Electron Microscope Hack to see Graphene

[@BreakingTaps]

*How can you form an image if electrons don't follow straight lines?* I accidentally left out some important details in an effort to keep it simple, my apologies! The secondary electron detector is a "single pixel" detector. So it only knows when it receives an electron, and doesn't have any kind of spatial knowledge. The machine will focus the beam onto a specific region, count how many electrons the detector receives, then move to the next pixel and repeat. So it doesn't matter what trajectory the electrons take to get to the detector, as long as they arrive before the beam moves to the next pixel. Hope this helps clear up some confusion!

[@MR-ub6sq]

Since the English language is not my strong point, I am wondering about something that, according to my knowledge, is not correct: The title of the video contains the indefinite article "an" followed by a consonant. I understand that "an" is used when its object starts with a vowel and "a" is used when the next letter is a consonant. So why is it like this: "...an SEM..." ?

[@kennybowman5139]

Great explanation! The only thing I would add is the way it moves from pixel to pixel within the substrate is an array of octapoles. the octapoles use positively charged capacitance to pull the beam from pixel to pixel similar to the rastering you would see in a slow motion capture of a cathode ray tube tv.

[@YodaWhat]

@@MR-ub6sq That issue seems to have been removed, but here's the answer anyway... F, L, M, N, R, S, X, and sometimes H are typically spoken as beginning with a soft vowel-like sound. Mostly the sound is like E, sometimes like A. As vowel mimics, the may be preceded by 'a'. Is that strictly proper English? I don't know, but it is commonly accepted. 'W' is something of an exception. And of course, other consonants are typically spoken with a harder, more percussive sound. English is weird.

[@MR-ub6sq]

@@YodaWhat Alright. You are right. You obviously meant acronyms, because I understand that SEM is like that and abbreviations are of course pronounced single letters in a row like "es ii em". So the "word" S is pronounced like "es" and as such begins with a vowel. I hadn't realized this possibility, but thanks for the tip. We're wiser now :) It's true that it LOOKS weird when written though.

[@YodaWhat]

@@MR-ub6sq It is not only the vowel-like case with acronyms. For many words, when the first letter comes from the list F, L, M, N, R, S, X, and sometimes H, that first letter is spoken somewhat softly, as is it were a vowel. The vowel-like H is heard in the way the British pronounce "Houston" like "yoo-ston". Americans begin "Houston" with a rather breathy exhalation, similar to the beginning of "huge". Here is a silly mnemonic for the list: Falcons Like Mice Not Rats So eXcuse Hawk Weakness.

[@otv9005]

Dude, i think i speak for many by saying i completely don't mind seeing longer videos. I could watch this for hours. Thanks for all your work.

[@kevinbissinger]

yeah, but not the algorithm :(

[@miles11we]

​@Kevin Bissinger yeah it's definitely a tough nut to crack with sciency videos on YT, like yeah a bunch of us would love 3hr videos but what if we don't watch the video all at once? Does the algorithm see that as viewer fall off/lack of retention? How much of a really dense video is the average viewer gonna watch? It's hard to get new viewers if the average person feels totally lost without watching a bunch of previous videos. So splitting your video up into parts might not be a good work around either. But then if you go too simplified is your core audience going to become disinterested? And all of that is ignoring the upload scheduling shit where (I think) making fewer more complex videos is less advantageous over more frequent shorter content.

[@Wtfinc]

cant watch it with headphones because of the high pitch whine

[@progi3751]

felt like 5 mins

[@jamesbaxter8914]

As someone who works with SEM professionally, the descriptions in this video are very well articulated.

[@evanbarnes9984]

I just love that you have an SEM in your shop and can really show us how it works. Seriously cool tool to have at your disposal. I also really like your use of the Concepts app for illustrating this! I just started using it similarly as my whiteboard for my math classes. It's really nice to be able to draw a triangle, duplicate it, and scale it, and then be like "that's what similar means."

[@BreakingTaps]

First time using Concepts, but I really like it so far! I've seen a few other channels use it to great effect (most notably Stuff Made Here) and this seemed like a good project to try it out on!

[@McImTheBear]

I need you to know what you're doing on KZbin is phenomenal. A lot of KZbinrs get caught up with making every video better than the last and then they stop. I hope you don't fall into that. Everything you share on your channel is so interesting to me and I love learning it all. Thank you.

[@WhenItComesToThisShit]

Wow man, seriously impressive you have the grit to even attempt such a notion in your personal machine shop. Your cohesive understanding of a variety of practices is very enthralling. I am just discovering how deep of an appreciation I have for those who actively seek knowledge, and I appreciate your candor. I'm glad the internet exists so I can live vicariously through those with the resources.

[@ThatGuy-fi9bm]

That’s cool! Never seen that before, I work with STEM a lot and really like the technique and what it can do, I guess it would be fun to see a resolution comparison between the “STEM” and regular SEM.

[@BreakingTaps]

I'll include one next time, that's a good idea! Have some improvements to make to the device then would be fun to see how it compares to SEM

[@bazoo513]

So, let me see if I got this right: - Your setup still scans a focused electron beam through the sample, as opposed to "real" TEM which works more like a, umm, slide projector? - It would be possible, _in principle,_ to collect separate images from backscattered electrons, "primary secondary" (I could not resist) electrons that normally get absorbed by that cup (that would require some quite bulky and unwieldy detector), and "true" secondary electrons from that plate you manufactured, _plus_ chemical composition of the sample via X-ray detector detecting X-ray fluorescence? (Of course, I understand that, in practice, tradeoffs in geometry, ideal beam energy and intensity etc make this impractical if not impossible, but still...) Man, I envy you! 😀

[@BreakingTaps]

Correct on both counts! - A traditional TEM uses a large, collimated beam of electrons just like a slide projector, and captures the "shadow image" on the other side. There are scanning TEMs that work similar to how this converted device works, by focusing on the sample and scanning. To be honest I'm not familiar enough with TEM/STEM to really know why you'd choose one vs the other. I think TEM is better for samples like tissue slices, while STEM might be better for nanoparticles/crystals/etc because you can derive additional information based on the angle the electrons are diverted. - Yep! In practice it could be quite tricky getting the photons out of the sample cup, and keeping the two types of secondary electrons from mixing. But in theory totally doable I think! There are also dedicated sensors you can place inside the chamber that go underneath the sample, which means you can collect the electrons directly instead of using the conversion plate. I've also seen schemes that use small electromagnet coils to curve the electrons after they come out the bottom of the sample, so you can "select" different electron energies to analyze. Lots of neat little tricks :)

[@bazoo513]

@@BreakingTaps Thanks!

