Application of Controlled Thermal Expansion in Diffusion Bonding for the

Application of ControlledThermal Expansion inDiffusion Bonding for theHigh-Volume Microlaminationof MECS Devices

Christoph PluessOregon State University

Corvallis

September 10, 2004

Master Defense

Table of Contents

2004 Christoph Pluess

Literature & Patent Review

Questions & Discussion

Theoretical Concept & Device Design

Introduction

Results & Conclusions

Experimental Approach

Finite Element Analysis (FEA)

1of 36

Introduction


Bulk -Fluidic Devices (MECS)

2of 36

Introduction


Microlamination (Paul et al., 1999)

Patterning• Laser Micromachining• Chemical Etching

Bonding• Diffusion Bonding• Diffusion Brazing

Registration• Pin Alignment• TEER (Thermal Enhanced Edge Registration)

p

p

T

t

OSU Device

3of 36

Introduction


Why is this topic relevant?

Example: Solid-State Diffusion Bonding within a Vacuum Hot Press

Vacuum Hot Press, Nano/Micro Fabrication Facility

• Pump-down: 0.75-1h• Ramp-up: 0.75-1h• Bonding: 0.5-1h• Cool-down: 2-3hCycle Time: 4-6h

Production Capability:

4of 36

Introduction


Why is this topic relevant?

Example: Solid-State Diffusion Bonding within a Vacuum Hot Press

• Large Substrate MECS Devices• Large Hot Press System ($)• Pressure Uniformity

Device Size:

5of 36

Vacuum Hot Press, Nano/Micro Fabrication Facility

Introduction


Using thermal expansion? Possible solution?

CTE Fixture• CTE, free source of pressure• low expanding frame• high expanding inner parts

Continouous Furnace System• Similar to microelectronics industry• high-volume microlamination

Conveyor Furnace “Sequoia”,MRL Industries

6of 36

Introduction


CTE Bonding Fixture

• Is this a plausible approach for microlamination?

• Can a particular fixture design provide control over:

• Pressure magnitude?

• Pressure timing?

• Pressure sensitivity?

7of 36



Is this idea unique?

•Use of thermal expansion for pressure application first stated by PNNL in 1999

•Most papers/publications try to minimize effects of thermal expansion

•Patents where thermal expansion is used to apply pressure/force:

• Clamping ring for wafers US 5,460,703 (1995)• Belt press using CTE US 6,228,200 (2001)

8of 36



Is this idea unique?

Patent Application Publication US 2003/0221777 A1

McHerron et al. International Business Machines Corp. (IBM)

Method and Apparatus for Application of Pressure to a Workpiece by Thermal Expansion

2003, McHerron, IBM

9of 36



IBM Patent Application

• Lamination of multilayer thin film structures (MLTF)

• Use for continuous production and high throughput

• Reduction of capital expenditures for lamination of MLTF

• Is this a plausible approach for microlamination?

• Can a particular fixture design provide control over:




10of 36

Theoretical Study and Model Development

Theoretical Model


)()()()( ,,,,00 Rififieie TTzzgTzgTg Gap closure function:

)(

)(

Bff

Btotal Tzh

Tg

Resulting strain in z-

direction:Resulting pressure in z-direction :

lamtotallam E

11of 36

Sensitivity Analysis

Theoretical Model


Applying reasonable material properties and sizes:

• initial gap variations of 1m pressure change 1.6 MPa

assumed accuracy with feeler gage 5m 8.0 MPa

Geometrical Sensitivity: 1.6 MPa/m

• temperature fluctuations of 5°C pressure change 4.7 MPa

Thermal Sensitivity: 0.94 MPa/°C

• Lowering thermal sensitivity increases geometrical sensitivity

• Lowering geometrical sensitivity increases thermal sensitivity

12of 36

Fixture Concept


Low expanding fixture frame

High expanding engagement block

Pressure Timing:Initial gap adjustment

Pressure Sensitivity:Spring constant

Pressure Magnitude:Preloading and force storage with spring elements

13of 36

Fixture Design


Fixture Frame:

Inner Parts:

•Designed to fit in hot press ( 3”)

•Max. service temperature 800°C

•Bonding area 25x25mm (2 stations)

Initial gap

setting

Load Cell

Bonding

Platens

Engagement Block

Cu Laminae

InconelDisc Springs

14of 36

Fixture Model (FEM)


Purpose of FE-Model

•Validation of developed fixture design

•Proof of feasibility

•Theoretical assessment of pressure magnitude, timing and sensitivity

15of 36

Fixture Model (FEM)


Expansion Behavior (in z-direction)

CmCmCmT

Tz

T

Tzz fegap

/391.0/665.0/056.1

)()( • gap closure function:

