FLCC November 3, 2003 F.M. Doyle, D.A. Dornfeld, J.B. Talbot 1 Feature Level Compensation and Control: Chemical Mechanical Planarization Investigators

November 3, 2003 F.M. Doyle, D.A. Dornfeld, J.B. Talbot

1

FLCC

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


2

FLCC

Students Involved in CMP EffortMechanical 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)


3

FLCC

Outline

Objective of FLCC effort in CMP Background on CMP Mechanical phenomena Interfacial and colloidal phenomena Chemical phenomena Modeling efforts Future work


4

FLCC

Objective of FLCC effort in CMP


5

FLCC

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


6

FLCC

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.


7

FLCC

Background on CMP


8

FLCC

w p :pad rotation

tablepad

slurry feedconditioner

head

w w : wafer rotationOscillation

F : down force

Backing film

Retainerring

Wafer

Wafer Carrier

Pad

Pore Wall

Abrasive particle

CMP Process Schematic

Electro plated diamond conditioner Typical pad

Wall

Pad

Pore


9

FLCC

Via formation (Metal CMP)

Shallow Trench Isolation (STI CMP)Multilevel Metalization

Interconnection(ILD CMP)

Applications of CMP – ILD, Metal and STI CMPApplications of CMP – ILD, Metal and STI CMP


10

FLCC

Idealized CMP

Silicon wafer

Polishing pad

Abrasive particle

‘Softened’ surface by chemical reaction

Pad asperity


11

FLCC

CMP Parameters

PadFiber Structure Conditioning Compressibility Modulus

SlurrypH OxidizersBuffering Agents, Abrasive ConcentrationAbrasive Geometry andSize Distribution

Wafer Geometry and

Materials

ProcessPressure VelocityTemperatureSlurry Flow Polish Time

WIWNU(Within-Wafer Non-Uniform

Material Removal)

WIDNU(Within-Die Non-

Uniform Material Removal)

Material Removal

Surface Quality

Roughness, Scratching

CMP

Input Parameters Output ParametersWafer

Die

Surface


12

FLCC

Interactions between Input Variables

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

Interactions Wafer Pad Abrasive Slurry chemical

Wafer - √ √ √

Pad √ - √ √

Abrasive √ √ - √

Slurry chemical

√ √ √ -

Recognize that wafer is heterogeneous, and there may not be a single interaction between it

and other input variables


13

FLCC

Chemical Mechanical Planarization

Mechanical Phenomena

Chemical Phenomena

Interfacial and Colloid

Phenomena


14

FLCC

Mechanical Phenomena


15

FLCC

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


16

FLCC

Stribeck Curve and Characteristics of slurry film thickness

Fric

t ion

coe

f fic

ien t

Film

thi

ckne

ss

Pressure

VelocityViscosity ⋅Hershey 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


17

FLCC

Mechanical and Chemical Material RemovalEffects vs Slurry Film Thickness

Source: Yongsik Moon and David A. Dornfeld, “Investigation of Material Removal Mechanism and ProcessModeling of Chemical Mechanical Polishing (CMP),” Engineering Systems Research Center (ESRC),Technical Report 97-11, University of California at Berkeley (September 1997)


18

FLCC

Effect of gap on CMP - material removal

Preston’s equation:

(C=Preston’s coefficientP = pressure, V=velocity,h=removed height, s=sliding distantt= time)

• Material removal per sliding distance

MR

R v

s sl

idin

g d

ista

nce

, A

/mete

r


19

FLCC

Proposed Pattern-Pad Contact Mode

Oxide

Pad

Slurry

Wafer

Semi-direct or Hydroplaning Contact

Direct Contact

Increasing Hershey Number

Increasing Line Density


20

FLCC

Die LayoutNum of

FeaturesLine Pitch

(m)Density

(%)

1 2000 13

2 1250 24

3 440 60

3 500 66

5 200 75

A

B

C

D

E


21

FLCC

Results – Line Density Effect

Increasing Hershey Number

Increasing Line Density


22

FLCC

Scale effects – Abrasives/Pad

• UR100 from Rodel

1500m

100m

400m

End fibrils

Vertically oriented pores

Urethane impregnated polyester felt

Abrasive particle(<0.1m)


