performance of an entegris phasor x heat exchanger in

CTA pub #75.pptSlide 1

Performance of an Entegris pHasor® X Heat Exchanger in Cabot Semi-Sperse® 12

Mark Litchy, Dennis Chilcote and Don GrantCT Associates, Inc.

Bipin Parekh, Annie Xia, Michael Clarke, and Russ MollicaEntegris, Inc.

2008 CMP Users ConferenceFebruary 12, 2008


Introduction

• Heat exchangers can be used to efficiently remove heat from liquid delivery systems, such as those incorporating centrifugal pumps.

• The heat exchanger evaluated is constructed entirely of PFA to provide– High purity– Chemical compatibility

• Even though it is constructed of PFA, it is efficient at removing heat with an acceptable pressure drop.

• However, in slurry applications, there is concern that the heat exchanger may damage the slurry or clog.


Introduction

• Materials of construction and design are important:– Thermal conductivity– Corrosion resistance– Capable of withstanding thermal stresses and high pressures

• Since PFA has poor thermal conductivity properties relative to most metallic heat exchangers– Tube thickness must be thin– Exchanger must be well designed to prevent flow channeling and

streaming– Surface area must be large

• Tradeoffs:– Reducing the tube thickness reduces the operating temperature and

pressure rating of the device– Increasing the surface area increases cost and size of the device


Heat exchanger configurations


Experiments performed in water

• Measurement of differential pressure drop (∆ P) in the tube and shell as a function of flow rate.– PHX03U (U configuration, short: surface area = 0.3 m2)– PHX08S (S configuration, long: surface area = 0.8 m2)– PHX08U (U configuration, long: surface area = 0.8 m2)

• Measurement of heat transfer coefficients at a series of flow rates for 3 different heat exchangers.– PHX03U– PHX08S– PHX08U


Schematic of pressure drop versus flow rate test system

circulationpumpwater tank

T

heat exchanger

flow controlvalve

FM

P

∆P

flowmeter

thermistor

differentialpressure gauge


Calculation of heat transfer coefficient

• where:– U = heat transfer coefficient (Btu/hr/ft2/°F)– Q = flow rate on the tube side (lb/hr)– c = specific heat of water (Btu/lb/°F) = 1.0– ∆Ttube = Ttube in- Ttube out (°F)– A = surface area of heat exchanger (ft2)– ∆Tln = logarithmic mean temperature difference =

[∆T0 – ∆TL]/[ln(∆T0/∆TL)] • where: ∆T0 and ∆TL= temperature differences between the hot

and cold fluids at the two ends of the heat exchanger, x = 0 andx = L

ln

tube

TA T c QU

∆

∆=


Schematic of heat transfer coefficient test system

circulationpumpwarm

water tank

to drain

cold water supply

T1

T3

T4

heat exchanger

FM

T2

flow controlvalve

to mainwater loop

flow controlvalve

regulator

FM

inletfilter

FM

T1

flow meter

thermistor #1


Experiments performed in slurry

• Measurement of the effect of the heat exchanger on the slurry particle size distribution (PSD) of Cabot Semi-Sperse® 12 (SS-12).– PHX08U (U-line long)

• 38.5 lpm• 10 lpm

– Stainless steel coil (control test)• 38.5 lpm


Slurry test system schematic

BPS4

T2

T4T5

Heat Exchanger

Chiller

flow meter

∆P

flow meter

SS12

Humidified N2

T3

T1


Details of tests performed in slurry

• Tests at 38.5 lpm (PHX08U and control test):– Test system volume: 50L of SS-12– Duration: 5 days (or ~5,700 turnovers)– Pump outlet pressure: 34 psig

• Test at 10 lpm (PHX08U test only):– Test system volume: 29L of SS-12– Duration: 10 days (or ~5,700 turnovers)– Pump outlet pressure: 3 psig

• All tests:– Tank blanketed with humidified N2: RH > 90%– Slurry temperature: 21 ± 1°C– SS-25 was filtered using an Entegris Planargard CMP5 10” filter prior to the

test– SS-25 was then diluted with ultra pure water as it was added to the test

system tank.