16of 36

Fixture Model (FEM)


p-Timing, p-Magnitude, p-Sensitivity

CMPaC

MPa

T

pTFEM

/0065.020

13.0)(

•150 times less sensitive than simple fixture model (0.94MPa/°C)

17of 36

Fixture Simulation


18of 36

ExperimentalCTE-FixturePrototype


19of 36

Experimental


Experimental Overview

Validation Experiments

Test Article Orientation

Means and 95.0 Percent Confidence Intervals

Orientation

De

flect

ion

0 902.7

3.2

3.7

4.2

4.7

5.2

20of 36

Experimental





Load Cell Validation

• Theoretical: 10’960 N/mm

• Practical: 11’215 N/mm (+2.3%)

21of 36

Experimental






p-Uniformity Bonding Platens

• Fuji Pressure Sensitive Film

22of 36

before:

3.0

3.5

4.0

4.5

5.0

5.5

6.0 MPa

after:

3.0

3.5

4.0

4.5

5.0

5.5

6.0 MPa

Experimental







p-Timing during Lamination

• Experimental

• Theoretical (FEM)

0°C300°C

400°C

600°C

800°C

T

23of 36

Experimental







p-Timing of CTE-Fixture


•T-limit of Fuji film 180°C

•No contact situation with g0 > 63m based on theoretical model

•Experimental validation of CTE- Fixture at low temperature!

•Uniform pressure distribution

• At low temperature (180°C):

g0=0m

g0=30m

g0=50m

g0=70m

24of 36

Experimental









•T-profile optimization by connecting TC directly to temperature control unit:

TC Measurements

•Furnace cool down optimization with helium cooling:

25of 36

Experimental



Validation Experiments Final Experiment

Hot Press

CTE-Fixture

vs.vs.

Fin warpage (timing)

Void fraction (p-magnitude)






TC Measurements

26of 36

Experimental


DOE Final Experiment

• 24 full-factorial design with 1 replicate (32 runs)• Mode: Hot Press / Fixture• Pressure: 3 MPa / 6 MPa• Temperature: 500°C / 800°C• Time: 30’ / 60’

• 2 samples each run for a total of 64 test samples

• ANOVA on fin warpage (128 measurements)

• ANOVA on void fraction (160 measurements)

32 32

27of 36

Results


ANOVA Fin Warpage


Mode

War

page

Fixture Hot Press3.5

3.7

3.9

4.1

4.3

4.5

Interactions and 95.0 Percent Confidence Intervals

Mode

War

page

Pressure36

2.6

3.1

3.6

4.1

4.6

5.1

5.6

Fixture Hot Press

•No statistical significant difference observed (p-value 0.92)

•Average fin warpage hot press: 3.99 m 0.44 m

•Average fin warpage CTE-fixture: 4.02 m 0.44 m

28of 36

Results


Void Fraction Inspection

•Metallographic preparation of the 16 samples

•10 inspection locations for each DB set (160 measurements)

•Voids were marked on atransparency at 384X

•Calculation of void fractions(reference length 250 m)

%1000

l

lv

voidf

29of 36

Results


Metallographic Pictures Bond Lines (Void Shrinkage)

500°C3 MPa60 min

800°C3 MPa60 min

800°C6 MPa60 min

Hot Press CTE-Fixture

8.0%

9.6%

15.8% 23.2%

78.8%

70.5%

30of 36


Mode

Voi

d F

ract

ion

Fixture Hot Press40

42

44

46

48

50

52

Results


ANOVA & Table of Means for Void Fraction

BondingConditions

Void Fraction

Hot PressS.D.

Void FractionFixture

S.D.

500°C/3MPa/30min 100.0% 0% 100.0% 0%

500°C/3MPa/60min 78.8% 8.4% 70.5% 8.7%

500°C/6MPa/30min 65.7% 10.3% 80.0% 6.2%

500°C/6MPa/60min 35.9% 12.0% 62.0% 7.3%

800°C/3MPa/30min 20.1% 6.9% 30.3% 10.3%

800°C/3MPa/60min 15.8% 9.5% 23.2% 4.2%

800°C/6MPa/30min 15.5% 3.7% 19.5% 8.5%

800°C/6MPa/60min 8.0% 2.4% 9.6% 2.8%

Overall 42.5% 6.7% 49.4% 6.0%

•Statistical significant difference observed

•Since T & t are equal, variation is due to pressure applied

31of 36

Results


ANOVA for Void Fraction (Interactions)

•Less significant difference at low pressures (frame stiffness maintained)

•Less significant differences at high temperatures (void fraction less pressure sensitive at high temp.)


