development of thermal interface materials and impact of ... honeywell.pdfbased upon astm standard...

Development of ThermalInterface Materials and Impact

of Application Surface

Devesh Mathur, Martin Weiser, Ravi RastogiHoneywell Electronic Materials

February 16, 2006

MEPTEC: The Heat is OnSan Jose, CA

2 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06

Outline

• Who is HEM

• Motivation & definitions

• Heat spreader surface finish

• TIM2 results & characterization

• TIM1 results & characterization

• Summary


Honeywell Electronic Materials (HEM)

Headquarters: Chandler, AZ

Locations: 10 in the U.S.,Canada, China, Japan, Korea,Philippines, Singapore, Taiwan,Thailand, France, Germany

Employees: 1,100

• Dielectrics• Metals• Electronic Chemicals• Interconnect PackagingSolutions

Competitive Advantages

• 40+ Years of Industry Experience and Relationships

• Core Competencies:• Chemistry, Organic and Inorganic

• Metallurgy

• Packaging Material Science

• Process Integration Capability and OEM Alliances

• Acids, Solvents, Etchants and Cleaning Chemistries

• Direct Global Distribution

14% Specialty Materials31% Automation & Control 17%Transportation & Power Systems38% Aerospace


HEM Segments and Products

Products

• Spin-on glass (SOG)• Bottom anti-reflective coating (BARC)• Spin-on low-k• Low-k enablers

• Straights etchants• High-purity solvents• Selective etchants

• Thermal spreaders• Thermal Interface Materials• Evaporative materials, preforms, & wire• Sapphire Wafers

• Al, Ti, Cu, Ta targets• Thermocouples• Coils

Segment

Dielectrics

ElectronicChemicals

PackagingMaterials

Metals


• Keep the die below 90oC- Heat output of >150 W- Hot spot >300 W/cm2

- Heat removed via conductionTIM1 Spreader TIM2 HeatSink

• TIM1- Void free to prevent new hot

spots- Stable to handle high

temperature & thermal cycles

• Heat Spreader- Flat to achieve low BLT- Surface finish for markability

• TIM2- Accommodate non-uniform bond

line to the heat sink

Heat Transfer Modes- Conduction - Convection - Radiation

1.5 GHz ItaniumThermal Map

130W / 3.74 cm2

35 W/cm2

Designed Hot Spot

The Thermal Challenge

Hea t


Package – Part & Thermal Definitions

Si Die

Tjunction

Tcase

Tsink

Tambient

Heat Sink

TIM2

Spreader

TIM1

Die

_jc = (Tjunction – Tcase) / Die power

_cs = (Tcase – Tsink) / Die power

_ca = (Tcase – Tambient) / Die power


TIM Fundamentals

• Dry, smooth surfaces- Only 1 - 2% solid contact area- 98 - 99% low conductivity gas

TIM effectively improves heat transferperformance

TIM effectively improves heat transferTIM effectively improves heat transferperformanceperformance

TIM

• With thermal interface material- Near zero solid contact area- 70 - 95% TIM contact area

Much higher conductivity than gas


Thermal Contact ResistanceThermal Contact Resistance

Ref: Vishal Singhal et al., IEEE Trans. Components and Packaging Technologies, 27 (2) (2004) 244.

( )

+

−•

=↔TIMTIM

TIMTIMc kkP

ER

11

2

1

12

tan

tan55.1

1

1

94.0

2)1(ν

θθ

σ

σ

E1 , k1

E1 , k1

tan θ

TIM

y

x

Assume same material contact (E1 , k1) with TIM and ETIM<< E1

Contact resistance is critical for heat transferat the interface of two different materials

Contact resistance is critical for heat transferContact resistance is critical for heat transfer

at the interface of two different materialsat the interface of two different materials

_ : rms surface roughnesstan _ : average slope of surface asperityE1, ETIM : elastic modulus of contact pair and TIMk1, kTIM : thermal conductivity of contact pair and TIMP : contact pressure_ : Poisson’s ratio for compression, _xx / (_xx - _yy )


