Upload
others
View
15
Download
0
Embed Size (px)
Citation preview
Ali Shakouri Baskin School of EngineeringUniversity of California Santa Cruz
Ultrafast Nanoscale Electrothermal Energy Conversion Devices and Measurements
Acknowledgement: ONR, DARPA, AFOSR, CEA, NSF,
DOE/EFRC, NASA/UCSC National Semiconductor,
Intel, Canon, Wyle Lab, SRC-IFC
IEEE Power Electronics Society, Santa Clara Valley Chapter; 16 Feb. 2011
2
A. Shakouri 2/16/2011
US Department of Energy; Energy Information Administration (2010)
World Marketed Energy Use by Fuel Type 1990-2035
13TW 2050: 25-30TW
Quad
3
A. Shakouri 2/16/2011
S. Koonin, Former Chief Scientist BPnrg.caltech.edu
Oil Resources
4
A. Shakouri 2/16/2011
Source: Hansen, Clim. Change, 68,
269, 2005.
400,000 years of greenhouse-gas &
temperature history based on bubbles trapped in Antarctic ice
Last time CO2 >300 ppm was 25 million
years ago.
John P. Holdren, 2006
Thousand years before 1850
5
A. Shakouri 2/16/2011
David MacKay, Sustainable Energy –Without the Hot Air
6
A. Shakouri 2/16/2011
RejectedEnergy 61%
http://eed.llnl.gov/flow
1.3TWefficiency ~31%
20%
RejectedEnergy 58%
1.3TW
efficiency ~32%
25%Petroleum37.13
Coal22.42
Natural Gas 23.84
Hydro 2.45
Nucl. 8.45
Energy Use = 99.2 Quad = 105 EJ → Power ~3.3 TW
Solar 0.09
Biomass 3.88
Wind 0.51
Geo 0.35
Transportation
Industrial
Residential
Commercial
27.86
23.94
8.58
11.48
Quads
Electricity Generation39.97
US Energy Flow 2008
7
A. Shakouri 2/16/2011Direct Conversion of Heat into Electricity
DV~ S DT
Electrical Conductor
Hot Cold
Seebeck(1821)
I
8
A. Shakouri 2/16/2011
Impact of 15oC temperature increase on IC performance
• Leakage power exponential increase with temperature
– 60nm→50-70% of total power– Potential thermal runaway
• Lifetime exponential decrease with temperature (x ¼) – e.g. electromigration, oxide breakdown
• Interconnect delay (10-15%)• Crosstalk noise ( up to 25%)
Thermal integrity: a must for low-power-IC digital design, EDN 15 Sept. 2005
Measuring Power and Temperature from Real Processors, Javi Martinez, Jose Renau, et al. The Next Generation Software (NGS) Workshop (NGS08), April 2008
http://masc.cse.ucsc.edu
9
A. Shakouri 2/16/2011
Electric Current
HeatingCooling
Peltier Effect (1834)
Reverse of Seebeck effect (electric current produces cooling/heating)
= STc´ ILord Kelvin (Thomson)
10
A. Shakouri 2/16/2011
Fraction of
Carnot
Efficiency0.1
0.4
0.3
0.2
ZT1 2 3 4
Commercial TE’s
Typical CFC System
Thermoelectric (Peltier) Coolers
Conventional thermoelectrics have low efficiencies and low cooling power density <50-100W/cm2.
)()()( 2
2
tyconductivithermaltyconductivielectricalSeebeckZ
kSZ
=
=s
11
A. Shakouri 2/16/2011TEs for Telecom Cooling• Melcor, Marlow and many other TE manufacturers
provide coolers specifically designed for Telecom laser-cooling applications
Typical Distributed Feedback Laser:Dl/DT= 0.1 nm/oC
Heat generation kW/cm2
Cronin Vining, ZT Services
12
A. Shakouri 2/16/2011
Thermoelectrically-Cooled/Heated Car Seat
Lon Bell BSST
13
A. Shakouri 2/16/2011
Solid-State Thermoelectric Energy
Conversion using Nanostructured
Materials
14
A. Shakouri 2/16/2011
Ali Shakouri, Director
Thermionic Energy Conversion Center
4 nm (Zr,W)N
2 nm ScN
UCSC (Bian, Kobayashi), Berkeley (Majumdar), BSST Inc. (Bell), Delaware (Zide), MIT (Ram), Purdue (Sands),
UCSB (Bowers, Gossard)
Engineering current and heat flow using nanostructures
Goal: direct conversion of heat into electricity with > 20-30% efficiency
Funding: DARPA, ONR, DOE, NSF
15
A. Shakouri 2/16/2011
Two beneficial effects of metal nanoparticles on thermoelectric performance
Nanoparticle
ErAs
InGaAs
ErAs
e-
e-
e-
e-
e-
e-
e-
ErAs
e-
Metal nanoparticles scatter phononsÆ reduced thermal conductivity
Metal nanoparticles donate electronsÆ enhanced electrical conductivity
by increasing electron density in the semiconductor matrix
(Seebeck)2(electrical conductivity)(thermal conductivity)
Z =
D. Vashaee and A. Shakouri, Phys. Rev. Lett. 92, 106103/1 (2004)C. Vineis et al. Nano. Thermoelectrics: big efficiency gains from small features, Advanced Materials (2010)
16
A. Shakouri 2/16/2011Thermoelectric figure-of-merit
Largest measured ZT~1.33 at 800K
J. Zide et al. Journal of Applied Physics (Dec. 2010)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
300 400 500 600 700 800
ZT
T (K)
ErAs:InGaAlAs
n-InGaAlAs (control)
The majority of ZT enhancement is from thermal conductivity reduction.5% power factor enhancement at 800K.
