Dear Thomas, thanks your help.

Dear Cao,

generally, the lamella should be as thin as possible :-) But of course there are some trade-offs involving sample preparation technique (if you preparing lamellas in a "flag" or "frontal" way), that you have to take into account to avoid either bending, redeposition, Cu contamination etc. Another factor is a TEM that you are using (for instance 200 kV vs 300 kV), and of course the type of the sample you are preparing (ceramics, metals, ...). In my experience 100 nm metal lamellas are already good for STEM, while way too thick for HR-TEM, therefore if possible I'm trying to use another preparation technique for such analyses.

Another rule of the thumb that can be useful: if it is possible to observe Kikuchi bands, sample is usually too thick for HR-TEM. And since lamella (should have) uniform thickness, it will be challenging to satisfy both conditions. If I understand correctly, you are interested in SAED, so probably just using Kikuchi bands for finding the proper zone axis, right? Try to find some other way to properly align your sample (maybe in thicker way of lamella, where it is attached to the support?), and go for the thin as possible. Also SAED of very thin samples has it's own benefits!

Good luck!

Dear Janez,

Thanks so much for your detailed explanations. Indeed, in my several trails, some lamellas got bent during final thining. The TEM I am using is 200 KV. and samples are the carbides in metallic matrix. Indeed 100 nm seems kind of bit weaker for SAD (zone axis with two beams).Because the interested area is very localized and micronsized scale, Maybe I could try firstly SAD, then repolish the sample even thinner for HRTEM. Another following question i want to ask is that supposing that the sample is thin enough to see Kikuchi bands in transmission EBSD, would it be fine in TEM again? Thanks a lot.

Best.