions of diffe rent concentration of SDS we re prepared at pH 10.4. During the measurement, the solutions were added to the Si ATR crystal assembly. After adding surfactant solution the sample chamber was purged with dry N2 gas to remove any residual atmospheric moisture and CO2. After 20 minutes of purging, CO2 peaks disappeared, however H2O peaks could not be eliminated. Adsorption Behavior of SDS on Silica Adsorption isotherm of SDS on colloidal silica suspension (Klebosol 1501-50, 12 wt%) measured at pH 10.4 is given in Figure 5-1. Desp ite the fact that both SDS surfactant and silica surface are negatively charged at this pH, the is otherm appears to be similar to that of electrostatic interaction domina nt adsorption behavior [87-89]. In region I, where adsorption density is not high, adsorption is assumed to occur by electrosta tic attraction. In region II, a sudden increase in adsorption is attributed to hemimicelle formation. In region III, there is a decrease in the rate of adsorpti on as indicated by change in the sl ope, which is ascribed to bilayer formation. Adsorption in region IV reaches a consta nt value apparently due to micelle adsorption on the surface [90]. In the current study, at low equilibrium surfactant concentrations up to 1.6 mM, the adsorption is very small due to electrostatic repulsion between SDS and silica substrate. However, there was a measurable adsorp tion prrobably due to hydrogen bonding. Beyond 1.6 mM, the adsorption increases sharply and may be attributed to attractive hydrophobic interactions between alkyl ch ains of surfactant resulting in hemimicelle formation. Beyond 8 mM, adsorption density leveled off. Critical he mimicelle concentration (HMC) and critical

PAGE 88

88 micelle concentration (CMC), which can be infe rred from the adsorption isotherm in Figure 5-1 occurred at around 1.6 mM and 8 mM, respectivel y. In the previously reported SDS-alumina system with a background electrolyte of 0.1 M NaCl, HMC and CMC were reported to be around 0.05 mM and 1.6 mM, respectively [91]. In the curre nt system, electrosta tic repulsion plays a dominant role in controlling adsorption behavior at lower su rfactant concentration, thereby resulting in relatively higher HMC on silica. Beyond HMC, hydrophobic attractive forces govern surfactant adsorption process. Under saturation adsorption conditions, the average area per molecule was calculated to be 41.6 2 from adsorption isotherm, whic h compares favorably to 53 2 reported at the air-water interface for SDS [80]. In the cas e of SDS-alumina system, it was calculated to be around 23.7 2 indicating the formation of more compact surfactant aggregates due to attractive electrostatic interactions between SDS and alumina The area covered by the adsorbed SDS molecules on silica particles was cal culated to be 663.2 m2, assuming the area occupied by one SDS molecule to be 53 2. Considering that the total ar ea of silica particles is 520.2 m2, surface coverage by SDS molecules indicates the formation of a bilaye r, if this surface is assumed to be homogenous, or micellar type adsorption, otherw ise. In the latter case, using th e reported aggregation number (64) and the radius of SDS micelle, 20 [79, 92], it was calculated that there are total 1.95 1019 micelles adsorbed onto silica pa rticles. Therefore, in the st eady state, 47.2% of the silica surfaces is covered with SDS micelles. Using a theo retical density of 2 g/cm3, 15.3 g of silica particles in the slurry, and the particle radius of 45nm, the number of s ilica particles in 100 ml slurry was calculated to be 2 1016. This value indicates that approximately 9.75 102 SDS micelles coat each particle.

PAGE 89

89 To further understand the adso rption of SDS on similarly ch arged silica, adsorption energy was calculated using modified Stern-Graham eq uation (Equation (4-3)). Calculated adsorption free energy under saturation adsorp tion conditions was found to be -2.9 kcal/mol indicating that primarily physical adsorption is resp onsible for SDS adsorption on silica. In order to assess the effect of pH on SDS adsorption on silica, measurements were conducted at 1.6 mM and 16mM concentrations Results plotted in Figure 5-2 show the adsorption behavior of SDS on silica correlates we ll the zeta potential of silica indicating that surface charge of the silica pl ays an important role in SDS adsorption. Adsorption energy calculations revealed that at 1.6 mM SDS con centration, adsorption energy at pH 10.4 is 0.02 kcal/mol as compared to -1.17 kcal/mol at pH 2, indicating an energetica lly unfavorable process at pH 10.4 and a favorable one at pH 2. At 16mM, the electrostatic effect was probably dominated by the increased hydrophobic attractive interactions between al kyl chains resulting in adsorption energy of -3.14 kcal/mol at pH 10.4 a nd -3.59 kcal/mol at pH 2 indicating favorable adsorption at both pH values. Structure of Adsorbed SDS Molecules To investigate the structure of adsorbed SD S molecules on silica su rface, zeta potential was measured as a function of added SDS concen tration at pH 10.4 (Figur e 5-3). At very low concentration of SDS (region I), zeta potenti al essentially remains unchanged. As the concentration increases (region II), sodium i ons are adsorbing on the silica surface resulting in less negative zeta potentia l. It is hypothesized that there are surfactant molecules weakly bonded to sodium ions. Surfactant molecules associated with sodium ion will not exhibit significant impact on the zeta potential measurements due to mutual charge neutralization. As the surfactant concentration increases, hemimicelles form and grow in size in region III, and the slope of zeta potential increase becomes smaller than region II. It seems that free SDS starts to adsorb on the

PAGE 90

90 hemimicell coated surface forming bilayers in region III resulting in a lower slope change. Finally in region IV, when surfact ant aggregates form micelles, zeta potential reversal occurs by incorporating a number of free monomers in the solution. When a background electrolyte was added to the system, overall zeta potential was less negative and the slope change was less pronounced, but a similar trend was observed. A bove hypothesis is shown schematically in Figure 5-4. The correlation between SDS adsorption and ze ta potential is clear er from Figure 5-5, where adsorption density and zeta potential ar e co-plotted as a function of equilibrium concentration of SDS. The change in zeta pote ntial follows the adsorption isotherm and zeta potential reversal occurs at CMC. However, due to the low surface coverage of micelles (4%) on silica surfaces the change was not significant. SDS adsorption behavior on silica, as determ ined in the present study, is contrary to electrostatic considerations, since both the s ubstrate and surfactant molecules are similarly charged. There have been several reports of SDS adsorption on negatively charged silica or sepiolite, a hydrated magnesium silicate (Si12Mg9O30(OH)6(OH2)4H2O) [76, 77, 93]. Possible mechanisms for this observation were hydroge n bonding between silanol groups and SDS, and counter ion mediated surfactant adsorption. Several noticeable th ermodynamic properties of the surfactant were reported by zdemir et al. throug h the adsorption study of SDS on sepiolite. At saturation adsorption, the adso rption free energy calculated from Frumkin model was -3.1 kcal/mol at 25 oC [93]. It is comparable to the results in the present study (-2.9 kcal/mol). This low energy of adsorption indicate s that weak physical forces are responsible for adsorption. Calculated adsorption free energy from the adso rption isotherm in Chandars report was -4.18

PAGE 91

91 kcal/mol, indicating that the dr iving force for adsorption involv ing electrostatic attraction is higher than hydrogen bonding alone. To further investigate the mechanism of SDS adsorption on negatively charged silica particles, FTIR ATR (attenuated total reflection) measurements were conducted. It should be mentioned that quantitative analysis by FTIR is not very accurate and it is not well understood how the electrostatic interaction affects the FT IR spectra. On the other hand, adsorption of various molecules via hydrogen bonding has be en well observed and documented [94, 95]. One of the noticeable research on hydrogen bonding be havior for silica and dibenzodioxin was done by Guan et al. [95]. They reported that as the adsorption of dibenzodioxin on silica surface increased, the peak of geminal silanol group de creased and that of is olated silanol group increased, indicating that the mo lecular adsorption occurs at th e expense of the silanol group by hydrogen bonding. Their measurement was done using dry powder samples. In the current study, all the measurements were conducted in aqueous surfactant solution by using ATR crystal. Figure 5-6 shows the spectra of SDS at 1, 2.5, 5 and 10 mM concentration in the CH2 stretching region measured at pH 10.4. As was discussed by Pankaj et al., the absorbance intensity increased up to 5mM, a nd it decreased at 10mM, which is higher than CMC of SDS (8 mM) [96]. The reason for the decrease of absorbance intensit y upon micelle formation is not well understood, however, it confirms the adsorption of SDS on silica surface at high pH. Due to the overlapping of the peaks from silanol groups a nd water, changes in si lanol groups were not confirmed in this experiment. SDS adsorption on silica can also impact th e dispersion stability. Figure 5-7 shows the particle size distribution of Stber silica without and with SDS 12 hours after the pH was changed to 2. Without SDS, there was an additio nal peak due to particle coagulation since the

PAGE 92

92 isoelectric point (IEP) of the si lica particle is known to be around pH 2.7. However, with SDS addition, no additional peaks were observed. In summary, SDS adsorption beha vior at low concentration was small due to electrostatic repulsion, however, limited adsorption was obser ved due to hydrogen bonding. At intermediate concentrations, it was hypothesized that sodium ion mediated charge ne utralization along with hydrophobic attractive force resulted in higher adso rption of SDS. The slope of the adsorption density decrease is attr ibuted to bilayer formation. Adsorp tion free energy calculations and zeta potential measurements as a function of SDS co ncentration were supportive of the proposed hypothesis. It was observed that SDS adsorption on silica surface resulted in a stable dispersion.

PAGE 93

93 Figure 5-1. Adsorption isotherm of SDS on collo idal silica (Klebosol 1501-50, 12 wt%) at pH 10.4. Critical hemimicelle concentration (HMC ) and critical mice lle concentration (CMC) was marked. 10010110210-810-710-6 pH 10.4 HMC IV III II ICMC Adsorption density (mol/m2)Bulk Equilibrium SDS Concentration (mM)

PAGE 94

94 Figure 5-2. Adsorption density of SDS on colloid al silica (12 wt% Klebosol 1501-50) at SDS concentration of 1.6 mM and 16 mM and zeta potential as a function of pH. 246810 0.0 2.0x10-84.0x10-82.0x10-62.2x10-62.4x10-62.6x10-6 246810 -100 -80 -60 -40 -20 0 Zeta potential (mV) Adsorption Density (mol/m2)pH 1.6 mM SDS 16 mM SDS Zeta potential

PAGE 95

95 Figure 5-3. Zeta potential of Kl ebosol slurry as a function of SDS concentration at pH 10.4. -74 -73 -72 -71 -70 -69 -68 -67 -66 -65 100101102 Klebosol 12 wt% no Salt Zeta potential (mV)Concentration of SDS (mM) pH 10.4IV III II I 1 mM NaCl

PAGE 96

96 Figure 5-4. Pictorial depictions of the possible surfactant aggreg ates films at concentrations corresponding to I-IV in Figure 5-3. Na Na Na Na Na Na S S i i O O2 2 Na Na Na Na Na Na Na Na Na S S i i O O2 2 Na Na Na Na Na Na Na Na Na Na Na Na Na S S i i O O2 2 Na Na Na Na Na Na Na Na Na Na Na Na Na S S i i O O2 2 Na I II IV III

PAGE 97

97 Figure 5-5. Adsorption characteris tics of SDS on Klebosol silica slurry and zeta potential as a function of concentration of SDS at pH 10.4 10010110-810-710-6 -72.5 -72.0 -71.5 -71.0 -70.5 -70.0 -69.5 -69.0 HMC CMC Adsorption Density (mol/m2)Bulk Equilibrium SDS Concentration (mM) Adsorption densitypH 10.4 IV III II I Zeta potential

PAGE 98

98 Figure 5-6. FTIR/ATR Spectra of SDS solution at 1, 2.5, 5 and 10 mM bulk concentration in the CH2 stretching region (2921, 2924) measured at pH 10.4 using Si ATR crystal. 3100 30002900 2800 2700 Wavenumber (cm1 )0.000 0.002 0.004 0.006 0.008 0.010 0.012 pH 10.4 SDS 1mM 2.5mM 5mM 10mM log (1/ R )

PAGE 99

99 Figure 5-7. Particle size di stribution of Geltech SiO2 at pH 2 with and without 16 mM SDS 12 hours after pH change. 0.51.01.52.02.53.0 0 5 10 15 20 25 30 pH 2 Differential volume (%)Particle size (m) without SDS with SDS

PAGE 100

100 CHAPTER 6 APPLICATION OF DENSITY FUNTI ONAL THEORY BASED MODELING FOR SURFACTANT ADSORPTION STUDY There have been numerous modeling efforts to develop reliable tools for predicting colloidal systems behavior. There are two broad areas of modeling for investigating the structure of molecules and their reactivity: molecular m echanics and electronic st ructure theory. They perform the same basic calculations: i) compute the energy of a particul ar molecular structure and ii) geometry optimization to produce the lowest energy molecu lar structure [97]. In addition, electronic structure model is capable of calculating vibrational frequencies of molecules resulting from interatomic motion. Molecular mechanics based models use the laws of classical physics, and each one is characterized by its particular forc e field. In general, it does not expl icitly treat the electrons in a molecule. They perform computations based on the interactions am ong the nuclei, while interactions involving electrons are implicitly in cluded in force fields through parameterization. This approximation enables the molecular mechanic s modeling to be fast and cost effective, and applicable to large systems. Ho wever, it also has several limita tions, e.g., each force field is system specific, and it is unable to calculate chemical problems where electronic effects predominate (i.e. bond formation and breakage), sin ce interactions among el ectrons are neglected [97]. Electronic structure methods use the laws of quantum mechanics. There are two major classes in the area, i) semi-e mpirical methods such as AM1 and PM3, which utilize parameters derived from experimental data, ii) ab initio methods, which utilize no experimental parameters, instead, computations are based solely on the la ws of quantum mechanics and the values of several physical constants. Semi-empirical calc ulations are relatively inexpensive and produce

PAGE 101

101 reasonable qualitative descriptions, while ab initio modeling can provide high quality quantitative predictions for a br oad range of systems [97]. Recently, newly developed electronic structur e methods termed density functional theory (DFT) methods, have been widely used. DFT me thods are attractive si nce they include the effects of electron co rrelation, while pure ab initio methods take it into account in an average sense. In general, electronic st ructure methods are known to requi re high computational time and are relatively costly. Many efforts to model collo idal systems use molecular mechanics methods, since the system involves relatively large molecules or molecules/particles in water, which is too large to be calculated by elec tronic structure model [98, 99]. However, electronic structure methods are also applied to coll oidal systems in many cases due to their ability to produce FTIR and Raman spectra and that they can be used w ithout parmeterization-a must have for molecular mechanics modeling [100-102]. In this study, DFT method was applied to th eoretically calculate adsorption and compare it with the experimental data. Methodologies DFT methods compute electron co rrelation via general functiona ls of the electron density. DFT functionals partition the electronic energy into several components which are computed separately: the kinetic energy, th e electron-nuclear interaction, the Coulomb repulsion, and an exchange-correlation term accountin g for the remainder of the el ectron-electron interactions. Various DFT methods are distingu ished by the way that they trea t the exchange and correlation term. In addition, there are several hybrid f unctionals, which combines the component of ab initio method and DFT methods. Among them the best of these hybrid functionals is Becke-style 3-Parameter Density Functional Theory using th e Lee-Yang-Parr correlation functional (B3LYP). This method has proven to be superior to the tr aditional functionals in terms of computational

PAGE 102

102 cost and accuracy of predicting experimental results. Due to these advantages, B3LYP method was utilized throughout th e current study [97]. Another important component of the theore tical calculation by electronic structure method is basis set. Basis set is the mathematical description of the orbitals within a system used to perform the theoretical calculation. Standard ba sis sets for electronic structure calculations use linear combinations of Gaussian functions to fo rm orbitals. Basis sets assign a group of basis functions to each atom within a molecule to appr oximate its orbitals. Ther e are several basis sets [97]: Minimal basis sets (STO-3G): they contain th e minimum number of basis functions needed for each atom. They use fixed size atomic-type orbitals and three Gaussian primitives per basis function. Split valence basis sets (3-21G and 6-31G): they increase the number of basis functions per atom. Split basis sets have two or more sizes of basis function for each valence orbital. They allow orbitals to change si ze, but not to change shape. Polarized basis sets (6-31G* (6-31G(d)) and 631G** (6-31G(d,p))): they allow orbitals to change shape. They add d functions to car bon atoms and f functions to transition metals. 6-31G(d) basis set, which adds d function to heavy atoms is becoming very common for calculations involving up to medium-sized systems. Another popular basis set is 631G(d,p). It adds p functions to hydrogen atom s in addition to the d functions on heavy atoms. In the current study, medium sized 6-31G* basis set was utilized for both optimization, single point energy calculation, and fr equency calculation in the curr ent study, consider ing calculation time. The most unique and useful advantage of th e electronic structure methods over molecular mechanics model is its ability to do frequency ca lculation. It can serve a number of different purposes, such as to predict Raman and IR sp ectra of molecules and to produce thermodynamic properties such as free energy, enthalpy and en tropy of the system. Energy calculations and geometry optimizations discussed so far ignore the vibrations in molecular system, which is not

PAGE 103

103 true of real systems. In geometrically optim ized states, these vibr ations are regular and predictable, and molecules can be identified by their characteristic spectra. Specifically, it can predict the direction and magnitude of the nucle ar displacement that occurs when a system absorbs a quantum of energy. In addition, all fr equency calculations include thermodochemical analysis of the system. By default, it is ca rried out at 298.15 K and 1 atmosphere of pressure. Gaussian which is used in the present study provides thermal correction for enthalpy and free energy through frequency calculation. Using the sums of electronic a nd thermal enthalpies and electronic and thermal free energies, enthal pies and free energies of reactions can be calculated for a model system as described below [103]. 1) (6 D C B A 2) (6 ) ) H ( ) H (( ) H ( ) H ( ) H ( ) H ( ) K 298 ( HB corr o A corr o D corr o C corr o ts tan reac corr o products corr o o r 3) (6 ) ) G ( ) G (( ) G ( ) G ( ) G ( ) G ( ) K 298 ( GB corr o A corr o D corr o C corr o ts tan reac corr o products corr o o r where, o is the total electronic energy, Hcorr and Gcorr are thermal correction of enthalpy and free energy, respectively produced through fre quency calculations. The above method works since the number of atoms of each element is th e same on both sides of the reaction, hence all the atomic interactions cancel out, requiring only mo lecular data for calculation. Adsorption free energy presented in the current study was calculated using the above methods. The above calculations are carried out in vacuum. The properties of molecules and transition states can differ considerably between the gas phase and in solution. For example, electrostatic effects are much less important for species placed in a solvent with a high dielectric constant than they are in the gas phase. There are methods developed to incorporate the solvent effects. All models consider the solvent as a continuum of uniform dielectric constant and the

PAGE 104

104 uniform reaction field. The solute is placed into a cavity within the solvent. The effect of polarization of the solven t continuum is determined numerically [97]. The optimization of each molecule and comp lex structure was done using 6-31G* B3LYP self-consistent field (SCF) computation. For comp lex structures, adsorben t and adsorbate were placed close to each other, around 5 initially. Af ter optimization, according to the interactions between two molecules, the bond length and angl e between them was adjusted. Single point energy and adsorption free energy calculations were done after structure optimization in vacuum state. Adsorption (electronic) energy of the complex is calculated by the following equation: 4) (6 ) E E ( E EA S SA ads where, ESA is the energy of the complex struct ure of adsorbate-adsorbent pair, and Es and EA are the energies of the constituent molecules. Frequency calculation was conducted to calculate adsorption free energy. Single poi nt energy calculation by polar ized continuum model (PCM) was performed on optimized model system in vacuum to calculate the energy in aqueous environment. Structures and Resources The basic molecular structures used in this research are shown in Figure 6-1. To model silica and silicon nitride surface, SiO4 and SiN4 tetrahedral units were used, which are the minimal building blocks of silica and silicon nitride. There are se veral reasons for selecting these structures, firstly simple struct ure can save time and computati onal resources, se condly, as the structures become complicated, optimization with 6-31G* basis se t was not always possible. The importances of the frequency calculation are discussed in detail in a later section. The optimization of the structure is the prerequisite for frequency calculation. It was one of the most important considerations for the current study. Ad ditionally, when a complicated structure (i.e. 1-

PAGE 105

105 dimensionally bonded five tetrahedrons of silica) was used, deformation of original structure occurred as a result of optimization providing multiple bonding sites with other molecules, which is not possible in a real system. Even t hough there are multiple bonding sites, calculated adsorption energy values had the same order of magnitude. Finally, materials for the experimental study are amorphous, which have only short range or der individual te trahedron is bonded to each other randomly. Therefore, actual re peat unit of silica and silicon nitride can be considered as individual tetrahedron. The basic structure used in this study consists of one Si and four oxygens for silica and one Si and four nitrogens for silic on nitride. All oxygens and nitroge ns are protonated to represent the condition at isoelectric point (IEP). To simula te different pH values individual tetrahedron was deprotonated or protonated to be positively or negatively charged. Zhmud et al. reported similar approach using ab initio, Hartree-Fock (HF) method on clusters of -Si3N4 and -cristobalite structures and nitrogen gas as an adsorbate [102]. Although they were able to succes sfully correlate the experime ntal results and theoretical prediction, their adsorption energy values were all positive due to non-equilibrium structures. There also have been reports using simple stru cture similar to the cu rrent study [104, 105] that were successful in prediction of FTIR spectra wi th small deviations from experimental results, and can be considered to be in support of our approach us ing minimal sized structures. Sodium dodecyl sulfate (SDS) structure was used without any modification. Triton X-100 (TX-100, C14H22O(C2H4O)n, n ~ 10) structure was simplified to n = 1 to reduce calculation time. All the computations for the current study we re done by Gaussian on a node of 4 AMD Opteron Cores (2 x 275, 2.2GHz) with 4 GB DDR400 RAM in Unix system. Computational time

PAGE 106

106 varied according to the complexity of the syst em. Usually, it took several hours for optimization and frequency calculation for complex structure of surface and surfactants. Results and Discussion After structure optimization, the bond length of Si-O in Si(OH)4 and Si-N in Si(NH2)4 was measured to be around 1.660 and 1.746 respectiv ely. Experimentally determined values for the bonds were reported to be 1.6041 ~ 1.6066 and 1.704 ~ 1.767 for quartz and -silicon nitride, respectively [106, 107]. It appears that despite the us e of minimal structure, the optimized bond length values are comparable to the crystallographic data. This may be attributed to the high symmetry structures used in the current study. Electronic energies for different complex molecu lar structures were calculated in vacuum, in water were using polarizable continuum model (PCM), and th e Gibbs adsorption free energies were calculated with the method described before. These data are presented in Table 6-1, and the experimental values are listed in Table 6-2. Due to the lack of data for Triton X-100 (TX-100), adsorption free energy was calculated from the da ta reported by Denoyel and co-workers [108]. In further discussion, sodium dodecyl sulfat e (SDS) surfactant will be denoted as Na+ and DSto describe the dissociated state in aqueous solution. Negative energy values indicate that interaction (adsorption) between two molecular species is energeti cally favorable and vice versa. Calculated adsorption free energy, which consid ers electronic vibrati onal motion at room temperature, showed values similar to the electronic energy calculated in vacuum at O oK implying that the contributions from the motion and elevated temperature are very small. In contrast, PCM correction resulted in significant effect on the ener gy values. As mentioned in the methodologies section, polarizable water medium reduced the energy significantly providing data more comparable to the experimental results.

PAGE 107

107 SDS Adsorption on Silica at, below, and above the Isoelectric Point (IEP) Initially, Si(OH)4 at isoelectric point (IE P) of silica and in pr esence of surfactant (DS-) was simulated. After structure optimization, DSadsorbed on neutral Si(OH)4, forming a bond between one of the oxygens in su rfactant and a hydroge n on silica. It was found that a relatively weak bond (Eads(PCM) = -8.46 kcal/mol) is responsible for the adsorption and equilibrium bond length was 1.881 (Figure 6-2). The low energy va lues or weak interactions indicate that electrostatic force or hydrogen bonding is responsible for adsorpti on. However, considering that the surface neutrality, it may be attributed to the hydrogen bonding. When compared to the experimental results, PCM energy calculation resulte d in the same order of magnitude but much higher value. This could be due to an intrinsic error when elec tronic structure method deals with low energy interactions. Below IEP of silica, when SiO4H5 + interacts with DS-, electrostatic attraction as well as hydrogen bonding is plausible (Fig ure 6-3). The additional hydroge n is more tightly bonded to the DSthan the surface. The bond length was cal culated to be about 1.001 The bond length of hydrogen bonding was calculated to be 1.895 which is similar to the neutral surface. Due to these two interactions, the cal culated PCM energy was much hi gher (-20.97 kcal/mol) than the neutral surface at IEP. However, experimenta lly the difference was not noticeable, since the measured zeta potential had a negative value even at pH 2. It is likely th at in a real system, SiO4H5 + is difficult to exist and neutral Si(OH)4 (considered to be neutra l silanol group) is the major surface species at low pH. Above IEP of silica, when SiO4H3 interacts with DS-, structure optimization was not possible since both of the surfaces and DSmolecu les kept diverging away from each other to lower the energy of the system (Figure 6-4). Th is must be caused by th e electrostatic repulsion between them. Although the energy is not calcul ated from optimized structure, the positive

PAGE 108

108 energy value calculated in vacuum indicates that th e adsorption is not favorable in this case. Due to the non-equilibrium structure, PCM energy showed slightly negative value, however the final geometry clearly indicates that the adsorption is not favorable. Calcul ated energy from the experimental data was also positive (above IEP at SDS concentration of 1.6 mM at pH 10.4) and is similar to the theoretical prediction. The interaction behavior dr amatically changed when Na+ ion was introduced into the system (Figure 6-5). Sodium ion resided betw een two species resulting in an equilibrium complex structure. The two molecules formed tighter bond than hydrogen boding and the energy (-22.05 kcal/mol) was similar to the value from a ttractive electrostatic in teractions below IEP. Sodium ion was found to be more tightly bonde d to the surface. The bond length between Na+ and Oon surface was around 2.184 and that of Na+ and Oin DSwas around 2.34 ~ 2.36 SDS Adsorption on Silicon Nitride at IEP Current theoretical approach was attempted to predict selective adsorption of surfactant onto a specific surface. For this purpose, silicon nitride tetrahedron structure was constructed and protonized for charge neutrality (Figure 6-6). In the case of silicon nitride surface, optimization of ionized structure was not possi ble probably due to instability of the structures. Therefore, comparison between with neutral silicon ni tride and silica at IEP is attempted. At IEP of silicon nitride, when Si(NH2)4 interacts with DS-, hydrogen bonding occurred with a bond length of 2.158 Ca lculated PCM energy was determined slightly lower (-8.29 kcal/mol) than with Si(OH)4. This trend is not consistent w ith the experimental results, where silicon nitride showed higher adso rption density at pH 2. Regarding this discrepancy, it should be noted that the current modeling is to simulate th e interactions of surfact ant and the representative of the major surface reactive sites under given cond itions. In reality, each surface has a different charge density due to different number of active sites constituting the surface.

PAGE 109

109 TX-100 Adsorption on Silica at IEP To verify the applicability of the current appr oach to another surfact ant, adsorption of TX100, a nonionic surfactant, onto neutral Si(OH)4 was investigated (Figur e 6-7). It is well known that TX-100 adsorbs onto a silica surface at different pH levels via hydrogen bonding [108]. After structure optimization, they were determined to form equilibrium complex structure. Low values of calculated PCM energy indicate form ation of hydrogen bonding on the same order of magnitude as the experimental results. The bond length between oxygen in SiOH4 to hydrogen in TX-100 and oxygen in TX-100 to hydrogen in SiOH4 was measured to be 1.9 and 1.849 respectively, similar to the value of DSand SiOH4. It is known that the electronic structure method is not capable of accurately describing low energy interactions [101, 109]. C onsequently, calculated energy values from the DFT method are usually overestimated [109]. In addition, calcu lated intermolecular in teractions are also overestimated due to the basis set superimpos ition error (BSSE) [102]. Th ese intrinsic errors seem to also exist in the current study. Various approaches have been attempted to reduce the calculation errors. Volkov et al. reported that electrosta tic energy can be successfully calculated with minimum error when DFT B3LYP was used w ith higher order basis sets. To reduce BSSE, counterpoise (CP) correction, which calculates each of the units usi ng just the basis functions of the other, has been reported to be successful [102]. To refine our cu rrent modeling study, these elements needs to be incorporated to give more realistic energy values. The ideal setting would be frozen surfaces that do not change their ge ometry after optimization, with multiple surface sites, using B3LYP and a higher order basis set than 6-31G*, with CP correction for calculating complex structures. However, this may significantly increase computational costs. Overall, the current relatively simple setting has been show n to describe the given system with reasonable success.

PAGE 110

110 In summary, electronic structure method B3 LYP with 6-31G* basis set was applied to describe the interaction of molecules with different surf aces. Among various energy values calculated, PCM energy, which takes into account the presence of solvent (water), provided the most realistic results. It was found that adsorption of DSonto neutral silica (at IEP) and silicon nitride occurs via hydrogen bonding, and the positively charged s ilica surface and DS(below IEP) resulted in a stronger ad sorption energy due to electrosta tic attraction. The adsorption of DSonto negatively charged silica surface (above IEP) was not energetically favorable. However, the introduction of Na+ greatly facilitated these interactions yielding similar adsorption energy values as the electrostati c attraction. Adsorption of DSonto the neutral silicon nitride surface was also attempted. It was found that adsorpti on is energetically favorable via hydrogen bonding, however, calculated values did not successfully correlate to the selectivity of surfactant adsorption measured experimentally. To verify the validity of the current approach, adsorption behavior of another surfactan t, TX-100, onto neutral silica wa s also investigated. It was determined that the adsorption was favorab le via hydrogen bonding and the theoretical predictions agree well with the experimental results reported for this system. Computational optimization of the structures in the presence of solvent, and the interactions of surfactant with the resultant surface systems, is required to more realistically simulate systems of practical significance.

PAGE 111

111 Table 6-1. Adsorption energy (kcal/mol) calculat ed by density functional theory (DFT) based method (B3LYP) using 6-31G* basis set. All the values were obtained after optimization with the same basis set. Eads (Vacuum) is the electronic energy calculated in vacuum, Eads (PCM) is the electronic ener gy calculated with polarizable continuum model (PCM) correction, which cons iders the effect of water solvent, and Gads (Vacuum) is the adsorption free energy calculated in vacuum. SiO2 Si3N4 pH Condition at IEP below IEP above IEP at IEP at IEP Complex Si(OH)4 /DSSiO4H5 + /DSSiO4H3 -/DSSiO4H3 /Na+/DSSi(OH)4 /TX-100 Si(NH2)4 /DSEads (Vacuum) -30.05 -142.73 4.87 -63.61 -12.72 -16.76 Eads (PCM) -8.46 -20.97 -0.56 -22.05 -8.20 -8.29 Gads (Vacuum) -17.15 -128.44 N/A -52.55 -0.91 -5.95

PAGE 112

112 Table 6-2. Adsorption free energy (kcal/mol) of SD S on silica calculated from adsorption density data in Ch. 5 at different pH and two di fferent added concentrations (1.6mM and 16mM). The energy of Triton X-100 (TX-100) on silica at pH 2.2 calculated from data in Ref. [108]. SDS TX-100 Gads (kcal/mol) pH 2 pH 3 pH 10.4 pH 2.2 1.6 mM -1.17 -1.16 0.018 16 mM -3.59 -3.59 -3.14 -2.79

PAGE 113

113 Figure 6-1. Optimized (a) Si(OH)4, (b) Si(NH2)4, (c) Sodiumdodecyl sulfate (SDS), and (d) Triton X-100 (TX-100) struct ure using B3LYP method a nd 6-31G* basis set. (a) (b) (c) (d)

PAGE 114

114 Figure 6-2. Optimized SiOH4 and DScomplex structure using B3 LYP method and 6-31G* basis set.

PAGE 115

115 Figure 6-3. Optimized SiOH5 + and DScomplex structure usin g B3LYP method and 6-31G* basis set.

PAGE 116

116 Figure 6-4. Sturcture of SiO4H3 and DScomplex. Optimization is not complete, since two molecules are being separated to decrease energy.

PAGE 117

117 Figure 6-5. Optimized SiO4H3 -, Na+, and DScomplex structure using B3LYP method and 631G* basis set.

PAGE 118

118 Figure 6-6. Optimized Si(NH2)4 and DScomplex structure using B3LYP method and 6-31G* basis set.

PAGE 119

119 Figure 6-7. Optimized SiOH4 and TX-100 complex structure using B3LYP method and 6-31G* basis set.

PAGE 120

120 CHAPTER 7 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK Conclusions Colloidal silica, exhibiting high dispersion stabil ity in the pH range of 2 to 11, was utilized to develop a high selectivity slurr y. The addition of SDS at pH 2 resulted in more than ten times higher selectivity than conve ntional slurry. Additionally, AFM roughness measurements indicated acceptable surface finish. Adsorption de nsity measurements revealed that there is preferentially higher adsorption of SDS on silic on nitride possibly due to more favorable electrostatic interactons as compared to silica. SDS adsorption behavior is believed to result in differential passivation (lubrication) and hence, lower polishing efficiency of silicon nitride as compared to silica. The CMP characteristics exam ined as a function of added SDS concentration showed that the decrease in MRR and increase in selectivity leveled off at about twice the surfactant CMC and remained unchanged thereafter. Surfactant requirement appears to be driven primarily by adsorption on silica abrasive particles. It app ears that selective surfactant passivating coatings on substrates can yield higher selectivity without any adverse impact on surface finish. Attempts to economize on surfact ant amount revealed that longer alkyl chain length surfactants yielded higher selectivity at a lower dosage. However, the addition of long chain length alcohols as a substitute for the surfac tant resulted in lower selectivity possibly due to passivation of silica. Mixe d ionic and nonionic surfactant syst ems resulted in poor selectivity due to larger decrease in MRR of silica and sm aller decrease for silicon nitride. These findings can be used as a guide for develo ping selective polishing CMP slurry. In order to find an alternative method to achieve high selectivity, the e ffect of different salt addition on CMP performance was investigated. Passivation (lubrication) effect by hydrated cations was found not to be a dominant factor in determining MRR. However, slurry with LiCl

PAGE 121

121 yielded lower MRR than one with CsCl, sugge sting that passivation by hydrated cations is dependent on the nature of the added ions. Attempts to study the role of ma terials properties in STI CMP us ing a colloidal silica slurry indicated that differences in Young s modulus of the substrate materi als play an important role in MRR since modulus is related to the bond strength of the materials. Electrostatic repulsive forces imposed by pH cha nge in the slurry play s a dominant role in determining MRR of silica and silicon nitride substrates. These interactions were also manipulated by monovalent salt addition to the slurry. A linear relationship between MRR and electrostatic forces implies that such repulsive interactions pr obably resulted in governing the number of particles engaged in the polishing process. In order to assess the effect of dissolution in CMP, dissolu tion studies were conducted by measuring thickness changes of substrates im mersed in 0.1 M NaOH solution. The dissolution rate of silica was found to be much higher than that of silicon nitride. However, it was too low to make any significant contribution to MRR. It seems that the att ack of hydroxyl ions at high pH levels is responsible for the high MRR and poor su rface finish due to formation of a softer top layer. SDS adsorption behavior on silica was investig ated systematically. At low concentration, SDS adsorption was opposed by electrostatic repu lsion, however, some adsorption did occur and was attributed to the hydrogen bonding. At interm ediate concentrations, it was hypothesized that sodium ions mediated charge ne utralization along with hydrophobic attractive forces resulted in higher adsorption density. Adsorption free energy calculations and zeta potential measurements as a function of SDS concentra tion seem to support the proposed mechanism. It was observed

PAGE 122

122 that SDS adsorption on silica surface, even at a pH below its IEP, resulted in stable dispersion. FTIR/ATR measurements conf irmed the adsorption of SDS on silica substrate. The electronic structure modeling method B3LY P with the 6-31G* basis set was applied to assess the interaction of molecules with di fferent surfaces. Among various energy values calculated, PCM energy, which takes into account the presence of solvent (water), provided the most relevant results. It wa s found that adsorption of DSonto neutral silica (at IEP) and silicon nitride occured via hydrogen bonding and interact ion between positively ch arged silica surface and DS(below IEP) resulted in hi gher adsorption energy due to electrostatic attraction. The adsorption of DSonto negatively charged silica surface (above IEP) was not energetically favorable. However, the introduction of Na+ greatly facilitated these interactions yielding adsorption energy values similar to those of the electrostatic attr action. Adsorption of DSonto the neutral silicon nitride su rface was also attempted. It was found that adsorption is energetically favorable via hydrogen bonding, however calculated values did not successfully correlate with the selectivity of surfactant adsorption as measur ed experimentally. To validate the current modeling approach, adsorption behavi or of another surfactant, TX-100 (nonionic), onto neutral silica was in vestigated. It was determined that the adsorption occurred via hydrogen bonding and theoretical predictions agreed with the experimental results for this system. Additional efforts such as computational optimizati on of structures in the presence of solvent and interactions of surfactant with the resultant su rface systems are needed to simulate practical systems. Suggestions for Future Work Frictional force measurements using lateral fo rce microscopy are needed to investigate and optimize the formation of surfactant mediated se lective passivation coatings. The effectiveness of the high selectivity slurry developed in th e current study has been limited to chemical

PAGE 123

123 mechanical polishing around pH 2. Development of robust high selectivity slurry for CMP at higher pH levels is needed for practical applications. It has be en well documented that alkali metal introduced to semiconductor device can be a source of device failur e. The high selectivity slurry developed in the present investigation involves sodium ions, which can be a reason for device failure. To avoid this possibility, applicatio n of ammonium based surfactants is suggested. Besides surfactants, there are a number of ot her potential passivating agents such as polymers, surface complexing agents, and protei ns that exhibit pref erential adsorption on specific substrates. These materials may be good candidates for developing high selectivity slurry for applications over a wide pH range. In addition, applications of high selectivity slurry can be extended to other CMP system such as copper, low k materials (i.e. porous silica, Cudoped silica, etc.) for high speed devices, and high k materials (Hf based metal) for sub 45 nm devices. When surfactants and polymers are used to develop passivations layers, it needs to be ensured that there are no residual molecules left after the cleaning pro cess. Application of a passivating agent concept to high selectivity abra sives, e. g., ceria, is recommended. In this application, since ceria is highly positivel y charged under the pH conditions suggested, surfactant may preferentially co at the abrasive particles. Quantitative investigation of the number of abra sive particles participating in the polishing process is also recommended for reliable prediction of MRR. With regard to modeling efforts, a systematic study using different methods, basis sets, and other corrections, such as CP method, is recommended to refine current modeling approach for colloidal systems. Increase in the size of molecules is recommended to accommodate cluster effects. A method that can deal with flat su rfaces with reactive atoms is recommended. In the proposed method, optimization of structure occu rs only on the surface and not under the surface

PAGE 124

124 in a way to maintain the surface structure. To de velop a predictive tool for selective adsorption of surfactant, molecular mechanics based m odeling is recommended, since it can incorporate effects of individual water molecules and zeta po tential of the surface. Molecular mechanics has advantages for systems involving weak interac tions (physical adsorpti on) and the electronic structure method has advantages for strong intera ctions from chemical adsorption. Therefore, a combination of both modeling methods is recomm ended to develop guidelines for formulating selectively polishing CMP slurries.

PAGE 125

125 LIST OF REFERENCES [1] ITRS (2007) Internati onal Technology Roadmap for Semiconductors 2006 Update Overview and Working Group Summaries. http://www.itrs.net/Links/2006Update/ FinalToPost/00_Exe cSum2006Update.pdf. [2] Moore, G.E. (February, 2003): No Exponential is Forever.but We Can Delay 'Forever'. In International Solid State Ci rcuits Conference (ISSCC), San Francisco, CA, pp. 1-19. [3] Stix, G. (1997) Under the wire. Scientific American 276 : 38. [4] Evans, D. (2002) The future of CMP. MRS Bulletin 27 : 779-783. [5] Murarka, S.P. (2000): Directions in the Ch emical Mechanical Planarization research. In Mat. Res. Soc. Symp., pp. 3-11. [6] Singh, R.K., Bajaj, R. (2002) Advances in chemical-mechan ical planarization. MRS Bulletin 27 : 743-751. [7] Sorooshian, J., Borucki, L., Timon, R., Stein, D., Boning, D., Hetherington, D., Philipossian, A. (2004) Estimating the Effective Pressure on Patterned Wafers during STI CMP. Electrochemical and So lid-State Letters 7 : G204-G206. [8] Boning, D., Lee, B. (2002) Nanotopography i ssues in shallow trench isolation CMP. MRS Bulletin 27 : 761-765. [9] Peters, L. (April 1999) Choices and Ch allenges for Shalllow Trench Isolation. Semiconductor international: http://www.reedelectronics.com/semiconductor/ article/CA164565?pubda te=164564%164562F164561%164562F 161999 [10] Zhao, E., Xu, C.S. (June 2001) Direct CMP for STI. Semiconductor International: http://www.reedelectronics.com/semiconductor/ article/CA151781?pubda te=151786%151782F151781%151782F 152001 [11] Bonner, B.A., Iyer, A., Kumar, D., Oste rheld, T.H., Nickles, A.S., Flynn, D. (March 2001): Development of a Direct Polish Process for Shallow Trench Isolation Modules. In CMPMIC. [12] Kim, S.Y., Lee, K.J., Seo, Y.J. (2003) In-s itu end point detection of the STI-CMP process using a high selectivity slurry. Microelectronic Engineering 66 : 463-471. [13] Martin, A., Spinolo, S., Morin, S., Bacchet ta, M., Figerio, F., Bonner, B.A., Mckeever, P., Tremolada, M., A., I. (2003): The Developmen t of a Direct-Polish Process for STI CMP. In Mat. Res. Soc. Symp, pp. F5.10.11-F15.10.16.