[@janctrnacty1215]

Great video! As somone working with these diffraction grating replicas, my heart shipped a beat when I saw grid on your finger, then I read note on bottom😆 These samples are so fragile than even high temperature during shipping can destroy them.

[@BreakingTaps]

Haha yeah, I was reviewing the footage and was like "Hmmm, I should put a disclaimer on this so folks don't have a heart attack 😂". Although unfortunately I did ruin the grating after imaging it, was handling it with a bad set of tweezers and it fell onto a not-super-clean surface, contaminating it pretty heavily 😢The grating is still fine, but the spheres are mixed with random junk now. Lesson learned and bought some dedicated TEM tweezers

[@janctrnacty1215]

@@BreakingTaps We are using vacuum tweezer to manipulate with all types of grids, but this type is really fragile, so even with vacuum tweezer the film on grid get demaged sometimes, but still better than normal tweezers, which usually deform the grid.

[@KallePihlajasaari]

@@janctrnacty1215 Aha, I was wondering what kept the samples in the spaces between the grid. Now there is mention of a FILM, what is it made of and how thin is it and HOW is that film prepared. Why is it called a diffraction grating when it looks like a basic TEM grid????? So much specialised knowledge hiding in just one tiny grid.

[@suspectsn0thing]

This is (as usual; can't overstate how much I love this channel) a good explanation that answered most of my questions, but I'm confused regarding one thing: early on when you're talking about backscattered electrons you say they don't lose any energy when they swing around the nuclei, but I'd always assumed changing direction usually comes with a change in energy? The two examples that jump two mind are the orbital slingshot/gravity assists you mentioned, which are done explicitly to add large amounts of energy to the craft without having to spend much fuel. The more particle physics-y situation that jumps to mind is synchrotron radiation, which is caused by charged particles shedding energy when their trajectory is changed. Why don't the electrons lose or gain any energy in the telescope's process? Or do they just not lose enough for it to really matter?

[@exvils]

19:00 yes its interesting, we had scanning electron microscope in Uni, and also normal microscopes, so i have some practice looking at crystalline layer of metals

[@jimsvideos7201]

As an aside, your electron on the left with the cowlick is adorable.

[@jonathanlebon9705]

Long video, and I enjoyed every minute of it. You're a good narrator. Clear, concise, and a scientist at heart by the way you speak. Always enjoy hearing what you have to say. For me I really want to see more of all that, it's a whole other world. You should really pull up obscure things under that SEM too. So lucky to get to play around with a tool like that. Can't wait for v2

[@maleeshapriyanjana7604]

as science students we have never got a chance to see an electron microscope in front of eyes at least. thank for your effort to show and explain how they are working

[@はいこれはロボ子の婚約者]

"Its kind of a niche and esoteric rabbithole" well yes, that's exactly why it's interesting. Even though I have absolutely no application for an SEM at the moment, who knows when all this knowledge might come in handy. Thanks for your great content!

[@BRUXXUS]

Super cool! Only think I wish I could see if a comparison between the TEM images and what the same thing looked like with the backscatter detector. Although, I have no idea if it's as easy as toggling between detectors. Probably not.

[@BreakingTaps]

Ahh, that would have been a good idea! Easy to switch, just takes a few seconds. I just didn't think about it because in basically all the samples it is completely black! 🙂 The particles are so small they don't generate much backscatter at all, so end up being an entirely black field. The only exception was the tungsten disulfide... because tungsten is so heavy (and some of the particles were kinda large) they can generate enough signal to be seen on BSE.

[@BRUXXUS]

@@BreakingTaps That makes sense. :D You did explain the benefits of TEM, but it would be kind of neat to see the advantage over what the SEM can do for the same sample. 🤔 Not a big deal, still a very cool video!

[@KallePihlajasaari]

@@BreakingTaps Can (S)TEM view through an already back thinned camera sensor or is the structure still too massive. Just thinking what sort of things can one view that are thin enough, a flies wing perhaps?

[@C-M-E]

Getting true graphene isn't quite at everyday accessibility, but nanoplatelets are, which is more or less what you're seeing on your samples. There's a few more steps involved to get a bit closer, though it's made somewhat more approachable if you start with a non-graphitic source and do the graphitization phase yourself. I've got a week-long process that I'm streamlining with every breakthrough, but have to send my samples to a specialist as I do not (currently) have a TEM-capable microscope. Still trying to decide if it's worth picking up a Raman spectroscope versus waiting on independant results for purity certification.

[@BariumCobaltNitrog3n]

You're one of the few people who has a very technical skill set using machines most people don't have access to, somehow brought them all home (surplus probably) and play with them like we would. If we knew how. I approach long (over 30 min) videos with a finger on the L key but not yours. You have interesting images and a great voice and always new (to me) topics. 👍🏽

[@dirtdart81]

I would watch an hour long video of the rabbit holes you find, please don't worry about their length. Great stuff!

