Feature Level Compensation and Control: Chemical Mechanical Planarization

1. Feature Level Compensation and Control: Chemical Mechanical Planarization Investigators Fiona M. Doyle, Dept. of Materials Science and Engineering David A. Dornfeld, Dept. of Mechanical Engineering University of California, Berkeley Jan B. Talbot, Dept. of Chemical Engineering, University of California, San Diego F.M. Doyle, D.A. Dornfeld, J.B. Talbot

2. Mechanical Aspects Andrew Chang (RAPT Technologies) Yongsik Moon (Applied Materials) Jianfeng Luo(Cypress Semiconductor) Inkil Edward Hwang Sunghoon Lee Jihong Choi Chemical Aspects Serdar Aksu (Suleyman Demiril University, Isparta, Turkey) Ling Wang Interfacial and Colloid Aspects Tanuja Gopal (UCSD) Students Involved in CMP Effort F.M. Doyle, D.A. Dornfeld, J.B. Talbot

3. Outline • Objective of FLCC effort in CMP • Background on CMP • Mechanical phenomena • Interfacial and colloidal phenomena • Chemical phenomena • Modeling efforts • Future work F.M. Doyle, D.A. Dornfeld, J.B. Talbot

4. Objective of FLCC effort in CMP F.M. Doyle, D.A. Dornfeld, J.B. Talbot

5. Aim • Insure uniform, global planarization with no defects by means of optimized process recipes and consumables • Idealized single-phase CMP processes are now well understood in terms of the fundamental physical-chemical phenomena controlling material removal • The challenge is to extend this to heterogeneous structures that are encountered when processing product wafers with device features F.M. Doyle, D.A. Dornfeld, J.B. Talbot

6. Approach • Develop integrated feature-level process models, encompassing upstream and downstream processes • These models will drive process optimization and the development of novel consumables to minimize feature-level defects • We will require the capability of faithfully predicting defects and non-idealities at feature boundaries. F.M. Doyle, D.A. Dornfeld, J.B. Talbot

7. Background on CMP F.M. Doyle, D.A. Dornfeld, J.B. Talbot

8. CMP Process Schematic F : down force Oscillation w w : wafer rotation conditioner Wafer Carrier slurry feed head Retainer ring Backing film Wafer table Wall Pore pad Wall Pore w p :pad rotation Pad Pad Abrasive particle Electro plated diamond conditioner Typical pad F.M. Doyle, D.A. Dornfeld, J.B. Talbot

9. Applications of CMP – ILD, Metal and STI CMP • Interconnection (ILD CMP) • Via formation (Metal CMP) • Shallow Trench Isolation (STI CMP) Multilevel Metalization F.M. Doyle, D.A. Dornfeld, J.B. Talbot

10. Idealized CMP ‘Softened’ surface by chemical reaction Silicon wafer Abrasive particle Polishing pad Pad asperity F.M. Doyle, D.A. Dornfeld, J.B. Talbot

11. CMP CMP Parameters Input Parameters Output Parameters Wafer Pad Fiber Structure Conditioning Compressibility Modulus Material Removal WIWNU (Within-Wafer Non-Uniform Material Removal) Slurry pH Oxidizers Buffering Agents, Abrasive Concentration Abrasive Geometry and Size Distribution Die WIDNU (Within-Die Non-Uniform Material Removal) Wafer Geometry and Materials Surface Surface Quality Roughness, Scratching Process PressureVelocity Temperature Slurry Flow Polish Time F.M. Doyle, D.A. Dornfeld, J.B. Talbot

12. Interactions between Input Variables Velocity V Vol Chemically Influenced Wafer Surface Wafer Abrasive particles on Contact area with number N Abrasive particles in Fluid(All inactive) Pad asperity Polishing pad Active abrasives on Contact area Source: J. Luo and D. Dornfeld, IEEE Trans: Semiconductor Manufacturing, 2001 Recognize that wafer is heterogeneous, and there may not be a single interaction between it and other input variables F.M. Doyle, D.A. Dornfeld, J.B. Talbot

13. Chemical Mechanical Planarization Mechanical Phenomena Chemical Phenomena Interfacial and Colloid Phenomena F.M. Doyle, D.A. Dornfeld, J.B. Talbot