PAGE 126

126 [14] Singh, R.K., Lee, S.M., Choi, K.S., Basim, G.B., Choi, W.S., Ch en, Z., Moudgil, B.M. (2002) Fundamentals of slurry design for CMP of metal and dielectric materials. MRS Bulletin 27 : 752-760. [15] Schlueter, J. (October 1999) Trench Warfare: CMP and Shallow Trench Isolation. Semiconductor International: http://www.reedelectronics.com/semiconductor/ article/CA169382?pubda te=169310%169382F169381%169382F 161999 [16] Jin, R.R., David, J., Abbassi, B., Osterheld, T., Redeker, F. (February 1999): A Production-Proven Shallow Trench Isolation (ST I) Solution Using Novel CMP Concepts. In CMP-MIC, p. 314. [17] Yu, C., Fazan, P.C., Mathews, V.K., Doa n, T.T. (1992) Dishing Effects in a Chemical Mechanical Polishing Planarization Pro cess for Advanced Trench Isolation. Applied Physics Letters 61 : 1344-1346. [18] Kim, K.H., Hah, S.R., Han, J.H., H ong, C.K., Chung, U.I., Kang, G.W. (2000): A Study of the Planarity by STI CMP Erosion Modeling. In Mat. Res. Soc. Symp., pp. 33-42. [19] America, W.G., Srinivasan, R., Babu, S.V. (2000): The Influence of pH and Temperature on Polish Rates and Selectivity of Sili con Dioxide and Nitride Films. In Mat. Res. Soc. Symp., pp. 13-18. [20] Fabtech (2005) High-perf ormance CMP slurry for STI. http://www.fabtech.org/content/view/368/93/. [21] Basim, G.B. (2002) Form ulation of Engineered Partic ulate Systems for Chemical Mechanical Polishing Applications. University of Florida. [22] Vakarelski, I.U., Brown, S.C., Rabinovic h, Y.I., Moudgil, B.M. (2004) Lateral force microscopy investigation of surfactant-medi ated lubrication fr om aqueous solution. Langmuir 20 : 1724-1731. [23] Bergstrom, L., Bostedt, E. (1990) Su rface-Chemistry of Silicon-Nitride Powders Electrokinetic Behavi or and Esca Studies. Colloids and Surfaces 49 : 183-197. [24] Sonnefeld, J. (1996) Determination of surface charge density parameters of silicon nitride. Colloids and Surfaces a-Physicoc hemical and Engine ering Aspects 108 : 27-31. [25] Sonnefeld, J. (1993) An Analytic-Expre ssion for the Particle-S ize Dependence of the Surface-Acidity of Colloidal Silica. Journal of Colloid and Interface Science 155 : 191-199. [26] Hackley, V.A. (1997) Colloidal processing of silicon nitride with poly(acrylic acid).1. Adsorption and electrostatic interactions. Journal of the American Ceramic Society 80 : 23152325.

PAGE 127

127 [27] Liu, D., Malghan, S.G. (1996) Role of pol yacrylate in modifying interfacial properties and stability of silicon nitride particles in aqueous suspensions. Colloids and Surfaces aPhysicochemical and Engineering Aspects 110 : 37-45. [28] Malghan, S.G. (1992) Dispersion of Si3N4 Powders Surface Chemical Interactions in Aqueous-Media. Colloids and Surfaces 62 : 87-99. [29] Malghan, S.G., Premachandran, R.S., Pei, P.T. (1994) Mechanistic Understanding of Silicon-Nitride Dispersion Using Ca tionic and Anionic Polyelectrolytes. Powder Technology 79 : 43-52. [30] Yanez, J.A., Baretzky, B., Wagner, M., Sigmund, W.M. (1998) The adsorption of tri alkoxy silane on silicon nitrid e for colloidal processing. Journal of the European Ceramic Society 18 : 1493-1502. [31] Bergstrom, L. (1992) Surface-Chemistry of Silicon-Nitride Powders Adsorption from Nonaqueous Solutions. Colloids and Surfaces 69 : 53-64. [32] Philipossian, A., Rogers, C., Lu, J. (2001) : Tribology, Fluid Dynamics and Removal Rate characterization of Novel Slurries for ILD Polish Applications. In VMIC Conference, Santa Clara, CA. [33] Hibi, Y., Enomoto, Y. (1995) Lubricati on of Si3N4 and Al2O3 in Water with and without Addition of Silane Coupling Agents in the Range of 0.05-0.10 Mol/L. Tribology International 28 : 97-105. [34] Palla, B.J., Shah, D.O. (2002) Stabilization of high ionic strength sl urries using surfactant mixtures: Molecular factors that determine optimal stability. Journal of Colloid and Interface Science 256 : 143-152. [35] Hu, Y.Z., Gutmann, R.J., Chow, T.P., Witc raft, B. (1998) Chemi cal-mechanical polishing for giant magnetoresistance device integration. Thin Solid Films 332 : 391-396. [36] Hu, Y.Z., Yang, G.R., Chow, T.P., Gutma nn, R.J. (1996) Chemical -mechanical polishing of PECVD silicon nitride. Thin Solid Films 291 : 453-457. [37] Palla, B.J., Shah, D.O. (2000) Stabilizati on of high ionic strength slurries using the synergistic effects of a mixed surfactant system. Journal of Colloid and Interface Science 223 : 102-111. [38] Allen, L.H., Matijevi.E (1969) Stability of Colloidal Silica.I. Effect of Simple Electrolytes. Journal of Colloid and Interface Science 31 : 287-&. [39] Allen, L.H., Matijevi.E (1970) Stabil ity of Colloidal Silica.2. Ion Exchange. Journal of Colloid and Interface Science 33 : 420-&. [40] Allen, L.H., Matijevi.E ( 1971) Stability of Colloidal Sili ca.3. Effect of Hydrolyzable Cations. Journal of Colloid and Interface Science 35 : 66-&.

PAGE 128

128 [41] Depasse, J., Watillon, A. (1970) Stab ility of Amorphous Colloidal Silica. Journal of Colloid and Interface Science 33 : 430-&. [42] Langmuir, I. (1938) The role of attractiv e and repulsive forces in the formation of tactoids, thixotropic gels, prot ein crystals and coacervates. Journal of Chemical Physics 6 : 873896. [43] Israelachvili, J., Wennerstrom, H. (1996) Role of hydration and water structure in biological and colloid al interactions. Nature 379 : 219-225. [44] Vigil, G., Xu, Z.H., Steinberg, S., Israelac hvili, J. (1994) Interacti ons of Silica Surfaces. Journal of Colloid and Interface Science 165 : 367-385. [45] Cook, L.M. (1990) Chemical Processes in Glass Polishing. Journal of Non-Crystalline Solids 120 : 152-171. [46] Ito, S., Tomozawa, M. (1982) Cr ack Blunting of High-Silica Glass. Journal of the American Ceramic Society 65 : 368-371. [47] Nogami, M., Tomozawa, M. (1984) Effect of Stress on Water Diffusion in Silica Glass. Journal of the American Ceramic Society 67 : 151-154. [48] Chi-Wen, L., Bau-Tong, D., Wei-Tsu, T., Ching-Fa, Y. (1996) Modeling of the Wear Mechanism during Chemical-Mechanical Polishing. Journal of The Elect rochemical Society 143 : 716-721. [49] Senden, T.J., Drummond, C.J. (1995) SurfaceChemistry and Tip Sample Interactions in Atomic-Force Microscopy. Colloids and Surfaces a-Physicoche mical and Engineering Aspects 94 : 29-51. [50] Habraken, F.H.P.M., Kuiper, A.E.T. ( 1994) Silicon-Nitride and Oxynitride Films. Materials Science & Engineering R-Reports 12 : 123-175. [51] Callister, W.D.J.: Materials Science a nd Engineering: An Introduction. John Wiley & Sons, New York 1999. [52] Choi, W., Lee, S.M., Singh, R.K. (2004) pH and down load effects on silicon dioxide dielectric CMP. Electrochemical and Solid State Letters 7 : G141-G144. [53] Holland, L.: The Properties of Glass Surfaces. Chapman and Hall, London 1964. [54] Pietsch, G.J., Chabal, Y.J., Higashi, G. S. (1995) Infrared-Absorption Spectroscopy of Si(100) and Si(111) Surfaces after Chemomechanical Polishing. Journal of Applied Physics 78 : 1650-1658. [55] Yeruva, S.B. (2005) Paticle scale modeli ng of material removal and surface roughness in chemical mechanical polishing. University of Florida.

PAGE 129

129 [56] Mahajan, U. (2000) Fundamental studie s on silicon dioxide chemical mechanical polishing. University of Florida. [57] Tabor, D.: Microscopic aspects of adhe sion and lubrication. El sevier, New York 1982. [58] Taran, E., Donose, B.C., Vakarelski, I.U ., Higashitani, K. (2006) pH dependence of friction forces between silica surfaces in solutions. Journal of Colloid and Interface Science 297 : 199-203. [59] Qin, K., Moudgil, B., Park, C.W. (2004) A chemical mechanical polishing model incorporating both the chemical and mechanical effects. Thin Solid Films 446 : 277-286. [60] Wonseop, C., Jeremiah, A., Seung-Mahn, L ., Rajiv, K.S. (2004) Effects of Slurry Particles on Silicon Dioxide CMP. Journal of The Elect rochemical Society 151 : G512-G522. [61] Israelachvili, J.N.: Intermolecular and Surface Forces. Academic, New York 1991. [62] Bouvet, D., Beaud, P., Fazan, P., Sanjin es, R., Jacquinot, E. (2002) Impact ot the colloidal silica particle size on physical vapor deposition tungsten removal rate and surface roughness. Journal of Vacuum Science & Technology B 20 : 1556-1560. [63] Zhou, C.H., Shan, L., Hight, J.R., Danyl uk, S., Ng, S.H., Paszkowski, A.J. (2002) Influence of colloidal abrasive size on material removal rate and surface fi nish in SiO2 chemical mechanical polishing. Tribology Transactions 45 : 232-238. [64] Choi, W., Mahajan, U., Lee, S.M., Abiade, J., Singh, R.K. (2004) Effect of slurry ionic salts at dielectric silica CMP. Journal of the Electrochemical Society 151 : G185-G189. [65] Iler, R.K.: The Chemistry of S ilica. John Wiley & Sons, New York 1979. [66] Hulett, G.: Solubility and Size of Pa rticles. Chemical Catalog Co., New York 1926. [67] Bressers, P.M.M.C., Kelly, J.J., Garden iers, J.G.E., Elwenspoek, M. (1996) Surface morphology of p-type (100)silicon et ched in aqueous alkaline solution. Journal of the Electrochemical Society 143 : 1744-1750. [68] Kovacs, G.T.A., Maluf, N.I., Peterse n, K.E. (1998) Bulk micromachining of silicon. Proceedings of the Ieee 86 : 1536-1551. [69] Palik, E.D., Glembocki, O.J., Heard, I ., Burno, P.S., Tenerz, L. (1991) Etching Roughness for (100) Silicon Surfaces in Aqueous Koh. Journal of Applied Physics 70 : 32913300. [70] Donose, B.C., Vakarelski, I.U., Higashit ani, K. (2005) Silica surfaces lubrication by hydrated cations adsorption fr om electrolyte solutions. Langmuir 21 : 1834-1839. [71] Raviv, U., Klein, J. (2002) Fluidity of bound hydration layers. Science 297 : 1540-1543.

PAGE 130

130 [72] Taran, E. (2006) Molecular-Scale Friction Characteristics of Silica Surfaces in High pH Solution Studied by Atomic Force Microscopy. Kyoto University. [73] Bremmell, K.E., Fornasiero, D., Ralston, J. (2005) Pentlandite-lizar dite interactions and implications for their separation by flotation. Colloids and Surfaces a-Physicochemical and Engineering Aspects 252 : 207-212. [74] Celik, M.S., Hancer, M., Miller, J.D. (2002) Flotation chemistry of boron minerals. Journal of Colloid and Interface Science 256 : 121-131. [75] Basim, G.B., Vakarelski, I.U., Moudgil, B.M. (2003) Role of interaction forces in controlling the stability and polishing performance of CMP slurries. Journal of Colloid and Interface Science 263 : 506-515. [76] Mohamed, M.M. (1996) Adsorption prope rties of ionic surfactants on molybdenummodified silica gels. Colloids and Surfaces a-Physicoche mical and Engine ering Aspects 108 : 3948. [77] Wu, Z., Lee, K., Lin, Y., Lan, X., Huang, L. (2003) Effect of surface-active substances on acid-base indicator re activity in SiO2 gels. J. Non-Cryst. Solids 320 : 168-176. [78] Somasundaran, P., Huang, L. (2000) Adsorp tion/aggregation of su rfactants and their mixtures at solid-liquid interfaces. Advances in Colloid and Interface Science 88 : 179-208. [79] Bruce, C.D., Berkowitz, M.L., Perera, L., Forbes, M.D.E. (2002) Molecular dynamics simulation of sodium dodecyl sulfate micelle in water: Micellar structural characteristics and counterion distribution. Journal of Physical Chemistry B 106 : 3788-3793. [80] Rosen, M.J.: Surfactants and Interfacial Phenomena. John Wiley & Sons, New York 2004. [81] Patist, A., Oh, S.G., Leung, R., Shah, D.O. (2001) Kinetics of micellization: its significance to technological processes. Colloids and Surfaces aPhysicochemical and Engineering Aspects 176 : 3-16. [82] Philipossian, A., Mitchell, E. (2004) Mean Residence Time and Removal Rate Studies in ILD CMP. Journal of the Elect rochemical Society 151 : G402-G407. [83] Zeng, X., Osseo-Asare, K. (2004) Part itioning behavior of silica in the Triton X100/dextran/water aqueous biphasic system. Journal of Colloid and Interface Science 272 : 298307. [84] Clark, S.C., Ducker, W.A. (2003) Excha nge rates of surfactant at the solid-liquid interface obtained by ATR-FTIR. Journal of Physical Chemistry B 107 : 9011-9021. [85] Madejova, J. (2003) FTIR t echniques in clay mineral studies. Vibrational Spectroscopy 31 : 1-10.

PAGE 131

131 [86] Yokoyama, Y., Ishiguro, R., Maeda, H., Mukaiyama, M., Kameyama, K., Hiramatsu, K. (2003) Quantitative analysis of protein adso rption on a planar surface by Fourier transform infrared spectroscopy: lysozyme adsorbed on hydrophobic siliconcontaining polymer. Journal of Colloid and Interface Science 268 : 23-32. [87] Harwell, J.H., Roberts, B.L., Scamehorn, J.F. (1988) Thermodynamics of Adsorption of Surfactant Mixtures on Minerals. Colloids and Surfaces 32 : 1-17. [88] Somasund.P, Fuersten.Dw (1966) Mechan isms of Alkyl Sulfonate Adsorption at Alumina-Water Interface. Journal of Physical Chemistry 70 : 90-&. [89] Wang, W., Kwak, J.C.T. (1999) Adsorption at the alumina-water in terface from mixed surfactant solutions. Colloids and Surfaces a-Physicoc hemical and Engine ering Aspects 156 : 95110. [90] Goloub, T.P., Koopal, L.K. (1997) Adso rption of cationic su rfactants on silica. Comparison of experiment and theory. Langmuir 13 : 673-681. [91] Chandar, P., Somasundaran, P., Turro, N. J. (1987) Fluorescence Probe Studies on the Structure of the Adsorbed Layer of Dodecy l-Sulfate at the Alumina-Water Interface. Journal of Colloid and Interface Science 117 : 31-46. [92] Leung, R., Shah, D.O. (1986) Dynamic prope rties of micellar solu tions: I. Effects of short-chain alcohols and pol ymers on micellar stability. Journal of Colloid and Interface Science 113 : 484. [93] Ozdemir, O., Cinar, M., Sabah, E., Ar slan, F., Celik, M.S. Adsorption of anionic surfactants onto sepiolite. Journal of Hazardous Materials In Press, Corrected Proof [94] Dines, T.J., MacGregor, L.D., Rochester, C.H. (2003) Adsorption of 2-chloropyridine on oxides an infrared spectroscopic study. Spectrochimica Acta Part a-Molecular and Biomolecular Spectroscopy 59 : 3205-3217. [95] Guan, Y., Liu, Y., Wu, W., Sun, K., Li, Y., Ying, P., Feng, Z., Li, C. (2005) Dibenzodioxin Adsorption on Inorganic Materials. Langmuir 21 : 3877-3880. [96] Singh, P.K., Adler, J.J., Ra binovich, Y.I., Moudgil, B.M. (2 001) Investiga tion of selfassembled surfactant structures at th e solid-liquid interf ace using FT-IR/ATR. Langmuir 17 : 468-473. [97] Foresman, J.B., Frisch, A.: Exploring Ch emistry With Electronic Structure Methods. Gaussian Inc., Pittsburgh, PA 1996. [98] Derecskei-Kovacs, A., Derecskei, B., Schelly, Z.A. (1998) Atomic-level molecular modeling of the nonionic surfact ant Triton X-100: The OPE9 com ponent in vacuum and water. Journal of Molecular Graphics and Modelling 16 : 206.

PAGE 132

132 [99] Dkhissi, A., Brocorens, P., Lazzaroni, R. (2006) Molecular mechanics study of the influence of the alkyl substituents on the packing of the conjugated PEDOT chains. Chemical Physics Letters 432 : 167. [100] Dominik, A., Tripathi, S. Chapman, W.G. (2006) Bulk and interfacial properties of polymers from interfacial SAFT density functional theory. Industrial & Engineering Chemistry Research 45 : 6785-6792. [101] Wu, X., Vargas, M.C., Nayak, S., Lotrich, V., Scoles, G. (2001) Towards extending the applicability of density functiona l theory to weakly bound systems. Journal of Chemical Physics 115 : 8748-8757. [102] Zhmud, B.V., Sonnefeld, J., Bergstrom, L. (1999) Influence of chemical pretreatment on the surface properties of silicon nitride powder. Colloids and Surfaces a-Physicochemical and Engineering Aspects 158 : 327-341. [103] Ochterski, J.W. (June 2, 2000) Thermochemistry in Gaussian. http://www.gaussian.com/g_whitepap/thermo/thermo.pdf. [104] Andersson, M.P., Uvdal, P. (2002) The vibrational spectra of methyl groups in methylthiolate and methoxy adsorbed on Cu(100). Langmuir 18 : 3759-3762. [105] Andersson, M.P., Uvdal, P. (2005) Ne w scale factors for harmonic vibrational frequencies using the B3LYP density functi onal method with the triple-xi basis set 6311+G(d,p). Journal of Physical Chemistry A 109 : 2937-2941. [106] Glinnemann, J., King, H.E., Jr., Schulz, H ., Hahn, T., la Placa, S.J., Dacol, F. (1992) Crystal structures of the low-te mperature quartz-type phases of Si O2 and Ge O2 at elevated pressure. Zeitschrift fuer Kristallographie 198 : 177-212. [107] Grun, R. (1979) The crystal structure of [beta]-Si3N4: structural and stability considerations between [a lpha]and [beta]-Si3N4. Acta Crystallographica Section B 35 : 800804. [108] Denoyel, R., Rouquerol, J. (1991) Th ermodynamic (Including Microcalorimetry) Study of the Adsorption of Nonionic and Anionic Su rfactants onto Silica, Kaolin, and Alumina. Journal of Colloid and Interface Science 143 : 555-572. [109] Volkov, A., King, H.F., Coppens, P. (2006) Dependence of the intermolecular electrostatic interaction energy on the level of theory and the basis set. Journal of Chemical Theory and Computation 2 : 81-89.

