Emerging materials for Thermal ManagementAl und Cu based diamond composites

L. Weber Laboratory for Mechanical Metallurgy

Ecole Polytechnique Fédérale de Lausanne (EPFL)CH-1015, Lausanne, Switzerland

EIDGENÖSSISCHE TECHNISCHE HOCHSCHULE LAUSANNE

POLITECNICO FEDERALE DI LOSANNA

SWISS FEDERAL INSTITUTE OF TECHNOLOGY LAUSANNE

ÉCOLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE

0.01

0.1

1

10

100

1000

10000

sun set on a beachsun bath at noonlight bulb 100W

cooking plate

LD module package

Pentium 4

IGBT power moduleLD module cavity

First wall ITERsurface of the sun

heat flow density [W/cm

2]

equivalent black body surface temperature [K]

205

364

648

2050

3640

6480

1152

The heat is on!

The heat is on!

Solution:

phase change materials

heat pipes

small active componenttransient heating

cold air flow

small active componentpermanent heating

cooling plate/circuit

spreading/absorbing the heat

large active componentpermanent heating

spreading and transfer mostly transfer

Solution:

High in plane

Medium/high through plane

Solution:

High through plane

Typical requirements on substrate or base-plate materials

• CTE similar to that of GaN and Si (3-5 ppm/K) (passive cycling) or slightly higher (active cycling).

• High thermal conductivity, [W/mK]

• High thermal diffusivity

• Sometimes: electrical conductivity

• Structural properties (stiffness, strength)

δ=λ

cp ⋅ρ

Candidate materials

Metals:

CTE too high

Ceramics:

“no” electrical conductivity, too brittle, CTE too low

0

1

2

3

4

5

6

7

8

9

10

0 5 10 15 20Coefficient of thermal expansion [ppm/K]

metals

ceramics

metalsceramics

area of interest

=> obvious choice:

composites

Composite concepts using carbon material

Common forms of Carbon

Continuous Carbon fibres

Chopped Carbon short-fibres

Graphite flakes

Carbon nanotubes and nanofibres

Diamond (particles and fibres)

Diamond price

Raw material prices 2007:[US$/litre]

Platinum 800’000.-Gold 380’000.-Palladium 150’000.-C-Nanotubes 12’500.-Silver 4’100.-CBN 3’000.-HC carbon fibres 2’400.-Tungsten carbide 1’300.-Tungsten 750.-Ni-Superalloys 700.-Molybdenum 680.-Titanium diboride 500.-Nickel 450.-Aluminium nitride 256.-Titanium 225.-Tin 100.-Copper 72.-Silicon carbide 50.-Alumina 40.-Aluminium 6.-

Industrial diamond price 1994 (after Ashby&Jones):

>1’000’000.- [US$/litre]

Industrial diamond price 2005:

10’000.- down to 600.- [US$/litre]

The making of diamond composites

Liquid metal infiltration process

Alternative routes:

• hot pressing of powder mixtures

• hot pressing of coated particles

Pressure infiltration apparatus

• Induction heating(using a graphite susceptor)

• primary vacuum pump (0.1 mbar)

• Crucible can be lowered on quench (directional solidification)

100 mm

• Cold wall vessel (250 bar, 200°C)Inner side of the wall in contact with a water cooled heat shield

Selected diamond grit

• Mono-crystalline diamond

• Low nitrogen level

• Relatively large size (>100µm)

Net-shape fabrication

Ag-Diamond composites

1. Pure Ag + 60 %-vol diamonds (100µm)

• Low thermal conductivity (270 W/mK)

• High coefficient of thermal expansion (≈17ppm/K)

2. Ag-Si alloy + 60 %-vol diamonds (100µm)

• High thermal conductivity (>700 W/mK)

• Low coefficient of thermal expansion (≈7ppm/K)

Cu-Diamond composites

1. Pure Cu + 60 %-vol diamonds (200µm)

• Low thermal conductivity (150 W/mK)

• High coefficient of thermal expansion (≈16ppm/K)

2. Cu-B alloy + 60 %-vol diamonds (200µm)

• High thermal conductivity (>600 W/mK)

• Low coefficient of thermal expansion (≈7ppm/K)

Matrix alloy development

• What is it that makes an alloying element an

“active” element

• How much active element do we need to get the right interface?

• And what does this quantity of active element do to the matrix properties?

