Measuring sub-50nm particle retention of UPW filters – The particles are very uniform in size (CV...

CT Associates, Inc. CT Associates, Inc.

Measuring sub-50nm particle retention of UPW filters

Don Grant and Gary Van Schooneveld

CT Associates, Inc.

7121 Shady Oak Road, Eden Prairie, MN 55344

May 2, 2011

1 CTA 1286 2519

CT Associates, Inc. CT Associates, Inc

Introduction

• The critical feature size of state-of-the-art semiconductor devices is on the order of 40 nm and expected to decrease to < 20 nm by 2015.

• Particles half the size of critical features can reduce finished device yield and reliability.

• Particles in UPW that contacts wafer surfaces during processing can deposit on the wafer surface.

• Microfilters and ultrafilters are used to remove particles from these liquids.

• Test methods are needed to measure the filter particle removal efficiency of particles smaller than 50 nm.

CTA 1286 2519 Slide # 2


Outline

• Introduction

• Desired test method characteristics

• Comparison of candidate challenge particle properties

• Examples of retention of candidate test particles by a 30 nm UPW filter

• Summary and conclusions

CTA 1286 2519 Slide # 3


Desired test method properties

• Testing should simulate “real-world” worst case conditions. – Particle size should be well characterized.

– Particle capture should be by sieving only.

– Particles used in testing should be representative of particles found in UPW systems.

• The test procedure should simulate filter performance during its projected lifetime in a reasonable test period – Operate at a face velocity (flow rate/surface area) similar to actual use

conditions.

– Example of a reasonable test duration - simulate 1 year in a 16 hour test.

• The cost per cartridge test should not be prohibitively expensive.

CTA 1286 2519 Slide # 4


Particle capture mechanisms

• Particle capture by filters can result from several mechanisms including: – Diffusion

– Interception

– Impaction

– Electrostatic attraction

– Sieving

• Particle capture should be by sieving only – Worst case capture mechanism

– Capture by diffusion, interception, impaction and electostatic attraction and adsorption should be absent (or nearly absent).

– Desire a strong repulsive force and a weak attractive (Van der Waals) force between the particles and the membrane surface to minimize the potential for adsorption.

CTA 1286 2519 Slide # 5


Removal of particle from UPW by a 0.2 µm rated filter by different capture mechanisms

CTA 1286 2519 Slide # 6

Particle diameter (µm)

0.01 0.1 1

Filt

er

rete

ntio

n (

LR

V)

0

10

20

30

40

Capture by interceptionplus diffusion

Capture bysieving


This study

• Focused on particle type.

• Measured the retention of different types of particles by a 30 nm commercially-available UPW filter cartridges.

• Particle types evaluated – Polystyrene latex (PSL)

– Colloidal gold

– Colloidal silica

• Test conditions employed – Filters were operated at a face velocity of 0.11 cm/min (equivalent to

~1.1 liters/min in a 10” cartridge). • Lower than actual use conditions.

• Chosen due to the high cost of gold particles.

– Inlet particle concentration – 2E8/mL (~6 ppb)

– Total challenge resulted in a fractional filter coverage of 0.2 monolayers

CTA 1286 2519 Slide # 7


Particle comparison

• Size distributions

• Anticipated capture mechanisms

• “Real worldliness”

• Cost

CTA 1286 2519 Slide # 8


Liquid Nanoparticle Sizer Description

Dynamic Mobility Analyzer (DMA)

Condensation Particle Counter (CPC)

Operating Principle •Nebulizer converts the hydrosol to an aerosol. •DMA separates particles according to size. •CPC measures concentrations of particles of each size.

Slide # 9 CTA 1286 2519

FilterChallenge Solution

CDA

UPW

UltrafineNebulizer

Scanning Mobility Particle Sizer (SMPS)ParticleAerosol

Pressureregulator

Drain

Neutralizer DMA CPCmonodispersed

aerosol

CT Associates, Inc. CT Associates, Inc CTA 1286 2519 Slide # 10

50 nm PSL 3050A-3490 - Multiple dilution ratios

Particle diameter (nm)

5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

ren

tial num

be

r co

ncen

tra

tion

d (

#/m

L)

/ d log(D

P)

0.0

5.0e+14

1.0e+15

1.5e+15

2.0e+15

1,800:1

6,000:1

18,000:1

60,000:1

180,000:1

100nm PSL - Multiple dilutions


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

ren

tia

l n

um

be

r co

nce

ntr

atio

n

d (

#/c

m3)