14. Mechanical Phenomena F.M. Doyle, D.A. Dornfeld, J.B. Talbot

15. Silicon wafer Abrasive particles Polishing pad Gap effects on “mechanics” Eroded surface by chemical reaction --- softening Silicon wafer Delaminated by brushing ‘Small’ gap Abrasive particle Polishing pad Pad-based removal ‘Big’ gap Slurry-based removal F.M. Doyle, D.A. Dornfeld, J.B. Talbot

16. Hershey number(= ) Stribeck Curve and Characteristics of slurry film thickness Slurry Wafer Direct contact Film thickness Polishing pad Direct contact Semi-direct contact Hydroplane sliding Elasto- hydrodynamic lubrication Hydrodynamic lubrication Friction coefficient Semi-direct contact Boundary lubrication Hydroplane sliding Stribeck curve F.M. Doyle, D.A. Dornfeld, J.B. Talbot

17. Mechanical and Chemical Material Removal Effects vs Slurry Film Thickness Source: Yongsik Moon and David A. Dornfeld, “Investigation of Material Removal Mechanism and Process Modeling of Chemical Mechanical Polishing (CMP),” Engineering Systems Research Center (ESRC), Technical Report 97-11, University of California at Berkeley (September 1997) F.M. Doyle, D.A. Dornfeld, J.B. Talbot

18. Effect of gap on CMP - material removal • Material removal per sliding distance Preston’s equation: MRR vs sliding distance, A/meter (C=Preston’s coefficient P = pressure, V=velocity, h=removed height, s=sliding distant t= time) F.M. Doyle, D.A. Dornfeld, J.B. Talbot

19. Increasing Line Density Proposed Pattern-Pad Contact Mode Direct Contact Increasing Hershey Number Semi-direct or Hydroplaning Contact Pad Slurry Oxide Wafer F.M. Doyle, D.A. Dornfeld, J.B. Talbot

20. A B C D E Die Layout F.M. Doyle, D.A. Dornfeld, J.B. Talbot

21. Increasing Hershey Number Results – Line Density Effect Increasing Line Density F.M. Doyle, D.A. Dornfeld, J.B. Talbot

22. Scale effects – Abrasives/Pad • UR100 from Rodel Abrasive particle(<0.1m) 100m End fibrils 400m Vertically oriented pores Urethane impregnated polyester felt 1500m F.M. Doyle, D.A. Dornfeld, J.B. Talbot

23. Wafer Pad Abrasives Wafer Wall Pore Pad CMP pad topography (IC1400 K-Groove) 3.Groove (side view) 1.Pore 2.Wall 300um Topography of pad surface (source : Rodel (Zygo profile)) After Conditioning (open pores/rough wall) After CMP (glazed pores/smooth wall) New pad (open pores/smooth wall) After Stabilization (open pores/rough wall) F.M. Doyle, D.A. Dornfeld, J.B. Talbot

24. The Need for Conditioning • Pad becomes glazed during polishing • Pad Porosity and Surface Roughness are affected • Slurry transport, contact area and by-product removal deteriorate • Conditioning needed to break up glazed areas New Conditioned Glazed Source: B. J. Hooper, UCD, Dublin, 2001 F.M. Doyle, D.A. Dornfeld, J.B. Talbot

25. 50um(space) Schematic of SMART pad 200um Top View Groove & Pore Hard Material (i.e. high stiffness) Wall Side View Soft Material (i.e. low stiffness) Protrusion for rough wall F.M. Doyle, D.A. Dornfeld, J.B. Talbot

26. Wafer Pad Abrasives Wafer Wall Pore 200um Pad Model Implementation - Pad Design SMART pad surface Polymer pad surface Polyethylene pad surface F.M. Doyle, D.A. Dornfeld, J.B. Talbot

27. Interfacial and colloidal phenomena F.M. Doyle, D.A. Dornfeld, J.B. Talbot

28. Mass Transfer Process • (a) movement of solvent into the surface layer under load imposed by abrasive particle • (b) surface dissolution under load • (c) adsorption of dissolution products onto abrasive particle surface • (d) re-adsorption of dissolution products • (e) surface dissolution without a load • (f) dissolution products washed away or dissolved Dissolution products Abrasive particle Surface Surface dissolution • Ref. L. M. Cook, J. Non-Crystalline Solids, 120, 152 (1990). F.M. Doyle, D.A. Dornfeld, J.B. Talbot

