T. Sandersa, T. Baumhöfera, D. U. Sauera, A. Schröderb,

K. Wippermannb, J. Mergelb, D. Stoltenb

aElectrochemical Energy Conversion and Storage Systems

Prof. Dr. rer. nat. Dirk Uwe Sauer

Institute for Power Electronics and Electrical Drives (ISEA) RWTH Aachen University

Univ.-Prof. Dr. ir. R. W. De Doncker

bInstitute of Energy Research, IEF-3: Fuel Cells

Forschungszentrum Jülich GmbH

Apparatus for spatially resolved impedance

spectroscopy on DMFC and PEM Fuel Cells

with special regard to fuel concentration

oscillations in gas channels

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 2

Motivation

“Flipping” effect observed in measurements

Caused by measurement method

Not a consequence of MEA stateSource:

Oscillations in Gas Channels II.I. A. Schneider et al.

Journal of The Electrochemical Society

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 3

Outline

The cause for the measurement effect

Influence of various factors

Model based

An advanced measurement setup

“EIScell” measurement hardware

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 4

Commonly used measurement setup

The full cell is excited by one voltage source

Spatially resolved current measurement n

ACn

I

VZ =

Oxygen flo

w

VAC

I1

I2

I3

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 5

Influence of concentration oscillations

Oxygen flo

w

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 6

Influences

Absolute oxygen concentration

Frequency

Limited Diffusion

Excitation amplitude No influence

1

22

1

21

CqFz

RT

dI

dU=

Spectra: 10 Segments, linear channel

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 7

Advanced measurement setup

Single excitation, sequential measurement

Problem: in-plane resistance

Smaller concentration oscillation amplitude

Influence only from neighboring segments

11 +− ++=

nnn

nn

III

VZ

V1 I1

I2

I3

V2

V3

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 8

Comparison

Cell excitation Single segment excitation

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 9

Comparison (Zoom)

Cell excitation Single segment excitation

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 10

EIScell

54 independent channels

± 2.5 V, ± 2.5 A per channel

AC response from all I,V simultaneously

DC measurements

I-V curves

Automated

Integration of test stand control

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 11

Contacting

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 12

Shunt Resistors

Voltage drop across shunt

resistors leads to an equalization

of measured currents

For the given test cell

resistances below 10mΩ would

be needed

Not suitable for measurement of

small currents

ShuntResistors Segments

In-planeResistances

25 mA

25 mA

25 mA25 mA

25 mA

25 mA

VmmAVmeas µ10101 =Ω⋅=

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 13

Shunt Resistors

Voltage drop across shunt

resistors leads to an equalization

of measured currents

For the given test cell

resistances below 10mΩ would

be needed

Not suitable for measurement of

small currents

ShuntResistors Segments

In-planeResistances

50 mA

25 mA

25 mA32 mA

36 mA

32 mA

VmmAVmeas µ10101 =Ω⋅=

Vdiff

Vdiff

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 14

Compensated Shunt Resistor

Measurement signal range can be optimized (variable shunt resistance)

No influence by wire resistance (Kelvin connection)

Parallel measurement up to high frequencies

mVmAVmeas 111 =Ω⋅=

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 15

First Measurement ResultsO

xygen flo

w

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 16

Summary

“Flipping” in spectra is no result from the MEA state, but a measurement influence caused by concentration oscillations.

By using another measurement setup the effect can be reduced.

Magnitude of improvement dependent on relation between in-plane and

segment impedance

A compensated shunt resistor allows for precise spatially resolved current measurement without affecting the current

distribution.

6/23/2009

T. BaumhöferInstitute for Power Electronics and Electrical Drives Slide 17

Thank you for your attention

bh@isea.rwth-aachen.de

T. Sandersa, T. Baumhöfera, D. U. Sauera, A. Schröderb,

K. Wippermannb, J. Mergelb, D. Stoltenb

aElectrochemical Energy Conversion and Storage Systems

Prof. Dr. rer. nat. Dirk Uwe Sauer

Institute for Power Electronics and Electrical Drives (ISEA) RWTH Aachen University

Univ.-Prof. Dr. ir. R. W. De Doncker

bInstitute of Energy Research, IEF-3: Fuel Cells

Forschungszentrum Jülich GmbH

Apparatus for spatially resolved impedance

spectroscopy on DMFC and PEM Fuel Cells

with special regard to fuel concentration

oscillations in gas channels