/ d

lo

g(D

P)

0

200

400

600

800

55.4 uL/mL 12/2/09

55.4 uL/mL 12/3/09

50.3 uL/mL 12/3/09

54.3 uL/mL 12/3/09

66.4 uL/mL 12/4/09

63.7 uL/mL 12/4/09

30nm PSL Suspension Stability


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

ren

tial n

um

be

r co

nce

ntr

ation

d (

#/m

L)

/ d

log

(DP)

0.0

5.0e+12

1.0e+13

1.5e+13

2.0e+13

2.5e+13

Initial

18 days

20 nm PSL - Multiple dilution ratios


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

ren

tial num

be

r co

ncen

tra

tion

d (

#/m

L)

/ d log(D

P)

0.0

1.0e+16

2.0e+16

3.0e+16

4.0e+16

5.0e+16

155,000:1

310,000:1

710,000:1

1,430,000:1

Examples of PSL particle size distributions


What size are the 20nm PSL particles?????

CTA 1286 2519 Slide # 11

20 nm PSL - Number weighted distribution


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

ren

tia

l n

um

be

r co

nce

ntr

atio

nd

(#

/mL

) / d

lo

g(D

P)

0.0

1.0e+16

2.0e+16

3.0e+16

4.0e+16

5.0e+16

155,000:1

310,000:1

710,000:1

20 nm PSL - Volume weighted distribution


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

ren

tial volu

me

co

ncen

tra

tion

d (

nm

3/m

L)

/ d log(D

P)

0.0

1.0e+19

2.0e+19

3.0e+19

4.0e+19

5.0e+19

155,000:1

310,000:1

710,000:1


What’s the small stuff in the PSL????

CTA 1286 2519 Slide # 12

50 nm PSL 3050A-3490


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

ren

tia

l n

um

be

r co

nce

ntr

atio

nd

(#

/mL

) / d

lo

g(D

P)

0.0

5.0e+14

1.0e+15

1.5e+15

2.0e+15

1,800:1

6,000:1

18,000:1

60,000:1

180,000:1

Measurement of 0.1% TM40 particles and 1% sucrose


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100D

iffe

ren

tia

l n

um

be

r co

nce

ntr

atio

nd

(#

/mL

) / d

lo

g (

DP)

0

1e+14

2e+14

3e+14

4e+14

5e+14

6e+14

7e+14

8e+14

2,000:1

6,700:1

20,000:1

67,000:1

200,000:1

Particles!!!!


Diafiltered 10nm (9.3nm) Gold particles


5 6 7 8 9 20 30 40 50 60 70 80 9010 100

Diffe

rentia

l num

ber

concen

tration

d (

#/m

L)

/ d lo

g (

DP)

0.0

2.5e+13

5.0e+13

7.5e+13

1.0e+14

1.3e+14

1.5e+14

20 nm (20.3) BBI gold particles


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

rential num

ber

con

centr

ation

d (

#/m

L)

/ d log (

DP)

0.0

2.0e+12

4.0e+12

6.0e+12

8.0e+12

1.0e+13

1.2e+13

1.4e+13

1.6e+13

Slide # 13 CTA 1286 2519

Sizing of gold nanoparticles from BBI

30 nm (30.3) BBI gold particles


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

rential num

ber

con

centr

ation

d (

#/m

L)

/ d log (

DP)

0.0

1.0e+12

2.0e+12

3.0e+12

4.0e+12Mixture of 20 (20.3) and 30 (30.3) nm BBI gold particles


5 6 7 8 9 15 20 25 30 35 40 50 60 70 80 9010 100

Diffe

rential num

ber

con

centr

ation

d (

#/m

L)

/ d log (

DP)

0.0

5.0e+11

1.0e+12

1.5e+12

2.0e+12

2.5e+12


Fraction of particles exceeding size (%)

1 2 5 10 20 30 50 70 80 90 95 98 99

Pa

rtic

le d

iam

ete

r (n

m)

0

5

10

15

20

25

30

35

40

9.3 nm

20.3 nm

30.3 nm

Mean (nm) CV (%) Mean (nm) CV (%)

10 9.3 < 15 8.4 13

20 20.3 < 8 20.8 7.4

30 30.3 < 8 30.5 7.3

Claimed size Measured sizeNominal Size (nm)

