Dear Venkat,

the quick answer to your question is NO. The frequency 10^5 Hz at which you see " dielectric loss" can not be taken as some demarcation line between the dielectric loss and Maxwell-Wagner (MW) "relaxation".  The both effects can overlap in the frequency ranges, usually measured.

I have couple of general comments in regards MW "relaxation" though :

1. MW effect in most cases that I have looked at , is caused by the spatial inhomogeneity of dc conductivity profile and HAS NOTHING TO DO with possible spatial changes of the dielectric constant. As soon as the electrical current in the sample has to flow through two spatially separated regions, with "appreciably" different dc conductivities (but identical dielectric constants !!!), you get "appreciable" MW type effect("relaxation").

2.  There is an easy way to demonstrate this. The electrical response in all condensed systems can be modelled (in first approximation) as series connection of two parallel R,C terms. The first term describes response of the thin region near the electrode-sample interface, let us assume that this thin layer is about 0.01 of the sample thickness. The second term then describes the response of the sample bulk. The respective resistances are dc resistances of the two regions and the respective capacitances are  the corresponding GEOMETRICAL capacitances having the SAME DIELECTRIC CONSTANT. In this geometry, the interface resistance will be 0.1 of the sample bulk resistance and the interface capacitance will be correspondingly 10 times the sample bulk capacitance. There will be no MW effect(relaxation), the two R,C terms can be lumped together exactly into a single sample bulk term with the full thickness.

Change now the interface  resistance by some factor (let us say factor ten). The impedance of the system will now show MW effect (low frequency response - interface R,C term and high frequency term - sample bulk R,C term), all this with NO DIFFERENCE IN DIELECTRIC CONSTANT in the two regions (for example the same piece of Silicon). In the real part of the complex capacitance data you will now see two capacitances, the larger one is the geometrical capacitance of the interface region and the smaller one is the sample bulk geometrical capacitance. But the reason that you see the interface capacitance IS NOT DUE TO DIFFERENT DIELECTRIC CONSTANT, it is purely due to higher dc resisistivity of this layer, when compared with sample bulk resistivity. I have named this type of response (MW type) "mobile charge relaxation" to clearly separate these type of effects from bound charges "dielectric response" that determines the dielectric response function in the materials. Therefore ...

3. The so called "Giant Dielectric Constant" , often discussed in connection with MW effect (as for example in Phys.Rev. 70, 144106 (2004) is a fallacy, originating from fundamental misunderstanding, when plotting the measured data as "complex dielectric constant/function"  (multiplying the measured complex capacitance by thickness of the sample and dividing  it by the sample area).

4. Plotting the measured electrical impedance in spatially inhomogeneous systems as "complex conductivity" and/or "complex dielectric constant" has no physical meaning and interpreting the derived conductivities and/or dielectric constants as relevant physical electrical material parameters is wrong and very misleading !!! 

Finally, the "good" think about the MW effect is that it is related directly to the geometry and the electrical material parameters of the sample and the corresponding electrodes and can be , at least in principle, calculated directly from Maxwell equations (see for example US patent n. 5 627 479, May 7th 1997).  

I hope this will help you on the way

With regards

Petr 

The best thing to do is to use some model to predict MW relaxation time and compare it with the experimental value, as well as the predicted relaxation strength. For this you need to know the geometric distribution and the permtiivities and conductivities of the corresponding components. If there is a reasonable agreement (order of magnitue) you may speak of MW relaxation - otherwise it is better to look for some other mechanism. MW relaxation is frequently invoked without any support using models. If your data are public, please load them up. If you want to have them treated confidentially, you may send it to me to the following email address: [email protected]

Mexwell - Wagner losses, in general, occour at lower frequencies as compared to frequencies at which dipolar type relaxation occurs.This is due to large difference of relaxation times. In Maxwell -Wagner type losses, polarization has a relaxation time is  in milliseconds while for dipolar type polarization it is in microseconds. Loss peak at 0.1MHz, as you asked, is expected from dipolar losses. However, you have to verify that peak position in loss vs frequency curve should shift with temperature in case of dipolar type relaxation.

Dear Venkat,

the quick answer to your question is NO. The frequency 10^5 Hz at which you see " dielectric loss" can not be taken as some demarcation line between the dielectric loss and Maxwell-Wagner (MW) "relaxation".  The both effects can overlap in the frequency ranges, usually measured.

I have couple of general comments in regards MW "relaxation" though :

1. MW effect in most cases that I have looked at , is caused by the spatial inhomogeneity of dc conductivity profile and HAS NOTHING TO DO with possible spatial changes of the dielectric constant. As soon as the electrical current in the sample has to flow through two spatially separated regions, with "appreciably" different dc conductivities (but identical dielectric constants !!!), you get "appreciable" MW type effect("relaxation").

2.  There is an easy way to demonstrate this. The electrical response in all condensed systems can be modelled (in first approximation) as series connection of two parallel R,C terms. The first term describes response of the thin region near the electrode-sample interface, let us assume that this thin layer is about 0.01 of the sample thickness. The second term then describes the response of the sample bulk. The respective resistances are dc resistances of the two regions and the respective capacitances are  the corresponding GEOMETRICAL capacitances having the SAME DIELECTRIC CONSTANT. In this geometry, the interface resistance will be 0.1 of the sample bulk resistance and the interface capacitance will be correspondingly 10 times the sample bulk capacitance. There will be no MW effect(relaxation), the two R,C terms can be lumped together exactly into a single sample bulk term with the full thickness.

Change now the interface  resistance by some factor (let us say factor ten). The impedance of the system will now show MW effect (low frequency response - interface R,C term and high frequency term - sample bulk R,C term), all this with NO DIFFERENCE IN DIELECTRIC CONSTANT in the two regions (for example the same piece of Silicon). In the real part of the complex capacitance data you will now see two capacitances, the larger one is the geometrical capacitance of the interface region and the smaller one is the sample bulk geometrical capacitance. But the reason that you see the interface capacitance IS NOT DUE TO DIFFERENT DIELECTRIC CONSTANT, it is purely due to higher dc resisistivity of this layer, when compared with sample bulk resistivity. I have named this type of response (MW type) "mobile charge relaxation" to clearly separate these type of effects from bound charges "dielectric response" that determines the dielectric response function in the materials. Therefore ...

3. The so called "Giant Dielectric Constant" , often discussed in connection with MW effect (as for example in Phys.Rev. 70, 144106 (2004) is a fallacy, originating from fundamental misunderstanding, when plotting the measured data as "complex dielectric constant/function"  (multiplying the measured complex capacitance by thickness of the sample and dividing  it by the sample area).

4. Plotting the measured electrical impedance in spatially inhomogeneous systems as "complex conductivity" and/or "complex dielectric constant" has no physical meaning and interpreting the derived conductivities and/or dielectric constants as relevant physical electrical material parameters is wrong and very misleading !!! 

Finally, the "good" think about the MW effect is that it is related directly to the geometry and the electrical material parameters of the sample and the corresponding electrodes and can be , at least in principle, calculated directly from Maxwell equations (see for example US patent n. 5 627 479, May 7th 1997).  

I hope this will help you on the way

With regards

Petr 

thanks professor Petr Viscor for spending lot of time to my question. your elaborated answer is useful for my research purpose.