20
Phase Change Materials: Alternative to Thermal Grease for Networking Applications Susheela Narasimhan Juniper Networks, Inc. & Hyo Xi and Chris Lee Honeywell Electronics Materials

Phase change-materials-networking-applications

Embed Size (px)

Citation preview

Phase Change Materials: Alternative to Thermal Grease for

Networking Applications

Susheela Narasimhan Juniper Networks, Inc.

& Hyo Xi and Chris Lee

Honeywell Electronics Materials

2

Discussion

• Device Thermal Trends

• Thermal Requirements & Test Criteria

• Results & Material Selection

• Reliability and Phase Change Materials

3

Power Increases More Than Seven Times After 2000

Merchant Silicon Power in Networking Applications

4

Power Cycling Trends in Networking Applications

5

Temperature Variations as a Result of Power Cycling

6

#   Customer  Need   Descrip1on   Requirement   Tests  

1   Longevity  Long    warranty  periods;    

long  term  thermal  reliability        

Constant  thermal  impedance  through    cycling  and  at  least    

8-­‐10  years  

Temperature  and  power  cycling  

2   Opera1ng  Condi1ons  

Performance  under  harsh  opera@ng  condi@ons;  external  factors  including  temperature,  

humidity,  shock,  etc.  

Ambient  up  to  40°C  at  customer  site  and  

55-­‐61°C  in  NEBS  tes@ng  

-­‐5°C  to  61°C  with    chip  /  heat  sink  interface  at  

100  -­‐105°C  

3   Thermal  Performance  Thermal  dissipa@on  of    extreme  heat  generated  by  high  power  

and  high  density  devices        

Lowest  thermal  Impedance  

Required  Impedance  of  about  0.01  –  0.02°C  in^2  /W  

4   Bond  Line  Thickness   Bond  line  thickness  

~7-­‐8  mil  thickness  in  order  to  meet  lid  and  heat  sink  flatness  requirements  

7-­‐8  mils  

Thermal Needs, Requirements & Test Criteria

7

Issues with Greases in Power Cycling Applications

• Greases are subject to thermal expansion of the heat sink and ASIC lid during power cycling which can cause pump out and result in dry-out scenarios of the interface between the heat sink and the chip

Name of Sensor  Grease

Application  Phase Change

Application  Chip 1   91.0   92.0  Chip 2   92.0   91.2  Chip 3   98.0   97.5  Chip 4   100.0   99.0  

PCM Can Provide Equal Thermal Performance as Grease w/o Issue

8

Illustration of Pump Out

Grease pump out and creation of voids

Phase change application forming a continuous interface between heat

sink and ASIC lid

9

Key Factors for Thermal Dissipation

• Thermal Impedance - The lower impedance, the faster heat dissipation -  Interface contact resistance + TIM resistance

• Bulk Conductivity - The higher conductivity, the faster heat dissipation

• Bond Line Thickness - The thinner BLT, the faster heat dissipation

• Pre-load Pressure - Different material has different pressure reflection

•  Interface Condition Impact - Better interface surface condition, faster heat dissipation due to

less void trapped

10

Visc

osity

Melt temp

Solid

Liquid/gel state •  Optimal Surface wetting •  Low Contact Resistance •  Low Thermal Impedance

Increasing Temperature ! 45 °C

Theoretical Curve: PCM Viscosity vs. Temperature

11

PCM Polymer

                                                           

                                     

• Higher Molecular Weight - Structural integrity

• Long Chain Polymer Structure

vs.

PCM

• Short Chain • Good Flow-ability, wetting

but… • Potential for Migration, Dry-

Out and Pump-Out Issues

• Long Chain • Stable and Consistent Filler • Minimizes Filler Migration /

Separation Over Accelerated Life Test (HTB, Temp Cycle)

Grease

PCM Polymer Structure Enables Reliable, Long-Term Performance

12

Honeywell PCM

Initial 1000 cycles

Silicone Grease

Thermal Cycling Test Condition: •  -55°Cx10min + 125°Cx10 min, for 500 to 1000 cycles •  Sandwich PCM & grease between aluminum and glass plates set at 200µm gap •  TI Test : ASTM D5470

Thermal Cycle (-55 °C to 125 °C) vs. Grease

HEM PCM vs. Silicone Grease

Grease breaks down Grease TI degrades

Silicone Grease

Honeywell PCM

Honeywell Provides Stable Polymer Structure with No Pump-Out Issue

13

Honeywell PCM

Initial 400 hr

Silicone Grease

High Temperature Baking (HTB) Test Condition: •  HTB = 150°C •  PCM & thermal grease dispensed on smoked glass plates •  TI Test : ASTM D5470

High Temp Bake at 150 °C vs. Grease

Oil bleeding from grease Grease TI degrades

0.09 0.11 0.12

1.78

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

2.00

Initial Baking 400 Hr

Honeywell PCM

Competitor grease

HEM PCM vs. Silicone Grease

Silicone Grease

Honeywell PCM

Honeywell PCM Does Not Suffer From Silicone Oil Bleed

14

0.00

0.20

0.40

0.60

0.80

1.00

1.20

0 200 400 600 800 1000

Ther

mal

Impe

danc

e (°

C-c

m2/

W)

Hours

High Temperature Bake at 150 °C TI vs. Hours of Exposure

PTM3180

Grease A

Grease B

Test Condition: 150°C continuous baking Test Method: Laser Flash, ASTM E1461

PTM5000

Significantly Better Reliability Than Silicone Grease

Thermal Reliability: PCM vs. Silicone Greases

15

Honeywell PCM Technology

• Fundamentally three primary components in PCM

• Each are vital to robust polymer matrix integrity and filler optimization

• Filler - Thermal

• Wax/Polymer - Structural integrity

• Additives - Cross linking, ATO, etc

TIM Performance

and Reliability

Thermal Filler

Wax/ Polymer

Additives

PCM Formulation is Critical to Performance

16

Extended Reliability: PTM6000

ASTM E1461

•  Thermal Stability >3000 hrs @ 150°C •  HAST > 192hrs@ 130°C/85%RH •  Superior Reliability

High Temperature Baking TI vs. Time

Polymer Chemistry Enables Improved Reliability

HAST TI vs. Time

17

Improved Thermal Impedance: PTM7000

•  Lower Thermal Impedance: <0.06 oCcm2/W @ Thinnest BLT • Wider Process Window for Pre-Load Pressure Range •  Test Data demonstrated on Thermal Test Vehicle

Test Method: Cut bar, ASTM D5470

PTM7000 Demonstrates Lower Thermal Impedance

18

TTV Results

TTV (thermal test vehicle) study

Design of TTV

CPU/GPU chip height delta 50um

PTM5000 PTM7000 Grease

19

Summary

•  Increased power densities of network devices challenge the performance and reliability of Thermal Interface Materials

• Accelerated life tests are critical to TIM selection criteria

• Robust polymer chemistry of Phase Change Materials enables low thermal impedance with proven long term thermal stability and reliability

Thank You