7/25/2007 1

Challenges and Opportunities in CMP Consumables – A Chemist’s View

Yuzhuo LiCenter for Advanced Materials Processing

Department of ChemistryClarkson University

Potsdam, New York 13699yuzhuoli@clarkson.edu

July 2007 CMPUG

Some Interesting Questions

At which technology node CMP will exhaust its usefulness?At which technology node copper will reach its limit as interconnect?What materials could/would be used to replace copper? Still need CMP?What practical considerations one must take for CMP consumables in 32 nm/450 mm processing?

7/25/2007

July 2007 CMPUG

2

Functions of CMP

PlanarizationDielectric (PMD and IMD) CMP, NiP CMP, etc

Formation of micro/nano structuresCu CMP, W CMP, STI CMP, MEMS CMP

Surface conditioningNiP, Sapphire, MgO, etc

7/25/2007

July 2007 CMPUG

3

Which function(s) of the CMP is most sensitive to technology node?

CMP Consumables

SlurryPadpCMP clean solutionPad conditionerRetainer ringCarrier filmFiltersetc

7/25/2007

July 2007 CMPUG

4

Cu CMP Slurry Formulation Strategy- in what order?

OxidizerPeroxides, persulfates, periodates, etc

pHConsider stabilizer stability and Pourbaix diagram

Complexing agentAssist copper dissolution

Passivating agentSuppress isotropic copper dissolution

“Abrasive” particlesSofter the better? What are the roles of particles?

7/25/2007 5

July 2007 CMPUG

Technology Node and PS

7/25/2007

July 2007 CMPUG

6

Fumed silica

Colloidal Silica

Organic?

Organic/inorganic composite can be larger?

Abrasive Particle Size

Removal rateStep height reduction efficiency

Surface QualityDefect count The key to a good polished surface

A balancing act between mechanical and chemical

Mechanical effect

Che

mic

a l e

ffec t

The perfect balanced line

H. MRR, G. SQ

M. MRR, M. ScratchingL. MRR, Chem. Corr.

L. MRR, Scratching

Strong Chemical/Weak Mechanical Strong Chemical/Weak Mechanical

7/25/2007 8

July 2007 CMPUG

Weak Chemical/Mild Mechanical Weak Chemical/Mild Mechanical

7/25/2007 9

July 2007 CMPUG

When there is a balance

Representative results from past studies

7/25/2007 10

July 2007 CMPUG

Importance of Abrasive Particles

7/25/2007

July 2007 CMPUG

11

No abrasive particles

3% 80 nm silica

Optimal Particle Size?

7/25/2007

July 2007 CMPUG

12

3% 20 nm silica particles

3% 130 nm silica particles

7/25/2007 13

Functions of an abrasive

To enhance the mechanical effect of a padMinimum requirement in hardness?What about abrasive-free or abrasive-diet systems?

To serve as a carrierChemically interact with pad, surface to be polished, and all slurry componentsPhysically remove the polishing debris away from the surface

Serve as a particulate lubricantWell-known in classic tribologySpecial effects when the particles are in nanosize

July 2007 CMPUG

7/25/2007 14

Particulate adsorption

Cu

SiO2

Cu2+ Cu2+ Cu2+

CuO

July 2007 CMPUG

Notice the relative size ratio:

Abrasive particle/polishing debris

Passivation Film Formation

7/25/2007

July 2007 CMPUG

15

Cu Surface

BTA molecule

BTA-Cu interaction

BTA-BTA interaction

Where other things fit in here (complexing agent, surfactant, etc)

Interesting question

Will higher MRR/SER ratio always translate to better SHRE?

Is it true that tougher the passivation film the better the SHRE?

Case 1: non-BTA based passivating film, zero static etch rate, > 5000A/min MRR, no step height reduction efficiencyCase 2: surfactant based passivating film, very low static etch rate (<50A/min), high removal rate (>5000A/min), SHRE < 30%

7/25/2007

July 2007 CMPUG

16

Pad

disconnect

wafer

Kaufman Model

Pad

wafer

Step height reducedTrench width kept the same

7/25/2007 177/25/2007 17

July 2007 CMPUG

Pad

wafer

Delamination model

Pad

wafer

Polish debris

No step height reduction orIncreased step height“Trench” width slightly increased

No disconnect

7/25/2007 187/25/2007 18

July 2007 CMPUG

7/25/2007

July 2007 CMPUG

19

Slurry with Higher Viscosity Gives Lower SHRE

Simulated linear velocity difference between the fluid in a recessed area and near the pad. The reduced flow, to certain extent, helps the preservation of the passivating film in the recessed area.

slurry Relative viscosity SHRE (%) Dishing at 100/100 um lines (A)

original 1.00 89 600

Original plus IPA 1.35 65 1200

20

Porous vs. Non-Cell PadsPorous Non-cell

21

Porous vs. Non-cell PadsAfter breaking-in and conditioning

Porous Non-cell

22

Temperature profiles during polishing

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0Polishing time (Sec.)

