Particle Agglomeration Mechanisms in
CMP Slurries
Mark Litchy and Don Grant
CT Associates, Inc.
CMP Users Conference 2006
February 16, 2006
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 1
Introduction
•
•
•
•
Some CMP slurries are said to be “shear-sensitive”, implying that if
the slurries are exposed to excessive shear stresses, the particles in the
slurries will agglomerate and the slurry will be “damaged.”
Traditionally, bellows and diaphragm positive displacement pumps
and vacuum-pressure systems have been widely accepted means of
bulk slurry delivery. Positive displacement pumps are generally
accepted as low shear devices due to their relatively low speeds of
operation, while centrifugal pumps, which typically operate at higher
speeds, are usually perceived as high shear devices.
Centrifugal pumps have been thought to impart high shear on the
slurry causing it to form gels.
However, the fluid dynamic conditions that lead to high shear stresses
often also increase the probability of fluid cavitation. This test work
was performed to try to separate the effects of shear and cavitation by
holding shear stresses nearly constant while changing the probability
of cavitation. The results suggest that cavitation may play a more
significant role in agglomeration of slurry particles than shear.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 2
Shear?
•
•
•
High velocity gradients may enable agglomeration by imparting
enough energy to overcome the particle repulsive forces thereby
enabling particle collisions.
However, shear forces from too high a velocity gradient can break up
loosely bonded agglomerates.
Many industrial processes use shear to break up agglomerates.
– The velocity differential results in shear stresses being imposed on an
agglomerate, causing it to break apart.
•
•
Thus, there may exist a shear threshold beyond which some
agglomerates are broken up.
Competition between coagulation and fragmentation during shear may
result in a self-preserving size distribution as has been observed during
shear-induced flocculation of PSL particles. (Spicer et al., Wat. Res. 30(5) 1996)
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 3
Cavitation?
•
•
•
The fluid dynamic conditions that lead to high shear stresses also
increase the probability of cavitation.
The forces in collapsing bubbles formed by cavitation are much larger
and may be more likely to cause agglomeration.
High fluid velocities over surfaces result in reduced pressures in the
liquid.
– If the pressure in the liquid is reduced below that of the liquid vapor
pressure, vaporous cavitation can occur in which bubbles of the liquid are
formed.
– If the pressure is reduced below the equilibrium vapor pressure of
dissolved gases in the liquid, gaseous cavitation can occur in which
bubbles of the dissolved gas are formed.
•
If the pressure in the liquid is subsequently increased, the bubbles will
collapse violently.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 4
Effect of charge
• The chemical composition of CMP slurries is such that the
particles in the slurries carry a high surface charge in order
to minimize agglomeration.
• Substantial forces are required to “push” particles in the
slurries close enough together to overcome repulsive
electrostatic forces and cause the particles to agglomerate.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 5
Double Layer
Schematic
Electronegative Particle
Nernst Potential
Zeta Potential
Electric Potential
Surrounding
the Particle
Stern Layer
Diffuse Layer
Ion Concentration
Profile
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 6
DVLO Theory
3
Relative interaction energy
Electrostatic repulsive forces
2
Combined forces
1
0
-1
Vander Waals forces
-2
Primary minimum
-3
0
1
2
3
4
5
Relative separation distance
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 7
DVLO Theory
3
0.010
2
Relative interaction energy
Relative interaction energy
Electrostatic repulsive forces
Combined forces
1
0
-1
Vander Waals forces
-2
Electrostatic repulsive forces
0.005
0.000
Secondary minimum
Combined forces
-0.005
Vander Waals forces
Primary minimum
-3
0
1
2
3
4
5
-0.010
0
Relative separation distance
5
10
15
Relative separation distance
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 8
20
Effect of ionic strength on interaction energy
Hamaker constant 10
-13
ergs, surface charge 50 mv, 50 nm particles
Repulsion
1.0e-14
10-7 molar
10-5 molar
10-6 molar
5.0e-15
10-4 molar
0.0
Attraction
Interaction energy (ergs)
1.5e-14
-5.0e-15
10-3 molar
-1.0e-14
-1.5e-14
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
Distance from surface (µm)
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 9
Rate of agglomeration dependent on many
factors including:
•
•
•
•
•
Particle concentration
Size distribution
Velocity gradients
Particle zeta potential
Solution ionic strength
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 10
Other factors affecting particle agglomeration
in slurries
•
•
•
•
Lack of humidification
Absorption of CO2
pH shock due to improper dilution