29. Electrical Double Layer of Abrasive Particles + + + + + + + + + + + a + + + + + + + + + + + + + Diffuse Layer Shear Plane Particle Surface • Potential at surface usually stems from adsorption of lattice ions, H+ or OH- • Potential is highly sensitive to chemistry of slurry • Slurries are stable when all particles carry same charge; electrical repulsion overcomes Van de Waals attractive forces • If potentials are near zero, abrasive particles may agglomerate Potential  Distance 1/ F.M. Doyle, D.A. Dornfeld, J.B. Talbot

30. Colloidal effects Glass polishing rate (mm/min) Oxide Isoelectric point Polishing rate (A/min) Colloid oxide • Maximum polishing rates for glass observed compound IEP ~ solution pH > surface IEP • (Cook, 1990) • Polishing rate dependent upon colloidal particle - W in KIO3 slurries • (Stein et al., J. Electrochem. Soc. 1999) F.M. Doyle, D.A. Dornfeld, J.B. Talbot

31. F.M. Doyle, D.A. Dornfeld, J.B. Talbot

32. Chemical phenomena F.M. Doyle, D.A. Dornfeld, J.B. Talbot

33. H-O-H N-H2 C-H2 H2-C H2-N Cu2+ H-O-H O C=O O=C O Chemistry of Glycine-Water System pKa1=2.350 pKa2=9.778 +H3NCH2COOH+H3NCH2COO-H2NCH2COO- Cation: H2L+Zwitterion: HL Anion: L- • Cu(II) glycinate complexes • Cu(H3NCH2COO)2+ : CuHL2+ • Cu(H2NCH2COO)+ : CuL+ • Cu(H2NCH2COO)2 : CuL2 • Cu (I) glycinate complexes • Cu(H2NCH2COO)-2 : CuL2- Potential-pH diagram, with {CuT} = 10-5, {LT} = 10-2 F.M. Doyle, D.A. Dornfeld, J.B. Talbot

34. Cu-Glycine Using electrochemical control of oxidation, see passivation only at high pH, where a solid oxide forms Using hydrogen peroxide as a chemical oxidant, see passivation at pH 4 and 9, where no solid oxide expected F.M. Doyle, D.A. Dornfeld, J.B. Talbot

35. Copper concentration, mg/l, and dissolved copper, nm, in unbuffered aqueous glycine (pH 4.5) F.M. Doyle, D.A. Dornfeld, J.B. Talbot

36. Modeling Efforts F.M. Doyle, D.A. Dornfeld, J.B. Talbot

37. CMP Modeling Roadmap Objectives from Industrial Viewpoint - VMIC 2001 • Models are not reliable enough to be used as verification of process • Usefulness of modeling is the ability to give feedback for “what-if” scenarios (predicting “polishability” of new mask designs) in lieu of time-consuming DOE tests • Models should give some performance prediction for realistic, heterogeneous pattern effects • Models should predict not only wafer scale phenomena but also have some capability to capture feature/chip scale interaction F.M. Doyle, D.A. Dornfeld, J.B. Talbot

38. Interaction Scale Modeling Aspect Estimated Completion Global Environmental Issues < 5 % Wafer-Level Material Removal – Slurry/Abrasive/Pad 50% Material Removal – Chemistry 50% Kinematics 90% Pad Stress Dynamics < 5% Chip/Die-Level Hydrodynamics 30% Pattern Density Effects – Oxide 90% Pattern Density Effects – Copper 20% Pattern Density Effects – STI 40% Feature/ Particle- Level Abrasive Particle Interaction 50% Current Status of Modeling Efforts F.M. Doyle, D.A. Dornfeld, J.B. Talbot

39. MRR Non-uniformity Slurry/Water usage Abrasive Concentration Other Chemical Effluent Energy References Wafer Level Feature Level Die Level Empirical Model Preston Boning Others Individual Model Tribology Kinematic Pad Chemical Integrated Model Literature Review of Modeling of CMP  Preston, 1927      Boning, et al, 1997-1998 (4) Various (10) incl. Burke, Runnels, Zhao     Various (11)        Various (2)    Various (4)   Various (6)         F.M. Doyle, D.A. Dornfeld, J.B. Talbot

40. Motivations for a Comprehensive Material Removal Model • Identify the most important input parameters related to Slurry Abrasives, Wafer, and Polishing Pad except the down pressure P0 and velocity V • Investigate the interactions between the input parameters • Develop material removal rate formulation to consider the roles of the input parameters and their interactions • Model as a basis for process design and optimization (including environmental impacts) F.M. Doyle, D.A. Dornfeld, J.B. Talbot

