Thermoelectric Conversion of Waste Heat to Electricity in an IC Engine Powered Vehicle

Principal Investigator: Harold SchockPrepared by:

Harold Schock, Eldon Case, Jonathan D’Angelo, Andrew Hartsig, Tim Hogan, Mercouri Kanatzidis, James Novak, Fang Peng, Fei Ren,

Tom Shih, Jeff Sakamoto, Todd Sheridan, Ed Timm

08/06/2008

Supported By: Acknowledgement: ONR support US Department of Energy under MURI ProgramEnergy Efficiency Renewable Energy (EERE) Mihal Gross, Project MonitorJohn Fairbanks and Samuel Taylor, Contract Monitors

IOWA STATE

UNIVERSITY

3D CFD Analysis

Iowa State / MSU

• Couple and Module IssuesConvection and radiation between

legs with and without insulationCurrent distribution, Joule heating,

Heat fluxes

• Electrical energy production

• Unsteady heat transfer analysis to and from modules (3D, pulsatile, comp.)

• Complete engine system- f(x,t)

• Temperatures and heat flux

• EGR energy

• Energy in exhaust (T, P, m)

• Turbine work, inlet/outlet temperatures

TEG Design and Construction

MSU/JPL

• Generator design• TEG materials selection• Mechanical and TE material property characterization including Weibull analysis• FEA analysis• Leg and module fabrication methods

• Design of electrical energy conditioning and utilization system

• Control system design and construction

• Inverter, Belt Integrated Starter-Generator Selection

Systems for Utilizationof Electrical Power Recovered

MSU

Implementation of a Thermoelectric Generator with a Cummins ISX Over-the-Road Powerplant

Engine-TEG Simulation and Experimental Verification

MSU / Cummins

6 Cyl. Engine Test Data

Cummins

P2 - Single cylinder +TEG Demo

MSU

Goals and ObjectivesGoals and ObjectivesUsing a TEG, provide a 10% improvement in fuel economy by converting waste heat to electricity used by the OTR truckEvaluate currently available thermoelectric materials to determine optimum material selection and segmentation geometry for this applicationDevelop TEG fabrication protocol for module and system demonstrationDetermine heat exchanger requirements needed for building TEGs of reasonable lengthDetermine power electronic/control requirementsDetermine if Phase 2 results make an engine demo in Phase 3 reasonable

Important BarriersImportant Barriers• Design of heat exchanger is a major challenge with heat transfer

coefficients needed which are 5x higher than without enhanced heat transfer modes

• Reliable thermoelectric module fabrication methods need to be developed for the new high efficiency TE materials

• Status at MSU – Routine skutterudite production of hot pressed legs underway– Techniques for fabricating hot pressed LAST/LASTT still being developed– Module production methods are still under development

• Material strength and thermoelectric properties must meet life cycle performance criteria ….nanostructures in LAST have survived for six months at 600C

• Temperature dependant material properties are critical in order to conduct a detailed and accurate generator design

• Powder processing methods are being refined to provide increased strength while maintaining thermoelectric properties of ingot forms of the material

• ZT for the temperature ranges (700K) for best TE materials are about 1.4 and need to be closer to 3.0 to reach the efficiency goals requested by DOE

Accomplishment to dateAccomplishment to dateSystems for ingot synthesis and leg preparation demonstrated

70 hot pressed pucks of LAST, LASTT and skutterudite at MSU since 4/07Tube furnaces (~500 gms) and extensive powder processing facilities at MSUSegmented legs with an efficiency of 14.5% demonstrated at JPL

Module fabrication during past year at MSUFabrication of 2, 4 and 8 leg LAST/LASTT modulesFabrication of 2, 4 and 16 leg (40 watt) skutterudite modulesDemonstrated aerogel insulation technique for 16 leg module

Power electronic modules being designed and tested at MSUTemperature dependent elastic moduli and thermal expansion coefficients have been measured for LAST in collaboration with Oak Ridge National LabTransport measurements conducted by MSU have been verified by Northwestern, JPL, Iowa State and the general literatureNew material systems based on skutterudite composites and PbTe-PbSsystems have been examined and appear to offer promising improvementsAnalytical studies performed for various operation modes and conditions

Geometries for high efficiency heat transfer rates evaluatedFinite element analysis of pressing, contact metallization has been conducted and generator design is nextEfficiency improvements for various operational modes for the Cummins ISX engine evaluated for various geometriesHeat transfer studies excellent insulation and hot side convection required

Mechanical and thermal characterization of LAST and LASTT

•In service, the TE elements will be subjected to both thermal gradients and thermal cycling.