17
A. Shakouri 2/16/2011Module power generation results
140 mm/140 mm AlN
400 elements (10-20 microns ErAs:InGaAlAs thin films, 120x120mm2), array size 6x6 mm2
0
0.5
1
1.5
2
2.5
3
0 20 40 60 80 100 120 140
10 mm module20 mm module
Out
put P
ower
(W
/cm
2 )
DT (K) Peter Mayer and Rajeev Ram MITTE module requirements
18
A. Shakouri 2/16/2011
Co-optimization of thermo-electric module with heat sinks
K. Yazawa and A. Shakouri, Int. Conf. on Thermoelectrics 2010, IMECE 2010
np
np
p
np
np
n
n
Source Heatexchanger
Cold sideHeat sink
ThermoelectricModule
current
heat
SubstratesElectrodes
Leg thickness
Fractional coverageof element
19
A. Shakouri 2/16/2011
0
200
400
600
800
1000
1200
1400
1600
1800
2000
1.E+02 1.E+04 1.E+06 1.E+08
Material cost of TE power generation system
Cross over
Air cooling (Aluminum heat sink)
Water cooling (microchannels in
Copper)
Heat flux [W/m2]
$/m
2
2nd generation TE modules
3rd generation TE modules
K. Yazawa and A. Shakouri, International Mechanical Engineering Congress Nov. 2010
Current TE modules
Today’s TE power generation cost:
$1-2/W
Potential to bring this down to:
$0.1/W
20
A. Shakouri 2/16/2011Microrefrigerators on a chip
Heterostructure Integrated Thermionic Coolers; A. Shakouri and John Bowers, Appl. Phys. Lett. 1997
Nanoscale heat transport and microrefrigerators on a chip; A. Shakouri, Proceedings of IEEE, July 2006
1 µm
Hot ElectronCold Electron
Si (10nm)
Si0.89Ge0.1C0.01
21
A. Shakouri 2/16/2011Microrefrigerators on a chip
Nanoscale heat transport and microrefrigerators on a chip; A. Shakouri, Proceedings of IEEE, July 2006
Featured in Nature Science Update, Physics Today, AIP April 2001
• Monolithic integration on silicon• DTmax~4C at room temp. (7C at 100C)
22
A. Shakouri 2/16/2011
High resolution thermal imaging of microcoolers
• Temperature resolution: 0.006oC• Spatial resolution: submicron• 256 channel lock-in camera <20kHz, 123dB (1sec)
J. Christofferson and A. Shakouri, Review of Scientific Instruments, 2005
23
A. Shakouri 2/16/2011Transient thermal imaging
• Maximum cooling can be increased under pulsed operation • Reason: Interface Peltier cooling and distributed Joule
Bjorn Vermeersch, Je-Hyeong Bahk et al.
24
A. Shakouri 2/16/2011
Thin Film Microrefrigerator Optimization
Current SiGe material
Decrease thermal conductivity
Increase Seebeck coef.
Increase Seebeck coef.
Decrease thermal conductivity
Younes Ezzahri, Ali Shakouri et al. InterPACK07, Vancouver, Canada
• 10 microns thick, 50x50mm2 monolithic microrefrigerator with ZT~0.5 can cool a 1000W/cm2 hot spot by >15C.
25
A. Shakouri 2/16/2011
Ultrafast Thermal Characterization and
Simulations Techniques
26
A. Shakouri 2/16/2011High Power MOSFET Transistor Array
“Thermal Characterization of High Power Transistor Arrays”, K. Maize et al, 25th IEEE SEMI-THERM Symposium, 2009.
