Do you have data (or some infos) about the temperature (T) changes vs. flow rate[1,2]  (u) ? 

1. Note1 : I presume that there is a local temperature increase with the increase of flow rate (dT/du>0).

2. Note2 : if possible, show the modulus (of the impedance) in the docx file, also.

Hi Ioannis, thank you for your answer. I have updated the modulus data in the word file. Please have a look. Regarding the temperature, I do not have data about the temperature change. But I do not think temperature change is the main cause since I get this results only in concentrated solutions (for instance 10X, 1X PBS around 2 or 0.2 M NaCl solution). If I diluted it, I do not see modulus, impedance or phase angle change.  

Concentrated solutions might create (pitting) corrosion nanosites, acting as additional (mechanical) nano resistors (heat nano sources) to the flow. Also, the semi-plateau in modulus (50Hz0).

are the main factors for the electrical impedance changes[1].

1. Note : Similar electrical impedance changes might be shown, also, with the accumulative time of flow (as an aging time parameter, instead of the flow rate parameter already shown in you case study).

Thank you so much for the answer, loannis. Here, do you think where the main corrosion occurs? On gold electrodes which is metal or PDMS channel or both?

I have also studied the same electrode pairs using su8 instead of PDMS as channel, the results are totally different from these. In the case of SU8 channel, I can not see flow-rate dependent impedance at high concentrated solution. Comparing, I think the corrosion mainly occurs on PDMS channel walls. Is it possible since I see many papers using PDMS as the protection layer to slow down metal corrosion. Could you send me some information about the model of distributed Cd formed by corrosion along the channel flow?  

Both metal[1,3]  and  PDMS[2] channel :  (Q : where the main corrosion occurs?)

Chemically inert platinum[1] is preferred as an electrode material for microfluids (thin Ti-layer between the Pt and PDMS).

1. Patterning of platinum microelectrodes in polymeric microfluidic chip http://www.mstf.wsu.edu/articles/pros_JMMM2.pdf

2. Free-flow isoelectric focusing microfluidic device with glass coating by sol–gel methods   http://www.sciencedirect.com/science/article/pii/S1567173909001229

3. Note :  At low freq. F  a  2nd one semi-plateau starts in modulus (end of F

 Hello Loannis, Thank you!

However, I do not totally agree the idea of impedance change with flow is mainly from metal corrosion even though the modulus results support the model of metal surface modification. It is because: 

1) Gold electrodes are on glass substrate with a Ti seed layer, not on PDMS.

2) I have bonded the same electrodes on glass/gold electrodes on PDMS c hannel and SU8 channels. These flow-rate dependent impedance only occurs on PDMS channel device, not all of them. I mean if the reason is from metal surface modification, it should happen to both systems.

3) I mean around 10 different flow rate in 20 minutes with the flow rate changing randomly, impedance decreases with the increase of flow rate. If the reason is from corrosion, the results should be un-reproducible. But my results are reproducible.

Therefore, I keep thinking PDMS channel is the main reason not metal modification (note 2). Also, I do not think corrosion is the main reason since the results are reproducible (note 3).  But I do not know what kind of PDMS change results in the results.        

How Z-reproducible is it ?

Also, show, please, some device specs (Ti-layer, ... etc.,   or Message, or send : [email protected]), if possible.

Dear Pengfei,

I am not an expert in your system and in fact I even do not know what it is that you measure (remember always to explain in short what it is that you measure), so my will be just some general comments and you must excuse my ignorance of some, apparently well known things within the electrochemistry community. Also, Ioannis has already suggested number of things, so I keep to some general aspects:

1. I am happy that we all agree that the changes you see, are caused by the electrode-system interface. It suprises me that you talk about the corrosion of Au electrode. Can it realy be ? In Solid State Physics it is a real noble metal, completely chemically inert (apart from one particular solution that dissolves it - I forgot the name). The very first thing I would do is precisely what Ioannis suggests, try Pt and see whether changes disappear. The other thing I would do is to do some sutface examination before/after EIS experiments (Electron Scanning Microscopy, Atomic Force Microscopy) to see how the surface really looks like.

3. Looking at your data, it seems to me that the data looks pretty clean, a clear contribution from the bulk (solution itself) and contribution from the interface, where the real part of the impedance decreases with increasing flow. That does not surprises me (well, at least at this point), the redox reactions and the electrical transport/response within the interface region does depend on the concentrations of various electrically active species (and do not have to be electrically charged !!).

4. Imaginary part of the impedance decreases as ~w^-1, something that one expects when there are no distributions of various equivalent R,C,L elements that one uses when modelling the response. In dielectric relaxation literature this refers also to process of "Debye" type, a simple exponential relaxation. Here, it is something a bit different ( not temporal, but spatial variations), but the result is the "same", "non-Debye like" response. It will be very useful to see how the complex capacitance looks like.

5. In conclusion then, my feeling is that rather than the metal corrosion, the flow dependence of the interface impedance (resistance) might be caused by the changes in the solution layer next to the metal-sample interface ( inner, outer Helmholtz layer (plus the diffuse layer; again real part of the complex capacitance would really clarify the issue).

I hope this will also help

With best regards

Petr 

Possible[1] corrosion of Au electrode has a minor effect (the Au//electrolyte effect in Z) in this study, due to the small time scale (say in Hours) of the experiment[1]. However, this minor electrode effect becomes a major one, if it concerns the (mechanical[2], mainly) stability (upon channel flow) of the partial interface Au//PDMS [3].

PS: About  a comment of Petr Viscor's : Quotation "... It suprises me that you talk about the corrosion of Au electrode. Can it realy be ? "    Petr, please, see ref[1], and there are more.

1. Gold dissolution: towards understanding of noble metal corrosion http://pubs.rsc.org/en/content/articlepdf/2013/ra/c3ra42684j

2. In dentists' context or glossary of Dental Health Terms

3. Not about the Au//electrolyte effect :it is a minor effect for Z.

Dear Pengfei and Ioannis,

thank you for references, I will study them carefully. It seems that one learns every day. Very good !!!

Petr