Entrainment of air during mixing
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 11
Particle agglomeration caused by bellows and
diaphragm pumps
108
0 turnovers
30 turnovers
101 turnovers
314 turnovers
574 turnovers
1105 turnovers
107
106
Cumulative Concentration (#/mL)
Cumulative Concentration (#/mL)
108
105
104
103
0 turnovers
26 turnovers
86 turnovers
260 turnovers
512 turnovers
823 turnovers
107
106
105
104
103
102
102
1
10
1
10
Particle Diameter (µm)
Particle Diameter (µm)
Diaphragm pump
Bellows pump
Litchy MR and Schoeb R. Semiconductor International, 27(12) 2004.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 12
No agglomeration observed with BPS-3 pump
Cumulative Concentration (#/mL)
108
0 turnovers
30 turnovers
96 turnovers
316 turnovers
757 turnovers
1218 turnovers
107
106
105
104
103
102
1
10
Particle Diameter (µm)
Centrifugal pump
Litchy MR and Schoeb R. Semiconductor International, 27(12) 2004.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 13
Comparison of particle agglomeration from
the pumps
Ratio of Measured Particle Concentration to
Initial Particle Concentration
100.0
Bellows pump
Diaphragm pump
Levitronix pump
10.0
> 0.56 µm
1.0
As-received slurry plotted at 1.1 turnovers
0.1
1
10
100
1000
10000
Turnovers
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 14
Centrifugal pumps used as artificial blood
pumps
• Shear-optimized centrifugal pumps were found to cause
significantly less hemolysis (destruction of blood-cells)
than peristaltic pumps.
• Blood damage caused by shear forces, is remarkably low
as long as the shear level stays below a threshold of
approximately 400 Pa. If this threshold is exceeded, the
hemolysis rate increases abruptly.
• In a shear-optimized centrifugal blood pump the shear
level always stays below this hemolysis threshold. In a
peristaltic pump however, the blood in the occlusion area
is exposed to forces which massively exceed the critical
level and rupture the blood cells.
Paul et al., Artificial Organs 27(6) June 2003.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 15
• The same problem occurs in the valves of diaphragm and
bellows pumps. Most of the fluid, which is pumped by
traditional CMP “dispense engines”, is exposed to very
low shear-force levels.
• However, a very small fluid portion, which is trapped in
the valves during closure, sees tremendously high
pressures of up to 106 Pa.
• In addition, the operation of check valves within the pumps
are also areas where cavitation may occur.
• In a centrifugal pump, the whole fluid volume is exposed
to moderate shear levels of 102-103 Pa generated by the
rotating pump vanes.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 16
Cavitation
• Cavitation is the process of bubble formation resulting
from a reduction in pressure on a liquid.
• Vapor cavities form when the ambient pressure at a point
in a liquid falls below the liquid’s vapor pressure.
• The vapor cavities collapse when they reach regions of
high pressure. The shock waves that are created when the
cavities collapse causing damage to components near the
collapse.
• Cavitation can be identified by following characteristics:
–
–
–
–
Noise
Vibration
Loss in performance
Material failure
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 17
Bernoulli’s Equation
P1 + ρu + ρgh1 = P2 + ρu + ρgh2 = constant
1
2
2
1
1
2
2
2
• Where:
–
–
–
–
–
P = pressure
ρ = density
u = velocity
g = gravitational acceleration
h = height
• Assumptions:
– Steady, inviscid, incompressible flow along a streamline
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 18
D1 < D0
u1 > u0
P1 < P0
Young FR (1999). “Hydrodynamic Cavitation,” in Cavitation, Imperial College Press, pp 187-317.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 19
The cavitation number
P0 − PV
NC =
1 2
ρu
2
where NC = Cavitation number
P0 = System pressure
PV = Vapor pressure
½ρu2 = Dynamic pressure
ρ = Fluid density
u = Fluid velocity
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 20
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 21
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 22
Places where cavitation can occur
•
•
•
•
•
•
Pumps
Venturies
Flow around fixed objects
Orifices
Valves
Pipe bends
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 23
Cavitation in valves
Young FR (1999). “Hydrodynamic Cavitation,” in Cavitation, Imperial College Press, pp 187-317.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 24
Experimental
• Objective: Vary probability of cavitation while
maintaining a constant shear stress
• Method: The BPS-3 inlet pressure was varied while
maintaining a constant pump speed. This allowed the
probability of cavitation on the pump inlet to be varied
while having a minimal effect on shear stress.
• This was accomplished by changing the length and
diameter of the tubing between the feed tank and the pump.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 25
Test conditions investigated
• The following parameters were varied in water to define
the range of operating conditions to be investigated in
slurry.