[@edgeeffect]

Where I used to work they were doing CryoEM single particle analysis, which I didn't really/completely understand (I'm ONLY a web developer) so it's always cool to see you doing stuff with TEM, I might understand the images I was displaying in my software someday. ;)

[@defneozturk]

I watch this every now and then just to refresh I use a Zeiss Sigma500 FESEM with HDBSD, SE2, and sTEM detector so all of this is great for review. Thank you for your work.

[@CircsC]

Regarding video length - I'm here for as much as you want to make. It's very interesting and you're good at explaining it.

[@cdburner5911]

I have always been so fascinated by electron microscopes. This is so neat! Once, years ago, I had the chance to acquire an old (70s era) electron microscope from a salvage yard. At the time, I was mostly concerned with, how would I ever get this going, as it was all CRT and really old transistors, but now I am kicking myself for not getting it. Even if it would have been a huge pain to get working, getting one for basically scrap value would have been awesome.

[@christopherleubner6633]

I would have grabbed it if nothing else for the HV and vacuum goodies inside.

[@cdburner5911]

@@christopherleubner6633 Indeed. That also would have been a good idea, and more of a reason to kick myself. Lol. After applied science got his olde one working, I was so disappointed in old me. I console myself with that I got some other awesome goodies there, like some lasers I ended up selling on to someone at nasa, which is awesome.

[@DD-ze7qm]

Ouch 🙄

[@23chaos23]

do a part 2 with cut content deep dive stuff, first half of this one was an awesome sem/tem/stem explanation. I pretty sure the i'm not the only one that thinks this kind of content should be longer format or at least multi part videos. You are doing amazing stuff, keep it up

[@byronwatkins2565]

At 8:30, where the electron exits is irrelevant; the video monitor plots the gathered electron current vs. the position of the beam's entrance location. That teardrop receives the charged beam, however, and builds up an equilibrium charge to repel the beam. The charged volume in TEMs is much smaller and deflects the beam much less.

[@BreakingTaps]

Cheers for the clarification! I definitely muddled that explanation up a bit (as well as not really explaining how TEM and STEM are different).

[@gracesonaufderheide4743]

Super cool! I work with AFM-IR mostly so it is cool to see other non-optical microscopy techniques. Very few channels do similar work! Also, do you take sponsorship opportunities? I think you’d really like playing with what AFM-IR type techniques can do

[@krishnajayaswal6051]

Electron microscopy is still optical just not visible light but electron waves/particle

[@JohnSmith-kf1fc]

You're one of my favorite content creator. Videos are packed with educational value, smoothly narrated and fueling curiosity. Tipping my hat to you sir

[@MrHichammohsen1]

We even need longer videos man! We appreciate it so much.

[@akshay6018]

I have a question,won’t the electrons interact with the coating in the metal

[@jeolman1]

I have been a field service engineer working on SEM/FIB and TEM/STEM systems for the past 29 years. Your presentation was a great explanation for those that have no knowledge or a very limited understanding of what these systems do. I have been teaching and training other Field Service engineers, since the late 1990s. I often struggle with explaining it in laymen terms to non professionals, as I tend to get a bit too technical and followers get lost.

[@tiagotiagot]

Since you're seeing thru samples, would there be a way to rotate the samples in a controlled manner to get multiple images you could use to calculate a volumetric scan of the sample? If it's too hard to control exactly how much you rotate the sample, could you have a grid mask or some other more complicated pattern around the sample that could be used to calculate the rotation each individual scan was taken at?

[@hamjudo]

Alternatively, add a second beam source so you can take stereo images.

[@BreakingTaps]

I bet you could! Sorta like a very shallow CT scan :) There'd be a limit to how much you could rotate the sample, since at some point it'll either be too thick, the beam start hitting the grid or other parts of the sample. But in principle it seems like it should work?

[@ikocheratcr]

Your video are never too long ;) More info and details: yes!

[@fvoisard]

you might be able to boost your contrast if you try a dark-field detector, all you would need is a hole in the center of your detector. A simple washer could work, if you could polish and gold-plate it, even better

[@Baleur]

Good video, but the full screen white paper is killing my eyes at 3am.

[@EricDalgetty]

This was an awesome explanation! We just got a SEM at work, ill have to try this

[@CafeBikeGirl]

lol, I'm game.

[@BreakingTaps]

Congrats! What kind did y'all get?

[@triggerfingerstudios]

I'm always surprised how well the reflection method works in the SEM. I helped to invent a new type of reflection holder that used a scintillator and visible light instead. The results were great signal, less noise, higher resolution. It rivaled full sized STEM detectors. It's called the UVD STEM holder, and is only available from Hitachi.

[@BreakingTaps]

Just looked that up, super cool device! UVD in general looks really neat, need to do some more reading on that, and the STEM variant looks very cool. It makes sense though, you won't get electron cross-contamination from SE off the sample itself, the scintilator I assume provides some amplification and with direct line of sight to the sintilator the photodetector can be optimized to capture all the photons. Neat! I wonder if there's space in my chamber for a photodiode... Cheers for the note, I love seeing ingenious devices like this!

[@namibjDerEchte]

@@BreakingTaps If you can optionally resolve the pattern on the scintillator, you could do holographic scans of the sample. Not sure what beam voltages end up favorable, but a classic analog CCD should be rather resilient to the beta radiation of the beam. Well, I say "holographic"; the actual realtively-easy-seeming (from a mechanical standpoint, at least) would be CBED, where you capture a 2D diffraction pattern for each STEM pixel. Just beware of power density in the sample (CCDs rarely exceed 100 frames per second; this would be the speed at which you could step the STEM beam).

[@AdlersAesthetics]

You have the best content, keep it up!