Particle size measurement

• “Working” particle size distribution– Measured using dynamic light scattering– Instrument used – NICOMP 380ZLS (Particle Sizing Systems)– All particles in a defined volume illuminated simultaneously– Particles are sized by measuring their diffusion coefficient– Measures relative concentrations– Sensitive to about 1% by volume

• “Large particle tail” size distribution– Instrument used:

• AccuSizer 780 sensor (Particle Sizing Systems)– Uses a combination of light scattering and light extinction to size

particles ≥ 0.56µm– Requires dilution

• CMP Slurries contain >1014 working particles/ml• The large particle tail contains ~106 particles/ml (≥ 0.56µm)


AccuSizer dilution system schematic

CirculationPumpUltrapure Water

Tank

Static Mixer

Drain

PIAccuSizer

Sensor

SlurrySample

FilterInjection Pump

Bypass Loop


Working particle size distributions (PSDs)

Particle Diameter (µm)

0.02 0.03 0.04 0.06 0.08 0.2 0.3 0.4 0.6 0.80.01 0.1 1Rel

ativ

e V

olum

e-W

eigh

ted

Diff

eren

tial C

once

ntra

tion

(%)

0

5

10

15

20

25


Results


Comparison of ∆ P vs. flow rate: shell side

Flow rate (lpm)

0 5 10 15 20 25

Diff

eren

tial P

ress

ure

Acr

oss

Hea

t Exc

hang

er (p

sid)

0

1

2

3

Diff

eren

tial P

ress

ure

Acr

oss

Hea

t Exc

hang

er (k

Pa)

0

5

10

15

20PHX03U-shellPHX08U-shellPHX08S-shell


Comparison of ∆ P vs. flow rate: tube side

Flow rate (lpm)

0 5 10 15 20 25

Diff

eren

tial P

ress

ure

Acr

oss

Hea

t Exc

hang

er (p

sid)

0

1

2

3

Diff

eren

tial P

ress

ure

Acr

oss

Hea

t Exc

hang

er (k

Pa)

0

5

10

15

20PHX03U-tubePHX08U-tubePHX08S-tube


Summary of differential pressure results

• Differential pressure drop versus flow rate were similar for both the PHX08U and PHX08S configurations.

• The ∆ P across tubes for the PHX03U configuration was significantly lower.


Heat transfer coefficients vs. flow rate: PHX03U

Flow Rate (lpm)

0 5 10 15 20 25

Hea

t Tra

nsfe

r Coe

ffic

ient

(Btu

/hr/f

t2 /F)

50

75

100

125

150

175

200

Hea

t Tra

nsfe

r Coe

ffic

ient

(W/m

2 /K)

400

600

800

1000Vary tube flow rate, constant shell flow rate = 12 lpmVary shell flow rate, constant tube flow rate = 12 lpm


Heat transfer coefficients vs. flow rate: PHX08U

Flow Rate (lpm)

0 5 10 15 20 25

Hea

t Tra

nsfe

r Coe

ffic

ient

(Btu

/hr/f

t2 /F)

50

75

100

125

150

175

200

Hea

t Tra

nsfe

r Coe

ffic

ient

(W/m

2 /K)

400

600

800



Heat transfer coefficients vs. flow rate: PHX08S

Flow Rate (lpm)

0 5 10 15 20 25

Hea

t Tra

nsfe

r Coe

ffic

ient

(Btu

/hr/f

t2 /F)

50

75

100

125

150

175

200

Hea

t Tra

nsfe

r Coe

ffic

ient

(W/m

2 /K)

400

600

800



Summary of heat transfer coefficient results

• The S configuration is slightly more efficient than the U configuration.

• The long heat exchangers were more efficient than the short heat exchanger.

• Approach temperatures for these tests were ~10°F, which is a reasonable operating condition.

• Typical heat transfer coefficients in shell and tube heat exchangers ~30-300 Btu/hr/ft2/°F

Average Heat Transfer Coefficients (Btu/hr/ft2/°F) Configuration PHX08S PHX08U PHX03U

Average of 4 heating and cooling tests 134 124 110

Std. dev. 2.6 2.8 12.1


Cumulative PSDs of the large particle tailControl Test (38.5 lpm)

Particle Diameter (µm)1 10

Cum

ulat

ive

Con

cent

ratio

n (#

/ml)

103

104

105

106

01130105319101214522256367447755694

pHasor X Test (38.5 lpm)


Cum

ulat

ive

Con

cent

ratio

n (#

/ml)

103

104

105

106

01029973189891410236534225704

pHasor X Test (10 lpm)


Cum

ulat

ive

Con

cent

ratio

n (#

/ml)