PAGE 133

133 BIOGRAPHICAL SKETCH Kyoung-Ho Bu was born on April 7, 1970 in Sout h Korea. He graduated from Kyungnam high school in Busan, Korea. He received his B. S. and M.S. degrees in inorganic materials science and engineering from Seoul Nationa l University, Seoul, Korea in 1994 and 1996, respectively. From 1996 to 1998, he was a research engineer in the Electr onic Materials Division of the Institute for Advanced Engineering (IAE), where he performed research in the fabrication and characterization of Pb(ZrxTi1-x)O3 thin film micro-mirror a rrays for projection display applications. From 1999 to 2002, he was assistant manager in the plasma display panel (PDP) R&D center of Orion Electric Co., Korea. His research interests we re cell design and gas discharge physics for high efficient plasma display. In fall 2002, he started his Ph.D. study with Professor Brij Moudgil at the University of Florida where he has worked at the Engineeri ng Research Center for Particle Science and Technology. His dissertation rese arch concentrated on selective chemical mechanical polishing. He graduated from the University of Florida wi th a doctorate degree in materials science and engineering with electronic materials and part icle science and technol ogy specialties in May 2007.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101208_AAAAJJ INGEST_TIME 2010-12-08T19:13:14Z PACKAGE UFE0019608_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 3465 DFID F20101208_AABPFI ORIGIN DEPOSITOR PATH bu_k_Page_113thm.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
c477b94fbffd794565ed6895f550c0b6
SHA-1
0d3613cf4be47474686cc65368679b86141e5395
24912 F20101208_AABPEU bu_k_Page_106.QC.jpg
d996d80a5f549b37d312b1df27af4366
572b6f63bf11aa82cb823b86ecfb6ad5d5ae6eae
3561 F20101208_AABOYZ bu_k_Page_020thm.jpg
f5abf48a5cb51115c9960f16fab976c7
3c4c6695e38183398c0a382bd5a21fac7961ee5a
73217 F20101208_AABOCG bu_k_Page_021.jpg
bc9bb58b27a45aea7e55b3fcb939cbda
ef1354d8c85ce13f166479db15a19ed3731dcd78
24730 F20101208_AABNXA bu_k_Page_004.QC.jpg
e9ff600e6b19650dc135744462212e12
d899d400891b63193deb95dbc6f7a3f906690afa
42217 F20101208_AABOBS bu_k_Page_007.jpg
0935ed3d8902771ae5d3a32b1ed41372
338d3492dc05b9f0977109b3ab88915e0e3dbffc
5352 F20101208_AABOZO bu_k_Page_028thm.jpg
845f2b7203e976d267f5f4576780e590
42c599554bf845722efb39b32368e97fc5c9be8b
2470 F20101208_AABPFJ bu_k_Page_114thm.jpg
afafa641a0bf022299ee2725e4488631
21297562ffa712259554a3fa99fb7e60613192cb
6887 F20101208_AABPEV bu_k_Page_106thm.jpg
eed17c1e2db6ff84d830985b90e06f72
fc1df64dd6c84e9e90e0fafaa62aae717db67f2d
81150 F20101208_AABOCH bu_k_Page_022.jpg
be4b88fbd9f9d7c88fee2810cceca6c1
4c866c836ee0a90f2e89a4260953a7dd341a08e4
1053954 F20101208_AABNXB bu_k_Page_129.tif
f6254b2b47139226f100c3bf8d02eac6
b34ecc7b965313e8dc2f22575be524753e4e3d71
40956 F20101208_AABOBT bu_k_Page_008.jpg
2ef1caf3d95aee35d0e47042c6f9e2a1
56b4d53b29d695d20051e7869e73fecfd4381b63
15573 F20101208_AABOZP bu_k_Page_029.QC.jpg
2977cec3a7f06bcb1216597c0e086b98
d03902dbaf7419d1a09466dc29face069651e53a
8020 F20101208_AABPFK bu_k_Page_115.QC.jpg
91faf90765d29a0dabf9be795d4c3f13
94c6302719ac7826a99d667db4401e9fb134fb1d
26536 F20101208_AABPEW bu_k_Page_107.QC.jpg
7df4c7233cd1c7e7f6b6b408d2486fd3
56094471fe1d2fe51382c19a8c3e2c81fd03bf8b
75816 F20101208_AABOCI bu_k_Page_023.jpg
5bb872c9a44a7a9f8cf50601d37e0650
b21b9226d96398e9131f563c1e0079b3d20c7812
28315 F20101208_AABNXC bu_k_Page_070.jpg
d80d10c2995a7f1d8bc2f413a59a402d
94d710340ff8ccc0aa54e15cc0fb2c0cf8b0e979
95135 F20101208_AABOBU bu_k_Page_009.jpg
8b3d97e0b47b2454ea517973cc011846
47965ab36dcddadd03199ed7dd891b78dd30730c
9775 F20101208_AABOZQ bu_k_Page_030.QC.jpg
978ea8e2aac50dd80925da79349a8826
d9c96872f2d64b427ccfbe0c986e8016b45299e1
2904 F20101208_AABPGA bu_k_Page_124thm.jpg
9be4cced1d11d0ef9c10d65234059991
3484acf38aca1f117cabfa6f098577f577679861
2660 F20101208_AABPFL bu_k_Page_115thm.jpg
f92ef2a769a2e3ebda8cf858a7e64879
7e4eb7ab503c0ded1d740235a7efb432560ffb02
7037 F20101208_AABPEX bu_k_Page_107thm.jpg
f674542c57e738b0ae207c431cbaa7d0
eca91ff1599dfd561e413766745fb398d48a3786
75390 F20101208_AABOCJ bu_k_Page_024.jpg
7c07beeaa747a94a546af17db6c08201
a1b9b322f025e5e4f265298b189703823c9e4623
3293 F20101208_AABNXD bu_k_Page_011.txt
3eb7c416b1673573129b61889c8cc259
114fc15fdc31784958b33400b3eddd42c7cc2491
102120 F20101208_AABOBV bu_k_Page_010.jpg
d26b99f3e11b45127145b0e9d88f764a
aabb696621b37b1f1ff933de7a0fc6071411a0dc
13248 F20101208_AABOZR bu_k_Page_031.QC.jpg
137ad73f0a91a5401ae09b662963b7b1
8aaccad872e6ba36ed1f9c18488030b9d81ebf2b
26501 F20101208_AABPGB bu_k_Page_125.QC.jpg
973638b60cc033c38bd467a5659f0385
4f1c5b59bdd6750147f4c7b10dfe1680966c68ea
5244 F20101208_AABPFM bu_k_Page_116.QC.jpg
2152a8e35e0eef95887ce9efbebf60f8
298febd804c0dd664670456da465616a8b0b1f5b
7161 F20101208_AABPEY bu_k_Page_108thm.jpg
b93993c8e25ab930882f3dc8ed28fe8d
f085979d7a8847a220b01cac3611d54edd746692
80461 F20101208_AABOCK bu_k_Page_025.jpg
ce6b612bbf52373b312d90da5ccbac29
9cde0dc5feee5411cd597a380edbd88957a10425
25271604 F20101208_AABNXE bu_k_Page_011.tif
ae0a6f38f48c4a5ee75488dd9f254d07
210665c9c565f2353f6e42a864d7b54104fa9a0c
105665 F20101208_AABOBW bu_k_Page_011.jpg
0e76f197f7667e269d9ca93bb71d6cfa
0caefa586a506a1a2d1027e7d500489e4c5c872b
4077 F20101208_AABOZS bu_k_Page_031thm.jpg
6a0d83dfadd43b96b0c5c0581f59b133
7e258dfb50a8264652ae8bdf29a20a0f556f1d92
7219 F20101208_AABPGC bu_k_Page_125thm.jpg
4a10ef9eefaec02efd5627d51f70b9df
a5e6d2a796259cd4f2dca2ffd67ff9d5a2cb4dd2
2056 F20101208_AABPFN bu_k_Page_116thm.jpg
65fc24b7dacfcabb887304a94a028a65
b1ccbab87cbd351aed78ce83b8db3afaf6b4d04e
26143 F20101208_AABPEZ bu_k_Page_109.QC.jpg
c5ed8526d597d2cd29fbdc62d812685d
4197d0aee16b6fc8a548ec10137655e4948441b7
83211 F20101208_AABOCL bu_k_Page_026.jpg
679713fe6d6bc1bb96b2894dbfe6c399
1b51c056e52de1744e2edc9a9530eb4fc218cb6f
120235 F20101208_AABNXF bu_k_Page_108.jp2
863c369a1323c1eda4ce0677f912880a
1963751cd3aacbe25d2fc5d19fa632941fa48c50
14541 F20101208_AABOBX bu_k_Page_012.jpg
92272f311e16b9f6e4a33b1dd8e39e34
477af019781df98ae89e26f1fbf6110f59673c36
10119 F20101208_AABOZT bu_k_Page_032.QC.jpg
c939e1b5bcefdf41e8679b836b396fe2
6455710dc285ff44eb7124fd381f2f10b8af9747
64322 F20101208_AABODA bu_k_Page_041.jpg
efd3afdfeb71fe26603749dd49c8b016
0f56a9e599be9dabc969258d5b025c426ba1e184
27240 F20101208_AABPGD bu_k_Page_126.QC.jpg
e3f91e29e244d3acd8d5f30ea25b14a1
bafc923d8db3ec3fa1745f3486424c233f59ad99
6948 F20101208_AABPFO bu_k_Page_117.QC.jpg
e90f0ca7f0a6d22966e29ed8b638421f
a06a4be4c7c13292fdf1a5dcb66c4d2f60fd9334
74728 F20101208_AABOCM bu_k_Page_027.jpg
80be30d30b90e581109bf6e3fe30c480
8fd78d72fd4fb2e087bc9e3c39c1ff9df182a856
F20101208_AABNXG bu_k_Page_039.tif
e13d6703116fd4ce96fc7ff4258be819
019092519654c021637b5b3525799c206a701c39
67915 F20101208_AABOBY bu_k_Page_013.jpg
e62d4481cdf69796504e6f54035fd8f7
35fcd48062126f05b678a20dc9352ded98094025
4141 F20101208_AABNWS bu_k_Page_079thm.jpg
a3dba18e4ec8b8d6b72cb89ecf51d08e
a44f41aa0b371d4f3794b06f94e756011c59c5d4
3474 F20101208_AABOZU bu_k_Page_032thm.jpg
96246dcca98d3c4f86fecbd5b2c3c6af
c20e95a56c964fab85da05154bf64260bcf1f265
80416 F20101208_AABODB bu_k_Page_042.jpg
ac746880d7bdcc373159dff5e9c78609
89aa2c7f34fd84b885823434e38c2d8f3d981e02
7455 F20101208_AABPGE bu_k_Page_126thm.jpg
d54a6b564f3e692203f6b87021ca6389
370686e68d082c98c80174e5d7282b89ac42f665
2345 F20101208_AABPFP bu_k_Page_117thm.jpg
c7f46db1a746dd8a2abfb02ec586cfb5
49e79c88f822571a61fbf69e7085531f4e3971c3
60524 F20101208_AABOCN bu_k_Page_028.jpg
aaf909b40aaad022f62d032c3311bbaf
43c0f20e6c5c925e74b3faa9a1ff53c06bdadacb
2355 F20101208_AABNXH bu_k_Page_118thm.jpg
0eace407f4d68f21683dd21f3d63798f
68be5398a0c2da0e4cab6443a60fbaf2c9d3ae50
21233 F20101208_AABOBZ bu_k_Page_014.jpg
e985eef234da7718f30abbbd3b490d97
36f1533a527640a31dc4f168a22203c5eb919069
F20101208_AABNWT bu_k_Page_024.tif
bba33f7831d4164d5e695ec1db4f6520
09f180da0efd93ab5e671f0bf475d3860e0099f6
12317 F20101208_AABOZV bu_k_Page_033.QC.jpg
b9d1f4912eb95591f22389e7f7901c66
282681a3defffd2010d333669eec833db44816e8
72405 F20101208_AABODC bu_k_Page_043.jpg
bd42fb553b8cd64c7b79e89c31872ab6
e7fedda4dc5e7f1be185a8fe1306f42732a439df
F20101208_AABPFQ bu_k_Page_118.QC.jpg
c6eda25fefe31a529043da86e61721ad
cd1acb825b3fe68593112bb71aeb60bf32419307
50516 F20101208_AABOCO bu_k_Page_029.jpg
01d3b2ef24fac47e0721c71dfbabf933
386871fb64d5ec3be4820e695a50952e1d3fdbe6
F20101208_AABNXI bu_k_Page_124.tif
ee7e56e6cc75763b2ff3b51197d2637f
761de2cfc9d31c264dadecea908013504b81e2dd
F20101208_AABNWU bu_k_Page_087.tif
f84d066c1636ad93bf1700288028fbd2
82cfe7161c82bbefb43969f25cfea7a81e9bbee9
3869 F20101208_AABOZW bu_k_Page_033thm.jpg
3caa06994e2f828584a6f54d2f714865
a9a86278acfffc965cfd0695a5ddbefe5020ebf2
26627 F20101208_AABPGF bu_k_Page_127.QC.jpg
e0acb55abf20d61740bfaf156d98dc2b
29160230be098deced23175070685436476c1856
2536 F20101208_AABPFR bu_k_Page_119thm.jpg
69b7698bd81890eba92793ce6079e191
276c7c409a5ffda89b95e417167c0537e4c1e1fe
32174 F20101208_AABOCP bu_k_Page_030.jpg
ae5c6138432ab2c9a1600dda0b194ee4
7097fa17e027e4b44290c69caed2ac401237afe6
21079 F20101208_AABNXJ bu_k_Page_013.QC.jpg
a2464c6b51e64c7d63572bf007e31ac6
0559750e0131415862a37162e2fcdd37fe250511
F20101208_AABNWV bu_k_Page_016.tif
2fe754de0fc77f71c8d8d80754dcfd3d
2fcc77f478cd518a14e9dded4917571c3cb83765
8276 F20101208_AABOZX bu_k_Page_034.QC.jpg
537617218e8275abfee351f05e062c5c
1924d46dbe1c98b6ea551ea418c590ee6ae38193
70516 F20101208_AABODD bu_k_Page_044.jpg
111ce97d0d1824ca9502c4888bb3f077
64e6a9a0d23e3dc4e45567d8f140f4e73ebd72ed
7553 F20101208_AABPGG bu_k_Page_127thm.jpg
e9bfecdbad169e2e19051c61cb73ed46
87188b8d8cf0dc73744c09c35bf6e809de15bebf
25523 F20101208_AABPFS bu_k_Page_120.QC.jpg
d9830d9872ed7d2ce3761bcbce86d584
c906405d613b34e658e21c4ebbff8807a7802924
43356 F20101208_AABOCQ bu_k_Page_031.jpg
5edaa5aa081c8bb4a245caedf797c4be
370fbf1d62e2f91fa97e9a7f39a88eb17f3c358d
92589 F20101208_AABNXK bu_k_Page_131.jpg
499961fe4a2873f51d5805d541079583
f5e092d7b3cde8ba5346a7f04827cd34a28b2a72
3118 F20101208_AABNWW bu_k_Page_118.pro
d49ec84c494bbbbe84ec913e1f891aff
a393e97fd0ad8cafb31a866af42afc2151af5a84
8485 F20101208_AABOZY bu_k_Page_035.QC.jpg
a530788aac1fb3a579f79690833f0421
fb88afa49561fbd79161bf1e8578b105130784d9
78057 F20101208_AABODE bu_k_Page_045.jpg
1f28aa0bf53e9b0125ce3314e7bc9457
e5d50ff7924bc23edc35f49016dbf81c62f5461f
25371 F20101208_AABPGH bu_k_Page_128.QC.jpg
db05e210fdcb7a8bf95d7398ed8a9e5c
32b7c5740ba5d03afc68c88e6dcda891e1cbb4b5
6876 F20101208_AABPFT bu_k_Page_120thm.jpg
9bb28bcec17c11094bbb093687966cc1
4131a93a65f08f50bb243800b54d4809195ff2aa
29366 F20101208_AABOCR bu_k_Page_032.jpg
f8ef53c84320a8eed504482324bb443f
7e11a58b678330315a37106aa68fe671f82e5510
F20101208_AABNXL bu_k_Page_061.tif
be0d92a3558a4f8b05cbabc0b55b8afc
c55e1828f309e5535a7e075f76745efd427162c6
79853 F20101208_AABODF bu_k_Page_046.jpg
1b3f403b3ba3fb05cfaf29c54c58fa26
45d006840de8d0207c3b3f013146f8b38f9ea32a
7126 F20101208_AABPGI bu_k_Page_128thm.jpg
3316c149b93f602839fbdf5c78504dbc
9a59455131a689e78cb03b8f7d27e01a4ecb15dc
23173 F20101208_AABPFU bu_k_Page_121.QC.jpg
d1840cfdbe79309dda9e585c8655b217
4f9bf7f71558c8cfcab4c7cee5c32e79115156c6
F20101208_AABNYA bu_k_Page_118.tif
e465602a5b7537b2ccc42ce3bf8eaeef
c103f481a7b1a5d03f9163f27fd2cb91fcd9913c
37644 F20101208_AABOCS bu_k_Page_033.jpg
999ff5c4080d9d06073c6b0600f256c4
d2155f693470d8475669ac342b5730ff8336aa3d
6983 F20101208_AABNXM bu_k_Page_122thm.jpg
e2a9cd542b2a2daef73e38ef815cee62
a50edb671dfd44fe83e6d5dd42d9011dec537000
6626 F20101208_AABNWX bu_k_Page_047thm.jpg
72e9e491f4e9b649aac7d0ec84dca6b9
2672735e59be6e52c652f7c2231611ea5d5a275c
3041 F20101208_AABOZZ bu_k_Page_035thm.jpg
87f9c02b97dcb5805a115bc5641d7a5a
02748b1aee6600d075a2710c35ca4197be715b6d
76609 F20101208_AABODG bu_k_Page_047.jpg
7a690b7d9c579e93d3845d7dbb28d30c
d4e93524c9b556f00846fed7b44bc20e4ecdd8b4
7024 F20101208_AABPGJ bu_k_Page_129thm.jpg
a211bc47d09dcda44edd8a64c639b7f4
aa2c1a238d7d7b8f03d2b022e78ff37ff4e4ddf7
6695 F20101208_AABPFV bu_k_Page_121thm.jpg
fafb0e95dbe270ce8524d0d35b09e121
5e0830311f57b91ab6fc8ec8c5cffb4d90af223c
9724 F20101208_AABNYB bu_k_Page_002.jpg
aa030becff31bb966b8de3d41e28ada4
2eb24432c43ed3b65b8645d8c75e4b483bf430de
23930 F20101208_AABOCT bu_k_Page_034.jpg
50385510de184ddb792ed74a49edef4a
e8bc81ff4ab9944dc5be32a72547444d3084445a
91616 F20101208_AABNXN bu_k_Page_078.jp2
fda1137c9e4fa0933ed1e7075aaf02aa
a6c0adaf5e1723cb2cca5faff93519aa5718ca1a
F20101208_AABNWY bu_k_Page_105.tif
3a7f8a137a3d91b75a27c7488fc318a6
8ae20b228c6ba1b3e6eeacf34af2939260072366
79392 F20101208_AABODH bu_k_Page_048.jpg
7e42afbfca966d6c1f742e0cb5b4e5c6
c25b89629621613289632ba3e546dfcfb658ff5c
27606 F20101208_AABPGK bu_k_Page_130.QC.jpg
f8e9d161c617b3a5ef725ca0a4aca776
c4d3aa23656acf6f0f1318d601684164bee93f5e
26053 F20101208_AABPFW bu_k_Page_122.QC.jpg
0a20c25cf2dfd2174c4f6bbd83c1aca7
d1e54acbc0e88bbcc13a8cfb3ba71d742f343815
119566 F20101208_AABNYC bu_k_Page_089.jp2
edf8275d1ce4cc9fbd25176f4ef3b185
a952da9305c362e545fd4010294c7e41556749af
26044 F20101208_AABOCU bu_k_Page_035.jpg
6373dc42c03203b97ba0e7420f55c684
8912b5a12f7120ab4014908797e8039eaed82d0d
3320 F20101208_AABNXO bu_k_Page_010.txt
d378777bb1c1de0f010b185e63b9f0b5
0bc4b966596d9637c11b8b3bf4154fcf9ea736dc
F20101208_AABNWZ bu_k_Page_022.tif
1f8215fa9ae7bf1b4eaf3de99dab28f5
4801ae98e60d475f8dde8771ba7dce1c6980225c
76108 F20101208_AABODI bu_k_Page_049.jpg
522c8ad76bfc525f6c46a3c0d7cf2f3e
ca2d943bc589f6c1acf7881f65d0d8a0b94b3551
7136 F20101208_AABPGL bu_k_Page_130thm.jpg
daba69d0cb3f28f9150f668409619ef4
ee85b4fba81477c35ca02d2c8a9a211503a079ec
25679 F20101208_AABPFX bu_k_Page_123.QC.jpg
750d8e12e94cfde8cbd42303bbdcd4da
a9be391d6279d882c758d095430e151666fc31ce
2186 F20101208_AABNYD bu_k_Page_123.txt
1fbe56180d09d5c6c019f848b79933e6
f0b2ef96050c09e1336e4d95d700b35379aa4165
46834 F20101208_AABOCV bu_k_Page_036.jpg
91347a9531d7bb1d500752e3418da6d8
d663d94ad6844f9466ba2ad65925f5810e9223d3
F20101208_AABNXP bu_k_Page_028.tif
e51dff054d5ad26c5981c106b73d6b9c
63b4d323ca68bbd7e611b4ca36006d90c9c4c658
69361 F20101208_AABODJ bu_k_Page_050.jpg
b8585750307511118b5b40dc5dc7ab41
978d4212e6c2bb0f575a5757197dedb0e05854ff
26349 F20101208_AABPGM bu_k_Page_131.QC.jpg
253ecca47089685ef4d15042fd8df2a3
f9350522e18c30c01a3e8dc0f17efc3517868c64
6928 F20101208_AABPFY bu_k_Page_123thm.jpg
93efb4957d4d463564a23af9e89369d0
94ba1cf12ecea71809434520354fca47b69872c0
528 F20101208_AABNYE bu_k_Page_080.txt
2bcf16db96425a7c90cc55fee2c20984
5b186549704ebc7099ccc2fc54ff5d856b988f54
73673 F20101208_AABOCW bu_k_Page_037.jpg
382ef78adbf966fa7dc6996479e8eb4a
641e9bc1a2f3441b9657605af5e36fec3f521428
F20101208_AABNXQ bu_k_Page_036.tif
b8164522cdee04705adfc566d15a01ac
bc7a4fdc4e362fed07100de3b0dfc2a69524aa7d
74686 F20101208_AABODK bu_k_Page_051.jpg
c7a94e2344b509499478bae2000a77e0
aa6d66d3b1e99173da79856dabe2a18f766c4221
7348 F20101208_AABPGN bu_k_Page_131thm.jpg
fa16867181778ebd2d13788d91d5ad9e
ad35de230e79d64e54f08ed79b3df03793ad9139
9609 F20101208_AABPFZ bu_k_Page_124.QC.jpg
f91fb22c2e829146c56a8e3e8394b49c
dfe424ac2cc5737c246c5df6819e33ead4fdca92
1051960 F20101208_AABNYF bu_k_Page_125.jp2
c283b94629e41141710781dbaa162c04
c82a7de9840e2e7e0865a2225cf351ce31ba1234
80125 F20101208_AABOCX bu_k_Page_038.jpg
2b83eba8c7e312d9f74b6039b7c3319f
6c7876c2954dca85e6730a8a9828d9c4a3177441
126465 F20101208_AABNXR bu_k_Page_026.jp2
5ca4c2bf08d0e60a9bfe8da7d4211665
25dc24cc561380eae663fefad132c89247907b91
34670 F20101208_AABOEA bu_k_Page_068.jpg
b1d9dab3624d6e7ad27d0fca28c1efcb
8284538d9145a64b217aae7c2a8ba0695151bf80
77279 F20101208_AABODL bu_k_Page_052.jpg
b179d0cdba5269f568f72822cae6d5af
8f3931289fbe3b54bc22280a2b56916e543a7322
25789 F20101208_AABPGO bu_k_Page_132.QC.jpg
e13be2049347775df599ad6c1e58aac0
2eaf1065fb71c858b718febb147734b4005381fa
20653 F20101208_AABNYG bu_k_Page_080.jp2
e4b7131cbd026836b2b99a0c8fcc36a3
82b2538fc9b18a394b3408f46d39bc783f0dabe2
71201 F20101208_AABOCY bu_k_Page_039.jpg
c5e3dd920cbffd0119b857718daa220f
cb2af7556295c1f2caf6cf431f5bbf17b5cd82a1
22951 F20101208_AABNXS bu_k_Page_054.jpg
2716aed83d15bce7bf00b82055f2c2b4
a78e59eddbc210a0fd566e76afb0f873ce502ac9
44680 F20101208_AABOEB bu_k_Page_069.jpg
6a3a9e7e7c5c7819b137d67f53e75eec
945c1edf3f9249a8f200416807dbaed55e17fe4d
64415 F20101208_AABODM bu_k_Page_053.jpg
f55de230ae06f2a7e07a5edb82943652
ddfa8f2844fac57aa1971dce1a8cff8135afbd58
6816 F20101208_AABPGP bu_k_Page_132thm.jpg
b3cba5fc225bbce2db4796085234ab1d
008a30aa5a93fa19791c3e0a653b9009fd98b9e8
673209 F20101208_AABNYH bu_k_Page_030.jp2
7454b3463b4606cf473eb2fc108dd746
5a01120cdbc9caaca0a9f472c6beecdc65714c2a
64099 F20101208_AABOCZ bu_k_Page_040.jpg
85d337e0c7aea53f04cee3584eeaea85
5e15517a019d8ef99f5a298b8675fa129fdb6a54
F20101208_AABNXT bu_k_Page_020.tif
dd820727b8184b2d22ab485a26d70dc6
a677539761fa56f947c3d92eaf5ee8dc7cea9959
31538 F20101208_AABOEC bu_k_Page_071.jpg
d5c489cf53070b10414aed19fabe2cc8
60e04acdee564176d86a3ce729786f2a4d4a9524
31418 F20101208_AABODN bu_k_Page_055.jpg
01afee3b1b720370ad3c4b6b6774db22
9bfe1375f69323174fda9d3b5d96b23898cc3476
4974 F20101208_AABPGQ bu_k_Page_133thm.jpg
071d0ee66e53f9143e7afd2f2602471a
4be14baab2f2f713c47bf406b189a53c7308e4b5
6245 F20101208_AABNYI bu_k_Page_073thm.jpg
40d54d313ad8bc415aadb12e42ca31d5
98a2cc66f1c26d37d14adecdc451507e6c02c757
1064 F20101208_AABNXU bu_k_Page_056.txt
5e1b8e35fe1f413b1e9dffe6d03d65d3
d3f2e80f8bfb234cac6149cb904b6c5ece9366b9
68712 F20101208_AABOED bu_k_Page_073.jpg
abe1d70614f4d857885f81e920d7b559
88959ccbf137812c029d637efa6233a769884f1c
38814 F20101208_AABODO bu_k_Page_056.jpg
ecadadff554dceeb5284af4c945a902b
c73d1cd03605819cd8b2f968cc4413172dd8f3d8
151164 F20101208_AABPGR UFE0019608_00001.mets FULL
06da9a59129b8181b7fc0ba42034c9db
6e11bb24d27bb64ce92a58701e468ac753da7ec1
1051986 F20101208_AABNYJ bu_k_Page_007.jp2
1fa75a96d80e4f1017fcaafb501cf351
db798910dd0bf6b3db624a9e9ec9970660a673c2
393 F20101208_AABNXV bu_k_Page_035.txt
75c6fb374103dcadcdafb1756af28338
e72251a1705ba6874a3bf33f85b3a2eba964875a
25156 F20101208_AABODP bu_k_Page_057.jpg
51c00fa6ff0659cdadefab2a51d94638
b716b8def6c9597be3c9e79135b20221cbcc06c7
78633 F20101208_AABNYK bu_k_Page_086.jpg
bba5a7a9c78f7db56d8b7fc49a5b732f
a4a32992ed268f09da94fc4104c8c5fc1aaa096a
2137 F20101208_AABNXW bu_k_Page_027.txt
031b5f01e1caec5ac65b8eaa5d2090b8
7e6e7084fffe12e88fea4cf55f443f873b27b007
71799 F20101208_AABOEE bu_k_Page_074.jpg
093767f76b7f196bbe509f68107063a4
dfbf784874ad0323a60fb085557eea93f5342181
30461 F20101208_AABODQ bu_k_Page_058.jpg
bc8340201dded24638d0cdd796b5e661
0667a58a0a72df01b0cfb9f9af495435bc470031
195 F20101208_AABNYL bu_k_Page_114.txt
dee9d33a7b03f0b2f10125a2c2e727c9
19a0d5ee198d01f68c12817a878f500c49ff5399
10630 F20101208_AABNXX bu_k_Page_055.pro
70e0bff229598df0f50aabaf1de874de
5fc56a04027664685f5cc823d6cff556da939d00
78100 F20101208_AABOEF bu_k_Page_075.jpg
9ad25a1180e58347ab0b28e1acec7faa
fc64579c5b312fa5a5d0f1671f09beb62ccb9b79
39570 F20101208_AABODR bu_k_Page_059.jpg
9923c0bede52ff3d54bd7cfa1ceea017
6dcb6b4f33e8742050c6664a544102c88512bd99
240 F20101208_AABNYM bu_k_Page_116.txt
5c0266b1a53554b57ab0b3f48cafc47f
8f36124544f3d9e8f4bd1f36c0c272c5a2282f60
72687 F20101208_AABOEG bu_k_Page_076.jpg
3fd2b6305aa9ba6091dc57f93883acea
63d603f4fe3e5d3bf4735b3b067397c5778d892c
3124 F20101208_AABNZA bu_k_Page_084thm.jpg
ba0febf629b24f9d34a4ea85161ccb90
ac3a173f09d6821943b5b4c3abbf6ff49cd01428
31152 F20101208_AABODS bu_k_Page_060.jpg
8047e6a6064caf542c74bdf105a7e10b
d56491895c337933e163da79fa645c5a8050fefe
56275 F20101208_AABNYN bu_k_Page_038.pro
ca5aa78eb5cf5ce98eedaf61171f9407
60a4a2affeb28dc01b39c21929767653e3a7933e
1844770 F20101208_AABNXY bu_k.pdf
a7a7d6ea48f2fd34c441e3327bde7330
e9405c86298084b389e89e8f09a27f07f5492357
77961 F20101208_AABOEH bu_k_Page_077.jpg
b992d29dc8e7c12a21e629cac02df97b
988809b158ce7cd6faa4231774e980ea8a0e8433
25185 F20101208_AABNZB bu_k_Page_047.QC.jpg
00108c5d9549ba5db9bfd683eaca7f42
9d309a6fe68ab2e5aed152366cd1e4c065763fb5
24265 F20101208_AABODT bu_k_Page_061.jpg
dc62236185a57112940ebded9ba3477a
b0a39364cd0493bceff08001b2a70dd74df92de8
2245 F20101208_AABNXZ bu_k_Page_038.txt
0b36a99d4d3b1f209d8b1a8cd6fcb2e3
2c1b5b7a02d0d9a92674c0142140b86a8592db9c
61394 F20101208_AABOEI bu_k_Page_078.jpg
d22b63e5b9984b2b6183ded5d150541e
00918ae316feedda0588c6c2293dcaf7b3b78cb4
257 F20101208_AABNZC bu_k_Page_070.txt
ab497eb2e87c215d7c5bd520c5c137d4
83028c1d3fb2068f7a6f245fc869a647d4deb7af
43108 F20101208_AABODU bu_k_Page_062.jpg
d72358f6ae7cbf5bed029abb35477617
6f3c773a4b5f88804a5361e883e5aa78aa9949d1
6996 F20101208_AABNYO bu_k_Page_052thm.jpg
3dcfbb8946bac7090919445e2a5bdc03
fb33c5728ed415ac47e9f69f3f3b2094dd8b7e2f
42432 F20101208_AABOEJ bu_k_Page_079.jpg
024209ed17d99bcbde990dc8629bdeea
e39f0ac256c4fbc9402507897a8ed31e6b5d409e
57986 F20101208_AABNZD bu_k_Page_102.pro
137699f2204629e630217dea5cf8e248
5284d5da0c4fabd43f955c05015e193e6bc0db63
25750 F20101208_AABODV bu_k_Page_063.jpg
34673d5e28d082440cf635c4c1e6656a
e286933e96548b2abf2034837b91b920293a06ee
82101 F20101208_AABNYP bu_k_Page_011.pro
c53638460e525d12d8c77e7ac7331659
13ff7a28f89ad2101ef5448fd0cf148b10135c97
24072 F20101208_AABOEK bu_k_Page_080.jpg
9f87486164f48e823acffeea2b00236c
a054417240299abf76f739367a09c683d8cee015
3269 F20101208_AABNZE bu_k_Page_030thm.jpg
ed8d7a91e1d81f5cd5542062dbf11d36
292c961be215c7deec30b7d6292be49db5aac7e9
42741 F20101208_AABODW bu_k_Page_064.jpg
b6ff99d337c27498ddea78c2bdd8a49b
393fea63412ec51dbce0b39ea3f1566097ad8d6c
7253 F20101208_AABNYQ bu_k_Page_119.QC.jpg
e40689270b779421b789c9fc3141b465
042e2665f7261ec4df381a94f8da0e77a815e930
37460 F20101208_AABOFA bu_k_Page_097.jpg
323cc531f67dc217e621667f1b6891a9
6f4648abe0e3f50970686d257c8a4a626a43aa39
36316 F20101208_AABOEL bu_k_Page_081.jpg
b31c04328142f4f0c54dac21380717fe
f43ba071b539784b6f02b9218e69adf7c18d67b6
463709 F20101208_AABNZF bu_k_Page_081.jp2
a88d22923153ef161ac37bbd293a4d35
04f495af8626f3d4e0693f4902574588292614a9
36267 F20101208_AABODX bu_k_Page_065.jpg
f9255efbf204525613382030afc8c6a7
93408b50a880f3747b94e5ebd4d245cf6865c74e
94 F20101208_AABNYR bu_k_Page_002.txt
3c5e2e14b8b19905c3d9e5bc99fff037
e5505aac284548af55c467acc0023d89a24675f8
38230 F20101208_AABOFB bu_k_Page_098.jpg
17c0375c78fe10f7d692ce6d1e8defcd
fa2f78e12bcdd666dfced2f3d5b6cb9b4b18236a
36378 F20101208_AABOEM bu_k_Page_082.jpg
beaec852338e5ca2a2bc03fe63f689ee
1c67b66d49ca1ad9b6c2af2385fa976849fda3d8
51968 F20101208_AABNZG bu_k_Page_105.pro
98d58872fbf72f029f59a627c5a93fcf
f7db968e065f2c5631b01b15a38b31b419077b77
44409 F20101208_AABODY bu_k_Page_066.jpg
bb50e459e56075bd61296e32ec2fde34
8f6ffe0e4555222681d1e89bcac0105b9a0f078a
1136 F20101208_AABNYS bu_k_Page_008.txt
92ff734d1016742a3a2d6d862e440104
c252bd2331f63beae6c9d10115187db22e6bfd36
28865 F20101208_AABOFC bu_k_Page_099.jpg
5f497c96523d9a8c77398ebc14c2cb6e
a0f679d4d04b5c99f1d9ec0b2bb84f0912b5faf4
20702 F20101208_AABOEN bu_k_Page_083.jpg
cfc300bf37054de53bd1a1b34343329b
211b925095dfc59b257047ad85f9203ffee755c6
25703 F20101208_AABNZH bu_k_Page_034.jp2
2c059d14ba19eb87b09aae2b2f47dc41
7f422668b7695272ccaa997fbadcde4ce40b6894
23449 F20101208_AABODZ bu_k_Page_067.jpg
586030cc2af784bf548f518aaf9400e9
2225b13129ea4791e61345ef05a394e0ac4e70d4
1051980 F20101208_AABNYT bu_k_Page_011.jp2
09f8a53e7d4a1345ddcffb3876719bce
138d77e90c2eca03ac62a969286bc71028ac0b7e
72137 F20101208_AABOFD bu_k_Page_100.jpg
ba9fe3edb706bd96edce218e1b3a3a3f
0366c1be196990cdf58a5886e4f22eae808dc000
30298 F20101208_AABOEO bu_k_Page_084.jpg
7fed14b7b49e92a10652ae85fd0ee231
132a5d983a1530f6382fdbe26c71b9e39211d54d
F20101208_AABNZI bu_k_Page_014.tif
e88398623fbc8df1fbbcdc5e7187f236
5f2f9af1ac757a2654d60454ff7e3098e4a25a12
588 F20101208_AABNYU bu_k_Page_034.txt
3dc52ac2e3486e66a3231c1b46b3c5f2
f1098f312199f35f79540b6584b22b3f49483bfd
77647 F20101208_AABOFE bu_k_Page_102.jpg
aa8864b2c7a8aff6a2b91a9f79536718
1b1539f4303d09abe5ddfd2e98774e58b91cfa19
31609 F20101208_AABOEP bu_k_Page_085.jpg
c8fdcb7c2b6f73537cef1ae337b1270c
fb224c2690628f9c53cf1792c5d2de0901275c1f
F20101208_AABNZJ bu_k_Page_055.tif
1da3d0d22dca9cd8bf40bf2e6dd263b4
87a5b4cd1ba8934a9234f9bb5ad11354fbb79fd3
F20101208_AABNYV bu_k_Page_005.tif
59c9467d2dd3d0d40760b9b832965e92
141e2514437a3256e3fd342556a4883e75b7b9e8
77414 F20101208_AABOEQ bu_k_Page_087.jpg
a1ef2867b1093388e586de646ab8cc93
f60a1d19bc9b0b75eea8397712f296b61b55258a
488 F20101208_AABNZK bu_k_Page_083.txt
d41e3b0d97548c3cc7d7e35b47ad10e2
a1077fbd688cfbbdad8bdcb459e1cb52b39c6f0f
17307 F20101208_AABNYW bu_k_Page_133.QC.jpg
92b468750781cb876f498d0631d4aec4
5e6c5a83e7b6a20adfbfb75136402249c6ae03aa
75551 F20101208_AABOFF bu_k_Page_103.jpg
cb357bbdb6074c9ada50a28a6a0a2261
0c17313bd9e71c3c4323532849bdaeb5ce957b64
83066 F20101208_AABOER bu_k_Page_088.jpg
138a4e77c5cb40153768756994313266
bca2c596d0103e0554182e32c20c490edcb9e3c5
10677 F20101208_AABNZL bu_k_Page_094.pro
0c851207e79b65c51e8d545fa26b54e8
c29d6986d3d7b69ac1564c6acb8c7469f5ccc4ef
51767 F20101208_AABNYX bu_k_Page_088.pro
69d6a1b7c2cd5fedbbe1a532e6a314da
136b936c273822927b3bd1fe3c93c9371aea7f19
71929 F20101208_AABOFG bu_k_Page_104.jpg
5ce6c39189fa655fb20ee57ab5fe6301
3a0c6cc28f9509cdb355eb6008d12323ee15ae27
77929 F20101208_AABOES bu_k_Page_089.jpg
6f208e3b62bc94f794e40ab4ddef39ac
59ae3109c9aac050ac3c4efc6766693dfbdbeeab
34619 F20101208_AABNZM bu_k_Page_113.jpg
1356add9ce4084b18238fd5186eb88a7
2e2e2fead2188c2eea12f35b2a2f9dfb8ba24ee7
49876 F20101208_AABNYY bu_k_Page_044.pro
95b46d2c858139c2105abba1eaf7c6e1
0c9f951f0089dbdc623c795c6ebc52763248b8bf
75024 F20101208_AABOFH bu_k_Page_105.jpg
9e33d107c6f4c3ebdccc965ebb034c63
2e65901db0bcd26d72ef1ccdfb3176918ef428a5
72334 F20101208_AABOET bu_k_Page_090.jpg
b44504a861edb68dc6ff587e271be87a
6150edca152f802a8bf55a1bb9d6423d911c57e0
3532 F20101208_AABNZN bu_k_Page_058thm.jpg
c542dc5a6ba832eba4636cc37b6606eb
4dc1158a02ae41201e9950136d75d9f349db9ab6
75472 F20101208_AABOFI bu_k_Page_106.jpg
f0594a88b90573fba278cacb764df758
6d63e70af63e308badf481a99e3aae56311da85b
76253 F20101208_AABOEU bu_k_Page_091.jpg
7e0728edce122b2df479ede53c2007a5
e450f8b4ca0136972bab5f9b41f2bcb0d3101a8f
25232 F20101208_AABNZO bu_k_Page_049.QC.jpg
176a9ef30cc9be97dee60a0571800d03
a515671ec45df1d9507c9d8b2322177f944030d4
40527 F20101208_AABNYZ bu_k_Page_028.pro
57fb9216c05b2b1f8c3a68d5134fd413
88e6444376a4570968a4d3535b154860f0f76f27
80245 F20101208_AABOFJ bu_k_Page_107.jpg
59d5f779c3cb0f4108f4df7268bd4c86
841b19930225da2edccb6207fdf36041c93fb35c
35244 F20101208_AABOEV bu_k_Page_092.jpg
fb8ea03281a17a486df9d16eecb393e9
055ca6a122ed8103b67b16e6bb0d63ee48cb50d7
458732 F20101208_AABNZP bu_k_Page_113.jp2
86c479274425b7e75b57806963188fde
74005c35e137564ff891c76cb6fa3d2256be4677
80499 F20101208_AABOFK bu_k_Page_108.jpg
d931f103da49d6fc261134c0dd23e566
139c77541078106817f631af129ae0b61400d6f4
32250 F20101208_AABOEW bu_k_Page_093.jpg
fbcc1be7452bd9a66b2077c2bac0b8f9
55c7fbc0725cfb736046f4779177d6262962f2c7
7061 F20101208_AABNZQ bu_k_Page_045thm.jpg
2ea9ade84559df0448d01c51c9ff0f61
a24ec0614197868bbc51e51f77cb4046eabf29b9
38286 F20101208_AABOEX bu_k_Page_094.jpg
5568f7a23f000b38edebca7d2ad96bbb
e083ff065a8ef717b30fef4d7a1bc6ea981cdb10
1145 F20101208_AABNZR bu_k_Page_029.txt
9dd56f6c0632b031fc7745d716f00068
a032615cec3f77d7fe1c12296563caa835266724
92885 F20101208_AABOGA bu_k_Page_126.jpg
0ba27998ac9576157cdf1d314b82c7cd
27e4c01c5b0e19fd2e91dac20827215af82c41a2
79417 F20101208_AABOFL bu_k_Page_109.jpg
58ef1cff6d27467be8fadfc8ebadf56d
7108ab00dcf4034c7bc3ac7564acc82636adde2b
28134 F20101208_AABOEY bu_k_Page_095.jpg
60519c2ef8f969ca69dbd6d2a0fef164
3942a3345bf76f1c789788e09a081bbb48e11274
2283 F20101208_AABNZS bu_k_Page_001thm.jpg
784c1503198bf346cacc8a62bd4af077
c52d1ddde852763d9dd086816c436ef9040cbda0
92215 F20101208_AABOGB bu_k_Page_127.jpg
98bf48e9e8a49859f3a1d966054f844d
1321ccdd8f2935fc813066baf5d74dc830c1aad9
62123 F20101208_AABOFM bu_k_Page_110.jpg
4f4b92ba2901c1c326c5010d0844bfcb
7179b0b6e4a3eec3e305091fbec9c1f6ba21962d
36723 F20101208_AABOEZ bu_k_Page_096.jpg
d7433109283b57a1ee35a20269a5b8fa
3e95a583f8a2eb7e4815df1e2447f67ec29d7d94
27154 F20101208_AABNZT bu_k_Page_029.pro
d28bbd1148e53f36eae108081672f58f
163f413bcc1abfdb7f7d64927e20aa80e5021902
81183 F20101208_AABOGC bu_k_Page_128.jpg
40cba09db9096744afc8071b4af2a7ca
36dae90a329b2b8b136978317fa1973373b21902
38857 F20101208_AABOFN bu_k_Page_111.jpg
1e5e1b14d11fb5af98f4781e2a1690ca
c6660acda75fe8a188e1962d8d30921956741cc8
F20101208_AABNZU bu_k_Page_108.tif
2b06398c956561dd0756b5c4d5a15e0f
bd1ed54bbaa7b60e5374517e8d54c17a631d7297
87841 F20101208_AABOGD bu_k_Page_129.jpg
6f7f470677c39ed35c4db23ec2b84aa0
aa9b97d5bcb99cb04a42a62e1bee8be760998772
25573 F20101208_AABOFO bu_k_Page_112.jpg
2de9d8cc6f6e516375bf4d3b9e156b4c
e23e26b232f73f1d2379981634f33790406ec515
2558 F20101208_AABNZV bu_k_Page_018thm.jpg
5a674a77af5b7cc5bc367d277d0df29f
853a9a471703f3fc7a05ba75fa4516a6ddb2e777
97269 F20101208_AABOGE bu_k_Page_130.jpg
f6856a69de339c910ab172d10d1510da
b26deff3a5b1c39df65850920eefea764b4e431d
22392 F20101208_AABOFP bu_k_Page_114.jpg
64735480c46ad0fba95b1e9154ef8c31
c2e6389e943f7e7be2dcfb57e1f6b6f884f46fde
2681 F20101208_AABNZW bu_k_Page_034thm.jpg
552bedff40521360ee4b67623a616ff8
7ba7763f29fa3dbf9a1d606ebdb9eb948797c0cd
93707 F20101208_AABOGF bu_k_Page_132.jpg
9dcdd9dc65a1b644f62369c74f8e4852
00716752078aeb67bee6287c583b78c1f17c5398
23566 F20101208_AABOFQ bu_k_Page_115.jpg
2d76d0c7f8d6f3255340e8e22f3e7017
969b5df2be221a099a31b7dcaab75ef58c2f12e7
2086 F20101208_AABNZX bu_k_Page_043.txt
1c8d6e5b8c53db8ed8aa98b5df6a791d
47608650d58ddd55b2a4fd34491d087f0ec151b0
17257 F20101208_AABOFR bu_k_Page_116.jpg
9725c0057f134f1656a8c984905635df
a38481a366865c5c84c61bd65ded37d7f1ec5834
6699 F20101208_AABNZY bu_k_Page_024thm.jpg
51e9ae0cc1437aac63873b38c10ee9a1
5950ed964ab7ad5b33093ea1cfbd8fa796934677
53560 F20101208_AABOGG bu_k_Page_133.jpg
1ae25f3ba1eb02b2d8af06c93b80e9fb
8296d8dd6205a995fd5ac7e37e3e82324eb6ff95
21130 F20101208_AABOFS bu_k_Page_117.jpg
68cd01724523e96b5cb413d2a8fbb19a
ff9b4b51a67d3477f9a51691f1ac84b7afe34536
12183 F20101208_AABNZZ bu_k_Page_097.QC.jpg
dc0b7a75f9a4a5052755a7c7da554256
f08f2c7ebcc40cfdd8fd58d074bb1f561b521219
24652 F20101208_AABOGH bu_k_Page_001.jp2
8acb3f0835b02f1e29fb59c2be340758
aff1c13164b49daf1a0bed996626cbcf4247d332
22178 F20101208_AABOFT bu_k_Page_119.jpg
953de2b6df483b7a4140cbda85705bb1
bd85679b516d0698787cfc22c26d2d1cf7d1e48f
4926 F20101208_AABOGI bu_k_Page_002.jp2
436443c2fa0a266ede3d0124f9e91409
edd90c2dae56a67ca6a5eb60bf71acbffaf40d41
77137 F20101208_AABOFU bu_k_Page_120.jpg
075226640d07754d4d5b842df71ec079
7486316fdf3b028cf175cfc6a036f99a257569e4
7287 F20101208_AABOGJ bu_k_Page_003.jp2
1b4df271eb186d1f24dff93778c2d312
f09be0dcd7f31b6f8be7dc0c2960269b523c49ef
72360 F20101208_AABOFV bu_k_Page_121.jpg
fdf089b253f9031ec1f6acae3cde7caf
f503687de05fafcebf08e04c2439de2c5ceaa177
113274 F20101208_AABOGK bu_k_Page_004.jp2
5331667ebd02535a1b61a2d56db60ae3
d491fb07decf6c7a947df01c9156d44c2aa7bc30
78424 F20101208_AABOFW bu_k_Page_122.jpg
afb3ca4f7a97c652f78a8c1eb264a384
20e15e260955072848948080568eb9af85924c09
120651 F20101208_AABOHA bu_k_Page_022.jp2
b52866f6170471c74649fd0b03e4b720
a0104e28681521c82fd5039a1defba64d11d9be7
33520 F20101208_AABOGL bu_k_Page_005.jp2
c4c0e784af4c670d235ff3cf521937f6
3f334dbadebcec784335df005c4afe8cc0f96633
78135 F20101208_AABOFX bu_k_Page_123.jpg
2ca7f04e7612cdf14100a2e65c5c02b1
4a012dcc0220722c722985d8cd5d9398d455af45
114827 F20101208_AABOHB bu_k_Page_023.jp2
62afbaeec88d8396c7bc45631d26a7cd
ed0c795dcb8ee38cce59c3f511a5685f339db0ac
1051978 F20101208_AABOGM bu_k_Page_006.jp2
28dbf925f60aa28d728c621847ed66e6
b70a616e6c29fa1330c3962287b9cbe4edb3265c
28425 F20101208_AABOFY bu_k_Page_124.jpg
ec1627289394e7670f861214a20dfea7
e4fed26fb40a0d28ded4f9d9cd9ccdeab14efe06
114062 F20101208_AABOHC bu_k_Page_024.jp2
e304c94d62ab25184f8a5de5719f1e0d
a62e528e96e806ae2509ab8ab75a7937f420c960
963829 F20101208_AABOGN bu_k_Page_008.jp2
5b6f901d1f59aedb96304edf0709a7f5
ffe8d43be0cc6f32127fbe4f10b3eda7dec4ddc5
89431 F20101208_AABOFZ bu_k_Page_125.jpg
91eb160c1bf4229d23f28399ea15f44a
865604b9e35b99fac04f27cd303e00597653005b
118509 F20101208_AABOHD bu_k_Page_025.jp2
61a55d27b6f34563d9322d8aa8271047
43ba0ce75726874c44d4a4db62190d0178cc60b3
1051971 F20101208_AABOGO bu_k_Page_009.jp2
21083c630d90ee5b8f591165f4bca9c3
5d65fd8777db64103295158cbd8516ed76693d93
113401 F20101208_AABOHE bu_k_Page_027.jp2
9d8435eada2c118b6fd8ad3006fa58b8
fc891f228fb8a6fe46772df4b05a81206132b4d8
1051984 F20101208_AABOGP bu_k_Page_010.jp2
69a884b910e35d480530fad74a9dfcca
87f95f01b11193b80c756e83a01d851c95d405ab
88925 F20101208_AABOHF bu_k_Page_028.jp2
9e68b89e68bc8291eb455a535b24e64a
3dc41c32eaf2413d513d263c419622e4b1839690
185164 F20101208_AABOGQ bu_k_Page_012.jp2
a93f5a704a17b7927882334e309c785b
7b8ee50e8a00b1f818191bab5afda31d43d4076b
699137 F20101208_AABOHG bu_k_Page_029.jp2
6fbe4c2a683089f4a3217b1991e8ae23
538a3ed8a72647e711667cc216571c5976ab863e
98076 F20101208_AABOGR bu_k_Page_013.jp2
e92b60197eedd86429a23f5d2907e065
95f8f976b1bdf474aa036094bef7f6ce12c87c23
27071 F20101208_AABOGS bu_k_Page_014.jp2
280f64f5b73f214f3bb9e7d328359ab1
0a0171c7a02fc7909ff64eb2a1c3c8f0d5c79695
841622 F20101208_AABOHH bu_k_Page_031.jp2
312809d8d3f261236acbd87e8f9d4bf6
f0733999ef43907418374406afbb04e7b315fa54
109029 F20101208_AABOGT bu_k_Page_015.jp2
105f01762d1b5ca0f843184b5245e011
195cc9c5311f7aa1078fbb2fe7c0ee74b43b7802
31033 F20101208_AABOHI bu_k_Page_032.jp2
4a3bfb42efa56472fc7bde376dbf77ab
ca59e3bc623e25b0ad1eb50d173ed49e5afb7cf2
101157 F20101208_AABOGU bu_k_Page_016.jp2
21b448885079b782f838d3682793c9ef
c8ae9a48403844523b9ec8d3975fb1b1955715e8
388879 F20101208_AABOHJ bu_k_Page_033.jp2
496dac25d887f20dc3b79dbfbd901bea
9f2779f9cda3804c08f8911204d511637423ecdd
41961 F20101208_AABOGV bu_k_Page_017.jp2
9608ea79c41b4bd00f769e83137fb91a
32ac07df34b1a0d8e3e8f57afe80469b6c234e76
245464 F20101208_AABOHK bu_k_Page_035.jp2
286a7af6c7bbc0eb0a93a24cb4ec65f8
f144c003d7190de53d08d7ba47b921f76b971ef1
235855 F20101208_AABOGW bu_k_Page_018.jp2
0178cab2dc18b7ff8fa5cf86dcd55a5b
452c050534d3a5f42edb73267cedaaea276478db
650523 F20101208_AABOHL bu_k_Page_036.jp2
1ef3f53a5bb4394b8ac500ae693d6cdf
84832ed11685677a2e13867fb03f37ef97b39869
1051917 F20101208_AABOGX bu_k_Page_019.jp2
5bf07f5292a690ef190d31f9af0ee613
27ac6587d2f8dc73b364572e17b380b9e33a20b1
109573 F20101208_AABOIA bu_k_Page_051.jp2
40f4eb475cbb55a1ce2983a8d29372b1
f36b8fbd98923faeb580ca1bad4b5a0aa10997ac
111798 F20101208_AABOHM bu_k_Page_037.jp2
cbdcb607a5ab80546f078cfe3d934e6d
1257b95ea6768f9d6af6ec6b95875006e82012de
931622 F20101208_AABOGY bu_k_Page_020.jp2
121e7df05814917db102169215166b00
79e9e42b02c932ed6a50b507e1752f61296d9e6a
115503 F20101208_AABOIB bu_k_Page_052.jp2
ac6bb12d4be3dfadff3f3b1093dfb117
49f71de3e0c2c61fd295fa2c4c0d22a951ded64b
120653 F20101208_AABOHN bu_k_Page_038.jp2
88e859310bee086f0bfc905a201f2723
a017e089d415637f9828ea148c42ca45d6440190
109380 F20101208_AABOGZ bu_k_Page_021.jp2
9e547bb4e26f37099889baafccb533ef
03aea9fe9c94b63843607f3d9a920bf55298bd28
94733 F20101208_AABOIC bu_k_Page_053.jp2
404b265eae6f53f7e6ce2e17513991b9
8cead2c65985e96803c788cb9dd4e1abf54a0bef
108008 F20101208_AABOHO bu_k_Page_039.jp2
f8b704236694e4f0bc519efd043b6ecf
cb9d44dfd7fe0a103cbfdcaf25ed4843e4387f25
24484 F20101208_AABOID bu_k_Page_054.jp2
dd0470c6c7d9faba0715b2f468ab2557
e0fef6443545175ce63b2ba718f340dda5ee601a
97181 F20101208_AABOHP bu_k_Page_040.jp2
01203c6f16de3549220c1963361df631
b853c75e0a67e0b941abe086e1b08914c534b8dc
451564 F20101208_AABOIE bu_k_Page_056.jp2
40f1039a184378d6401d101b71720d5b
a288df2641dbe88a5ea79299ed5c4dc081fd87e1
97580 F20101208_AABOHQ bu_k_Page_041.jp2
cd058c883d4046fd69ee97ad8b08c481
16a77a53ac3e1dadc4adea5d379dc23f00a5dea0
266152 F20101208_AABOIF bu_k_Page_057.jp2
53aff2721402bbc3e858459c9f1b5d1e
a981112dbb1cf213ef083a96bd7690711b62f816
118345 F20101208_AABOHR bu_k_Page_042.jp2
9a5a6aaea1a8e360c06706f52155a5de
790615a2a2ad5a75ade3e868e411090e77608234
350468 F20101208_AABOIG bu_k_Page_058.jp2
b5dd56ab2f421afd5f098183ba387f9e
a009afd073a58fcb98fe0e91ada827c3dca13771
108629 F20101208_AABOHS bu_k_Page_043.jp2
f455300aa8a416e0dd28157e4637f00d
33622eb1ca8a03115a30457989bb7ed0741e1f77
448364 F20101208_AABOIH bu_k_Page_059.jp2
29657cd3683972cd3158f69b299194bb
23e793f2c94aa7cfc646830e800dac7a59d91fde
104528 F20101208_AABOHT bu_k_Page_044.jp2
a367c45554e00d83fd012749cadf8155
