This problem was discussed here https://www.researchgate.net/post/Which_is_the_best_method_of_sample_preparation_for_Magnesium_alloy_EBSD_Mechanical_polishing_or_electro-polishing

Hello Gert Nozle.

Pre-cleaning, activation surface of Al and aluminum parts can be carried out using method surface activation by cathode spots in vacuum. This method makes it possible to obtain a clean juvenile surface.

Yours faithfully

Konstantin

Personally, I have never used electrolytic polishing for EBSD as it can attack and remove second phase particles and also produce a non-flat surface which is a disaster for EBSD. I am attaching a large pdf file listing many aspects of specimen prep to get good EBSD results on many metals and alloys, including high purity Al. I hope it reaches you as it is ~38 MBs as a pdf.

Dear George, I did not use it as well for many years, but as you know, any sample prepration may affect the surface, i.e., especially any kind of mechanically based preparation. The major problem is, whether you can recognize artefacts or not, and here the currently limited precision of EBSD hides a lot of them. In other words, the higher the precision of EBSD will improve in future the more you will see, how "terrible" you sample preparation can be, especially in case of any mechanically based technique, even vibration polishing. Chemical or electrochemical polishing has the huge advantage that during preparation you do not touch your surface, however, as you argued the attack might be not that homogeneous, especially for multiphase materials. But this is also the case for any other material since each phase has a slightly different resistence against a specifically used agent.

Sample preparation is a widely underestimated factor for EBSD measuerements, and the often used argument that a high hit rate is a sufficient indication for a good preparation is simply a terrible misbelief.

Sample prep is like cooking. You did this for years always with the recipe of your mother, but accidentally it turns out that the taste can become better if you change something. However, the problem with EBSD (and also cooking) is, what does it mean: better? Since nowbody exactly knows the uneffected microstructure (therefore we investigate it) who can tell that the produced surface quality is sufficient, especially with regard to minor orientation variations recently used for the characterisation of deformations? If the error is not obvious we simply decide that the preparation is satisfying (never perfect...this we also pretend to know as final excuse).

Thanks for the PDF. There are many slides which are worth to discuss in more detail, especially the KAM images which display a lot of orientation artefacts not only coming from sample preparation.

Gert:

Go ahead and do chemical attack polishing or electrolytic polishing. I do not do that. For numerous years I worked with EDAX and Oxford Instruments giving lectures, like what I sent you, to start seminars on EBSD. You saw my results. They have been excellent, unlike many efforts by other. If you do not want to try my methods, that is your option. But, many people have used them with great results. George

When results can be called excellent or great? When your maps looks nice?

Mechanically based sample prep has a big probability to insert artifacts especially when we consider the information depth of some tens of nanometers only. You are one of the biggest expert on this field but I do not understand how you can per-se exclude another technique which have of course drawbacks as well. Their advantage is at least that they are visible whereas your preferred mechanical prep hides or even produce them. I think all techniques including ion etching or thermal etching have their advantages and one should try to find this technique which prevents microstructure as less as possible. Beautiful looking maps do not say anything...

Gert: I answered your question based on about 8 years of intense activity with EBSD with the two major companies, EDAX and Oxford Instruments, at that time period (2002 to 2009), plus experts at a number of universities here and abroad plus at Max Planck Institute. When I started working with EDAX and Oxford, both told me that they would prepare specimens mechanically for customers, send them the results, the customers would buy their systems, they would install the equipment and train the customers. Next thing the customers told them was identical nearly every time: "Your equipment is no good, I cannot get any patterns!" It was not the equipment's fault, the problem was the inability of the customer to prepare the samples properly. I have been doing color etching of specimens for more than 40 years (I started this to selectively color phases or constituents for image analysis measurements) and that requires removal of all damage introduced in sectioning the specimen and grinding and polishing. My results are superb - which means there is very little residual damage left. I apply the same techniques when doing EBSD. The results are great as verified by all of the EBSD experts, but not by you. That is your problem. I told you, if you want to do chemical attack polishing (which is not harmful if you know what you are doing), or electropolishing, go ahead. Personally, I do not like to do EP. Look at the literature on these methods. Many times, the authors tell you to grind the surfaces to 600 grit SiC paper (but sometimes they say to stop after using a much coarser grit paper) and then EP for 1 minute. But, the removal rate with most EP solutions is only around a micrometer per minute, but the residual damage left after grinding many different metals and alloys is in the neighborhood to 8 to 12 micrometers! That is inadequate. If I were to use EP, I would fully prepare my specimens with the methods I sent you and then maybe do a 1 minute EP, just to remove any minor residual damage left. But, for most metals and alloys, I do not need to do that as my EBSD results are perfect! A vibratory polish after the standard method always improves the results, and is absolutely necessary for specimens such as lead and lead alloys. But, you do not accept that as proof that mechanical polishing is a valid technique! Well, you have a serious problem. My specimens exhibit no pull out of second-phase particles, no residual damage of any significance and no loss of flatness. If I color etch them, the microstructures are perfect as proven countless times. But, you cannot accept that as proof that mechanical polishing is a valid procedure for preparing specimens for EBSD! Well, you are a member of a very small minority in your belief! Yes, if people perform mechanical polishing poorly, they can damage the surface - but I can say that just as well about electropolishing! No matter how good a method is, people can do that method poorly when they do not know what they are doing. George

