UK Thermoelectric Network Meeting Herriot Watt University

Gao Min School of Engineering, Cardiff University, UK

Key Issues in Design and Fabrication of

Thermoelectric Modules

Gao Min

School of Engineering, Cardiff University

N P

UK Thermoelectric Network Meeting

Herriot Watt University, Edinburgh

13th February 2018


Design and Fabrication Processes

Design

Fabrication

Evaluation

SPECIFICATIONS

Materials Selection Geometry Design

Contact Resistance Barrier & Brazing

Power Output Conversion Efficiency


Selection of Thermoelectric Materials

• Large average ZT

• Matching N and P

• Thermal stability

• Non-toxicity

• Abundances

• Manufacturability

TEMPERATURE (K)

0 200 400 600 800 1000 1200 1400

TH

ER

MO

ELE

CT

RIC

FIG

UR

E-O

F-M

ER

IT, Z

T

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

SnSe

SKD

300K - 850K

h ~ 9%

ZT~ 0.5

h ~ 12%

ZT~ 0.8

(e.g. SKD vs Silicide)

(e.g. Half-heusler vs SKD)

(e.g. Half-heusler vs SKD)

(e.g. Silicide vs PbTe)

(e.g. Silicide vs Bi2Te3)


Design of Thermoelement Geometry

Nature of Heat SourceConstant DT Constant heat flux

l

A

A/l = ?

Pmax =(a ×DT )2

4r

A ×2N

(n + l)(1+ 2rlc / l)2

f =

DT

TH

æ

èç

ö

ø÷

1+ 2rlc

l

æ

èç

ö

ø÷

2

2 -1

2

DT

TH

é

ëê

ù

ûú+

4

ZTH

é

ëê

ù

ûú

l + n

l + 2rlc

é

ëê

ù

ûú

æ

èç

ö

ø÷

G Min, CRC Handbook of Thermoelectrics, 1996, 11-1

Shorter legs (down to ~500mm)

G Min and Rowe, 2002, Energy Conversion & Management , (43) 221-228

P =s

(1+ s)2×

Z

(1+ ZTM )2×

˙ Q 2

l×

l

A×

l

(n + l)(1+ 2rlc / l)2

f =P

˙ Q =

s

(1+ s)2×

Z

(1+ ZTM )2×

˙ Q

l×

l

A×

l

(n + l)(1+ 2rlc / l)2

H Hashim, JJ Bomphrey, G Min, Renewable Energy, 87(2016), 458-463

Longer legs (typically 3-8 mm)

G Min, “TE Design for a Given Thermal Input” J.Electr. Mater. (2013), 42(7), 2239-2242.

Temperature difference (K)

0 20 40 60 80 100

Po

we

r o

utp

ut

(W)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4Theory

Best commercially available module

optimised

cooling water

Solar radiation

Solar absorber

TE generator

cooling water

Solar radiation

Solar radiation

cooling water


Thermoelectric

Materials

Thermal contacts

N P Solder N P Cu

Electric contacts

Thermoelectric

n-p junctionThermoelectric

Module

Materials to Module – Interfaces Challenge

Requirements:

To join and hold the N- and P-type thermoelements together.

• Mechanical strength.

• Temperature stability and durability.

• No reduction in thermoelectric properties.

Role:


oTD

Q

l

oTDTD

2cTD

2cTD

ReR

cR

cR

Q

Influence of Contacts – Theoretical Consideration

Bulk properties:

Contact properties:

ZT will be reduced due to the interfaces. The degree of reduction

depends on contact properties, materials properties, and leg length

ZT =a 2T

RK

ZTe =ae

2T

ReKe

= ?

G Min and M Phillips, 2016, “Preparation and Characterisation of TE Interface” Chapter 6 in Thermoelectric Energy Conversion (Eds. Pineda & Rezania, Wiley-Vch), 111-125


l (mm)

0.1

0.5

1.0

1.5

2.0

2.5

3

5

Manufacturing Requirements

N P

Case study based on Bi2Te3: aims to limiting ZT reduction within 10%.

Assume:

It requires: andL = 2mm


Sample Preparation for Interface Studies

Skutterudite with barrier layer

Copper block

1.3- 1.5 mm

Solder/braze

120-200 mm

Brazing holder

Brazing (or soldering) in argon gas or vacuum


High-resolution Scanning Voltage Probe for Measuring Electrical Contact Resistance

R(x) =VS (x)

Vref

× Rrefr = dR(x) / dx

rc = DR × A

DR

DR

Dx

x

R(x)

perfect contact

- contact DR

- no alloy

- contact DR

- alloy region Dx High Resolution Scanning Voltage Probe (1 mm)

The key is appropriate spatial resolution (~1mm for for bulk modules)



Characterisation of Skutterudite Interfaces

1 2 3 4

0.028

0.030

0.032

0.034

0.036

0.038

0.040

0.042

Resi

stance

/

Distance along sample/mm

Copper region

Unaffected region

Inter-diffusion region

SVP Measurement

High resolution scanning voltage probe

measurement indicates the formation of

alloying region.

