#431 – An Interview with Adam McCombs

Adam Mccombs: This is The Amp Hour Podcast. Released February 24th, 2019. Episode 431. An interview with Adam McCombs. Welcome to the Amp Hour. I'm Chris Gammell of Contextual Electronics. And I'm Adam McCombs, and I play with Electron Microscopes. Welcome, Adam. I'm glad to talk to you here. We first met at Hackaday Supercon when you were talking about scanning electron microscopes. How did you start this? Why? Why do you do this?

Adam McCombs: It's fun. It's one of the big reasons for me. But back in, it must have been 8th or 9th grade, I saw Ben Kravsno's video on his Electron Microscope. And pretty much was like, I want one of these things really badly. And so, immediately set up an eBay alert. And that ran for years and years and years. And then my senior year of high school, one popped up. And it was only two and a half hours away and 400 bucks. And about two minutes later, it was mine. So, it all kind of started there.

Adam Mccombs: Okay. Okay. And I mean, you're still pretty young. I mean, you seem like you came up on this pretty fast. How long did this kind of take you to get into it? Let's see here.

Adam McCombs: So, I started in... Where did you start? I got the microscope beginning of March 2017. Wow. And then just kind of went from there. And I had done some kind of high vacuum and high voltage work before. But nothing on the level of actually like trying to lens electrons into a coherent beam to image something. So, it was still a pretty big learning curve. I had done not as many electronics projects as you would suspect for the stuff I was trying to accomplish on the microscope beforehand. A couple of bigger things. I did a build of Dan Berard's scanning tunneling microscope, which uses piezos and tungsten tips to image things at possibly atomic resolution. And that was a fun project to build. And then before that, kind of RC planes. A couple of kind of maker projects here and there. But nothing that would have prepared me for what I was about to get into.

Adam Mccombs: Yeah. Yeah. And it's interesting, too. So, I've been following you on Twitter for a while now. And you post a lot of great stuff there. We'll definitely have people follow you. It's Nanographs on there, right? Yep. That's your Twitter name. But it seems like it's... These systems seem like they're not just electronics. It seems like they're kind of... It seems like a lot of disciplines. What's actually involved in one of them?

Adam McCombs: Oh, absolutely. Well, the big things, of course, are the electronics and the vacuum system. The microscope has to run at high and, in some cases, ultra-high vacuum to be able to function. The electrons just can't travel through the air. They hit things. So, you've got to get all the air out of the way of the electrons. So, getting an instrument to a high vacuum is shockingly challenging. Okay. It's a single fiber, a single speck of dust on an O-ring is enough to drastically degrade the vacuum. What's the level we're talking about here?

Adam Mccombs: What is the level? Like, so, just to give us relative stuff. Yeah, the level of vacuum. Assume I'm an absolute beginner because I am.

Adam McCombs: Okay. Yeah, yeah. No problem. When we're talking about vacuum here, you may think, oh, yeah, I can get the Harbor Freight vacuum pump and I can pull a vacuum. And, yeah, you can pull a vacuum. And that vacuum is going to be, we'll say, between 99% vacuum on a full scale of, you know, 100% being absolute perfect deep space vacuum, which I can't actually achieve on Earth. But, well, that Harbor Freight pump will get to 99 to 99.9% vacuum, maybe a bit further. But to get to the number of nines you need to run an electron microscope, you probably need at least three to four more orders of magnitude of vacuum. And, yeah, and getting that is incredibly challenging. And it stops, the air stops behaving like a liquid. It starts behaving like a bunch of discrete molecules because they just can bounce around. They don't actually hit each other. So, there's no, nothing to direct those molecules down a tube or pressure doesn't start to, doesn't behave like we think pressure does at regular, at atmospheric pressure. So, you have to use things like molecular pumping technologies like diffusion pumps or turbo molecular pumps that more or less beat the individual air molecules down the pump to get them out of the system. But when you're trying to get to that low of a vacuum, everything matters. A single speck of dust on an O-ring is enough to cause a very slight deformation that will let enough air molecules in to degrade your vacuum by two or three orders of magnitude sometimes.

Adam Mccombs: Wow. Yeah, because I guess when I think of like vacuum, I think of like pressure levels. But pressure levels, it seems like it's almost, I mean, it's not the way you're talking about it. It sounds like you're just trying to get stuff out, including all gas molecules, everything.

Adam McCombs: Yeah, pulling an ultra-high vacuum or something doesn't actually apply any more force to that system. You think you pull a vacuum, something can implode. But once you get the vacuum most of the way there, you've pulled, we'll say, 14.4 psi pressure on the vessel. But the difference between 14.399999 and 14.4 isn't that much change in pressure, but it's a massive change when you're looking at it from the perspective of these vacuum systems that need to achieve this crazy low level.

Adam Mccombs: And what is with that number, the 14.4? Is that a specific number?

Adam McCombs: Oh, that's atmospheric pressure in pounds per square inch.

Adam Mccombs: Okay. Yeah.

Adam McCombs: Sorry. Yeah. Yeah. So you're not taking away a massive number of particles at this point. You're just trying to get away individual particles in the vacuum system.

Adam Mccombs: Yeah. I think my closest, I guess I have used electron microscopes and stuff when I used to work in a fab, but I never had any. If I went to go open up the thing and they would have been like, what are you doing? You're fired immediately. And I worked with etch chambers. They had turbo pumps too, but I don't think it was the same level that you're talking about here. Maybe it was.

Adam McCombs: I don't know. Turbo pumps are definitely that level. So that's what the turbo pump does is if you try to start a turbo pump up in atmosphere, it will not work. It only works when you pulled a pretty good vacuum already.

Adam Mccombs: Oh, interesting.

Adam McCombs: Yeah. The turbo pump will actually is so finely balanced and so well engineered for operating at these high vacuum levels where there's barely any air molecules. If you try to run around an atmosphere, it will probably destroy itself. Just because there's so much stuff that gets in the way. There's so much stuff. Yeah. They don't operate in fluid flow.

Adam Mccombs: What is a turbo molecular? Yeah. What does it look like though inside? Is it like a turbine or what?

