Advanced Measurement and Modeling Tools for Improved SOFC Cathodes

Yunxiang Lu, Jamie R. Wilson, Lilya Dunyushkina, Stuart B. AdlerUniversity of Washington

SECA Core Technology ProgramOctober, 2003

Acknowledgements• University of Washington Department of Chemical Engineering

– Yunxiang Lu (PhD student)– Jamie Wilson (PhD student)– Lilya Dunyushkina (IHTE)

• Collaborators– Dan Schwartz (UW Chemical Engineering)– Olga Marina, Peter Rieke (PNNL)– Allan Jacobson (University of Houston)

• Support– DOE/NETL SECA Core Technology Program– NSF, UW Provost Fund

Motivation: How do we identify and improve physical processes limiting electrode performance?

The Problem

Current electrochemical techniques are limited in the information they can provide.

– Difficulty isolating the electrode in a meaningful environment.– Limitations of impedance spectroscopy (i.e. “interpreting blobs”).– Lack of models linking performance to properties, microstructure.– Inability to implement broadly in a development environment.

Our Approach

• Microelectrodes for improved cathode measurements

– Better resolution and isolation than standard half-cells.– Allows testing under more realistic conditions.– Higher experimental throughput via. miniaturization.

• Analysis of nonlinear harmonics (NLEIS, EFM)

– Helps identify physical processes via. nonlinearity.– Broader spectrum of information without more experiments.

• Mechanistic modeling using finite element analysis

– Quantitative evaluation of proposed mechanisms (based on data). – Includes nonlinear and 3-D effects. – Links performance to properties and microstructure.

Materials of Interest

Porous Perovskite Electrodes:

electrolyteCe0.8Sm0.2O2-x

electrodeLa1-xSrx(Co,Fe)O3-δ

1µm

Thin-Film and patterned electrodes:

La0.5Sr0.5CoO3-don single-crystal YSZ

Purpose of a half-cell measurement

Typical Experimental Goals:

• Measure voltage loss associated with a particular electrode in an operating fuel cell.

• Test an electrode under a current load in a particular environment.

• Isolate the electrode frequency response (impedance).

Reference Electrodes are Prone to Error on Thin Cells

FEA calculation Measured Error

Error Includes Distortion of Impedance Spectra

• “Cross-contamination” of WE and CE response.

• Distortion even with perfect electrode alignment.

• Impossible to avoid with thin cells having 1-D geometry.

Principle of a 2-D “strip” microelectrode

Potential benefits:

• Excellent isolation of WE.• Representative materials.• Ability to test at high bias.• Reduced IR and inductive effects.• Multiplicity through miniaturization.

Phase 1 Cell Design and Implementation

SDC

MgO/spinel

WE RE

Search for a Better Mask Material

Problems with MgO/spinel:

• Voids allow some electrode contact.• Must be >50 micron thick.• Must align multiple layers.• Too course to pattern well.

electrodematerial

MgO/spinel(“boulders”)

SDC

220,000c-MgO(2 layers)

200,000c-MgO(1 layer)

25,000MgO/spinel(3 layers)

3000MgO/spinel(1 layer)

Resistance(Ω-cm2)

Maskmaterial

Current solution: colloidal MgO powder.

• 10 times better insulation with one layer.• Only 10 microns thick.• Results more repeatable. • Less reactive with SDC.

Phase 2: Sputtered MgO films

Preliminary results with LSCO on SDC.

• Microelectrodes clearly reveal differences between cathodic and anodic polarization.

• Impedance exhibit clean and repeatable ohmic intercepts, indicating good frequency isolation and reduced effect of lead induction.

• Technique not quantitative with MgO/spinel mask.