0c1f2d86b305e68bb9ca6d77c04c9acd98f17cf9
116279 F20101208_AABOHU bu_k_Page_045.jp2
ffa4bc80ccfa9110885a5c961709a8ae
e10f4c22e6767c60cb7c57ae6e3fbc179f0c19ba
439156 F20101208_AABOII bu_k_Page_060.jp2
f7e0c1f03d78e133a91884f777b81569
562ad395878815fc0f3dbf6322ee05b398fbe4d5
120869 F20101208_AABOHV bu_k_Page_046.jp2
4a4962e290908de1738971835d4def13
b84db67b3c0fa33d7a5dabaa8c6c7c07602569c6
199359 F20101208_AABOIJ bu_k_Page_061.jp2
5fd8cde97fc9aeb68aa4b640ec689788
4107e59dbd47aa5dd197770f39ef17032025d44d
115122 F20101208_AABOHW bu_k_Page_047.jp2
5fee30db20c46bea164ee102985b9c4a
4c52ae7c7aee616492c0913e58f207a8b70f60fe
514410 F20101208_AABOIK bu_k_Page_062.jp2
f694a7c11b649c2424d0a5c499b3ac32
24cccb1986fa2ae927c78b7bcfd559d68e5769ea
118329 F20101208_AABOHX bu_k_Page_048.jp2
01f6b12023135eefaa0fa8ff19282a60
f17e58a9a1983b26d3b999155453695921f88236
313845 F20101208_AABOJA bu_k_Page_084.jp2
2b6ee8aed61f3f4cf3d8746e5900f3a0
a4185502e2fc74f72689d79653e7db7dbe9c8a2f
242145 F20101208_AABOIL bu_k_Page_063.jp2
4a30342cd0084c06d45cbf022560c26d
3513dc72807ffb2c22c4086075358b89d0332adb
114625 F20101208_AABOHY bu_k_Page_049.jp2
3d69b08d5e9ed81566560abc2b8428f8
e12dac32b360f9d0a414bfad216e8c6cd1aea17e
305412 F20101208_AABOJB bu_k_Page_085.jp2
21c34147ea3285582af80ce0136c254b
87f6a9ce7b0cf48141c64fdfeac8f7eca8329962
894606 F20101208_AABOIM bu_k_Page_064.jp2
18504adb866a57ad4f41fc603aba5a48
ca6e55a0aeb55e7def20b988ad39ab576dd72400
104735 F20101208_AABOHZ bu_k_Page_050.jp2
b9c692c89ee8d36e4936ac699608df2d
f3fa104879bca1dab1578d449c6e1157753669b0
118833 F20101208_AABOJC bu_k_Page_086.jp2
03f1777fc63a57123384e72ba4953bf3
f5a703991b0f7b1cacf1193602f1e7cf7b0a7fa3
413875 F20101208_AABOIN bu_k_Page_065.jp2
59525296b88c60eb6b6d5dda8c375730
b2848986909279084abff9d8f5f4d66aa1fcae11
117081 F20101208_AABOJD bu_k_Page_087.jp2
629420b6c58efe8f9bf761c81519208c
63bcd1b77d33c9a0810194d3015b3823ae0670dd
1050133 F20101208_AABOIO bu_k_Page_066.jp2
56734d77acda5f63c6e23420e2f0b1af
c5383fe331ce54abe7162446d76cea5a7e5a6e28
1051967 F20101208_AABOJE bu_k_Page_088.jp2
ffcbfb39da7a616196e8f4ab86a6e3e4
e7046270c840db33ce4dccf36ce28c6a37e7887f
339800 F20101208_AABOIP bu_k_Page_067.jp2
863dd02416c4e8b965bb554a4568784d
f933f2e3d211a0665fd0e286b18946cfd86c4940
110576 F20101208_AABOJF bu_k_Page_090.jp2
11cad021804c8defb1b4414c294fa4b6
95f45f78a9b8997cc079c98350e58af0d70ee1e5
420226 F20101208_AABOIQ bu_k_Page_068.jp2
b5e27aa71f91f83e4c39a9ea5508b346
e2f69dd0cc0ac5872f36232d8f7aa75a9620e5d1
114919 F20101208_AABOJG bu_k_Page_091.jp2
ba642f3d9e76ed0df7aacc602dbd1323
080c2e1f1b0a48642e59d062abcbf650309a31bc
367869 F20101208_AABOIR bu_k_Page_071.jp2
a16def0bc7a7020aca422604be76e76c
f205e587a149e227753441faafe1f86fbfa098f0
49054 F20101208_AABOJH bu_k_Page_092.jp2
142a17b37e7b1df12f68cdc4f2c718f1
4b23d14baac5ad15f5e06ea700b6f55127460754
103537 F20101208_AABOIS bu_k_Page_073.jp2
a2f7f18e3f2af0e1e0261bd7d68e23be
15a8b4d0e8823686783d2c425619b85c5f9624c4
307121 F20101208_AABOJI bu_k_Page_093.jp2
554e8fc40ac7050e4b011b303ace90a1
1704854b26c03abda78cc3e2bf7e6debcd09d12b
108377 F20101208_AABOIT bu_k_Page_074.jp2
28b4837dc2e2c14eef65fb1027b5a3c5
bc6d7b8e30cf31e5004d6e8ab3654616b826e59d
115640 F20101208_AABOIU bu_k_Page_075.jp2
d75224c21d05675e58ded49c844bdeae
ef24c7dc816b58d132e55bb3ff0ef048108c87ca
438155 F20101208_AABOJJ bu_k_Page_094.jp2
47bd7526a6c4e0d9366a80b3b2b4885a
299107302ca7981911f2bdf78877fda98b96910b
109868 F20101208_AABOIV bu_k_Page_076.jp2
ec45e47203160d777968a355556a3110
3b0583658a1b7070ec324e2b77138f3c7ce8547f
285304 F20101208_AABOJK bu_k_Page_095.jp2
687ee1e74ff4e260ff1da6fc4ef20970
0cf9e7776728ccd3ae72cc82df0836c593a4072c
118003 F20101208_AABOIW bu_k_Page_077.jp2
a47191a16a8a0e2cfaf8228be78ffdb6
8fb7b897f64298dd44dd33d1ab6a6c73890550e1
26634 F20101208_AABOKA bu_k_Page_112.jp2
afb94be14d72db17970c83674d650a08
eb4c1a9972aa80f663b0826fc8abe52e2da2e712
826558 F20101208_AABOJL bu_k_Page_096.jp2
661efee129588a873cea7a475b07f184
878c35a884d93392361de7d10be51054ff3e81a1
521913 F20101208_AABOIX bu_k_Page_079.jp2
008806f2fb76ac7bee263a9ba6afd6f5
ce12f7405f170b429a6491c4eb0c4fed9ae01e6a
240335 F20101208_AABOKB bu_k_Page_114.jp2
7c9f0671f797d66df9247a4cace949f9
f0d8f1728ff55a77ea31edbd4937bf8c79f20e65
420937 F20101208_AABOJM bu_k_Page_097.jp2
f35cb4aa71875ed7399816a4f4c5792d
396444f7febcceed21ca6b72ba176fb4db20bb75
401352 F20101208_AABOIY bu_k_Page_082.jp2
eb6270ddd62ab0844cf57aac10b63e59
09d546491cf76873c7fd280a7dd706121d4c5ed2
275327 F20101208_AABOKC bu_k_Page_115.jp2
4ed709403f0178588eb33b4394c647f7
4d688b2e54a958469066ff6f413b747f6847241d
544196 F20101208_AABOJN bu_k_Page_098.jp2
eec77c3d23e9454de8b8f659cbe85870
61e7a28b7be23d5c3d956b17eb678b1a72d33802
20097 F20101208_AABOIZ bu_k_Page_083.jp2
7a5daac2b49e6d128ea941a1b592135b
f5810800eb84f59a344b56a85213bf2cbd3f940f
159442 F20101208_AABOKD bu_k_Page_116.jp2
e8d0afc0207495d0eafdf42e395361ba
46c336785d8f186a94bacafee15ef768d4ca17a6
315582 F20101208_AABOJO bu_k_Page_099.jp2
d9e778d348e8096a48a934e900b26570
ab48fac75d0d738e09b154ba5c99e444e3fbaf37
218578 F20101208_AABOKE bu_k_Page_117.jp2
242e5714cb5b1facf8029942175dd25c
6bcf14b623fae208f240637e62d5119d1f809f1e
110914 F20101208_AABOJP bu_k_Page_100.jp2
8292b24e7649ddcb8c5bb10587f38cd4
b5e27dfc5e27bba93ed8107d1a942b8815de0a48
225677 F20101208_AABOKF bu_k_Page_118.jp2
5e2503a885e3a1e254875272e1264d47
e5363ec5c31baeab374ec50d73645201ee8d2930
114467 F20101208_AABOJQ bu_k_Page_101.jp2
8a7985f597cbee2a4ec328d50a97d6cc
4ed169cfdc15cb5479cb75d8e18da9f255a7e36d
255481 F20101208_AABOKG bu_k_Page_119.jp2
1d073c5ab86189fe50795a866f3f223c
4e18733654943bf0281c371ac7b41cf7225fe78e
122195 F20101208_AABOJR bu_k_Page_102.jp2
6df2d7c5a199bea7da44f86823aaecad
ee8f7bcb2091d41aa31e79c16de58fac63597b39
116726 F20101208_AABOKH bu_k_Page_120.jp2
e52248b4d4f410e2c2b94c1b192cde47
52ed44375541a041c175b8adc50cfc860b500aa7
113070 F20101208_AABOJS bu_k_Page_103.jp2
327e2eccc8581ae68f7533a6c92de176
d72e744f7bb97234be8ca19b2ea0625a8bff7a71
107166 F20101208_AABOKI bu_k_Page_121.jp2
70fbb53c79ba78b7a5f2b9ef6574fcee
91b2aa56dd3029fb88b8a4f0b7fec6e8418a57bb
109211 F20101208_AABOJT bu_k_Page_104.jp2
9a744c0a4dbb6b500aff255bc226a0f0
1349225a836c59b03667649e279c132f2d8a2c1f
110848 F20101208_AABOJU bu_k_Page_105.jp2
1aba2fdef843d96c6e44d280788ecf97
729017513c5cba913c73fef3e93f4b41670fca00
119219 F20101208_AABOKJ bu_k_Page_122.jp2
6942ff9537aec853ae1af2c34c5a26aa
71858c6f0dab54dec164a1c946172ef88c8d13df
114105 F20101208_AABOJV bu_k_Page_106.jp2
ca78e9c1bd69653b6517eb4e8bb61ce2
8dffca51619d8d596b2b0fdd998e165b87b9ef32
118916 F20101208_AABOJW bu_k_Page_107.jp2
e62d2ee2192a12d2ab235d67ecd2e60a
277f618fba94eebf31cea0d5bdc3b2c32fe82514
119491 F20101208_AABOKK bu_k_Page_123.jp2
2ab6fe32190b69aa8c6fd3e8c5a0b4ea
ab0993279869d780fe8ec54ee88ed122ea641f1f
118491 F20101208_AABOJX bu_k_Page_109.jp2
f6ec096e096165f2716e9fe0c9397222
16d617a9413d05203194e15ad899d871916acb9d
F20101208_AABOLA bu_k_Page_008.tif
a51163e5c30256789c0e47036553cdd6
1c435dc581237644dac552509e40d3a889a1078e
38319 F20101208_AABOKL bu_k_Page_124.jp2
7fa9cd0846608e72ba7d51198f06b1f7
a9dabf9bfada8150f7ca706b5f182217f5cdc803
93078 F20101208_AABOJY bu_k_Page_110.jp2
6d9d13142001d2d06c6fa74eb75de052
eb91b046c2db9cc20cbe66ad2d7ce86f382a6bb5
F20101208_AABOLB bu_k_Page_009.tif
4300f51686eb3474b115139dc4065f4d
eed9db51d33de8aecced08ff6e62361562fcbd3e
1051972 F20101208_AABOKM bu_k_Page_126.jp2
7cc5b3d3fd68f71fddc7c95a7029fa8f
1366c294b66a70374f3047151c1f611f1f793b67
49562 F20101208_AABOJZ bu_k_Page_111.jp2
0dbb5ada5c0516b6514a9cb6a83fbc2b
078d141d8f2fbd1a03984d9d83919979e8e9c429
F20101208_AABOLC bu_k_Page_010.tif
3ac2752d05787418b912e4159ac45f7a
55f22cc6b8b03118e178ca0cd656f552969aacfa
139985 F20101208_AABOKN bu_k_Page_127.jp2
2b138635ce69ad2cbe5465b6e15cca47
c5b8f6165f4af4ae604a802eba0820385727df8d
F20101208_AABOLD bu_k_Page_012.tif
5fb113d2ab3817b3f9b38cffd3a1b5b1
48bc829bb8ef76f014d4a1024a47992143b75ec6
129129 F20101208_AABOKO bu_k_Page_128.jp2
24cef8a5332301b8914d1d621f084ff0
e958201336bd165ee840e9146f35792969a3b1b8
F20101208_AABOLE bu_k_Page_013.tif
e799c9a768683afc8c19929c29efecf1
df18e6e052785f821493ba7b62bbfd851009036f
137610 F20101208_AABOKP bu_k_Page_129.jp2
a72828854484c2c0014831221622c8e7
33741a21219bad155ad244b81abb80f59fc6d488
F20101208_AABOLF bu_k_Page_015.tif
bf4e16e03fe926b54442daba3277da9e
c3d387cde7c4f14c234e938af76b8d6aba95e4db
139312 F20101208_AABOKQ bu_k_Page_130.jp2
ab9faf212125a5f79785e83a95a81ac0
fcd012529ee4c9a0e9041368c06f7e30f15c7b93
F20101208_AABOLG bu_k_Page_017.tif
a6c8fef5081fc0556b851635dcb1b382
b162520ae3589e27505bc1bacf320c164669f9f1
138791 F20101208_AABOKR bu_k_Page_131.jp2
8b9c88366346d044981f8e3009d38bbb
aa163f87341664eb63a36114f3aa968bc7a4100b
F20101208_AABOLH bu_k_Page_018.tif
582c905bf1220a82a18da2f939b743a1
efe762e2d76b1baa9d86469aae6ff5f78df7ed7e
F20101208_AABOKS bu_k_Page_132.jp2
b8c676c31743731b070811c134ea41e0
9b1a6021314e80338431fdc2df3a9253c0e7d8da
F20101208_AABOLI bu_k_Page_019.tif
a82cc36d7443d6ea97afab3a4d442d80
9438efca0f1390e439d58e345eb0645e52ce5f00
76521 F20101208_AABOKT bu_k_Page_133.jp2
2d6d327e10e36157e0ca8cd94d9d3564
6e99a0b74172a0f004a6eea709017b60330431f5
F20101208_AABOLJ bu_k_Page_021.tif
2a6255ac71819ba504fca792768e3e80
daa98dee3c52fc7ed7980f717e6f3ab10c2a44e1
F20101208_AABOKU bu_k_Page_001.tif
11acfcd4cc8523cdf4c1a89513364ab3
708b51f80cc18e5d578bf54c1e505dd6b4b4e282
F20101208_AABOLK bu_k_Page_023.tif
e0f54965388d704aa77f9a94bba9e7b6
ccb4e2d54158c27e4529ed703581b5cb8ba02c3b
F20101208_AABOKV bu_k_Page_002.tif
f5db4e608c39fe1409f056348ef038f3
2d10f5c49e9457eea7d459e9c3fd58d16a2e5012
F20101208_AABOKW bu_k_Page_003.tif
e2a530c824891dd6ce6543d39495c444
8c031fa01f94e8926eb2185fb0de88b175c68d74
F20101208_AABOMA bu_k_Page_045.tif
3a37f54b2d72a0926b2226ce32ffd64a
01089a973c095520df807221df2518f8139b40b5
F20101208_AABOLL bu_k_Page_025.tif
eb83dcdcd099a012ca0c7f14c2af5788
6adae38c27364839379e1012ba3bf78213a7e2ad
F20101208_AABOKX bu_k_Page_004.tif
c495aefb79c653d3a1528ceab26e5e6a
ea23c2d3a85e0a7db9023de6043c5b706de2fa5b
F20101208_AABOMB bu_k_Page_046.tif
1089275e09bc1214a04a04cfd3e59be9
d104a58b16f9bb9862297a141e9c5349f2c60d29
F20101208_AABOLM bu_k_Page_026.tif
5d3b2bcbed6fc336c8b0fdfe3b1d4deb
64b93e8b99949166ebec46ccf3eecb81106c59ae
F20101208_AABOKY bu_k_Page_006.tif
048ec11865781bae6b4733f9f71347ef
345836c05b9fd4609b618271667ace9a049ee5bd
F20101208_AABOMC bu_k_Page_047.tif
0b6957c0889118ddd9284914cd3353a4
efe0c2bf29b13ff851505538e87edf90e7d15b98
F20101208_AABOLN bu_k_Page_027.tif
f5be19034ccbd309563a8d5ae99570b7
85f6451eaf5b98db4a62256359f1be5527975853
F20101208_AABOKZ bu_k_Page_007.tif
751680cd1bf26be65dad84893f8d013a
2776233bf2f2d9f8caec4ce5cdb6fe6bc0e43ad0
F20101208_AABOMD bu_k_Page_048.tif
2c2a172355dba6e397abd10543f8c7a5
ee572c06dee562961c3d2a1c4d414142774fcfa8
F20101208_AABOLO bu_k_Page_029.tif
b18c3a237535dc58ac5ca92c2a188229
61d07767e707fcc78bdcbf8f4ac53330ec0c6648
F20101208_AABOME bu_k_Page_050.tif
d55258ad3855430493d304d393d5a72c
308360124e5b53bd76d14b1309773967e44d64d9
F20101208_AABOLP bu_k_Page_031.tif
7d55fa8e784264b9208156499273c5ca
611b9cc1151be8141a1b3d0b897794ecb33fd51b
F20101208_AABOMF bu_k_Page_051.tif
8fa8dc6eddef2a72f832c14fe6a053dd
77336daca2600651205753b1063d5e77818d9fb8
F20101208_AABOLQ bu_k_Page_032.tif
29b4bcffe832bead56de1f5261c5f52c
265f5976ede18a39ed6a7944e5f8b66f0bdb4ca9
F20101208_AABOMG bu_k_Page_052.tif
39e67ca7fc6f7ffbbc888db08f3afd77
ba280c21e632308f395808fac049838ea7f427b3
F20101208_AABOLR bu_k_Page_033.tif
2e4cf47558a5e4dbc0814495d0c4cedd
29d9ba0489677d2deecdfc48544b34981a0b24ef
F20101208_AABOMH bu_k_Page_053.tif
fd2a5b4d6cd507d800eb66545b8aa3f9
d40833520726f97513f7fe12c07ef473a314d6c4
F20101208_AABOLS bu_k_Page_034.tif
d243c4c05e99a80cd0e4ce826a1cec14
cad9329bd36f8cde5de25d4410c447d9bcb25ca8
F20101208_AABOMI bu_k_Page_054.tif
cef21e5315719548c9889adb82ef200e
03bb38937e8ddfeaa871a80975cbfc8b62aa8947
F20101208_AABOLT bu_k_Page_037.tif
df121e046748cba53a9932db4d75d28d
2b63feff561549de99e8a0cea64e21edd730f641
F20101208_AABOMJ bu_k_Page_056.tif
59512f0212a5f548d199119fba585fdd
b3b22af304e8c0a46bc254292c374ee25fa6b4c0
F20101208_AABOLU bu_k_Page_038.tif
38484a8ad344a72902e0026abd428f7b
a5a260b161dcb70ff3ffc5b2bf9b5bae65261d79
F20101208_AABOMK bu_k_Page_057.tif
14c3d77a7aaf814e9105c4d0d56fdbf4
3a96b7ee5cefa546f00f9d1625c5922035b66bdf
F20101208_AABOLV bu_k_Page_040.tif
949d00e7aaad9d9f9774cdd1c7b548c4
8d5390506d9ae599437be1beb1f063eec1ab2e9e
F20101208_AABOML bu_k_Page_058.tif
ea54437ae2d882b6dda9c8964f665850
4b29d55752546e7c4d4b84fcb671eaf8b71f977a
F20101208_AABOLW bu_k_Page_041.tif
01d2ce60cdcefdf2861a06b6ea6651a7
f3ac18e358ef3098e9f77ac08271099ff1034ec1
F20101208_AABOLX bu_k_Page_042.tif
d63a1a45b6f6dec8dfcafc2229e0a05c
6210f739bca5785f8bd3ec3a63d4c0a1ccaf0af0
F20101208_AABONA bu_k_Page_074.tif
205d51567377c7a11f8cdbe559b12bea
0e8457832707577efcbb146ecdecffce84a9ee4a
F20101208_AABOMM bu_k_Page_059.tif
fb96faf5080a804d1ed36453b686435a
71c416c47bf1cb19f532af4f2ff91db1dd3998e9
F20101208_AABOLY bu_k_Page_043.tif
b0f881856e5a20f9245944af0689cc2c
ab601e344ee7be3eabba266b215bc15d05765034
F20101208_AABONB bu_k_Page_075.tif
2b425387b58f7f586af30289bfbb4fe9
0aad311303ace2a98e85baf0f0bcdebe9e457ace
F20101208_AABOMN bu_k_Page_060.tif
616bf389003d8d0224adcc15cb036bd3
c1ee3a89e55f075c5a09c660d99798058c437712
F20101208_AABOLZ bu_k_Page_044.tif
44a9ee6f449ea0f6d0292029fd02d906
ebdc3c1420ff49e1abd78047c54c2ea1ee21b815
F20101208_AABONC bu_k_Page_076.tif
16f209dad530ad27de4df4ecc8787217
e0199a45ce607d22fae99649659d4cb58415c9d5
F20101208_AABOMO bu_k_Page_062.tif
56dc3956c4761eb8fabf3874b8b5ef68
50f728c8596f8280c5c4ee5b0e231d245f732ae9
F20101208_AABOND bu_k_Page_077.tif
c569399736c89f367fa27fa5f9250e45
a88fa4e7bc855bcb454fef5795120f3c49d5a46b
F20101208_AABOMP bu_k_Page_063.tif
80b2d58f5915d8256779f18364434cb9
a96535623204f94e6912c51b3fac04281b9a106f
F20101208_AABONE bu_k_Page_078.tif
0d466950e06131398a1a6f7b6dee4e9f
e70bbbcd1f9ed3ab4b86399445088ff41c34104a
F20101208_AABOMQ bu_k_Page_064.tif
b09c89df3945e219d58aaec917ce8aea
46daf63fcfddf7cc20f3f0c4960692264a763d1b
F20101208_AABONF bu_k_Page_079.tif
09ef6dc9d6d48db881bf19e19db3a5fe
8a2a6a398fb7c1c1712066afb47f164d3944a390
F20101208_AABOMR bu_k_Page_065.tif
a774bf345df3820ecc837e2dc50c7903
34283be515da9f7ae6855f18e12d057380c2ca8f
F20101208_AABONG bu_k_Page_080.tif
75a2955e02d822f15aa5ad9f442f5c11
b41c8134a50a2df6847ef026a21f1f37f171a004
F20101208_AABOMS bu_k_Page_066.tif
a66612b33ccadb0e8ba2d52b573bcc68
ce11d9c8b626a6dba08dd7dee9fba8d201f79984
F20101208_AABONH bu_k_Page_081.tif
f4a1e42d7e199380eee80e848213cde4
f31901a43f0f1615930257e1cc867c994d680a04
F20101208_AABOMT bu_k_Page_067.tif
278664dfab4f427eeedec052167663e2
aff2672c068790365dddacf479db760e92692628
F20101208_AABONI bu_k_Page_082.tif
86a1a07e59a52a469465306fbf7b3477
075eabb97f01a22842bc86dde3e79970ce7045e0
F20101208_AABOMU bu_k_Page_068.tif
de949da93c4d668d0ac7ce0bf0f442d6
bf98160e1722aa8ca0b4c0f04485f4f20288006c
F20101208_AABONJ bu_k_Page_083.tif
d301a8bb702eb31e5a5aad4373092663
ffb18d11c8c4e221b1bae1aa0dea8b353c202cfc
F20101208_AABOMV bu_k_Page_069.tif
05ed42ceb64beb4e3d65d911493650ba
fc42fb4f74ea8579f3731bbfa522e197ca118514
8423998 F20101208_AABONK bu_k_Page_084.tif
6c85f828074860ada08eb12f43ce861b
836cf35704f22dc45ab2b196154e65ffd3777d3a
F20101208_AABOMW bu_k_Page_070.tif
35d8fb98d0e41e4a30c14ae53149b5ba
5fe3d83839af47ab2a0088de4a390b3bea1af2fe
F20101208_AABONL bu_k_Page_085.tif
f84769407b6853b3f3bde218be4978c8
131b803abb41e3e42dbdfe333c880ee57387a7ae
F20101208_AABOMX bu_k_Page_071.tif
dfe60f01dc0d21fa0738ca769a9da18c
9c53ee7486e039f7cc7333e2ad640355b0ab6d1d
F20101208_AABOOA bu_k_Page_104.tif
fe95879e591bfcd6847579b004fea8e5
39fd0219bccd66e516255acb616105f7c25293e7
F20101208_AABONM bu_k_Page_086.tif
25537882b4c589e764e92c41a88c77f1
fea92a686c37b17a0cca8bd8c9c0522fc2a972af
F20101208_AABOMY bu_k_Page_072.tif
6b25e97cb399a13ee8be209866fcf08f
07b6ff40ea0aa81e40051905038e990eedade6e0
F20101208_AABOOB bu_k_Page_106.tif
c691d88da6146071c2bc8ad96623d218
bb0ac3c401c014e1ed63f37808e5e36fce9920eb
F20101208_AABOMZ bu_k_Page_073.tif
56d7e1da0a09bd8dfd471c58715b5d86
2fc1eab9953b631f3e42c529a05bad043747dac0
F20101208_AABOOC bu_k_Page_107.tif
2225e1349899c8e6a2b702241f1f303b
b65324f6d3e9e9d1744121bb2d4468827c1ddf40
F20101208_AABONN bu_k_Page_088.tif
1dddca5b866389bfeca6b1ef6839110d
348c077fdebe0774285113b1180c2fc6c911a16a
F20101208_AABOOD bu_k_Page_109.tif
654d9883e906cdc3e11eb7de0ade4ba8
b8da922e40013a7a59dcdedd2df5ccf645d2b6c2
F20101208_AABONO bu_k_Page_089.tif
683846c44a7c5515fde66a50178d1278
1ff0c4a8d26a83f617bbc425347dd65a001809aa
F20101208_AABOOE bu_k_Page_110.tif
67175d4508aa87f0927781f99ad99e6b
b5e840438d94d115a9261742e810dfcc06f09825
F20101208_AABONP bu_k_Page_091.tif
87331d9a7245a463571ae2103d18fb69
a4bea241b2aff217bdceaa38d2938779a408c553
F20101208_AABOOF bu_k_Page_111.tif
6bb92328c29342331708377f097a293a
88e05e1fa9da1463c126fec0b2acbeb31c4c04f5
F20101208_AABONQ bu_k_Page_092.tif
1de9837c1b59166cf7bbe0db00c78118
ac7177646509f041930be8105f69ba369407050b
F20101208_AABOOG bu_k_Page_113.tif
493d63e929910dd2f115c926c17ec455
fb3c5feee7ef0110a668f4fdbeaf9a80a3e9f34d
F20101208_AABONR bu_k_Page_093.tif
75eab1de262c929c24c36f6a8b33df25
24298fa2ad20be3ab56006b7019cc90e44802971
F20101208_AABOOH bu_k_Page_114.tif
b1810575a28184e1a15a7c836fe008c6
43556f9d40dbf84ac42fa7e36dee2c3793bc5549
F20101208_AABONS bu_k_Page_094.tif
fb469bba43dd8b4844595f8ae1f7442a
2555ed0525599ca137e4c90e0f30dc8e0e749a20
F20101208_AABOOI bu_k_Page_115.tif
a890c99159ca68222f850fe2fd2a38ab
7b8d721a41daac08399618137f78287dc7c651a4
F20101208_AABONT bu_k_Page_095.tif
eefe5331194db5fe539f1c78a265d245
a5225a8f37eef1e888d813a1a3b534d0db3b7af1
F20101208_AABOOJ bu_k_Page_117.tif
fd9ca306f58c7548e37fd91c47e24685
81d7ccb2c0457ed1f50a982ad22f55640c538470
F20101208_AABONU bu_k_Page_096.tif
93da93efceee3d7730c67209e0df7847
1e1c51ffb1228eaff3f19d1d28dcf03a0b72ee79
F20101208_AABOOK bu_k_Page_119.tif
c54f44cb114293e3e3b0dda9b3edee4e
747bc04d1cce3507a75b0146c4069f3dee9d8073
F20101208_AABONV bu_k_Page_097.tif
1321c936e0ce2031c097b49d823e72a2
4744a5449c5486782ea7b6a0b724c6952f6a0711
F20101208_AABOOL bu_k_Page_120.tif
166291e0d8e3849f8841e7ae79f861ef
6f82184fdd3f5239d60ad7046af5a9f46e6eeff2
F20101208_AABONW bu_k_Page_098.tif
0ead492caedcd866bd76575b18d3cb15
053b91c719a66d1e32235cb8c88c4f7f221b562d
14294 F20101208_AABOPA bu_k_Page_005.pro
b83582db6d2fdc5e0d719b7d3b2cd046
be8f55e5e067fa9c5534d874e1ed397d764dfdbf
F20101208_AABOOM bu_k_Page_121.tif
f30758ff3937cc44dad36e9dc16658fe
3c651200552afa60b98142d6fb7ba9e541dfda2c
F20101208_AABONX bu_k_Page_100.tif
cd5d721fa9e64ad566e5bd5ca4047726
218daeacfa2c6e675939d939083c3db6c2367a0e
88027 F20101208_AABOPB bu_k_Page_006.pro
0d4c18389644d6f43372eb5be4964f5e
5ef68a2e540895c1bde147a19e432f349e23b453
F20101208_AABOON bu_k_Page_122.tif
cfa4a2a89911263bc83b24a3ee86aaab
32071d37375c957349faf260c650f68ad6591cb3
F20101208_AABONY bu_k_Page_102.tif
54068ec2a4dcb3f188fec45666f47a29
d40fed20557a993957cf391393ba95c90071a5a7
38952 F20101208_AABOPC bu_k_Page_007.pro
dde3845361068c5c8b06f30b01613edf
e65db50d8e5efa50df26245cdada2d4df87e8fd0
F20101208_AABONZ bu_k_Page_103.tif
29ab618430fcd3d16f8b85935166af68
882b6652cbe17a28ed0be702642a6f80d7fc5206
78464 F20101208_AABOPD bu_k_Page_009.pro
e5d00188067d3431d0f5b72d51ef66eb
144f084242659298eccef0c6cdc7611cc6c4bf10
F20101208_AABOOO bu_k_Page_123.tif
2e4d04973c991b5ca7bbfdd660c53e16
7a9c9de0f71dc81e1352f5045bce8a174b2dc0c4
83115 F20101208_AABOPE bu_k_Page_010.pro
af81ca78c9e67f6905acdad0eba4539d
595002950d411cc62383ab163bd1af125d26d4b4
F20101208_AABOOP bu_k_Page_125.tif
e0f253fec84723f98838592e914da69b
81eaa024151fa7ef8da28b551e5f30eae669fb8c
4412 F20101208_AABOPF bu_k_Page_012.pro
79b727025271be0d9d570677ca51ecf1
054763e91164611c026b026c4958c75fdf1813f7
F20101208_AABOOQ bu_k_Page_126.tif
812419c6a55154ec9cbfbfa6c94e0403
dcb138a78c85fde2a08d6052c873c6a8b441125c
F20101208_AABOOR bu_k_Page_127.tif
0062716ed80d1630374db77bacef5bd8
00c5bbdda55b6ff7f662d411e7cb56d495d16737
44997 F20101208_AABOPG bu_k_Page_013.pro
bcc762d21d4a4fa343b04b4612eeceee
846bd8d2e01d9dc97d245a84504c9f906d79ecc4
F20101208_AABOOS bu_k_Page_128.tif
ff9b224f62a1b10ab0322da513e6820e
d05270008ee508b1d9756b4909568d9561cd8579
10639 F20101208_AABOPH bu_k_Page_014.pro
855ae7b1a12ea354c94b1e3c3fdfc6c6
a9f77c0ee51024d1b1cde27be1124ea9366bb2d8
F20101208_AABOOT bu_k_Page_130.tif
5b28729279ebd6cc5c376c5f0b99cc49
8b161da993ad7c28dbf36b9d17c8b93b97a07a3b
50375 F20101208_AABOPI bu_k_Page_015.pro
7bb0b890ca59e31988bd077e6a12b210
30ad296657ba274d9d330cfde5819a76d8d27571
F20101208_AABOOU bu_k_Page_131.tif
80509329124b5e98884b8eb874e714c2
46481662641eca628cde317960f011520ce9e64b
46340 F20101208_AABOPJ bu_k_Page_016.pro
bd59274b1483579a5168d45436dfaea7
7cd032357f63d1067d424a2125da57de5d4c9ca4
F20101208_AABOOV bu_k_Page_132.tif
b3507789c1178f194cc39bc9cd83e65b
ae06d0d37baa957548f973bacc45ce63fa39984d
5300 F20101208_AABOPK bu_k_Page_018.pro
98aa9182fbaf97bbcf442aaffee9fc91
dd24f7a6fb6c4684fcb1f388a002094804b8b385
8002 F20101208_AABOOW bu_k_Page_001.pro
5d42591187dd71d2afbe09b4ab1c750c
33dcc44a0ae021e0a1a3f98fac523110650ce765
5097 F20101208_AABOPL bu_k_Page_019.pro
d98f933fad0f9bae78c108b73f4fc85e
a695b361148f5ec8ca4fa113791788f94d929ccb
779 F20101208_AABOOX bu_k_Page_002.pro
5479b6fb193bf9465b51f3a120088a05
defa65e79bfbe2a318d9473f7f9067deb9fa5914
3986 F20101208_AABOQA bu_k_Page_036.pro
a8630d06bd8443dd9d6cd9cafdde27f3
c7ba224837491703f927a5737f6ecb74c5240a23
4233 F20101208_AABOPM bu_k_Page_020.pro
d8456529bce704e6b258fc12ffe2871b
a407b884927a16ac2c660c82268bf667fae7c258
1713 F20101208_AABOOY bu_k_Page_003.pro
c8b509f09d9f82cd26e4177362bb9583
b9d7326fe7d2abaf633a5916a951acf8c4b09d58
52419 F20101208_AABOQB bu_k_Page_037.pro
9c3998ecbdd15b699c9a51054329b48e
a9c1c0977ae2f71a84c3f657188db5cde07ec6e2
51745 F20101208_AABOPN bu_k_Page_021.pro
3508d1677c1c70cfaef3a9a89fd15f8b
88eaa726022b6552e62297acc60d8119ea5819a6
51083 F20101208_AABOOZ bu_k_Page_004.pro
bee1a1cf158bcacfcb3ab2fe82a2e172
b6e07726016402fb5cfb6110fc80f5baaacadc22
50032 F20101208_AABOQC bu_k_Page_039.pro
9001aa70e72bdf263e3078c49af0c634
e7a42129e655d3230e2cd20945e8fc354807fac8
58179 F20101208_AABOPO bu_k_Page_022.pro
f1492159cc93cf2d8ad19f60b4e3ed65
eb96be9d5a72a2aa99d7b20d3a94af0d1dcdb608
46165 F20101208_AABOQD bu_k_Page_040.pro
755c2233a69bd80a09a4c6a26b3fd2f2
84d2e3085f609a63483e79f83c3ac9b323c8bd6f
45874 F20101208_AABOQE bu_k_Page_041.pro
90297d22da96ea74211fb9d23a4c93fd
ca8e8cca47bc8fa002b6397ce9f0dcbe9401c8a7
54575 F20101208_AABOPP bu_k_Page_023.pro
052aac8d5ddead8e16f5f478b662daf1
65c4b655a3f7851500274e45953fc58ace2afa44
56557 F20101208_AABOQF bu_k_Page_042.pro
1fdecf273e5d1aba25c0c6c6c0b87d62
07d85d5953b46c2964e19a3f4b7fa4170f4841ac
52904 F20101208_AABOPQ bu_k_Page_024.pro
2612d13371c96eb41de35f6089810fb0
63ff46b0ccbf6eb8088911326b5910968323281e
51955 F20101208_AABOQG bu_k_Page_043.pro
e59b1096e8279f350901be3bc2a88c40
b56b795076ae58c2c735d71e418fb7f1ef1f009b
56686 F20101208_AABOPR bu_k_Page_025.pro
9a9b832de92a18285160fff1ba141bdd
93978ede93c992d672751bf4b1e15da66ec59578
55113 F20101208_AABOQH bu_k_Page_045.pro
2a883338932b31f40090690bfb1d403f
bc21da13533a789a3e5d0f63be24c6d3f8850b3f
59024 F20101208_AABOPS bu_k_Page_026.pro
9f8ca301afef655d4de78ee8cf001373
94aa95cf7a20216cd68c4c6466a4d9254b333bc8
57241 F20101208_AABOQI bu_k_Page_046.pro
1492282511f41885c184e7c6458ce750
7117ef78fc6eb38b9c45b75482a92bc33ff83a00
53427 F20101208_AABOPT bu_k_Page_027.pro
5e336ab875b3e9ff2690cfff6ee84889
bbed60b3c5ddda5a453753960f6fa10eb3803868
53991 F20101208_AABOQJ bu_k_Page_047.pro
9442d77eab73fbae9809c17027fb06d4
e5fe9d48cc612fa7e38f3f8d5470188381d38ae5
11394 F20101208_AABOPU bu_k_Page_030.pro
70b1c11ca5121f7d96cb604623afdd81
87db0ac47f43f182fbb5e8cb9a7cd94fc467f188
56891 F20101208_AABOQK bu_k_Page_048.pro
ac4fdc41ae77e6d65e0acfd8c286defc
6ce5d107a52f24b165164941b38693ba12dd5ca8
17433 F20101208_AABOPV bu_k_Page_031.pro
66ec51f76264d3bb95ca1d3b1f27d4b1
bd0d3be1652b433602789992323870511dcb0abd
53958 F20101208_AABOQL bu_k_Page_049.pro
22ef3395fe9b6abae5fec62c2920bd70
37839fa1355fe4ac791364e07a48239062a10a81
7431 F20101208_AABOPW bu_k_Page_032.pro
d06f7f0ec4a086c494a4b97f1c9d2745
4217e702d222971d33bdc2bbf924d10872ca6ba5
14173 F20101208_AABORA bu_k_Page_065.pro
a446c62d7dc4038ad44616d5d91dd057
53d0b42c0dde7f9b4672b3d8670d566d98460c1b
49373 F20101208_AABOQM bu_k_Page_050.pro
41a43233912dd7d01136c24bc4219175
add981581253cda08c104a592db6b135a62146e8
10852 F20101208_AABOPX bu_k_Page_033.pro
2274c7b45db20b2221ce0c02155cb55c
c4e1ff5cb2de25e0795675a0ffacbf3f853d0d1d
F20101208_AABORB bu_k_Page_066.pro
2c589aa13bdb0b67e1d1521090626a17
0395048f0fb2c41cfe16ba95c86e87c2b8057aca
51533 F20101208_AABOQN bu_k_Page_051.pro
30cf9bbf673523dab2621e4fc2b1c4c9
2c0319d3b5990ec86b04936b0b84afdaf4913a5c
7483 F20101208_AABOPY bu_k_Page_034.pro
1c06316f5d01880e23886e1c27db3262
f3fb0b71b30c40cc7f7a5415fec84471a482c304
4225 F20101208_AABORC bu_k_Page_067.pro
c2131f1e2b0509020736bf25f448d1aa
e1b099bf461fbd95e938b6a46a0d41820e123a20
54213 F20101208_AABOQO bu_k_Page_052.pro
2af8aec392fe9c898422ec0b19e9d635
1da6abf4140d0525955e12b891b7fe5480828c9c
7342 F20101208_AABOPZ bu_k_Page_035.pro
86f498c42ace9296ddd4cad13ddd65c3
6275ed890eac248a32ba5161256e9fd7b10495e8
9542 F20101208_AABORD bu_k_Page_068.pro
1d92258e9eab57f62d4734a680a150ff
e82b38f914ec173418cfd1bfa962fc15071a6280
43647 F20101208_AABOQP bu_k_Page_053.pro
9fcb974629278303909e962087ac313f
b9fbbc4494d955b07e72034bb7beb1a55ab2261d
5951 F20101208_AABORE bu_k_Page_069.pro
155431da78399521674fdf410c5eb27d
fb7e8441187821484beb75f881d0212716f0c889
4647 F20101208_AABORF bu_k_Page_070.pro
a6edabd612cfd6ae60d51b1f0449bf53
dfe4557c80a885f0fee268e8fba4375b8c2e4259
9437 F20101208_AABOQQ bu_k_Page_054.pro
c20b43dbf67b635a1ff98c30cff60d52
50b52afd126497dec50d1fa6d86971088173bcaa
12020 F20101208_AABORG bu_k_Page_071.pro
2f8326d8884892e47062be056a76e5d5
3f3de077537e01cc53e5a12609ec3d9f6c1fd869
19369 F20101208_AABOQR bu_k_Page_056.pro
01cc98976cd19e351a05f183c3b76504
4ccb1e2892bca85a56c09e4c0e1d92d989d39843
52835 F20101208_AABORH bu_k_Page_072.pro
618d321ac090abf107ef45481241df3e
0c0fa6d0b0ca8f4b83246d9d52c21e7dcc362c35
5394 F20101208_AABOQS bu_k_Page_057.pro
fa0c6c339b4da28a241c6f13a316c756
81d1d0c68a1933c23ec49dc911648c3fd52b39e0
48390 F20101208_AABORI bu_k_Page_073.pro
e8c78f27073f64db838f00c06affaa9e
1bab2774fb0432bb92384d5ecd4032214ff15626
17490 F20101208_AABOQT bu_k_Page_058.pro
20549a4b6dd4aada5f98678d534645df
2a0a4a479bcc9b5c075c2af1baa2d30605213ce5
50236 F20101208_AABORJ bu_k_Page_074.pro
5b9b6ee06264f6bade9fc4b0001244dd
79b61a9ba1e8ec81b57bb03e0abede42b55fbb6b
19546 F20101208_AABOQU bu_k_Page_059.pro
b3d80e350bfa25ed3052216313a93957
483418fa97df8723104d6b338c650d6cc405ef88
53422 F20101208_AABORK bu_k_Page_075.pro
0567a9aef24be7e39ff45ca13ec68091
070e8e13ce5f8a3948cc79d9cb58ad26a5ffdc47
9620 F20101208_AABOQV bu_k_Page_060.pro
e7428945daf1d2b1b805eb808d69345d
5556aee1335c8bc5d8e4e2404eb37d18e8a457cb
51721 F20101208_AABORL bu_k_Page_076.pro
b41a9b664f65a4da86890777c4289bb4
39ae09b8ea8d964cc29c5162e2de21c4592d5da1
6657 F20101208_AABOQW bu_k_Page_061.pro
7f428f3f6db17c52712063925eae139d
9fee23ad7fed9e092beaea5f98fb46d543ef7b31
54474 F20101208_AABORM bu_k_Page_077.pro
ffd5872cb67a96f490750cc29d40c2b4
a1332d670fbf333763b106c6a022791914a53a6d
15799 F20101208_AABOQX bu_k_Page_062.pro
3b39c6329e44ceef31cd5207cea7080b
c44adbd436ec1e02f32ba45a8c4d2430d07df019
21099 F20101208_AABOSA bu_k_Page_092.pro
c3dc6c561a102c40adba268459f245dd
918cdd3c143f7a4ca775bbb15586f3728a3f7352
41983 F20101208_AABORN bu_k_Page_078.pro
fc1d133cd761e9ce7775b7f82c3ebf7a
9102fa3b7b14a481a67f0fb0dd019f44d2ffe461
8180 F20101208_AABOQY bu_k_Page_063.pro
d411452e35252bb54bcaf1e7f964ea43
c883019905f97dc5962d7eaa513c4f335d0fd1c7
10987 F20101208_AABOSB bu_k_Page_093.pro
e612ab5ee7f11f2d791182296f6055cb
378f217c1ccd339ade2c36594fff4c3a1988088d
14737 F20101208_AABORO bu_k_Page_079.pro
0c8986fefa0248ec41707a82f7920f31
2e9e3671c9ef8b222cd925a9345ce70b8b90bbe1
14433 F20101208_AABOQZ bu_k_Page_064.pro
1bcdb41ade82c9967f204299baa02a9c
f5648399db207ba2488a41a1e62290741ad3debb
11908 F20101208_AABOSC bu_k_Page_095.pro
243c24c4e69e1b22ab8d2ba7963dc3c3
eef61d4d3aa2565515b17c57081516f245e40b4e
7284 F20101208_AABORP bu_k_Page_080.pro
f9832dff351599ff2c7df3c11bb9a901
e7a379a1179c6ce3f6d03c6ed5ef449e493b5710
9168 F20101208_AABOSD bu_k_Page_096.pro
e71753451a7b854f5f31428f5db114f5
daeb105b0834918f0a9c6194ac991265039ec460
22377 F20101208_AABORQ bu_k_Page_081.pro
7a95a2822cc639bbf46212fee0d8c24a
fc64d3256c4a09fe65a258c9e221c9a18a5f1675
17083 F20101208_AABOSE bu_k_Page_097.pro
a05793178c8c30ff911c3595776248de
71d1327c06e6fbe81f7bc0c0294771286e614b3b
14727 F20101208_AABOSF bu_k_Page_098.pro
c688b90594fd746f041ed6f3b2986d6e
c547b81b27fc3cc86644e1c1d7c64b35918b170b
14518 F20101208_AABORR bu_k_Page_082.pro
ca6c095ed627389c41028f360375b063
a925541f950d90b4e1090bb378d0b7823f2d9184
4845 F20101208_AABOSG bu_k_Page_099.pro
227a461a48f6a8d4c972e283f8a456bf
e18e6fc5597bbb46339d7088e5490b824db5221a
5757 F20101208_AABORS bu_k_Page_083.pro
4335fd36cb62c041de36e1f35af59389
cb82cf9b488c582e882746350b3409e882de4f0a
50936 F20101208_AABOSH bu_k_Page_100.pro
60fe18595d28d16d3368db84fafe761b
8bff6a132ef702f9c3f2ba45812e54381e22df18
13574 F20101208_AABORT bu_k_Page_084.pro
ec03f7a24f34a74a50a306bb24fb1183
c6d626b93c23a5a5b96aa4ec3e34535ada05994e
53479 F20101208_AABOSI bu_k_Page_101.pro
0b30b76298812f7734023efdce38426d
3fc76d1759b695049fd821da6d7f5b083f4dace8
10574 F20101208_AABORU bu_k_Page_085.pro
899cfac8fe703105150ba424536dff90
f12affe74096f1bec31e17e994fba69ae6f5e849
54185 F20101208_AABOSJ bu_k_Page_103.pro
543043ba677fcd477ef3e2ce9f045a6c
e5690529544d3bded0d23854e98c90645c439d76
54033 F20101208_AABORV bu_k_Page_086.pro
183f8d83ee8c9fd7c61160f008e8b8d7
c4af831c814f177fd6d59770232bf7544a2665ec
51827 F20101208_AABOSK bu_k_Page_104.pro
c415b97370e744482fbac747fe2a22cb
6ddbb1e26bac8b5505a9265fc5e6bb6602b228a1
54442 F20101208_AABORW bu_k_Page_087.pro
0a64b1a5bcedb361ad1c61b8de5fde6d
6c0d6b3c9532dc1d65011c798792adc02cc560d2
52849 F20101208_AABOSL bu_k_Page_106.pro
fdcc9f9bb56c8bf636bb10567dc67798
f2fa45f9e59a47a9112c2d64ab3b68d61af28260
55890 F20101208_AABORX bu_k_Page_089.pro
612d679a342e110bfe4f1e45d72fd9c0
4aebda5caa2e6a8a69eba7e35752679c97f516db
15921 F20101208_AABOTA bu_k_Page_124.pro
ceeb2e4946b83c5beeb4c79ab377d5e4
891f369a11a2900da47a5f635816103484885755
55940 F20101208_AABOSM bu_k_Page_107.pro
13012dbaa50b0dcf93d06bdda139852d
103a10c5f847b37e8938b2ff5574abf7af8ebd58
50511 F20101208_AABORY bu_k_Page_090.pro
1c2fe8ddadfc1b90d0f01668dc6de338
97c86cc41c745719aa0f1337fe1549e741538743
56328 F20101208_AABOTB bu_k_Page_125.pro
6570ac62805d80c0c3e666ab79716e60
15b4f108cf301f76707d1adec5f0782ce8e686a1
56782 F20101208_AABOSN bu_k_Page_108.pro
889061992ad16c4e8c017eb05d91b878
52a7b7b6ac1bcb70eaccecec362c9cff9570a9c9
53135 F20101208_AABORZ bu_k_Page_091.pro
2ee85755348b931c54044a30d8fdf943
ba84764367d0dcdfa96e57ee3d3738c4ced6e1ac
62352 F20101208_AABOTC bu_k_Page_126.pro
5b59c03cf17add237804c14a4d65dc64
25ca00c5abff65f2589838117c97623a342408f3
55261 F20101208_AABOSO bu_k_Page_109.pro
6cb8898421cf8e42582f5b6d885220d3
e2b5f6aca0b04391474bb2871aef4b3084ff9bd1
65853 F20101208_AABOTD bu_k_Page_127.pro
1055c4a903245d8b021afb9f56c28cf8
6e7d074664e09e8ecc39fadb5d26982275cf6ac0
42503 F20101208_AABOSP bu_k_Page_110.pro
deb310ddcb4ff8b29509207eb83871a1
43fd9660e60984c39cf84f1bb2d3c38ed1c203e9
59100 F20101208_AABOTE bu_k_Page_128.pro
62930670ed81980ab379985ea383b168
9ff6ec9a87890ad97ebd4e36ec500f3d00a1ee39
22742 F20101208_AABOSQ bu_k_Page_111.pro
593a6129a40770a0a1e8e08e1475baec
4aa2fe92f526c874d3e7993acc60bd9dc84339a4
65467 F20101208_AABOTF bu_k_Page_129.pro
b7360084d461d7f5792be2e277ce46d7
3ba3412fb77a9bf54210d678d6210f6f5d4e6474
11805 F20101208_AABOSR bu_k_Page_112.pro
79fba64bc8fd4bc656e2003dddf8cd96
71187a119f429694a6d03c193c0b3d2b912a7b85
64747 F20101208_AABOTG bu_k_Page_130.pro
4fb045b1f09434bd59f8e5b89a9095d3
90cbd4e7039831be3ec25ea35fcc14933737f053
65628 F20101208_AABOTH bu_k_Page_131.pro
42c02a936cbe9c79c98d26333281d64a
bd53393b4daad4998d0ebb9150fa006c73199815
5361 F20101208_AABOSS bu_k_Page_113.pro
a184ff653e6baf00e8fff29277cdf27d
8f46061568ae0eeffc28b856a6c53a720cb92dbd
58321 F20101208_AABOTI bu_k_Page_132.pro
044c812dacd16ac0d9e5626c25cfcc63
499faa1d0260e3de4e026623648f8058e84788fe
2715 F20101208_AABOST bu_k_Page_114.pro
dc435278b4f3975dc53687ee85171e78
03943fbaafffed55d627a7193b52a8394ac4c0b8
33321 F20101208_AABOTJ bu_k_Page_133.pro
a7080f9c7800d720cdd814b9e44bf8c1
245d104a790c9deda6daa994f52e19d0f8159ccc
2870 F20101208_AABOSU bu_k_Page_115.pro
7c64719b814148acc36a0c8c53c3a538
75f501a3169bbbca7745baaaba66977146e6f719
F20101208_AABOTK bu_k_Page_001.txt
6e1f37893fda10f0a05157ae7fda4264
42d4d3ac8eb8cb6e4dc6e488ceb37d17035689b9
4172 F20101208_AABOSV bu_k_Page_116.pro
e6e65c5c8370cdfea15eee6890d4800e
1ddba4f47532994c76b9af2bbf6a8e491fc14e80
120 F20101208_AABOTL bu_k_Page_003.txt
f251b451be4b7058e220142b25597653
d2be2c7abce7669ed9f68511b04c4745071b0757
2976 F20101208_AABOSW bu_k_Page_117.pro
243dddc04a3aaf4ccdc8e52b9accbf3e
abace47f03935854f90acff8f71cb4ab2733e1bb
2288 F20101208_AABOUA bu_k_Page_022.txt
b66b53d2acdd8821f6f998f56293dc98
4793e97024f196816b66c6f5168ad9bbc572690e
2048 F20101208_AABOTM bu_k_Page_004.txt
2e5d3a931027065e99cf521eb1c51e43
f68c721a9ff4fe1ca3c5ec50ad2b6e29e38a66bb
53876 F20101208_AABOSX bu_k_Page_120.pro
8c4a2d0853e617cc96cb46e696312c9e
066f0fd6b57d6134fad119833c6bdcbf09ae62aa
2148 F20101208_AABOUB bu_k_Page_023.txt
476d61402b64c722db5dd49a5c662f4e
c2dc526f594c7427bedadd981c2abc8562356646
578 F20101208_AABOTN bu_k_Page_005.txt
f03ef6238cdaf517f53753f86fce5723
c94e0d13663815f533c240ed4f9ac324039e1e61
55418 F20101208_AABOSY bu_k_Page_122.pro
b50b60b83a3bc9eae634d379f58f4d31
22e9a7a478234534aa4f9b198629dad9030d7eb2
F20101208_AABOUC bu_k_Page_024.txt
959d88880af03623f279264d43a8a3e0
086fecb05d6787c5eab9dd35b6bf811e2dde7e0c
3593 F20101208_AABOTO bu_k_Page_006.txt
04336a47cb6dae38baf0373c36ca7961
149cf1eac143d9c4dc925eb72b6b9547eb6c1c41
55552 F20101208_AABOSZ bu_k_Page_123.pro
f26fd7c64b16b27b26ebe71a4dc603f3
f5090ecaffae149954a0c8372e7ff14a9cedddf9
1575 F20101208_AABOTP bu_k_Page_007.txt
8d662d5413f00fa5268b32846903de76
45d8f24c9f2d0d0abb571f5eb37668c444e02ac3
2234 F20101208_AABOUD bu_k_Page_025.txt
f26cd58884fa97d934980df59e57ab42
b0f8436a9b7fc95916ead3926daaf66e7ae755a5
3177 F20101208_AABOTQ bu_k_Page_009.txt
b7677b9d39c106fdb50a219d40c97cb6
68149dca4c3ecc9ef8c3561fc6da87c470b68435
2321 F20101208_AABOUE bu_k_Page_026.txt
f1d49e10bbc2149f1a95e4dc803f1a15
4e8cd59507ba8d2996d913249274412f1c8253f3
185 F20101208_AABOTR bu_k_Page_012.txt
6719019afdbca7f302d8c6c29da4b093
b31849f841fb86f3b9d9be6b21564c8f3c029f04
1660 F20101208_AABOUF bu_k_Page_028.txt
3d7d96f5654b285f29d48210875e9039
11b2092db5188f4ce522353ea8e28080034aea6a
432 F20101208_AABOTS bu_k_Page_014.txt
242745f8c2a073752039697e6cc3cbf8
22f05d6d5f1bb0b0543ff80e9e1a95f4cdd66752
14918 F20101208_AABPAA bu_k_Page_036.QC.jpg
63b7fda77e1e828e9f642159ff7f4c15
d3e4f09c92ce460999feb4c5445ed806b48fbd97
1011 F20101208_AABOUG bu_k_Page_030.txt
780af3b33e78dffbbab0e8977060aa8a
fc0294a41a2450879f2d05a2a22a620954a8a2dd
4920 F20101208_AABPAB bu_k_Page_036thm.jpg
cca7e6b179448aa6b0e2fe6d12887f37
42a09642ff4ffe2f6cb94ddd91855544bb80ec25
963 F20101208_AABOUH bu_k_Page_031.txt
8c0c67169afe1cac06346d14e6aa19ad
b5a50455bbb102169f3e78584dfc748ead0a8824
2067 F20101208_AABOTT bu_k_Page_015.txt
25b018c2390a67087e891f795c385e0c
a32ef521a3fbe67016639a0d8840a8fd1c936ed1
24846 F20101208_AABPAC bu_k_Page_037.QC.jpg
f7bea2927ee6a275a3c1cc6f06a64b97
e97baea16faca9bd81e9ff77779541b8cfba9794
536 F20101208_AABOUI bu_k_Page_032.txt
05bb7752f3b09635a84dbf95d10e2d55
f59436be61f3ec24a2f2bea0acccf347862aef36
1844 F20101208_AABOTU bu_k_Page_016.txt
01e3fb9606d10c31f79c8f335bb8c8c0
971d3229bc9b36833f90fba489070a5cdf6dd53b
6831 F20101208_AABPAD bu_k_Page_037thm.jpg
350b477947a49f01b3ab0817b99f5c9e
677758600b786b93b3214214f0d74eea1d2ef4ab
758 F20101208_AABOUJ bu_k_Page_033.txt
cb9a5487b3be77efe300e255af321096
2727f3299d1901fb6f9a5e4b4be56dacdf7e7247
1075 F20101208_AABOTV bu_k_Page_017.txt
17c30360c0da701a106aa29f72e9ac4d
31d9e99d50e967c1b993c47d170a2277d7678342
