LEVM - multiple ZARC elements in series

[jkuhn]

Hi,

I'm new to LEVM and would like to fit an equivalent circuit to measured impedance data from a solid oxide fuel cell.  The literature(1,2) often suggests modeling with a series of 4-5 ZARC elements.  None of the LEVM models provide more than 3 DEs in series. 

Is there a way of using existing models to have 5 ZARC(+ 1 warburg in some cases) elements in series?  If so, can someone provide instructions or point me to a section of the LEVM manual? 

Or do I need to recompile a new circuit model using subroutine XSUB? 

Thanks.

1.  Detailed Characterization of Anode-Supported SOFCs by Impedance Spectroscopy, 2007, Journal of the Electrochemical Society 154, 4, pp.B371-B378
2. Electrochemical Analysis of Biogas Fueled Anode Supported SOFC, 2011, ECS Transactions, 35, 1, pp.2961-2968

[Macdonald, James R]

Thanks for your inquiry.  LEVM does not indeed allow more than three DE's in series, although one can add to that several RC composite elements.  So to get more, you would indeed have to complile a new circuit using a Fortran program.

In my experience, however, I find it rare that any set of data even one with 7 to 10 decades of data requires something like 4 or more ZARC elements in series.  In one or more of my papers, I have shown that the ZARC is not actually a very good circuit if one wants to be able to interprete a fit in meaningful physical terms. Usually, a composite circuit using two more physical models (K1, K0, PNP, PNPA etc) in parallel or series-parallel, possibly including a
Davidson-Cole one (DCD,DCC for dielectric or resistivity levels) is usually able to fit the data well.  You can find out more from my website and from all my papers listed there.  Also, you might find that Dr. Barsoukov of Texas Instruments might be able to provide you with more detailed suggestions.

Best wishes,

Ross Macdonald

------------------------------------------------------------------
Dr. J. Ross Macdonald is William R. Kenan Jr. Professor of Physics, Emeritus; Department of Physics and Astronomy of the University of North Carolina at Chapel Hill, and a member of NAS and NAE. Website: http://jross macdonald.com

---

From: impedance-s...@googlegroups.com [impedance-s...@googlegroups.com] on behalf of jkuhn [jnk...@gmail.com]
Sent: Thursday, April 26, 2012 2:28 PM
To: impedance-s...@googlegroups.com
Subject: LEVM - multiple ZARC elements in series

[Yevgen Barsukov]

I quite agree with Prof. Macdonald that if you need to use too many elements

to describe some system, you are dealing with a distributed system. So it is better

described with one element but such that is specially designed for this distributed

system and has parameters that captures the distributed property. For example

you can describe porous material with 4-5 RC elements, but you will lose the information

about conductivity along the thickness of material. Or you can describe it with one

transmission line model which will have the conductivity as a parameter

which can be fitted and gives you additional information about the system.

    You can model some distributed models directly in LEVM, or you can compile your own

models from the literature information (see LEVM manual about compiling additional functions). Alternatively you can use a front-end for LEVM such as MEISP which comes with a library with most common distributed elements (and also allows to combine them in any way you like, or add your own

functions as DLLs without recompiling LEVM).

Regards,
Yevgen