Hi Ruby Aslam ,

just as food for thought, without mathematical treatment:

At 0 Hz, Z = 100 Ohm, so obviously there must be a resistor or a combination of resistors resulting in 100 Ohm. Since impedance and phase depend on the frequency, there must be Cs or Ls or both, too, but surely the Cs are not in series to R (otherwise Z -> unlimited @ f = 0 Hz), and the Ls are not parallel to the rest of the circuit (otherwise Z = 0 @ f = 0 Hz).

Increasing the frequency up to a certain point results in a reduced Z; looks like a C (C2 in the sketch) parallel to R.

During diminishing of Z, the (negative) phase first increases then decreases; the cause for decreasing must be an L, in series to R and C.

Beyond a certain frequency, both Z and the phase are increasing. The increasing phase looks like a second C (C1 in the attached sketch) coming into play.

The question remains: Why does Z increase, too? Possible answer: XC2 is now so small that R || C2 has a very low impedance. The currents in the circuit are now mainly determined by C1 and L, and these form a band-stop filter with a resonance frequency beyond the limit of the Bode diagramm.

BTW, C2 and L form a band-pass filter, damped by R, and I guess its resonance frequency is somewhere between the first local maximum of the phase and the following local minimum; probably nearer to the minimum. (Without C1, at resonance the phase would change from negative values to positive ones.)

Hope this helps a bit.

Hi Ruby Aslam ,

fiddling around a bit with an online simulator like this one:

https://www.multisim.com/content/i5giQfciZBMpaNyK82mtXZ/bode-plot-circuit/

convinced me that your circuit is in need of two more resistors because the resonance effects were more pronounced than I expected. While R2 is still 100 Ohm, you can attenuate the effect of the resonance with the values of R1 and R3.

Now you have to find the correct point in a 5D space! :-) I would start by determining the product C2 * L on which the resonance frequency near the Z minimum depends.

From my point of view, it is preferable to start from the electrochemical system you are studying to determine an equivalent circuit that is representative of physics. Otherwise the resistances, capacitances, inductances that you will estimate will be difficult to interpret.

From the Nyquist plot it seems simple, but the Bode plot shows a feature at high frequencies that should be studied in detail, as it may be only due to the cell configuration or to the reference electrode.

I agree wih Gaël Maranzana that you show have previus info on the electrochemical system. In fact, many circuits may fit your results, but perhapes only one has a physical meaning.

Anyway, without a datafile it is difficult to help.

Good luck.

Dear Joerg Fricke , Gaël Maranzana, J.C.S. Fernandes

Thank you all so much!!!

My research work is focused on the corrosion inhibition behavior of some organic compounds for mild steel in acidic medium. To evaluate the anti-corrosion properties of synthesized compounds, I have performed EIS measurement using Metrohm Potentiostat/Galvanostat. I am adding the measurement data here, I was operating on the same system under the same experimental conditions recently and a year earlier. It does not appear the system behavior is the same. What factors do affect my measurement? Secondly on what basis I can choose an equivalent circuit to fit the EIS data. Chi-square can be used on the basis of literature to determine the compatibility of the curves. On this basis, my old measurement fits the circuit of the type (R(RQ)). But the recent data fits another circuit model if I consider the low chi-square values. However, showing ‘NO’ convergence with 300 number of iterations. I tried most of the circuit available in Nova 2.1 software. But I am unable to find the appropriate and meaningful circuit for my system. And what is the significance of each added circuit element for the system studied?

Please have a look at the circuits and the plots along with the fit in the below figures. It would be great if anyone can give me some inputs.

Regards,

I have the same problem as well. I have this EIS data and I can't quantify the EIS parameters in Nyquist because it needs electrical circuit. Is there any references that will help me predict and determine the EC? Because most answers I saw were "It depends on the system", which is right but how do I know the proper EC given a particular EIS data? I badly need help thank you

Dear Miguel Mansilungan,

since most answers you saw, about a proper ECM[1], were:

"It depends on the system",

then say, "what is your system" ?