Thermal Impedance (Z) Measurement

Heater

Chiller Block

Load Cell

Pn

eum

atic

Cyl

ind

er

Temperature

ΔT

Slope

Q = kCu× Slope /AreaZ = ΔT / Q

Standard test for thermal pastes includesconductivity and contact resistance

Standard test for thermal pastes includesStandard test for thermal pastes includes

conductivity and contact resistanceconductivity and contact resistance

BLT

Based upon ASTM Standard D5470Everyone implements differently


Flatness and Roughness Definitions

• Flatness = distance between highest and lowest point parallel to theaverage line or plane

• Rq = root mean square roughness = standard deviation of heights• Ra = average roughness = distance between average positive and average

negative height

-1.2

-0.8

-0.4

0

0.4

0.8

1.2

0 0.2 0.4 0.6 0.8 1

Relative Position (mm)

Hei

gh

t (m

icro

ns)

Flatness

RqRa

xxy

l

xyxydxxyl

R

xyn

dxxyl

R

n

nn

l

q

n

nn

l

a

point aat height surface :)(

length sampling :

)()(n

1or )(

1

Roughness SquareMean Root

)(1

or )(1

Roughness Surface Average

0

2

0

2

00

??

??

?

?

???

??

Ref: Jukka Rantala, Electronic Cooling, May 2004

Laser profilometery used to measuresurface roughness

Laser profilometery used to measureLaser profilometery used to measuresurface roughnesssurface roughness


Heat Spreader: Selected OptionsOne Piece Two Piece

Different Variations

Growing trend for thermal managementoutside of CPU needs

Growing trend for thermal managementGrowing trend for thermal managementoutside of CPU outside of CPU needsneeds


Coined Heat Spreader

-6

-3

0

3

6

9

0 5 10 15 20 25Position (mm)

Prof

ile (m

icro

met

er)

Critical region for die attach

Ra = 0.36 µm

-2

-1

0

1

10 12 14Position (mm)

Clo

se u

p of

Pro

file

(mic

rom

eter

)

25 µmCavity side

100 pts/mm & 1 mm/s

8 µm flatness


Coined & Deburred Heat Spreader

Cavity side100 pts/mm & 1 mm/s

25 µm-4

-2

0

2

4

6

8

10 12 14Position (mm)

Prof

ile (m

icro

met

ers)

-4

-2

0

2

4

6

8

0 5 10 15 20Position (mm)

Prof

ile (m

icro

met

ers)

8 µm flatness

Ra = 0.53 µm


Effect of Ni Plating on Heat Spreader

Bare Cu Ni Plated

5 µm5 µm

Plating smoothes out the irregularitiesPlating smoothes out the irregularitiesPlating smoothes out the irregularities

Coined &Deburred

Ra = 0.54 µm

CoinedRa = 0.36 µm


TIM2 Materials and Testing

MATERIALS• PCM45F

- Pad format- <0.25 oC-cm2/W at 0.002”

BLT

• PCM45SP- Screen printable format- <0.25 oC-cm2/W at 0.002”

BLT

TESTING• Test effects

- Pressure- Block surface profile- Shim to control minimum

BLT


TI vs. Pressure for PCM45SP

0.05

0.08

0.10

0.13

0.15

0.18

0.20

0.23

0.25

0 20 40 60 80

Pressure (psi)

Th

erm

Im

pe

d (

C-c

m2

/W)

Cupped blocks

Domed blocks

Tested without thestandard 0.002” shim

TI is much lower with the domed blocksTI is much lower with the domed blocksTI is much lower with the domed blocks


0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014

Shim Thickness (cm)

Ther

mal

Impe

danc

e (C

-cm

2 /W)

