Flexible Thermal Ground Planes with Thicknesses Below Quarter‐mm
Ryan Lewis, Ph.D.
Shanshan Xu, Y.C. Lee, Ronggui Yang, Li‐Anne Liew
University of Colorado at Boulder
rjlewis@colorado.edu
IMAPS ‐ Thermal 2014
1
Future smart phones are thin
30
Thickness (mm)
25
20
15
10
Future smartphones?
5
0
2004
2006
2008
2010
2012
IMAPS ‐ Thermal 2014
http://pocketnow.com/2012/05/16/remember‐how‐thick‐phones‐used‐to‐be
2014
2016
2
3mm flexible smartphone
Flexible by design!
Phones crack with 90‐150 lbs force
IMAPS ‐ Thermal 2014
http://www.consumerreports.org/cro/news/2014/09/consumer‐
reports‐tests‐iphone‐6‐bendgate/index.htm
3
Flexible 1‐mm smartphone as future microsystem
OLED; 0.05 mm
Glass or Polymer Cover; 0.05 mm
Chips (0.075mm)
PCB TGP- Thermal Ground Plane (0.20 mm)
Anode/Solid State Electrolyte (0.150mm)
> 4X increased storage density
New Cathode/Solid State
Electrolyte (0.4 mm)
Aluminum (0.05 mm)
Polymer Case; 0.025 mm
Three enabling technologies being developed at CU‐Boulder
• Flexible thermal ground planes
• All solid state battery
• Atomic layer deposition
IMAPS ‐ Thermal 2014
4
Air Gap? Alternative Needed!
chip‐side
Back‐side (i.e. skin surface)
2 W/(4mm x 4mm)
Air Gap
Air Gap
Polymer
Copper
Polymer
Copper
200 μm Cu
200 μm air
Polymer
62 oC
Polymer
66 oC
45 oC
Natural convection
on top and bottom sides
with Tair = 25oC Total Thickness = 1 mm + Tair = 25oC
Temperature distribution
on back surface, i.e. skin.
43 oC
Air Gap
For thin electronics:
•Copper and Graphite + Air‐gap not good enough!
IMAPS ‐ Thermal 2014
5
Vacuum-Enhanced Heat Spreader:
Skin Temperatures Reduced
Copper or Graphite vs. Vacuum-Enhanced Heat Spreader
10cm x 5cm x 250 um
Copper Heat Spreader:
Lowest temperature: 34.8°C
Highest temperature: 40.2°C
Temperature difference: 5.4°C
Graphite Heat Spreader:
Lowest temperature: 35.4 °C
Highest temperature: 38.2 °C
Temperature difference: 2.8 °C
Note: this 2.8oC could be increased substantially for 100um thin graphite.
2.5 Watt
Vacuum-Enhanced Copper Heat Spreader
Lowest temperature:35.5 °C
Highest temperature:36.7 °C
Temperature difference: 1.2°C
Vacuum-Enhanced Graphite Heat Spreader
Lowest temperature: 35.6 °C
Highest temperature: 36.4 °C
Temperature difference: 0.8 °C
Vacuum-Enhanced Heat Spreader:
Junction Temperatures?
Max. Junction Temperature ~ 50 oC
Vacuum-Enhanced Heat Spreader – Design A
Max. Junction Temperature ~ 90 oC
Vacuum-Enhanced Heat Spreader – Design B
•Vacuum‐enhanced heat spreading: a short term solution that is always
better than copper/graphite + air gap.
•Trade‐off analysis for an optimum design required.
•Best solution: flexible thermal ground planes.
IMAPS ‐ Thermal 2014
7
Thermal Ground Plane
10 cm x 6 cm total area • Printed Circuit Board substrate
• Built‐in “vacuum” insulator
• Latent heat of phase‐change for heat removal
IMAPS ‐ Thermal 2014
8
Video
IMAPS ‐ Thermal 2014
9
TGP and Copper IMAPS ‐ Thermal 2014
10
TGP working physics
(2-D Heat Pipe or Vapor Chamber)
Evaporator Region
Condenser Region
Vapor Core
Heat Source
Sintered powder
Y. Sungtaek Ju et al. I. J. Heat Mass Transfer, 60, 2013, 163‐169
C. Ding et al. JMEMS 19, 4, 2010
Liquid Returned thru
a Wicking Structure
Micro‐pillars
Woven mesh
C. Ding et al. JMEMS 19, 4, 2010
IMAPS ‐ Thermal 2014
Heat Sink
Grooved Channels
M. Sigurdson, et al. I.J. Heat Mass Transfer, 62, 2013, 11
178‐183
TGP and Graphite
Flexible TGP
Graphite
In‐plane thermal conductivity
1,000 – 4,000 W/m‐K
400 ‐ 1,500 W/m‐K
Through‐plane thermal conductivity
<0.03‐100 W/m‐K
3 – 15 W/m‐K
Integration with PCB
Can be part of PCB interconnects and vias
Separate piece.