26436 F20101208_AABPAE bu_k_Page_038.QC.jpg
8d74111b71d99ad18427d73c41fcaf22
42d4045bc10e67b227f0945a08effd294c9bf79e
321 F20101208_AABOUK bu_k_Page_036.txt
412d7f1ff05490cb3afb58b24600e404
45db99731db080e79e05bbd6dbf9a12bcfcf04ac
297 F20101208_AABOTW bu_k_Page_018.txt
9d051eeebce6dba9993e1331cf29fb01
516dd4a060629513ffbce8e2efad567c777bdb0e
7177 F20101208_AABPAF bu_k_Page_038thm.jpg
a8b0930d2c8f1de81f925fbdb33db212
aec987340d13dd4b6d92e78b273683e0979a1a97
F20101208_AABOUL bu_k_Page_039.txt
a288044fd3ce5b0150da744e1249e942
495656a01e9a439e2504127cd0615cc43ff5b6ec
281 F20101208_AABOTX bu_k_Page_019.txt
6a4e58704cb333a623aa828f1e11e1c6
7a2ede58b9495d798d4ee77db6413272e737e70d
24124 F20101208_AABPAG bu_k_Page_039.QC.jpg
c06719baddcb2c571af6dbfd61e0bf49
b608ce8a857b489939a810669f3f37c65870fe69
734 F20101208_AABOVA bu_k_Page_055.txt
93edbbedbb8307d9fbd757154ccf1c5e
bf0b24190b10ccb7ecd365a38782f0d7b3c78ab0
1930 F20101208_AABOUM bu_k_Page_040.txt
d4bfe9459e90f7477c086acdf168b7ee
70464ecc1b33d502dbad04e0b7125f810ccfc8e4
292 F20101208_AABOTY bu_k_Page_020.txt
761e170a395777936dd9c9e805d54886
b6ef8989399d0e5f1d969312e36a1a5f1a699c6a
6559 F20101208_AABPAH bu_k_Page_039thm.jpg
29cf62a944bb1a4a176520bf91ebfdc1
7f8cd2f6e8e269dfbcd0f936f29a0b4bccf557ed
405 F20101208_AABOVB bu_k_Page_057.txt
345d44ff11ddb94eb0dd3276ee861eaa
a6ea2d9a8a93a949a699d9f53e10572fff73695c
1865 F20101208_AABOUN bu_k_Page_041.txt
921cd702e7435f3d89760f283ea2ff5d
9a1b821d60e9680c6b2050a789a372feb1cfbb91
2194 F20101208_AABOTZ bu_k_Page_021.txt
3bcb87f073e9f7d10cf05bf11ded91ab
cd945c51b0471f4e56b599df9011ea3c59b52871
22013 F20101208_AABPAI bu_k_Page_040.QC.jpg
a9ca7bc816b79df00b30a0cf3a692a77
7d47265b9c8be373e6aa097d27a9f9edeb17fd9d
1498 F20101208_AABOVC bu_k_Page_058.txt
5b939f3cb0f99d4bfe7481ce3b7045a5
9687c0d1ac8147a52bc6d48376272782c8589cc9
2213 F20101208_AABOUO bu_k_Page_042.txt
496cdfc60b9c78138313e0a0773e1263
7379322ebb9c026c52d8196e0b598abcc1bbbbf6
6137 F20101208_AABPAJ bu_k_Page_040thm.jpg
fb25949464a7c6395c518841d016cb6d
aacc40845dcd9b613511fd879cbd2664eaf9cbea
1182 F20101208_AABOVD bu_k_Page_059.txt
77f4c9158aad81e19ee1f6e50af2700a
63ba58158757e691524cd6e674936b503640355c
2000 F20101208_AABOUP bu_k_Page_044.txt
cb5929e051f0dc106e5e537a404d9491
f3f9932c4745e0f2ee9f6c5faa68fcb8f87c5d8e
21746 F20101208_AABPAK bu_k_Page_041.QC.jpg
d002d15c3fa572cbfe4caaeb3210da1e
cc82cefe5ef15be0daf7c9e78b46d7297b06a3d4
710 F20101208_AABOVE bu_k_Page_060.txt
26036761f1201aae9fa91fe83356e6ee
26b009898fbc346b62bbfe19490bd87e709e6542
2177 F20101208_AABOUQ bu_k_Page_045.txt
90f8935fff9066e264e30d7b7713cf68
0709d29fad76d27cdf1c8709d67a32919253c077
6142 F20101208_AABPAL bu_k_Page_041thm.jpg
3d2a683c1a0b0f99a989da49c41fdff0
69a714fd2946aabcd158ffaf47d218b7c3e404d9
360 F20101208_AABOVF bu_k_Page_061.txt
f720bc46774061c82652ceaf1d0fd3d7
5e3e8f05b74a8a289d13803314d22400096f1a39
2239 F20101208_AABOUR bu_k_Page_046.txt
33118e3edd6654c8a6333064ead481ec
2a73402baa7be8f7419fff1c004a2d17dca2f108
6744 F20101208_AABPBA bu_k_Page_051thm.jpg
c08f118d308228928ea61ba8f644ca15
3295b83305296ea7aa0a547cc6db0ff5e3b95ddc
26255 F20101208_AABPAM bu_k_Page_042.QC.jpg
bdb49affd16c2c23a79705e55b76a914
c1d6c9fa3d256502e06213eb7f668c2a1324d719
1148 F20101208_AABOVG bu_k_Page_062.txt
aad9aef5d2e494366f01ed18108e0ae6
d65d3497196cc77bc10db3fa2d7abd6411d39e1c
2134 F20101208_AABOUS bu_k_Page_047.txt
ac5f288f7ccfefb7fb74eed6bef67e6c
09eab683bc69be44d4d03f9ecc745255fbd3cb5f
25067 F20101208_AABPBB bu_k_Page_052.QC.jpg
50c903536dd50ba6d903e8e1e6d1242c
874e7cdf4fc657493a291e6d880a5f900c5e0365
6913 F20101208_AABPAN bu_k_Page_042thm.jpg
6102e8d416096612379c2b53ec543d5b
833a2e1cb661585ac011ec2e55c1a853e21059d0
447 F20101208_AABOVH bu_k_Page_063.txt
e794e23bc2ce448ae481874222d30d53
fd34f87119cea995f73cc68470f726ae95b45430
2249 F20101208_AABOUT bu_k_Page_048.txt
11106f6e9b4f91fd3bee01a08104722b
2359f52db96df7fc014e7478b14ab995c23eb074
21252 F20101208_AABPBC bu_k_Page_053.QC.jpg
8013f795fb55a7b518df666cf040feb8
0297a14e7d020221676cf3f970ed58666d0fce86
23655 F20101208_AABPAO bu_k_Page_043.QC.jpg
1bba864c26a2a25f97c21ddd30a5f39a
d3c0568744127883d51a60fcb22f1460dc368edf
934 F20101208_AABOVI bu_k_Page_065.txt
cd5bb66797fcfa03fbc3113df2c37ead
fc665b784e7c2ee1ae0afb09f3d303f94f3c6a10
5823 F20101208_AABPBD bu_k_Page_053thm.jpg
f0aa954fdf2c5baf071a92daa5a9646e
a62c4e4510f28580701d5f58a3ff45654002f16b
23175 F20101208_AABPAP bu_k_Page_044.QC.jpg
8d7e74078044a00c50dc6212ec3f43b4
6b708c555337fb8cdbf76f3daf77f7181a0727c0
637 F20101208_AABOVJ bu_k_Page_066.txt
d8af8024c035022afbbe80d99139f54e
f5b60ffb117e19a9eaa0eff8cc81dbf29c48319e
2120 F20101208_AABOUU bu_k_Page_049.txt
3593587e39f15b2944a953ed487da407
cb835dac81e8c07e760cb220c6863c67ed663fd5
6922 F20101208_AABPBE bu_k_Page_054.QC.jpg
9d9a353d019ef9fd5de696354b24793b
dda46eac4f03f4d1d8d8c0c960d44c014a9bd962
6541 F20101208_AABPAQ bu_k_Page_044thm.jpg
24c1c5f9029d82d405b4b8b1f8250724
d69827326618646b1d21faa0548e1a0118a5a5ef
395 F20101208_AABOVK bu_k_Page_067.txt
1aaf6fb89549a694d991916b42a0f656
e888d2b3198b7022ab1853c5d52517e41be6e5a9
2019 F20101208_AABOUV bu_k_Page_050.txt
54c1ec602465edd658585fc1dd7dc7b5
5768162e08164c26eed33422400c1925ee2a8dc2
2204 F20101208_AABPBF bu_k_Page_054thm.jpg
84a81b66b3c4bb53ae23a04be7d12ae9
0c3bb814d508b41964dd23c99bce7114eba5c813
25202 F20101208_AABPAR bu_k_Page_045.QC.jpg
3f63e66f59075071371b70d21ba23c0a
810e8ad135a3e04b1dd6613729c6f98950e9fee9
529 F20101208_AABOVL bu_k_Page_068.txt
4b694d902b2e881eacb6908c9a0ca4af
c0b4277fd329f269cfcb1301843f1ba80bf5c69f
2063 F20101208_AABOUW bu_k_Page_051.txt
87d4b946784a1ce823c10f198fa6f22f
4a49cb8843c15fd3eb4cf1d0b134e2b307260f41
9848 F20101208_AABPBG bu_k_Page_055.QC.jpg
2e9e53b34cbea8dddd5ecb46920084a6
3abc6b96503e1d2df455f11a7b03616f71dd589b
F20101208_AABOWA bu_k_Page_086.txt
c8561bb0e3a0d18c1e71b79dea54d9aa
9cc17dacf8d4d77a35f9f07f7491103384966d70
26208 F20101208_AABPAS bu_k_Page_046.QC.jpg
bebfaf55e409a4534dd079accebd564c
69b1e2d70b7808f35b2403bd654946de178f933d
302 F20101208_AABOVM bu_k_Page_069.txt
092c6d0d427d49629e1e689d75efc6a2
0eb883788f5843da6fa8f70b504c72ea99660780
F20101208_AABOUX bu_k_Page_052.txt
2d549203e2f0761a69d7427e6c93f20a
e8280adbf4c08788be487b3b30226db4f117b526
3203 F20101208_AABPBH bu_k_Page_055thm.jpg
a9360c6f2d9d69e4fd1a9d51f79922d4
10a03d0e3f9622cd05f93e4145636bde2ef685fa
2162 F20101208_AABOWB bu_k_Page_087.txt
e017292223703db0ebb7fb41099841b3
6d2c3f20749f0a7bf6ec61d321ca963cd29b431f
7030 F20101208_AABPAT bu_k_Page_046thm.jpg
ad075ab43804a027c2b9ac1c75fb9401
0d921a87f989347e92fcdad9980d442a9f978bff
753 F20101208_AABOVN bu_k_Page_071.txt
f9c6ce7c344e1df3be091c1901393557
1afff9b12026b3cecef978943f0c339851b448de
1738 F20101208_AABOUY bu_k_Page_053.txt
00133803d099152e75e261c82ab63093
a0c1a2f0fbbc082b27d8e0a87e185fff667a1e55
12444 F20101208_AABPBI bu_k_Page_056.QC.jpg
a59ea7928eda7d5173e7144b9e9de8f5
d15584b8142d6552ce2e881251717a15f596d41c
2037 F20101208_AABOWC bu_k_Page_088.txt
1d17a7a62e1c3ea9c3993094c9dc6dd0
21b14e8afed98e2ff8711acde9030d5acb0260a9
26492 F20101208_AABPAU bu_k_Page_048.QC.jpg
5a21a234104b248be75bea17c2b1199a
d3902ab9179b58b68581a09911e7144bb7043d78
F20101208_AABOVO bu_k_Page_072.txt
2f5e772c87b5b274762f4a9c5722d054
db9e08cfb6d1b50dc5688f4121e7f8652a110f9e
503 F20101208_AABOUZ bu_k_Page_054.txt
b394bb2130e8e7d5b247d90fe8e1ae92
cc9f24a9a8b1247512dc291afe308a10a070b917
4095 F20101208_AABPBJ bu_k_Page_056thm.jpg
713688bb876ad582ac1da3b980a5691c
23528aa268ef37f9414a7cc5890b626b5d49f5eb
2221 F20101208_AABOWD bu_k_Page_089.txt
46ecd128aad8f1f333cd9c3a7926e016
176eb5f07995c0fa3eb6338591265e5887c520c3
7058 F20101208_AABPAV bu_k_Page_048thm.jpg
7f7f13123afdab0ba76e114bd5be92fe
08afcc297a207e775f1dc83b721a33b3a401b4ca
2010 F20101208_AABOVP bu_k_Page_073.txt
4579713b835fec1d561e07f3f8ab7c97
ddfc003c103ecc8ce787d455556f058ec8307871
8427 F20101208_AABPBK bu_k_Page_057.QC.jpg
7ccb04fe4d27d5b513e001b49d3b37e8
eda72e355668286f0cacce8e6a4eefdb94d691c4
1993 F20101208_AABOWE bu_k_Page_090.txt
d653dd1e93905be399c36528f104cc2c
7146badc796e6f88a6087c48cce09dfcbc030d52
6786 F20101208_AABPAW bu_k_Page_049thm.jpg
75b4522e15744623352024bdea6b3908
bc6adc38f758e623a5ad54075b18d292249b6d2b
1982 F20101208_AABOVQ bu_k_Page_074.txt
462aad43693b6978643cf649f9008756
5b94574999930974b17ea1009335905174b1d20d
13239 F20101208_AABPCA bu_k_Page_066.QC.jpg
7612d456458b3fbab8d4584cfd9298d4
8257148d32bbe21611c5ba848f82836fb6755220
2912 F20101208_AABPBL bu_k_Page_057thm.jpg
1ec0f96b309e7d0807773b5454f42280
8af341b58f9b6f560020fe7810580b3c0313aace
2098 F20101208_AABOWF bu_k_Page_091.txt
271d5210594768f6c2836a7df4ca6083
12f8271581ec9f2d742e7f64431c30c3b49d27f1
23405 F20101208_AABPAX bu_k_Page_050.QC.jpg
97229274c7595345806c928f9454e85d
8c0ed5a5814e18c24256af70dd38a8cb0abc52b5
2111 F20101208_AABOVR bu_k_Page_075.txt
98f05bc5220d87790fa0e53143daed1a
99a1dfba564708743fa1e10078778bee449be569
10404 F20101208_AABPBM bu_k_Page_058.QC.jpg
8c00e361a3be42b7196d41bdcaede9dc
7c84aa51fced360e277d7605ba5278af8e8bf39a
6219 F20101208_AABPAY bu_k_Page_050thm.jpg
39403d731dbb0eb9143c7a883705eb74
3ffc40e00cb8bcb9a786c8d695aa57bf9745729e
843 F20101208_AABOWG bu_k_Page_092.txt
4e1de269f4f40035a777a59d7a5c8fb1
6ecad6cf1b71dd315ffb2a20755a224793422aa2
2093 F20101208_AABOVS bu_k_Page_076.txt
7687c43bbc53f7fbb13da65e2640931c
05f66562d9a71b10d1106e4c901c762cd237fd0c
4781 F20101208_AABPCB bu_k_Page_066thm.jpg
f160242af617afc5883944ca70a60ad8
ad07f82aeb395bb16d83182f9dbad57ec03f488e
12728 F20101208_AABPBN bu_k_Page_059.QC.jpg
e66744b269b58dd446a82b94bdeac46a
7433235612f7e1e642b425fb9a237a0bd60c13e7
23858 F20101208_AABPAZ bu_k_Page_051.QC.jpg
aebb7647c63311497d48532ab22f797c
a63e941eaa643c3493376b876d42652c58b93b45
650 F20101208_AABOWH bu_k_Page_093.txt
bd62fbe3e7250cdb1cb4748aaeb2d1d8
cc67e18fe37d474828695def87287afec1c53802
2143 F20101208_AABOVT bu_k_Page_077.txt
6cf6406176582f0b570d3b1cd78de667
d9a9187deef958aad91ca5956110ad4a73f2c367
8969 F20101208_AABPCC bu_k_Page_067.QC.jpg
44d835c747204b537c935332582c3ac3
10f60b39c6dc4035c2ea4a2daebf30d9b601587d
4147 F20101208_AABPBO bu_k_Page_059thm.jpg
864d3dad0a843d795637aced78fbfe79
2f5be9651afe633e3f69ad24d85fa1f44e2982e8
531 F20101208_AABOWI bu_k_Page_094.txt
0a9cf30b7261db8b726c6648dc2a81b8
5e98cc4319001f2169e5613a4dec22b0c7b96a59
1668 F20101208_AABOVU bu_k_Page_078.txt
cfe54901c7386b61cd85b8a69a3de253
535a89693d72c499b1f06d66aa56f587d1907c40
3756 F20101208_AABPCD bu_k_Page_067thm.jpg
3cdb547488f6f2eae0811e0e7413f227
340c4ff728cff88ead8aa1cf963410ae114df149
3481 F20101208_AABPBP bu_k_Page_060thm.jpg
28a4143642547526a07c62c82789908f
b487b81c15ccfe113eacaff70a5f1a5f13a72b83
801 F20101208_AABOWJ bu_k_Page_095.txt
87bfe795cffa2c56eb16b27491586f00
a27c186af8cd289dd58843380555b882eb9297c0
11346 F20101208_AABPCE bu_k_Page_068.QC.jpg
78de6985f0d1ee8fc1bf4bc3c1b3daf4
28cbd01d81e311be234560fd4bac0bafa895d5f1
8371 F20101208_AABPBQ bu_k_Page_061.QC.jpg
b11c2e3c543b69e5ee34aa7e520118ba
5d296f35d7d99af708e28785c29dfa6f97c13f3b
456 F20101208_AABOWK bu_k_Page_096.txt
21c3e11155ceefa7156084e843b481cb
96b4b3598a76c79aa0d3d6f05a092378981d2288
774 F20101208_AABOVV bu_k_Page_079.txt
804406de317e239900b16e145801d415
42ceae504277361765eb78eef2500796f4a01e8d
3882 F20101208_AABPCF bu_k_Page_068thm.jpg
bbc811a0f550a0d2726980edabcf16f1
b9823ab20bf9a53529214e8fcb1fa88eac8b4cc1
2883 F20101208_AABPBR bu_k_Page_061thm.jpg
df287ee9a2bf9d78b618d39af90eea67
e6cd88304697574ba16b34d279972f2effcf59d0
935 F20101208_AABOWL bu_k_Page_097.txt
47fc8e0bdfd7736ac20b62874b2a40cc
cfeac001e135ab01f358a677966a05cc3356e82e
1675 F20101208_AABOVW bu_k_Page_081.txt
de7ef69e4dff71b77247692e61935321
20ae3f8fd3ab33661824890f7fb470b430d0eed6
14029 F20101208_AABPCG bu_k_Page_069.QC.jpg
cca4af4ecac1092248d52eb0b71736c6
99af5c5c01faa57bee0ac403dcd58f3ef1bede7f
13362 F20101208_AABPBS bu_k_Page_062.QC.jpg
b94680a3d28d904269952ed95de4b13b
2f25901573220450b8e4983ce4181937877b4148
283 F20101208_AABOWM bu_k_Page_099.txt
534130dfde683c37178ec56987559de4
18beb775ac046cd512b415883adcdbbe6662398b
970 F20101208_AABOVX bu_k_Page_082.txt
4eefa8d1b0d78bc9a076eb00bec5609e
91df0d4902769da5fafb1175dc0e8631081e51d1
404 F20101208_AABOXA bu_k_Page_113.txt
786976513f07da9ee8ae557a3eaa835d
283fd8ef871b71bcbd0182bbf081953489a223ab
4369 F20101208_AABPCH bu_k_Page_069thm.jpg
9c95b3e739d7e7c3654dd68fc881feb9
7fba5a6f06b87a3bd79cd28ee4b06fae27b8c1f1
4218 F20101208_AABPBT bu_k_Page_062thm.jpg
c3fb6e9024e5ec7164a2440bed2c6ecd
ad9a394cf49d9d74954e46814c2cc82b43ef01c1
2079 F20101208_AABOWN bu_k_Page_100.txt
d96ed345778ccd2908f76b434da394e5
3fc3e5a5515aba38291496edded6997b1c3ec632
928 F20101208_AABOVY bu_k_Page_084.txt
acec24985fc7c50a3041b2a72ccfbe97
76d4f20d732192925eccab978dc615f66ddcfcb1
273 F20101208_AABOXB bu_k_Page_115.txt
1df0a59fd4981d0fd15eb3c4a31e7aa3
08bc51e9f44b454c1dfa2b5aa85e137d9527c613
8520 F20101208_AABPCI bu_k_Page_070.QC.jpg
7535df60fbfd5d5aa993b4925f8c0c24
98de69a5cfb8a3ba56eb34ce7b9f4abc0542d319
8768 F20101208_AABPBU bu_k_Page_063.QC.jpg
6421de012a22c14f9d123074945e4ac4
ac06fb9e9931c14032aa6fac351478892c7866bc
F20101208_AABOWO bu_k_Page_101.txt
2ef5dada11bb7ebab862c56bd589ef7d
e9d9c8c389546d580cc69d7c3677ada71cf3b225
490 F20101208_AABOVZ bu_k_Page_085.txt
f06430126cdf9804ae0a651a7e172491
27c85116b6a4955ba87957695c0cd9324570b385
187 F20101208_AABOXC bu_k_Page_117.txt
4508b75edf3e236769918566879c681e
7c19ce8a3d58eab3c1167d52065af49ee30e644f
2971 F20101208_AABPCJ bu_k_Page_070thm.jpg
cb7c81c425e8c0c79b00fbff3e4de17c
7abc42eb864fa4e68f95decbe0348dc345e43bc4
2788 F20101208_AABPBV bu_k_Page_063thm.jpg
d62e97f84fd3c7909ac2f099ced990d9
64b5f5cbecedc6d689cd021491d29455ceff632b
2329 F20101208_AABOWP bu_k_Page_102.txt
d5b9d7fa6a0cc29503cd7441c29ad306
de54576897002b07d610ec7afe3e6b364c740746
287 F20101208_AABOXD bu_k_Page_118.txt
98b05b7a6ac753495cbd6512eb53f4d6
db575f11af0584dca9f708bf05c976b34917b512
9916 F20101208_AABPCK bu_k_Page_071.QC.jpg
e63f91426ecdc255844909b53b651259
35eec33a3bbc48d0dbc929f338d621d5aa2a1a6b
13880 F20101208_AABPBW bu_k_Page_064.QC.jpg
73b71b0348eb7502eba2f4b98a8024a3
5fc3dfec724b70f03b6964b0874869139641b2ec
F20101208_AABOWQ bu_k_Page_103.txt
6d95ec3b5a257f2b56ae39c03ce9189e
f89299f24cb02d9a618f76c6b6ef2edecef096f1
199 F20101208_AABOXE bu_k_Page_119.txt
924135595bfb50159662ca90cf81d437
7eaf4b593d34598bbeaa8053f3dfaf152f00dc0f
3871 F20101208_AABPDA bu_k_Page_081thm.jpg
96487c25ebd96ffc6f18b7006985c23d
21c72483b1ddd159228bf60d35236b0987c2a613
3213 F20101208_AABPCL bu_k_Page_071thm.jpg
5e86871a78428764b9fbcf745f7f5431
75de8bd6ea19aa793d8898daef44338ba35cd9f6
4582 F20101208_AABPBX bu_k_Page_064thm.jpg
934337ad352154077c0a0cea791c2e6d
300ae8fca68038adc65822e2fbc2879d7155ef80
2102 F20101208_AABOWR bu_k_Page_104.txt
4891a6c09e59c2882e14bed9950ed426
bfa278b1e1fe6eda68d3fa7458c5f58642a8cedd
2212 F20101208_AABOXF bu_k_Page_120.txt
7501fcdeea5138ca6ae247e91c6ea983
ff675b79da5866818785d7fcc019e6549ba9661e
11323 F20101208_AABPDB bu_k_Page_082.QC.jpg
cef3f212f43cedf89d187903ac040218
1278bfcbfece7e2bc9c58f2e525a7de025cedda9
24575 F20101208_AABPCM bu_k_Page_072.QC.jpg
ff23656198d92d7dd926d4796b3da2a4
f7233484ef9b089d6698faf7f7714fbcac21a405
11841 F20101208_AABPBY bu_k_Page_065.QC.jpg
7c676ad0fb26cf06e920fb5ed42ea75a
38bf2307187d20e5dc35471f805baa5448c5ccea
2060 F20101208_AABOWS bu_k_Page_105.txt
9fe59587bed784589de9fbcad933627a
fbb600c048281eccfe9efc6d7611dedafac45396
1960 F20101208_AABOXG bu_k_Page_121.txt
ffdc6e57df0b26336b26bd0954472be5
7aaadc43ee0ecbad8e33bf02eea579a541d312fb
6886 F20101208_AABPCN bu_k_Page_072thm.jpg
1d853af1db49c9a7aa73329a56faf416
45cd94da4603ff2040490b3ff0500c67f1808454
3574 F20101208_AABPBZ bu_k_Page_065thm.jpg
161cdba942339286b288bfad2c1692ee
b746e4b655ee5d3522683f9e923f6551ba31fdbd
2123 F20101208_AABOWT bu_k_Page_106.txt
b9ccad6427eb2ff5b7249c7c1b5ab255
fe183f2b61e4182bdbe9c2fdbdbbbff5f54d0592
F20101208_AABOXH bu_k_Page_122.txt
7ee43d5ef48593b5df9cf97880180bea
2f9a189df5c7e8b72d1b954e71b588527b4890ef
3820 F20101208_AABPDC bu_k_Page_082thm.jpg
2a7d0b228c7038dbcca430ed27c60d93
19c78eeb3993deb697d32045fe285234ec331715
22817 F20101208_AABPCO bu_k_Page_073.QC.jpg
25b48f5240b673d12b37e36ebab535db
0cedf8f2e73ea8da327938b79a7e7080fed04fde
2201 F20101208_AABOWU bu_k_Page_107.txt
2a3be7f6ba0ba86ac92fc76066fd9ad5
43d2ae949bb75c3ed0dafc0146b2105f331cc352
76044 F20101208_AABOAA bu_k_Page_101.jpg
8facaf71f51432722990ac9a33ce1f04
e09d424016ee4cded7e50115d4d1a0e1dbb420eb
633 F20101208_AABOXI bu_k_Page_124.txt
b57886ef1922a0e461a4ae1a7e433670
ac7f74f0c69c93e248092c2658358e3c5eed17cb
6610 F20101208_AABPDD bu_k_Page_083.QC.jpg
b38e64c3b60e612ff350d9d0ca097d02
195d50920c4e8d0b348c43cc070bbf8f930e93fe
23301 F20101208_AABPCP bu_k_Page_074.QC.jpg
a09acf9b640b973b304c631b21112a2f
78cffa748073661c3a1f476f07974c698733fc30
2229 F20101208_AABOWV bu_k_Page_108.txt
1d8d57a9081163f5c4d6922512ae1f54
12f8ab1514fe8d7513d88e120d17d272ef15d4bc
6721 F20101208_AABOAB bu_k_Page_114.QC.jpg
6c2050573ea50696f783d1c130d3f2b4
a6d42952823af4a3f3d43c7f13c31d8e7d5a99be
F20101208_AABOXJ bu_k_Page_125.txt
35a94dbcf679208022d62c2df998a51c
2e9c7bca546f371575935c9e5600c785db1040fc
2473 F20101208_AABPDE bu_k_Page_083thm.jpg
7a65d5f4976a91fcb7e55f089b9c633f
b8c00c331824306a6ed468db7609160fccce0e34
6465 F20101208_AABPCQ bu_k_Page_074thm.jpg
726ea5105018af4f2bf24e211252e76d
19bb7692ee0f54984cad4ec3421360044e6e3b33
375660 F20101208_AABOAC bu_k_Page_070.jp2
bca2bdbda9fc187485abb0301f12b3ce
b7a0c0fc3f765e402462c40d3dae0452faa66c44
2431 F20101208_AABOXK bu_k_Page_126.txt
cf76670fdd792b93ee38a39a1247344a
b4626a60103342ce79cb1dcd2cc9b598e5608ba7
9315 F20101208_AABPDF bu_k_Page_084.QC.jpg
4fabd8e8f41ca74768f51f45d74a0a34
d70a10bf964bd4a8a7aa06d407c6e7b1505dad2a
25831 F20101208_AABPCR bu_k_Page_075.QC.jpg
51bde930d4455017c8b8bc9f4de7b5f4
2e0a6afa1146964bd0e9111ac4269f6825eba0fd
2171 F20101208_AABOWW bu_k_Page_109.txt
2185a93c9975bba9c88c060d032b49d5
31a66fd23606fd91cead2ebd0b50b54a39a8dd90
49394 F20101208_AABOAD bu_k_Page_121.pro
6ea3417473e1b00a0a57140825177942
382b789cf15776099e57a1da532c1d2f18405525
2566 F20101208_AABOXL bu_k_Page_127.txt
5ee1ea8eb5d04e1ec3616f3d37b751dc
881bd39adea437b267df754e01afb6b80fb4ae89
10001 F20101208_AABPDG bu_k_Page_085.QC.jpg
7feceab938e425c211b143ca90c2109f
e5da610b82fd28dd638c8ece6a95da5691c0dc11
6893 F20101208_AABPCS bu_k_Page_075thm.jpg
68d4a42e23468d9d9e0fb3970d2747e5
242defa50bc54e732dc62cdeb3d246873c80cb03
1688 F20101208_AABOWX bu_k_Page_110.txt
825d18670eb5c948667b28bb52345952
5a19bf30fe95e3096ab8760edac85b92abc4199b
2115 F20101208_AABOAE bu_k_Page_037.txt
7da1bb8e618c26350abea8ae8b4b250e
d157c72e85855371bf94ae7f23dbbdd7e312d82f
6159 F20101208_AABOYA bu_k_Page_006thm.jpg
45a76e928ff9fd2fd263a14eb7c65695
efcf816558524e7f5f9db3e47f7eabca6d402c2f
2316 F20101208_AABOXM bu_k_Page_128.txt
f7f2ada1e227a53735361128c9614410
73ec7576ab62f37c6347f4c5729d041766475150
2972 F20101208_AABPDH bu_k_Page_085thm.jpg
a35ab801d0438fdf9ffbe1666001c512
e0c8dfce0ad21906ae78ad39e3535ea44e8e0ed1
6927 F20101208_AABPCT bu_k_Page_076thm.jpg
224d738f309f7c5f2abdfcf768614d19
7a33628c19537bc3f744bafa703149d191a265b7
1029 F20101208_AABOWY bu_k_Page_111.txt
ad915411cee02cc2858e660e486c77ee
23a7d76d67064bcebc6da313aeba791a14a3c033
113887 F20101208_AABOAF bu_k_Page_072.jp2
8a9ab73f3bbc83291e4f6b8a83e83d42
914acc22258fc7cae007f07da78b5fe1146b3468
11744 F20101208_AABOYB bu_k_Page_007.QC.jpg
f9d44f3b80c1b37aa01158a42d086f23
fd94eefed6e6a211910f615fdd204c70a353a99c
2562 F20101208_AABOXN bu_k_Page_129.txt
6d03380e06767001d2c01196a996a4dd
2d77c45d59c8f852a0e633763ba8fb594341bc21
25769 F20101208_AABPDI bu_k_Page_086.QC.jpg
1b5e78af29da06e2a1ba9a5a8a42a2fd
51c128959d6fe18d8af7910694111f9b84e1cb82
25887 F20101208_AABPCU bu_k_Page_077.QC.jpg
7a15bc50a590d23b529db2b0bbadaaac
038c3ea87d1807050689e6837df8886a56ed15e1
735 F20101208_AABOWZ bu_k_Page_112.txt
b96272b2cac864885d1e92c19a2ab493
7896ed287f33593137573574a70e50e86fd2da7b
1987 F20101208_AABOAG bu_k_Page_013.txt
aa80e0d1f0c112ac508b1aa3c516ae1c
2afc97aa12ad824e4a7587b23e5afd33e944d0a6
3199 F20101208_AABOYC bu_k_Page_007thm.jpg
6a703fc3ae97fe0f310e826a5ec64d27
367ea8dcf3b08f2a3c4363fa88ae7e97e7aa02f8
2532 F20101208_AABOXO bu_k_Page_130.txt
0702a954e5747fb870d9b15fd286c1e1
1420df1c0626a23ddb911d128b07ba0e68dd0933
6924 F20101208_AABPDJ bu_k_Page_086thm.jpg
59b14a0b6776f5e278c6a1b788640e33
7c8dba3d501f5e2e14393cbc5db39d26ab643540
5506 F20101208_AABPCV bu_k_Page_078thm.jpg
d5760710a4a96327cd4920b0ca897e6b
e9db544ad56b6f9958a4fe65fccb2afc01956db6
22137 F20101208_AABOAH bu_k_Page_017.pro
985d58d40c4568517925efd869ea07d8
f5e24550ece7cacaa265126bab8b8f19367b19cd
11219 F20101208_AABOYD bu_k_Page_008.QC.jpg
bbe6417f125ba508925113cc984f1a37
19e9dae955748abbce9ec4ea35226e03fa345213
2560 F20101208_AABOXP bu_k_Page_131.txt
7ae8934235668cf7a42db5bb99575c86
94c1bfca309149c9628837eaf0d44ec4b2012816
25274 F20101208_AABPDK bu_k_Page_087.QC.jpg
9fe60b02eb586afa6b1c834e538337f5
88e12707041c161bc7b2dd19dbb7d4842a649ad4
12967 F20101208_AABPCW bu_k_Page_079.QC.jpg
4d61fd25b6c9ad6e56618c25221eba02
5607d78989b12ae105f682c1131343cbdfbba9a3
2277 F20101208_AABOAI bu_k_Page_132.txt
685e0b1871f7cecf07cfe036d72bf78e
a6603021c5606ebce987ea8f0df93e853717cf4d
3256 F20101208_AABOYE bu_k_Page_008thm.jpg
959864919631e6c8a1f12562c21812b2
e1e2b20a44306dc1d6efba2cf8d949f899a60021
1370 F20101208_AABOXQ bu_k_Page_133.txt
ca982e1ad81fcc58442d84f6be03b636
7ee0639110d39be5a1e7a9aed0ed66f377636901
3367 F20101208_AABPEA bu_k_Page_095thm.jpg
9bdcdafc54cb2eb82be9d67a121f04a1
fe0dc3d73b8746c4864158fc0e0c919d66d6d616
6941 F20101208_AABPDL bu_k_Page_087thm.jpg
bcb37dc5e9b0cf447122fef3b5cfb24c
402984e2eda22873dd27f6f22dd4d53b881785e5
7665 F20101208_AABPCX bu_k_Page_080.QC.jpg
5a7ba3b61db2e8d074b970698adf166e
a2ded1d9082654d024b9a1fd108f24a4d1a3a2da
F20101208_AABOAJ bu_k_Page_090.tif
c5d42d6e4bf392c9a3c8614adc6ae86e
c2ef750733f1a2a98826df7b21a0976f0f003839
27162 F20101208_AABOYF bu_k_Page_009.QC.jpg
7fd9bb91c9dd0e0fbe3e50e4e53ad0c7
1eb51bd3938de4a0880e9f80628c878e34b1a18a
7383 F20101208_AABOXR bu_k_Page_001.QC.jpg
1caaba43b5bca3e4525cac5018d77e97
616964c12ede146ae1f61670ababf41ae84b03e6
11654 F20101208_AABPEB bu_k_Page_096.QC.jpg
11172bc78603e65f2214a06199bbd4da
4bc0acbc3e739196eadb6a029f6db3a153bfc535
7105 F20101208_AABPDM bu_k_Page_088thm.jpg
b107b7d252bd14539094f1046b85876a
048b96165d59c21beb0c10e304e63938e21331e0
2974 F20101208_AABPCY bu_k_Page_080thm.jpg
277fc4b69dd77f265537b93cb1d21808
144d8d24e34a810accff05a6db4f813832e3bc86
F20101208_AABOAK bu_k_Page_112.tif
2bb2d9030cc4b719a853a7000164e764
59b35bf559ec8753cbd2e3941f1cf29c9019071a
6875 F20101208_AABOYG bu_k_Page_009thm.jpg
f1122417b6be153f848bcd9dfa8af017
34909cdb6421bd5062fc176c5327de254abc383d
3161 F20101208_AABOXS bu_k_Page_002.QC.jpg
414adb957ed2e3db761381219abd9e48
227c31567273d22e6b8d64863b5e737dcf2e0bb7
4080 F20101208_AABPEC bu_k_Page_096thm.jpg
5e8921043b639928a6637625a9469157
db581e8ed4f00b279e9824f4ddbfa2fd1ae9a8b9
25922 F20101208_AABPDN bu_k_Page_089.QC.jpg
f0baaf55aaff3851a3acef25a87b562e
f4b2c299a43926ae856f3352449b79ab3f1579a1
11833 F20101208_AABPCZ bu_k_Page_081.QC.jpg
48faea38961adc9d7296da7a21c2dd40
aeabf845e9588a90aa1aa3c3250c870a65d2c147
980 F20101208_AABOAL bu_k_Page_064.txt
a7c881cd2fb249f70944196d67c97f1d
f27b3ab09577ecc244ac144a14ac06efd194f1af
27641 F20101208_AABOYH bu_k_Page_010.QC.jpg
f2f14efe3e06a0c9f52c7c8550de6259
c7ce1ece57e4aa8fd1902fb31d69b7371be297a5
1338 F20101208_AABOXT bu_k_Page_002thm.jpg
8e63adfac6d027e42ff1a21570ad9fe5
736f1494d4debd11c8f56b21fe0f1300cf7d3844
6979 F20101208_AABOBA bu_k_Page_077thm.jpg
9976166c0d5763f0afbbd5b60d27809d
0be78a35d933882c063be8448bfbede56c34bac6
7133 F20101208_AABPDO bu_k_Page_089thm.jpg
cc8f2c399f18a19f35c6aaa271b11e04
c79c0227424a62c0d2edfbd08c5c29cc54fc223b
3886 F20101208_AABOAM bu_k_Page_119.pro
000d8522aafeb2a389d0fc1d3bce07f4
9324ecd1082c96d6981a47dd89b1f6c66252a615
7147 F20101208_AABOYI bu_k_Page_010thm.jpg
8af676f87224a5e71f35ac295494082d
3398d16933d2dd80c6f56264bb1b027a4149c120
3387 F20101208_AABOXU bu_k_Page_003.QC.jpg
d642d4ee911070ac0a6caf0fd78a6efd
1c23aa7aadac25f0a00f40390189788723afe48b
4042 F20101208_AABPED bu_k_Page_097thm.jpg
da4e8d6656aa92c7f6b3ce1f02937a45
304b7784b94360ec6177f3b6be7807a7984397a4
23351 F20101208_AABPDP bu_k_Page_090.QC.jpg
65d738aa79799d1fc0782cac5c66a444
cd7b58ac85edf9ab213e3d085de2697576b1bccb
26588 F20101208_AABOAN bu_k_Page_088.QC.jpg
7c92438df777289a844ef132a70a7bb4
2684031596ccd72f61909eee2257772becddccf4
28830 F20101208_AABOYJ bu_k_Page_011.QC.jpg
aa20216c61fe07f2e3c8c1c782b0413c
5f87d6f7d027526914607ac787805589739493ea
1426 F20101208_AABOXV bu_k_Page_003thm.jpg
734844c97235798816215ec8c4bd741f
7dffc4bf65ef8cdba5a6873ae00ff475f653b86f
26789 F20101208_AABOBB bu_k_Page_108.QC.jpg
75142d98d3884e8a15047cc14048be8b
c2f7ec641ef5fcc5876d49cdcd54ccd9c547345b
12308 F20101208_AABPEE bu_k_Page_098.QC.jpg
d34019da58d49956a0a826585860d9ac
5dd852cc6709c199433e7bd37142734acabec2fb
6571 F20101208_AABPDQ bu_k_Page_090thm.jpg
a461c4f98a832cf0e174ef54bcb1e0d1
f6971f5914e6813081989f8e9d3b3748b8af82aa
7564 F20101208_AABOYK bu_k_Page_011thm.jpg
47238539839a391d2c2e28fdd9c7a1c5
96be4b06b354855930572afd47a36dba1fac5f74
6639 F20101208_AABOXW bu_k_Page_004thm.jpg
3069c6814176b68031eafee7bf9bbf7c
65692307e9d4aa3eac6e0d5767aba047fe9a91d9
F20101208_AABOBC bu_k_Page_049.tif
51d6876007d6a1b219ee08966f6c71b2
afae434b83886a0d4483bb4380f32d796de4b2f3
F20101208_AABOAO bu_k_Page_035.tif
38673eb5da1b47c4b9d37726e5d3f778
6d7f1f78ff1ac24afbc4bb1f73ef08fcfc260f30
3844 F20101208_AABPEF bu_k_Page_098thm.jpg
3ed05ef26bf0af732587f90c68e55094
16c32de1350d0d13b151f8cd5ad0899b26db6894
24986 F20101208_AABPDR bu_k_Page_091.QC.jpg
55aba2b4d65a78c29bbc199bbeec9b61
6e5547afbb62e38ded64470a5984fca89e2220e7
4423 F20101208_AABOYL bu_k_Page_012.QC.jpg
5d3ddda00ebeacccac50a9bcbca97acb
18f17cc10e6b1d0ec3df08039dadabf8f0dffafc
24574 F20101208_AABOBD bu_k_Page_076.QC.jpg
783b7233ff11590a2372d5293fc6f069
da1ff36eb71837c2330d56e584560330c7968110
6654 F20101208_AABOAP bu_k_Page_043thm.jpg
aef4a378ff5422f6cc190733e889b22b
4f0146a4342b617736aabd348f8796319e3a035d
10092 F20101208_AABPEG bu_k_Page_099.QC.jpg
f42ba78fdb187d274bcb8e20eba48488
c9faef55e9e02893dbbc6cfda567ac45d455dee3
6661 F20101208_AABPDS bu_k_Page_091thm.jpg
b088e50676f910b4a1ff20f84a45dfd5
904dd82db1fbc34a4fa377c7f188a1dec674aed6
1588 F20101208_AABOYM bu_k_Page_012thm.jpg
e14acf8160f84e2010bb1c59588ccfc0
106885cd745cfe72c05b4fa92655ae04f15fa001
9245 F20101208_AABOXX bu_k_Page_005.QC.jpg
5f8eb391ac358c1c36e28cb614f95c65
dfc26dfe6d2b6df2cc0506bc1ab435f084b6c8b9
1051965 F20101208_AABOBE bu_k_Page_069.jp2
dd847e5cfa6ffc83bfaf869d038eb383
0de3e1cc0fde979e2e2ce2aea99e6179538e4a21
26940 F20101208_AABOAQ bu_k_Page_008.pro
d4d03fd3e2ca79baf1848a78b04811d5
0202c1d0c92b3e0157509dd0a0d827c59098c626
24010 F20101208_AABOZA bu_k_Page_021.QC.jpg
0706cfdb863f4b232cdb68ee8424d7ee
e6f84575cb6251ed975268174faa832eae804104
F20101208_AABPEH bu_k_Page_099thm.jpg
cad632c9291dda9ae2ba8ee14549f741
5a9bec2b813aef647761f8023e15279acc0ebd70
11181 F20101208_AABPDT bu_k_Page_092.QC.jpg
14bf8dcb38e3d79100633163359e17ad
90bbbd53957e083f1c7f1a8839a49562e69fcdf7
6044 F20101208_AABOYN bu_k_Page_013thm.jpg
dc4608a4e3dfb623cbf1cbec0f150a0c
018588719cc0623d3b5d064ec32c9faf9e6c228f
2760 F20101208_AABOXY bu_k_Page_005thm.jpg
34f4aaf019c04530b90eac9cf88b6cec
8f4ce7567f11becc5a288494fb2bece367b7ab28
20049 F20101208_AABOBF bu_k_Page_078.QC.jpg
924f0f6a156f4d098b7af7480fc8fae2
07b8c1590a58fbe67d57c3ca3e793c62e4e13b1f
F20101208_AABOAR bu_k_Page_030.tif
b69caa68c77d2ad0f9670a4cff69fbb6
f3bea46a79a841a615c1b907c73470a4dda842e6
6753 F20101208_AABOZB bu_k_Page_021thm.jpg
70cfe647a53997f740de365498351ce0
760b1f754d000893cd73cdb6051bee21db3eac28
23489 F20101208_AABPEI bu_k_Page_100.QC.jpg
6f977b857d2cfd6faaa34b5ac9486c8b
f334a95a72a06790b50b48327be0ce3489e20567
3492 F20101208_AABPDU bu_k_Page_092thm.jpg
40ccb6c6be65b493a0b0f9078ecd2c82
a2b91727b23663b264db7e9e1aed8748802144f0
6967 F20101208_AABOYO bu_k_Page_014.QC.jpg
7f5346141bcbd07479c362d62526493b
512f81bbec78c8944fb55090a9fe0667bb9575f5
24086 F20101208_AABOXZ bu_k_Page_006.QC.jpg
ac336bd6b8377bf17591abf3b676c767
ba4bffdd78f59adc2fe43f6ea12ec01305a90b98
21160 F20101208_AABOBG bu_k_Page_118.jpg
774c1e2c5a42e0430838b1caf169f72a
e201522674d131ad5d193e22ee92cc16c3bb556f
4880 F20101208_AABOAS bu_k_Page_029thm.jpg
3b52554206f44f0251f4358d1436330b
a84de110d57215efeb946a393a9b749ed7272e83
26891 F20101208_AABOZC bu_k_Page_022.QC.jpg
58eb7c2c3eeed7fda04bfc3bfce4a4ab
0dde792c5b24297973a7d915b29c0a902c22a2db
6708 F20101208_AABPEJ bu_k_Page_100thm.jpg
3e745dd39d1e04baa9eeb77a9ebe0b45
b6d1c3fe5043af54a6b14d0362b467a8a7e35b37
10068 F20101208_AABPDV bu_k_Page_093.QC.jpg
7d668450563463dd8f5d67aaeac7d9db
e6e5f5213c1cd3fee73d4f427dde4d7da555b492
2253 F20101208_AABOYP bu_k_Page_014thm.jpg
d464869a5bdc1eaf0ca03d329fcd5caa
48527af348b534f4478f3230334e4973035e47dd
76306 F20101208_AABOBH bu_k_Page_072.jpg
59cc31e48878affd25920082516cb7d1
413c6d452523959b25b8f8e7b819914b37e2c929
27245 F20101208_AABOAT bu_k_Page_129.QC.jpg
471dc9b57c081e732101be31998e890e
e24137ebc43b6da984a8574bc5c75ea45aac0658
7035 F20101208_AABOZD bu_k_Page_022thm.jpg
7aaee109df3ed6c6891cd83f0eeeb941
cd0cb8842680758b2ddaca4e90194cffd7f6fcf3
24984 F20101208_AABPEK bu_k_Page_101.QC.jpg
c3d23329ef5163628f3be832dde248b5
a022306913d2ca4a8bca4d49fb320208e0588187
3583 F20101208_AABPDW bu_k_Page_093thm.jpg
49da82265c68b27021c70f10c8cf9e4b
cbe8e66cf40823ce80baff2ab6bc0238d181f912
23432 F20101208_AABOYQ bu_k_Page_015.QC.jpg
840284b3b97896c5930e0196ec9177c3
01356d7e77436f6aa1ca757072889c79c8224b68
F20101208_AABOBI bu_k_Page_133.tif
76e1dbf8a3ce5d169ea750f7c6f04724
234b077cfd3a54bc4e1a554219af06e23073fdd5
22073 F20101208_AABOAU bu_k_Page_016.QC.jpg
90d1e1ad0a8bc5bfa551c8d705baa945
e6921802b08e6d1b3457732de8d37a826b8ea5e5
24893 F20101208_AABOZE bu_k_Page_023.QC.jpg
8821f9ea4584d3876aa5ead8173c21de
b5fa1e1cb1332cfcec507c857cfe4f32c4a23d36
F20101208_AABPFA bu_k_Page_109thm.jpg
bb4e05d40c990a452b7159c5370cb3e8
bde9a396a73c0e6c821f11f26439ab2ae9f025ea
6804 F20101208_AABPEL bu_k_Page_101thm.jpg
b4b326151713782174744f14cce42fe6
c2e6691e33a942dcbdee0066cfc87dcab3f88061
12875 F20101208_AABPDX bu_k_Page_094.QC.jpg
27f0ae9c0b28cff07547ecd663989e0c
48bd4427fa4b1696485a6577e09956144e9276a8
6469 F20101208_AABOYR bu_k_Page_015thm.jpg
3e17f53d5bb312e6e4f0184ba94544b2
3d387c68ba5dfce92b4d9f32be31ac79c48fe7db
F20101208_AABOBJ bu_k_Page_116.tif
0f3d48c21d03db70edb3ef6c853bc3a2
398d7d4120843dbe0c3f6f2ca8e53a211d959cc7
F20101208_AABOAV bu_k_Page_099.tif
a5a5c66b14003da21e152e83a5e000fe
8489c4e817a9cdb73a6d791db57443409cf5df76
6656 F20101208_AABOZF bu_k_Page_023thm.jpg
2c8a26c3cd0f52307788bc79fd8af789
ad14f295ab5cc69560e39539bc946ff24c619f7a
20542 F20101208_AABPFB bu_k_Page_110.QC.jpg
5aa5afbfe52a35516ab11bd570c59b19
b5fc97aba9b7baff86e11e7ba2a2f37cf2ecc802
24383 F20101208_AABPEM bu_k_Page_102.QC.jpg
3777794d163a89b6b86bacde60998390
314adc3bef4d8e09ae32afa20634998501e97c89
3866 F20101208_AABPDY bu_k_Page_094thm.jpg
bc4658acdd80932cd7487ad2ba5ec340
f93186d6ca3fe8ef51808fbbbaa0f4233008db04
6023 F20101208_AABOYS bu_k_Page_016thm.jpg
d6f80b2abb5c6d8f7eb4fe632a4a37a3
b00a5aa834b11d87155d658e6e8e5f5194a99beb
195428 F20101208_AABOBK UFE0019608_00001.xml
bd9b8b0baf2792ecede404b93c2a54d2
59ca1efde15e8faf111dee2a8bb9085ba5444f8d
321217 F20101208_AABOAW bu_k_Page_055.jp2
0aa0dfec9b291fab4d636588f3cb3d38
e62890010c44b278d16eea15bbdb7ad17d0325d3
24860 F20101208_AABOZG bu_k_Page_024.QC.jpg
4b00e4f4171f970beb374379fb7d1009
3ae097f56428154426f74c1d34e3bc8f6df47584
5517 F20101208_AABPFC bu_k_Page_110thm.jpg
f88f1a29072a4ba3d52fa8ff730bce15
eab1d0e8b355458345af982f7020b8fe31214d1e
7009 F20101208_AABPEN bu_k_Page_102thm.jpg
0d6caf591a41f58849cc773916be7de2
a100d309d353ba2fee305a8e193cc510839810e8
8939 F20101208_AABPDZ bu_k_Page_095.QC.jpg
1a3b5ad569a9f048490aa696af8e80b3
3649500381678aac417d02768163f083a0a4651a
11108 F20101208_AABOYT bu_k_Page_017.QC.jpg
8dd23305ab5829ad5b577fc086551005
d57d2b337957fdc3a47abbd6d76dc60c6846ef77
73417 F20101208_AABOCA bu_k_Page_015.jpg
c5bcdc663c423f1fff1244fb1a48a1e2
fa44a29af7923ef442309c5e420bbb7312810476
F20101208_AABOAX bu_k_Page_101.tif
2a35e7ae15152201cf33ff5d1f78feaa
7f9b4640dea29f8ed5baf8d6546a0d99125ea59d
26322 F20101208_AABOZH bu_k_Page_025.QC.jpg
d8960d81115ce7230d546b13ac4275f4
04cd1a186fca42bdc418c2689863352c59b33811
12095 F20101208_AABPFD bu_k_Page_111.QC.jpg
38ef0cb2d0009662b19019de65106e4c
fbeda7a15c41e3ad962f932c813d9da2efc93c83
24753 F20101208_AABPEO bu_k_Page_103.QC.jpg
1c7eb89677993e4d0be0abcea8835e5e
cd29d717cbe82b3a11d11eb18e423ec0d3e23db2
3103 F20101208_AABOYU bu_k_Page_017thm.jpg
fab88184f756f08d65bcc4f0c6da6a67
5689935273100b175ba8f70f788a6f709d8d00b0
66792 F20101208_AABOCB bu_k_Page_016.jpg
e91b42123278b2f2797d3f056d9580fc
c42dace25396f25152520a9d07ec56db30356d7b
1093 F20101208_AABOAY bu_k_Page_098.txt
0920f903fa0bb5db7d30a4a2765a0885
41c810ced068604dd7dde867a8560be3c0aad5b9
7275 F20101208_AABOZI bu_k_Page_025thm.jpg
9c68ce1272213a2005d2395257fb2a12
d80fb47b6f56d21a337b0094c668b1662586d54f
6767 F20101208_AABPEP bu_k_Page_103thm.jpg
0a46d1406f0972df53088fc24e13444d
39de2ac9071cdd0ebb1f2863a10a9097c34ad796
6634 F20101208_AABOYV bu_k_Page_018.QC.jpg
1a678ef120e3d2eac9b8e8f6bfb14cec
7a14d170deaee977ee248c730a32b574033b6973
24429 F20101208_AABOBN bu_k_Page_001.jpg
786b6a862dc7d8b9898b7677e2c06a6c
b4439417eabd18fcc1abb7eef94890752012ffd1
10736 F20101208_AABOAZ bu_k_Page_060.QC.jpg
15601b87709da884347e2d61ada70a77
117befa3268da273676ff8c8ce9fd2be7aa98924
27537 F20101208_AABOZJ bu_k_Page_026.QC.jpg
eca37e6a19a721df8990a12a174cbd3c
cc284a7f5c713c1dd2157412f21cc7c995247fd7
3598 F20101208_AABPFE bu_k_Page_111thm.jpg
6ceb246b3b02343970093376f8bc66cd
98733b2215bc242ea11fc7c06f8cad920ed67ba3
23385 F20101208_AABPEQ bu_k_Page_104.QC.jpg
78fd90413151adcbd3ce363e77541f85
26eafac1eef7d411a92aeef02e8b85b49294c223
13526 F20101208_AABOYW bu_k_Page_019.QC.jpg
c2892fda36253c3817ed5767124b54c7
f43e6d69e3a398774a55d86c40355b06fc25adef
37120 F20101208_AABOCC bu_k_Page_017.jpg
4e9d20eb08078a2867477616c115dee3
e2eb5e8622bbcd9d97ef5778485ce3d964a9ea12
11171 F20101208_AABOBO bu_k_Page_003.jpg
6b9b4da8f02b3d6dd66149e8aba1734c
e464b6daba90b7f9ca62abc9cf5d03795b40d6eb
7409 F20101208_AABOZK bu_k_Page_026thm.jpg
9afe0e1be7e5fc1e0bd381c7ce9ed61f
f96cb023f16f30b6d5e5e63dadd1e760f7d52f81
7791 F20101208_AABPFF bu_k_Page_112.QC.jpg
87f6dc24091ba5e75923a7bb608d81dc
fefcc99effdd209dc55cef9f9ea453ba03b3809a
6570 F20101208_AABPER bu_k_Page_104thm.jpg
89f98a5c37e559ec37446f330fbf53fe
e265136de6c6c9aeb21ab5fd307411047b71d199
4205 F20101208_AABOYX bu_k_Page_019thm.jpg
c1b22320bf66374360f64d3fdce45819
57171d52f2466cb29a435726c825fa0feafe1cad
20945 F20101208_AABOCD bu_k_Page_018.jpg
6c73911762a5118cb9cec34f9c9cf75b
cb29c0fac0a5dc7fd9b7edbed4fcd9858cdfe15f
75772 F20101208_AABOBP bu_k_Page_004.jpg
6f76c3897a52aacf665a727c7f6e1f5b
ba593d1b4e30bde89967b6e73542b133013c0f00
24847 F20101208_AABOZL bu_k_Page_027.QC.jpg
1a4978e41f79022686f27cb552c10393
339d36aec52c5d3a45b580668a1dc12680ce3968
2580 F20101208_AABPFG bu_k_Page_112thm.jpg
4164a4aeec05e05343e78b8d47439a9a
d3bcdcfb034c8ba2923b10b8908c5675d5cce36c
24306 F20101208_AABPES bu_k_Page_105.QC.jpg
4b0f3652633e546c15618193c19d0d33
fe096eb9a98091fb8cdbdddab96ec9030cde3bd6
43162 F20101208_AABOCE bu_k_Page_019.jpg
ec039ac0c6b4867ebf7bea499918aa38
8ae6b3d2395db1f8641d22b9f6fa3f7978dd103f
26766 F20101208_AABOBQ bu_k_Page_005.jpg
092a055d9fb8fa96dd8dc6e533904d7e
386ea786b082c6e5e92e8c29fb3f280af352d5fe
6493 F20101208_AABOZM bu_k_Page_027thm.jpg
382da1b569380f0b7dc5536b0cb63ea7
dcbc1686907ba6b4d6d2a9aed4d60da07b1e17f1
10470 F20101208_AABPFH bu_k_Page_113.QC.jpg
ab4befb81861998e4563bc6c3f444a67
4aaf1a3f35a19848d09f3b48cb8093716c5c8eb9
6636 F20101208_AABPET bu_k_Page_105thm.jpg
fb89666d4389af8336b54fc0501d2673
ecb2e58e9416c8c1c5d4e86321e4ee55171a0cc4
11821 F20101208_AABOYY bu_k_Page_020.QC.jpg
6867d38dda90e37228da10070a83ab5d
4e96aefcaaeac1d96fdb89761db942110f2a5374
39641 F20101208_AABOCF bu_k_Page_020.jpg
c6c2fe7b6255be32c4d2dc2a392a06e3
703f1811c824418a195b59be7c2347aaa2225ab6
88447 F20101208_AABOBR bu_k_Page_006.jpg
058bb86045914c80b3b9e82c58ca0a76
90b715bb915dee333af5579dd96f72d2a85fc562
19400 F20101208_AABOZN bu_k_Page_028.QC.jpg
28f2a3582775c5d95dc1bab1d91c26ee
0838983f06d14632090350ca58039603e3b9f852