Adam McCombs: Yeah. It looks, it looks a lot similar. It looks very similar to a jet turbine. Um, and it has a bunch of different stages and it uses blades that vary an angle from kind of really steep to shallow and the really steep blades, when they spin around at, we'll say 10,000 RPMs, they'll, um, hit individual air molecules and direct them downwards into the base of the pump. And you keep doing that and you keep changing the angle of the blades as the kind of pressure increases as you go through the pump. And then you have another vacuum pump, pull all the, or pull the air molecules that are at the base of the pump out now. Hmm. And that's called the roughing pump. And that's kind of the, that's this type of vacuum pump people normally think of is what you have to use in addition or behind the turbo pump to keep it running.

Adam Mccombs: Okay. How do you get it to that point where it's, it's okay to start operating it though? Is that using that roughing pump as well? Like with the bike? Yeah.

Adam McCombs: You use the roughing pump. You, um, you, you valve the system, you valve the system in such a way that at first you connect the roughing pump to the chamber, you pull it down to, we'll say, uh, 10 to the negative three tour. That's just a measurement of vacuum. Um, negative three tour is kind of what you can usually get to with one of those rotary pumps or those roughing pumps. And then, um, you, you change the plunk, you change the valving. So you, instead of connecting the rotary pump directly to the chamber, you connect the rotary pump to the turbo pump. And then the turbo pump spins up and brings it down the rest of the way. And the turbo pump can go to like, uh, depending on how good your seals are anywhere from negative seven tour to negative nine tour.

Adam Mccombs: And that's 10 to the 10 to the minus, 10 to the minus negative nine.

Adam McCombs: Yeah.

Adam Mccombs: Okay. These are like, exactly. Like you said, orders of magnitude. That's, that's crazy.

Adam McCombs: It's your, it's everything in vacuum is orders of magnitude. Wow. Or a high vacuum, I'll say.

Adam Mccombs: Yeah. Okay. And so, so then maybe can we take a step back and can you, so you said there's, there's two things. There's a beam and there's the pump, right? Uh, can you kind of like walk us up and down the stack of what it looks like? I think I was looking at your Hackaday talk and I think you have a good, a good on the Hackaday page for your talk that you gave at Supercon. I think there's a good, uh, diagram here, but could you kind of talk us through what that is?

Adam McCombs: Yeah. So the electron microscope has to start with the electrons and that, and those come from the electron gun. And in the most basic form, that's a, it's called a thermionic emission gun. And we're going to talk specifically about a tungsten one. There's a whole bunch of different variants that all kind of improve performance in different ways. But the most basic form is you take a tungsten filament, you heat it up and you apply high voltage across, you apply high voltage to it to get it above ground. And then due to the chamber geometry, like where the filament is and then where the ground is, um, the electrons get extracted from the filaments and set down the column. Um, it's, it's just creating a huge voltage differential and basically using that as an acceleration. Exactly. Exactly. So it's just accelerating electrons off of the filament and down the rest of the column. The filament gets so hot that if you, if you were to look in the chamber when it was running, it would look like a light bulb. Oh, okay. Yeah. Classic. And when you do that, when you do that in vacuum, you can boil the electrons off and then you can accelerate the electrons and that's what you form the beam with. Okay. And then from there, we, I'm going to talk about lenses and these are going to be magnetic lenses and not optical lenses because electrons are charged particles. So you can affect them with a magnetic field. So these microscopes generally use magnetic lenses to form the beam and to scan the beam across the sample. And so we send this through a couple of magnetic lenses. The first one is to kind of condense the beam. So those are called the condenser lenses. Then your next set will focus the beam. So that's the objective lens. And then to scan the beam across the surface, you use some, um, electromagnetic scan coils and those just deflect the beam in a raster pattern or any pattern you want, usually raster pattern across the surface. Uh, when the beam hits the surface, it generates more electrons from the surface. Uh, it generates a whole bunch of different types of electrons, but normally secondary electrons you're looking at. And secondary electrons are electrons that the primary beam that we shot out of the electron gun hits the surface and knocks the secondary electrons free. And from, from the sample, and then we can pick up the secondary electrons with the detector. And if you get more secondary electrons at one point in the sample, that's a brighter part of the image. If you get less electrons, that's a darker part of the image. So you scan the whole sample recording where it's bright and dark and generate an image. So what is the detector? Does that have its own like sub assembly of stuff? Yeah. Yeah. So there's a lot of different types of detectors for a lot of different types of electrons, but we'll talk the secondary electron detector, uh, uses about, um, 400 volts ish to kind of attract all these secondary electrons towards the detector. And once they come towards the detector, you use about 10,000 volts, um, to suck the, to really accelerate the electrons towards a phosphor. That phosphor converts the electrons into photons. And then those photons get picked up by a photo multiplier tube. And that's how you generate your signal. Hmm.

Adam Mccombs: Yeah. So like kind of the, the, the stuff you're talking about here, I'm thinking of a CRT tube just because of like, yeah, no, absolutely.

Adam McCombs: It's very similar. Okay. Yeah. The, uh, the optics in a CRT tube will commonly use static deflection instead of magnetic deflection, where you use voltages to attract the electrons and deflect the beam. But it's the same principles regardless. Okay.

Adam Mccombs: All right. And it sounds like the, like some of these voltages you're talking about are kind of crazy. Like what, what are some of the ranges of voltages for each element you're talking about here?

Adam McCombs: Yeah. So the, in an SEM, the electron guns in older systems will commonly go up to 40 KV and newer systems that go up to 30 KV. Uh, the reason for that is you used to need more, uh, you used to have, you used to need higher energy electrons to get higher resolution, but optics have gotten so good today that we can do it with lower energies. And in scanning electron microscopes using a lower voltage is almost better or is almost always better. It causes less damage to the sample. Uh, just because, I mean, you apply more energy to an electron. It's got more kinetic energy. It can do more things. And then one of those things that can do is damage the sample. Yeah. So you, you, if you can get a lower voltage, you'll get better results, but it's kind of like light where with infrared, it's a really large wavelength, but with ultraviolet, it's a really short wavelength. Well, infrared is less energy. Ultraviolet is more energy. It's a similar thing in the electrons where a low KV. So let's say like five KV is going to result in kind of a, a, a less coherent beam than 30 KV, which is going to be a very tight, coherent beam just because you got the more energy, the shorter wavelength and everything's more uniform. So what does the, when you say damage, what does that actually mean?

Adam Mccombs: Does it like degrade the material inside or, or what?