27
A. Shakouri 2/16/2011
thermal image - m1-10 Vd = 500mV, Id = 545, J = 6A/mm2
50 100 150 200 250
50
100
150
200
2500
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.2
0.3
0.4
0.5
0.6
0.7 g
cha
nge
in te
mpe
ratu
re D
eg C
b
0.250.30.350.40.450.50.550.60.65
a
a’
-1500 -1000 -500 0 500 1000 15000.2
0.3
0.4
0.5
0.6
0.7
Distance (um)
Tem
pera
ture
Ris
e fro
m A
mbi
ent (
K)
M1-10, Vds=500mV, Temperature Profile
The magnitudes of temperature rise at both ends (source and drain ) are in a good agreement. Potential role of the probe tips.
6 amps/mm2Vd = 0.5V, Id ≈ 0.5A
Thermal image vs. simulation at low bias
C°
28
A. Shakouri 2/16/2011Array heating at low & high current densities
Δ Te
mpe
ratu
re [°
C]
47 A/mm2
VD=5V, ID=4.2A
b b’
Δ T profile from b to b'
Distance 30
70
50
60
40
ΔT [°C]
HIGH BIAS
6020 400
Δ Te
mpe
ratu
re [°
C]
b b’
400μm6 A/mm2
VD=500mV, ID=545mA
a a’Distance
LOW BIAS
400μm
a a’
0.4 0.60.20ΔT [°C]
Δ
Tem
pera
ture
[°C
]
0.4
0.6
0.2
0.5
0.7
0.3
Δ T profile from a to a'
Thermo-reflectance
thermal images
Tempera-ture
profile along array
“Thermal Characterization of High Power Transistor Arrays”, K. Maize et al, 25th IEEE SEMI-THERM Symposium, 2009.
29
A. Shakouri 2/16/2011
Through-substrate thermoreflectance –Comparing topside and backside heating
Thermal images
Topside
Backside (InGaAs CCD)
drain contact source contact
100µm 6X
Nea
r-IR
Li
ght
K. Maize et al, SemiTHERM 2010 and IHTC 2010.
30
A. Shakouri 2/16/2011Heating in Electro Static Discharge Devices
300ns
40µm Anode
Cathode
10
ΔT [K]
0
5
10
15
20
0
20
Snapback current =1.22A.
30µs 170µs
ΔT [K]600
400
200
0
K. Maize, V. Vashchenko et al, To be presented at IRPS, 2011.
31
A. Shakouri 2/16/2011
0
5
10
15
20
0 50 100 150
Heating on Metal Above Via
550nm 600mA350nm 400mA
Temp
eratu
re C
hang
e fro
m Am
bient
Delay in Nanoseconds
High Speed Thermal Imaging (800ps)
350nm Via, 400mA
550nm Via, 600mA
J. Christofferson et al., Proceedings of Int. Heat Transfer Conf., August 2010
Study of heating in submicron interconnect vias
32
A. Shakouri 2/16/2011
Thermoreflectance Imaging for Reliability Characterization of
Copper ViasShila Alavi, Glenn B Alers, Kaz Yazawa
and Ali Shakouri Department of Electrical EngineeringUniversity of California, Santa Cruz
Advanced Studies LaboratoryUniversity of California, Santa Cruz
33
A. Shakouri 2/16/2011Copper via chains
S. Alavi et al, To be presented at SemiTHERM, 2011.
34
A. Shakouri 2/16/2011
20
25
30
35
40
45
50
55
60
0 50 100 150
Resi
stan
ce (Ω
)
Time (h)
10vias
100via
Resistance shift for different via chainsafter aging at 200C
35
A. Shakouri 2/16/2011
0 hour
83 hours
S. Alavi et al, SemiTHERM, 2011.
Thermal map of 10 via chain under bias
0
5
10
15
20
25
36
A. Shakouri 2/16/2011
0
5
10
15
20
25
0 hour
120 hours
S. Alavi et al, SemiTHERM, 2011.
Thermal map of 10 via chain under bias
37
A. Shakouri 2/16/2011
Relative change in R and DT
0
1
2
3
4
5
6
7
8
9
10
0 20 40 60 80 100 120 140
Rela
tive
chan
ge t
o t=
0
Time t [hour]
Rvia(t)/Rvia(0)
DTvia(t)/DTvia(0)