Effect of active element on CTE

Active elements are needed to form carbides at the Metal/diamond (carbon) interface

0

100

200

300

400

500

600

700

0.000001 0.00001 0.0001 0.001 0.01 0.1

Cr concentration in Cu [at/at]

Thermal conductivity [W/mK]

0

2

4

6

8

10

12

14

16

18

matrix thermal conductivitycomposite thermal conductivitycomposite CTE

CTE [ppm/K]

insulating inclusion

carbide stabilitylimit

0

100

200

300

400

500

600

700

800

0.00001 0.0001 0.001 0.01 0.1 1

B concentration in Cu [at/at]

Thermal conductivity [W/mK]

0

2

4

6

8

10

12

14

16

18

matrix thermal conductivitycomposite thermal conductivitycomposite CTE

CTE [ppm/K]

insulating inclusion

carbide stability limit

Ag-Si: thermal conductivity

L.Weber, Metall. Mater. Trans. 33A (2002) 1145-50

0

50

100

150

200

250

300

350

400

350 450 550 650 750 850

annealing temperature [°C]

Ag-4 Si

After infiltration

Ag-Si-X: alloy requirements

The ternary alloying element X should have/generate

• “no” solubility in solid Ag

• some solubility in liquid Ag

• reduced Si-activity in the solid state

weak silicide-forming element

Ni Fe Mn

Ag-Ni binary system

• Ni content limited to 0.3-0.4 at-%

• Resistivity increase due to Ni<0.05µΩcm (after HT @ T<700°C) and is maximum about 0.4 µΩcm after HT @ 950°C.

800

900

1000

1100

1200

1300

1400

0 0.2 0.4 0.6 0.8 1

Ni-content [at-%]

temperature [K]Ladet (1976)

Stevenson & Wulff (1962)

this study

(Ag)

liquide

liquide + (Ni)

(Ag) + (Ni)

Ag-Ni-Si: Si activity

-60

-50

-40

-30

-20

-10

0

0 0.2 0.4 0.6 0.8 1

xSi [-]

∆Gf [kJ/mol]

Acker (1999)

Tokunaga (2003)

Kaufmann (1979)

700°C

Ni3Si2

NiSi

NiSi2

Ag-Ni-Si: thermal conductivity

600

700

800

900

1000

1100

1200

0 1 2 3 4 5

∆ρ [__ ]cm

[ ]temperature KAg-2at-%SiAg-0.5SiAg-0.25SiAg-0.3Ni-0.25Si

∆ρ [µΩcm]

0

50

100

150

200

250

300

350

400

450

700 800 900 1000 1100

temperature [K]

thermal conductivity [W/mK]

Ag-2at-%Si

Ag-0.5Si

Ag-0.25Si

Ag-0.3Ni-0.25Si

measurements

Typical situation after infiltration

GPI SC

Thermal conductivity

660 110

CTE 10-12 17-25

Kinetic effects: Al-diamond

Interface study of Al-Diamond composites

Comparison of GPI and Squeeze Casting

Influence of diamond volume fraction on CTE

Interesting CTE range can be achieved with mono-modal particle size distribution

Low pressure infiltration is possible

4

5

6

7

8

9

0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85

volume fraction diamond [-]

bimodalmonomodal

Al-SiC

Influence of diamond volume fraction electrical conductivity

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

35 40 45 50 55 60 65 70

fraction non-conducting phase [vol.-%]

normalized el. conductivity [-]

5 µm, angular12 µm, angular30 µm, angular58 µm, angular100µm, angularbimodal 3µm/30µm, angularbimodal 5µm/30µm, angularbimodal 5µm/58µm, angular bimodal 12µm/30µm, angular3-P SCS (spheroids, aspect ratio 0.275)mean-field approach (spheroids, aspect ratio 0.275)differential scheme (spheroids, aspect ratio 0.275)5 µm, acicular5µm acicular29 µm angular5 µm, slip cast

Going from 60 to 75

pct vol diamond

reduces the el.

conductivity by a

factor >2!

Importance of the interface transfer problem

Electrical conductivity: • High phase contrast• No effect of interface resistance=> no effect of phase region size and field-line distortion

Thermal conductivity: • low phase contrast=> Effect of interface resistance

Various models (extension to finite volume fractions):

Effective particle thermal conductivity:

Effective particle properties

€

d ,eff =λ d

1+λ dhbd r

=λ d

1+ B

pλmh

€

c = f λ m ,λ dλm

,hbd ,r,Vp ⎛

⎝ ⎜

⎞

⎠ ⎟= g λm ,

λ d ,eff

λ m,Vp

⎛

⎝ ⎜

⎞

⎠ ⎟

Indirect measurement of the ITC —

size effects

0

100

200

300

400

500

600

700

800

900

1000

0 20 40 60 80 100

particle radius [µm]

composite conductivity [W/mK]

Exp Ag-Si/diamond

DEM; h=6.6 10^7 W/m2K

Small particles:

• Higher strength

• Better machinability

• Lower thermal cond.

• Metal diamond composites are a promising material for next generation thermal management solutions.

• They can exhibit twice the conductivity of pure silver, while having a coefficient of thermal expansion similar to semiconductor devices.

• The interface is extremely important for both, thermal conductivity and coefficient of thermal expansion.

Conclusions