Cupped

Domed

Profile Impact on Shimmed TI Measurement

40 psi test pressureTwo Chromel wire shims

Block profile has less impact when min BLT iscontrolled

Block profile has less impact when min BLT isBlock profile has less impact when min BLT is

controlledcontrolled

Equivalent to a 8 µm BLT difference

No shim results – effective BLT from least squares fit of the TI vs. shim thickness


Cupped & Domed Cut-Bar Block Profile

Cupped blocks result in a fairly thick sampleExplains the higher TI seen previously

Cupped blocks result in a fairly thick sampleCupped blocks result in a fairly thick sample

Explains the higher TI seen previouslyExplains the higher TI seen previously

-25

-20

-15

-10

-5

0

5

10

15

20

0 5 10 15 20 25Position (mm)

Pro

file

(µm

)

-25

-20

-15

-10

-5

0

5

10

15

20

0 5 10 15 20 25Position (mm)

Pro

file

(µm

)

Cupped – As Machined Domed – Hand Lapped


Profile of Ground and Plated Blocks

Surface grinding dramatically reduces cuppingSome cupping due to dog-boning during plating

Surface grinding dramatically reduces cuppingSurface grinding dramatically reduces cuppingSome cupping due to dog-boning during platingSome cupping due to dog-boning during plating

-25

-20

-15

-10

-5

0

5

10

15

20

0 5 10 15 20 25Position (mm)

Pro

file

(µm

)

Surface Ground Ground and Ni Plated

-25

-20

-15

-10

-5

0

5

10

15

20

0 5 10 15 20 25Position (mm)

Pro

file

(µm

)


0.05

0.08

0.10

0.13

0.15

0.18

0.20

0.23

0.25

0 20 40 60 80

Pressure (psi)

Th

erm

Im

ped

(C

-cm

2/W

)Cupped blocksDomed blocks'Flat' Blocks

Flat Block Profile Impact on TI

Flat blocks result in TI between cuppedand domed blocks

Flat blocks result in TI between cuppedFlat blocks result in TI between cupped

and domed blocksand domed blocks

Tested without thestandard 0.002” shim


Package Thermal Resistance

0.00

0.02

0.04

0.06

0.08

0.10

0.12

Ther

mal

Res

ista

nce

(C/W

)

Old New

Θcs

Θ jc

Flattening the chiller block dramaticallyreduces the thermal resistance

Flattening the chiller block dramaticallyFlattening the chiller block dramatically

reduces the thermal resistancereduces the thermal resistance


TIM2 Application Surface Summary

• Surface profile impacts measured TI- Cupped blocks give higher value than reality- Domed blocks give lower value than reality

• Test pressure impacts measured TI- Higher pressure causes more squeeze out – thinner BLT

• Shims give the most realistic test results- BLT in practice is 0.001 – 0.003” due to spreader and heat

sink flatness- Minimizes the effect of block surface profile

• Surface metal finish impacts results- Customer feedback- Al, Cu, Ni plating change the thermal resistance values


TIM1 Materials and Testing

HEM MATERIALS• Solder

- Preforms and wire format- As low as 0.025 oC-cm2/W thermal

impedance

DEVELOPMENTAL MATERIALS• TM200

- Dispensable filled curable polymer- <0.15 oC-cm2/W thermal impedance

• Grease- Printable filled polymer- <0.1 oC-cm2/W thermal impedance

• PSH- Dispensable polymer-solder-hybrid- <0.075 oC-cm2/W thermal impedance

TESTINGTest effects• Surface roughness• Surface chemistry


What is PSH?