[@Nobe_Oddy]

HELLS YES THIS IS INTERESTING!!!! That EXACTLY why I'm subbed!!! You do some of the COOLEST SHIZZ and it just blows my mind!!! I WISH I had stuck with academics. It's like now that I'm older I just can't learn enough!!! lol THANK YOU for all the amazing stuff you bring to KZbin and allow anyone and everyone to learn about.... :)

[@Ferrarimangp]

12:27 yeet !

[@Matt1979321]

"I hesitate to go too much into detail" Dude, that's why we're here! Long live the painful details!

[@mvadu]

Can you please make a followup on the kind of detector you have? How does the point in space where the electron hit the sample is mapped out? 🤔 I. E. How does it map x/y coordinates? Is it a scanning system like CRTs and scanner moves to next spot after either getting the secondary emissions or after a specific timeout?

[@AndrewZonenberg]

The beam is being scanned across the sample one pixel at a time. Signal from the detector is captured in sync with the beam, so if the beam is at the top left pixel of the scan area then the signal from the detector goes to the top left pixel of the image, etc. The detector itself in a SEM has no spatial selectivity, it picks up any secondary/backscattered electrons heading its way. So you have to know where the beam was. In the old school analog CRT scopes, they'd literally drive the imaging beam and CRT scan coils with the same waveform (scaled appropriately in magnitude) and then the detector signal would go straight into the brightness input of the electron gun.

[@mvadu]

@@AndrewZonenberg this makes so much sense Andrew. Thank you..

[@miles11we]

Always know I'm gonna learn a bit about a subject or process iv never even heard or when there is a new breaking taps video.

[@bowang5390]

A really smart way to get aim information from materials, cool

[@liamjoyce22]

Very cool video, I actually work at Nanoscience instruments where all the service and intakes are done for the XL2s

[@trevise684]

1:25 i cant BELIEVE you didnt even mention something about the scanlines being similar to a crt television

[@BreakingTaps]

🤦‍♀️ Whoops! Trying to think of an analogy and the most obvious one was staring me in the face!

[@morkovija]

Thats how I very much like to spend my fridays, thank you

[@WhiteDoppler]

If I had half as much ambition as you do, my dad wouldn't have needed to go get cigarettes

[@timvw01]

Very cool, love the details, also how you see the imperfections of the wall of the gird.

[@tiagotiagot]

With the original arrangement, can you clock the time between the electron being emitted and being detected to separate the detection on the secondary detector and produce multiple "layers" based on how much the electrons bounced before coming out and hitting the detector? Or if you can't detect when the electrons get emitted, would it be possible to use the beam deflection as a sort of shutter, timing based on when you switch the aim from a dump direction to a direction that does hit the sample, getting an statistical sample over many shots to estimate the bounce time for each individual deflection angle?

[@BreakingTaps]

I'm not sure, maybe! Grain of salt, I'm garbage at EE so not entirely sure what's possible or not. I have seen some devices that use an electromagnet to alter the flight path of electrons after they pass through the sample. By varying the strength of the field, you can curve the electrons more or less depending on their energy, basically "filtering" them (if an electron has multiple collisions passing through a sample it'll have a lower energy, etc). You might be able do it in reverse too and vary the accelerating voltage to see how much is required to get through the sample and estimate from that? Probably a way to do high speed capture of electrons though, not sure. The scintilator might be the limiting factor in terms of speed?

[@whathappenedattiananmen7484]

maybe the line from the line from the graphite comes from the thing under the sample during the nanoparticle dropping, i know absolutely zero about the thinks u talk about but i really like your channel

[@stuartmicklethwaite5427]

You could try putting a hole on the plate under the sample and it will give you dark field image. Gold coating it will give better signal as well

[@BreakingTaps]

In the works for version 2! Part of the rambling that got cut at the end was talking about that, the aperture is definitely too large and you can actually see it shift from brightfield to darkfield depending on where you're looking on the sample. Plan to add a little slot accept different size and shape masks/apertures under the sample so that it's easier to test. Noted on gold! Been avoiding buying a sputter target because 💸 but it would be really useful to have in general

[@stuartmicklethwaite5427]

You can get away without the aperture if you use a smaller plate underneath and coat everything else in carbon paint. It's easier to change between bf/df/haadf just by changing the plate. If you remove the sleeve on the top and have a small WD you will get better resolution and electons shouldn't get to the detector because the pole piece will get i the way. Thats how we run ours at Leeds uni

[@prbmax]

Am I the only one who doesn't think your videos are too long?

[@irkedoff]

Great video. I like the variety.

[@mattlong4102]

nice job diy tem. never would have guessed it was possible

[@chrzoc]

As the sr. Mech/App eng for an aerospace company, I find your youtube name hilarious (: Addon: P.S Glad to see you using CAT40xSFC’s with your 5x for the 3+2 and 4+1/axis substitution micromilling. Good man 👍🏻

[@tysonion01]

A scintillator? That's Brilliant. I could really use something like that in my life.

[@Veptis]

So we now have multiple different imaging modes. And essentially different signals. Can you map them to RGB or other color challenge to display some of the difference between them as a color coded image? Sure there isn't any chromatic order, but there is plenty of ways to map 3 channel (or more) data to color (RGB, YUV, HSV, ...) I have seen people color TEM images with additional information like from an XRF probe

[@Anar10n]

Wow man makes clever piping for electrons. Awesome. I know it's probably dumb but could you make and observe positron-electron collision in your magical mystery box? No idea how hard is it to find some positron emitting stuff though

[@practicalmicrofabrication1858]

Super neat add on for the SEM! Did not know this was a thing.

[@smellycat249]

I love these videos. They are so different than most things out there and the videos are super entertaining. And since you never swear in the first 15 seconds of your videos, I think you should never have to worry about monitization. Lol.

[@bazoo513]

Swearing allowed only when a tap breaks 😀

[@simonwatson2399]

Graphene has a very low transmittance for visible light. Single layers are visible under a light microscope. I wonder if the same holds true for electrons at the energy involved here.

