Thanks for sharing this awesome tutorial. It is fantastic with a lot of details, especially the parameters you've been using for the final cleaning steps.
You're welcome, Zelong; yes, it is definitely a matter of debate/preference as to what degree of over/under tilting should be used when final thinning with the I-beam.
Thank you! To be fair, a few year later I feel I can now make a better sample than this one! It might be worth doing another lamella prep video with some updated techniques and tricks. Maybe I can even figure out how to record it via screen capture instead of using crappy iPhone video!
@@NicholasRudawski Sounds great! I was trying to follow your settings on my G4 Helios PFIB system yesterday, but I realized it is challenging to directly copy your recipe on a different system, even though I am using the same HT voltage and beam current. I do have a question: you are trying to cut your sample at 57 and 47 degrees on each side, and you mentioned that you do not want to damage the Pt layer. However, the shape of the lamella is like a "V." Will this cause problems after your lift-out and TEM imaging? Also, I do not use the cleaning protocol; instead, I use a 30kV 6.7 nA ion beam to image the edge for 3 minutes, which also helps me to make the edge smooth. Have you tried this? Thanks for your answer.
@@yaaasg You are correct that over and under tilting by 5 degrees will tend to taper the lamella towards the bottom. This is intentional to help preserve the Pt straps and the near-surface part of the lamella, which tends to be more important, and it also makes it easy to perform the "J" cut. When I do the final thinning, I'm not using tilt angles this large, so the final lamella is much less tapered. I'm not familiar with your technique used to smooth out the lamella edge, but I will as always freely admit that there are so many possible wrinkles, techniques, and approaches that there will never be any large scale agreement on what is the "correct" or "standard" way to make a lamella. Ultimately, it really comes down to the quality of your final lamella and if it is of sufficient quality for you to obtain the information you were intending to obtain. As long as you are getting the results you want to get, you're doing something right (though it may still be possible to further improve upon what you have already done); so, even if you feel you have everything all figured out, it's always best to keep an open mind!
@@yaaasg hey, sorry for the late response. You are correct that using my method of over and under tilting will taper the lamella down towards the bottom. This is deliberate to preserve more of the Pt layer, but also to make it easy to undercut the sample and lift it out later (as there is less material at the bottom of the lamella to now undercut). When I do the final thinning, the amount of over and under tilt is reduced, so the lamella become less tapered during final thinning. If there is one thing I've learned over my years of dual FIB/SEM lamella prep, it's that there is no one standard way to make a lamella and there are countless variations and tweaks that people have adopted over the years; this is why I feel it is so important to learn how others make lamellas, because you can (and often will) learn something new to add to your own lamella prep recipes.
@@NicholasRudawski That is true; previously, I preferred to trim the lamella after mounting it on the grid. One of my systems has a flip stage, which makes it much easier, but it has some alignment issues. After I use 30kV and 24 nA to live image the sample and set the trenches area, and then go to a higher beam current for milling, it causes a shift (the cutting square is no longer at my designated place). We have two systems; the PFIB one does not have this issue, but the Helios has a shift of about 15 µm. I talked with the engineer, but he said it is normal. Have you seen this kind of issue on your system? Currently, I have to do a quick snapshot before the final cutting to confirm the location. Thanks!
what a wonderful tutorial, let me learn a lot However, I would like to ask some questions. My sample is a metal sample, which needs to be observed for dislocation in TEM. After making the sample with fib, there are always curved lines on the surface of the sample under TEM observation, which may be the interference introduced by fib. How can I minimize this interference in fib?
Thank you and glad to hear the video was helpful. Regarding your question, it's tough to determine the origin of the "curved lines" on your FIB-prepared samples without seeing any images If you can send me an email (ngr@ufl.edu) with some SEM and/or S/TEM images showing the issue, I should be able to help you out. Looking forward to hearing from you soon.