Permanent Link: http://ufdc.ufl.edu/UFE0019608/00001

Material Information

Title: Surfactant Mediated Passivation to Achieve Chemical Mechanical Polishing Selectivity
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0019608:00001

Permanent Link: http://ufdc.ufl.edu/UFE0019608/00001

Material Information

Title: Surfactant Mediated Passivation to Achieve Chemical Mechanical Polishing Selectivity
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0019608:00001


This item has the following downloads:


Full Text





SURFACTANT MEDIATED PASSIVATION TO ACHIEVE CHEMICAL MECHANICAL
POLISHING SELECTIVITY




















By

KYOUNG-HO BU


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2007





































2007 Kyoung-Ho Bu

































To my beloved family, Mineok, Minji, and Seongah Byeon.









ACKNOWLEDGMENTS

It is a privilege to work with intelligent and committed individuals. Too many people to

mention have influenced my work and provided inspiration and useful suggestions over many

years, but I would especially like to express my appreciation to my advisor, Dr. Brij Moudgil, for

his invaluable research guidance and constructive support through intense discussions and

productive feedback on this study. His sincere dedication to science, discipline in conducting

research and considerate attention to details have always kept me moving forward and made

significant contributions to this dissertation.

I would also like to acknowledge the other members of my advisory committee, Dr. Rajiv

Singh, Dr. Stephen Pearton, Dr. Dinesh Shah, and Dr. Wolfgang Sigmund, for their

indispensable support. I also wish to acknowledge Dr. Susan Sinnott, Dr. Chang-Won Park, Dr.

Yakov Rabinovich, Dr. Ivan Vakarelski, Dr. Parvesh Sharma, and Dr. Manoj Varshney who

have informed and elaborated this work, with special appreciation to Dr. Ko Higashitani for his

valuable insights.

I am grateful to the National Science Foundation's Engineering Research Center for

Particle Science and Technology for financially supporting this research (Grant EEC-94-02989).

To Gary Schieffele, Gill Brubaker, and all other ERC staff, faculty, and administrators, I extend

my hearty thanks for making my time there productive.

Colleagues and friends who have contributed to this research through critical discussions

as well as friendship include Scott Brown, Vijay Krishna, Madhavan Esayanur, Rhye Hamey,

Marie Kissinger, Monica James, Dushyant Shekhawat, Suresh Yeruva, Kalyan Gokhale, Amit

Singh, Debamitra Duta, Stephen Tedeschi, Sejin Kim, Takgeun Oh, Sangyup Kim, Won-Seop

Choi, Seung-Mahn Lee, Kyo-Se Choi, Suho Jung, and Inkuk Jun. I also thank Bryce Devine and

Bryan Op't Holt for training me how to use modeling tools.









I have been blessed with Father Sangsun Park in Gainesville Korean Catholic Church who

helps me have peace in mind, and blessed with my children, Minseok and Minji, who encourage

me to overcome obstacles and motivate me to try my best in life. In addition, I owe particular

debts to my parents and my parents-in-law for their strong confidence in my family.

Finally, I am always grateful to my wife, Seongah, for her patience and support in spite of

all ups and downs during my study. This work would not have been possible without her.









TABLE OF CONTENTS

page

A CK N O W LED G M EN T S ................................................................. ........... ............. .....

L IST O F T A B L E S ..................................................................................................... . 8

LIST OF FIGURES .................................. .. ..... ..... ................. .9

A B S T R A C T ............ ................... ............................................................ 13

CHAPTER

1 INTRODUCTION ............... ..................................................... ..... 15

2 L ITE R A TU R E R E V IE W ........................................................................ .. .......................2 1

Shallow Trench Isolation (STI) Structure and Selectivity of Slurry ....................................21
Influence of Selectivity on Global Planarization in STI CMP Process.............................22
N anotopography ................................................................................24
Surfactant M ediated Lubrication Effects........................................................... ..................24
Surface Chemical Characteristics of Si02 and Si3N4 Surfaces in Aqueous Solution.............25
Surfactants Adsorption on Silicon Nitride and Lubrication Effect .....................................26
M ixed Surfactants Sy stem ............................................................................ ....................27
Research Approach ................... ......................................... .... ........ .. 28

3 CMP CHARACTERISTICS OF SILICA AND SILICON NITRIDE............................. 37

E x p e rim e n ta l ...................... .............. .... ......... ........................ ........................... 3 8
Relationship between Material Removal Rate (MRR) and Young's Modulus....................39
Role of Electrostatic Interactions on M RR ........................................................ ...............41
Effect of pH ................................................................. .............. ...... ........... 42
Effect of Salt A addition .............................. ........................................ ............... 45
Parameters Affecting Surface Finish in STI CMP ...................................... ............... 48
Salt M edited Lubrication ................. .................................. ...... ........ .......... ....... 1

4 ROLE OF SURFACTANTS IN DEVELOPING SELECTIVE PASSIVATION LAYER
IN C M P ............................................. ................................. ..........................72

High Selectivity Slurry Using Surfactants.................. .................................. 73
Surfactant Mediated Boundary Layer Lubrication for Selective Polishing..........................75
Optimization of High Selectivity Slurry.......................... ....................... ...............76

5 ADSORPTION STUDY OF SODIUM DODECYL SULFATE ON SILICA ......................86

A dsorption B behavior of SD S on Silica........................................................ ............... 87
Structure of Adsorbed SD S M olecules........ ......... ..................... ............... .... .......... 89









6 APPLICATION OF DENSITY FUNTIONAL THEORY BASED MODELING FOR
SURFACTANT ADSORPTION STUDY ................................................ ..................... 100

M eth o d o lo g ie s ........................................................................... 10 1
Structures and R sources ........................................................................... .......... ........... 104
R results and D iscu ssion ............... ..... ... ...... ... .. ...... ....................................... ... 106
SDS Adsorption on Silica at, below, and above the Isoelectric Point (IEP)...............07
SD S A dsorption on Silicon N itride at IEP ........................................ ..................... 108
TX -100 A dsorption on Silica at IEP ........................................ ........................ 109

7 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK.................................... 120

C o n c lu sio n s ........................................................................................................................... 1 2 0
Suggestions for Future W ork ......... ................... ....................................... ............... 122

L IST O F R E F E R E N C E S ......... ............................................................................ .................. 125

B IO G R A PH IC A L SK E T C H .............................................................................. .. ............. 133









LIST OF TABLES


Table page

1-1 A Product Generations and Chip Size Model Technology Trend Targets-Near-term
Y e a rs ............................................................................................. 1 7

3-1 Young's modulus, hardness measured by nanoindentation method, material removal
rate (MRR), ratio of MRR (CMP pressure of 7 psi), and ratio of Young's modulus
for silica and silicon nitride........... .......................................................... .. .... .. ........54

6-1 Adsorption energy (kcal/mol) calculated by density functional theory (DFT) based
m ethod (B3LYP) using 6-31G* basis set .................................................. ..................111

6-2 Adsorption free energy (kcal/mol) of SDS on silica calculated from adsorption
density data in Ch. 5 at different pH and two different added concentrations (1.6mM
and 16mM ). ................................... ................................... ......... 112









LIST OF FIGURES


Figure page

1-1. Schematic representation of chemical mechanical polishing (CMP) process. .................18

1-2. M oore's Law M eans M ore Perform ance................................................. ....... ........ 19

1-3. Multilevel metallization, cross section with silica dielectric and aluminum
m etallization ......... .... ..... ......... ....................................... .............................20

2-1. Schem atic shallow isolation structure .......................... ........................... ................29

2-2. Nanotopography (a) Top view and (b) cross-section graph of substrate
nanotopography............................................................... .... ..... ......... 30

2-3. In-situ friction force and material removal rate responses of the baseline slurries (12
wt%, 0.2 mm primary particle size) and the slurries containing C12TAB, C1oTAB and
CsTAB surfactants at 32, 68 and 140 mM concentrations in the presence of 0.6 M
N aC 1 at pH 10.5. ........................................................................... 3 1

2-4. Lateral force as a function of loading force in the presence of surfactant [22] ...............32

2-5. Zeta potential behavior of silica, silicon nitride, cerium oxide (ceria), and polishing
pad (polyurethane) with respect to the pH ............................................. ............... 33

2-6. Maximum surface concentration of benzoic acid (*) and pyridine (0) obtained by
fitting the adsorption data to a Langmuir-Freundlich equation. ......................................34

2-7. Friction coefficient of silicon nitride ceramic as a function of load in pure water (o)
and silane aqu eou s solution ( ) .............................................................. .....................35

2-8. The mechanism of high-ionic-strength slurry stabilization by the synergistic mixture
of anionic and nonionic surfactants .............................................................................36

3-1. Variations of mateiral removal rate (MRR) for silica and silicon nitride substrate as a
function of applied pressure by using undiluted (30 wt%) colloidal silica slurry at pH
1 0 .4 ........................................ ......................... ..........................5 5

3-2. Variations of MRR of silica and silicon nitride substrate and calculated electrostatic
force between two abrasives as a function of pH of the diluted (12 wt%) colloidal
silica-based slurry (K lebosol 1501-50) ........................................ ......................... 56

3-3. Particle size distributions of colloidal silica slurry at two different pH conditions..........57

3-4. Zeta potential of colloidal silica slurry and electrostatic force between silica abrasive
p article s. ........ ........ ......................................................................... 5 8









3-5. Variations of MRR and calculated electrostatic force between two abrasives as a
function of slurry NaCl salt concentrations in the slurry at pH 10.4...............................59

3-6. Particle size distributions of colloidal silica slurry (Klebosol 1501-50, 12 wt%) as a
function of salt concentrations at pH 10.4. ............................................. ............... 60

3-7. Surface roughness of silica and silicon nitride substrate after CMP as a function of
added salt (N aC1) concentration at pH 10.4.................................... ....................... 61

3-8. Material removal rate of silica and silicon nitride as a function of repulsive
electrostatic force between silica abrasives. ........................................... ............... 62

3-9. Surface roughness of silica and silicon nitride substrates after CMP as a function of
slurry pH .........................................................................................63

3-10. Surface morphologies and profiles of substrates from two pH conditions......................64

3-11. Material thickness change of silica and silicon nitride substrates as a function of
immersed time in pH 13 NaOH solution. ........................................ ....... ............... 65

3-12. Surface morphologies and profiles of substrates before and after etching in pH 13
N aO H solu tion s.............................. .......................................................... ............... 6 6

3-13. Etch pits formed on (a) silica and (b) silicon nitride substrate immersed in 0.1 M (pH
13) N aO H solution for 12 days ................................................ .............................. 67

3-14. Lateral force of a 6.8 |tm silica particle interacting with a silica substrate in pure
water and CsC1, NaC1, and LiCl solutions of 1 M..............................................................68

3-15. Schematic representation of the hypothetical frictional mechanisms.............................69

3-16. Particle size distributions of colloidal silica slurry (Fuso PL-7) without salt and with
1 M L iC 1 and 1 M C sC .......................... ...... ................................... .. .....70

3-17. Material removal rate of silica substrates by CMP using diluted (9.6 wt%) colloidal
silica slurries (PL-7) without salt and with 1 M LiCl and 1 M CsCl as a function of
applied polishing pressure.......... ................................................................... ..... .... .. 7 1

4-1. Influence of SDS addition on CMP performances. ........................................................79

4-2. Surface finish of silica and silicon nitride substrates processed with standard and
high selectiv ity slu rry ............ .... .................................................................. ........ .. ... 80

4-3. Variation of zeta potential of silica and silicon nitride substrate and adsorption
density of 16mM SDS on silica and silicon nitride powder measured by total organic
carbon (T O C ). ............................................................................ 81









4-4. Variation of MRR and accompanying selectivity of Klebosol slurry (12 wt%) as a
function of added SDS concentration at pH 2. ...................................... ............... 82

4-5. Adsorption density of SDS on 12 wt% Klebosol slurry with 16 mM SDS as a
function of pH ............................................................................................................ ...... 83

4-6. Effect of alkyl chain length of sodium alkyl sulfate on MRR and selectivity at pH 2......84

4-7. MRR and selectivity obtained by slurries with various surfactant and surfactant
m ix tu re s at p H 2 ...................................... ............. ................. ................ 8 5

5-1. Adsorption isotherm of SDS on colloidal silica (Klebosol 1501-50, 12 wt%) at pH
1 0 .4 ........................................ ......................... ..........................9 3

5-2. Adsorption density of SDS on colloidal silica (12 wt% Klebosol 1501-50) at SDS
concentration of 1.6 mM and 16 mM and zeta potential as a function of pH. ..................94

5-3. Zeta potential of Klebosol slurry as a function of SDS concentration at pH 10.4.............95

5-4. Pictorial depictions of the possible surfactant aggregates films at concentrations
corresponding to I-IV in Figure 5-3 ....................................................... ............... 96

5-5. Adsorption characteristics of SDS on Klebosol silica slurry and zeta potential as a
function of concentration of SDS at pH 10.4.......................................... ............... 97

5-6. FTIR/ATR Spectra of SDS solution at 1, 2.5, 5 and 10 mM bulk concentration in the
CH2 stretching region (2921, 2924) measured at pH 10.4 using Si ATR crystal ............98

5-7. Particle size distribution of Geltech SiO2 at pH 2 with and without 16 mM SDS 12
hours after pH change. ....................................... .... .. ..... .............. .. 99

6-1. Optimized (a) Si(OH)4, (b) Si(NH2)4, (c) Sodiumdodecyl sulfate (SDS), and (d)
Triton X-100 (TX-100) structure using B3LYP method and 6-31G* basis set..............113

6-2. Optimized SiOH4 and DS- complex structure using B3LYP method and 6-31G* basis
set. ................................................................ ...........................1 14

6-3. Optimized SiOH+ and DS- complex structure using B3LYP method and 6-31G*
basis set. ......... .... .............. ..................................... ............................115

6-4. Sturcture of SiO4H3- and DS- complex. Optimization is not complete, since two
m olecules are being separated to decrease energy................................... ... ..................116

6-5. Optimized Si04H3-, Na and DS- complex structure using B3LYP method and 6-
3 1G basis set. ......... ..... ............. ................................... ......................... 117

6-6. Optimized Si(NH2)4 and DS- complex structure using B3LYP method and 6-31G*
basis set. ......... .... .............. ..................................... ............................118









6-7. Optimized SiOH4 and TX-100 complex structure using B3LYP method and 6-31G*
b asis set. ........................ ....... ................. 1 19









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

SURFACTANT MEDIATED PASSIVATION TO ACHIEVE CHEMICAL MECHANICAL
POLISHING SELECTIVITY


By

Kyoung-Ho Bu

May 2007

Chair: Brij M. Moudgil
Major: Materials Science and Engineering

Chemical mechanical polishing (CMP) is an indispensable technique in the

microelectronics industry to achieve planarization and patterning of metal and dielectric layers.

Device fabrication using high density and small pattern size requires precise control of CMP

slurry properties.

In this study, the performance of a colloidal silica CMP slurry for silica/silicon nitride,

which consists of the shallow trench isolation (STI) structures, was investigated. Factors

determine material removal rate and surface finish were examined. It was found that electrostatic

interactions can have significant effects on CMP performance. Emphasis was placed on selective

removal of material. More than 10-fold increase in selectivity over conventional colloidal silica

slurry was achieved with the addition of sodium dodecyl sulfate (SDS), an anionic surfactant.

Adsorption characteristics of SDS on silica and silicon nitride were measured as a function of

slurry pH and surfactant concentration. It was determined that the preferential adsorption of SDS

on silicon nitride by electrostatic attraction results in the formation of a material-selective self-

assembled passivation (boundary lubrication) layer leading to selective polishing. It was found

that the adsorption density of surfactant plays a dominant role in determining selectivity.









Accordingly, material-targeted boundary layer lubrication concept may be used to develop

selective CMP polishing slurries.

A theoretical approach based on density function theory was attempted to model various

aspects of surfactant adsorption. Through this approach, it was possible to predict adsorption

behavior and related thermodynamic properties to assist selection of passivating molecules.









CHAPTER 1
INTRODUCTION

Chemical mechanical polishing (CMP) is the planarization technique predominantly used

for the fabrication of multilayer devices. Main components for CMP process include the

substrate to be polished, the slurry that provides the chemistry and abrasives for mechanical

removal-and the polishing pad. A schematic of CMP system is shown in Figure 1-1. Due to the

demand for the faster and smaller devices, the number of devices (density) on a single wafer is

expected to grow constantly as depicted in Figure 1-2. Accordingly, the size of components of a

device is expected to become smaller as listed in Table 1-1. Hence the requirements for large

scale integration are becoming more challenging.

Current semiconductor devices are composed of multilayers as shown in Figure 1-3. Due

to the planarity requirement for lithography processes, further processing is not possible if the

required planarity is not achieved. In addition, the standard for global planarization is becoming

more demanding due to the high degree of device integration.

Among the various structures requiring CMP, shallow trench isolation (STI) is one of the

most challenging, due to its large variation in pattern density. There are a number of possible

approaches to accomplish global planarization in STI CMP process. Among these, the

development of high selectivity slurries has been gaining more significance in order to

accomplish a one-step CMP process for global planarization. State of the art, high selectivity

ceria based slurry has several drawbacks such as problems with coagulation and high defectivity,

whereas conventional silica based slurries are known to be free of those problems, but they

exhibit low polishing selectivity between silicon nitride and silica substrates. In this dissertation,

silica based slurries were modified to achieve the targeted selectivity of 15 or higher.









The overall objective of the proposed investigation is to improve the selectivity (ratio of

material removal rate of silica to silicon nitride) of the STI CMP slurry. Specific objective is to

differentially modify surface states of silicon nitride and silica with surfactant or polymer

adsorption, thereby selectively minimizing silicon nitride polishing, and thus leading to enhanced

global planarization in STI CMP process. A synopsis of the various research tasks constituting

this study is organized as follows.

Chapter 2 reviews the literature on the STI CMP process and slurry selectivity. Different

defects, hampering device performance will be addressed. The selectivity of the CMP slurry will

be defined and its effect on global planarization will be discussed. Finally, strategies to increase

the selectivity will be suggested. Chapter 3 covers the CMP characteristics of silica and silicon

nitride substrates by colloidal silica slurry with respect to the material removal rate (MRR) and

surface finish. Variables affecting the polishing process have been studied with special emphasis

on electrostatic interactions. Chapter 4 presents the methodologies to increase the selectivity of

the slurry. Specific mechanisms for observed results will be discussed. Chapter 5 discusses the

adsorption behavior of sodium dodecyl sulfate (SDS) on silica substrates, since it was found that

SDS adsorption on silica abrasive particles determines the necessary dosage of surfactant to

fabricate high selectivity slurries. Chapter 6 describes the modeling efforts to develop

methodologies based on density functional theory to predict optimal conditions for selective

surfactant coating. Chapter 7 summarizes the conclusions of this study and suggests future work.