103

104

105

106

0103299509114515882526355652315709

Turnovers

Turnovers

Turnovers


Concentrations relative to initial concentration

> 0.56 µm> 0.7 µm> 1.0 µm> 2.0 µm> 3.0 µm


Turnovers10 100 1000 10000

Con

cent

ratio

n R

elat

ive

to In

itial

Con

cent

ratio

n

0.1

1.0

10.0

Control Test (38.5 lpm)

Turnovers10 100 1000 10000

Con

cent

ratio

n R

elat

ive

to In

itial

Con

cent

ratio

n

0.1

1.0

10.0

> 0.56 µm> 0.7 µm> 1.0 µm> 2.0 µm> 3.0 µm


Turnovers10 100 1000 10000

Con

cent

ratio

n R

elat

ive

to In

itial

Con

cent

ratio

n

0.1

1.0

10.0


Change in concentrations for selected sizespHasor X Test (10 lpm)

Turnovers

0 1000 2000 3000 4000 5000 6000

Cha

nge

in C

once

ntra

tion

(#/m

l)

-10000

-5000

0

5000

10000

15000

20000

Control Test (38.5 lpm)

Turnovers

0 1000 2000 3000 4000 5000 6000

Cha

nge

in C

once

ntra

tion

(#/m

l)

-10000

-5000

0

5000

10000

15000

20000

> 1.0 µm> 2.0 µm> 3.0 µm


Turnovers

0 1000 2000 3000 4000 5000 6000

Cha

nge

in C

once

ntra

tion

(#/m

l)

-10000

-5000

0

5000

10000

15000

20000


Discussion of large particle test results

• A slight decrease in the large particle tail was observed duringthe control test.

• An increase in the concentrations of particles ≥ 1 µm was observed during the heat exchanger test at the higher flow rate.

• Agglomeration at smaller sizes probably occurred as well, but since the concentration of these particles was higher, no increase was apparent.

• Particle concentrations tended to increase roughly linearly withincreasing turnovers.


Discussion of large particle test results

• For particles ≥ 2 µm, the concentration increased by roughly a factor of 5 by the end of the test. This equates to a 0.07% increase in particle concentration ≥ 2 µm per pass through the heat exchanger. In a typical slurry delivery system application, slurry is used with ~100 turnovers, thus only a fairly small increase, ~7%, in concentration would occur.

• This level of increase is dramatically lower than generated by some pumps. (Previous studies have shown up to a 500% increase particle concentrations within 100 turnovers.)

Diaphragm Pump

Turnovers

1 10 100 1000 10000

Con

cent

ratio

n R

elat

ive

to In

itial

Con

cent

ratio

n

0.1

1.0

10.0

100.0

> 0.56 µm> 0.7 µm> 1.0 µm> 1.5 µm> 2.0 µm> 5.0 µm

2007 CMP Users Conference


Working particle size measurementsControl Test (38.5 lpm)

Turnovers

1 10 100 1000 10000

Vol

ume-

Wei

ghte

d D

iam

eter

(nm

)

0

50

100

150

200

250

300

350

Mean99th Percentile Size


Turnovers1 10 100 1000 10000

Vol

ume-

Wei

ghte

d D

iam

eter

(nm

)

0

50

100

150

200

250

300

350

Mean99th Percentile Size


Turnovers

1 10 100 1000 10000

Vol

ume-

Wei

ghte

d D

iam

eter

(nm

)

0

50

100

150

200

250

300

350

Mean99th Percentile SizeError bars represent ± 3σ

Error bars represent ± 3σ

Error bars represent ± 3σ


Discussion of working particle results

• No significant change in the mean or 99th percentile size was observed due to the presence of the heat exchanger.

• A small increase (25-35 nm) in the 99th percentile size was observed during the test at the high flow rate. However, this increase was also observed during the control test, thus it may be attributed to the test system rather than the heat exchanger.


Summary of slurry test results

• No indication of clogging or settling of slurry in the heat exchanger during these tests.

• The heat exchanger was capable of efficiently removing heat so that a constant temperature could be maintained.

• No significant change in the mean or 99th percentile size was observed due to the presence of the heat exchanger.

• Minimal change in the large particle tail was observed during the control test and heat exchanger test at 10 lpm.

• During the heat exchanger test at higher flow, an increase in particle concentrations were observed for particles ≥ 1 µm in size.

• Since slurry is typically used within 100 turnovers in a typicaldelivery system, only a small (~7%) increase in particle concentrations ≥ 2 µm would occur.

• This level of increase, although significant, is dramatically lower than that generated by some pump systems.

performance of an entegris phasor x heat exchanger in

Documents