In fact you cannot get this. During sample praparation you always affect the surface-near layer which is responsible for the signal. As written above, if the lattice is hardly desctroyed you don't get any signal. But it is clear that between nothing and ideal a wide range exist which reflects the influence of sample preparation. You can only try to reduce the impact of any mechanical (meachanical polishin, vibro-polishing) or high-energetic damaging (ion polishing). The only technique I presently know is chemical "etching", or electropolishing. 

Of course, as long as you can still see linear crossing orientation variations it is highly possible that they are residual deformations from former scratches. Often they are the remaining parts of grinding and cutting, i.e., you should always very gently cut and grind. Use proper cooling and don't cut too fast. The same with grinding. It is often not that bad to grind with several papers of decreasing grain size (higher numbers). And the final polishing with OPS or silica. They are differently working although based on silica. A good way is to make a test with the same material and you increase stepwise the polishing time. If you scan the sample you should see how your maps changes. If no change is visible anymore you should be close to the optimal surface (you can reach with this technique). The last clause I only give to point out again, that especially for soft materials like Ag, Au, Bi etc. you will never be able to get the real bulk conoditions since even a careful preparation will  generate a new deformation layer which can reach several µm in depth. We did tests with such samples in FIB to see the deformation layer. Any othe material is simply again something in between.

Therefore, any calculations focused on intragranular misorientations are always a bit speculative since the effect observed can be the result of your insufficient preparation. However, the best way to have a check is: post a map and we will have the opportunity to give you some advice. There is a wide range of interpretation. The image quality, or band contrast map is often sufficient, but an additiional misorientation map would help.

@ s, Choudhury: backscattered electrons will always reach the detector. The difference is that the Kikuchi pattern as small intensity modulation on a big background disappears because of the missing diffraction caused by an incrystallinity. I also don't get what do you mean with uneven surfaces? You don't need flat surfaces per se. You can measure also on quite rough surfaces as long as you reach the surface with your electron beam, and the signal has free access to the EBSD screen. A martensitic transformation of a formerly flat (polished) surface generates a lot of topography which is not that high so that you cannot map it.

By 'Bad Sample preparation' are you referring to scratches and uneven surfaces? If there are scratches and assuming they are created during grinding, polishing steps. You should be able to distinguish. The scratches marks should not give you any data (black pixel) as the backscattered electron would not rich the detector and there will be no pattern.

Hopefully it helps.

-SFC

In fact you cannot get this. During sample praparation you always affect the surface-near layer which is responsible for the signal. As written above, if the lattice is hardly desctroyed you don't get any signal. But it is clear that between nothing and ideal a wide range exist which reflects the influence of sample preparation. You can only try to reduce the impact of any mechanical (meachanical polishin, vibro-polishing) or high-energetic damaging (ion polishing). The only technique I presently know is chemical "etching", or electropolishing. 

Of course, as long as you can still see linear crossing orientation variations it is highly possible that they are residual deformations from former scratches. Often they are the remaining parts of grinding and cutting, i.e., you should always very gently cut and grind. Use proper cooling and don't cut too fast. The same with grinding. It is often not that bad to grind with several papers of decreasing grain size (higher numbers). And the final polishing with OPS or silica. They are differently working although based on silica. A good way is to make a test with the same material and you increase stepwise the polishing time. If you scan the sample you should see how your maps changes. If no change is visible anymore you should be close to the optimal surface (you can reach with this technique). The last clause I only give to point out again, that especially for soft materials like Ag, Au, Bi etc. you will never be able to get the real bulk conoditions since even a careful preparation will  generate a new deformation layer which can reach several µm in depth. We did tests with such samples in FIB to see the deformation layer. Any othe material is simply again something in between.

Therefore, any calculations focused on intragranular misorientations are always a bit speculative since the effect observed can be the result of your insufficient preparation. However, the best way to have a check is: post a map and we will have the opportunity to give you some advice. There is a wide range of interpretation. The image quality, or band contrast map is often sufficient, but an additiional misorientation map would help.

@ s, Choudhury: backscattered electrons will always reach the detector. The difference is that the Kikuchi pattern as small intensity modulation on a big background disappears because of the missing diffraction caused by an incrystallinity. I also don't get what do you mean with uneven surfaces? You don't need flat surfaces per se. You can measure also on quite rough surfaces as long as you reach the surface with your electron beam, and the signal has free access to the EBSD screen. A martensitic transformation of a formerly flat (polished) surface generates a lot of topography which is not that high so that you cannot map it.

Thank You Soud for your answer. But as mentioned by Dr. Nolze, my concern is not really corresponding to the extremes and it is rather about the wide range in between "nothing and ideal".

Thanks a lot Dr. Nolze for your comments. I agree with your suggestions about cutting slowly, grinding with a number of papers having decreasing grain size and giving a greater amount of time on cloth polishing. These are the fundamentals which we should never overlook. But I still have a few more queries. These are summarized as follows:

1. As mentioned by you, Electropolishing is a very good technique for reducing the impact of any mechanical damage caused by previous sample preparation steps of cutting/grinding. So if we are using Electropolishing as a final step, there should not be any damage left at the top surface layer which is mainly responsible for the BSE signals used during EBSD. Moreover, as per my knowledge, Electropolishing removes the top 20-40 micron layer and the damage caused by 600 grit size SiC paper is also around 20 micron deep. So theoretically, electropolishing step should remove the damage caused by SiC paper 600 or greater sized grit (i.e. 800,1000, 1500 and 2000 grit size). And if we are using sample preparation till 2000 grit size followed by alumina cloth polishing before moving to electropolishing, it should have a totally damage free surface. Am I missing something here?

2. You mentioned that the "calculations focused on intergranular misorientations are always a bit speculative". So does it mean that the results that we are obtaining related to grain boundary fractions should also be doubted because these are also essentially calculated using intergranular misorientations?

3. Last but not the least, can we not say that in a well annealed material, there should be a cut-off value for an average value of band-contrast intensity, MAD number, local misorientation and any other corresponding parameter for each EBSD map below which the data that we are capturing is not reliable because all of these parameters will be affected greatly by the damage at the surface?

You are right, I know several people who apply electropolishing directly after grinding. And I would also use a further polishing with alumina or diamond since we need preferably a flat surface. Not absolutely required but is can help to exclude irritations which are perhaps then shifted to the missing polishing.

Regarding intragranular misoriens I ment GNDs. Sorry for mixing "inter" and "intra". I will correct it!

3. This might be possible if you always have the same material, same preparation, same measurement conditions, single phase... since these parameters depend on several factors. I think we have to make compromises. If we assume our prep is OK, we should simply take into account, that this might be incorrect.

We also have to take into account that we don't really know how EBSD patterns are formed and how big the spatial resolution really is. We have models which seem to match, but they often only consider the depth and not the lateral resolution. Also the SEM design is important. A specific positioning does not mean that the beam is fix there. For EBSD it means that the effective beam size is bigger than the real one as function of the magnification. This is for instance the reason that we observe grain boundaries although our beam diameter is perhaps only 10nm wide  but we are scanning with 10µm step size. It would be very inlikely to hit there a grain boundary, but you see them all :-). This has nothing to do with the scattering of electrons beneath the surface, and does not mean that your real beam diameter is bigger.