[@pibus12]

what a clear explanation, thx for video, but i didnt catch on what surface was graphite laying, if we were cathcing all the electrons passing through the grid, where there is no surface, graphite should have fallen on platinum plate further

[@BreakingTaps]

Thanks! And that's my fault, totally glossed over that. The grids I'm using have an ultra-thin layer of Formvar (polyvinyl formal) polymer, think it's 20-50nm thick. It's thin enough to be mostly transparent to electrons, but robust enough to hold things like nanoparticles and flakes. There are a bunch of variations you can order depending on your sample (silicon monoxide layers, silicon nitride, pure carbon, thin metals, etc)

[@pibus12]

@@BreakingTaps that is absolutely amazing, i am curios how a such thin layer of polyvinyl is made

[@anullhandle]

@@pibus12 dissolve it in a solvent and let that coat the grid? The polymer is sold for diy tem lab use.

[@lalitasharma6687]

I have a question sir My knowledge of stem is next to zero so it's a kinda stupid question. I read in quantum chemistry course that STEM works on tunning current where the beam is going up and down just like tip of a pen an then it create image of sample using that like change in tunneling current Please help me out with this question thanks 😊

[@AndrewZonenberg]

Is there any way to get diffraction images out of a STEM? When I took electron microscopy of materials back in grad school, a big part of the TEM portion was using diffraction apertures to get spot patterns for looking at crystal structure, or imaging only a certain diffraction pattern so you can e.g. see oriented grains as light and everything else dark.

[@BreakingTaps]

I'm not sure to be honest! My gut says that STEM can't generate those nice crystallographic diffraction patterns... but I'm really not sure. Hopefully @AlphaPhoenix or someone equally knowledgeable shows up with some answers :)

[@stuartmicklethwaite5427]

Probably not. TEMs have lenses after the sample that asre used to form the image/diffraction pattern.

[@namibjDerEchte]

Well, going by the Xray crystallography experiment we were shown back in high school (AP physics, senior year iirc), it's enough to hit a monocrystalline region with the beam, and record the angular diffraction pattern that comes out. And the transmission that the video uses with the secondary leectron detector should have that angular diffraction pattern from the crystal structure visible, but the single-pixel detector (and lack of focusing optics between the plate and the pixel) can't resolve it.

[@joaquins90]

That's great!! What other cool gadgets you could add to the SEM to add to your toolbox??

[@BreakingTaps]

There's a neat little mechanism that adds a magnet under the sample, been wanting to try that. It creates an "immersion lens" which boosts the resolution due to better channeling of electrons. Also lots of neat sensors/detectors, but those might be beyond my skills :)

[@legoworks-cg5hk]

In simple optical microscope terms: SEM = darkfield or top lit TEM = brightfield

[@Hexlattice]

"Any sufficiently advanced technology is indistinguishable from magic" -Arthur C. Clarke

[@bcddd214]

Thank you the very intriguing 19 + min!

[@seeigecannon]

Great video. You might want to double check the 3D printed part though as it would have a ton of surface area and voids to offgas. I would think the material itself may also be gas permiable which may screw up the vacuum in the area.

[@BreakingTaps]

Thanks! I've run small printed parts in my machine a few times, tends to work out ok as long as you're careful with the design (no huge interior volumes, dense infill, etc). But mainly because my SEM operates at a pretty low vacuum in the main chamber, just 100mbar. Compared to larger floor-standing machines it's practically atmospheric. So it's very tolerant of outgassing and otherwise dirty samples :)

[@seeigecannon]

@@BreakingTaps oh wow, that is high pressure. I have a vacuum oven I have been playing around with at home (installed an ISO100 flange with a 3" pipe on it). That bottoms out at 0.0mTorr, but that is just because that is how low my convection vacuum gauge reads.

[@nickploum9681]

@@BreakingTaps The sample chamber is definitely not 100mbar. It would be in the range of 0.1Pa-60Pa, as the UI also says. (0.001mbar-0.60mbar). Outgassing is definitely something you want to keep in mind!

[@morgan0]

can you change the field strength on the secondary electron detector? i’m curious how that would affect the image. i assume at high levels it would see electrons from anywhere, and at low levels it would see only electrons moving kinda towards it

[@BreakingTaps]

Not on my machine, but probably on nicer / academic machines! I think your intuition is 100% correct. I've seen some papers that also add an electromagnet under the sample to curve the electrons, allowing them to "filter" it through a second aperture. Same idea, just different way to do it.

[@alexanderzintler]

You are probably recording new data already and I presume you are changing this parameter as well, but what is the minimum working distance of this microscope? With a shorter sleeve you might be working at the optimum WD for highest resolution. Also the grid you were using probably had a thin polymer foil on it already to support the particles. Holey or lacey carbon grids might be worth a shot. (And yes, I work in EM). As others said, I really like the longer format and the animations you showed!

[@BreakingTaps]

You're correct, I switched to a shorter sleeve and it definitely helped! Not entirely sure the minimum working distance to be honest, my machine is a Phenom and they try hard to idiot-proof the machine (the stage won't let you insert a sample that is too tall, so it's not possible to hit the pole piece). I think the working distance is 1-3mm though, something in that range. Also correct on grids! Mine were formvar stabilized with carbon. I'll pick up some lacey carbon and give it a shot! To be honest I was a little overwhelmed choosing grids, so many different varieties. So I just grabbed what seemed like the "simple" or "default" choice based on ted pella descriptions :) Appreciate the tips!