Mode

Vo

id F

ract

ion

Pressure36

28

38

48

58

68

Fixture Hot Press


Mode

Vo

id F

ract

ion

Temperature500800

0

20

40

60

80

100

Fixture Hot Press

32of 36

Results


Comments

• Timing of pressure within the CTE-fixture did not show

any problems

•Although void fractions showed a difference, bond quality is comparable

•Source of p-variation due to the use of high expanding stainless steel bolts (potential loss of preload)

•Level of pressure was dampened due to spring implementation in frame (loss of rigidity)

33of 36

•Is this a plausible approach for microlamination?

•Can a particular fixture design provide control over:




Conclusions


( yes)

yes

yes yes

34of 36

Future Research


• CTE-Fixture design for large substrates

• Validation of large substrate fixture design with FEM:

• Structural, p-uniformity• Thermal, T-gradients

• Process optimization for continuous production line

• Experimental investigation of large substrate bonding

35of 36

Thanks for your attention!M.S. Defense Presentation

Christoph Pluess

September 10th 2004

Oregon State UniversityCorvallis

USA

Special thanks to:Major Professor Dr. Brian K. Paul

Committee Members:Dr. Sundar V. AtreDr. Kevin M. Drost

Dr. Timothy C. KennedyDr. Zhaohui Wu

Special thanks to:Steven Etringer

Questions & Discussion


36of 36

Additional Slides


Experimental


First CTE-Prototype

• 7 MPa at 800°C

• 5 Cu-layers bonded: 02/20/2004

• 1st successful CTE-bond

Fixture Model (FEM)


FE-Model Features (ANSYS)

•¼ model 6655 elements (17’000

nodes)

•LINK10 prevented use of contact elements

•One solid structure (SOLID95)

•Spring constant, initial gap, preload of load cell defined over real constants

•Input of bonding parameters over Scalar Parameter Menu

•Automatic load step definition and execution with APDL-macro

FE-Model Settings


Real Constant Settings

Important CommentsRoom Temperature 20 °C Active Warpage Temperature Range 0 °C Check with fin buckling limit equation!Bonding Temperature 200 °C Load Cell Force at Bonding Temperature 2500 NTemperature of Contact 200 °C Load Cell Force at Contact Temperature 2500 N

Bonding Area (Test Article) 625 mm2 Load Cell Force at Room Temperature 3034 NDesired Bonding Pressure 4.0 MPa Preload Force (Hot Press) 682 lbs Min. adjustable load of hot press is 400 lbs

Gap Closure Function 0.391 m/°C Additional Load Cell Compression -49 m (+): add. compression / (-): relaxation

CTE Load Cell Fasteners 16.2 10-6/°C Spring Constant per Stack 2740 N/mm

CTE Load Cell Platens 6.5 10-6/°C Active Force per Spring Stack at Room T. 758 N Flat load of disc springs 1620 NActive Expanding Bolt Length 16.7 mm Active Force per Spring Stack at Contact 625 N Flat load of disc springs 1620 NThickness Load Cell Top mm Active Force per Spring Stack at Bonding T. 625 N Flat load of disc springs 1620 N

Tensile Stress Area Bolt 20.5 mm2 Tensile Stress of Bolts at Room Temp. 37 MPa Max. stress limit of ceramic bolts 55 MPaNominal Load Disc Spring 800 N Tensile Stress of Bolts at Contact Temp. 30 MPa Max. stress limit of ceramic bolts 55 MPaNominal Compression D.S. 0.292 mm Load Cell Compression at Room Temp. 277 m >nominal compression, increase # of springsNumber of Springs per Stack 1 Load Cell Compression at Contact Temp. 228 mNumber of Spring Stacks 4 Load Cell Compression at Bonding Temp. 228 m

Expansion Potential CTE Fixture 70 mActive CTE Compression of Load Cell 0 mStarting Pressure at Contact Temperature 4.0 MPaPressure Sensitivity CTE Fixture in f(T) 0.0069 MPa/°CPressure Sensitivity CTE Fixture in f(z) 0.0175 MPa/m

Adjustable Initial Gap 70 m

Element Length LINK10 (T.o.) 13.668 mm ISTRN Value for LINK10 (Tension only) 0.02025 277 RC Set #3, Element Type 5 (Fasteners)Element Length LINK10 (C.o.) 7.332 mm ISTRN Value for LINK10 (Compr. only) -0.02816 -206 RC Set #4, Element Type 6 (Set Screw)

FEM Model Data Input

Bonding Data Input

CTE Fixture Data Input

CTE Fixture Data Output: Preload Load Cell

CTE Fixture Data Output: Initial Gap

FEM Model Data Output: Real Constant Settings

Application of Controlled Thermal Expansion in Diffusion Bonding for the

Documents

christoph pluessis

christoph pluesswhy

christoph pluess13of

christoph pluess2of

christoph pluess15of

christoph pluess19of

christoph pluess1of

christoph pluess18of