S. Alavi et al, SemiTHERM, 2011.
38
A. Shakouri 2/16/2011
D. Kendig et al, To be presented at
SemiTHERM, 2011
Thermal/ electroluminescence imaging of LEDs
39
A. Shakouri 2/16/2011
D. Kendig et al, IPRS 2010
Thermal/ electroluminescence imaging of Solar cells
40
A. Shakouri 2/16/2011
Thermal Imaging Analyzer NT200TMicrosanj.com
Technology Transition:Stand alone thermoreflectance imaging system (resolution: 200nm, 100ns, 0.2C, pixels: 1024x1022)
41
A. Shakouri 2/16/2011
Ultrafast Thermal Simulation Techniques
42
A. Shakouri 2/16/2011
THERMAL SIMULATION APPROACHES
vHeat Equation
vApproachesvFinite Difference MethodvFinite Element MethodvGreen’s Function Method => Analogy with image blurring (power blurring)
volumeunitperrategenerationheatcapacityheat
densitytyconductivithermal
timeetemperatur
*
*
2
2
2
2
2
2
======
+¶¶
+¶¶
+¶¶
=¶¶
qcρkt
T
qzTk
yTk
xTk
tTc
p
pr
4W 4W 4W 4W
4W 16W 16W 4W
6W8W4W4W
4W4W4W4W
Power Map Temperature Profile
T. Kemper, Y. Zhang, Z. Bian, A. Shakouri, 12th International Workshop on THERMAL INVESTIGATIONS of ICs and SYSTEMS (THERMINIC), Nice, France,
27-29 September 2006
43
A. Shakouri 2/16/2011
Package Model
HEAT SPREADING FUNCTION (THERMAL MASK)
Finite Element Analysis
Apply Point Heat Source
Thermal Masknormalization
44
A. Shakouri 2/16/2011BOUNDARY EFFECTS
Due to the finite dimension of the Si-Chip, boundary effect is important. => simplification: Use Method of Images and uniform heat compensation
Center Side Edge Corner
V. Heriz, J.-H. Park, T. Kemper, S.-M. Kang, A. Shakouri, International Workshop on Thermal Investigation of ICs and Systems (Therminic), pp.18-25 Sept. 2007
A. Shakouri et al. US Patent 7,627,841
45
A. Shakouri 2/16/2011Inverse problem (temperature -> power)
Measured temperature profile (include noise)
IMAGE DEBLURRINGREGULARIZATION:lTikhonov Regularization (does not
preserve edges) lBilateral Total Variation (Maximum
a Postiori Cost function) Farsiu, et al. IEEE Trans on Image Processing, 13 (10), p. 1327 October, 2004.
(Semitherm 2007 -Xi Wang et al.)
46
A. Shakouri 2/16/2011Inverse problem (temperature -> power)
Measured temperature profile (include noise)
(Semitherm 2007
-Xi Wang et al.)
47
A. Shakouri 2/16/2011
Transient Temperature Change in IC
t (sec)0.1
Input
0.2 0.3 0.4
P2P1
Input Pattern
48
A. Shakouri 2/16/2011
Computation Time
ANSYS:10056 (sec)
Power Blurring: 100 (sec)
0.1 0.2 0.3 0.4 0.535
40
45
50
55
60
65
70
Time (sec)
Tem
pera
ture
(deg
ree
C)
Pow er BlurringANSYS
Transient temperature at the chip center
Virginia Martin Heriz, A. Shakouri et al. THERMINIC 2007
49
A. Shakouri 2/16/2011
ADAPTIVE POWER BLURRING
Length X(cm)
Wid
th Y
(cm
)