Impedance of LSCo electrode at 750C in Air

-5.00-4.00-3.00-2.00-1.000.00

0 2 4 6 8 10 12 14 16 18 20

Z'

Open Circuit PotentialCathodic polarization 100mVAnodic polarization 100mV

120Hz

1KHz

150kHz

LSCo Electrode Polarization Curve

-0.05000

0.05000

-0.30000

-0.20000

-0.10000

0.00000

0.10000

0.20000

0.30000

Voltage (volt)

Half Cell Cathodic polarization

Half Cell Anodic polarization

Verification of Microelectrode Overpotentials

Procedure:1) Measure i-V characteristics of

microelectrode.2) Subtract iR losses (based on

impedance).3) Add anode and cathode OP.4) Compare to V(i)-iR for a full-

sized symmetric cell having the same electrodes.

(air)

Conclusions:

• Sum of anode & cathode OP’s yield values close to a symmetric cell.

• Anode and Cathode OP’s have behavior consistent with literature observations.

Applications and Future Work

• Incorporate Microelectrodes into our perovskite electrode studies.

- Mechanistic studies employing controlled microstructure. - Development of tailored microstructures for improved cathode performance.- Co-development of NLEIS and EFM.

• Sputtered thin-film mask

- Thin mask = smaller dimensions.- Reduced risk of geometric effects.- Thinner electrolytes

• Develop methods to test multiple cells on a single substrate.

- Materials Screening- Massively Parallel Testing- Design of Experiments- Long-term Degradation

How Electrochemical Impedance Spectroscopy (EIS) works

Plus: EIS facilitates separation of complex overlapping physics via. time scale.

Minus: multiple models tend to predict similar (or identical) linearized response.

What are harmonics, and why measure them?

• All nonlinear systems generate responses at multiples of the excitation frequency (harmonics).

• The magnitude, sign, and phase of the harmonics are highly sensitive to the details of the underlying physics.

Types on nonlinear harmonic measurements

Our apparatus for NLEIS and EFM measurements

How NLEIS Data is Acquired and Processed(example: LSCO on SDC at 750°C in air, 10 Hz)

V(t)

(V)

0.6

-0.6

-0.4

-0.2

0.0

0.2

0.4

1.00.0 0.2 0.4 0.6 0.8

Plot 0

0.6

-0.6

-0.4

-0.2

0.0

0.2

0.4

2.50.0 0.5 1.0 1.5 2.0

Plot 0

V(t)

(V)

Time (sec)

0.2

-0.2

-0.1

0.0

0.1

1.00.0 0.2 0.4 0.6 0.8

Plot 0

Plot 0

I(t) (

A)

Time (sec)

Raw Data

Apodized Data0.2

-0.2

-0.1

0.0

0.1

2.50.0 0.5 1.0 1.5 2.0

I(t) (

A)

Time (sec)Time (sec)

9E-4

-9E-4

-5E-4

-3E-4

0E+0

3E-4

5E-4

1000 20 40 60 80

observed d

fitted data

9E-4

-9E-4

-5E-4

0E+0

5E-4

1000 20 40 60 80

observed d

fitted data

9E-4

-9E-4

-5E-4

-3E-4

0E+0

3E-4

5E-4

1000 20 40 60 80

observed data

fitted data

9E-4

-9E-4

-5E-4

-3E-4

0E+0

3E-4

5E-4

1000 20 40 60 80

observed data

fitted data

Imag

inar

y V

olta

geR

eal V

olta

ge

Rea

l Cur

rent

Imag

inar

y C

urre

nt

Frequency (Hz) Frequency (Hz)

Frequency (Hz)

How NLEIS Data is Acquired and ProcessedComplex Fourier Transformed Data

Current Voltage

Frequency (Hz)

1E-4

-2E-4

-2E-4

-1E-4

-5E-5

-1E-20

5E-5

6035 40 45 50 55

observed data

fitted data

2E-2

-2E-2

-1E-2

0E+0

1E-2

600 10 20 30 40 50

observed data

fitted data

V1real

V4mag

V5mag

Real part of ComplexFourier Transformed Voltage

5E-4

-1E-3

-8E-4

-5E-4

-3E-4

0E+0

3E-4

4015 20 25 30 35

observed data

fitted data

V2imag

V3imag

Gaussian Fit to Determine Magnitude and Phase

Frequency (Hz)