1. See Fig.5, in: Experimental and theoretical studies on mild steel corrosion inhibition by the grieseofulvin in 1M HCl https://www.researchgate.net/publication/315474882_EXPERIMENTAL_AND_THEORETICAL_STUDIES_ON_MILD_STEEL_CORROSION_INHIBITION_BY_THE_GRIESEOFULVIN_IN_1_M_HCl

Dear Ruby Aslam ,

also, "it would be great if" you are willing to show some (extra) exper.-conditions[1], about the EIS[2] (and cell's) differences, apart 'EIS-aging'.

1) The cell(s) is a 2 or 3-electrodes configuration ?

2) The WE-'aging' is in a dry (or in a wet) cell ?

3) The CE was aged, also ?

etc., etc.

1. Storage Temperature, Humidity..., etc.

2. If some (two) ECMs are (both) representative of the WE-'physics', then Chi-square, doubtlessly, can be used to determine the correlation of the ('physics' vs ECM's-fitting) curves.

I have an aqeous solution for detection of analyte. Can I have an idea what is the importance of frequency? What kind of process it symbolize?

Miguel Mansilungan, having two diffewrent questions in the same discussion is a little bit confusing. I would recommend that you submit your questions with the full description of your system in a separate section.

Ruby Aslam, in your last post you mention that your present results are differnt from those of one year ago. As I mentioned before, the behaviour that is present at high frequencies (and that deviates form the R(RQ) circuit that you have used before) may be due to the reference electrode. Some reference electrodes or, alternatively, some salt bridges (eg. the Luggin capilar) may introduce this type of features in the high frequency domain.

Best regards

Joao Salvador Fernandes

Choosing an appropriate equivalent circuit may require a deep understanding of the corrosion processes of your system, the use of complementary characterization techniques such as weight loss method , SEM/EDX, XRD XPS, and support by EC from published similar works

Dear Ioannis Samaras,

The electrochemical measurements were performed using a conventional three-electrode cell assembly with the mild steel (MS) specimen (exposed surface area of 1.0 cm2) as the working electrode, Ag/AgCl electrode (saturated KCl) as reference electrode and Pt wire as the counter electrode. An Autolab Model 128 N Potentiostat/Galvanostat with an inbuilt impedance analyzer FRA2 was used to perform the measurement. To minimize IR drop, a Luggin- Haber capillary with its tip very close to the surface of the working electrode was included in the cell set up. Before performing the EIS measurement, the MS specimen was immersed in the test solution (1M HCl) for 30 min to obtain a steady potential. The steady-state potential was confirmed when no significant variation in rest potential was detected. The measurements were made at 30 ± 1 °C under aerated, non-stirred conditions.

Before the measurement, test coupons were mechanically abraded with different grit SiC papers (#320 to #1200), rinsed with double-distilled water, degreased with acetone and dried in warm air, and then used with no further storage.

Dear Ruby Aslam ,

thanks for giving all these additional experimental details.

I suggest a minor modification of your old[1] ECM (Equivalent Circuit Model) with an intermediate ECM1Coat. class that is used in most cases of (quasi-perfect) organic Coatings and Paints[2]. However, for cases of imperfect (very 'bad') Coatings, the ideal capacitors (of ECM1Coat.) elements should be replaced by (imperfect capacitors) CPE.

So, let me suggest[3] the ECM2bad_Coat.[4], as your next ECM. It will be useful, not only for this case, but for other (similar) case studies, in order to fit better (lower Chi-square), both your old and new, EIS data.

1. Your Quotation: "... old measurement fits the circuit of the type (R(RQ))..."

2. Figure 4. Equivalent Circuit for a Damaged Coating, in: EIS of Organic Coatings and Paints https://www.gamry.com/application-notes/EIS/eis-of-organic-coatings-and-paints/

3. The useful ECM2bad_Coat. (class) is, actually, a generalization of the ECM1Coat (class).

4. Figure 3. The equivalent circuit ... , in: Novel pyrazole derivatives as inhibitors of stainless steel in 2.0M H2SO4 media: Electrochemical Study https://www.researchgate.net/publication/339959385_Novel_pyrazole_derivatives_as_inhibitors_of_stainless_steel_in_20M_H2SO4_media_Electrochemical_Study

Having had a look at the shared data, the main difference is that one goes an order of magnitude higher in frequency, i.e., to 100 kHz and not 10 kHz (all from your plots.docx document). I think you would get a good fit with a simple R(QR) circuit if you limited the fitting software to use data until 10 kHz like in your old data.