[@alexanderzintler]

@@BreakingTaps its interesting to learn about these simpler Seams and how they are made for a broader user base. Thanks for the explanation! And yeah, there are grids in every shape and "coating" available, the ones you had are actually also very good for the experiment you showed. Thank you for doing these videos, these are the ones I show to my less sciency friends to explain what I'm doing 😅

[@alexanderzintler]

@@BreakingTaps I just pick this up here again: I recently learned that there is an open source and low cost option for transmitted electron detection in an SEM (BSE as well). doi: 10.1016/j.ohx.2023.e00413 The only thing your SEM would need is electrical feedthroughs 😬As there are not a lot of things open source or low cost in EM, I thought I'll leave this here.

[@RobertElderSoftware]

If you don't mind mentioning numbers, what's the approximate $ investment required to get to a point where you can make images likes the ones that you see in this video? I would imagine there is also a huge amount of setup and calibration to get everything working just right. Is it feasible to buy a hodgepodge of parts (scintillators, optics, etc.) off of eBay and connect them together, or do you have to buy into a specific proprietary ecosystem to have a hope of getting it working?

[@electricalychalanged4911]

About 200k

[@electricalychalanged4911]

That is for a new SEM if you want to have an used system you might be able to get it for half or less especially if you know what you are doing and already have the required pumps.

[@BreakingTaps]

Yeah so my system is 60k-150k new, depending on the model and company. It's a "desktop SEM", so designed for less rigorous analysis than a big academic floor-standing model. They also don't show up on the used market very often because they are relatively niche. New floor standing models are probably what @electricalychalanged4911 said, 100-200k or higher. But you can find old analog SEMs for pretty darn cheap these days! 2-10k, depending on location and condition. They still take great images, just often are purely analog and are being sold by universities as they upgrade. Might need to tinker with them or debug analog electronics though, and parts are probably more difficult to find (but also, easier to replace since old analog stuff, often come with complete schematics, etc) Building one from scratch is probably quite a feat, although Ben at Applied Science did it :) The old analog ones are pretty amenable to tinkering though. Usually standard fittings (ISO flanges) so you can plug/play a lot of different things

[@RobertElderSoftware]

@@BreakingTaps Awesome, thanks for the detailed reply :) A used one for 2-10k sounds tempting. I guess it would be a bit of a gamble though. If you get unlucky, I suppose you might end up spending enough time/money 'fixing' it that it would be 'cheaper' to have just bought the desktop one. Great stuff, keep it up.

[@hamjudo]

While he was a high school student, Sam Zeloof bought a broken electron microscope off of eBay, and fixed it. So we have an existence proof that it was possible for at least one person to get their own working electron microscope with very little financial resources. Just a thousand hours of work. He has posted videos on KZbin about his microscope.

[@malachilandis9542]

I feel like I missed an important video somewhere where you all of a sudden have a machine shop... I swear you were working with a CNC router not too long ago, now you are on the sick 5-axis! Congrats

[@BreakingTaps]

Hah yeah, it's been slowly accumulating over the years. I basically run a prototype jobshop in the background now, just doesn't show up on the channel much. Hoping to incorporate some more manufacturing stuff in the future though!

[@moki123g]

They make a esd safe pla. I wonder if that would also work for your absorber? Another great video! Thanks!

[@electricalychalanged4911]

Can you actually get higher resolution that way? When we use TEM in our group, it is moastly to measure distances of atomic layers. What is the benefit for such a conversion. And why doesn't the 3d printed destroy the vacuum in the SEM chamber. I always had huge problems withe vacuum when I used 3d printed stuff. Thanks for the Video this is great

[@MrTheomossop]

You can see in the video that he is using a working pressure of 0.1Pa, which is WAY higher pressure than a TEM requires, and is relatively high for some high resolution SEMs. I like these lower-end SEMs because they are much much more resistant to dirty materials and usually "good enough", depending on the analysis, and massively speeds up throughput if you have a lot of samples, as pump-down times are much lower. Check out Environmental SEM, many machines can be made to work with volatiles and dirty or outgassing samples, and water.

[@BreakingTaps]

Yep, @MrTheomossop nailed it! My SEM is a desktop machine, so the "high vacuum" mode is actually pretty low vac (and the actual low vac mode is practically atmospheric, iirc it's like 60Pa / 0.6mbar). But I can cycle the chamber in about a minute so it's pretty convenient for most things I work on, and tolerates all kinds of awful samples and outgassing materials 😁 To the question of why it's (potentially) useful, I'd struggle to resove many of these 50nm'ish sized particles if they were just placed on a silicon wafer. Very low contrast on SED, and almost invisible to my BSE. Still hard to resolve with this converter but there are more electrons to work with so the signal is a bit better. Can't speak to more professional machines, but I imagine the difference is more stark if you have a proper, floor-standing SEM :) You can also play with bright/dark-field depending on how you mask the aperture which I'm told can be useful, although I don't fully understand the ramifications of when you'd want each yet (I did somewhat misconstrue TEM, STEM and STEM-in-SEM in the video though. Was trying to keep it concise and didn't properly define each, which probably resulted in some confusion for folks that actually know about the different instruments)

[@randomname4726]

Great video, but don't worry that it's too long. I would love longer vids, and I'm sure lots of others would too.

[@arcrad]

In an SEM, why does interaction volume blur the image? If you know where you shot the beam, then once you get an event and subsequent signal, why does it matter exactly where that electron came back out? Is it a timing issue between entering and then coming out and being detected? I figure since you know where you shot the beam, once you get a return signal you still know where it was supposed to originate from.

