1.  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley Modeling of CMP David Dornfeld CMP researchers: Jihong Choi, Sunghoon Lee, Dr. Hyoungjae Kim, Dr. Dan Echizenya Department of Mechanical Engineering University of California Berkeley CA 94720-1740 http://lma.berkeley.edu

2. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 2 Overview Background on modeling Review of work to date Some new developments pattern/feature sensitivity pad design

3. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 3 New Book on Modeling Chemical Mechanical Planarization (CMP) “Integrated Modeling of Chemical Mechanical Planarization for Sub-Micron IC Fabrication: From Particle Scale to Feature, Die and Wafer Scales,” J. Luo and D. A. Dornfeld For information: www.springeronline.com/east/3-540-22369-X. Written by researchers at UC-Berkeley, this monograph reviews CMP modeling literature (from Preston to present day efforts) and develops, with a strong emphasis on mechanical elements of CMP, an integrated model of CMP addressing wafer,die and particle scale mechanisms and features. Special emphasis is on abrasive sizes, distributions and resulting material removal rates and uniformity resulting over all scales. 175 Figures and 14 tables ISBN 3-540-22369-x Springer-Verlag 2004 Or contact: dornfeld@berkeley.edu

4. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley Chemical Mechanical Planarization Mechanical Phenomena Chemical Phenomena Interfacial and Colloid Phenomena CMP Team in FLCC Dornfeld, et al Doyle, et al Talbot, et al

5. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 5 Scale Issues in CMP From E. Hwang, 2004 Scale/size nm µm mm Material Removal Mechanical particle forces Particle enhanced chemistry Chemical Reactions Active Abrasives Pores, Walls Grooves Tool mechanics, Load, Speed critical featuresdies Pad Mechanism Layout wafer

6. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 6 w p :pad rotation table pad slurry feed conditioner head w w : wafer rotation Oscillation F : down force Backing film Retainer ring Wafer Wafer Carrier Pad Pore Wall Abrasive particle CMP Process Schematic Electro plated diamond conditioner Typical pad

7. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley Bulk Cu CMPBarrier polishing W CMPOxide CMPPoly-Si CMP Physical models of material removal mechanism in abrasive scale Chemical reactions Bulk Cu slurryBarrier slurryW slurryOxide slurryPoly-Si slurry Mechanical material removal mechanism in abrasive scale Abrasive type, size and concentration [oxidizer], [complexing agent], [corrosion inhibitor], pH … Pad asperity density/shape Pad mechanical properties in abrasive scale Pad properties in die scale Slurry supply/ flow pattern in wafer scale Wafer scale pressure NU Models of WIWNU Models of WIDNU Topography Wafer scale velocity profile Wafer bending with zone pressures Better control of WIWNU Reducing ‘ Fang ’ Small dishing & erosion Ultra low-k integration Smaller WIDNU Reducing slurry usage Uniform pad performance thru it ’ s lifetime Longer pad life time Reducing scratch defects Better planarization efficiency E-CMP Pad groove Pad design Fabrication Test Fabrication technique Slurry supply/ flow pattern in die scale Cu CMP model design goal Pad development Pattern MIT model Dornfeld modelDoyle An overview of CMP research in FLCC Talbot

8. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 8 The 4-component system Hypotheses: –all polishing processes can be described as a 4 component system; –Understanding the components and their interactions (pair-wise, triplets, etc) provides a structure to catalog our knowledge (and ignorance) Lap (rigid) WorkpieceLapGranuleCarrier fluid Platen Pad } “Granule”? Deliberately sought a word that covers the range of particles used without implying anything about size, hardness, or removal mechanism:  m to nm size range; from hard (diamond) to soft (rouge); Source: 86. Evans, J., Paul, E., Dornfeld, D., Lucca, D., Byrne, G., Tricard, M., Klocke, F., Dambon, O., and Mullany, B., “Material Removal Mechanisms in Lapping and Polishing,” STC “G” Keynote, CIRP Annals, 52, 2, 2003.

9. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 9 Six possible pair-wise interactions Fluid-workpiece Workpiece-pad Workpiece-granule Granule-pad pad-fluid Fluid-granule

10. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 10 Three-way interactions (triplets) Workpiece-fluid-granule Workpiece-fluid-pad Workpiece-granule-pad Fluid-pad-granule

11. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 11 Stribeck Curve and Characteristics of slurry film thickness Friction coefficient Film thickness Hersey number(= ) Hydrodynamic lubrication Elasto- hydrodynamic lubrication Boundary lubrication Direct contact Semi-direct contact Hydroplane sliding Stribeck curve Polishing pad Wafer Slurry Direct contact Semi-direct contact Hydroplane sliding

12. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 12 Gap effects on “mechanics” Pad-based removal Slurry-based removal ‘Small’ gap ‘Big’ gap Silicon wafer Polishing pad Abrasive particle Delaminated by brushing Eroded surface by chemical reaction --- softening Silicon wafer Polishing pad Abrasive particles

13. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 13 Idealized CMP Silicon wafer Polishing pad Abrasive particle ‘Softened’ surface by chemical reaction Pad asperity Mechanical Aspects of the Material Removal Mechanism in Chemical Mechanical Polishing (CMP)

14. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 14 Interactions between Input Variables Four Interactions: Wafer-Pad Interaction; Pad-Abrasive Interaction; Wafer-Slurry Chemical Interaction; Wafer-Abrasive Interaction Polishing pad Abrasive particles in Fluid (All inactive) Pad asperity Active abrasives on Contact area Vol Chemically Influenced Wafer Surface Wafer Abrasive particles on Contact area with number N Source: J. Luo and D. Dornfeld, IEEE Trans: Semiconductor Manufacturing, 2001 Velocity V

15. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 15 Framework Connecting Input Parameters with Material Removal Rate Slurry Abrasive Weight Concentration C Fraction of Active Abrasive: 1-  ((g-X avg )/  ) where g is the minimum size of active abrasives Force F & Velocity Wafer Hardness H w / Slurry Chemicals & Wafer Materials Vol Active Abrasive Size X avg-a Basic Equation of Material Removal: MRR= N  Vol Average Abrasive Size X avg Proportion of Active Abrasives N Pad Topography & Pad Material Abrasive Size Distribution  Down Pressure P 0 g X avg  Fraction of Active Abrasives X avg-a

16. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 16 K e1 (K 1 =84148, K 2 = 0.137) Experimental Verification of Pressure Dependence of Material Removal Rate (MRR) Advantage over Preston’s Eq. MRR= K e PV+ MRR 0 : What input variables and how they influence K e is predicable MRR= N Vol= K 1 {1-  (1- K 2 P 0 1/3 )}P 0 1/2. K e2 (K 1 =8989, K 2 = 0.3698) SiO 2 CMP Experimental Data from Zhao and Shi, Proceedings of VMIC, 1999

17. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 17 Abrasive Size Distribution Dependence of MRR: Particle Size Distribution [1] Five Different Kinds of Abrasive (Alumina) Size Distributions for Tungsten CMP 1. Bielmann et. al., Electrochem. Letter, 1999 Abrasive Size X (Log Scale) (%) Frequency

18. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 18 Relationship between Standard Deviation and MRR Based on Model Prediction Std dev influenced Size influenced

19. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 19 Pattern-Density Dependency Model InterLevel Dielectric Case (single material) K K/density Up Area 0 Down Area Time pad oxide pad oxide Same Pattern Density Different Orientations Source: MIT MRR

20. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 20 Framework of a CMP Topography Evolution Model

21. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley Dishing and Erosion in Copper Damascene Process Via Trench SiO 2 SiN (a) (b) (c) (d) Fabrication steps in dual damascene process (a) deposition of SiN, SiO 2 and etching trenches and vias in SiO 2 (b) deposition of barrier layer (c) copper fill (d) CMP and deposition of SiN (courtesy of Serdar Aksu)

22. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 22 Definition of Feature-Scale Topography (a) (b) (a) Feature scale topography before dielectric material is exposed and (b) feature scale topography after dielectric material is exposed W cu S H W ox Copper Dishing d = S Oxide Erosion e H= H ox H ox = H ox0 H cu Copper Thinning

23. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 23 KdKd KfKf (a) (b) (c) (d)  E E  E1E1 11 E2E2 22 E Models of Polishing Pad Linear Elastic and Linear ViscoElastic Models Separated Models of Pad Bulk and Asperities

24. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 24 Dishing d 3 Erosion e 2 1 Df S 1 =Df 1 S=S 0 H=H cu0+ H ox0 H= H stage 1 H cu0 H ox0 Three Stages of Wafer-Pad Contact Only upper part of step is in contact Both upper and bottom parts of step is in contact Two different materials are removed simultaneously

25. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 25 Simulation Results of Step Height Evolution for Different Pattern Density Linear Elastic Pad Linear Viscoelastic Pad Step height S (nm) Planarization time (sec) W cu = 100 microns

26. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 26 Copper Dishing as a Function of Pattern Density using commercial pads

27. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 27 Copper Dishing as a Function of Selectivity

28. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 28 Effect of Pattern Density - Planarization Length (PL) ILD Metal lines Planarization Length High-density region Low-density region Global step

29. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 29 Effective pattern density a=320um a=640um a=1280um Modeling of pattern density effects in CMP Planarization length (window size) effect on “Up area”

30. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley Initial pressure distribution Topography evolution New pressure distribution Contact wear model MRR model Iteration with time step Die scale modeling of topography evolution during CMP

31. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 31 PAD Z(x,y) Reference height (z=0) Z_pad Z(x,y) Z_padz dz Feature level interaction between pad asperities and pattern topography F_tent > F_die ?F_tent < F_die ? ++Z_pad--Z_pad No Yes No Yes Z_pad

32. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 32 k1 k2 Chip level interaction between pad and pattern topography r PL w 40um Pattern 40um x 40um cell MIT model : approximation of contact wear model

33. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 100% 50% 33% 20% 50%33%20% t=0 sect=10 sect=20 sec t=30 sect=40 sect=50 sec t=60 sect=70 sect=80 sec Simulation result

34. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 34 Pattern orientation effect on on copper dishing SiO 2 TiCu Si Kinetic analysis of sliding direction during process time pad rpm < wafer rpm pad rpm = wafer rpm

35. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 35 R a = 12.5µm R z = 96.7µm Pore diameter : 30~50 µm Peak to Peak : 200~300µm 100µm 45µm -45µm 100µm 300µm500µm (SEM, x150) 200~300µm (White light Interferometer, x200) Pad Characterization

36. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 36 Asperity: Real contact area 10~50 µm Pores 40~60µm Simplified Pad Model Peak to Peak 200~300 µm 1. Reaction Region (10~15 µm) 2. Transition Region 3. Reservoir Region Pad modeling

37. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 37 3 Dimensional analysis Reaction region Reservoir region Transition region

38. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 38 2D and 3D image of reaction region Contact area : 10-50µm Ratio of real contact area : 10-15% Spherical or conical shape edge Stress concentration when compressed 2 dimensional image (w/o pressure) 3 dimensional image (w/o pressure)

39. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 39 10 – 50 µm Reaction region (asperity) Reaction region – ILD CMP Over polishing on recess area Smoothing, not planarization Defects of a conventional pad 50 µm Large asperity wafer ILD Rounding 10 µm Small asperity wafer ILD Over polishing

40. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 40 Reaction region – Cu CMP wafer Pressure Position Stress concentration ErosionDishing Fang Cu-CMP defects (due to stress concentration in conventional pad) Pad asperity Nominal pressure Avg. contact pressure wafer

41. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 41 NewIn 3minutes In 5minutes In 7minutes Pad degradation

42. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 42 Design rules for a pad Macro scaleMicro scaleNano scale  Stacked layer (Hard/soft)  Slurry channel  Constant contact area (width:10-50um)  The ratio of real contact area (13-17%)  Conditioning-less CMP  High slurry efficiency  Compatible features to abrasive  Constant re-generation of nano scale surface roughness Design rules for a pad

43. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 43 Soft Layer (i.e. low stiffness) Hard Layer (i.e. high stiffness) ChannelNano scale features A pad design based on the rules 50-70µm 50-200µm

44. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 44 Wafer Pad Wafer Pad Conditioning-less process High planarity & good uniformity in ILD CMP Without stress concentration Less defects in Metal CMP Advantages ILD CMP Cu CMP Expectations

45. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 45 Design of new pads Type 2 – With slurry guidance Type 1 – Without slurry guidance 50µm Slurry flow direction 20µm

46. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 46 Type 1 Type 2 Area : 4.294^-10 m 2 Flow rate : 3.24^-10 kg/sec Area : 4.3^-10 m 2 Flow rate : 3.93^-11 kg/sec 8 times more flow rate On contact area Simulation result

47. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 47 Pad fabrication 1. Master2. Silicone Rubber Casting 3. Silicone Rubber Mold 4. Hard Layer Casting 5. Soft Layer Casting 6. Demolding New pad

48. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 48 Performance of a new pad – Planarity in ILD CMP ILD pattern (MIT mask Version 1.0) 50% 50um/50um 20% 20um/80um Si wafer SiO 2 0.77µm 1.7µm Pad IC1000/SUBA400New pad 60rpm Wafer 3inch wafer (12-100% density,1.7 µ m SiO2) 30rpm Slurry D-7000 (Cabot Co.) 100ml/min Pressure1.6psi 2.7psi Polishing machine Experiment condition

49. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 49 Density 20% - under same pressure:1.6psi Time : 17minutes Over Polishing : 2200Å Time : 40minutes Over Polishing : 400Å Good planarityHigh removal rate IC1000/SUBA400 (1.6psi)New pad (1.6psi)

50. LMA  Laboratory for Manufacturing Automation, 2005 University of California at Berkeley 50 IC1000/SUBA400 (1.6psi)New pad (2.7psi) Time : 17minutes Over Polishing : 2200Å Time : 20minutes Over Polishing : 800Å Good planarity & removal rate Density 20% - under different pressure:1.6psi &2.7psi