Comparison between claimed and measured gold nanoparticle sizes (from BBI)

CTA 1286 2519 Slide # 14


Size distributions of Ludox® Colloidal Silica Particles

CTA 1286 2519 Slide # 15


5 6 7 8 9 20 30 40 50 60 70 80 9010 100

Diffe

ren

tia

l n

um

be

r co

nce

ntr

atio

nd

(#

/mL

) / d lo

g (

DP)

0.00

2.50e+16

5.00e+16

7.50e+16

1.00e+17

1.25e+17

1.50e+17

SM30

HS40

TM40

® Grace Davison


Particle size distributions

• Polystyrene latex – Available in “mono-dispersed” sizes of 20, 30, 40, 50 nm, etc. – Particles smaller than about 50 nm are not very uniform in size.

• The standard deviation of particle diameters is approximately 6 nm.; regardless of mean size. • This results in CVs of ~6% and ~30% for 100 and 20 nm particles; respectively.

– Suspensions contain high concentrations of small particles. – Suspensions contain moderate surfactant concentrations to stabilize the particles.

• Colloidal gold – Available in “mono-dispersed” sizes of 5, 10, 15, 20, 25, 30 nm, and larger. – The particles are very uniform in size (CV around 8%.) – Suspensions contain high concentrations of dissolved species.

• Colloidal silica – Available with median sizes of 12, 18, and 28 nm (possibly also smaller sizes). – The 28 nm particles are very uniform in size (CV < 10%); the 12nm and 18nm are

less uniform. – Contain low concentrations of dissolved species.

CTA 1286 2519 Slide # 16


Particle capture mechanisms

• Particle capture by filters can result from several mechanisms including: – Diffusion

– Interception

– Impaction

– Electrostatic attraction

– Sieving

• Particle capture should be by sieving only – Worst case capture mechanism

– Capture by diffusion, interception, impaction and electostatic attraction and adsorption should be absent (or nearly absent).

– Desire a strong repulsive force and a weak attractive (Van der Waals) force between the particles and the membrane surface to minimize the potential for adsorption.

CTA 1286 2519 Slide # 17


Anticipated particle capture mechanisms • PSL

– Particle zeta potential (-16 mV) predicts a moderate particle-membrane repulsive force in most cases.

– Particle composition predicts a large particle-membrane attractive force in most cases.

– Tests with multiple membrane types indicate significant non-sieving particle capture occurs.

– Non-sieving capture can be eliminated if surfactant is added to the challenge – not a real-world situation.

• Colloidal gold – Particle zeta potential (-11 mV) predicts a low to moderate particle-membrane repulsive force in most

cases.

– Particle composition predicts a moderate particle-membrane attractive force in most cases.

– The solution in which the purchased particles are suspended contains a high concentration of dissolved conductive material.

– Tests with multiple membrane types indicate significant non-sieving particle capture occurs.

– Non-sieving capture can be reduced/eliminated by modifying the surface of the particles.

• Colloidal silica – Particle zeta potential (-16 mV) predicts a moderate particle-membrane repulsive force in most cases.

– Particle composition predicts a weak particle-membrane attractive force in most cases.

– Tests with multiple membrane types indicate little if any non-sieving particle capture.

CTA 1286 2519 Slide # 18


Particle “real-worldliness”

• PSL

– The particles are plastic spheres and are not believed to be representative of particles in UPW systems.

• Colloidal gold

– The particles are not believed to be representative of particles in UPW systems.

• Colloidal silica

– UPW systems are known to contain colloidal silica.

CTA 1286 2519 Slide # 19


Particle comparison

Particle Type

Sizes Available

“Real World”? Sieving only? Cost of

particles per gram

PSL Yes No Can be achieved by adding

surfactant. $1,800

Colloidal Gold

Yes No Can be achieved by surface

modification. $23,000

Colloidal Silica

Yes Yes Yes $0.08

CTA 1286 2519 Slide # 20

Colloidal silica appears to be the best choice.


Filter cartridge testing

• Cartridges were challenged with 3 types of 30 nm particles (PSL, colloidal gold, colloidal silica)

• Three separate cartridges were tested.

• Each cartridge was challenged with multiple particle types. The challenge order was varied amongst the cartridges.

• One cartridge was also challenged with a mixture of 15 nm colloidal gold and 30 nm colloidal silica particles.