Tem

pera

ture

(C)

Reference

NCP B

NCP pad temperature is lower during the polishing.

Lower local temperature

Slower to develop glazing effect?

Pad Surface Temperature

23

Due to Greater Heat Capacity?

20

25

30

35

40

45

0 2 4 6 8 10 12

sec

Tem

para

ture

(C)

NCPReference

24

Results and DiscussionBlanket wafer data as a function of pad life

Mipox padperformancestable evenafter 1200 wafers

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 200 400 600 800 1000 1200 1400

Wafer number

MR

R(A

/min

)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Nor

mal

ized

WIW

NU

Mipox IC 1000 Mipox WIWNU IC WIWNU

IC 1000 experiencedsteady drop in MRRfrom wafer 300

Conditioner ruled outas cause for drop inremoval rate

Carrier film changed

25

Blanket wafer data as a function of pad life (contd.)

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 100 200 300 400 500 600

Wafer number

MR

R(A

/min

)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Nor

mal

ized

WIW

NU

Mipox IC 1000 Mipox WIWNU IC WIWNU

MRR starts to decrease after 300 wafers in case of IC-1000

26

Surface quality of polished wafers

0

0.5

1

1.5

2

2.5

3

0 200 400 600 800 1000 1200 1400

Wafer number

RM

S (n

m)

IC 1000 Mipox

Comparable surface roughnessvalues for wafers polished onboth the pads

27

Pad temperature measurement

0

5

10

15

20

25

30

35

40

45

0 200 400 600 800 1000 1200 1400

Wafer number

Tem

pera

ture

(C)

IC 1000 MipoxMaximum pad temperature also suggesting end of life after 300 wafersfor the IC 1000 pad

Consistent maximum temperaturesattained with the Mipox pad

28

End-point times as a function of wafer number

40

50

60

70

80

90

100

110

0 200 400 600 800 1000 1200 1400

Wafer number

End

poi

nt ti

me

(sec

)

IC 1000 Mipox

Consistent end-point times with the Mipox padSteady increase in the end-point times with the IC-1000 pad after wafer 300

29

Patterned wafer resultsPlanarization efficiency as a function of wafer number

0

0.2

0.4

0.6

0.8

1

1.2

0 200 400 600 800 1000 1200 1400

Wafer number

Plan

ariz

atio

n ef

ficie

ncy

Edge(100 micron) Middle(100 micron) Center(100 micron) Edge(100 micron) IC pad Middle(100 micorn) IC 1000 Center(100 micron) IC 1000

Imporved planarization efficiency with Mipox pad due to the higher hardness

30

Dishing as a function of wafer number

0

200

400

600

800

1000

1200

1400

1600

1800

2000

100 200 300 400 500 600 700 800 1150

Wafer number

Dish

ing(

A)

Mipox IC 1000

Lesser dishing in most of the runs with the Mipox pad due to the higher stiffness incomparison to the IC-1000 pad

31

SURFACE IMAGE

Reference ID MD OD

Reference ID MD OD

Magnification : X200

Angle : 45degreesMipox pad

IC 1000 pad

Reference: New unconditioned pad

After 1200 wafers

After 500 wafers

32

SURFACE IMAGE

Reference ID MD OD

Reference ID MD OD

Magnification : X50

Angle : 45degreesMipox pad

IC 1000 pad

Lower Retention of Slurry Chemistry on NCP Pad

33

34

Potential advantages

Higher hardness and modulus valuesImproved pad-lifeHigher planarization efficiency and lower dishing

Non-porous natureFeasibility of process development at lower flow-rates due to the non-porous natureFeasibility of rapidly changing slurry chemistry during Cu CMP process

For more information, see Mipox poster

Summary

For technology node sensitive CMP processes, abrasive particle size will continue to reduce

Smaller the particle size, the true abrasiveness is reduced.Larger the surface area, the surface adsorption property becomes more importantThe abrasiveness of these particles will be transmitted or express via the tips of the padsNon-cell type pads may offer several potential advantages including longer life time

7/25/2007

July 2007 CMPUG

35