Transient Through-Silicon

Hotspot Imaging

MEPTEC “Heat Is On” Symposium

March 19, 2012

K. Yazawa* Ph.D. Research, Microsanj LLC.,

D. Kendig (Microsanj), A. Shakouri (Purdue Univ.)

Info@microsanj.com

+1 (408) 256-1255

www.microsanj.com 1

mailto:Info@microsanj.com

Thermal characterization needs

Wakil, J., et al

“Thermal Development, Modeling and Characterization Of The Cell Processor

Module” Proc. of the ITHERM 2006, 2006, pp.289-296

1st gen Cell Broad Band Engine

Hot spot loads - Changing in around 1,000 -10,000cycles ~ nano seconds

Simulation

2

Transient thermal imaging techniques

MethodResolution Imag-

ing?Notes

x(mm) T (K) t (sec)

m thermocouple 50 0.01 0.1-10 No Contact method

IRThermography 3-10 0.02-1 1m Yes Emissivity dependent

Lockin IR

Thermography3-10 1m NA Yes Need cycling

Liquid Crystal

Thermography2-5 0.5 100 Yes

Only near phase transition

(aging issues)

Thermo-reflectance 0.3-0.5 0.08800p-

0.1mYes Need cycling

Optical

Interferometry0.5 100m

6n-

0.1mScan

Indirect measurement

(expansion)

Micro Raman 0.5 1 10n Scan 3D T-distribution

Near Field (NSOM) 0.05 0.1-1 0.1m ScanS/N dependent

Tip/sample interaction

Scanning thermal

microscopy (SThM)0.05 0.1

10-

100mScan

Contact method surface

morphology

Limitations of conventional infrared

thermography

• Spatial resolution limitation by wave length

• Time resolution limitation by sensor response

• Large power dissipation of a whole chip

– need heat sink

Based on the theory of

Planck blackbody emission

T

R

R and vary sharply

due to interference

Spatial selectivity: a few mm

Spectral resolution: ~1 nm

Sensitivity: DR/R~ 3.10-5 in 1 min

Tessier et al. (2006)

Thermoreflectance coefficient for different surfaces

Basic set up of Thermoreflectance

imaging

Temperature Controlled Stage

Camera

TEC

Scientific grade CCD 30fps

probes

Microscope objective 1x-100x

Computer

Sample

Pulsed LED light source

T/C for calibration

Beam Splitter

Function Generator

Image/signal processing software

probes

CCD speed of 30fps

Applications

Non-invasive surface temperature measurement

of electronic & optoelectronic devices

Thermal design validation of semiconductors

Microelectronic component analysis & quality

control

Device defect & failure analysis

Production line testing

7

• Computer (SanjVIEWTM 2.0 with Transient Thermal

Imager SW Module)

• High Speed Signal

Generator

• Thermoreflectance Imaging

Module & Biasing (TIM II)

II. Transient Imaging System

I. CCD Microscopy Head

Monitor

NT210A TIA System Diagram

• Microscope • CCD Camera • LED

GP

-IB

8

Key Features of the equipment

Submicron spatial resolution (compared to 5-10

µm for infrared microscopy)

High temperature resolution (0.1 ºC)

High speed transient imaging (100ns resolution)

Through-the-Substrate imaging if near IR is used

Fully featured software package

Low cost solution

Use visible light for foreside (no IR objectives)

9

Lock-in Imaging Result

DC Reflection

AC Reflection

◦ Phase

◦ Amplitude

Mask

◦ Identify different materials,

cooling/heating

regions

Normalization Phase

Raw AC data Image Mask

DC Reflection AC Phase

AC Amplitude Mask

Acquisition time: 5 minutes

Through-the-Substrate Imaging

Top view (visible light)

Through-the-substrate

view (IR light)

Si substrate 200 microns

Gold Contact Layer 3 microns Superlattice Micro Cooler Layer

3 microns

Current Flow Causing Joule Heating

Bottom Substrate Thermoreflectance

Visible Light

Backside Thermoreflectance

1310nm Laser

Topside Thermoreflectance

Visible Light

Small optical absorption by using below bandgap

energy laser

Bottom Contact

R2, Cth2, DTD

R1, Cth1, DTB

Backside Thermal Characterization Using near IR light, temperature profile from the

backside through silicon is obtained

0 100

2 10-5

4 10-5

6 10-5

8 10-5

1 10-4

0

0.5

1

1.5

2

2.5

3

0 50 100 150 200 250 300 350 400

TopSide and Backside Thermoreflectance

Device in Cooling Mode

Backside Cooling

Topside Cooling

Ba

cksi

de

Th

erm

ore

flect

an

ce S

ign

al(

Vo

lts)