Dear Nicholas Rudawski. I would like to watch this video for my TEM sample preparation by FIB. But as you recommended in the start to watch another video of "FIB basic operation", so, could you please share that link so that i can watch it first before coming to this video. I am unable to find that video on youtube. Thanks!
You're welcome; I think the stage is a piezo-driven stage, but I'm actually not sure; the attachment on it that holds the stub and grid is called a universal mounting base.
Yes, that is possible. I have not worked on any FEI Dualbeam for some time, but I think you can do so with the 3 buttons below the milling progress bars (On the right side visible during milling). On Zeiss Dualbeams that is also possible. The distance the beam jumps per click can be set by the operator.
Questions! -why not flag mount the sample to the slide of the post to reduce redep? -why no e-beam C depo so that TEM can see the contrast difference….to see where the depo ends and the Si begins -Are you guys reusing grids? And last why such a large sample ?
Hi MonkeyBrainSushi: I don't like flag mounting because it leaves the sample more susceptible to bending because one end is free, but many people do indeed use this approach. I'm not sure what you mean regarding the "e-beam C depo", but generally, you are trying to avoid beam-induced contamination on your sample. You could reuse the grids, I suppose, but the cost of the grid is exceedingly small compared to the instrument time cost, so there really is no financial reason to do this. The size of this lamella is actually on the smaller side; I know of many people who use a default size that is basically twice as wide as this sample (~20 um wide).
@@NicholasRudawski Hi Nicholas, I think there is one more reason why flag mounting is not preferred. When you rotate the flip stage 180 degrees and try to weld the back, you might not be able to see the joint because the grid blocks the way (if the lamella is not welded in the perfect position).
Nicholas Rudawski , in FIB TEM preparation I typically go to lower KV to minimize amorphous layer and at the final step I use Gatan PIPS polishing machine at 0.5 kV ... it improves samples quality much more in comparison with 2 and even 1 kV ;) your videos are pretty nice !!! We need more !
@@АлександрПриходько-ъ8г hello, Alexander! Could you give a rough guide for 0.5 kV polishing? Particularly, how do you choose the polishing time for different materials?
@@alexandergolubnichy8588 I would say it is quite tricky stuff and I messed up a lot of lamellas. So what I would to recommend. 1) You need to prepare in the TEM grid a special trench and put your lamella there in order to avoid the redeposition from the TEM grid during polishing procedure in the PIPS mashine. 2) You lamella should be thinner at the botom side and iteratively I perform polishing procedure untill the botom side starts to desappear
Good question; I think it may depend on the type of pattern. With something like the cleaning cross-section or regular cross-section, I think rotating the pattern or changing the scan direction may effectively do the same thing. With something like the rectangle, I'm not so sure. The next time I get on one of the dual beams, I'll do some digging and I should be able to answer this for you; please look for another reply soon, thanks.
Hey, so I had a chance to look into this on our dual FIB/SEMs and for all three of the patterns I mentioned (R, RCS, and CCS) rotating the pattern 180 degrees does the exact same thing as reversing the scan direction (which makes sense). Thanks for asking about this and helping me learn something new in the process!
I’m working on a very similar device, but of course with way less experience on my side, so I was just wondering if there might be some advantage with this. Thank for looking into it!
I have one question here, why ion beam image needs to be rotated 180 deg? I notice from AutoTEM manual, the probe is from the left side, down to up position for probe approach, but for almost all manual operation manual, the probe is from right up to down side approach, is there any special consideration? Thanks a lot.
Hi Guo: the 180 degree scan rotation for the I-beam is more of a preference than a requirement; I personally prefer it because it maintains a top-down orientation in the I-beam image (particularly when using the Omniprobe), but I know of many users who prefer to keep the I-beam scan rotation at 0 degrees. So it's really more of a person preference than anything else.