23

FLCC

CMP pad topography (IC1400 K-Groove)

New pad(open pores/smooth wall)

After Stabilization(open pores/rough wall)

After CMP(glazed pores/smooth wall)

After Conditioning(open pores/rough wall)

300um

3.Groove(side view)

Topography of pad surface(source : Rodel (Zygo profile))

1.Pore 2.Wall

Pad

Wafer

PoreWall

Abrasives

Wafer

Pad


24

FLCC

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

The Need for Conditioning

100 µm

Source: B. J. Hooper, UCD, Dublin, 2001

New

Conditioned

Glazed


25

FLCC

Groove & Pore

Wall

Protrusion for rough wall

200um

Soft Material(i.e. low stiffness)

Top View

Side View

Hard Material(i.e. high stiffness)

50um(space)

Schematic of SMART pad


26

FLCC

Model Implementation - Pad Design

SMART pad surface

Polymer pad surface

200um

Polyethylene pad surface

Pad

Wafer

PoreWall

Abrasives

Wafer

Pad


27

FLCC

Interfacial and colloidal phenomena


28

FLCC

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

Surface

Dissolution products

Abrasive particle

Surface dissolution

Ref. L. M. Cook, J. Non-Crystalline Solids, 120, 152 (1990).


29

FLCC

Electrical Double Layer of Abrasive Particles

+ +

++

+

+ + +++

+

++

++

+

+

+

+

+

+

a

+

+

+

Distance

Pote

ntia

l

1/

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


30

FLCC

Colloidal effects

• 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)

Pol

ish

ing

rate

(

/min

)

Colloid oxide

Gla

ss p

olis

hin

g ra

te (m

/min

)

Oxide Isoelectric point


31

FLCC

6

EKC Tech alumina slurry

-35

-25

-15

-5

5

15

25

35

45

55

65

3 4 5 6 7 8 9 10 11

pH

Zeta Potential (mV)

0

500

1000

1500

2000

2500

3000

3500

Effective Particle Size (nm)


32

FLCC

Chemical phenomena


33

FLCC

Chemistry of Glycine-Water System pKa1=2.350 pKa2=9.778+H3NCH2COOH +H3NCH2COO- H2NCH2COO-

Cation: H2L+ Zwitterion: HL Anion: L-

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

0 2 4 6 8 10 12 14 16pH

E, V vs. SHE

Cu2+

CuL2CuL

+

CuO

22-CuO

Cu2OCu

Potential-pH diagram, with {CuT} = 10-5, {LT} = 10-2

Cu(II) glycinate complexes

•Cu(H3NCH2COO)2+ : CuHL2+

•Cu(H2NCH2COO)+ : CuL+

•Cu(H2NCH2COO)2 : CuL2

Cu (I) glycinate complexes

•Cu(H2NCH2COO)-2 : CuL2

-

H-O-H

H-O-H

N-H2H2-C

OO=C

C-H2H2-N

C=OO

Cu2+


34

FLCC

Cu-Glycine

i, A/m2

10-4

10-3

10-2

10-1

100

101

102

103

E mV vs. SHE

-800

-600

-400

-200

0

200

400

600

800

1000

1200

1400

1600

1800

pH 4

pH 9

pH 12

0

20

40

60

80

100

120

140

160

0 1 2 3 4 5 6

H 2 O 2 , wt%

Removal Rate, nm/min

Dissolution Rate

Polish Rate

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


35

FLCC

H2O2 concentration, wt%0 1 2 3 4 5 6concentration of copper ion, mg/L0

20406080100120

Dissolved Copper, nm05001000150020002500

50 minutes40 minutes30 minutes20 minutes10 minutes

Copper concentration, mg/l, and dissolved copper, nm, in unbuffered aqueous glycine (pH 4.5)

sampling time, minutes0 10 20 30 40 50 60

copper ion concentration, mg/L020406080100120

Dissolved Copper, nm050010001500200025000.5 wt% H2O20.75 wt% H2O21 wt% H2O23 wt% H2O25 wt%H2O2