•Analytical or numerical analysis of the stress-strain behavior requires knowledge of the elastic moduli and the coefficient of thermal expansion as functions of temperature

•For both LAST and LASTT, we have determined the elastic moduli by the Resonant Ultrasound Spectroscopy technique and the coefficient of thermal expansion by thermal-mechanical analysis and by high temperature x-ray diffraction.

Temperature dependent elastic moduli and thermal expansion coefficient

300 400 500 600 70035

40

45

50

55

/ Heating/Cooling, MSUHP6 (n-type, Ag0.43Pb18Sb1.2Te20)

/ Heating/Cooling, MSUHP4 (p-type, Ag0.9Pb9Sn9Sb0.6Te20)

You

ng's

mod

ulus

(GPa

)

Temperature (K)300 350 400 450 500 550 600 650

16

20

24

28

32 MSUHP18 (p-type, Ag0.9Pb9Sn9Sb0.6Te20) MSUHP6 (n-type, Ag0.43Pb18Sb1.2Te20)

CTE

(x10

-6/K

)

Temperature (K)

• The study on temperature dependent elastic moduli and thermal expansion has been conducted in collaboration with the High Temperature Materials Laboratory, Oak Ridge National Laboratory.

FEA Analysis of Metallization Process

1. Materials in this analysis: N,P type skutterudite, Cobalt and Titanium

2. Study’s goal was to determine stress concentrations in regions of metallization

3. Temperature dependent material properties including CTE and elastic modulus needed for FEA and later material fatigue properties

4. Heat transfer rates on the mold surface can influence cooling rates and stress concentrations

Bulk Thermoelectric Materials

0

0.5

1

1.5

2

200 400 600 800 1000 1200 1400

ZT

Temperature (K)

Bi2Te

3

Pb18

Ag0.86

SbTe20

PbTe

CoSb3

La2Te

3

SiGe

PbTe-PbS(8%)

n-type Materials

Ba0.30

Ni0.05

Co3.95

Sb12

Mg2Si

0.6Sn

0.4Mg2Si

0.4Sn

0.6

0

0.5

1

1.5

2

200 400 600 800 1000 1200 1400

ZTTemperature (K)

CsBi4Te

6

Bi2Te

3

Na0.95

Pb20

SbTe22

Zn4Sb

3

Ag0.5

Pb6Sn

2Sb

0.2Te

10

Tl9BiTe

6

CeFe4Sb

12

PbTeYb

14MnSb

11

(AgSbTe2)0.15

(GeTe)0.85

p-type Materials

Ce0.28

Fe1.5

Co2.5

Sb12

5nm5nm

Best Hot Pressed Samples

0

0.5

1

1.5

2

300 400 500 600 700 800 900

ZT

Small Ingot

Large Ingot

PbTe

Hot Pressed(Increasing #of Cycles)

Temperature (K)

Ag0.86Pb18SbTe20

0

0.5

1

1.5

2

300 400 500 600 700 800 900

ZT

Small Ingot

Large Ingot

PbTe

Hot Pressed(Increasing #of Cycles)

Temperature (K)

0

0.5

1

1.5

2

300 400 500 600 700 800 900

ZT

Small Ingot

Large Ingot

PbTe

Hot Pressed(Increasing #of Cycles)

Temperature (K)

Ag0.86Pb18SbTe20

• Hot pressed samples initially exhibit lower ZT than ingot, but improve with repeated temperature cycling (healing of grain boundaries? stress relaxation?)

• Nanostructures persist after powder processing and hot pressing

• ZT calculated above using thermal conductivity from measured cast samples.