[@BreakingTaps]

Hopefully a more knowledgeable microscopist chimes in, but my understanding is that it's mainly due to "side effects" of that increased interaction volume. For example heavier elements will generate more backscattered electrons than lighter elements so have a higher signal. But if the electron beam enters at a Carbon location (light) but bounces around and hits a neighboring Tungsten atom (heavy), it'll generate a high signal at the Carbon location because the SEM thinks the return signal is from where it sent the beam. Sorta similarly, electrons might drift over to a nearby edge where they are more easily emitted as secondary (you often see glowing/charging on edges and sharp features), so you'll get an increased signal within a certain distance of an edge due to that, even though the spot you actually scanned would have lower signal otherwise. Also knock-on effects like electrons in the interaction volume repelling other incoming electrons, etc etc. So it's sorta a combined effect causing a reduction of signal:noise ratio. At least, that's my relative amateur understanding :)

[@AndrewZonenberg]

@@BreakingTaps Exactly. The exit point of the electron doesn't matter, what matters is where the interaction took place. You're essentially unable to distinguish between an interaction that happened right under the beam impact site, and one that happened a small distance away from it.

[@electricalychalanged4911]

It depend on the acceleration voltage and the beam intensity. In Both cases higher means more interaction withe the sample because the beam gets in deeper with higher energy settings. This gives you more over all signal strength but worse signal to noise ratio. If you want topology you use low energy settinggs if you want chemical information you use higher energy settings. In general if you go down the periodictable you get more elektrons per atom so there are more chances of an interaction. Same withe the positive chaarge of the cores of the atoms. So this why you get more Signal from lower down in the periodic table. Also secondary electrons are more easely produced in the surface layer of the sample. I am not sure why. Maybe because there the beam hits withe the highest energy and focus. And regarding your comment about were you shoot the sample. The fokus of our elektronbeam is not perfect so you are always limited by the beam diameter. In moast cases this does not matter as the topology of you sample is much rougher than your beam, but when you talk about Atoms or super small nanoparticles this is a real issue. Greetings from Germany

[@arcrad]

@@BreakingTaps Thank you for the clear explanation! That makes a lot of sense. I was thinking of it in too binary terms, either got a signal or didn't get signal. The additional considerations you outlined greatly help me understand the negative effects of the interaction within the bulk material.

[@jhonbus]

If my understanding of it is correct, it's not so much that the _image_ is blurred, it's that the sample itself is inherently "blurry" to this technique. You know where on the surface of the sample the beam was pointed at for a given detection, but the interactions within the sample that either reflect electrons or don't could be happening somewhere that's not exactly where the beam is pointed. It's maybe like trying to read a piece of text behind a pane of privacy glass. It's not that there's anything wrong with your eyes, they are still forming a perfectly good image showing you where on the glass surface every bit of light that makes up the image is coming from. It's just that every spot on the surface of the glass is showing you a part of the text from a randomised area so the image appears blurred. Not sure if that analogy makes any sense, given the way electron microscopes work is kind of "backwards" compared to the way eyes work. But it was interesting to think about!

[@Mwwwwwwwwe]

Mr Taps 😃the god of science!

[@parkerengines]

Now I need to know why the electrons only make a U turn around the nucleus instead of coming off in all directions, depending how close together they got

[@BreakingTaps]

Hehe, yeah I omitted that scenario for brevity. There are plenty (most probably!) of electrons that zip by and get flung off into random directions. That's partially why there's such a big "interaction volume" for an electron. It'll cruise near a nucleus and get diverted, go near another nucleus and get diverted again, on and on until it smacks into a nucleus, loses enough energy and is captured, or gets flung back out as a backscattered electron.

[@CarterColeisInfamous]

1:16 isnt that just a television?

[@BreakingTaps]

Yeah! Basically exactly how CRTs work, just with the electron optics configured to produce a very small and highly circular beam over a small field of view, while a CRT is optimized for a big area. Otherwise essentially the same in principle!

[@mvadu]

"electron optics" nice choice of words.. For someone who isn't following might be mislead to thinks it's glass (like regular optics) where you got buch of fine tuned coils producing magnetic fields acting as an optical system on beam of electrons.

[@GLITCH_-.-]

can you add a comparison in your next video? I have no idea how much better the resolution is, if a all. :(

[@BreakingTaps]

Whoops, that was pretty silly of me. The improvement isn't _huge_ in my machine, mostly because it's a "lower end" machine and partially because the geometry of the holder isn't perfect yet. Once I tweak the holder and make an updated video I'll be sure to include some comparisons! For my machine the difference is basically "dirty smudgy blobs" using this device and "noise" when using it regularly. Really hard to see stuff this small on my machine, so seeing it at all was a win :)

[@GLITCH_-.-]

@@BreakingTaps Ah, that's great. I'm eagerly awaiting the next video (ANY video you make is amazing) :)

[@gsuberland]

I spent like 10 minutes chuckling at the little electron faces. Check out cool guy on the left with his quiff.

[@lidamullendore6166]

Can't wait for part 2!!!

[@nobodynoone2500]

Parts 2-17 please.

[@RoyaltyInTraining.]

I never knew that 3D prints can be used in high vacuum systems. Wouldn't there be pockets of air trapped inside due to the layered structure? Seems like it would outgas like crazy.

[@BreakingTaps]

They aren't too bad if you're careful. The material itself doesn't outgas a ton (there are some outgassing studies by NASA and similar, it's not amazing but not bad for a polymer), so as long as you're careful about walls and internal voids it's ok'ish. Either single walls with access to the interior, or dense to avoid trapped pockets. The other consideration is that my SEM actually operates at a pretty low vacuum (100mbar) so it's very tolerant of outgassing and otherwise dirty samples. :)

[@omegasalmonfish]

How the hell do you get any kind of focused image from those secondary electrons floating all over though?

[@BreakingTaps]

I pinned a comment regarding this, did a poor job explaining! tl;dr: it's a single-pixel detector, so it only knows when an electron shows up and not where it came from. So the machine hits a certain location, records the number of received electrons and then moves to the next location. Apologies for the confusion!

[@maplaboratories]

I noticed that your design (as well as TedPella PN# 16412) use conductive polymer as the anti-scatter sleeve. Do you know what the advantage is of using conductive polymer here? Would there be a disadvantage to using something more conductive like a piece of aluminum or a brass tube?