Spectrum Co Cu Zn Ag Sb Total

Spectrum 1 12.29 2.39 0.00 1.93 83.39 100.00

Spectrum 2 3.64 34.17 7.50 14.78 39.90 100.00

Spectrum 3 0.00 97.59 0.00 0.00 2.41 100.00

SEM and EDX Analysis

EDX analysis confirms the formation of

the alloying region, which needs to be

minimised in good TE modules.

Pd/Cu/Ag50Cu20Zn28Ni2


Influence of Brazing Temperature on Interface Quality

-0.10 -0.08 -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08 0.10

0

50

100

150

200

250

300

Conta

ct

resis

tivity/m

cm

2

Sample position/cm

654 oC

663 oC

668 oC

2 x 10-5 /cm2

7 x 10-5 /cm2

1.5 x 10-4 /cm2

• Brazing temperature

• Thermoelectric materials

• Brazing atmosphere

• Brazing materials

• Barrier layers

INFLUENTIAL FACTORS


Thermal Contact – Principle of Measurement

x

)(xT

Perfect contact

TE TECu

CTD2

HT

CT

xD

TD

Ensure constant heat flow

through the sample:

QQ



Evaluation of Thermal Contact using IR Microscopy

Hot contact Cold contact

Sample

(a)

Bi2Te3

Pressure contact

Bi2Te3

(b)

Bi2Te3

Sn/Pb solder

Bi2Te3

Similar results obtained using indium or heat sink compound


Module Assembly

DESIGN

Thermoelectric Module Assembly Process

Assembly rig

Bracing materials

Ceramic substrate

Metallisation

High Temperature Skutterudite Module

by CU-RU-LU

Cutting TE legs

Coating barrier layers

ASSEMBLY

J Garcia-Canadas, A Powell, A Kaltzoglou, P Voqueiro and G Min, Journal of Electronic Materials, (2013), 42(7), 1369-137


Rapid Screening by Thermoelectric Impedance Spectroscope

J. Garcia-Cañadas and G. Min, “Impedance Spectroscopy for Thermoelectric Characterisation” J. Appl. Phys. 116 (17), 174510 (2014).

ZT =RTE

Ri

TE

module

ReferencemoduleZT=0.170

~620oCbrazebothsidesZT=0.084

~650oCbrazebothsidesZT=0.117

~620oCbrazeonesideZT=0.103


Determination of Power Output by I-V Curves

0 1 2 3

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Voltage/V

Current/A

50K

100K

150K

200K

250K

300K

350K

400K

440K

0 1 2 3 4 5

0

200

400

600

800

1000 50K

100K

150K

200K

250K

300K

350K

400K

440K

Pow

er/

mW

Current/A

Be aware of errors when used for

determining conversion efficiency


Determine Conversion Efficiency by Triple I-V Curves

hmax =1

4×DTo

T× 1-

VOCO

VOCS

æ

èç

ö

ø÷

Fast Scan

Slow Scan

VOCO

VOCS

Fast Scan – constant temperature difference

Slow Scan – constant heat flux

VOCO - corresponding to the first fast scan

VOCS - corresponding to the second fast scan

s

Open-circuitShort-circuit


Ring-structured Thermoelectric Module

Hot

FluidCold

Fluid

Hot-pressing

Copper

Insulator

N-type

thermoelement

Copper

Insulator

P-type

thermoelement

Insulator


Ring-Structured Thermoelectric Tube

Gao Min and Rowe, 2007, Semicond. Sci. Technol., (27) 21-28


Micro/Nano Converters for Integrated Circuits

Insulating layer/silicon substrate

Membrane

Cold junction

Hot junction

IR focal plane

Metalization

P-type

N-type

Bonding pads

IR focal planeInsulating layer

(membrane)Substrate

Horizontal structure

Ref: G Min and D Rowe, Electronics Letter, 1998, 34(2), 222-223

L.M. Goncalves. P. Alpuim, G. Min, D. M. Rowe, C. Couto, and J. H. Correia, “Optimization of Bi2Te3 and Sb2Te3 thin

films deposited by co-evaporation on polyimide for thermoelectric applications”, Vacuum, 82 (12), (2008), 1499-1502


Thank You

Acknowledgements

Dr. Matthew Philips, Dr. Jorge Garcia-Canadas and Dr. Tanuji Singh

The work is funded by the United Kingdom Engineering and

Physical Science Research Council (EPSRC) under the projects of

EP/K029142 and EP/K022156.

Documents

UK Thermoelectric Network Meeting Herriot Watt University