PSH = Polymer Solder Hybrid = smart materials• High conductivity filler• Low melting solder alloy• Low modulus polymer matrix

FillerPolymerMatrix

Before Cure

Solder

Heat Conduction through a continuous network &lower contact resistance

Benefits of solder conductivity without thecompliance issues

Heat Conduction through a continuous network &Heat Conduction through a continuous network &

lower contact resistancelower contact resistance

Benefits of solder conductivity without theBenefits of solder conductivity without thecompliance issuescompliance issues

After Cure

Heat Spreader

Die


Machine Ground TI Block & Spreader

Machine Ground TI BlockNi Plated 200X

Coined SpreaderNi Plated 200X

TI blocks need to match spreader surface finishto obtain the most useful data

TI blocks need to match spreader surface finishTI blocks need to match spreader surface finish

to obtain the most useful datato obtain the most useful data


Metalized Si wafer & fly cut blocks

0

12

3

45

6

7

89

10

5 6 7 8 9 10

Si waferFlyCutyFlyCutx

Fly cut is similar to theetched and plated

wafer

Fly cut is similar to theFly cut is similar to the

etched and platedetched and plated

waferwafer


0 0.05 0.1 0.15 0.2 0.25

SmoothAu plated

RoughNi plated

GroundNi plated

SmoothNi plated

TI (C-cm2/W)

Surface condition impacts performance

Surface treatment has a large impact in TI valueSurface treatment has a large impact in TI valueSurface treatment has a large impact in TI value

Conductive phase was 80% solder & 20% Filler BExample formulation


PSH on Ni & Au Plated Spreaders

PSH interacts quite strongly with Au plated spreaderPSH interacts quite strongly with Au plated spreaderPSH interacts quite strongly with Au plated spreader

Ni Plated Spreader Au Plated Spreader


Impact of Spreader Surface Plating

Au plating of the spreader gives betterperformance

Au plating of the spreader gives betterAu plating of the spreader gives better

performanceperformance

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

Formula A Formula B

Ther

mal

Impe

danc

e (C

-cm

2 /W)

Ni-Ni blocksNi-Au blocks


Effect of Ni Surface Aging

T I (

°C- c

m2

/W

)

F Au

-Au B

locks

E 3 hr

s afte

r lap

D 5 hr

s in va

cuum

afte

r lap

C 5 m

in after lap

B 2 m

in after lap

A Au

-Std Ni b

lock

0.12

0.10

0.08

0.06

0.04

0.02

Exposure to Air Increases the Thermal ImpedanceExposure to Air Increases the Thermal ImpedanceExposure to Air Increases the Thermal Impedance

Ni surface was lapped with 6 µm diamondNi + Au plated block except as noted


Competitive Benchmarking

PSH is significantly better than competitive pastesPSH is significantly better than competitive pastesPSH is significantly better than competitive pastes

Thermal Conductivity (W/m-K)Product Reported§ Measured*HON Devel PSH 6.3Competitor A 11 5.5Competitor B 60 7.8Indium 84

§ Laser flash of bulk sample

*K = 1/slope from TI measurement vs. BLT-

0.050

0.100

0.150

0.200

0.250

0.300

0.350

0 50 100 150 200BLT (microns)

TI (C

-cm

2 /W)

HON PSH

Competitor A

Competitor B

Indium


TIM1 Application Surface Summary

•Surface roughness is important-Fine, deformable particles lock into the crevices

•Surface chemistry is important-TIM1s are often designed to react at the interface

Lower contact resistance due to bondingLess pump out due to bonding

-Surface chemistry interaction specially strong forPSH type TIMs


Conclusions

• The performance of TIMs depends upon- Surface flatness- Surface roughness- Surface chemistry

• Lab testing conditions for designing TIMs mustmimic the application to obtain useful information- TIM2 materials should be tested at application BLT

ωDue to both component flatness and TIM squeeze out in use- TIM1 materials should be tested using representative

surface roughness and chemistryωThey often rely on significant mechanical and chemical interactions to

obtain their high thermal performance

Honeywell has TIM1, TIM2, and spreaders alongwith the testing capabilities to fulfill these

requirements

Honeywell has TIM1, TIM2, and spreaders alongHoneywell has TIM1, TIM2, and spreaders along

with the testing capabilities to fulfill thesewith the testing capabilities to fulfill theserequirementsrequirements

development of thermal interface materials and impact of ... honeywell.pdfbased upon astm standard...

Documents