To

psid

e C

oo

ling

(de

gre

es C

)

Device Current(mA)

Backside

Signal (V)

Topside

Cooling (ºC)

Topside

Thermal Map

Backside

Thermal

Map

Visible

Light

reflected

from Top

Surface

Near IR Light

reflected from

Back Surface

(metal)

12

Polysilicon via chain shows uniform power dissipation. If a single element was causing the higher resistance in the chain, it would have a higher temperature compared to the other vias. a) Optical image b) Thermal image c) Temperature profile d) Merged optical/thermal image

to show location of heating

Verify interconnect and via integrity

Determine if defects are due to single elements or if they are uniform throughout the whole chain. (2D temperature/ power map instead of a 1 point electrical measurement)

Aa’ Aa

Aa’ Aa

(a)

(b)

(c)

(d)

13

High Speed Thermal Imaging (800ps)

J. Christofferson et al., Proceedings of Int. Heat Transfer Conf., August 2010

0

5

10

15

20

0 50 100 150

Heating on Metal Above Via

550nm 600mA350nm 400mA

Tem

pera

ture

Cha

nge

from

Am

bien

t

Delay in Nanoseconds

350nm Via, 400mA

550nm Via, 600mA

Study of heating in submicron interconnect vias

14

Heating in Electro Static Discharge Devices

K. Maize, V. Vashchenko et al, IRPS, 2011.

VBNC

= 88V, VSCR STEADY STATE

= 3.4V, ISCR STEADY STATE

= 1.33A

D R/R

0

5

10

15

20Optical Image 300ns

40µm Anode

Cathode

VBNC

= 88V, VSCR STEADY STATE

= 3.4V, ISCR STEADY STATE

= 1.33A

D R/R

0

5

10

15

20

10

ΔT [K]

0

20

Snapback current =1.22A.

Transient Thermal Image: 30ms

Calibrated for metal, Cth=1.0e-4

DT [K]

0

200

400

600

Transient Thermal Image: 170ms

Calibrated for metal, Cth=1.0e-4

DT [K]

0

200

400

600

30µs 170µs

Transient Thermal Image: 30ms

Calibrated for metal, Cth=1.0e-4

DT [K]

0

200

400

600

ΔT [K] 600 400 200 0

300ns

15

High Magnification Comparison

0 20 40 60 80 100 120 14030

35

40

45

50

Distance along profile

20

25

30

35

40

45

50

55

60

0 50 100 150 200

Tem

pera

ture

(C

)

Distance (µm)

Te

mp

era

ture

(C

)

Distance (pixels)

50 100 150 200 250

50

100

150

200

25020

25

30

35

40

45

50

50 100 150 200 250

50

100

150

200

25020

25

30

35

40

45

50

120μm

15x 20x

Performing AC measurement & pulsing the DUT, much more localized peak temperatures can be found since diffusion length is inversely proportional √f. Thermoreflectance images show sharp peaks on top of the 4 μm wide heater lines.

Thermoreflectance 3 V, 50 μs pulse 5 min average

Infrared 2 V DC

16

High Magnification Images

• Hot-spot defect in Multi-finger MOSFET gate using Thermoreflectance. The hot-spot FWHM is 1.4 μm.

• Overlay of the optical & thermal image shows precise location of defect on the transistor.

0

10

20

30

40

50

60

70

-3 2 7 12

Tem

pera

ture

(a.

u.)

Distance (μm)

50x

17

Cu via chain interconnects

Si

Cu

TaN ~10nmSiN

via via via via

Cross section zoom view

500nm

f500nm

JEDEC standard JESD22-A103D High Temperature Storage Life 200oC

S. Alavi1,2, K. Yazawa*1, G. Alers2, B. Vermeersch1, J. Christofferson1 and A. Shakouri1,2, “Thermoreflectance Imaging for Reliability Characterization of Copper Vias”, Semi-Therm27, San Jose California, 2011

Initial mechanical stress

Thermal expansion +

Void nucleation growth?

Suspected mechanism

18

Thermal Images of 10-via chain

0 6h 20h 43h

Before and after thermal treatment

83h 100h 120h 150h

Temperature scale: DT=0-30 degC

19

Relative change in R and DT

Rvia(t)/Rvia(0)

0

1

2

3

4

5

6

7

8

9

10

0 20 40 60 80 100 120 140

Re

lati

ve c

han

ge t

o t

=0

Time t [hour]

DTvia(t)/DTvia(0)

10-via chain

Time [Hour]

20

Diode Sensor

Resistive Heater block ( 7.6 Ω )

The TEA Thermal Test Chip TTC-1002 is based on a unit cell with two resistors and four diode temperature sensors in each cell. The two resistors in each unit cell are laid out to occupy 86% of the available area within the electrical contact pads. 1m

m

IR vs Thermoreflectance

TTC-1002 Thermal Test Chip

21

Thermoreflectance & IR Images

Thermoreflectance 29 V, 0.5 Hz, 2 min average

TEA 30V

20

30

40

50

60

70

80

90

100

Infrared 30 V DC

95

20 1mm

Non-uniform heating due to packaging is seen in both Thermoreflectance & IR imaging.

