1

Reinhold H. Dauskardt (dauskardt@stanford.edu)

Department of Materials Science and Engineering

Understanding Stress, Chemistry and Molecular Diffusion for Optimized CMP of ULK Dielectrics

ULK Thin-Film MaterialsTaek-Soo Kim, Andrew Thiel, Yusuke Matsuda

Polymers and NanomaterialsMark Oliver, Jeffery Yang, Ruiliang Jia, Ani Kamer

Ultra-Thin Barrier FilmsRyan Birringer

Chi P k I t tiChip Package InteractionsAlex Hsing

Photovoltaic and Flexible Electronic MaterialsVitali Brand, Fernando Novoa

Collaborators: T. Konno and T. YamanakaJSR Micro, Sunnyvale, CA

Road Map for Optimized CMP of Nanomaterials

Defect EvolutionCMP Damage

Diffusion Removal Rate (RR)

k Increase

optimized

CMP

ULK damage

PAD

abrasiveCMP Slurry

Applied CMP shear loadSi

ULK RR

diffusion

nanoporous glass

nanopore

surfactant molecule

micelle

2

Applied CMP down force

Applied CMP shear loadSi

ki

σf

P

Crack Driving ForceGGGG

elastic-plastic contact

Crack Driving Force and Subcritical Cracking

PAD

abrasiveCMP Slurry

ULKcracking delaminationσf

damage initiated by abrasive or pad asperity

CMPcontactfilmtotal GGGG ++=

f

fffilm E

hZG

2σ=

3

2

aEPG

fcontact ⋅

⋅=χ elastic-plastic

contact

film stress

),( pressureshearfnGCMP =

P

In the absence of chemically active environmental species, crack propagates if

2( / )total cG G J m≥

2( / )total cG G J m<

In the presence of chemically active species during CMP, crack propagates if

CMP slurry accelerates defect evolution

nce,

g(S

)

surface or interfacialdefects

volume defects

Reliability and Implications for Processing Yield

nano-scaledefect

uenc

y of

Occ

urre

n

device or processstress

crackingprobability

damageinitiation

0.1 μm

processing stress d h i t

Strength Level, S

Freq

u probability

Depends on defect size and fracture energy, Gc

and chemistry

3

Automated Crack Velocity TestingVapor environment Liquid environment

Port for vapor injection

adhesive/cohesivecrack

Accelerated Cracking

aqueous

aqueous pH 11

pH 4.5 3%H2O2ci

ty, d

a/dt

(m/s

)

10-7

10-6

10-5

10-4

Po

Load

, P

dP/dt

Load Relaxation Crack Growth Technique

Solution containerThermocouple

DTS Delaminator System

aqueouspH 3

10-111 2

Cra

ck G

row

th V

elo

Applied Strain Energy Release Rate, G (J/m2)

10-10

10-9

10-8

10 7

threshold crucial for reliability

LC

rack

Len

gth,

a

Time (s)

da/dt

fracture

path

cap

liner

Low kOSG

compliance analysis

NH4OH 79Citric acid 70TMAH 68DI water 65H2O2 50

Relevance to CMP Damage Thierry FARJOT – Patrick LEDUC. LETI

45 s CMP

NH4OHDI waterCitric acidTMAH

10-4

10-3

a/dt

(m/s

)

NH4OH DI Water Citric Acid TMAH H

2O

2

10-3

NH4 OH

y, d

a/dt

(m/s

)

Characterize crack growth rate to predict damage…

TMAHH2O2

1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.810-7

10-6

10-5

Cra

ck G

row

th R

ate,

da

Applied Strain Energy Release Rate, G (J/m2) 40 45 50 55 60 65 70 75 80 85 9010-5

10-4 DI water

citric acidTMAH

H2 O2

Ave.