Adam McCombs: Uh, it, it, it can, I have literally melted things under the electron microscope. Um, like in real time, but you can see it. Oh yeah. No, I was, um, uh, one of the first times I was on like a real brand new system. Um, uh, uh, part of it was calibrated. So I hit a button to like slightly increase the amount of electrons coming down the beam, which is called the probe current. And it drastically increased it. And then the, where I was focusing, I was, I was looking at a fly under the microscope and where I was focusing, I just started to boil under the microscope. Uh, gross. Yeah. You can literally, and I've, I've, uh, I've melted hair with one at some point. Um, like left a, just a, like a, a spot where the hair used to be. And it's not great because you kind of deposited everywhere inside the vessel. Yeah. Um, so it literally like there's a, there's enough energy there. You can really damage stuff. I mean, if you take an, if you take these, uh, the electron beam, make it super powerful. You actually turn it into what's called an electron beam welder and they use electron beam welders in aerospace and really sensitive applications to weld metal together. So we're using a much weaker version of that. That's awesome. Yeah. Like for what, what, uh, final applications that use for like the electron beam welding, um, aerospace is really common when you're trying to assemble really complex assemblies, um, that need to have incredibly strong penetrating welds. Um, so the, like metal, like larger scale macro things or like tiny, tiny scale things. Oh, no, larger scale macro things. Like you can like weld a rocket nozzle on with this technology. That's awesome. Okay. Yeah. Yeah. No, it's crazy. And it's the electron microscope is just a much weaker version of that in some ways.

Adam Mccombs: Um, yeah. We should, we should say too. So I remember actually, I'm remembering a little bit of biology class too. I think when they showed us like the picture of like a, a tardigrade or a, you know, a fly's face or something and they're like, yeah, no, everything, everything's dead when it goes in there. Right. Cause it, it can't survive the vacuum.

Adam McCombs: In most cases, uh, the kind of one of the cutting edges of scanning electron microscopy is in situ microscopy or scanning wet samples. Um, so that's majority of the time. Yeah. It's very, very dead. Um, the high vacuum, the electron beam, uh, it's, it's, it's, everything's going to be dead and probably needs to be dead before you put it in. But there's new microscopes out there that use a lot of different technologies such as variable pressure where they can actually increase the pressure in the sample chamber. Without increasing the pressure in the column and doing that, you can get just enough kind of atmosphere in there. You can kind of keep some wet samples in there. And so while cellular functions are going to be degraded, it's still going to be a little bit wet and can be a lot closer to what it looks like in real life.

Adam Mccombs: And that's just so like the, all the liquid doesn't boil off. Is that kind of the idea?

Adam McCombs: Exactly. Cause when, when you, uh, if you pull a vacuum on some water, the liquid, the water is going to, uh, either boil off or it's going to sublimate off. And that's going to, in a, in a regular electron microscope, that's going to degrade the vacuum and the water vapor will ruin your vacuum and you can't run it. So you can't normally run wet samples, but if you freeze the sample with liquid nitrogen and hold it at that temperature and put it in the vacuum, the water doesn't sublimate off as fast. You can run it or if you can using a variable pressure or environmental microscopes, you can, uh, keep it just enough pressure around the sample. You can image a wet sample.

Adam Mccombs: That's crazy.

Adam McCombs: And yeah. So what's really crazy though, is that the electron optics column still has to run at a high vacuum. So you use what's called differential pumping to through the same orifices, the electron beam travels through, um, have different stages of vacuum up that it was with a pump at each stage.

Adam Mccombs: Does something happen when it hits the barrier between the high, the high vacuum and the lower vacuum stage or.

Adam McCombs: The beam does degrade slightly, but if you're imaging biological stuff, that a slight decrease in optical performance is worth it to be able to image something that's wet and closer to real life.

Adam Mccombs: And I mean, yeah, maybe we could talk about the scale of things too. Cause it sounds like biological stuff's a little bit, a little bit bigger. Obviously not if you're doing like DNA or something, but.

Adam McCombs: Yeah. Yeah.

Adam Mccombs: But like, could you maybe compare like biological stuff to like electronic stuff? Cause my experience is electronic stuff on, on the.

Adam McCombs: Oh yeah. So the, the, the newest stuff or the newest, the newest semiconductor devices, um, are coming out around the like 14 nanometer processes or the seven nanometer processes. Um, and in biological stuff, there are details down out 14 and seven nanometers. It's this year at that point, very, very subcellular. Um, I'm talking about individual organelles, uh, mitochondria, that type of thing. Um, as you start to go up in size, um, having details in the one micron range is really common or in the sub micron range on biological stuff is really common. Um, if you're trying to look at the, say the surface of some bacteria, then you're gonna, you're probably going to want around, uh, 20 nanometer resolution on the microscope. That's kind of the minimum resolvable detail. Um, you keep going up from there. Um, it depends on what you're looking at. Uh, the microscopes don't just do a great job at high resolution. They also do a great job at high depth of field. So an optical microscope that could do 1500 times magnification on something, it's gonna be doing that 1500 times magnification. And it's pretty close to the optical limit on a really thin section on the sample of mersin oil. And it's not, it's gonna be a cross section of what you're seeing and you have no depth of almost no depth of field. So you can't look at really 3d things. Whereas an electron microscope, 1500 times mag is nothing. So you can stick something in there. You can image the surface of it. You can view it in kind of 3d and you have a really large depth of field. So they're useful for a lot more than, than just high magnification. As long as you don't need the stuff to be alive. As long as you don't need the stuff to be alive. This is true. That's, that's the, that's the big advantage of optical microscopes is it doesn't matter if the thing, or you can view things at atmosphere that are alive. You really can't do that in the microscopes. So for biological stuff, you always have to be careful that what you're seeing under an electron microscope might not be reality. You have to prepare the sample somehow to get it dried out and to get it under the beam and the way you prepare it might've changed it.

Adam Mccombs: Yeah. Let's talk about that too, because I remember seeing something about like silver coating in the past, but also maybe that was a different type of scope, microscope as well.

Adam McCombs: Yeah, actually you're, um, it's usually gold coating or gold palladium, but the samples in electron microscope generally have to be conductive. And you're going to hear me say generally a lot because there's a lot of technologies out there that can get you around certain problems and we could talk for hours on those. So I'm just going to kind of go with the basic microscope here. Okay. Um, so the, the surface, the surface has to be conductive. The electron beam is made of electrons. If you shoot an electron beam at an insulator, the electrons are going to get stuck in the insulator.