Dear George, I did the same for more than 15 year and I would agree if I would not have found some experimental observations recently. Austenitic steel 304L is one candidate where even a surface cleaning with Cs transformed a lot of austenite. For this material we knew before that it should be austenite, but what would happened if we did not know it (as usual for unknown material)? With 304L EP worked fantastic, or better, it looked more like a microstructure we expected for this material (e.g. from colour etching). Ion polishing also worked but required hours of polishing which is economically not acceptable if other techniques are available. My message is only that EP can produce better results if it works for the specific material. Maybe that there are better techniques for the here discussed Al alloy. I realized that in literature EP has been applied and as you can see from recommendations above, there are some people in honourable universities and institutes who are using EP. If one observes remarkable artifact the question arises which technique generates bigger or more critical for the investigated property? For soft materials I would always test different techniques as long as I do not know the result already since only the general aquivistion of EBSD patterns as required from you from Oxford and EDAX is only a proof that they can index patterns where before no pattern was available. To make this clear, if you generate by color etching or another technique an oxide which produces patterns it is no proof for a correct description of the mictostructure since the phase is wrong. Maybe the orientations are even correct, but certain effects like lokal misorientations are at least questionable since because of the lattice mismatch misorientations will appear. For a mechanical polishing I only want warn that if you look with higher orientation precision (what is not available at the moment using standard indexing, you don't see these effects. In your KAM maps different grains are higher deformed than others, but I would bet a wager that these are artefacts from the indexing. I observed the same but looking with a more advanced processing they all disappered. And, the displayed misorientation variation is in practice about ten time smaller...and subgrain boundaries appear in the scale of 0.1 -0.2° appear. Additionally you will see many remaining deformation zones around scratches you do not see anymore in your images as long as you do not have large grains an using orientation contrast imaging (FSE) which is much more sensitive (comparable to the applied pattern matching) than speed optimized indexing.

Only as final summary: There are no doubts that your contributions are amazing and they showed many people how powerful standard polishing can be. However, if people start to look for tiny orientation variations like GNDs they should be aware that that what they are measuring might be more affected by their preparation than from really inherent deformation in materials. Last investigations (thanks to pattern matching) experimentally proved this for several materials, especially for small grains where FSE does not help that much anymore. It is hard to see...and therefore it is hard to evaluate whether a polishing is "sufficient".

Best greetings, Gert

Gert: You mentioned that cleaning with Cs transformed a lot of austenite. What is "Cs"? I have never seen any cleaning process transform austenite in a stainless steel and I worked heavily with stainless steels for many years, especially when I worked for Carpenter Technology Corp and supervised their metallography labs in the R&D Center. George

Cs sputtering before TOF-SIMS investigations in SEM, see page 17 or 18 in

Presentation Preparation artefacts: When do we know that EBSD results are reliable?

The topic of our German annual EBSD meetings was in June "Sample prep". There were several contributions about 304L and phase transformation during polishing. I mean, the problem is not that you possibly find a preparation technique which prevents artefacts. The problem is that you do not always know the material and how it reacts. And as you can see from page 10 (mechanical and vibration polishing) there are no visible artefacts (except from a scratch during vibration polishing...). My fear is that if one does not know a material it is hard to predict when a sample preparation is "perfect"...since the collection of 100% solutions it is obviously not.

Let me refer asl well to

Presentation EBSD sampling, preparation and area selection

Gert

It is one of the best discussion about the EBSD sample prep that I'm honored to read and learn a lot from you !

Gert Nolze did you have any success EP the Al alloy? I too am trying to EP a 2000 series Al alloy for EBSD. I have done up to 1 um diamond polish for 15-20 minutes. Then 1 hour colloidal silica. I did not get good EBSD (0.01 um step size, 2x2 speed) on a FEG SEM for a microstructure of 6 um grain size and many coarse precipitates. I tried 3 hours and still did not get a good scan. I am trying EP now. I have done 15 seconds of polishing, no etching, 15V, -28C, and a flow rate of 9 (Struers LectroPol-5) using 30/70% nitric acid/methanol. It seems as if I am getting pitting. Please advise.

Hi, no...I am not doing EP by myself. But you can contact Grzegorz Cios. He is doing EP for us in case we need it. On the other hand...I think you are doing silica polishing too long.

If you take a simple Al foil (for kitchen) you will already get without any preparation nice patterns. The problem with alloys is practically that there are second phases precipitaded. They you wash out and then they generate damages again like grinding paper again. This is the reason why EP might be a good idea. If you do not have it, use short time silica and prolong it. We usually never do longer than 20min. By hand! This is certainly not the best way if you have many samples but we commonly don't.

Hi we routinely do EBSD of Aerospace Alloys and our final mechanical polish is with a soft silk cloth using 0.25micron. We then always ion beam mill usually at 8kV for 15min at 4degrees followed by 2kV at 5degrees for another 8min. Finally 400V at 5degrees for 5minutes. This consistently produces good results on the bulk Al and intermetallics.