Post on 17-Sep-2020
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
3 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
4 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
5 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
• 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
6 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
7 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
8 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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 )
9 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
10 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
11 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
12 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
13 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
14 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
15 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
16 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
17 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
18 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
19 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
)
20 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
21 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
22 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
23 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
24 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
25 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
26 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
27 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
28 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
29 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
30 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
31 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
32 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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
33 Development of Thermal Interface Materials and Impact of Application Surface – Mathur, Weiser, & Rastogi – Feb 06
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