36

FLCC

Modeling Efforts


37

FLCC

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


38

FLCC

Current Status of Modeling Efforts

Abrasive Particle Interaction

Pattern Density Effects – STI

Pattern Density Effects – Copper

Pattern Density Effects – Oxide

Hydrodynamics

Pad Stress Dynamics

Kinematics

Material Removal – Chemistry

Material Removal – Slurry/Abrasive/Pad

Environmental Issues

Modeling Aspect

Feature/ Particle- Level

Chip/Die-Level

Wafer-Level

Global

Interaction Scale

50%

40%

20%

90%

30%

< 5%

90%

50%

50%

< 5 %

Estimated Completion


39

FLCC

MRR

Non-uniform

ity

Slurry/W

ater usage

Abrasive

Concentration

Other C

hemical

Effluent

Energy

References

Wafer Level

Feature Level

Die Level

Empirical Model

Preston

Boning

Others

Individual Model

Tribology

Kinematic

Pad

Chemical

Integrated Model

Preston, 1927

Boning, et al, 1997-1998 (4)

Various (10) incl. Burke, Runnels, Zhao

Various (11)

Various (2)

Various (4)

Various (6)

Literature Review of Modeling of CMP


40

FLCC

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)


41

FLCC

Past 2-D Material Removal Rate (MRR) Models

•Experimental Model [1]: Preston’s Equation MRR= KeP0V + MRR0 where Ke an all-purpose coefficient, MRR0 a fitting parameter, P0, down pressure, and V, the relative Velocity.

•Analytical Model Considering Wafer-Pad Contact Area[2]: Zhao’s Equation

V

MRR= Ke(P0-Pth)2/3V, where Pth a fitting parameter.

Active abrasive number is proportional to contact area. Contact area P0

2/3

*All with an all purpose factor Ke to 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

Contact Area A

V

P0

P0

Wafer

Pad

Wafer


42

FLCC

Sub-Model not Included in Current Model

Mechanical Model

Wafer-Pad Interaction

Abrasive-Pad Interaction

Abrasive-Wafer Interaction

Model of Wafer Properties

Model of Pad Properties

Model of Slurry Abrasive Properties

Force Applied on Abrasives

Number of Active Abrasives

Material Removal by a Single Active Abrasive

Chemical Model

Mechanical-Chemical Interaction Model:

Relationship not Included in Current Model or Unimportant Relationship

Strong Relationship Included in Current Model Weak Relationship Included in Current Model

Sub-ModelModel Output

MRR

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.

Wafer-Chemical Interaction: Passivation Rate Model

Abraded Material-Chemical Interaction (Dissolution) Model

Inputs and Outputs

Wafer Surface Hardness Model

Chemical-Pad Interaction Model

Chemical-Slurry Abrasive Interaction Model

Enhancement of Mechanical Elements (Indentation, Leading Edge Area) on Passivation

Competition between Mechanical Removal and Passivation


43

FLCC

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)

abrasives

growth rateremoval rate

F 2a2

(a) (b) (c)

2

Shaped area is leading edge with aggressive chemical reaction

softersurface

hardersurface

Three Different Regions of Material Removal Rate With Increase of Material Removal Rate


44

FLCC

Process Modeling “Roadmap”

Particle-Scale Material Removal Model

Wafer-Scale Pressure and Velocity Distribution

Die-Scale Pressure Distribution Model

Consumable Parameters

Layout Density

Macroscopic Contact Mechanics Model (Weight Function)

Pattern Density Window and Effective Pattern Density

Material Removal Rate

Surface Quality

Die-Scale Topography (both vertical & lateral directions)

Upper Stream Wafer-Scale Topography

Wafer-Scale Topography Upper Stream

Die-Scale Topography

CMP Tool Configurations

Pattern Density Effect

Dishing Erosion

Dummy FillingCircuit PerformanceECAD

Include device design


45

FLCC

Process Models - Basis for Environmental Analysis


46

FLCC

Future Work


47

FLCC

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

Documents

FLCC November 3, 2003 F.M. Doyle, D.A. Dornfeld, J.B. Talbot 1 Feature Level Compensation and Control: Chemical Mechanical Planarization Investigators