ultrafast nanoscale electrothermal energy conversion

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


S. Koonin, Former Chief Scientist BPnrg.caltech.edu

Oil Resources

4


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


David MacKay, Sustainable Energy –Without the Hot Air

6


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


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

http://masc.cse.ucsc.edu/

9


Electric Current

HeatingCooling

Peltier Effect (1834)

Reverse of Seebeck effect (electric current produces cooling/heating)

= STc´ ILord Kelvin (Thomson)

10


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


Thermoelectrically-Cooled/Heated Car Seat

Lon Bell BSST

13


Solid-State Thermoelectric Energy

Conversion using Nanostructured

Materials

14


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


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


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


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


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


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


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


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


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


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


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


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


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


0

5

10

15

20

25

0 hour

120 hours


Thermal map of 10 via chain under bias

37


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)


38


D. Kendig et al, To be presented at

SemiTHERM, 2011

Thermal/ electroluminescence imaging of LEDs

39


D. Kendig et al, IPRS 2010

Thermal/ electroluminescence imaging of Solar cells

40


Thermal Imaging Analyzer NT200TMicrosanj.com

Technology Transition:Stand alone thermoreflectance imaging system (resolution: 200nm, 100ns, 0.2C, pixels: 1024x1022)

41


Ultrafast Thermal Simulation Techniques

42


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


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


Transient Temperature Change in IC

t (sec)0.1

Input

0.2 0.3 0.4

P2P1

Input Pattern

48


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


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


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/

http://quantum.soe.ucsc.edu/research/hsc.html

54


• 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


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

ultrafast nanoscale electrothermal energy conversion

Documents