Dear Nicholas Many thanks for these valuable videos... But I have a question: my samples (oxide) are coated with silver (sputter coating), and always when I prepare a TEM sample a big part of the silver layer get re-deposited over the cross-sectional surface of interest (it looks like clouds over what I need to investigate) and l have performed cleaning steps at 5kv and 2kv but it didn't solve the problem. Do you have any advice PLZ?
Hi Samer: glad you find the videos helpful. Re-deposition when you are preparing a sample can definitely be troublesome. How do you know the re-deposition is actually Ag and not something else? The most common culprits for re-deposition I find are the grid material and/or the protective Pt. The grid issue can be rather easily dealt with by using a Mo grid instead of Cu. The Pt issue can sometimes be avoid by thinning past the point where the entire protective cap is gone but often you can't do this if you are trying to protect the near surface region.
@@NicholasRudawski Thanks for your reply. I confirmed it using EDS .... I tried to post a link for TEM image but my comment was deleted every time I post a link to image.
Hi Samer, indeed silver can be a tricky material especially when it comes into contact with air. To be precise with the sulfur in the air. That can create problems if you store the sample some time out of vacuum and then for example plasmaclean it before taking it into the TEM. The plasmacleaning leads to redeposition. When we went directly (about 10 min handling time) from FIB vacuum to TEM (w/o additional plasma cleaning) we had no issues. Or do you see the redeposition already during FIB preparation? Best regards Alex
This shows the FIB is a technological beauty. Thank you for sharing.
You're welcome.
Thanks for sharing this awesome tutorial. It is fantastic with a lot of details, especially the parameters you've been using for the final cleaning steps.
Sorry for the late reply; glad you liked the video!
Hi, that's the very nice technic proceed lamella. Thanks for good & nice knowledge in operate FIB.
Thanks, Peter.
Thanks for the video. The best tilted angle for different ion beam current is sometimes hard to find.
You're welcome, Zelong; yes, it is definitely a matter of debate/preference as to what degree of over/under tilting should be used when final thinning with the I-beam.
Thanks sooooo much for this great video! It helps a lot!
Hi Lulu: thank you, I'm glad you found the video helpful.
Astonishing skills
Thank you! To be fair, a few year later I feel I can now make a better sample than this one! It might be worth doing another lamella prep video with some updated techniques and tricks. Maybe I can even figure out how to record it via screen capture instead of using crappy iPhone video!
@@NicholasRudawski Sounds great! I was trying to follow your settings on my G4 Helios PFIB system yesterday, but I realized it is challenging to directly copy your recipe on a different system, even though I am using the same HT voltage and beam current. I do have a question: you are trying to cut your sample at 57 and 47 degrees on each side, and you mentioned that you do not want to damage the Pt layer. However, the shape of the lamella is like a "V." Will this cause problems after your lift-out and TEM imaging? Also, I do not use the cleaning protocol; instead, I use a 30kV 6.7 nA ion beam to image the edge for 3 minutes, which also helps me to make the edge smooth. Have you tried this? Thanks for your answer.
@@yaaasg You are correct that over and under tilting by 5 degrees will tend to taper the lamella towards the bottom. This is intentional to help preserve the Pt straps and the near-surface part of the lamella, which tends to be more important, and it also makes it easy to perform the "J" cut. When I do the final thinning, I'm not using tilt angles this large, so the final lamella is much less tapered. I'm not familiar with your technique used to smooth out the lamella edge, but I will as always freely admit that there are so many possible wrinkles, techniques, and approaches that there will never be any large scale agreement on what is the "correct" or "standard" way to make a lamella. Ultimately, it really comes down to the quality of your final lamella and if it is of sufficient quality for you to obtain the information you were intending to obtain. As long as you are getting the results you want to get, you're doing something right (though it may still be possible to further improve upon what you have already done); so, even if you feel you have everything all figured out, it's always best to keep an open mind!