Cra

ck V

eloc

ity

Area fraction of Delamination

4

• Low crack growth rates critical for growth of nano-scale defects

• Dominated by threshold behavior in v-G curves

Synergistic effects of CMP slurry chemistry and stress on defect

evolution/crack growth

Relevance to CMP Damage

g

10-5

10-4

10-3

h R

ate,

da/

dt (m

/s)

NH4OH DI Water Citric Acid TMAH H2O2

threshold crucial for CMP

0.1 μm

nano-scaledefect

Crack growth rates <10-10 m/s (below threshold) necessary to

achieve reliable integration1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8

10-7

10-6

Cra

ck G

row

th

Applied Strain Energy Release Rate, G (J/m2)

threshold crucial for CMP defects

CH3 CH3 CH3 CH3I I I I

CH3 – CH – CH2 – C – C ≡C – C – CH2 – CH – CH3I I

CH3 CH3 CH3 CH3I I I I

CH3 – CH – CH2 – C – C ≡C – C – CH2 – CH – CH3I I

Cm En

Dimeric (Gemini) surfactantLow foaming (defoaming) and rapid surface wetting

Linear (bridged) surfactantPolyoxyethylene Alkyl Ethers

Surfactant Effects on Defect Growth and Diffusion

Effects of surfactant molecules on the defect evolution/crack growth are unknown!

CH3 StrainedCH3

CH3 OHOHOHCH3CH3CH3OHOH

OH- OH-

Competition for adsorption sites

at high pH

I IO O CH2 CH2CH2 m CH2 n I I

OH OH

I IO O CH2 CH2CH2 m CH2 n I I

OH OH

Surfactant containing solution

CH3CH3

Crack Growth

Crack Tip Reaction Region

H2O

Bond

OH

OHCH3

3 OH

OHCH3CH3OHOH

HydrophilicHydrophobic

Hydrophilic interaction

Hydrophobic interaction

OH OH

OH-

at high pH

CH3

5

Surfactant Effects on Defect Growth and DiffusionCH3 CH3 CH3 CH3

I I I I CH3 – CH – CH2 – C – C ≡C – C – CH2 – CH – CH3

I IO O CH2 CH2CH2 m CH2 n I I

OH OH

CH3 CH3 CH3 CH3I I I I

CH3 – CH – CH2 – C – C ≡C – C – CH2 – CH – CH3I IO O CH2 CH2CH2 m CH2 n I I

OH OH

Cm En Dimeric (Gemini) surfactant

Linear (bridged) surfactantPolyoxyethylene Alkyl Ethers

10-4

0.1 wt. % surfactant solution10-4

slower growth faster growth

10-9

10-8

10-7

10-6

10-5

10

ck G

row

th R

ate,

da/

dt (m

/s)

pH 7

C8 En

10-9

10-8

10-7

10-6

10-5

10

ck G

row

th R

ate,

da/

dt (m

/s) 420

440

465

DimericpH 7

CH3 CH3 CH3 CH3I I I I

CH3 – CH – CH2 – C – C ≡C – C – CH2 – CH – CH3I IO O CH2 CH2

CH3 CH3 CH3 CH3I I I I

CH3 – CH – CH2 – C – C ≡C – C – CH2 – CH – CH3I IO O CH2 CH2none none

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

10-10

30oC

Cra

c

Applied Strain Energy Release Rate, G (J/m2)

E4E6

E9

Dimeric surfactant accelerates crack growth

CmEn surfactants suppressed crack growth!

Dimeric surfactant decreases diffusion

CmEn surfactants accelerate diffusion!

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

10-10

30oCC

rac

Applied Strain Energy Release Rate, G (J/m 2)

485CH2 CH2CH2 m CH2 n I I

OH OH

CH2 CH2CH2 m CH2 n I I

OH OH

none none

Micellar Bridging in Aqueous Solution

= −tip applied bridgingG G G

10-4

Micellar bridging reduces crack tip driving force

H2O

CH3 CH3 OHOHCH3CH3OHOH

OH-

OH OHCH3

HydrophilicinteractionHydrophobic

interaction

OH– OH–

Competition foradsorption sites

Crack growth

10-9

10-8

10-7

10-6

10-5

10

ck G

row

th R

ate,

da/

dt (m

/s)

pH 7

C8 En

Crack growthOH OH

CH3

OHCH3CH3OH OH OHCH3

OH Crack tip reaction region

OH

OHCH3

OH CH3CH3OH OHCH3

CH3

interactionStrained

bond

Bridging force

CH3

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

10-10

30oC

Cra

c

Applied Strain Energy Release Rate, G (J/m2)

E4E6

E9 micelle

OH

H2O

CH3

OHOH

OH CH3CH3OH

OH-

CH3 Crack tip reaction region

Strained bond

OH

bilayer

CH3

6

Probing Molecular Interactions with AFMDI water pH10 (NH4OH)

)

40

60)

40

60 AFM tip removal

pH10 (NH4OH) +0.1wt% C18E20

100 nm

WaterC18E20

Forc

e, F

(nN

)