[@BreakingTaps]

I'm not 100%, but I think it's to discourage the generation of additional secondary electrons. I.e. a backscattered electron might strike the conductive sleeve and since it has the same energy as the original beam, can liberate secondary electrons. If that happens near the top they could escape into the detector. So using a polymer instead a metal helps reduce "cross-contamination" because they are poor SE generators, and being vaguely conductive helps ground them and drain away any built up charge. That's just a guess, I haven't seen a specific explanation for that but it makes sense in my head :)

[@maplaboratories]

@@BreakingTaps That makes some sense. I guess Aluminum would potentially be a bad scatter shield just because it is so bad at attenuating X-rays and those X-rays could generate more secondary electrons. Not sure how big of an issue that would be (I suspect this would be a small effect). I kinda want to order that TedPella 16412 and play with it.

[@jeffgough8862]

Great video. Looks like you have progressed from your Avid CNC. What machine is this?

[@BreakingTaps]

Thanks! And yep, the shop has since grown to accumulate a few more (and much nicer machines). This was a Kitamura Medcenter 5ax, and I also have a small Haas OM2a now too.

[@jeffgough8862]

@@BreakingTaps Wow, that's great. Both look very capable for something that doesn't take up a lot of space. If you don't mind me asking, how do you go about financing that in a home workshop? Do you use your space for contract work?

[@lumotroph]

Brilliant to watch! Would love to do something like this one day. Do you have a video about how you came to get into these hobbies / this career? How come you have a machine shop and SEM? (Jealous). Greetings from Johannesburg ZA

[@BreakingTaps]

Thanks! No backstory video, although I'll probably do a shop tour at some point. I do prototype machining, microfab and microscopy analysis, so the SEM and CNC equipment support that side of the business but I get to play around with them to make videos when idle :)

[@lumotroph]

@@BreakingTaps that’s super cool. I’d love to get into prototype design / manufacturing! Dan Gelbart style

[@gizelle-s]

This was awesome! Thank you

[@qihaoliu3631]

Now I understand what I'm doing in my labs

[@andymouse]

Awesome ! would like to see more, if I still had access to a microscope I would be talking to the peep's in the machine shop right now !...cheers and Have a Great New year !

[@frantisekvaske2144]

The most convoluted way to save 1000 dollars. Luckily you had that CNC :D

[@PixelSchnitzel]

Amazing! Would it be possible to improve the image SNR by stacking 4 or 8 sequential images of the same view? . . . which leads me to also ask, how long does it take to capture a full resolution image like these?

[@BreakingTaps]

Yep! The images actually are a stack of several images (you set an exposure time, and it integrates a number of passes). But the software doesn't do it very intelligently... it just dumbly integrates per-pixel so if there is any vibration or shifting of the sample, you'll get a blurry image. I've been meaning to capture a lot of individual frames and stack it with some astrophotography software. I suspect that would do a lot better job A 4k image with reasonable quality usually takes 20-40 seconds, so pretty fast. If there's a particularly troublesome image I might drop it down to 1080p resolution and increase the exposure time. Ends up about the same time, but more integrations (and tends to not make the sample move around as much... it gets complicated). Technically the machine can capture 8k images but that takes forever, like 5+ minutes :)

[@ahfreebird]

I've seen the term *tunneling* electron microscope used. Is that something different from a transmission electron microscope?

[@BreakingTaps]

It is indeed different, although also a very cool instrument. Tunneling electron microscopes or more commonly Scanning Tunneling Microscopes (STM) are the historical precursor to Atomic Force Microscopes (but should be noted they are still in widespread use today too, despite being a precursor to AFM). An STM uses a very sharp, conductive tip and places it very close to the surface. The whole instrument is inside an ultra-high vacuum, and when a bias voltage is applied to the tip, electrons will "tunnel" through the vacuum to the tip. You can achieve angstrom-level detection of samples using an STM and literally map out individual atoms. Really cool stuff!

[@blahblahblahblah2933]

Are 3D printed parts straight forward to use under vacuum? do they take awhile to out gas moisture or anything?

[@BreakingTaps]

They aren't too bad if you're careful to avoid big internal voids. My microscope operates at a pretty low vacuum, even the "high vac" mode, so it's fairly tolerant of outgassing and virtual leaks. Under more stringent conditions with real high vacuum it can take a while for all the little voids between layers and walls to finish pumping down, basically a ton of virtual leaks. In other projects I've either done single-walled prints with access holes to the inside so it pumps quickly, or try to keep them small and dense so there aren't a lot of voids at all. The PLA itself doesn't outgas too badly (various sources, i.e. sosolik.people.clemson.edu/papers/Vacuum172a109061_2020.pdf quote rates like 7*10^-6 Torr L/(cm^2s) ). I think NASA has a study somewhere with similar results. Outgasses some but not the worst offender amongst plastics

[@ramous5182]

How do the TEM grids work? I expect the nanoparticles arent just floating between the rows?

[@BreakingTaps]

There is an ultra-thin layer of formvar polymer (polyvinal formal) adhered to the copper grid, usually like 5-25nm thick. It's thin enough to be mostly transparent to electrons, but can suspend particles in the gaps. There are tons of different options depending on what kind of sample you're looking at, silicon monoxide layers, metal film layers, etc: www.tedpella.com/Support_Films_html/Support_Films_and_Substrates_Overview.aspx

[@sethbracken]

What drawing software are you using? It looks really good.

[@BreakingTaps]

Concepts App!

[@TT-lf5hi]

Time is not really an issue as long as the material is not too heavy, don't leave details that are you think are interesting for the sake of video time. I don't think that engineering/science videos correlate a lot with normal entertainment videos on KZbin, for example shorts. The audience is completely different.