Table 1-1 A Product Generations and Chip Size Model Technology Trend Targets-Near-term
Years [1].

Year of Production 2005 2006 2007 2008 2009 2010 2011 2012 2013
DRAM I1 Pitch (nm) (contacted) 80 70 65 57 50 45 40 36 32
MPU. I iC Metal 1 (A/) 2 Pitch (nm) 90 78 68 59 52 45 40 36 32
MPUPrinted Gate I. irih (nm) 54 48 42 38 34 30 27 24 21
MPU Physical Gate l, ,grih (nm) 32 28 25 23 20 18 16 14 13

ASIC/Low 0,.. ,, g Power Printed Gate
76 64 54 48 42 38 34 30 27
I. ,i01li (nm)

ASIC/Low 0,,.. l,, Power Physical Gate
AS Low 0 Power Physical Gate 45 38 32 28 25 23 20 18 16
Flash Pitch m) (un-contactedPol 76 64 57 51 45 40 36 32 28(nm)
Flash 12 Pitch (nm) (un-contacted Poly) 6/) 76 64 57 51 45 40 36 32 28

















Substrate
Holder


Slurry Feed


Slurry Feed


HIder


k Polishing Pad






Figure 1-1. Schematic representation of chemical mechanical polishing (CMP) process. (a) Side
view; (b) Top view.






































Figure 1-2. Moore's Law Means More Performance. Processing power, measured in millions of
instructions per second (MIPS), has steadily risen because of increased transistor
counts [2].




















LAYER 5


LAYER 4



LAYER 3


LAYER 2


LAYER T






Figure 1-3. Multilevel metallization, cross section with silica dielectric and aluminum
metallization [3].









CHAPTER 2
LITERATURE REVIEW

Shallow Trench Isolation (STI) Structure and Selectivity of Slurry

Chemical mechanical polishing or planarization (CMP) is the key technology for shallow

trench isolation (STI) process. STI process can reduce the required area for the device isolation

and give better planarity relative to the local oxidation of silicon (LOCOS) process. Therefore,

existing sub-0.13 [Lm technologies device isolation techniques strongly depend on the STI CMP

process [4-7].

There are several drawbacks such as dishing of silica, erosion of silicon nitride and failure

to clear oxide that hamper global planarization in CMP process [8]. Typically the thickness

uniformity across the substrate (usually called within-substrate non-uniformity, or WIWNU)

must be below 3%, and dishing must typically be less than 20-50 nm. To minimize such defects,

current STI CMP process is comprised of multi-steps [9] or raw structure modifications such as

reverse mask, dummy active area, and additional active area [10]. For better productivity and

process simplicity, a minimum number of process steps are highly desired and accordingly,

approaches for "high selectivity single-step" slurry designs are being widely investigated [11-13].

Usually selectivity represents the ratio of material removal rate (MRR) of silica to silicon nitride:

Material removal rate of Silica
Selectivity = (2-1)
Material removal rate of Silion nitride

In general, conventional silica abrasive based STI CMP slurry exhibits selectivity in the

range of 3 to 4 [14]. According to the result reported by J. Schlueter, erosion of silicon nitride

could be minimized to less than 100 A using ceria based high selectivity slurry in a multi-step

STI CMP [15]. Besides the influence on planarization, high selectivity provides enhanced

endpoint detection capability. Generally, if oxide to nitride polishing selectivity is greater than









15, monitoring substrate carrier motor current can be utilized for endpoint detection [11].

Therefore, research on improving selectivity and understanding the polishing mechanisms to

achieve global planarization are needed. In this study, systematic approaches and strategies to

improve selectivity of the STI CMP slurry for single-step CMP process were investigated.

In the following sections, the detailed influences of selectivity on global planarization

will be introduced, and issues of nanotopography that justify a strong need for high selective STI

CMP process will be outlined. Next, a brief review of the polishing passivation/inhibition

mechanism (i.e., surfactant mediated lubrication effects) will be provided. Surface chemical

characteristics of silica and silicon nitridewill be reviewed, followed by examples of specifically

adsorbing surfactants on silicon nitride surface. As an alternative to inhibit polishing of silicon

nitride by surfactants, silane additives to form passivation layer on silicon nitride will be

introduced.

Influence of Selectivity on Global Planarization in STI CMP Process

As previously mentioned, several obstacles exist inhibiting global planarization in STI

CMP. Figure 2-1 shows a schematic of a typical STI structure. It consists of a silicon device, a

silicon nitride mask, and a silica insulating layer inside of the trenches. In the ideal CMP process,

the oxide should be removed completely in all active regions, leaving it only in the trench

regions (Figure 2-1 (b)) without eroding silicon nitride. In reality, there are three failure modes

such as failure to clear oxide, excessive removal of nitride, and excessive removal of oxide [8].

The former is primarily an end-point detection issue, whereas the other two mechanisms are

closely related to the pattern density of the device, selectivity of slurry, pad stiffness, imposing

pressure, etc. [12]. To minimize these barriers, several approaches have been evaluated. One

method is to use a stiffer pad and lower selectivity slurry [5], and the other is to use a softer pad

and higher selectivity slurry [16]. When a stiffer pad is used, which does not bend in the applied









pressure range, the highest portion of the surface will start to be polished ultimately resulting in

global planarization. However, there is also a possibility of poor surface finish and wafer

breakage. When a softer pad is used, which has a greater flexibility, all the structures on

substrate will be in contact to pad, and hence a high selectivity slurry will be required not to

preferentially polish the unwanted structure. In this case, the risk of poor surface finish and

substrate breakage will be reduced.

Current high selectivity slurries in STI CMP usually contain ceria abrasives showing

higher material removal rate for silica than silicon nitride [16]. In general, for higher pattern

densities of which the area of silica isolation layer is not large, dishing effect decreases, since the

pad bending is limited. For lower pattern densities, dishing effect increases because the pad

bending is high [17]. Therefore in each case, the pad materials and operating pressure should be

chosen appropriately.

Kim et al. investigated the influence of slurry selectivity of the slurry on erosion and

planarity by modeling. It was predicted that above 30% active pattern density, high selectivity

slurries show good planarity [18]. In these cases, planarity is defined as the difference of height

between the highest region and the lowest region on a substrate. Considering that higher pattern

densities of the device will be required with decreasing device size in the future, a systematic

research for a high selectivity slurry will be essential to meet these goals.

In general, current STI CMP processes use silica abrasives that show low selectivity (about

3 4) [14]. W. G. America investigated the influence of selectivity on material removal rate of

silica and silicon nitride using silica and ceria abrasives. In this case, the material removal rate of

silica and silicon nitride was determined to be about 2700 A/min and 700 A/min, respectively

[19]. Recently, ceria abrasives have shown higher selectivity (more than 5), and are being









investigated for further enhancement. According to W. G. America, material removal rate of

silica using ceria abrasives was more than 5700 A/min as compared to 800 A/min for silicon

nitride [19]. However in ceria CMP, the pH at which maximum polishing rate and maximum

selectivity are achieved is about 8, which also is the isoelectric point (IEP) of ceria. This results

in coagulation of ceria yielding poor surface morphologies with scratches and higher roughness

[13]. Recently, it has been reported that by decreasing the size of ceria abrasive particles, the

number of scratches can be decreased significantly [20].

Nanotopography

An emerging issue impairing global planarization in STI CMP is nanotopography. This

phenomenon is becoming a strong driving force for developing high selectivity slurries.

Nanotopography is a term used to describe relatively gentle (10-100 nm) surface height

variations occurring over lateral distances of 1-10 mm on unpatterned silicon substrates (Figure

2-2). Boning et al. have investigated this issue by modeling and verified it by experiments, and

have suggested that due to the height variation of blanket wafer, several defect mechanisms

come into play such as failure to clear oxide and excess nitride thinning (erosion). It has been

commonly believed that stiffer pad would yield acceptable planarization [5]. On the contrary, it

has been shown that softer pad and lower pressure is more effective in minimizing such defects

[5].

With respect to this phenomena, if the selectivity of the slurry is not high enough and

endpoint detection is not accurate, accompanying erosion will be unavoidable. Silicon nitride

erosion can be minimized if only additional protective layers exist on silicon nitride surface.

Surfactant Mediated Lubrication Effects

As a protective mechanism from polishing for silicon nitride, one of the approaches is to

incorporate surfactant mediated lubrication effects. Basim et al. have shown that the addition of









long chain cationic surfactant (e.g. C12TAB) produces an enhanced defect-free surface

morphology but the polishing rate was extremely small due to the lubrication effect of surfactant

(Figure 2-3.) [21]. Although this research was focused on dispersion of abrasive particles, it

implied that long chain surfactant can act as an anti-polishing agent.

Vakarelski et al. showed that the primary mechanism of lubrication is the formation of an

intervening surfactant aggregate film on solid-liquid interface largely by electrostatic interactions

[22]. In addition, the decrease in frictional force depends on the concentration of surfactant.

After the concentration reaches critical micelle concentration (CMC), there was no further

decrease in lateral frictionall) force. The effect of surfactant concentration on the lateral force is

illustrated in Figure 2-4.

Surface Chemical Characteristics of SiO2 and Si3N4 Surfaces in Aqueous Solution

Understanding the surface chemistry of substrates is the first step to implement the above

approach to create a selective passivation/lubrication layer. It is well known that silicon nitride

forms the same type of surface hydroxyl layer as silica in an aqueous solution. However, there is

a difference in the surface group compositions. Figure 2-5 illustrates the zeta potential variation

with respect to pH. Unlike silica (IEP of 2.2), silicon nitride exhibits an IEP of about 5.8. This

difference is explained on the basis of relative number of silanol (Si-OH) and amine (Si2-NH)

groups on the silicon nitride surface as compared to only silanol groups on silica surface [23].

The silanol groups are acidic in nature and thus result in a lower IEP, while the presence of

amine groups results in a higher IEP. In the case of silicon nitride powder with an IEP of pH 6,

the ratio of nitrogen to oxygen was calculated to be approximately 0.2, and it was nearly 1 for

powders with an IEP of pH 7.9 [23].

Sonnefeld et al. reported, based on potentiometric titration measurements, that the surface

site densities of amine group (Si2NH) and that of silanol group (SiOH) are 0.56 /nm2 and 1.83









/nm2, respectively on the silicon nitride surface [24]. Density of silanol groups on silica surface

was estimated to be 0.74 /nm2 [25]. From these values, converted area of amine and silanol

groups for molecular adsorption is 1.79 nm2 and 0.546 nm2, respectively on the silicon nitride

surface, and area of silanol group on silica was 1.35 nm2.

Surfactants Adsorption on Silicon Nitride and Lubrication Effect

There have been many reports on stabilization of silicon nitride powders using polymeric

dispersants [26-30]. Malghan et al. investigated the dispersion behavior of silicon nitride powder

using both cationic Betz 1190 (quaternized polyamine epoxychlorohydrin) and anionic -

Darvan C (ammonium poymethacrylate) polymers [29]. In the case of cationic polyelectrolyte

(CPE), there was strong electrostatic attraction between CPE and silicon nitride powder at pH 9

leading to stable dispersion, while in the case of anionic polyelectrolyte (APE), the adsorption

was very restricted due to the similar surface charge, consequently, small adsorption occurred

possibly due to the hydrogen bonding. According to Hackley et al., anionic poly acrylic acid

(PAA) adsorption on silicon nitride surface decreased from 100% at pH 3 to around 25% at pH

10, however, stable dispersion was achieved due to depletion forces in the presence of PAA [26].

Besides the sign of surface charge, hydrogen bonding plays an important role in adsorption

of organic molecules on silicon nitride. Bergstrom et al. investigated the adsorption behavior of

various organic probe molecules in cyclohexane [31]. They showed that benzoic acid and benzyl

amine prefer to adsorb on the basic amine (Si2NH) groups via hydrogen bonding (N-H) (Figure

2-6). To accomplish selective adsorption of surfactants or polymers on silicon nitride surface,

anionic surfactants should be investigated first considering that nitride shows higher negative

zeta potential at pH 10.5 for current CMP conditions. Philipossian et al. showed that by applying

anionic poly-carboxylate, the selectivity increased from 5 to 100 [32]. They used ceria abrasives

for silica polishing at pH 8. According to their results, most of anionic surfactant adsorbed on









silicon nitride with some amount of polymer adsorption on silica and ceria abrasive particles,

resulting in overall decreased MRR from 500 to 100 (a.u.).

Hibi et al. investigated the lubrication effect of silane coupling agents (3-(2-

aminoethylaminopropyl) dimethoxymethylsilane) on silicon nitride and alumina ceramics

(Figure 2-7) [33]. They reported that amino-containing silane coupling agents formed the cross-

linked polysiloxane by hydrolysis and dehydrative condensation, which was effective in reducing

both friction and wear of silicon nitride. In other words, the additives reduced the wear of silicon

nitride as a result of inhibition of silicon nitride reaction with water. In this case, the silane

agents reacted with the oxide (silanol group) on silicon nitride surface.

As mentioned above, since the density of silanol groups on silica and silicon nitride surface

was estimated to be 0.74 /nm2 and 1.83 /nm2, respectively [24, 25], the extent of the passivation

on silicon nitride and silica is expected to be different.

Mixed Surfactants System

Palla et al. investigated the use of mixed surfactants to disperse the alumina abrasive

particles in CMP. They reported that by applying anionic surfactant, sodium dodecyl sulfate

(SDS), mixed with various nonionic surfactants, the dispersion stability was highly improved

[34]. The schematic of the slurry stabilization of alumina abrasives is shown in Figure 2-8. In

this scheme, adsorption was attributed to strong adsorption of ionic surfactants on abrasive

particles, and association of nonionic surfactants with ionic surfactants via hydrocarbon chain

interactions (attractive hydrophobic forces). Alumina is known to have Lewis active site similar

to silicon nitride, hence, it can be envisioned that mixed surfactants concept can be applied to

silicon nitride-silica system. However, under the normal CMP pH conditions, zeta potential of

silicon nitride is negative, indicating the greater significance of electrostatic interaction.









Research Approach

Commercial ceria abrasive STI CMP slurries with selectivity of about 5 are known to

result in high defectivity and post-CMP cleaning problems, while colloidal silica slurries has a

lower selectivity of 3 to 4, although they exhibit acceptable defectivity. Therefore in the

proposed research, surfactants that selectively adsorb on silicon nitride will be investigated and

methods to inhibit polishing and the mechanisms will be studied to improve global planarity.

One of the major challenges is the fact that both materials have silanol group on their

surfaces in water and show negative zeta potential at the conventional CMP pH of 10.4. The

ideal solution is to find a surfactant, which has selective affinity only to silicon nitride. To

achieve this goal, several anionic surfactants and mixed surfactant systems will be investigated in

terms of adsorption with respect to pH and added surfactant concentration.

In using silica abrasives under current CMP conditions, anionic surfactants adsorption on

abrasive particles will be largely opposed due to the similar (negative) charge of the adsorbate

and adsorbent. Therefore, to increase the amount of surfactant adsorption on silicon nitride,

readjustment of CMP process pH to a lower value may be required. Since pH plays a dominant

role in determining surfactant adsorption through electrostatic interactions, detailed investigation

of the adsorption behavior of the anionic surfactant as a function of pH will be required to

achieve optimal surfactant adsorption.










(a) Before CMP


Si02
/


(b) Ideal result after CMP


- S i3N 4 ..................................................


-- Si


(c) Erosion
(d) Dishing




IIImmi


(e) Failure to clear oxide

I \


I I


Figure 2-1. Schematic shallow isolation structure: (a) Initial structure before CMP: typical trench
isolation structure used to isolate "active" regions on a substrate where devices will
be built. The nitride layer has been patterned and a shallow trench etched into the
silicon. An oxide has then been deposited into the trench, which also results in
overburden oxide above the nitride active areas. (b) ideal result after CMP: the oxide
is removed completely in all active regions, leaving oxide only in the trench regions.
Three key failure mechanisms may arise: (c) excessive removal (erosion) of nitride in
active areas, (d) excess removal of oxide (dishing) within the trench, and (e) failure to
clear oxide from nitride active areas [8].

















(a) (b) 100oo

80
100 nnm E 60
60
.- 40
0 20
W o
-20
W -40
-60
-100 nm 60
-10 -80 Nanotopography

-100 Length




Figure 2-2. Nanotopography (a) Top view and (b) cross-section graph of substrate
nanotopography. Dotted line in (a) shows path of scan. The x axis in (b) indicates the
distance along the scan path in (a), moving from left to right [8].















10000


0 6 6000



4 4000



2 2000 .



O 0
Baseline Baseline 140mM 68mM 32mM
W/O Salt W Salt C8TAB CloTAB C12TAB


Figure 2-3. In-situ friction force and material removal rate responses of the baseline slurries (12
wt%, 0.2 mm primary particle size) and the slurries containing C12TAB, CloTAB and
CsTAB surfactants at 32, 68 and 140 mM concentrations in the presence of 0.6 M
NaCl at pH 10.5. (Striped bars represent the Friction Force responses and the solid
bars represent the Removal Rate responses) [21].















300
0 Pure Water
1A mM C12TAB
0 8mM C12TAB
X 16mM C12TAB
S200 32mM C12TAB
(D0
o 150
LL.

S100
-J

50



0 500 1000 1500

Loading Force (nN)


Figure 2-4. Lateral force as a function of loading force in the presence of surfactant [22].

















inn


0 p

6o polyurethane

A flI


N I
-40
I l

-o silicon nitride

-so- silicon dioxide

-100
0 1 2 3 4 5 6 7 8 9 10 11 12
pH



Figure 2-5. Zeta potential behavior of silica, silicon nitride, cerium oxide (ceria), and polishing
pad (polyurethane) with respect to the pH [32].


- i


4tU


20



m


_ ceriym oxide

\ ,,


I
















E
0
E
3.0
r0
o

r

o 2.5
0



2.0 -
Cn



1.5 III
0 10 20 30 40 50
Amount amino groups (%)


Figure 2-6. Maximum surface concentration of benzoic acid (*) and pyridine (0) obtained by
fitting the adsorption data to a Langmuir-Freundlich equation [31].












SN4


0.9-
0.8
0.7
0.6-
0.5
0.4-
0.3-


0.2'
0.1
0.0


0 S TO I


Load (N)


Figure 2-7. Friction coefficient of silicon nitride ceramic as a function of load in pure water (o)
and silane aqueous solution (*) [33].


C H2O
S* Silanel aq.















































Figure 2-8. The mechanism of high-ionic-strength slurry stabilization by the synergistic mixture
of anionic and nonionic surfactants [34].














36










CHAPTER 3
CMP CHARACTERISTICS OF SILICA AND SILICON NITRIDE

The Shallow trench isolation (STI) chemical mechanical polishing (CMP) process involves

polishing of silica and silicon nitride layer. Therefore, the characteristics of the both materials are

very important for process optimization and overall STI CMP process performance. Besides,

silicon nitride is widely used for various applications such as giant magnetoresistance (GMR)

and ceramic ball bearings making the research on the CMP characteristics of silicon nitride more

significant [35, 36].

There are several abrasives used in STI CMP slurries according to its specific purposes [19,

36, 37]. Among them, colloidal silica is the traditional material, which has long been used for

various applications, and its dispersion stability towards various electrolytes is well documented

[38-41]. A unique property is that it shows high dispersion stability around its isoelectric point

(IEP, pH 2 4), unlike other materials. It has long been believed that hydration force due to

modified water structure at the silica surface or silanol (SiOH) groups give rise to a repulsive

forces, which is responsible for the observed phenomena [39, 42]. Another explanation is that the

formation of a surface gel layer or short polymer-like hairs protruding from the silica surface can

give rise to steric repulsion [43, 44]. In intermediate pH range, silicic acid chains (-Si(OH)2-O-

Si(OH)2-OH) or siloxane bonds (Si-O-Si) are reported to form silica gel relatively easily by

reaction between acidic ionized silanol (SiO-) and neutral silanol (SiOH). At a higher than pH 10,

colloidal silica shows stable dispersion again through electrostatic repulsion between almost

completely ionized silanol groups. As a result, colloidal silica suspensions are stored and used

usually under high pH conditions. When a lower pH application is required, the pH transition is

performed in a very short time period to avoid gelation.









Silica is a promising candidate for the STI CMP due to its high surface quality as

compared to other materials. However, the basic CMP characteristics for silica and silicon nitride,

which consist of the STI structure, are not completely understood. In this chapter, CMP

characteristics of silicon and silicon nitride by colloidal silica abrasives will be discussed with an

emphasis on the electrostatic interactions encountered in the system.

Experimental

The CMP slurry used in this study was Klebosol 1501-50 from Rodel Co. The original

slurry of 30 wt% colloidal silica abrasives was diluted with nano-pure water to 12 wt%. The

slurry pH was measured to be 10.4 after dilution. HC1 and KOH solutions were used for further

adjustment of the slurry pH. The study of lubrication by hydrated cations utilized PL-7 supplied

by Fuso Chemical Co., which is originally at 20 wt% colloidal silica abrasives. It was diluted

with nano pure water to 9.6 wt%/, with a final slurry pH of 7.3. Salt concentration was controlled

to 1 M by adding the proper amount of 5 M salt solution to the slurry. Concentrated 5 M solution

was prepared with analytical grade LiCl and CsCl purchased from Fisher Scientific Co. Silica

and silicon nitride wafers were purchased from Silicon Quest International. Two [tm thickness of

silica thin film was deposited on (100) Si substrate by plasma enhanced chemical vapor

deposition (PECVD) method using Tetra Ethyl Ortho Silicate (TEOS) as a source on (111) Si.

For the silicon niride wafers, 3000 A thickness silicon nitride film was deposited on the 3000 A

silica, which was used as a diffusion barrier on (100) Si by low-pressure chemical vapor

deposition (LPCVD) method using dichlorosilane (SiC12) and ammonia (NH4) as source

materials. IC 1000/Suba IV stacked pads supplied by Rodel Inc. and TegraPol-35 with

TegraForce-5 from Struers Co. tabletop polisher were utilized for CMP purposes. The rotation

speed was controlled to 150 rpm both for the pad and the wafer. Material removal rate (MRR)

was measured using ellipsometry (Woollam EC110 Ellipsometer) by dividing the decrease in









thickness by polishing time. In the present study, MRR reproducibility was within 5 %. Prior

to each polishing step, the pad underwent 30 seconds of conditioning with diamond conditioner.

The actual time for polishing was controlled to 30 seconds. Young's modulus and hardness were

measured by Nanoindentation method using Hysitron Triboindenter purchased from Hysitron Co.

Digital Instruments Nanoscope III atomic force microscope was used for the measurement of

surface roughness of substrates after CMP.

Zeta potential of the slurry was measured by Acoustosizer purchased from Colloidal

Dynamics Co. A variation in the zeta potential values (20 mV) at pH 10.4 was observed for

different batches purchased from slurry supplier. A decrease in zeta potential was also observed

with aging time (10 mV upon 1 year aging). Accordingly, zeta potential values at the same pH

were found to be different depending on the batch and aging time. However for a given sample,

the reproducibility of measurement was found to be within 3 mV over a month period.

Particle size distribution was measured by Coulter particle size analyzer (Coulter

LS13320). After dissolution, the pictures of the substrate surface were taken by optical

microscopy (Olympus BX60).

Relationship between Material Removal Rate and Young's Modulus

The MRR of silica and silicon nitride wafers as a function of polishing pressure is plotted

in Figure 3-1. In this experiment, original slurry (30wt% solids loading) was used without further

dilution. The MRR showed a linear relationship with polishing pressure, as predicted by the

empirical Preston equation [45]:

As
MRR = KP (3-1)
SAt

where, Kp is Preston coefficient, P is polishing pressure, and As is the relative travel between

glass surface and lap over in which the wear occurs (platen speed) during time interval At [45].









The MRR of silicon nitride was determined to be lower than silica. In CMP of Si-based

materials such as silica and silicon nitride, it is well known that water plays a significant role,

because no material removal occurs in non aqueous medium. It is commonly believed that water

attacks and breaks the siloxane bonds by the following reaction:

Si O Si + H20 = SiOH + SiOH (3 2)

It has been reported that the hardness of silica decreases to around 50% of the original value in

aqueous systems [46, 47]. The above reaction is believed to be controlled by the diffusion of

water in silica, which in turn affects surface hardness.

There have been several attempts to explain MRR theoretically [45, 48]. One of them is

Cook's model, assuming Hertzian penetration [45]:

1 As
MRR = P S(3-3)
2E At

where, E is the Young's modulus of the material. Considering that the modulus is the resistance

of the material to tensile or compressive deformation, above equation indicates that material with

high modulus should be harder to polish. A more elaborate model incorporating chemical effects

was proposed by Chi-Wen and co-workers [48]:

1 1 As
MRR = C( + -) P (3-4)
E, E At

where, C is the coefficient accounting for chemical effect of a slurry and other properties of

CMP consumables, Ea and Ew are the Young's modulus of abrasive particle and substrate,

respectively. Trends in experimental results with substrates of different moduli were in

aggrement with those predicted by Equation (3-4).

To evaluate the correlation between MRR and mechanical properties of substrate materials,

Young's modulus and hardness of both substrates were measured by the nano-indentation









method and are summarized in Table 3-1. The MRR and Young's modulus ratio indicated a

correlation between MRR and mechanical properties of the material. However, according to this

explanation, silicon nitride cannot be polished by silica abrasive particles, since silicon nitride

has a higher hardness than silica, in contrast to experimental evidence. In reality, the formation

of a thin silica layer (around 1 nm) on the silicon nitride surface by spontaneous oxidation

represented by the equation below and is expected to influence the polishing characteristics of

silicon nitride [23, 31, 49]

Si3N + 6H20 = 3SiO2 + 4NH3 (3- 5)

It has been reported that the rate-limiting step for the above reaction is the breakage of Si-N

bonds [50], with relatively faster breakage of Si-O bonds due to diffusion of water. In other

words, the reaction of water with silicon nitride for breaking the Si-N bond is slower than water

diffusion. As a result, the thickness of the newly formed silica layer on silicon nitride will be

very thin compared to that of the silica substrate, thereby resulting in different MRR of the two

substrates. Theoretically, Young's modulus reflects the bond strength of the material on an

atomic scale [51]. In other words, a higher modulus means stronger bonds, which will be harder

to break.

Role of Electrostatic Interactions on MRR

It has long been observed that MRR is dependent on the pH of the slurry in various

polishing processes including CMP. As was discussed by Choi and co-workers, electrostatic

interactions can influence the CMP performance. However, systematic approaches and

quantitative analysis to explain the effect and modulation have not been attempted until now.









Effect of pH

One of the best ways to modulate the electrostatic interaction is to change pH of the slurry.

Colloidal silica slurry is the best candidate for this purpose, since it shows stable dispersion

throughout a wide pH range, if only the pH was adjusted just before polishing. To investigate the

effect of electrostatic interaction on CMP performance, the MRR for both substrates as a

function of slurry pH was measured and plotted in Figure 3-2. Particle size distribution at pH 2

and 10.4 in Figure 3-3 confirmed that there was no measurable coagulation of silica particles at

pH 2.

MRR as a function of pH reached a maximum as slurry pH is reduced. At high pH beyond

11, MRR steeply increased for silica and remained constant for silicon nitride. The CMP results

of silicon and silica as a function of pH were reported by several authors [52-54]. Choi et al.

attributed the increase in MRR at lower pH to the electrostatic attraction between the oppositely

charged silica substrate and silica abrasive particles, and a higher MRR at higher pH to increased

softening of silica induced by its high solubility at higher pH. According to their report, the

electrostatic force between silica particles and substrate showed a maximum around 0.4 mN/m

(force/radius of particle) at pH 10.4. The contact area of the CMP pad and the substrate was

reported to be around 1% due to the asperity characteristics of the pad materials employed in

their study at the same pH [55]. Assuming that half of the individual abrasive particle will be

embedded in the substrate surface and the other half of the particle will be captured by pad

asperities during the CMP process, the contact area will yield the number of particles in contact

with the substrate. If 1% of a 1 x 1 inch wafer is in contact with the abrasive particles, then there

will be approximately 109 particles of diameter of 90 nm in the system. The total electrostatic

force is calculated to be 18 mN. According to experiments in the present study, if one assumes

that there is no electrostatic force contribution at pH 3 (due to its nearly zero value of zeta









potential), a pressure caused by repulsive force of 6.85 N on 1 x 1 inch wafer, is required to

make a difference in MRR. This is more than two orders of magnitude difference in electrostatic

force contribution between the abrasive and the wafer. It is, however, possible that induced

repulsion by electrostatic interactions may contribute to lubrication effects. According to Choi,

there was approximately 25% decrease of frictional force between colloidal silica abrasives and

the wafer when the slurry pH was increased from 2 to 10.4. Mahajan also reported that the

frictional force between pad and the wafer decreased at higher pH due to increased electrostatic

repulsion between them [56].

It is well known that in the case of boundary lubrication, friction follows the equation for

interfacial sliding, as proposed by Tabor et al.[57].

Ff,,hcon = SA (3-6)

where, Ff,,,cton is a frictional force, Sc is a critical shear stress that depends on the details of the

interfacial region, and A is the contact area. It is not clear which term is affected by the

electrostatic interaction for the current system. However, it seems reasonable that if electrostatic

repulsion between the abrasive and substrate is high, critical shear stress (Sc) will be reduced,

resulting in overall reduction in the frictional force. On the other hand, surface layer

characteristics can also change upon a shift in pH, resulting in changes in contact area (A)

between the pad and the substrate. Yeruva reported that there was no consistent evidence that the

Young's modulus of the pad, which is directly related to the contact area, changes with solution

pH.

Recently, Taran et al. have reported that a lubrication effect between silica particles and the

substrate resulted in reduced lateral force at high pH above 9.6, using lateral force microscopy

[58]. Below pH 9.6, there was no noticeable change. They correlated their observations with









solubility of silica and formation of surface gel layer, which is believed to form at high pH due to

high solubility [58]. It seems likely that the lubrication phenomena may play a role in explaining

low MRR at high pH, but it is not possible at present to explain high MRR below pH 8.6.

Another possibility is that the electrostatic forces between particles can change the number

of abrasive particles participating in the polishing process, depending upon their

dispersion/coagulation characteristics. It has been generally known that MRR is almost linearly

proportional to solids loading of the slurry [59, 60]. Zeta potential of the abrasive particle will

produce electrostatic repulsive forces that will resist the particles to come within a certain

distance of the substrate resulting in limited number of particles participating in polishing at a

certain pH. The repulsive force can be calculated using simplified Poisson-Boltzman equation

[61]

F /R = 2iSeegCg, e (3 7)

where, F/R is the electrostatic force per particle radius, K is the Debye-Huckel parameter, Fo is

surface potential, and D is the distance between particles which is assumed to be 1 nm. The

absolute force value can change as a function of distance, but the trend should be similar. Zeta

potential was assumed to be the same as the surface potential, since there were no specific

adsorbing ions in the slurry. Figure 3-4 shows the measured zeta potential of silica and the

corresponding electrostatic force between abrasive particles calculated from the potential as a

function of pH (also plotted in Figure 3-2). At pH around 3 (IEP of silica), the electrostatic force

leveled off and approached zero and MRR for silica also reached a maximum value at pH 3. In

the intermediate pH range (3 10), MRR and the electrostatic force were inversely proportional

to each other.









At pH above 11, the MRR of silica showed a sudden increase, probably related to the

solubility of silica. However, the MRR of silicon nitride, which has a lower solubility than silica,

showed the same trend as electrostatic force. Overall, it appears that there exists an inverse

correlation between the MRR and repulsive electrostatic forces between the abrasive particles.

The zeta potential of the substrate and colloidal silica should be similar, since both

materials are amorphous silica, therefore it may be safe to assume that the calculated electrostatic

force also represents the trend in the force between abrasive particles and substrate. It is clear

that the electrostatic forces induced by zeta potential of various materials has a significant effect

on MRR in terms of (i) opposing force against polishing pressure or (ii) number of particles

participating in the CMP process.

Effect of Salt Addition

It is well known that various salts reduce the surface charge of the particles in a colloidal

system, decreasing the electrostatic repulsion and thereby promoting their coagulation by

attractive van der Waals interactions [38, 39]. The minimum concentration of salt causing

coagulation of particles is called the critical coagulation concentration (CCC). This phenomenon

can be utilized to modulate the electrostatic force in the CMP process. Among various salts,

monovalent ions are most suitable for this purpose in terms of controllability, since multivalent

ions have far lower CCC than monovalent ions. Allen and co-workers have reported that CCC of

NaCl for colloidal silica was around 0.4 M and that of CaC12 was around 1 mM, at pH 9. CMP

was conducted as a function of NaCl concentration added to the slurry. The MRR for both

substrates and calculated electrostatic force between abrasive particles from zeta potential values

are plotted in Figure 3-5.

The first thing to be monitored is the coagulation of particles whenever salt is added into

slurry. Figure 3-6 shows the particle size distribution as a function of NaCl concentration. Below









0.5 M NaC1, the particle size maintained a narrow mono size distribution. When the

concentration reached 0.5 M, gelation occurred and particle size distribution showed multiple

peaks. It is not clear from Figure 3-6 if there is coagulation, since the additional peak(s) from

coagulation are not noticeable due to the multiple peaks from gelation. It is very likely that there

is some degree of coagulation at that concentration. Gelation usually occurs at intermediate pH

and high salt concentration in a colloidal silica system, and it is different from coagulation.

Gelation is reversible, i.e. the dispersion stability can be restored simply by stirring or dilution,

but if the coagulation occurrs, it is not usually reversible. In gelation, silica particles form a

network by siloxane (Si-O-Si) bonds. In coagulation, they do not form any network, but they

simply collide with each other by Brownian motion leading to very strong attractive van der

Waals interactions. It is not known how gelation of abrasive particles affects the CMP

performance. A colloidal silica slurry adjusted to neutral pH and kept for some time to promote

geltation without any salt can be a good candidate to isolate such effects.

Below a salt concentration of 0.5 M, NaCl addition to the polishing slurry showed the

same trend in MRR change as the pH change. There was a steep decrease in the MRR after the

salt concentration exceeded the CCC (0.5 M NaC1) for silica, however. The silicon nitride

substrate did not show such dramatic change. It has been reported that at fixed solids loading, the

MRR decreases as a function of particle size after reaching a critical size of particles [62, 63].

This leads to the explanation of how the coagulation might affect MRR. At a fixed solids loading,

coagulation leads to two possible effects, (i) reduction in the number of abrasive particles

participating in the polishing process thereby decreasing the contact area between particles and

substrate, (ii) increased penetration depth due to size enlargement resulting in higher MRR. As

was discussed by Yeruva, optimal indentation depth is determined by the thickness of the









modified surface layer of silica caused by reaction with water, which is believed to be on the

order of nm in thickness [55]. Besides, the optimum mean particle size resulting in maximum

MRR was reported to be around 75 nm experimentally [63]. In the present study, the

agglomerated particle size is larger than 100 nm, hence a decrease in MRR and poor surface

finish are expected and experimental results confirmed these predictions.

Choi reported that at intermediate salt concentrations, Stober silica slurry showed a broader

distribution with a larger particle size accompanying the MRR increase, and was attributed to

reduced electrostatic forces and increased particle size due to coagulation [64]. At a higher salt

concentration, they reported low MRR and high roughness values attributed to coagulation of the

silica abrasive particles. In the present study with a colloidal silica slurry, the increase in MRR

can solely be attributed to reduced electrostatic repulsion, since there was no particle size

increase.

Measured surface roughness values indicated that up to 0.3 M NaC1, there was not much

difference in surface roughness (Figure 3-7). However at 0.5 M, a rough surface with low MRR

on silica but not on silicon nitride was observed. On the silicon nitride substrate, the coagulation

of abrasive particles does not seem to have as high an effect as on the silica probably due to the

higher hardness of silicon nitride substrate as compared to the silica abrasive particles.

Salt addition has been reported to increase frictional force between the pad and substrate as

also observed by Mahajan [56]. This suggests that coagulation of abrasive particles is a major

factor in determining frictional forces, which in turn impact MRR.

In order to further establish a correlation between the MRR and electrostatic forces, the

MRR for both materials is plotted in Figure 3-8 as a function of electrostatic repulsive force

between colloidal silica abrasive particles at different levels of pH and salt concentrations.









Except under the extreme conditions such as pH 2, 11.5 and NaCl concentration of 10 mM,

where calculated electrostatic force was not sensitive to experimental variables, an inverse linear

relationship was observed between MRR and electrostatic forces.

Parameters Affecting Surface Finish in STI CMP

Figure 3-9 shows the surface roughness of silica and silicon nitride substrates as a function

of slurry pH. Selected surface morphologies and roughness profiles of the silica and silicon

nitride after CMP at pH 10.4 and 11.5 for both materials are plotted in Figure 3-10. CMP by

colloidal silica slurry improved the roughness of both materials below pH 11. 5 and silica

showed higher roughness values than silicon nitride over the entire pH range examined in this

study. At pH 11.5, CMP resulted in poor surface finish for both materials but the increase of

roughness was higher for silica. Scratches from the CMP process were not observed on either

substrate.

This variation of roughness follows exactly the same trend as the silica solubility results by

Iler [65]. It is known that the solubility of silica shows a steep increase in the basic pH condition.

Iler reported about a three orders of magnitude increase in silica dissolution rate as the pH value

changed from 2 to 11 [65]. The increase in solubility is believed to be due to the hydroxyl ion

(OH-) acting as a catalyst for attack by water on the siloxane (Si-O-Si) network. Specifically,

hydroxyl ions create an excess of electrons resulting in a higher negative surface potential and

consequently more attacks by H30 [21]. Therefore, it has been widely believed that the high

dissolution rate of silica at high pH is responsible for the high MRR [53, 54]. The effect of

solubility on surface roughness has not been well understood. It should be noted that solubility of

silica is known to depend on the curvature of the silica surface [66]. Hulett et al. reported that the

convex surface of colloidal silica shows higher solubility than the concave one, and a smaller

radius of curvature exhibits higher solubility [66]. This implies that surface convex impurities









will dissolve faster than flat substrates. However, this prediction is contrary to our experimental

observation of the effect of solubility on MRR and surface roughness, and requires further

investigation.

To evaluate the effect of solubility of silica and silicon nitride on CMP performance,

dissolution rate was determined by measuring the thickness of both substrates immersed in a

0.1M (pH 13) NaOH solution for 12 days without stirring (Figure 3-11). Surface roughness of

the substrates before and after dissolution is presented in Figure 3-12. The dissolution rate of

silica was three orders of magnitude higher than that of silicon nitride most probably due to

higher bond strength of the latter. Even though the experiment was conducted at pH 13, the

magnitude of dissolution of both substrates was relatively low. However, in a real CMP process,

dissolution can be increased by the imposed pressure resulting in higher tensile stress created by

the abrasive particles as they abrade silica surface. Nogami and co-workers reported a 50%

increase in solubility when 30 MPa compressive stress was applied compared with the stress-free

condition [47]. Additionally, when abrasive particles abrade the surface, the temperature can be

higher due to heat generated by friction. It has been reported by Iler that solubility of colloidal

silica increased by more than ten times at 200 C than at room temperature [65]. However,

incorporation of all those factors still gives a far less dissolution rate than the MRR increase at

pH 11.4 for silica.

Regarding this apparent discrepancy, it should be noted that the attack of hydroxyl ions

will be higher at higher pH resulting in a softer layer, which can be removed easily, and is prone

to damage by abrasion. Consequently, the attack of hydroxyl ions increases the solubility and

promotes formation of a softer layer on the substrate at high pH. The dissolution of silica itself

does not seem to play a bigger role in determining MRR. The extent of hydroxyl ion attack will









also be dependent on the bond strength, and according to Young's moduli of the materials, this

may explain the reason for low MRR and dissolution rate of silicon nitride.

Figure 3-12 illustrates the surface morphologies of the two substrates after dissolution at

pH 13 for 12 days. There was very small increase in surface roughness for both of the substrates.

The inverse pyramidal-type etch pits observed in Figure 3-13 are common phenomena when

highly concentrated alkaline solution is used for etching silicon in the micromachining of silicon

substrates [67-69].

The reason for the anisotropic etching is different reactivities of certain crystal planes of

silicon. In other words, anisotropic etchants etch much faster in one direction than in another,

which is usually (111) planes of silicon. Therefore, anisotropic etching of (100) silicon by

alkaline solution results in the inverse pyramidal-type etch pits, as was observed experimentally.

Since the thin film used in this research was deposited on (100) silicon, the silica film will have a

similar atomic arrangement as the underlying silicon. It has been well observed in silicon

anisotropic etching that when a dilute alkaline solution (20 wt%) is used, the etching produces

high surface roughness. Palik et al. reported that the high surface roughness is attributed to the

formation of hydrogen bubbles acting as a pseudomask, thus inhibiting uniform etching [69]. In

silica, the overall reaction of the dissolution can be described as follows:


-SiOH + H20 +OH =Si(OH) + IH2 (3-9)
2

Gas bubbles were observed during dissolution experiments and are believed to be hydrogen gas.

For the reaction shown in Equation (3-9) to occur, the nucleation of hydrogen bubbles is a

dominant step and it is much easier to nucleate them on high energy sites giving rise to surface

defects.









Salt Mediated Lubrication

Donose and co-workers have reported that various cations adsorbing on silica from

electrolyte solutions can induce lubrication through the formation of a hydrated cation layer [70].

Due to the difference in hydration enthalpy of different cations, resultant lubrication was

different for each added salt. Similar phenomena have been reported by Raviv and Klein by a

modified surface force apparatus [71]. They measured the shear force between mica surfaces and

concluded that hydration layers of adsorbed cations act as a highly efficient boundary lubricant.

Their research was mostly done by lateral force microscopy and the macroscopic effect on CMP

was not investigated.