Long range surfactant bridging

0

-20

-40

20

WaterC18E20

Forc

e, F

(nN

)

Long range surfactant bridging

0

-20

-40

20

• Si3N4 tips• Spring constant ~ 0.22N/m• Images taken in soft contact mode

surfactant self-assembly

100 nm

Displacement, δ (nm)0 100 200 300 400 500

Displacement, δ (nm)0 100 200 300 400 500

tip applied bridgingG G G= −

10-5

10-4

C18En

Molecular Bridging Contribution

Crack growthOH OH

CH3

OHCH3CH3OH

OH-

OH OHCH3

Bridging force

CH3

11

10-10

10-9

10-8

10-7

10-6

pH 10NH4OH

Cra

ck G

row

th R

ate,

v (m

/s)

100

20

10

w/o surfactant

From AFM measurements:

OH

H2O

CH3

OHOH

OH CH3CH3OH

OH

CH3 Crack tip reaction region

Strained bond

OH

bilayer

CH3

1 2

0( ) 0.93J/mbridging o oG dσ δ χ ε ε= =∫

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

Applied Strain Energy Release Rate, G (J/m2)

ΔGth ~ 0.75 J/m2

Fmax ~ 13.3 pN

Bridging stress: σo = Fmax · A = 2.6 pN/nm2

Bond Areal density: A ~ 1/[(π·(C·r)2] ~ 2 · 10–4 nm–2

7

CmEn Concentration Effects on Crack GrowthpH7 + 0.01wt% surfactant

10-6

10-5

10-4

a/dt

(m/s

)

10-6

10-5

10-4

/dt (

m/s

)pH7 + 0.1wt% surfactant

C18 En

Reduced conc.

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

10-10

10-9

10-8

10-7

30oC

Cra

ck G

row

th R

ate,

da

Applied Strain Energy Release Rate, G (J/m2)

pH 7

E20

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

10-10

10-9

10-8

10-7

30oC

Cra

ck G

row

th R

ate,

da/

Applied Strain Energy Release Rate G (J/m2)

E100

E20

E10

pH7

Phase diagram of water-C12E6 binary mixture

results in different

assembly near crack tip

Applied Strain Energy Release Rate, G (J/m )

Crack Tip Reaction Region

H2O

StrainedBond

OH-

Tem

pera

ture

(°C

)

Concentration

OH

Cra

H2O

CH3

OH

OH

OH

OHCH3

OH

OH CH3CH3

CH3CH3OH

OH

OH-

OH OH

CH3

CH3

Crack tip reaction region

OH

bilayer

Bridging force

CH3

CH3

Crack Growth in Commercial Slurries

10-6

10-5

10-4

a/dt

(m s

−1)

10-6

10-5

10-4

a/dt

(m s

−1)

k =2.35 k =2.6

0.4 0.8 1.2 1.6 2.0 2.410-11

10-10

10-9

10-8

10-7

B3 k = 2.35 23oC

Cra

ck G

row

th R

ate,

da

A li d St i E R l R t G (J −2)

B1 DIW

0.4 0.8 1.2 1.6 2.0 2.410-11

10-10

10-9

10-8

10-7

B3 k = 2.6 23oC

Cra

ck G

row

th R

ate,

da

A li d St i E R l R t G (J −2)

DIW

B1

Applied Strain Energy Release Rate, G (J m 2) Applied Strain Energy Release Rate, G (J m 2)

BMS-B1 BMS-B3

pH 10.5 9-11

Oxidizing agent 0.4 wt% H2O2 1.0 wt% H2O2

Inhibitor BTA BTA

Surfactant O O

Chelate O O

Abrasive (silica) 12.5 – 15 wt% 1 - 30 wt%

8

Effect of Low-k Film Density/Dielectric Constant

10-5

10-4 k = 1.9 2.2 2.6 2.9Bulk SiO

24.5

MSSQ/SiC3.2

CDO/TaN

a/dt

(m/s

)

10-6

10-5

10-4

(m s

−1)

10-10

10-9

10-8

10-7

10-6

Cra

ck P

ropa

gatio

n R

ate,

da

10-10

10-9

10-8

10-7

10

k=2.35DI Water

ck G

row

th R

ate,

da/

dt (

3.02.610

0.5 1 1.5 2 2.5 3 3.5 4

Applied Strain Energy Release Rate, G(J/m2)