– Pump speed (3000-8000 rpm)
– Restrictor between feed tank and pump: length (1.5 inches to 15
feet) and diameter (3/8” to ¾”) of tubing
• The BPS-3 was operated briefly at each combination of
above parameters to characterize their effect on:
–
–
–
–
Pump inlet pressure
Pump outlet pressure
Flow rate
Cavitation
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 26
System schematic
Humidified
Nitrogen
Chiller
P
Slurry
Tank
Restrictor
P
Flowmeter
Sample
Valve
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 27
Operating Conditions
•
Two extreme conditions are presented:
1. Low probability of cavitation:
•
No restrictor on inlet of pump, inlet pressure slightly positive (0-1 psig)
2. High probability of cavitation:
•
•
Slurry used: Cabot SS-12
–
•
•
•
•
•
15 foot length of ½” tubing used as restrictor on inlet of pump (-24 in Hg or –
12 psig)
Slurry from the same drum was used in each test.
Levitronix BPS-3 pump was operated at 7000 rpm.
Flow rate varied from 15 to 20 lpm over this range of conditions.
The restriction on the pump outlet was such that the pump discharge
pressure was 15 psig with no restriction on pump inlet.
The slurry particle size distribution (PSD) was monitored over time for
each inlet pressure tested. Samples were analyzed for mean particle
size using a PSS NiComp ZLS380 and for large particle
concentrations (> 0.56 µm) using a PSS AccuSizer 780.
Each test was continued until the slurry had passed through the pump
about 6,000 times.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 28
Normalized Cumulative Concentration
PSDs: Low probability of cavitation
Test conditions:
Pump inlet pressure: 0 in Hg
Pump outlet pressure: 27 psig
Pump speed: 7000 rpm
Flow rate: 20 lpm
Turnovers
10
32
72
105
286
1580
2260
3810
7000
1
10
Particle Diameter (µm)
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 29
Normalized Cumulative Concentration
PSDs: High probability of cavitation
Test conditions:
Pump inlet pressure: -24 in Hg
Pump outlet pressure: 15 psig
Pump speed: 7000 rpm
Flow rate: 15 lpm
Turnovers
10
30
106
360
1390
1930
3000
6230
9275
1
10
Particle Diameter (µm)
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 30
Change in concentration with time:
Low probability of cavitation
Test conditions:
Pump inlet pressure: 0 in Hg
Pump outlet pressure: 27 psig
Pump speed: 7000 rpm
Flow rate: 20 lpm
Normalized Cumulative Concentration
> 0.56 um
> 0.70 um
> 1.0 um
> 2.0 um
> 5.0 um
100
101
102
103
104
105
Turnovers
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 31
Change in concentration with time:
High probability of cavitation
Test conditions:
Pump inlet pressure: -24 in Hg
Pump outlet pressure: 15 psig
Pump speed: 7000 rpm
Flow rate: 15 lpm
Normalized Cumulative Concentration
> 0.56 um
> 0.70 um
> 1.0 um
> 2.0 um
> 5.0 um
100
101
102
103
104
105
Turnovers
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 32
Low probability of cavitation repeatability
Normalized Cumulative Concentration
Test 3: Initial
Test 3: Final
Test 5: Initial
Test 5: Final
1
Particle Diameter (µm)
CMP Users Conference 2006
10
Test conditions:
Pump inlet pressure: 0 in Hg
Pump outlet pressure: 27 psig
Pump speed: 7000 rpm
Flow rate: 20 lpm
CMPUC 2006.ppt
Slide 33
High probability of cavitation repeatability
Normalized Cumulative Concentration
Test 1: Initial
Test 1: Final
Test 4: Initial
Test 4: Final
Test 6: Initial
Test 6: Final
1
Particle Diameter (µm)
CMP Users Conference 2006
10
Test conditions:
Pump inlet pressure: -24 in Hg
Pump outlet pressure: 15 psig
Pump speed: 7000 rpm
Flow rate: 15 lpm
CMPUC 2006.ppt
Slide 34
Summary
• Minimal increases in the large particle tail were observed
during 2 tests conducted at higher inlet pressure (low
probability of cavitation).
• Significant increases (factor of 5-10x) in the large particle
tail were observed for supermicron size particles at a very
low pump inlet pressure (high probability of cavitation).
• No significant difference was observed in the working
PSD during tests performed at low and high probability of
cavitation.
• These results suggest that cavitation may play a more
significant role in agglomeration of slurry particles than
shear.
CMP Users Conference 2006
CMPUC 2006.ppt
Slide 35