Adam Mccombs: Yeah.

Adam McCombs: Exactly. And they're going to charge up the surface of the insulator and it's going to create a static field. And that static field is going to interfere with the beam. You're not going to be able to focus the image. You're going to get bright spots and streaks, and you're not going to be able to get to a high resolution at all. So to get around that, anything that's not conductive, we generally coat in gold using a device called a sputter coater. And that sputter coater, uh, accelerates argon ions, um, towards a gold target. And the argon is a big, heavy noble gas. So that noble gas doesn't react with the gold, but it knocks the gold off the target. And then the gold travels through the argon plasma and then hits the sample and deposits on the sample in a really, really thin layer. And that layer of gold is conductive and also generates a lot of secondary electrons. So it makes really bright images or it gets the images brighter. So you don't amplify it as much. So it's less, it's not as noisier. It's not as noisy. Right.

Adam Mccombs: And you know, makes for a great Valentine's day gift. If you, if you miss Valentine's day, you know, I think I saw something on your Twitter feed about a gold coated Cheerio. Is that right?

Adam McCombs: Ah, yes. The golden Cheerio. That one might not.

Adam Mccombs: I mean, it's golden and gold, so that's good, but it's a Cheerio. Yeah.

Adam McCombs: If you, if you dried it out, actually, you could probably image it. I haven't actually put it under the microscope yet. Um, and I, I, I did that when I first got my sputter coder, um, a friend of mine and I, we went and kind of picked this thing up from lab and brought it somewhere like in the, in the back trunk of their car and immediately set it up when we got it there. And, uh, we were just, we didn't have a mic. I didn't have a mic on site there. So you're just sputter coding things cause it was cool. So like we sputter coded, um, uh, quarters and, uh, just release and just random stuff just to make it gold. Cause it was cool. And there was a bunch of students there who were interested in it. And at the end of it, well, the, there was a box of Cheerios that had spilled in the back of their car. So kids, they must be the kids. No, no, no. They, it was just this random snack they were eating. Okay. Yeah. Oh yeah. They, it was stored back there for snacking at some point. And, uh, they really liked Cheerios. So we went and grabbed a Cheerio and put it in the spider coder and became the golden Cheerio and it's still sitting on their shelf today. And that was at least a year ago and it hasn't, hasn't, hasn't done anything yet, but it's

Adam Mccombs: the golden Cheerio. That's good. So, so like when you do sputtering, like how much, how much total thickness does that usually impart? I mean, I, I'm guessing it's time dependent on how long you do it for, but.

Adam McCombs: Oh, absolutely. Uh, you, you want, you want to choose how thick a spider of a coding you do depending on what you're going for. Because if you make the coding too thick, you're, you're going to degrade the resolution of the sample cause you're going to cover up details. But at the same time, thicker coatings generally make imaging it a lot easier. And once again, it can, it can yield more secondary electrons. So less noisy image. So we're talking anywhere from, uh, 40 angstroms on the very, very low end all the way up to four or 500 angstroms. If you're really putting a lot of gold on it.

Adam Mccombs: Oh man.

Adam McCombs: But angstroms.

Adam Mccombs: Yeah.

Adam McCombs: Yeah.

Adam Mccombs: Always with, always with the frigging angstroms. It's always with the angstroms. I know. Why, why, why, why you do this? Yeah. I know. Same thing with semiconductors. It's like, I think it's just, I mean, what is it? Is it 10 angstroms to a nanometer? Is that right? Or is that backwards? That sounds right.

Adam McCombs: Why don't they just say 0.1 nanometers? You're quizzing me on units right now. I'm not sure I should be doing that live on air.

Adam Mccombs: That's, it's fine. I mean, even if we're wrong, it's just, it's just always.

Adam McCombs: It's, it's thin.

Adam Mccombs: It's a, well, I think it's just a relic of, of the old days too, because it's.

Adam McCombs: Oh, absolutely. The, the, it's funny in microscopy, you'll always see nanometers everywhere and then they'll, they'll pull the angstroms out whenever you want to. Let's talk about coatings. That's the only time you use the angstroms. Nice. Nice. Yeah. Which is, I think kind of harkens to semiconductor where you're always talking about coatings and angstroms. Yeah. I think so.

Adam Mccombs: So how much, well, what you had basically gotten this, this sputter machine in order to do these readings and stuff. Like what is, what does that entail? You said that it has an accelerating factor of the argon and stuff. Like what, what does it look like practically?

Adam McCombs: At its most basic form, it's a gold disc. You apply high voltage to that gold disc. Um, my spot, I have one sputter coder that does around 1.2 KV, another one that does around 2.5 kind of varies on the design of the coder, but it's not, it's not, it's not a massive amount of voltage in the microscopy world. Um, and you apply voltage to that gold disc. The, it ionizes a partial atmosphere of argon. So you've, you've pulled a pretty good vacuum on it. You're about down to negative two, negative three tour, or once again, 10 to negative two, 10 to negative three tour. Backfill it with the argon and then ionize the argon using that voltage. And that's pretty much it. And it can get a lot more complex and you can start to put more advanced geometries in there, start to put magnets in there and get a lot better results. But its most basic level is voltage on gold. Interesting.

Adam Mccombs: Yeah. Okay. And then, and then do you have to do anything special to the sample in order to get it to, does the gold kind of go everywhere or is it, is it targeted?

Adam McCombs: The gold covers everything in the chamber. Um, so the walls of the chamber will eventually turn gold. Um, and you, you just end up taking a chem wipe and wiping the gold layer off and you actually need to see inside of it and then feeling bad when you throw it away, but also knowing there's not enough gold on there to be worthwhile. So. Right, right, right.

Adam Mccombs: Okay. Yeah. How much does the, I mean, like, what, like, it sounds expensive if you're plating some of gold, but maybe it's not that much at the time.