Guyer, Patz and Dauskardt, JMR 20060.4 0.8 1.2 1.6 2.0 2.4

10-11

10 DI Water 23oCC

rac

Applied Strain Energy Release Rate, G (J m−2)

Effect of Low-k Film Density/Dielectric Constant

10-6

10-5

10-4

(m s

−1)

SiN (200 nm)

Epoxy

Si

G = 3 1J/m2

10-10

10-9

10-8

10-7

10

2.35DI Water

ck G

row

th R

ate,

da/

dt

3.0k = 2.5delaminated

2.6Delaminated

Si

k=2.5 (555 nm)

SiN (200 nm) Gc = 3.1J/m2

(FPB)

Gc = 1.6 J/m2

(FPB and DCB)

Failure near the bottom interface

0.4 0.8 1.2 1.6 2.0 2.410-11

10 DI Water 23oCC

rac

Applied Strain Energy Release Rate, G (J m−2)

at the bottomDelaminated at the bottom Lower bond density

Fast diffusion path

Further accelerated environment assisted crack growth

9

Road Map for Optimized CMP of Nanomaterials

Defect EvolutionCMP Damage

Diffusion Removal Rate (RR)

k Increase

optimized

CMP

ULK damage

PAD

abrasiveCMP Slurry

Applied CMP shear loadSi

ULK RR

diffusion

nanoporous glass

nanopore

surfactant molecule

micelle

Diffusion of Solutions into ULK Films

Cleavededge

Diffusion front

Diffusion of BMS-B3400

(μm

)

Diffusion

Solutions diffuse into porous films

change RI

Nanoporous ULK

SiN x

Silicon

0 500 1000 15000

100

200

300

No diffusion for k = 2.6 and 3.0

2.06 x 10-8 for k = 2.35

Diff

usio

n D

ista

nce,

x

D [μm2/s] = 4.88 x 10-8 for k = 2.5

0.1mm

Distance, x

x Dt=Fick’s law:

Kim and Dauskardt - Stanford University

0 500 1000 1500Square Root Time, t0.5 (sec0.5)

10

23

Diffusion time, t [hour]

Diffusion of CMP Slurry: Effect of H2O2 AdditionWith H2O2 (1 wt.%)

- Uniform Fickian diffusion- No buckling

Without H2O2

- Severe buckling

69

64

140

235

23.5

64

Cleaved edge300μm 300μm

Analyzing Diffusion Front with XPS

• Carbon content is used to track extent of slurry infiltration

80

Porous DEMS

SiCN

• Chemical analysis shows a clear carbon maximumin the middle of the diffusion front

• Due to complexity of slurrythe exact diffusing speciescannot be determined

40

50

60

70

Car

bon

conc

entra

tion

(%)

Withindiffusion Ahead of

Bulk DEMS average carbon content

Silicon

cannot be determinedusing XPS

0 100 200 300 400 50020

30

C

Distance from sample edge (μm)

d us ofront diffusion front

Edge of sample

11

Diffusion of Solutions into Nanoporous Films

SiNx

Watch solutions diffuse,

change in RIcleaved edge

1.0

1.5

2.0

w/ surfactant

ista

nce,

x (m

m)

Cm En

Porous MSSQ

Silicon

Diffusion front

diffusion

0 200 400 600 800 10000.0

0.5

Diff

usio

n D

Square Root Time, t0.5 (sec0.5)

w/o surfactant

1.5

2.0

n = 4

x (m

m)

C10En

0.1 mm

distance, x

pore dia ~ 2.1 nm

0 200 400 600 800 10000.0

0.5

1.096

Diff

usio

n D

ista

nce,

Square Root Time, t0.5 (sec0.5)

Surfactant Solution Diffusion100 wt% surfactant

10-12

10-11

, D (m

2 s-1)

C E

0.1wt% solution

10-11

t, D

(m2 s-1

)

C12En

102 10310-15

10-14

10-13

Dimeric

Diff

usio

n C

oeffi

cien

t

Molecular Weight, M (g mol-1)

-1.85

-1.96

CmEn

102 103

10-13

10-12 C10En

Dimeric

Diff

usio

n C

oeffi

cien

t

Molecular Weight, M (g mol-1)

-1.31

-0.14

-1.25

Increase in k value after direct CMP

Nanoporous thin film

SiNx

Silicon

Ar ion etching(2 mm x 2 mm)

(1) (2) (3)

XPS scan(0.1 mm x 0.1 mm)