N-type skutterudite material development

Skutterudite crystal structure

• Background– High ZT reported in the 300-800K temperature range for

Bax Yby Co4 Sb12 skutterudite compositions1

– High ZT values mainly attributed to low lattice thermal conductivity due to the broad range of resonant phonon scattering provided by the Ba and Yb fillers

– Samples used for this study were prepared by a multi-step synthesis process, potentially difficult to scale-up

• Goal– Develop a scalable synthesis process for Bax Yby Co4 Sb12 skutterudite

compositions and evaluate TE properties in a first step– Evaluate applicability for integration into advanced TE couples for

waste heat recovery applications• Approach

– Ball milling • High-energy ball

mills: ≤ 15 g loads• Planetary ball mill:

≥ 50 g loads– Hot-pressing

• Graphite dies and plungers

1X. Shi et al. APL 92, 182101 (2008)

Planetary ball millHigh-energy ball mill

Hot-pressed pucks and disks of Bax Yby Co4 Sb12

Bax Yby Co4 Sb12 : initial transport properties results

• Ball milled Bax Yby Co4 Sb12 - initial transport properties– ZT ~ 1.3 at 873K (consistent with previous report)– ~ 40% improvement over n-type PbTe in the 873K-373K

temperature range– ZT improvement over doped-CoSb3 appears to be due to:

• Lower thermal conductivity (double rattler)• But also higher carrier mobility

0.00.10.20.30.40.50.60.70.80.91.01.11.21.31.41.5

300 400 500 600 700 800 900 1000

Temperature (K)

BaxYbyCo4Sb12

Pd & Te doped CoSb3

n-PbTe

1

10

100

1000

1E+18 1E+19 1E+20 1E+21 1E+22

Hall carrier concentration (cm-3 )

Pd & Te doped n-type CoS3

BaxYbyCo4Sb12

p-Ce

1 Fe3 R

u1 Sb

12

n-Ba

x Yby C

o4 Sb

12

p- Bi0.4Sb1.6Te3 n- Bi2Te2.9Se0.1

Cold-shoe Cold-shoe

Hot-shoe

≤ 873K

373K

~ 473KIllustration of skutterudite-

Bi2 Te3 couple

TH = 873K - TC =373K

TH = 773K - TC =373K

TH = 773K - TC =373K

With Bi2 Te3 segments 11.8 10.0 7.9

Without Bi2 Te3 segments

10.7 8.8 6.75

Couple efficiency (%)

At equivalent carrier concentration, the Hall mobility for Bax Yby Co4 Sb12 is higher than that

for doped CoSb3

PbTe – PbS system

1Androulakis, J. et al., JACS 2007, 129, (31), 9780-9788.

Background

•The material PbTe – PbS 8% has been shown to exhibit at enhanced ZT ~1.41 700 K.

•Reason for high ZT:

•high power factor at 700 K (17-19 uW/cmK2))

•Very low total thermal conductivity (0.8 W/mK)

•Low lattice thermal conductivity is the result of nanostructures formed by the spinodal decomposition and nucleation and growth phenomena

PbTe – PbS 8% nucleation and growth creates inhomogeneities

on the micro and naoscale

PbTe matrix PbS

S

TePb

EDS linescans show presence of PbS particles within the PbTe matrix.

SEM micrographs show presence of spinodal decomposition, nucleation and growth

EDS analysis

Kanatzidis effort

22 mm

Large-scale Synthesis of (Pb0.95 Sn0.05 Te) – PbS 8%

300 400 500 600 700 800

0.4

0.6

0.8

1.0

1.2

1.4

Pb0.95Sn0.05Te - PbS 8% : small scale large scale

ZT

Temperature, K

300 400 500 600 700 8000.40.60.81.01.21.41.61.82.0

Ther

mal

Con

duct

ivity

, W/m

K

Temperature, K

Pb0.95Sn0.05Te - PbS 8% κtot

κlat

Nanostructures in PbTe – PbS reduce lattice thermal conductivity, increasing ZT.

Thermoelectric properties are repeatable

•Successful preparation of ~100 g material per batch

300 400 500 600 700 8000

500

1000

1500

2000

2500

Elec

trica

l Con

duct

ivity

, S/c

m

Temperature, K

Pb0.95Sn0.05Te - PbS 8% small scale Pb

0.95Sn

0.05Te - PbS 8% large scale

Parasitic Contact Resistance and Efficiency

h

o

QP

=η

pn

chchccccc

RRRRRR

RR

++++

== 2121δ

M. H. Cobble, “Calculations of Generator Performance,” CRC Handbook of Thermoelectrics, CRC Press, 1995.

RRc=δ

0

0.05

0.1

0.15

0.2

0.25

0 0.5 1 1.5 2 2.5 3

η

ZTavg

δ = 0.0δ = 0.1

δ = 0.5

Th = 800KTc = 300K

2chR1chR

2ccR1ccRn p

oR

+-

I

cT

hT

hQ

Four Leg Modules – Soldered Contacts

0

0 .0 5

0 .1

0 .1 5

0 .2

0 .2 5

0 .3

0 .3 5

0 .4

0

0 .1

0 .2

0 .3

0 .4

0 .5

0 1 2 3 4 5 6

Out

put V

olta

ge (V

) Output Pow

er (W)