Adam McCombs: It's not that much at a time at all. It uses very, very little gold. Um, but when you do have to replace the targets, it can get pretty pricey, although I'm probably never going to run out of gold on one of my targets just because I don't, I'm not using my coders 24 seven. But if you're in a facility that's running this butter coder every day, multiple times a day, you're going to have to replace targets every so often. And the, depending on the target, cause different types of targets are having more or less gold looking at anywhere from 500 to $2,000. Well, that's not, not a small, but it's also not that much in the grand scheme of things. I mean, it's a very, very thin coding of gold. So the cost per sample is still very low.

Adam Mccombs: Yeah. And I guess if you compare it against like sending something out to get coated, it probably would have been a lot more.

Adam McCombs: So yeah. And it's when you're doing this stuff, you want to be able to run it in house because doing it kind of quickly is nice. Cause you can iterate on your process better and you can keep things dry. Uh, this entire time your samples got to stay dry. If the biological sample, it's going to really, really want to suck water out of the atmosphere. So shipping something off and then getting it back can be challenging.

Adam Mccombs: And so that that's because you keep it dry by having like the, the dry ice, like you mentioned.

Adam McCombs: Oh, uh, you, you, that, the, the, the cryo SCMs that is, that's, that's one way of keeping water in the sample without having it supplement off in vacuum. Most of the time you just dry it out and just get rid of all the water period. And so you have to either store it in a dry chamber with a bunch of desiccant or store it under vacuum just to keep the water out of the sample.

Adam Mccombs: Got it. Okay. Okay.

Adam McCombs: Cause something that was, something that was once wet and dried out really wants to pull a little bit of moisture back in.

Adam Mccombs: Yeah.

Adam McCombs: So it's like a different, it's like a gradient, right? Where really dry thing and wetness in the air. Exactly. It's not, and the wetness in the air really wants to get into the really dry thing.

Adam Mccombs: Yeah. Makes sense. You've mentioned a couple of times now, so you've said SEM, you've said STM and TEM. TMA. So could you, could you define each one?

Adam McCombs: So yeah, I've really only talked about the scanning electron microscopes. And that's kind of the one that everyone usually thinks of when they see an electron microscope. That's the one that images the surface of something. The other major type of electron microscope is a transmission electron microscope. And that operates relatively similarly to a SEM, but, or in the fact that uses an electron gun and optics, but instead of scanning the surface of something, it transmits the electrons through the sample. And when you do the transmission electron microscopy, you can achieve much, much higher resolutions than you can scanning electron microscopy. The cutting edge of transmission electron microscopes is not just atomic resolution, but is quarks. Saying subatomic is a bit deceiving, but it's subatomic orbital resolution where you can see two atoms step on top of each other. And, and, and see exactly where the atoms are. And even kind of see the cloud that is the covalent bonds between the atoms. That's really tricky.

Adam Mccombs: Yeah.

Adam McCombs: No, we're, we're, we're imaging atoms now. It's crazy. And the fun part is you have to get something that's atomically thin to do that with. And like, oh, it's like the sample on a regular basis. As you're like, for like preparing the sample you're saying or what? Yeah. When you're preparing the sample, you have to get something that is one atom thick to be able to image an atomic, a single atomic layer. Wow. Yeah. It's fun. How do they do that? Uh, there's actually a lot of different methods. One of the really commonly used ones though is called electro polishing. And that's where you use, um, once again, uh, we use argon a lot because argon's magic and does a lot of great stuff, but we, we use a jet of argon that has a really high etch rate at the center and a really low etch rate at the outside. And you point that at the sample. And as you etch through the sample at one point, you go punch all the way through the sample in the center. But because that's, that's, because that's where you have the high etch rate and on the outside, you're not going to punch all the way through the sample because that's where you have the lower etch rate. So somewhere in between where you've punched through and the outside of the sample, you're going to have that really thin section.

Adam Mccombs: Okay. And, and they just kind of say, they just do the math of saying out here, it's this thickness in the middle, it's nothing. And just do like a exponential gradient or whatever it would be.

Adam McCombs: Yeah. Yeah. And you can actually just put in the microscope and look at where's nothing. Where's the outside? Let's go right next to nothing. Probably atomically thin layers there. Oh, that's cool. Yeah. That's good thinking. Yeah.

Adam Mccombs: That's it. It's interesting too, because it does seem like there's a lot of, there's a lot of theory involved, obviously, but this feels like this is where a practical, like practical machining and, you know, chemical handling and all of the stuff that you're talking about. It's just like where it runs smack into the theory stuff because theoretic or theoreticians need, need this kind of technology to verify the results, but you need to be an engineer and like, you need to try a lot of different things to get that super crazy vacuum and get the samples you're talking about.

Adam McCombs: Yeah, absolutely. It really is a hands-on meets theory. And those atomic resolution microscopes, you're not just recording image at the surface. You're actually, or those is really, really high end atomic resolution microscopes. You're actually counting individual electrons that diffract in the sample on a detector that literally counts every single electron. And then using that data can reconstruct that really high atomic resolution. Wow. And yeah, that's a really crazy mix between really advanced computer modeling and really well engineered sensors that can do that.

Adam Mccombs: Yeah. It sounds expensive.

Adam McCombs: It's at that end, everything's expensive and you start talking to millions of dollars really quickly. Okay. Okay. So you said that's the tunneling one.

Adam Mccombs: So there's one more transmission.

Adam McCombs: Oh, that's transmission. Okay. So that's transmission. And that's just, and that's one thing transmission can do. Transmission can also do things like electron diffraction patterns. You can image in really weird ways. You can actually kind of image magnetic fields of samples or really thin sections of samples. Yeah. That's awesome. One of the, one of the, my favorite projects in this is a really high automated throughput on these microscopes, on these transmission electron microscopes. And so if you can image a whole bunch of something at a really high resolution, you can do crazy things. Like you can. Oh, like averaging and stuff.

Adam Mccombs: I really love.