@@yaaasg hey, sorry for the late response. You are correct that using my method of over and under tilting will taper the lamella down towards the bottom. This is deliberate to preserve more of the Pt layer, but also to make it easy to undercut the sample and lift it out later (as there is less material at the bottom of the lamella to now undercut). When I do the final thinning, the amount of over and under tilt is reduced, so the lamella become less tapered during final thinning. If there is one thing I've learned over my years of dual FIB/SEM lamella prep, it's that there is no one standard way to make a lamella and there are countless variations and tweaks that people have adopted over the years; this is why I feel it is so important to learn how others make lamellas, because you can (and often will) learn something new to add to your own lamella prep recipes.
@@NicholasRudawski That is true; previously, I preferred to trim the lamella after mounting it on the grid. One of my systems has a flip stage, which makes it much easier, but it has some alignment issues. After I use 30kV and 24 nA to live image the sample and set the trenches area, and then go to a higher beam current for milling, it causes a shift (the cutting square is no longer at my designated place). We have two systems; the PFIB one does not have this issue, but the Helios has a shift of about 15 µm. I talked with the engineer, but he said it is normal. Have you seen this kind of issue on your system? Currently, I have to do a quick snapshot before the final cutting to confirm the location. Thanks!
what a wonderful tutorial, let me learn a lot
However, I would like to ask some questions. My sample is a metal sample, which needs to be observed for dislocation in TEM. After making the sample with fib, there are always curved lines on the surface of the sample under TEM observation, which may be the interference introduced by fib. How can I minimize this interference in fib?
Thank you and glad to hear the video was helpful. Regarding your question, it's tough to determine the origin of the "curved lines" on your FIB-prepared samples without seeing any images If you can send me an email (ngr@ufl.edu) with some SEM and/or S/TEM images showing the issue, I should be able to help you out. Looking forward to hearing from you soon.
Dear Nicholas Rudawski. I would like to watch this video for my TEM sample preparation by FIB. But as you recommended in the start to watch another video of "FIB basic operation", so, could you please share that link so that i can watch it first before coming to this video. I am unable to find that video on youtube. Thanks!
Thank for the tutorial, very useful to see the steps. Somewhat les related: what is the sample stage you are using?
You're welcome; I think the stage is a piezo-driven stage, but I'm actually not sure; the attachment on it that holds the stub and grid is called a universal mounting base.
Is it possible to move (forward or backward) ion beam during milling? Or You have to stop milling process, move milling box and start process again?
Yes, that is possible. I have not worked on any FEI Dualbeam for some time, but I think you can do so with the 3 buttons below the milling progress bars (On the right side visible during milling). On Zeiss Dualbeams that is also possible. The distance the beam jumps per click can be set by the operator.
Questions!
-why not flag mount the sample to the slide of the post to reduce redep?
-why no e-beam C depo so that TEM can see the contrast difference….to see where the depo ends and the Si begins
-Are you guys reusing grids?
And last why such a large sample ?
Hi MonkeyBrainSushi: I don't like flag mounting because it leaves the sample more susceptible to bending because one end is free, but many people do indeed use this approach. I'm not sure what you mean regarding the "e-beam C depo", but generally, you are trying to avoid beam-induced contamination on your sample. You could reuse the grids, I suppose, but the cost of the grid is exceedingly small compared to the instrument time cost, so there really is no financial reason to do this. The size of this lamella is actually on the smaller side; I know of many people who use a default size that is basically twice as wide as this sample (~20 um wide).
@@NicholasRudawski Hi Nicholas, I think there is one more reason why flag mounting is not preferred. When you rotate the flip stage 180 degrees and try to weld the back, you might not be able to see the joint because the grid blocks the way (if the lamella is not welded in the perfect position).
It would be great if you demostrate this prepared TEM specimen quality in the HRTEM
Agreed; it wouldn't be too hard to make the video, just some static TEM images with some voice over; thanks for the suggestion!
Nicholas Rudawski , in FIB TEM preparation I typically go to lower KV to minimize amorphous layer and at the final step I use Gatan PIPS polishing machine at 0.5 kV ... it improves samples quality much more in comparison with 2 and even 1 kV ;) your videos are pretty nice !!! We need more !