Frequency (Hz)

Harmonic Response of LSF-82 on SDC at 750°C(symmetric cell)

3rd Harmonic 5th Harmonic

0.6

0.4

0.2

0.0

-0.2

Re(

V1)

, Im

(V1)

(vol

ts

100 101 102 103 104

ω (Hz)

-0.2

-0.1

0.0

-Im

(V1)

(vol

ts)

0.60.50.40.30.2Re(V1) (volts)

-30x10-3

-20

-10

0

10

Re(

V3)

(vol

ts)

100 101 102 103 104

ω (Hz)20x10-3

10

0

-10

Im(V

3) (v

olts

)

100 101 102 103 104

ω (Hz)

8x10-3

6420

-2Re(

V5)

(vol

ts)

100 101 102 103 104

ω (Hz)15x10-3

10

5

0

-5Im(V

5) (v

olts

)100 101 102 103 104

ω (Hz)

100%O2

air air

100%O2

1% O2

1% O2

1% O2

100%O2

air

!st Harmonic(Impedance)

Possible sources of nonlinearity in a perovskite cathode.

elect roly t e

1 / 2O2 + 2e- O2 -

O2-

O2-

V

mixed-conduct ing f ilm

Dense Thin FilmMixed-Conduct ing Elect rode

∂xv

∂t= Dv

RT ∇ ⋅ − 1

2∂ ln PO2

solid

∂ ln xv

∇xv

n ⋅Nv y=L= k

(PO2

gas)n

(PO2

solid )m − (PO2

solid )n−m

i = 2F (n ⋅Nv y= 0) = i± (e

α1Fη int

RT − e−α2 Fη int

RT ) + C∂η int

∂t

Can be solved in 1-D, 2-D, or 3-Dusing Finite Element Analysis (FEA)

Numerical FEA Modeling of Porous LSCO on SDC

60x10-3

40200-I

m(Z

)

0.280.220.16Re(Z)

-20x10-3-10

01020

Re(

V3)

, Im

(V3)

100 102 104

ω (Hz)8x10-3

4

0

-4

Re(

V5)

, Im

(V5)

100 102 104

ω (Hz)

-0.3-0.2-0.10.0-I

m(V

1)

0.60.40.20.0Re(V1)

80x10-3

40

0

-40-R

e(V

3), -

Im(V

3)

0.1 1 10 100ω

10x10-3

0

-10

Re(

V5)

, Im

(V5)

0.1 1 10 100ω

-0.3-0.2-0.10.0-I

m(V

1)

0.60.40.20.0Re(V1)

20x10-3100

-10

-Re(

V3)

, -Im

(V3)

0.1 10 1000ω

2x10-3

10

-1R

e(V

5), I

m(V

5)0.1 10 1000

ω

LSC/SDC/LSC750°C in Air

co-limited by:O2 dissociation, bulk transport

co-limited by:O2 adsorption, bulk transport

Applications and Future Work

• NLEIS measurements on half-cell LSF/SDC microcathodes.– 2nd and 4th harmonics.– Higher S/N and reduced inductive effects at high frequency.

• Measurements and modeling of patterned electrodes (collaborativewith PNNL and U Houston).

• Application to developing new materials and microstructures for cathodes (joint with MSE at UW).

• NLEIS as an alternative method for characterization and diagnosis of cells/stacks, with no additional experimental effort/time.– Look at materials in industrial development.– Build a database of characteristic behavior.

Summary

• Microelectrodes potentially offer a more reliable vehicle for both measurement and development testing.

• It may be possible to use microelectrodes for for design-of-experiments or other combinatorial applications.

• NLEIS appears to be a promising method for both electrode analysis and characterization/diagnosis of cells/stacks, with no additional experimental effort/time.

• These tools will be used in subsequent work to develop a deeper understanding of electrode mechanisms, and new materials and microstructures for cathodes.