CTA 1286 2519 Slide # 21


Filter test system schematic

LNS filter retention overview 03-17-11 Slide # 22

Particlesuspension

Testfilter

UltrafineAtomizer

ScanningMobilityParticle

Sizer

UPW

Pressureregulator

Vent

Liquid Nanoparticle

Sizer


Test Procedure

• The cartridges was flushed until the filtrate approached the system background concentration (106/mL > 10 nm).

• The filter was challenged with 2E8 particles/mL (~6 ppb).

• The challenge concentration was verified.

• The face velocity throughout the test was 0.11 cm/min.

CTA 1286 2519 Slide # 23


Fractional coverage (Monolayers)

0.000 0.025 0.050 0.075 0.100 0.125 0.150 0.175 0.200

Re

tention (

%)

0

20

40

60

80

100

Silica

PSL

Gold

High retention of PSL particles

Low retention of silica particles.

Retention of PSL followed by silica (Cartridge #1)



0.000 0.025 0.050 0.075 0.100 0.125 0.150 0.175 0.200

Re

tention (

%)

0

20

40

60

80

100

Silica

PSL

Gold

High retention of gold particles.


Retention of gold followed by silica (Cartridge #2)



0.000 0.025 0.050 0.075 0.100 0.125 0.150 0.175 0.200

Re

tention (

%)

0

20

40

60

80

100

Silica

PSL

Gold

High retention of gold and PSL

particles.


Initially counting silica rather than PSL particles.

Retention of multiple particle types (Cartridge #3)


Retention of different 30nm particles types by a commercially available

UPW filter cartridges

CTA 1286 2519 27


0.000 0.025 0.050 0.075 0.100 0.125 0.150 0.175 0.200

Re

tentio

n (

%)

0

20

40

60

80

100

Silica

PSL

Gold


0.000 0.025 0.050 0.075 0.100 0.125 0.150 0.175 0.200

Re

tention (

%)

0

20

40

60

80

100

Silica

PSL

Gold


0.000 0.025 0.050 0.075 0.100 0.125 0.150 0.175 0.200

Re

tention (

%)

0

20

40

60

80

100

Silica

PSL

Gold

•3 separate filters were tested. •Each was tested with a sequence of particle types. •In all cases the challenge concentration was 2E8/mL.


Retention of a mixture of 15 nm gold and 30 nm silica by a commercially available UPW filter

CTA 1286 2519 Slide # 28


15 20 25 30 35 40 45 5010

Diffe

ren

tia

l n

um

be

r co

nce

ntr

atio

nd

(#

/mL

) / d

lo

g D

P)

0.0

2.0e+10

4.0e+10

6.0e+10

8.0e+10

1.0e+11

Challenge (3-6 hours)

Filtrate at 3 hours

Filtrate at 6 hours

•Challenged with 3E8/mL gold particles only for first 3 hours •Then 3E8/mL gold + 3E8/mL silica particles for next 3 hours.


Retention of a mixture of 15 nm gold and 30 nm silica by a commercially available UPW filter

CTA 1286 2519 Slide # 29

Time (hours)

0 1 2 3 4 5 6

Re

ten

tio

n (

%)

0

10

20

30

40

50

60

70

80

90

100

15 nm Gold particles

28 nm Silica particles

•Challenged with 3E9/mL gold particles only for first 3 hours. •Then 3E9/mL gold + 3E9/mL silica particles for next 3 hours.

Moderate retention of 15 nm gold particles.

Low retention of 30 nm silica particles.


Particle retention comparison

• Retention of PSL and gold particles was significantly higher than silica particles.

• Retention of silica particles is predominately by sieving while PSL and gold particles are likely removed by several capture mechanisms.

• Silica is the preferred particle type for UPW filter retention testing.

CTA 1286 2519 Slide # 30


Summary and conclusions

• Test methods to measure the retention of sub-50nm particles by UPW filters are needed.

• The methods should:

– Use real-world particles that are removed by sieving ,

– Test the filter under representative conditions of particle concentration, particle loading and face velocity

– Be applicable to cartridges as well as filter samples.

– Not be cost prohibitive.

• Filter retention tests performed with PSL, gold, and silica 30 nm particles indicated that silica is the preferred challenge particle.

CTA 1286 2519 Slide # 31

Measuring sub-50nm particle retention of UPW filters – The particles are very uniform in size (CV...

Documents

Transcript of Measuring sub-50nm particle retention of UPW filters – The particles are very uniform in size (CV...