Effects of Pad Properties on Cu Dishing During CMP - PowerPoint PPT Presentation

Loading...

PPT – Effects of Pad Properties on Cu Dishing During CMP PowerPoint presentation | free to download - id: 95ea8-ZDg1M



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Effects of Pad Properties on Cu Dishing During CMP

Description:

Hydrophobicity (Surface Tension) Fibrous vs. Porous. Experimental Overview. Cu. Ta. Silicon ... Hydrophobicity DR for porous pads, for large features ... – PowerPoint PPT presentation

Number of Views:307
Avg rating:3.0/5.0
Slides: 25
Provided by: capric
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Effects of Pad Properties on Cu Dishing During CMP


1
Effects of Pad Properties on Cu Dishing During CMP
  • Caprice Gray
  • PhD Student, Mechanical Engineering
  • Tufts University, Medford, MA
  • Intel Intern Presentation
  • 8/18/05

2
Overview
  • Goal Investigate Cu line dishing during CMP
    using polishing pads with varying properties
  • Motivation How do pad properties affect dishing
    as design rules shrink?
  • Topics Covered
  • Automated dishing analysis for higher data
    throughput
  • Evaluated 9 different pad
  • Pad Properties Evaluated
  • Fiber size
  • Pad hardness
  • Hydrophobicity (Surface Tension)
  • Fibrous vs. Porous

3
Experimental Overview
Cu FW
Ta
ILD up region
4
Polishing Pads
Fibrous
Porous
Polyurethane impregnated Polyester felt
Soft
Silk
Mohair Microfiber
Soft
Hard
Polyurea/ Polyurethane Nanocomposit
Tangled Polyester
Microporous polyurethane
Elastomer/ WSP Alloy
5
Analysis Process
Data Modeling and effect screening
6
Automated Data Evaluation
Isolate region of interest and apply data
smoothing filter
Import Raw Data
Isolate feature width region
Isolate features and calculate dishing
Calculate average dishing and Erosion for each
feature width
7
Dishing Rate Results
8
Dishing is Predictable
Hardness Fiber/Pore Size Hydrophobicity Asperity
Geometry Grooving Compressibility Density Material
Pad is a convolution of many material properties
Dishing Rate is a function of Feature
Width Feature Density Pad
9
Evaluation of Variables
10
Summary and Conclusions
  • A LabView interface was developed to automate
    data analysis
  • Summary of design rule affects on Dishing Rates
    (DR)
  • Feature Width?? DR?
  • Large Features ? Large DR variation pad to pad
  • Small Features ? Small DR variation pad to pad
  • Feature Density has minimal affect on DR
  • Summary of pad property affects on DR
  • Hardness? ? DR? for large features
  • Fiber size? ? DR? more so for large features
  • Hydrophobicity? ? DR? for porous pads, for large
    features
  • Hydrophobicity? ? DR? for fiberous pads, more so
    for large features
  • Recommendations
  • Large Features (FW gt 2 um) Allied, JSR
  • Small Features (FW lt 2 um) Allied, SUBA IV, TWI
    711, (Polytex)
  • All Feature Scales (0.32um 20 um) Allied, TWI
    711, SUBA IV
  • Future Work
  • Determine the relationship between other pad
    properties and dishing
  • Examine erosion data
  • Is there different selectivity on the pads? (Cu
    vs. Ta vs. Oxide polish rates)

11
Bonus Features
12
Dishing Rate Normalized to Bulk Rate
Oxide Polishing Pad
13
Dishing Rate and Feature Density
14
Effect of Feature Density by Pad
15
Dishing Rate and Pad Hardness
16
Normalized Dishing Rate with Pad Hardness
17
Dishing Rate and Fiber Size
18
Dishing Rate and Contact Angle
Brushlon and TWI 711 behave completely
differently Brushlon- fiberous, high dishing
rate, contact angle 0 TWI 711 intermediate
dishing rate (between allied and SUBA,90 contact
angle
Porous
Soft Fiberous
19
Contact Angle Images
WSP Elastomer
Microfiber-initial
Tangled Polyester
Nanocomposit
Soaks in near different grove-drop locations.
Otherwise, no soak.
Soaks in
Does not soak in
Soaks in
Porous Felt
Polyurethane
Fibrous Felt
Soaks in quickly
Soaks in quickly
Soaks in
20
Processing Interface
21
Evaluating Erosion Data
  • Long Scans Length ? Scan Leveling Drift ?
    Inaccurate Erosion Data
  • Can examine erosion amplitude
  • See planarity in time across entire structure
  • Lose information about feature width effects
  • Must Evaluate shorter scans for accurate erosion
    data

22
Erosion Amplitudes
23
Erosion v Dishing
TWI 711
SUBA IV
24
Erosion Through Material Layers
  • Can optically identify the presence on Ta inside
    Cu lines, but not outside the lines

Ta no Cu
Unpolished
No Ta or Cu
About PowerShow.com