Increase in k-value after direct CMP

(Kondo et. al., IITC, 2007)

12

Mechanism Likely Related to Polymer Reptation~D M α−

α = 2 polymer reptation theory

10-12

10-11

nt, D

(m2 s-1

)

CmEnpolymer melt

surfactant in nanopores

-2

Experimentally,

Data for poly(butadiene)

2.3~D M −

102 10310-15

10-14

10-13

Surfynol

Diff

usio

n C

oeffi

cien

Molecular Weight (g mol-1)

-1.85

-1.96

Dimeric

Jones, Soft Condensed Matter

Molecular Weight (g mol )

diffusion

nanoporous glass

nanoporesurfactant molecule

micelle

Static tube (interconnected nanopores)

Dynamic tube (polymer entanglement)

Road Map for Optimized CMP of Nanomaterials

Defect EvolutionCMP Damage

Diffusion Removal Rate (RR)

k Increase

optimized

CMP

ULK damage

PAD

abrasiveCMP Slurry

Applied CMP shear loadSi

ULK RR

diffusion

nanoporous glass

nanopore

surfactant molecule

micelle

13

Correlations with CMP RemovalRR is inversely proportional to GTH.

103

104

of L

KD

C8En

RR is inversely proportional to D.

600

800

1000

1200

l Rat

es, R

R

# f C 8

~ −thRR G β

2 2

m2 )

100 101 102101

102

C12En

RR

Number of EO, n

m2 s-1

) ~ − −c dRR D M

0 10 20 30 40 500

200

400

Rem

oval

EO Length, n

# of C = 8

# of C = 12

0 20 40 601.2

1.4

1.6

1.8

2.0

2.2

C12En

C10En

Thre

shol

d of

G, G

TH (J

/m

Number of EO, n

Gemini102 103

10-13

10-12

10-11

C10

En

Gemini

Diff

usio

n C

oeffi

cien

t, D

(m

Molecular Weight, M (g mol-1)

C12

En

Defect EvolutionCMP Damage

~Damage v10-10

10-9

10-8

10-7

10-6

10-5

10-4

30oC

Cra

ck G

row

th R

ate,

da/

dt (m

/s)

pH 7

E EE9

0.1 wt% surfactant

10-10

10-9

10-8

10-7

10-6

10-5

10-4

30oC

Cra

ck G

row

th R

ate,

da/

dt (m

/s)

pH 7

E EE9

0.1 wt% surfactant

ect G

row

th R

ate,

da/

dt (m

/s)

Road Map for Optimized CMP of Nanomaterials

Diffusion Removal Rate (RR)d

~ a bthG D M ~ −

thRR G β

g

k Increase

~k D

optimized

CMP

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

Applied Strain Energy Release Rate, G (J/m 2)

E4E6

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

Applied Strain Energy Release Rate, G (J/m 2)

E4E6

Def

e

Driving Force, G (J/m2)

102 103

10-13

10-12

10-11

C10En

Gemini

Diff

usio

n C

oeffi

cien

t, D

(m2 s-1

)

Molecular Weight, M (g mol-1)

C12En

100 101 102101

102

103

104

C12En

RR

of L

KD

Number of EO, n

C8En

~ − −c dRR D M

14

Summary

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

10-10

10-9

10-8

10-7

10-6

10-5

10-4

30oC

Cra

ck G

row

th R

ate,

da/

dt (m

/s)

Applied Strain Energy Release Rate, G (J/m 2)

pH 7

E4E6

E9

0.1 wt% surfactant

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.610-11

10-10

10-9

10-8

10-7

10-6

10-5

10-4

30oC

Cra

ck G

row

th R

ate,

da/

dt (m

/s)

Applied Strain Energy Release Rate, G (J/m 2)

pH 7

E4E6

E9

0.1 wt% surfactant

Def

ect G

row

th R

ate,

da/

dt (m

/s)

Driving Force, G (J/m2)

• Defect Evolution and Damage– fracture of ULK materials– slurry chemistry effects on damage evolution

diffusion

nanoporous glass

nanopore

surfactant molecule

micelle• Diffusion of Chemically Active Solutions

– diffusion of aqueous solutions– effects of nonionic surfactants

• Correlations with CMP Removal Rate

PAD

abrasiveCMP Slurry

Applied CMP shear loadSi

ULK RRdefect evolution rates

– role of slurry chemistry and surfactants– removal, diffusion and