Figure 3-14 shows the lateral force as a function of loading force in the presence of various

salts such as LiC1, NaCl and CsCl reported by Donose and co-workers [70]. According to their

results, every salt showed higher lubrication effect than pure water. The thickness of the

adsorbed cation layer increases with increasing electrolyte concentration. Highly hydrated

cations such as Li can form a thick and soft layer resulting in higher lubrication than poorly

hydrated cation such as Cs+. It was observed that the degree of lubrication followed their order of

hydration, which is Li+ > Na+ > Cs+. Schematics shown in Figure 3-15 illustrate this concept.

For pure water, at least one layer of water molecules are bound to the silica surface, but this layer

is relatively thin and firmly adsorbed to the silica surface resulting in rigid interface. In the

presence of an electrolyte solution, there is a thicker hydration layer than pure water. The model

suggested by Raviv et al. states that the cations surrounded with water molecules are very hard to

remove and remain fluid like in a lateral direction and promote lubrication. It is well known that

smaller Li ion has the highest hydration enthalpy and hydrated radius among various









cations[72]. Accordingly, Li ions have a thicker and more effective lubricating layer on silica

surface, while Cs+ ions have a thinner and less effective lubrication layer.

To investigate how the variations in lubrication affects the real CMP performance, CMP

was conducted as a function of applied polishing pressure using three slurries with no salt, 1 M

LiC1, and 1 M CsC1. The salt concentration was selected corresponding to the results reported by

Donose and co-workers [70]. The particle size was measured to assess if the selected salt

addition causes any coagulation of the abrasive particles (Figure 3-16). When appropriate

amounts of 5 M LiCl and 5 M CsCl were added to change the salt concentration, there was no

particle size increase initially up to about 10 minutes after mixing. As time passed, gelation took

place slowly and the peak height of the particle size decreased and the size distribution became

broader. While it is not well understood how the gelation affects the CMP performance, CMP

was performed 5 minutes after the mixing of 5 M salt solution to avoid the possible effect of

gelation and ensure uniform mixing of added salt. Surface roughness measurement showed that

RMS surface roughness was around 0.15 nm for all the conditions at the same polishing pressure

and there was no increase from salt addition. Therefore, it appears that the variation in MRR is

due to the effect of salt on material properties and not necessarily from gelation and coagulation.

Figure 3-17 shows the variation in the MRR with and without 1 M LiCl and CsCl as a

function of polishing pressure. Increase in the MRR with added salt suggests that electrostatic

interactions play a dominant role in polishing. The MRR of the silica substrate using a slurry

with 1 M LiCl is lower than that of 1 M CsCl showing results in agreement with those from

lateral force microscopy measurements. Considering the same electrolyte concentration in both

experiments, the electrostatic forces should be similar. The lubrication effect of individual









particles should reduce the MRR, but increase in the number of abrasive particles due to reduced

electrostatic repulsive forces between them seems to have resulted in overall higher MRR.

In summary, CMP performance using colloidal silica slurry in a silica and silicon nitride

system revealed that Young's modulus of the substrate material is more likely the reason for the

differences in their MRR, with electrostatic repulsive force imposed by pH change in the slurry

playing a dominant role. The electrostatic interaction was validated by monovalent salt addition

to the slurry. A linear relationship between the MRR and electrostatic forces implied that such

repulsive interactions probably resulted in governing the number of particles engaged in the

polishing process. Dissolution rates were measured by immersing substrates into 0.1 M NaOH

solution for 12 days and the results showed that dissolution of silica was much higher than

silicon nitride, however, the rate of dissolution was too low to make any significant difference in

the MRR. It seems that the attack of hydroxyl ions at higher pH is responsible for poor surface

finish and higher MRR due to the formation of a softer top layer. Dissolution in alkaline

solutions produced a poor surface finish due to nucleation of hydrogen gas bubbles.

The effect of the nature of added ions on CMP performance was also investigated. The

Lubrication effect of hydrated cations was determined not to be a dominant factor in MRR.

However, a slurry with LiCl showed lower MRR than one with CsC1, which suggests that the

lubrication of the hydrated cations is playing a limited role in determining the MRR.













Table 3-1. Young's modulus, hardness measured by nanoindentation method, material removal
rate (MRR), ratio of MRR (CMP pressure of 7 psi), and ratio of Young's modulus for
silica and silicon nitride.


E (GPa) H (GPa) MRR (A/min) MRRsio2/MRRsi3N4 Esi3N4 /ESiO2
SiO2 84.6 3.0 8.5 0.3 4696
3.4 2.1
Si3N4 176.6 + 2.5 23.5 + 1.0 1382













9000
8000- pH 10.4 0 SiO2
E 7000- SiN,
) 6000-
S5000-
S4000-
0
E 3000
0 2000-
1000-
o 0
-1000 i i i i i
0 2 4 6 8 10 12 14 16 18
Pressure (psi)


Figure 3-1. Variations of mateiral removal rate (MRR) for silica and silicon nitride substrate as a
function of applied pressure by using undiluted (30 wt%) colloidal silica slurry at pH
10.4.













0.10
2200
S2000 -0.08
E z
S1800- E
MRR SiO 2
2 1600 -- MRR Si3N -0.06
S1400- 0
(0 L.
00
> 1200- 0.04 o
a 1000- A Force -
-* -0.02 2
-co 800 -
a 600- 0.00
2 400
2 3 4 5 6 7 8 9 10 11 12
pH

Figure 3-2. Variations of MRR of silica and silicon nitride substrate and calculated electrostatic
force between two abrasives as a function of pH of the diluted (12 wt%) colloidal
silica-based slurry (Klebosol 1501-50).






























O -

-2 ..- . 1
0.1 1
Particle Size (prm)


Figure 3-3. Particle size distributions of colloidal silica slurry at two different pH conditions.

















0.10
0-

-15 0.08 E
-15-
Z

> -30- 0.06 a
E

-45 o
S\ / 0.04 u
0
0--60- V
(D 0.02 -
N 0
-75- (D
.LU

-90- .00

1 2 3 4 5 6 7 8 9 10 11 12
PH

Figure 3-4. Zeta potential of colloidal silica slurry and electrostatic force between silica abrasive
particles. Force was calculated from the zeta potential values by constant surface
charge model. The distance between abrasives was assumed to be 1 nm.














2400
pH 10.4 Force 25
2200-
E E
S2000 z
1800 A
1600-
FU 15 2-
> 1400- 0
10 U_
), 1200- U MRRSiO o
-F 1000 0 MRR Si N, -
r--0
a) 800 \-
ro -5 a)
600-
400

101 102
NaCI Concentration (mM)

Figure 3-5. Variations of MRR and calculated electrostatic force between two abrasives as a
function of slurry NaCl salt concentrations in the slurry at pH 10.4.















16

14 -- No Salt

\ I 0.1 M NaCI
10.3 M NaCI
-- 0.5 M NaCI
S10-
E
d 8-

6-
4
2- V '




I 1 '
0.1 1
Particle Size (p(m)


Figure 3-6. Particle size distributions of colloidal silica slurry (Klebosol 1501-50, 12 wt%) as a
function of salt concentrations at pH 10.4.














0.5-


0.4-


0.3-


0.2-


0.1-


0.0 1


SiO2
Si3N4


I-.









0 0.3 0.5
NaCI Concentration (M)


Figure 3-7. Surface roughness of silica and silicon nitride substrate after CMP as a function of
added salt (NaC1) concentration at pH 10.4.













2400 (a)

E 2100-

a 1800-

S1500-
0 SiO
> 2
O 1200- 0 Si3N4

n," 900-

a 600

300-_
0.00 0.02 0.04 0.06 0.08 0.10
Electrostatic force/R (mN/m)



r (b) pH 10.4
S2400

S2100

o 1800

-F 1500
> Sio
0 2
E 1200- Si34

__ 900

I 600-

0 5 10 15 20 25
Electrostatic Force/R (mN/m)



Figure 3-8. Material removal rate of silica and silicon nitride as a function of repulsive
electrostatic force between silica abrasives: (a) pH effect and (b) Salt (NaC1) addition
at pH 10.4.


















0.55-
0.50-
0.45-
0.40-
0.35-
0.30-
0.25-
0.20-
0.15-
0.10-
0.05


* SiO2
* Si3N4


m - -


3 4 5 6 7 8 9 10 11 12 13
2 3 4 5 6 7 8 9 10 11 12 13


pH


Figure 3-9. Surface roughness of silica and silicon nitride substrates after CMP as a function of
slurry pH.











SiO2, pH 10.4, RMS Roughness: 0.24 nm







70 2.00 4.00 6.00


(b) SiO2, pH 11.5, RMS Roughness: 0.48 nm







o 2.00 4.00 6.00


Si3N4, pH 10.4, RMS Roughness: 0.14 nm







0 2.00 4.00 6.00


(d) Si3N4, pH 11.5, RMS Roughness: 0.22 nm







V0 2.00 4.00 6.00
JJM

Figure 3-10. Surface morphologies and profiles of substrates from two pH conditions; (a) silica
at pH 10.4, (b) silica at pH 11.5, (c) silicon nitride at pH 10.4, and (d) silicon nitride
at pH 11.4
















2000-
0.1 M NaOH (pH 13)
1800 SiO2 Dissolution rate: 0.109 A/min

1600

1400
Co,
C,
a) 1200

o 1000-
r SiO,
800-

> 600
E
u, 400
200 Si3 N Dissolution rate: 0.966 x 10-3A/min
0 I I I
0 2 4 6 8 10 12
Days

Figure 3-11. Material thickness change of silica and silicon nitride substrates as a function of
immersed time in pH 13 NaOH solution. Ellipsometer was used to measure thickness
change. Calculated dissolution rates were also shown.












SSiO2, RMS Roughness: 0.338 nm


o-.T


2.00


4.00


SSiO2 at pH 13 for 12 days, RMS Roughness: 0.434 nm


o -MV YX >MvTA^W"^A4


SSi3N4, RMS Roughness: 0.204 nm


0 tr T


2.00


4.00


Si3N4 at pH 13 for 12 days, RMS Roughness: 0.241 nm


o ~r t.>.


2.00


4.00


rF, n


Figure 3-12. Surface morphologies and profiles of substrates before and after etching in pH 13
NaOH solutions; (a) bare silica (b) silica after 12 days, (c) bare silicon nitride, and (d)
silicon nitride after 12 days.
































^i
'-I





















S












Figure 3-13. Etch pits formed on (a) silica and (b) silicon nitride substrate immersed in 0.1 M
(pH 13) NaOH solution for 12 days.










*PorwfRP

1I0 Wum r





0 ---- i -- i ----


0 100 200 300 400 5
Loading firce (nN)

Figure 3-14. Lateral force of a 6.8 jtm silica particle interacting with a silica substrate in pure
water and CsCl, NaC1, and LiCl solutions of 1 M: dependence of friction on the
applied load at a fixed scan rate of 2 jtm/s [70].








Pure Water


Tj i_ Lubriec tlon
iU \ :tT / \ ^ / L


l.OM CICJ
Lubriettion




1AOM LiCI

I~ Am37$!


* H /HXQ *LT *VC


Figure 3-15. Schematic representation of the hypothetical frictional mechanisms [70].






























0.1 1
Particle Size (pm)


Figure 3-16. Particle size distributions
1 M LiC and 1 M CsC1.


of colloidal silica slurry (Fuso PL-7) without salt and with














240-
E 210-
E pH 7.3 Fuso PL-7
5 180- 0 1M CsCI
-1- m 1 M LiCI
U 150 -LiI
f No Salt
-120
0
E 90-
(D
Or 60-

30

r 0
II II I I I
0 2 4 6 8 10 12
Pressure (psi)


Figure 3-17. Material removal rate of silica substrates by CMP using diluted (9.6 wt%) colloidal
silica slurries (PL-7) without salt and with 1 M LiCl and 1 M CsCl as a function of
applied polishing pressure.










CHAPTER 4
ROLE OF SURFACTANTS IN DEVELOPING SELECTIVE PASSIVATION LAYER IN CMP

In this chapter, a surfactant mediated passivation approach to increase STI CMP selectivity

is discussed.

Selective adsorption of a surfactant is necessary to develop selective passivation in CMP.

It can be achieved if there is an adequate difference in the surface charge characteristics of the

substrates. This concept has been successfully used to achieve selective coating of surfactants in

mineral flotation [73, 74]. Interactions between a solid surface and charged polar head of the

surfactant molecule determine the adsorption strength and the resultant adsorption density. In

CMP, surfactants have been used not only to disperse abrasive particles but also to create

lubricating layers, yielding passivation against polishing. In the present study, an anionic

surfactant, SDS, was used to create a selective passivating layer only on the silicon nitride and

not on the silica substrate. It is well known that isoelectric point (IEP) of silicon nitride is higher

than silica resulting in less negative potential for silicon nitride above the IEP [23]. The

concentration of SDS was adjusted to 16 mM, twice the critical micelle concentration (CMC),

which has been shown previously to yield stable dispersion of silica abrasives in a CMP slurry

[75].

For adsorption studies, silica from Geltech Co. and silicon nitride from Ube Co. (SN-E10)

were used to simulate silica and silicon nitride substrates. The particle size of silica was

measured to be around 0.53 [m by Coulter, and that of silicon nitride, which was measured by

centrifugal sedimentation was reported to be around 0.5 [im by the manufacturer. Their specific

surface areas were measured to be 8.1 m2/g and 10.4 m2/g, respectively by Quantachrome Nova

1200, BET surface area measurement technique. The specific surface area of abrasive silica

particles was measured to be 34 m2/g by Quantachrome Autosorb 1C-MS. The Phoenix 8000









UV-Persulfate TOC Analyzer was used to measure the SDS concentration. 99% sodium dodecyl

sulfate (SDS) surfactants from Acros Organics Co. and Fisher Scientific Co. were used as

received. 98% dodecyl alcohol from Eastman Kodak Co., 95% Sodium tetradecyl sulfate from

Acros Organics Co. and Tween 80 from Fischer Scientific were also used, as received.

High Selectivity Slurry Using Surfactants

The addition of SDS to the slurry was found to result in a lower value of MRR of silica and

silicon nitride in the entire pH range investigated in the present study (Figure 4-1). However,

significant increase in selective polishing of silica was measured below pH 3, yielding a

selectivity of 25 as compared to state-of-the-art ceria abrasives of 5. The silicon nitride surface

appeared to be fully passivated with the surfactant layer at a pH below its IEP of pH 4.5, with

minimal effect on silica CMP. The surface quality of substrates plotted in Figure 4-2 indicated

that surfactant addition did not cause any additional defects measured as root mean square

(RMS) roughness.

To understand the reasons for the observed selectivity, zeta potential and adsorption

density measurements were conducted as a function of slurry pH (Figure 4-3). The IEP of silicon

nitride and silica substrates were measured to be about pH 4.5 and pH 2.2, respectively. The

difference in the IEP results from the different surface groups constituting each material. As

mentioned earlier, acidic silanol (SiOH) are the major surface groups on silica, while the silicon

nitride surface consists of basic amine (Si2NH) and acidic silanol (SiOH) groups [23]. These

surface groups can acquire charge in aqueous solution according to following reactions:

SiOH = SiO + H (4-1)

Si2NH + H = SiNH+ (4-2)









Consequently, zeta potential of silicon nitride is more positive due to the positively charged

amine groups on its surface.

The adsorption density of SDS was measured to be higher on silicon nitride than silica at a

pH below their IEP. This is attributed to the resultant electrostatic interactions between the

substrate and surfactant molecules. At pH 2, the zeta potential of silicon nitride was measured to

be +40 mV, whereas, that of silica was around +3 mV. Accordingly, the adsorption density on

silicon nitride was determined to be six times higher than on silica resulting in complete

passivation of the former. At pH values above the IEPs for both materials, there was still

measurable adsorption on both materials, however, the adsorption density on silicon nitride was

higher probably due to more positive sites on silicon nitride from surface amine groups. There

have been several reports of SDS adsorption on the negatively charged silica surface. Hydrogen

bonding and sodium ion mediated surfactant bonding are proposed as plausible mechanisms [76,

77].

In order to measure the effect of surfactant concentration on selectivity, polishing was

conducted as a function of added surfactant concentration (Figure 4-4). The MRR for both silica

and silicon nitride started to decrease upon SDS addition and reached a minimum above 16mM.

The maximal decrease in the MRR for silica was around 20% from its original value, and that for

silicon nitride was more than 90%, resulting in 10 times higher polishing selectivity than without

surfactant addition. No further change in the MRR or selectivity was observed once the added

surfactant concentration exceeded 16mM. It has been reported that once the equilibrium

concentration reaches CMC, no more adsorption changes are observed due to electrostatic

repulsion between adsorbed micellar aggregates and free micelles in solution [78].









Surfactant Mediated Boundary Layer Lubrication for Selective Polishing

Vakarelski et al. have shown that beyond the CMC of the cationic surfactant

(dodecyltrimethylammonium bromide, C12TAB), there was no further decrease in the lateral

force on silica substrate [22]. Consequently, it is hypothesized that the maximum decrease in the

MRR will occur when the bulk concentration reaches the CMC of SDS (around 8mM) [79].

However, in the present study, two times higher concentration of surfactant than the CMC was

required to achieve maximum selectivity. The measurement of SDS adsorption on the CMP

slurry as a function of pH showed that about 91% of added (16mM) SDS adsorbed on the

abrasive particles at pH 2 as shown in Figure 4-5. The area per molecule using the Gibbs

adsorption equation, was calculated to be around 70 A2/molecule, which is higher area per

molecule than the literature value of 53 A2/molecule [79] at the liquid/gas interface.

The possible reasons for the higher dosage of surfactant than expected are that the

surfactant adsorption does not reach true equilibrium conditions due to process conditions

encountered in CMP. This phenomenon may also be related to the dynamic aspects of surfactant.

The reported 72 for SDS is around 2.32 x 10-3 s [80]. However, according to Patist and co

workers, when 15 mM SDS was used for foaming experiments, the dynamic surface tension

decreased as a function of bubble life time until it reached the saturation after about two seconds

[81]. Recently, Philipossian et al. have reported the mean residence time of colloidal silica slurry

between pad and substrate to be of the same order of a few seconds under the present

experimental conditions [82]. Assuming that other conditions are similar, the mean residence

time in our study is expected to be 2 3 seconds. Considering that these two numbers are

comparable, migration of surfactant to the newly formed substrate surface may be limited due to

the high speed rotation of pad and wafer in CMP.









The adsorption free energy is the driving force for surfactant adsorption and is the sum of

various molecular interactions [78]. In the current study, it can be categorized into two categories,

(i) interactions between the polar head of SDS and the surface through electrostatic and hydrogen

bonding, and (ii) hydrophobic interactions between alkyl chains of adsorbed SDS molecules. By

using the measured adsorption density plotted in Figure 4-5, and the radius of the SDS micelle

(20 A) [79], the adsorption free energy of SDS on silica abrasives was calculated to be -3.58

kcal/mol at pH 2 using modified Stern-Graham equation [78].


F = 2rCo exp Gd (4-3)


where, F is the adsorption density, r is the effective radius of the adsorbed ion, k is the Boltzman

constant, Co is the bulk concentration, Tis 298 K, and AG is the adsorption free energy. The

electrostatic component of the adsorption free energy was calculated to be -0.76 kcal/mol using

ze V/f [78], where, z is the valency of the adsorbate species, e is the charge of the electron, and the

45 is the potential at the 6 plane (assumed to be the zeta potential). These calculations indicate

that significant adsorption of the surfactant on the abrasive particles is more favorable and may

act as an additional energy barrier.

Optimization of High Selectivity Slurry

It is clear from the above discussion that the adsorption density of surfactant molecules on

the substrate is an important factor in determining the slurry selectivity. In order to reduce the

required dosage of the surfactant, longer alkyl chain length surfactants were examined, since it

was expected to exhibit better lubrication effects at a lower amount of added concentration. This

is attributed to the formation of more compact surfactant layers [75]. The MRR and polishing

selectivity as a function of alkyl chain length of the sodium alkyl surfactant are plotted in Figure









4-6. The surfactant concentration was selected to be twice the CMC value to compensate for the

loss of surfactant due to adsorption on silica abrasive particles. As expected, SDS with longer

alkyl chain length (C12) resulted in higher MRR decrease for silicon nitride with almost

negligible effect on silica, thus yielding higher selectivity than sodium decyl sulfate (Clo).

However, when sodium cetyl sulfate (C14) was examined, there was a smaller decrease in MRR

of silicon nitride resulting in lower selectivity. Considering that the Krafft point of C14 sodim

sulfate (30 C) [80] is higher than room temperature and higher than that of SDS (16 C) [80],

the surfactant was not completely solubilized and therefore failed to form a functional

passivation layer.

Another approach to decrease the dosage of the surfactant required to achieve desired

selectivity involved using mixed surfactant system (Tween 80/SDS and dodecyl alcohol/SDS) at

pH 2. The MRR and selectivity for the selected systems are plotted in Figure 4-7. In the case of

dodecanol and SDS, selectivity was lower for the mixed surfactant system than for 16mM SDS

alone. It is possible that the addition of a small amount of dodecanol promotes adsorption of SDS

both on silica and silicon nitride. Although there was no appreciable change in the MRR on

silicon nitride, the higher adsorption of SDS on silica also passivated its surface.

It has been reported by Pala and co-workers that surfactant mixture of SDS and various

nonionic surfactants can produce synergistic effects for dispersion of slurry under high ionic

strength conditions [34, 37]. When 8mM Tween 80 was added to 16mM SDS, the MRR of silica

was highly suppressed, whereas that of silicon nitride remained almost unchanged, thereby

resulting in poor selectivity. It is well known that nonionic surfactant such as Tween 80, which

has ethylene oxide groups (OC2H4), can adsorb on silanol groups (SiOH) on silica through

hydrogen bonding [83]. These observations strongly suggest that a surfactant or surfactant









system that exhibits strong preference only for silicon nitride is essential for developing

surfactant-based high selectivity slurries.

In summary, colloidal silica, which shows high dispersion stability in the range of pH 2 to

11, was utilized to develop a high selectivity slurry. The addition of SDS at pH 2 resulted in

more than ten times higher selectivity than the conventional slurry. Additionally, AFM

roughness measurement showed an acceptable surface finish. Adsorption density measurements

revealed that there is a preferential higher adsorption of SDS on silicon nitride, possibly due to

electrostatic attraction, as compared to silica. The SDS adsorption results in differential

passivation/lubrication and hence lower polishing efficiency of silicon nitride as compared to

silica. The CMP characteristics examined as a function of added SDS showed that decrease in

MRR and increase in selectivity leveled off at about twice the surfactant CMC and remained

unchanged, thereafter. The surfactant requirements appear to be driven by their adsorption

primarily on silica abrasive particles. To reduce the surfactant dosage, longer alkyl chain length

surfactants were tested, which yielded higher selectivity at lower dosage. However, the addition

of a long chain length alcohol to substitute for the surfactant resulted in lower selectivity,

probably due to higher adsorption of the surfactant on silica. Mixed ionic and nonionic surfactant

systems, on the other hand, resulted in poor selectivity due to passivation of both silica and

silicon nitride, although to a different degree.
















3200-
2800-
2400-
2000-
1600-
1200-
800-
400-
0-
-400-


MRR Without SDS
-- SiO2
--- 3N4


MRR with SDS
-0- SiO2
-0-SiN4


1 2 3 4 5 6 7 8 9 10 11
pH


(b) -- Selectivity without SDS
A --- Selectivity with SDS










A A


1 2 4 5 6 7 8 9 10 11
pH



Figure 4-1. Influence of SDS addition on CMP performances: (a) Variation of material removal
rate (MRR) as a function of slurry pH with and without 16mM sodium dodecyl
sulfate (SDS), (b) Accompanying selectivity of the slurry.














0.35


0.28- I SiO2
E Si3 N4

) 0.21

0)
S0.14
t--

0 0.07


0.00
Standard (pH 10.4) pH 2 16mM SDS at pH2

Figure 4-2. Surface finish of silica and silicon nitride substrates processed with standard and high
selectivity slurry.















60
5.0-
.- 0 SiO2 40
4.5 0 Si3N4 -20
5 4.0 0
E 4 -- ---- -* --- ----^--- -- -- -- ---------*--------------------------- *--------- -------- 0 >
> 2
S3.5- Si0 -40 N
3 3.0 -40 :
-6 2.5 60 0
2.0 -80 -
N
S1.5- -100

< 1.0- -- --120
0.5 --140
0.0-- ,------- -160
1 2 3 4 5 6 7 8 9 10 11
pH

Figure 4-3. Variation of zeta potential of silica and silicon nitride substrate and adsorption
density of 16mM SDS on silica and silicon nitride powder measured by total organic
carbon (TOC).























81

















2700-
pH 2 -50
2400 -
E 2100-
SI 40
1800
S1500-
S1200- Selectivity 30 ;
0 900-
I 20 (9
280
210- MRR SiO,
210
S140 /- MRR SiN 10
S140- 3 4
70

-3 0 3 6 9 12 15 18 21 24 27
SDS Concentration (mM)



Figure 4-4. Variation of MRR and accompanying selectivity of Klebosol slurry (12 wt%) as a
function of added SDS concentration at pH 2.















2.38
2^ 2.36

2.34

2.32

S2.30
S2.28

S2.26
~ 2.24

2.22

2.20

2 4 6 8 10 12
pH
Figure 4-5. Adsorption density of SDS on 12 wt% Klebosol slurry with 16 mM SDS as a
function of pH.


















M MRR Si3N4
E 2300- Selectivity
2200 40

M 2100-
H 30 >
2000
> O
o 60-
E 20 U)

cU 30 -
0 10



66 mM Clo 16 mM C12 4.2 mM C14

Sodium Sulfate Sodium Sulfate Sodium Sulfate


Figure 4-6. Effect of alkyl chain length of sodium alkyl sulfate on MRR and selectivity at pH 2.
The concentration was adjusted to 2 times the CMC to compensate the loss during
CMP process.














pH2 A


?000-

1500-

1000

60


SMRR SiO2
MRR Si3N
B Selectivity


II


-50

-40

-30>
UD
(D
-20 C/

-10

-0


A = 16 mM SDS
B = 0.8 mM Dodecanol/15.2 mM SDS
C = 8 mM Tween 80/16 mM SDS


Figure 4-7. MRR and selectivity obtained by slurries with various surfactant and surfactant
mixtures at pH 2. Slurry A (16mM SDS) was included for comparison purpose.


2500-


n










CHAPTER 5
ADSORPTION STUDY OF SODIUM DODECYL SULFATE ON SILICA

Considering the relevance of surfactants in developing selective CMP slurries, it is

important to understand their adsorption mechanisms on different substrates to optimize their

performance as passivating agents. Accordingly, adsorption behavior of sodium dodecyl sulfate

(SDS) on silica was studied. Special emphasis was placed on SDS adsorption on colloidal silica

particles at high pH where both constituents exhibit negative charges.

To measure adsorption density of SDS on colloidal silica particles, diluted Klebosol

colloidal silica slurry (12 wt%) was prepared. After dilution, the slurry pH was measured to be

around 10.4. Suspension pH was adjusted with HC1 and KOH solutions prepared with analytical

grade substances purchased from Fisher Scientific Co. A proper amount of 100 mM SDS

solution was added to obtain 1 to 5 mM SDS concentrations. Higher SDS concentrations were

achieved by adding dry SDS powder. Adsorption density measurement incorporated the

following steps: 1) add surfactant to silica suspension, 2) magnetically stirring for 10 min, 3)

centrifuge at 1500 rpm and 4) separate appropriate amount of supernatant and dilute with

nanopure water to yield a concentration within calibration range (around 50 ppm). Finally, 40

ml vials were loaded to total organic carbon (TOC) analyzer and measure the residual (bulk)

SDS concentration in the supernatant. Concentration of adsorbed surfactant on particle was

calculated from the difference in input concentration and residual bulk concentration. Specific

surface area of colloidal silica was measured to be 34 m2/g by Quantachrome Autosorb 1C-MS.

In order to gain insight into specific binding mechanisms, Fourier transform infrared

spectroscopy (FTIR) measurements were conducted. A nitrogen-purged Nicolet Magna 760

spectrometer equipped with a DTGS detector was used to conduct FTIR analysis.

FTIR/attenuated total reflection (ATR) method is well established for its sensitivity to the









surface property change [84-86]. Since it is well known that the surface of silicon is covered with

silica by spontaneous oxidation, Si ATR crystal and surfactant solution were used to investigate

the adsorption behavior. All the spectra were the results of 512 co-added scans at a resolution of

4 cm-1. Surfactant solutions of different concentration of SDS were prepared at pH 10.4. During

the measurement, the solutions were added to the Si ATR crystal assembly. After adding

surfactant solution the sample chamber was purged with dry N2 gas to remove any residual

atmospheric moisture and CO2. After 20 minutes of purging, CO2 peaks disappeared, however

H20 peaks could not be eliminated.

Adsorption Behavior of SDS on Silica

Adsorption isotherm of SDS on colloidal silica suspension (Klebosol 1501-50, 12 wt%)

measured at pH 10.4 is given in Figure 5-1. Despite the fact that both SDS surfactant and silica

surface are negatively charged at this pH, the isotherm appears to be similar to that of

electrostatic interaction dominant adsorption behavior [87-89]. In region I, where adsorption

density is not high, adsorption is assumed to occur by electrostatic attraction. In region II, a

sudden increase in adsorption is attributed to hemimicelle formation. In region III, there is a

decrease in the rate of adsorption as indicated by change in the slope, which is ascribed to bilayer

formation. Adsorption in region IV reaches a constant value apparently due to micelle adsorption

on the surface [90].

In the current study, at low equilibrium surfactant concentrations up to 1.6 mM, the

adsorption is very small due to electrostatic repulsion between SDS and silica substrate.

However, there was a measurable adsorption probably due to hydrogen bonding. Beyond 1.6

mM, the adsorption increases sharply and may be attributed to attractive hydrophobic

interactions between alkyl chains of surfactant resulting in hemimicelle formation. Beyond 8

mM, adsorption density leveled off. Critical hemimicelle concentration (HMC) and critical









micelle concentration (CMC), which can be inferred from the adsorption isotherm in Figure 5-1

occurred at around 1.6 mM and 8 mM, respectively. In the previously reported SDS-alumina

system with a background electrolyte of 0.1 M NaC1, HMC and CMC were reported to be around

0.05 mM and 1.6 mM, respectively [91]. In the current system, electrostatic repulsion plays a

dominant role in controlling adsorption behavior at lower surfactant concentration, thereby

resulting in relatively higher HMC on silica. Beyond HMC, hydrophobic attractive forces govern

surfactant adsorption process.

Under saturation adsorption conditions, the average area per molecule was calculated to be

41.6 A2 from adsorption isotherm, which compares favorably to 53 A2 reported at the air-water

interface for SDS [80]. In the case of SDS-alumina system, it was calculated to be around 23.7

A2 indicating the formation of more compact surfactant aggregates due to attractive electrostatic

interactions between SDS and alumina. The area covered by the adsorbed SDS molecules on

silica particles was calculated to be 663.2 m2, assuming the area occupied by one SDS molecule

to be 53 A2. Considering that the total area of silica particles is 520.2 m2, surface coverage by

SDS molecules indicates the formation of a bilayer, if this surface is assumed to be homogenous,

or micellar type adsorption, otherwise. In the latter case, using the reported aggregation number

(64) and the radius of SDS micelle, 20 A [79, 92], it was calculated that there are total 1.95

x 1019 micelles adsorbed onto silica particles. Therefore, in the steady state, 47.2% of the silica

surfaces is covered with SDS micelles. Using a theoretical density of 2 g/cm3, 15.3 g of silica

particles in the slurry, and the particle radius of 45nm, the number of silica particles in 100 ml

slurry was calculated to be 2 x 1016. This value indicates that approximately 9.75 x 102 SDS

micelles coat each particle.









To further understand the adsorption of SDS on similarly charged silica, adsorption energy

was calculated using modified Ster-Graham equation (Equation (4-3)). Calculated adsorption

free energy under saturation adsorption conditions was found to be -2.9 kcal/mol indicating that

primarily physical adsorption is responsible for SDS adsorption on silica.

In order to assess the effect of pH on SDS adsorption on silica, measurements were

conducted at 1.6 mM and 16mM concentrations. Results plotted in Figure 5-2 show the

adsorption behavior of SDS on silica correlates well the zeta potential of silica indicating that

surface charge of the silica plays an important role in SDS adsorption. Adsorption energy

calculations revealed that at 1.6 mM SDS concentration, adsorption energy at pH 10.4 is 0.02

kcal/mol as compared to -1.17 kcal/mol at pH 2, indicating an energetically unfavorable process

at pH 10.4 and a favorable one at pH 2. At 16mM, the electrostatic effect was probably

dominated by the increased hydrophobic attractive interactions between alkyl chains resulting in

adsorption energy of -3.14 kcal/mol at pH 10.4 and -3.59 kcal/mol at pH 2 indicating favorable

adsorption at both pH values.

Structure of Adsorbed SDS Molecules

To investigate the structure of adsorbed SDS molecules on silica surface, zeta potential

was measured as a function of added SDS concentration at pH 10.4 (Figure 5-3). At very low

concentration of SDS (region I), zeta potential essentially remains unchanged. As the

concentration increases (region II), sodium ions are adsorbing on the silica surface resulting in

less negative zeta potential. It is hypothesized that there are surfactant molecules weakly bonded

to sodium ions. Surfactant molecules associated with sodium ion will not exhibit significant

impact on the zeta potential measurements due to mutual charge neutralization. As the surfactant

concentration increases, hemimicelles form and grow in size in region III, and the slope of zeta

potential increase becomes smaller than region II. It seems that free SDS starts to adsorb on the









hemimicell coated surface forming bilayers in region III resulting in a lower slope change.

Finally in region IV, when surfactant aggregates form micelles, zeta potential reversal occurs by

incorporating a number of free monomers in the solution. When a background electrolyte was

added to the system, overall zeta potential was less negative and the slope change was less

pronounced, but a similar trend was observed. Above hypothesis is shown schematically in

Figure 5-4.

The correlation between SDS adsorption and zeta potential is clearer from Figure 5-5,

where adsorption density and zeta potential are co-plotted as a function of equilibrium

concentration of SDS. The change in zeta potential follows the adsorption isotherm and zeta

potential reversal occurs at CMC. However, due to the low surface coverage of micelles (4%) on

silica surfaces the change was not significant.

SDS adsorption behavior on silica, as determined in the present study, is contrary to

electrostatic considerations, since both the substrate and surfactant molecules are similarly

charged. There have been several reports of SDS adsorption on negatively charged silica or

sepiolite, a hydrated magnesium silicate (Sii2Mg9030(OH)6(OH2)4H20) [76, 77, 93]. Possible

mechanisms for this observation were hydrogen bonding between silanol groups and SDS, and

counter ion mediated surfactant adsorption. Several noticeable thermodynamic properties of the

surfactant were reported by Ozdemir et al. through the adsorption study of SDS on sepiolite. At

saturation adsorption, the adsorption free energy calculated from Frumkin model was -3.1

kcal/mol at 25 C [93]. It is comparable to the results in the present study (-2.9 kcal/mol). This

low energy of adsorption indicates that weak physical forces are responsible for adsorption.

Calculated adsorption free energy from the adsorption isotherm in Chandar's report was -4.18









kcal/mol, indicating that the driving force for adsorption involving electrostatic attraction is

higher than hydrogen bonding alone.

To further investigate the mechanism of SDS adsorption on negatively charged silica

particles, FTIR ATR (attenuated total reflection) measurements were conducted. It should be

mentioned that quantitative analysis by FTIR is not very accurate and it is not well understood

how the electrostatic interaction affects the FTIR spectra. On the other hand, adsorption of

various molecules via hydrogen bonding has been well observed and documented [94, 95]. One

of the noticeable research on hydrogen bonding behavior for silica and dibenzodioxin was done

by Guan et al. [95]. They reported that as the adsorption of dibenzodioxin on silica surface

increased, the peak of geminal silanol group decreased and that of isolated silanol group

increased, indicating that the molecular adsorption occurs at the expense of the silanol group by

hydrogen bonding. Their measurement was done using dry powder samples. In the current study,

all the measurements were conducted in aqueous surfactant solution by using ATR crystal.

Figure 5-6 shows the spectra of SDS at 1, 2.5, 5 and 10 mM concentration in the CH2

stretching region measured at pH 10.4. As was discussed by Pankaj et al., the absorbance

intensity increased up to 5mM, and it decreased at 10mM, which is higher than CMC of SDS (8

mM) [96]. The reason for the decrease of absorbance intensity upon micelle formation is not well

understood, however, it confirms the adsorption of SDS on silica surface at high pH. Due to the

overlapping of the peaks from silanol groups and water, changes in silanol groups were not

confirmed in this experiment.

SDS adsorption on silica can also impact the dispersion stability. Figure 5-7 shows the

particle size distribution of St6ber silica without and with SDS 12 hours after the pH was

changed to 2. Without SDS, there was an additional peak due to particle coagulation since the









isoelectric point (IEP) of the silica particle is known to be around pH 2.7. However, with SDS

addition, no additional peaks were observed.

In summary, SDS adsorption behavior at low concentration was small due to electrostatic

repulsion, however, limited adsorption was observed due to hydrogen bonding. At intermediate

concentrations, it was hypothesized that sodium ion mediated charge neutralization along with

hydrophobic attractive force resulted in higher adsorption of SDS. The slope of the adsorption

density decrease is attributed to bilayer formation. Adsorption free energy calculations and zeta

potential measurements as a function of SDS concentration were supportive of the proposed

hypothesis. It was observed that SDS adsorption on silica surface resulted in a stable dispersion.














pH 10.4 III


IV


---- --


U 7

S10-6 / \ CMC









10 10- 101 10
10-










(CMC) was marked.
< HMC


10 0 101 102
Bulk Equilibrium SDS Concentration (mM)

Figure 5-1. Adsorption isotherm of SDS on colloidal silica (Klebosol 1501-50, 12 wt%) at pH
10.4. Critical hemimicelle concentration (HMC) and critical micelle concentration
(CMC) was marked.














2.6x10-6
0~-6

E 2.4x10-6

E 2.2x106-
6
:. 2.0x10 -
c ,

Q 4.0x10 -
0 -
o

o 2.0x10 -
0.0


0.0-


A Zeta potential


* 1.6 mM SDS
* 16mMSDS


2 4 6 8 10


pH


Figure 5-2. Adsorption density of SDS on colloidal silica (12 wt% Klebosol 1501-50) at SDS
concentration of 1.6 mM and 16 mM and zeta potential as a function of pH.


-0


--20
E
--40 -m


--60 0


--80


--100












-65
-66 Klebosol 12 wt% pH 10.4
1 mM NaCI
-67 no Salt III

> -68-

-69-\1
-69 II IV

-70 I

oi -71 I o
S *-* \
S-72-

-73

-74
'' l 1 l i e I I I I '
100 101 102
Concentration of SDS (mM)

Figure 5-3. Zeta potential of Klebosol slurry as a function of SDS concentration at pH 10.4.











I ./
* a f f a a


7 *I II
W..........I.E...


'i"


SSiO,
1^~liiiii~..._~~~_~~ 1~ ~~ ;;;'"~~~~~~~ ///


IV


Si Sii 5' a.1m aaS1 Si.'
II~~~~~~~~~ II) II II PI II iO)


Figure 5-4. Pictorial depictions of the possible surfactant aggregates films at concentrations
corresponding to I-IV in Figure 5-3.


Si
RiO_


d











:- Adsorption density -69.0
S- + Zeta potential -e~ .Z
E pH 10.4 CMC IV --69.5
0 6 pH 10.4 ll
E 10 -
/ --70.0

SII -70.5
0
10.-7i -71.0

S- \ -71.5

10-8 -72.0
------HMC 72.5

100 101
Bulk Equilibrium SDS Concentration (mM)

Figure 5-5. Adsorption characteristics of SDS on Klebosol silica slurry and zeta potential as a
function of concentration of SDS at pH 10.4.












0.012- pH 10.4

0.010- SDS
1mM
---- 2.5mM
0.008- i
i 5mM
Y .------- 10mM
\ 0.006- 1 0

0) 'i
0 0.004- \\
_.'i ..
/ i \ /
0.002- \" ,
I\ I/

0.000 ,___ .....

3100 3000 2900 2800 2700
Wavenumber (cm-1)


Figure 5-6. FTIR/ATR Spectra of SDS solution at 1, 2.5, 5 and 10 mM bulk concentration in the
CH2 stretching region (2921, 2924) measured at pH 10.4 using Si ATR crystal.
































0.5 1.0 1.5 2.0 2.5
Particle size (am)


Figure 5-7. Particle size distribution of Geltech SiO2 at pH 2 with and without
hours after pH change.


16 mM SDS 12


3.0










CHAPTER 6
APPLICATION OF DENSITY FUNTIONAL THEORY BASED MODELING FOR
SURFACTANT ADSORPTION STUDY

There have been numerous modeling efforts to develop reliable tools for predicting

colloidal systems behavior. There are two broad areas of modeling for investigating the structure

of molecules and their reactivity: molecular mechanics and electronic structure theory. They

perform the same basic calculations: i) compute the energy of a particular molecular structure

and ii) geometry optimization to produce the lowest energy molecular structure [97]. In addition,

electronic structure model is capable of calculating vibrational frequencies of molecules resulting

from interatomic motion.

Molecular mechanics based models use the laws of classical physics, and each one is

characterized by its particular force field. In general, it does not explicitly treat the electrons in a

molecule. They perform computations based on the interactions among the nuclei, while

interactions involving electrons are implicitly included in force fields through parameterization.

This approximation enables the molecular mechanics modeling to be fast and cost effective, and

applicable to large systems. However, it also has several limitations, e.g., each force field is

system specific, and it is unable to calculate chemical problems where electronic effects

predominate (i.e. bond formation and breakage), since interactions among electrons are neglected

[97].

Electronic structure methods use the laws of quantum mechanics. There are two major

classes in the area, i) semi-empirical methods such as AM1 and PM3, which utilize parameters

derived from experimental data, ii) ab initio methods, which utilize no experimental parameters,

instead, computations are based solely on the laws of quantum mechanics and the values of

several physical constants. Semi-empirical calculations are relatively inexpensive and produce