Power Map 2D
0 0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
0.1
0.2
0.3
0.4
0.5
A. Ziabari et al, IMAPS, 2010.
Power Blurring
Adaptive Power Blurring
Average execution time for PB, APB_I, APB_II, ANSYS: 0.04s, 0.16s, 0.83s, 16s
50
A. Shakouri 2/16/2011
0 0.5 1 1.5 235
40
45
50
55
60
Diagonal Distance (cm)
Tem
pera
ture
( °C
)
Path along A--A'
Power BlurringANSYSHotSpot
ANSYS HotSpot PB
Computation Time 56 s 0.11 s 0.041 s
Relative error at hottest spot - 29% 0.3%
Absolute temperature
error- 2-9˚C <1˚C
LengthX(cm)
Wid
thY(c
m)
Power Distribution (Watt)
0 0.5 1 1.50
0.2
0.4
0.6
0.8
1
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
LengthX(cm)
Wid
thY(c
m)
Temperature Distribution °C (ANSYS)
0 0.5 1 1.50
0.2
0.4
0.6
0.8
1
45
50
55
60
A
A’
Architecture level thermal simulations
A. Ziabari et al, submitted to SemiTHERM 2011
51
A. Shakouri 2/16/20113D Chip Thermal Simulation
Top Chip
Bottom ChipBond
Heat Sink
Heat Spreader
TIM 1
TIM 2
52
A. Shakouri 2/16/2011Cross-section temperature profiles
0 0.2 0.4 0.6 0.8 185
90
95
100
105
LengthX (cm)
Tem
pera
ture
(deg
ree
C)
ANSYS vs. Power Blurring along b--b' (layer 1)
Power BlurringANSYS
0 0.2 0.4 0.6 0.8 162
64
66
68
70
72
LengthX (cm)Te
mpe
ratu
re (d
egre
e C
)
ANSYS vs. Power Blurring along b--b' (layer 2)
Power BlurringANSYS
Top Chip Bottom Chip
Je-Hyoung Park, Ali Shakouri, and Sung-Mo Kang, “Fast Thermal Analysis of Vertically Integrated Circuits (3-D ICs) Using Power Blurring Method,” Proceedings of the InterPACK Conference,
San Francisco, July 19-23, 2009.
53
A. Shakouri 2/16/2011VISIT OUR WEB SITE
Santa Cruz CampusExclusive research collaboration Sharing and transferring our techniques and tools via industrial affiliate program
Major EquipmentHigh Temperature Cryostat System (4-1000K)Raman and Luminescence Spectroscopy SystemThermal Imaging CameraFemto/Picosecond LaserAtomic Force MicroscopeMid/Near-infrared Thermal Imaging (InSb, InGaAs cameras)OLYMPUS FluoView FV1000-Confocal Microscope
Santa Cruz
NASA
Thermal Characterization Lab at NASA Ames Research Center300 sq feet space for optical/thermal characterization systemResolution: 6mK temperature, 500 nm spatial, DC/Transient -100ns
http://quantum.soe.ucsc.edu/
54
A. Shakouri 2/16/2011
• Metal / semiconductor nanocomposites can improve thermoelectric energy conversion– Mid/long wavelength phonon scattering, Hot electron energy filtering
• Micro Refrigerators on a Chip (need ZT~0.5 on chip to cool 1kW/cm2 hot spots by 10-15C)
• Fast transient thermal imaging (~200nm, 0.2C, 800ps)
• Thermal characterization of high power transistors, ESD devices, solar cells, LEDs and copper vias
• Fast thermal simulation of packaged IC chips (adaptive power blurring) –T-dependent material properties
SummaryA. Shakouri 7/7/2010
A. Shakouri and M. Zeberjadi, "Chapter 9: Nanoengineered Materials for Thermoelectric Energy Conversion," in: Thermal Nanosystems and
Nanomaterials, ed. by S. Volz, Springer, pp. 225-299 (2009)
55
A. Shakouri 2/16/2011
AcknowledgementSenior Researchers: Zhixi Bian, Kaz Yazawa, James Christofferson
Postdocs/Graduate Students: Kerry Maize, Hiro Onishi, Tela Favaloro, Paul Abumov, Phil Jackson, Oxana Pantchenko, Amirkoushyar Ziabari, Bjorn Vermeersch, Gilles Pernot, Shila Alavi, Je-Hyeong Bahk
Collaborators: John Bowers, Art Gossard, Chris Palmstrom (UCSB), Tim Sands (Purdue), Rajeev Ram (MIT), Venky Narayanamurti (Harvard), Arun Majumdar (Berkeley), Yogi Joshi, Andrei Federov (Georgia Tech), Kevin Pipe (Michigan), Josh Zide (Delaware), Lon Bell (BSST), Stefan Dilhaire (Bordeaux), Natalio Mingo (CEA), Sriram Shastry, Nobby Kobayashi, Joel Kubby, Ronnie Lipschutz, Melanie Dupuis, Steve Gliessman, Ben Crow (UCSC), Bryan Jenkins, Susan Kauzlarich (Davis), Steve Kang (Merced)
Alumni: Younes Ezzahri (Prof. Univ. Poitier), Daryoosh Vashaee (Prof. Oklahoma State), Zhixi Bian (Adj. Prof. UCSC), Yan Zhang (Tessera), Rajeev Singh (Sun Power), James Christofferson (Microsanj), Kazuhiko Fukutani (Canon), Je-Hyoung Park (Samsung) , Javad Shabani (postdoc, Harvard), Mona Zebarjadi (postdoc, MIT), Xi Wang (postdoc, UCSB), Helene Michel (CEA), Tammy Humphrey