O u tp u t C u rre n t (A )

0

0 .0 5

0 .1

0 .1 5

0 .2

0 .2 5

0 .3

0 .3 5

0 .4

0

0 .1

0 .2

0 .3

0 .4

0 .5

0 1 2 3 4 5 6

Out

put V

olta

ge (V

) Output Pow

er (W)

O u tp u t C u r r e n t (A )

Voc

0.43W

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0 1 2 3 4 5

y = 0.39537 - 0.09177x R= 0.99995

Vol

tage

(V)

Current (A)

Thot = 885 KTcold = 300 K

Isc

Thot = 876 KTcold = 300 K

Voc

Isc

0.45 W

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 1 2 3 4 5 6

y = 0.3171 - 0.057988x R= 0.99998

Vol

tage

(V)

Current (A)

8-Leg ModulesModule # Module RT

Resistancen-type leg

p-type leg

Module 1 0.026 Ω ETN204 ETP44

Module 2 0.026 Ω ETN204 ETP44

Module 3 0.027 Ω ETN204 ETP44

Module 4 0.027 Ω ETN204 ETP44

Module 5 0.026 Ω ETN204 ETP52

Module 6 0.027 Ω ETN204 ETP52

Module 7 0.025 Ω ETN204 ETP52

Expected module resistance = 24mΩ (~3mΩ per leg)

δ = 0.13

0

50

100

150

200

250

300

0 1 2 3 4 5 6

Vol

tage

(mV

)

Current (A)

Thot

= 387K

Thot

= 466K

Thot

= 637K

Thot

= 753K

Thot

= 814K 8-leg module #5

Tcold

= 300K

0

50

100

150

200

250

300

0 1 2 3 4 5 6

Volta

ge (m

V)

Current (A)

Thot

= 387K

Thot

= 466K

Thot

= 637K

Thot

= 753K

Thot

= 814K8-leg module #4

Tcold

= 300K

Module Resistance Values After Temperature Cycle

p n p n

10.19mΩ

3mΩ

3.3mΩ 6.73mΩ

8mΩ

7.76mΩ

6.13mΩ6.54mΩ

p n p

+ ‐

n

Total module resistance = 51.65 mΩ

δ = 0.91

Heater 700C

Cooling 100Cunicouple

p-leg n-leg

After testingTest set up

Skutterudite Unicouple made and tested at MSU

Belljar system (1atm argon)

•Unicouple power output ~ 2.5 Watts per 0.67cc of material

•Peak power at ~30 Amps, test limited by power supply current limitations

Skutterudite Unicouple 1

0

0.05

0.1

0.15

0.2

0 10 20 30 40 50 60

current (Amps)

pote

ntia

l (Vo

lts)

0

0.5

1

1.5

2

2.5

3

pow

er (W

atts

)

Skutterudite 16-leg module•Low temperature test confirmed Open Circuit Potential is in good agreement with what is expected

•Materials used in this module have ZT necessary for 40W power output: 700C- 100C assuming no parasitic contact resistance

16-leg Skutterudite module Fabrication and Testing

700W Heater

Water cooled

Cummins ISX 6 cylinder diesel engine

Thermal Power Split Hybrid – Options Using the electric power recovered from waste heat

I E

WheelsFD

Trans.

C

T

Σ

Induction Intercooler

InductionAir

EGR

EGRMixer

ESS, Batt+ Ultra-Cap

Th, qh

Pe @62% = Y kWPe @ 100% = Z kW

Pow

er E

lect

30 -

50 k

VA EGR Cooler

WP

Engine Coolant

Air

Exhaust

CoolantPump

TEG-1

Rad

iato

r

T/C or clutch

B-IM

G

ExcessElectricalPower

PoweredAncillaries

Cool Exhaust OutTEG-2

ηINV = 0.96ηBIMG =0.93ηmi = 0.89

X % of exhaust to EGR, (100- X) % of exhaust to turbine.