Adam McCombs: Like for, so like you're kind of like overlaying images and stuff or, or what? Kind of this one, uh, you can do computer tomography. So if you take a, if you take a sample, you cut it into, and this is a specific project I'm talking about. Take a sample, take a sample of brain tissue, cut it into 26,000 thin sections, load each one of those thin sections onto a reel of electron transparent tape, load that tape into a modified microscope that can automatically run this tape. Spend six months imaging that whole thing. And then you have 25,000 massive images at, uh, in, uh, and a resolution so high, you can distinguish the connection between individual neurons in this brain tissue, stack them together, run the, uh, uh, run the computer vision stuff on it. And you're able to generate a 3d map of neuron interconnections in a brain. That's bonkers. Yeah. So it is tissue scanning. I mean, dead, dead tissue though, right? Dead tissue. Not only is it dead tissue, it's tissue that has been, um, extracted from a brain. Uh, you have to, it's called fixing. You have to, uh, uh, kind of get a heavy metal into the sample to get enough contrast on the image. And use what's called osmine tetroxide for that. And osmine tetroxide job is to stain a sand stain tissue with, with heavy metals. And if you know anything about heavy metals, you know, they're not a super great thing to stain living tissue with. So the sample is very, very dead at this point. Got it. And to do that, you actually have to microwave the sample while you're staining it. So you've very thoroughly killed every cell in this thing. You just done it very carefully. So the cells don't change shape when they're killed.

Adam Mccombs: Right. You want to make sure you preserve what's connected to what, but otherwise. Exactly. Yeah. So it's like, it's almost like you're making like film or like a, you're like motion picture film. Like the, like that goes through a, a project.

Adam McCombs: Yeah. It's a, yeah. It's take, take a section, put it on the tape, image it, take another set or, and you keep doing that 26,000 times in a row.

Adam Mccombs: Wow. And then that, that basically builds, does that build width? So like, it's like, you're taking these slices and then you're basically reconstructing that slice in 3d.

Adam McCombs: Exactly. Okay. And it gives you that 3d map. Um, if you want to learn more about that project, uh, the, uh, what place that I kind of have worked on the Microsoft is doing that is the Allen Institute of brain research in Seattle. Um, I'm, you can go to their website and they'll have some stuff on that project. If you want to see more about it, it's really crazy. And it actually did image it fast enough. Uh, it's six microscopes running almost 24, seven to do that. So, oh my engineering. Yeah. It's, it's six of these massive things in parallel. That's a, that's, that's quite a legacy. Didn't Paul Allen pass away recently? I think that's his. He did, unfortunately. Yeah. Um, he, he set the Institute up so that he gave it the seed money and then it was always designed to be able to get external grants. So that's awesome.

Adam Mccombs: Yeah. I mean, that's, that's great. Wow. Okay. So, um, so that's, again, you said tunneling, so that's, or sorry, that's transmission. Transmission. So what's the STM then?

Adam McCombs: So tunneling and, um, so tunneling electron microscopes, that was the first one I ever built that's operates very differently than the ones we've talked about so far. So the, the tunneling electron microscope or the scanning tunneling microscope is what we're going to call it. So it's not the same acronym as transmission electron microscope. Um, so STM instead of TEM is, doesn't even operate at vacuum or just doesn't have to, a lot of them do. Oh, but what you do is you take a, um, the one I built at least takes a piece of tungsten. It etches an atomically sharp tip on it. Uh, just using sodium hydroxide to drain cleaner and like a five volt power supply. You can etch this atomically sharp tip in your kitchen. And then you scan that tip across the sample, monitoring the number of electrons that tunnel into the tip from the sample. So, uh, it is leveraging quantum tunneling to generate image of the sample. And if you're closer to the sample, you get more electrons that tunnel. If you're further away, you get fewer electrons that tunnel. But if you scan the tip just precisely enough, you can actually detect the differences in tunneling current for when you're above an atom and when you're not above an atom. And so if you scan it precisely enough, you can generate images of individual atoms using a device you can build for several hundred dollars on your workbench.

Adam Mccombs: I was going to ask you about this, actually. I think this is the course approach done. You'd basically have a picture of two metal plates. It looks like on Twitter. Oh yeah. Yeah. You said you can build out these parts of the hardware store, eBay and DigiKey. So that's, that's pretty crazy. So that's, that's all it takes? Seriously? Like that, that seems, that seems not right.

Adam McCombs: Yeah, I know. It's, it's, it's really funny. There's like a, uh, there's like a bell curve distribution of cost or cardio to attain a certain resolution here. It's like at the really, uh, to get like the SCM images of like several hundred thousand times magnification, it's really crazy expensive. But then you need this like $400 apparatus to image individual atoms. And it's not, it's, it's not as powerful of course as the really powerful transmission electron microscope. And you can only image stuff that doesn't oxidize. So carbon and gold pretty much, um, unless you put it in an ultra high vacuum chamber, but that gets really expensive and really finicky really quickly. Uh, so, but just to, if you just want to see individual atoms, the scanning tunneling microscope is the way to go.

Adam Mccombs: What would, so like, uh, could you give us an example? Use case, maybe, maybe for each of these, I guess you kind of already gave the TEM example, but like, and I know there's a ton, but just some examples. Yeah.

Adam McCombs: And I, I gave you the really high end examples. So I'll kind of give you the, the normal industrial or scientific examples that the majority of microscopes do. So the scanning electron microscope is whenever you want to do surface science. So you want to look at what the surface of something looks like, or in some cases, even with the surface composition is. So, uh, materials work, failure analysis, looking at how metals are sheared apart. Uh, biological stuff. You want us to look at, uh, what the surface of cells look like, maybe what cell organelles look like in 3d. Um, semiconductor work really loves to use scanning electron microscopes. They do a great job at, uh, diagnosing process stuff. So like, why didn't the transistor work? Well, you can take a transistor apart, put it under the microscope and you can, it'll probably show you what's wrong with it. Yeah.

Adam Mccombs: Yeah. The cracking it in half.

Adam McCombs: That's the cross sections are always the really useful ones. When you start looking at cross section polishing is really fun. Yeah. Um, transmission electron microscopes, um, uh, they're not as versatile as a scanning electron microscope because the samples have to be so thin, but when you need a transmission electron microscope, it's definitely the right thing for the task. The, I'd say the most common industrial application for TEMs just by number of microscopes doing this is probably asbestos actually. Um, if you want to know if there's an asbestos fiber in something, transmission electron microscopes are the way to do that. So there's hundreds of labs that do the, um, asbestos analysis. So anytime you have, if you have building material, you're tearing a building apart, you need to know if it's got asbestos in it. You're probably using a transmission electron microscope. They're also used really commonly.

Adam Mccombs: Well, they've got those really thin, those really thin fibers, right? That's like what causes all the problems.