@@АлександрПриходько-ъ8г hello, Alexander! Could you give a rough guide for 0.5 kV polishing? Particularly, how do you choose the polishing time for different materials?
@@alexandergolubnichy8588 I would say it is quite tricky stuff and I messed up a lot of lamellas. So what I would to recommend. 1) You need to prepare in the TEM grid a special trench and put your lamella there in order to avoid the redeposition from the TEM grid during polishing procedure in the PIPS mashine. 2) You lamella should be thinner at the botom side and iteratively I perform polishing procedure untill the botom side starts to desappear
@@АлександрПриходько-ъ8г thank you very much for your response! And how do you check the bottom side of lamella between iterations? With SEM?
is there a difference between rotating the pattern in the advanced tab and changing the scan direction in the basic tab?
Good question; I think it may depend on the type of pattern. With something like the cleaning cross-section or regular cross-section, I think rotating the pattern or changing the scan direction may effectively do the same thing. With something like the rectangle, I'm not so sure. The next time I get on one of the dual beams, I'll do some digging and I should be able to answer this for you; please look for another reply soon, thanks.
Hey, so I had a chance to look into this on our dual FIB/SEMs and for all three of the patterns I mentioned (R, RCS, and CCS) rotating the pattern 180 degrees does the exact same thing as reversing the scan direction (which makes sense). Thanks for asking about this and helping me learn something new in the process!
I’m working on a very similar device, but of course with way less experience on my side, so I was just wondering if there might be some advantage with this. Thank for looking into it!
I have one question here, why ion beam image needs to be rotated 180 deg? I notice from AutoTEM manual, the probe is from the left side, down to up position for probe approach, but for almost all manual operation manual, the probe is from right up to down side approach, is there any special consideration? Thanks a lot.
Hi Guo: the 180 degree scan rotation for the I-beam is more of a preference than a requirement; I personally prefer it because it maintains a top-down orientation in the I-beam image (particularly when using the Omniprobe), but I know of many users who prefer to keep the I-beam scan rotation at 0 degrees. So it's really more of a person preference than anything else.
@@NicholasRudawski Got it. That is helpful for my understanding. Thanks a lot!
Dear Nicholas
Many thanks for these valuable videos... But I have a question: my samples (oxide) are coated with silver (sputter coating), and always when I prepare a TEM sample a big part of the silver layer get re-deposited over the cross-sectional surface of interest (it looks like clouds over what I need to investigate) and l have performed cleaning steps at 5kv and 2kv but it didn't solve the problem.
Do you have any advice PLZ?
Hi Samer: glad you find the videos helpful. Re-deposition when you are preparing a sample can definitely be troublesome. How do you know the re-deposition is actually Ag and not something else? The most common culprits for re-deposition I find are the grid material and/or the protective Pt. The grid issue can be rather easily dealt with by using a Mo grid instead of Cu. The Pt issue can sometimes be avoid by thinning past the point where the entire protective cap is gone but often you can't do this if you are trying to protect the near surface region.
@@NicholasRudawski
Thanks for your reply.
I confirmed it using EDS .... I tried to post a link for TEM image but my comment was deleted every time I post a link to image.
@@samerdaradkeh1504 You can email it to me directly: ngr@ufl.edu
Hi Samer, indeed silver can be a tricky material especially when it comes into contact with air. To be precise with the sulfur in the air. That can create problems if you store the sample some time out of vacuum and then for example plasmaclean it before taking it into the TEM. The plasmacleaning leads to redeposition. When we went directly (about 10 min handling time) from FIB vacuum to TEM (w/o additional plasma cleaning) we had no issues. Or do you see the redeposition already during FIB preparation?
Best regards
Alex
@@Alex-qv7zd Hi Alex, thanks for the valuable info.
I didn't see any re-deposition during FIB preparation