Additional energy recovery opportunity

Pm =249+TEG@ 100% , TEG@ 62%

Single TEG with exploded view of a module

ISX Engine Operating Conditions for ESC Duty Cycle Modes

HX: Heat Transfer Enhancement

Goal:

How to get high heat-transfer rate?

high heat transfer ratelow pressure drop

ribsDimplesvortex generatorshybrid

(combinations of ribs, dimples, …)

no rib

with rib

New Design Concepts w/ Heat Exchanger Cont.

• Dual TEGs Concept also modified

• 50cm Heat Exchanger added in EGR circuit after TEG

• Previous concept EGR temp was ~100K higher than Cummins test data

• Goal was to reduce EGR temperature with hope of increasing BHP

0.0

1.0

2.0

3.0

4.0

5.0

6.0

A-25 A-100 B-62 B-100 C-100

Operating Point

BSF

C Im

prov

emen

t (%

TE material efficiency 9.1% TE material efficiency scaled with packing factor and local temperature

Calculated BSFC Improvement LAST,LASTT-BiTe Materials

Collaborations/InteractionsCollaborations/Interactions

• MSU, JPL, Tellurex, Northwestern, Iowa State and Cummins Team continue to partner in this effort

• Office of Naval Research sponsored effort has provided the basis for new material exploration and assisted in module fabrication developments

• Oak Ridge DOE (High Temperature Materials Laboratory) has provided significant assistance in material property characterization

Publications/PatentPublications/Patent• Pilchak, A.L., Ren, F., Case, E.D., Timm, E.J., and Schock, H.J., (2007). “The

Effect of Milling Time and Grinding Media Upon the Particle Size Distribution for LAST (Lead, Antimony, Silver, Tellurium) Powders,” Philos. Mag. A.

• Ren, F., Case, E.D., Hall, B.D., Timm, E.J., and Schock, H.J., (2007). “Young’s Modulus of N-Type Last Thermoelectric Material Determined from Knoop Indention,” Chemistry of Physics and Materials.

• A. L. Pilchak, F. Ren, E. D. Case, E. J. Timm, H. J. Schock, C. -I. Wu, T. P. Hogan, (October 2007). “Characterization of Dry Milled Powders of LAST (lead-antimony-silver-tellurium) Thermoelectric Material,” Philosophical Magazine, v.87, Issue 29, pp. 4567-4591.

• Ren, F., Case, E.D., Timm, E.J., and Schock, H.J., (2007). “Young's Modulus as a Function of Composition for an N-Type Lead-Antimony-Silver-Telluride (LAST) Thermoelectric Material,” Philosophical Magazine, v.87, Issue 31, pp. 4907-4934.

• Ren, F., Case, E.D., Timm, E.J., and Schock, H.J., (2007). “Hardness as a Function of Composition for N-Type Last Thermoelectric Material,” Applied Physics Letters.

• Ren, F., Case, E.D., Timm, E.J., Jacobs, M.D., Schock, H.J., (2007). “Weibull Analysis of the Biaxial Fracture Strength of a Cast P-Type Last-T Thermoelectric Material,” Philosophical Magazine Letters.

• Case, E.J., Ren, F., Schock, H.J., and Timm, E.J., (2006). “Weibull Analysis of Biaxial Fracture Strength of Cast P-Type LAST-T Thermoelectric Material,” Philos. Mag. Lett. 86 [10]: pp. 673-682.

Major Plans for Remainder of Phase 2 EffortMajor Plans for Remainder of Phase 2 Effort

July ~ Aug, 2008: 2,4,8 and 16 leg modules being fabricated, segmented concepts evaluated and powder processing method development ongoingAug ~ Nov, 2008: Module construction and performance testing Sept ~ Nov, 2008: Advanced heat exchanger design and numerical simulation of expected system performanceDec, 2008: Choose a thermoelectric system to demonstrate a 500 watt generatorJanuary 2009: Complete 500 watt generator designFeb-June 2009: Construct the 500 watt generatorJuly-Nov. 2009: Generator testingDecember 2009: Complete report on generator performance

SummarySummary• Systems for material synthesis, powder processing, hot

pressing, leg and module fabrication are operational at MSU

• Performance testing of legs and modules at MSU is in agreement with others doing similar measurements

• Can produce materials required for a 500 watt module in one month

• Power conditioning electronics for being designed and tested

• Improved head exchanger designs are critical to success of TE effort for waste heat recovery

• Using TEG technology, a 5% improvement in bsfc for and OTR truck is a reasonable 5 year goal …10% improvement possible with new TE materials