Adam McCombs: Exactly. So the fibers are already thin enough to image under the TEM. So it's pretty easy to prepare the sample. Another common use is histology work when you need to know if some tissue's got some certain virus in it. So there's a lot of really high throughput labs in hospitals that will take tissue samples, prepare them, stain them with the osmine tritoxide and image them under the microscope really quickly just to see if there's individual viruses under there. And so the analyst is running the microscope, scanning across a large surface area and just looking for those viruses. And it would just be able to say like, yes, no, there is a virus or actually what the virus is. A lot of times I can identify it. The viruses do physically look different under the TEM because you're at the resolution where you can see that.

Adam Mccombs: Yeah. And virus shapes are insane. Like the one that's like, there's like diamond ones and weird shapes. Yeah.

Adam McCombs: Almost all those shapes have been determined by a transmission electron microscope. Oh, okay. Okay. As far as in the more of the research side of things, looking at tissues is really common. Looking at materials is really common. You can do some crazy in situ stuff where you can actually have a tensile stage that pulls an atomicity thin sample apart. And if you can record it fast enough, you can image the individual atoms moving around as the, as the break is forming. My goodness. That's crazy. There's, there's atomic resolution, environmental TEMs that can image gases and reactions happening to gases at ultra high frame rates at atomic resolution and actually watch what's happening. Those are once again, a really high end microscope, but they're crazy technology. Right. Yeah. So what about, what about STMs then? So STMs are much more specialized. They can image atomic resolution of surf of the surface of a sample, which is something that TEMs can't do. They can image through a sample, but they can't look at like the surface of the top of a sample. Um, if you want to, and they're more commonly used for really fundamentals, fundamental materials research, where you're trying to determine like, where are the atoms of copper or the atoms of iron in this, uh, or in this, uh, alloy.

Adam Mccombs: So like lattice type work almost like seeing like how, yeah, really common for that.

Adam McCombs: Um, if you've seen IBM's video and Adam or a boy in his Adam or something, I think that's the title of it. It was, it was, uh, the image of the IBM logo and the boy kind of bouncing the Adam around that was done with a scanning, tunneling microscope that could, in addition to just imaging the atoms actually manipulate individual ones. Okay. But they're, yeah, they're not as commonly used just because they're limiting what they could run, but when you can use them and it's the right application, they're very powerful.

Adam Mccombs: Yeah. And each, so it sounds like you've said video a couple of times here too. So like, what is, what is the, what is the likelihood of having multiple frames or is it always, is it normally single frame?

Adam McCombs: So scanning electron microscopes, the faster you scan an image, the less time you spend on each point in the image, the noisier the image is going to be. A slower scan yields a better image. You can kind of think of it like a camera's exposure length, but it's a little bit different physics behind it. So very commonly SEM is only single frame. And if you're trying to do really fast stuff in SEM, it's probably going to be a really grainy image. And so TEM though, you're not scanning the beam at all. You're just transmitting electrons through the sample. And so you're imaging the entire sample at once, which. Oh, okay. So it's like raster versus exposure pretty much.

Adam Mccombs: Like exactly. Same thing happens in photolithography.

Adam McCombs: Yeah, exactly. It's, it's, it's scanning the sample versus just imaging the entire thing all in one go. So TEM is much better in general. It's much better for these, these high frame rate applications because it's transmitting any image of the entire sample at once. And you can just record that. And if your sensor is sensitive enough and you have enough electrons going through the going through the sample, you can get some crazy, crazy high frame rates. And the technologies that do high speed TEM videography or imaging are simply ludicrous. Like you take one image, one, one image sensor and you deflect the, as the, whatever's happening, you, instead of just taking one image of that sensor, you take four images or 16 images. And you do that by deflecting the beam onto a different part of the image sensor for each frame you want to do. So it's almost like a, one of those like rotating mirror type of things, but not really a mirror, but using, yeah, using electron optics, once again, not glass magnetic fields and static fields. Yeah. You can, you can do many multiple times the frame rate that your sensor can do by moving it across the sensor multiple times. And so that's where we start to get into close to enough timescale resolution to resolve atomic processes happening. Okay. Um, how did you learn all this stuff?

Adam Mccombs: I mean, it sounds like you're very into it, which is amazing, but still like, yeah, what, what, where did you start? I mean, like you said, you watched the Krasna video. I think I know one of the ones was it, it wasn't the toothbrush one, right? It was earlier than that.

Adam McCombs: Uh, I think I saw them kind of all at the same time. Uh, or actually it was a, no, it was a bit, a bit earlier than the toothbrush one. Actually that did come out afterwards after I saw it originally. Um, first he built his own one. Then he got his, uh, Joel, I think it was a Joel T 300 from Sweden. Um, got it shipped to him and that kind of started it. And I would read about it here and there and kind of learn the basics. But once I got my first microscope into my shop, it, uh, it wasn't quite as described on the eBay listing. Uh, the eBay listing had it as complete, right. Which it was indeed complete, but the picture on the listing was of a photo of a microscope sitting on a pallet ready to go. Um, when I showed up, I had discovered that that photo was taken in 2006. And since then the microscope had been entirely disassembled and stored on this guy's shelf. Wow. So I showed up with a, with my dad in an F one 50 and a trailer expecting to put the microscope on the trailer. In reality, I put the frame of the microscope on the trailer and then filled the entire back of the truck up with the individual components of the microscope.

Adam Mccombs: Well, um, well, I mean, transport's a little bit easier, so that's nice. But, uh, the rest sucks.

Adam McCombs: Yeah. The assembly, uh, was interesting. There's no instruction manual for this thing on how do you build this when somebody didn't know what they were doing, took it all the way apart. Um, so there, there was a lot of trying things out, failing, um, almost catching a part on fire once blowing a whole bunch of transistors apart. The getting that thing to vacuum was a nightmare. It's got hundreds or probably that one's not hundreds. I'm probably about a hundred O-ring seals in it. Every one of which has to be perfect. Most of the O-rings were turning to dust when I got it.

Adam Mccombs: Of course. Of course. That's probably the first thing. It's like replacing caps on old electronics, right? You just go through, you replace them all. But I'm guessing they also have some very specialized ones that are not.

