I think you are wrong. To present piezoelectricity, a ceramic sample MUST BE POLED before any measurement, ,otherwise the sample is not piezoelectric. For poled sample, the response is normaly butterfly loop.

Thanks for your prompt reply.

Usually for small signal d33 the sample must be pre-poled. However for large signal d*33 it is not necessary to poled it before measurements.

Here I asking about large signal d*33 in a relaxor ferroelectric.

Your suggestions are highly appreciated.

Thank you

Unpoled ceramics may exhibit a very small d33. Only after poling they exhibit a high d33 value, depending upon the composition.

Polycrystalline ceramics has to be poled, whereas a single crystal with a specific orientation need not be poled.

Now let us say you are measuring strain versus field for the first time, and you are able to estimate some d33 value, well the sample is getting poled in a way during the time you are applying the field to measure your strain.

K. Sreenivas

Dear Professor Sreenivas,

Highly thankful for your nice explanations.

Now in a relaxor materials after removing the applied electric field induced strain returned to the 0% but polarization does not decreased to zero?

Once again thank you very much.

Jam

Since the sample is poled, the domains are already aligned and when you measure the strain hysteresis of such sample, the butterfly curve will show some preferential orientation.

There will always a non-zero large signal strain in a ferroelectric irrespective of the sample history (just even due to electrostriction). In the hypotehtical case where there would not be back switching strain would be much smaller in a second cycle compared as to a first cycle. If you are in the large signal regime (as I think this question goes) poling or not is not a critical paramter since your sample is repoled after ecah cycle (you are above the Ec generally). Although of course your response will change if ones compared the virgin cycle with others. Also if you texturing increases with cycles your response will also change after each cycle and after some cycles you will get a stable state. For the small signal, indeed as Dr. Marchet said in order to obtain a macroscopic piezoelectric response your sample has to be poled so that you break your oveal isotropy and you have a resultant macroscopic spontaneous polarization. Note, however, that your sample is still ferroelectric/piezoelectric but you can just not measure it macroscopically. If you would measure only one grain with a PFM for isntance the piezoelectric response of your unpoled grain would be non-zero.

Dear Matias 

Thanks for your valuable comments. 

How about this ?

In a relaxor materials after removing the applied electric field induced strain returned to the 0% but polarization does not decreased to zero?

Thank 

Jam

Relaxor Ferroelectrics exhibit a gradual change from purely electrostrictive behaviour to the conventional butterfly type strain-field loop on cooling. Simultaneously, the P-E loop evolves from a very slim form to the typical strongly hysteretic shape of a normal ferroelectric. Therefore, the answer to your question depends on temperature. If you are observing electrostrictive behaviour with zero or little hysteresis then it is likely that the remanent polarisation is close to zero.

Another more subtle point is how you define the reference point for zero strain. In a conventional ferroelectric, the remanent strain values for the positively and negatively polarised states are the same but they are different from the unpoled state by an amount usually referred to as the poling strain.

Prof. Hall made an excellent point about having to define zero strain. Furthermore, I'd like to add that the remnant polarization in typical relaxor ceramics is close to zero and material behaviour is also highly dependent on temperature. So yeah the strain response is generally a function of temperature.

 Besides, if you are considering response from a high frequency bi-polar electric field set-up (high-field response) then the value of remnant polarization is negligibly small, so the sample is essentially considered to remain unpoled.