3cm x 3cm
PCB‐TGP
Thermal vias
9.5 cm X 5 cm X 0.1cm
IMAPS – Thermal ATW 2014
12
TGP Background at Colorado
1‐mm thick, 20 cm x 5 cm footprint
Copper (measured)
Keff > 2,000 W/mK
w/o thermal resistances across Kapton; Keff > 3,000 W/mK
IMAPS ‐ Thermal 2014
•Flexible, manufacturable, but too thick, too high power
13
Quarter‐mm TGP Results
thermal resistance(K/W)
• Carrying Heat (heat‐pipe)
20
Rcu ref
15
10
5
RTGP
R=ΔT/Qout
0
0
2
4
6
8
10
1600
1400
keff=(Rcu/RTGP)kCu
1200
1000
800
600
200
0
0
2
4
6
8
5cm
8 mm
400
8 mm
Effective Thermal Conductivity (W/m‐K)
input power(W)
10
2.5cm
input power(W)
IMAPS ‐ Thermal 2014
14
Vapor Core Design: Critical to Thin TGP
• Thermal resistance
Re, top
Re, vapor
Re, mesh
Re, bottom
Heat Source
Ra, top
Rc, top
Rc, vapor
Ra, vapor
Rc, mesh
Rc, bottom
Ra, mesh
Ra, bottom
Heat Sink
Thick TGP
vap
Rc,bottom
Rc,mesh
IMAPS ‐ Thermal 2014
Thin TGP
Rc,bottom Re,bottom
Rc,mesh
Re,mesh
Re,bottom
Re,mesh
Rvap
Rvap
15
TGP Limited by 50um Vapor Core?
1.E+06
• Vapor Core Effective Thermal Conductivity:
1.E+05
Fourier’s Law k=Q/(dT/dx)
‐Temperature Gradient proportional to pressure gradient dT/dx~dP/dx
‐Pressure gradient proportional to flow‐rate, which is proportional to heat applied dP/dx~m~Q
kvap (W/m‐K)
1.E+04
1,500 W/m‐K
50 μm limit
1.E+03
Copper Ref
1.E+02
20 μm limit to
“beat” Cu
1.E+01
‐NOT A DIRECT FUNCTION OF POWER
1.E+00
1
10
100
1000
Vapor Core Thickness
keff
IMAPS ‐ Thermal 2014
16
TGP Limited by 50um Vapor Core? No!
• As a Heat Spreader
– Distributed condenser
Natural Convection
• Measurement of uniformity of skin temperature
• TGP= 5x lower ΔT than Cu at 6W
Skin temperature difference – Copper VS TGP
T1
T2
T3
IMAPS ‐ Thermal 2014
Temperature difference (°C)
Heat in
30
25
20
Cu
15
10
TGP
5
0
0
2
4
6
Input power (W)
8 17
10
TGP’s Limit?
Overall Thickness
100 μm?
Heat Sink
Heat Source
100 μm TGP limits model
14
Heat Load (W)
12
10
8
6
4
Working area
2
0
0
50
100
150
Condensertemperature (oC)
IMAPS ‐ Thermal 2014
18
Summary
• Copper + Air‐gap not enough!
• Short‐term improvement via vacuum‐enabled heat spreaders: copper or graphite + vacuum
• Thermal ground planes!
–
–
–
–
PCB substrate
Built‐in insulator layers
Phase‐change heat transfer
<0.1mm TGP with Keff > 3,000 W/m‐K feasible.
• Quarter‐mm, demonstrated keff=1,400 W/m‐K.
Prototypes are to be distributed.
IMAPS ‐ Thermal 2014
19
Acknowledgements
• TGP research at CU has been supported by:
– University of Colorado
– State of Colorado AIA: award number: 2015‐3212
– Defense Advanced Research Projects Agency (DARPA)
Thermal Ground Plane Program N6601‐08‐C‐
2006
– Intelligence Community
This project was supported by a grant from the Intelligence Community Postdoctoral Research Fellowship Program through funding from the Office of the Director of National Intelligence. All statements of fact, opinion, or analysis expressed are those of the author and do not reflect the official positions or views of the Intelligence Community or any other U.S. Government agency. Nothing in the contents should be construed as asserting or implying U.S. Government authentication of information or Intelligence Community endorsement of the author’s views. 20