41. Past 2-D Material Removal Rate (MRR) Models V P0 • Experimental Model [1]: Preston’s Equation Wafer MRR= KeP0V + MRR0where Ke an all-purpose coefficient, MRR0 a fitting parameter, P0, down pressure, and V, the relative Velocity. Pad • Analytical Model Considering Wafer-Pad Contact Area[2]: Zhao’s Equation V P0 Wafer MRR= Ke(P0-Pth)2/3V, where Pth a fitting parameter. Active abrasive number is proportional to contact area. Contact area  P02/3 Contact Area A *All with an all purpose factor Keto represent the roles and interactions of other input variables except the down pressure and velocity 1Preston, 1917, J. of Glass Soc. 2. Zhao et. al., 1999, Applied Physics F.M. Doyle, D.A. Dornfeld, J.B. Talbot

42. Consumable Parameters including: Slurry Abrasive Concentration, Abrasive Size Distribution, Slurry Oxidizer Type and Concentration, PH, Pad Topography and Pad Material (Hardness and Young’s Modulus), Wafer Materials and Process Parameters including Down Pressure, Relative Velocity, Slurry Temperature and so on. Mechanical Model Chemical Model Model of Wafer Properties Wafer-Pad Interaction Force Applied on Abrasives Wafer-Chemical Interaction: Passivation Rate Model Number of Active Abrasives Abrasive-Pad Interaction Model of Pad Properties MRR Abraded Material-Chemical Interaction (Dissolution) Model Model of Slurry Abrasive Properties Material Removal by a Single Active Abrasive Abrasive-Wafer Interaction Mechanical-Chemical Interaction Model: Wafer Surface Hardness Model Competition between Mechanical Removal and Passivation Enhancement of Mechanical Elements (Indentation, Leading Edge Area) on Passivation Chemical-Pad Interaction Model Chemical-Slurry Abrasive Interaction Model Weak Relationship Included in Current Model Strong Relationship Included in Current Model Sub-Model not Included in Current Model Sub-Model Relationship not Included in Current Model or Unimportant Relationship Model Output Inputs and Outputs F.M. Doyle, D.A. Dornfeld, J.B. Talbot

43. removal rate growth rate abrasives softer surface harder surface (a) (b) (c) 2a F 2 2 Shaped area is leading edge with aggressive chemical reaction Three Different Regions of Material Removal Rate With Increase of Material Removal Rate Different layer removed with increase of abrasive weight concentration/material removal rate: (a) small MRR < GR of Upper Layer (removed material is upper layer) (b) MRR= GR of Upper layer (upper layer is removed as formed) (c) MRR> GR (part of removed materials is the upper softer layer which is removed as formed, and part of removed materials is the bottom harder layer) F.M. Doyle, D.A. Dornfeld, J.B. Talbot

44. Material Removal Rate Particle-Scale Material Removal Model Consumable Parameters Surface Quality Die-Scale Pressure Distribution Model Layout Density Pattern Density Window and Effective Pattern Density Die-Scale Topography (both vertical & lateral directions) Macroscopic Contact Mechanics Model (Weight Function) Pattern Density Effect Dishing Erosion Wafer-Scale Pressure and Velocity Distribution Wafer-Scale Topography Upper Stream Die-Scale Topography CMP Tool Configurations Upper Stream Wafer-Scale Topography Include device design Dummy Filling Circuit Performance ECAD Process Modeling “Roadmap” F.M. Doyle, D.A. Dornfeld, J.B. Talbot

45. Process Models - Basis for Environmental Analysis F.M. Doyle, D.A. Dornfeld, J.B. Talbot

46. Future Work F.M. Doyle, D.A. Dornfeld, J.B. Talbot

47. Chemical and interfacial effectsCMP of heterogeneous surfaces • Characterize the effect of different slurry additives on the etching of copper when adjacent to diffusion barrier materials, advanced low-k dielectrics and other relevant phases • Explicitly consider autocatalytic effects, e.g. dissolved metal ions catalyzing decomposition of H2O2 • Investigate the ability of different slurries to wet different phases, and the effect of surfactants in modifying wetting behavior • Consider coupling of chemical and mechanical phenomena through mechanisms such as: zeta potential modifications, hydration, dehydration, and sorption of species, particularly organic surfactants and polymers F.M. Doyle, D.A. Dornfeld, J.B. Talbot