22

- Similar edge-effect issue at high magnification due to sample movement - Thermoreflectance images show influence of surface roughness of material. Average calibration coefficient can be obtained for the rough regions

50100150200250300350400450500

50

100

150

200

250

300

350

400

450

50065

70

75

80

85

90

95

100

50 100 150 200 250 300 350 400 450 500

50

100

150

200

250

300

350

400

450

500 65

70

75

80

85

90

95

100

100

20

Thermoreflectance 30 V, 0.1 Hz, 2 min average

Infrared 30 V DC

5x

200um

Thermoreflectance & IR Images

23

Diode Temperature Map 4 5

93.5 99.2

1

93.5 3

95.7

2 6

92.0 98.2

Diode

Senso

r

4

1

3

2

5

6 TEA 30V

20

30

40

50

60

70

80

90

100

Diode sensor temperature measurements were measured at 6 locations. 8% temperature variation was seen across the test chip with the diodes.

24

Temperature Correlation Between Measurements

ROI

• Thermoreflectance & diode measurements within 0.3% of each other, while IR measurement was within 6%. • Error in IR measurement due to poor emissivity of metal surface & thermal expansion-induced image shift at high magnification.

0

20

40

60

80

100

120

0 5 10 15 20 25 30 35

Te

mp

era

ture

(C

)

Voltage (V)

Thermoreflectance

Diode

IR

25

Transient Thermoreflectance Image

• TTC heating in response to a 60 V, 100 μs device pulse.

• In first few milliseconds, heat from resistors has not been transferred to the substrate or other metal layers. (the blue region in center of the resistor is due to passivation artifact)

• At shorter time-scales the heating in the device is more uniform suggesting the temperature non-uniformity at DC is due to packaging.

Thermoreflectance 60 V, 100 μs pulse

2 min average (2 x 2 binned)

30

20

26

TTC Transient Response Short Time Scale

(Thermoreflectance) Long Time Scale (IR)

20

30

40

50

60

70

0 5 10 15 20 25

Tem

pera

ture

(C

)

Time (s)

20

25

30

35

40

45

50

55

60

65

-0.5 0 0.5 1 1.5

Tem

pera

ture

(c)

Time (s)

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0.0 1.0 2.0 3.0 4.0

Ch

an

ge i

n T

em

pera

ture

(C)

Time (ms)

Heater

Substrate

• (above) Thermal transient of TTC in response to a 30 V, 1 ms pulse. Note the slow response of the substrate

• (right) Thermal transient of TTC in response to 10 V step function. ~0.5 s for DUT to reach steady state.

27

Backside transient

Chip on PCB

LED

CCD

Microscope

28

5 x 5 array flip chip, through silicon 800um

Backside IR

Optical image Thermal image

29

Backside thermoreflectance

thermal Image

Heated section

TEA flip chip mounted on PCB

Si 500mm

PCB

30

Thermal transient profile

-2

0

2

4

6

8

10

12

0 0.002 0.004 0.006 0.008 0.01

Ch

an

ge in

te

mp

era

ture

[K

]

Time [sec]

Substrate

Heater

31

Identification of parasitic

capacitances

0

1

2

3

4

5

6

7

8

9

10

0 0.002 0.004 0.006 0.008 0.01

Te

mp

era

ture

ex

ces

s [K

]

Time [sec]

o Measured - Model

Heater metal

Heater body

Chip substrate

32

Conclusions There are limitations of a conventional thermography

that uses IR emission detection

◦ Spatial resolution limitation by wave length

◦ Time resolution limitation by sensor response

◦ Large power dissipation of a whole chip – need heat sink

Thermoreflectance imaging is capable of:-

◦ Submicron spatial resolution with 1.5mm wave length

◦ 100nsec of time resolution both for foreside and back side

◦ No heat sink is required while low duty biases are applied

Future study could combine the packaging impact

33

Acknowledgement

Thermoreflectance imaging technology

has been transferred from University of

California Santa Cruz.

Fundamental research publications are

available at Web page of the Quantum

Electronics Group at UCSC:

http://quantum.soe.ucsc.edu

34

Questions?

info@microsanj.com

35

http://www.microsanj.com/

Visit