Adam McCombs: You, you try as hard as you can to not replace them all. If one's not disturbed, it's probably okay. So you kind of leave it there. And if the system pulls down to vacuum, just leave them there. Got it. The O-rings are usually Vyton. So the Vyton ones tend to last a really long time. Never heard of it. It's a type of rubber or what is that? Yeah. It's a, uh, ooh, I think, I believe it's silicon based. Um, it's, it's a really good O-ring material. It lasts a really long time. It's good for really high vacuums. It's good for harsh chemicals and it's good for high temperatures. It's just expensive. Got it. But it's electron microscopes. So everything's already expensive.

Adam Mccombs: So nobody cares. Yeah. Well, give us an idea of that. So like, I mean, if you don't mind saying what your first one costs, uh, even on eBay.

Adam McCombs: Yeah. No, I, um, I tend to get microscopes for well under market value. Um, my first one was $400 and it was a terrible microscope. It was the ISI Super 3A. It was sold as the microscope that you would buy from ISI in the eighties when you couldn't afford something nicer. So in the eighties, it costs about a hundred and I think it was $15,000. Um, my second microscope was a Joel 35 C, which was Joel, uh, Joel is Japanese electron optics laboratory. I do, I have a lot of their microscopes. Um, they're cool. Uh, that one was in the seventies and it was their flagship model. So it was a really high end microscope when it was new. It was a quarter million dollar microscope in, in seventies. Yeah. Um, and then I got it for free. Naturally. Really? Really? I, you, these older microscopes, um, you get for free, very common. If you know where to look and you know who to ask.

Adam Mccombs: So are you willing to divulge that information? I'm guessing at least one person listening to this is gonna be like, well, Adam, Adam's seems like he's doing some interesting things here. So.

Adam McCombs: Yeah, no, absolutely. It's, I can say you can get through universities, you can get them labs, but it's not, there's not kind of one place you go to get a free microscope. Got it. It's really a, who, you know, game. Got it. Yeah.

Adam Mccombs: Um, right. Cause you need to know when it's on the dock anyways. Right.

Adam McCombs: Yeah. Or you need to know like when they really want it gone.

Adam Mccombs: Right. So. Right. And being willing to show up with the truck and, and yeah, exactly.

Adam McCombs: Yes. Yeah. Moving the bigger ones can be a bit expensive. I mean, these things are massive and they're very sensitive and you don't want to be taking the whole thing apart. If you, if you can at all avoid it. Um, so like last, um, last week, uh, helped one of my friends, Connor out at New York. Um, he, uh, you, you may remember him as the IBM mainframe kid. Okay. I, I don't remember. I can call that anymore, but, um, so helped him get a system. And, uh, so he paid a few hundred bucks for on eBay and we showed up with a trailer and had to use a forklift to load it in the trailer. Cause it was a much bigger system. Wow. When I moved my 35 C, I had to get that microscope up and over a six inch concrete bump. So we built a very large industrial teeter totter effectively, and then used come along to ratchet it up this thing. And then very slowly let down the other side. And we had to use a lift gate truck to get it actually up off the ground and lift gate trucks. When you have a good lift gate or a good way to move them, uh, forklifts. So, or another really great option. I could talk, um, for hours about how to safely move in a microscope. And if you're going to move one, I strongly recommend you get in contact with me or on the, uh, kind of the, we have, we have the discord server called vacuum hackers and we talk about moving the stuff a lot. So if you go to, uh, vacuum hackers.com, it'll link you to the discord and, uh, well, ask us any question you want. We're, we're, we're there to help. And we want to see more people get these things. And we just don't want to see you get hurt moving it because they can be dangerous. They're big, they're incredibly top heavy and can be a pain to move sometimes. So, but what you get, it's really cool.

Adam Mccombs: What are the rates, the weights we're talking about here? Like how much do these things weigh? Like maybe use your example of the 35 C. How about that?

Adam McCombs: My 35 C is 1500 pounds. When set up for transport, when assembled, it's about, or just, just the column section is 1500 pounds, which isn't too massive, but a thousand pounds of that are above the midpoint. Oh, wow. Okay. So it's super top heavy and that's, that's most microscopes. That's what, that's what makes them dangerous. It's just how crazy top heavy they are and kind of how skinny the bases are because the scan columns are mounted up high. So, okay. So it's just the, the, the big metal parts are up above ground. Yeah. It's kind of exactly. Exactly. The transmission electron microscopes are even worse. If you really commonly push 2000 pounds just for the scan column and then additional parts, you're looking at another thousand to 1500 pounds, just in like the console you put on the side, the power supply rack, the high voltage tank. I forgot to mention the transmission electron microscopes generally run in the neighborhood of 120 to 300 KV with the biggest systems running at 3 million volts. Wow. Okay. Yeah. Yeah. They're, they're big and the x-ray shielding is insane. Um, but anyways, back to the, back to the weights, they, the smaller ones, you can actually go all the way down to like a desktop microscope. Uh, and those are out there and those you can just pick up and carry. Um, they've got all the same systems. They're just a lot smaller. So you tend to lose resolution and flexibility with those systems. But like my super three a is nowhere near as powerful as my, uh, 35 C, but it also only weighs like five or 600 pounds total. So much easier to manage, much easier to move. So you can get into this at a lot of different levels. You don't have to go for the biggest, best microscope you can get. You could find, you can get a smaller one and still do a ton of things with it. And it'll still be very powerful. Yeah. And it sounds like that's a good way to, I mean, you got to get into it somehow anyways,

Adam Mccombs: right?

Adam McCombs: Yeah. You don't really want to get, you don't necessarily want to get into one of the really, really high end systems or even the older high end systems. Cause they are much more complicated than the lower end systems. And you can do that. It'll just take you a lot longer and it'll be a lot more finicky to learn on, but there's still great microscopes.

Adam Mccombs: So when I remember when Ben was, uh, I remember Ben's videos too. He would do that. It was like, you don't get the digital capture as much as you get just the display. And so are there a lot of conversions or what?

Adam McCombs: Yeah. If you get digital capture with an old microscope, you're doing really well. The digital capture systems can be fairly expensive. Even today. Um, there's two types of capture. There's passive and there's active control. Uh, Ben's system he built is passive. It monitors where the beam is and monitors what the brightness is and puts the two together and generates an image. The higher quality systems actually control the beam and tell the beam where to go and then monitor brightness and then produce an image. And that digital beam control gives you a lot more flexibility and scan times, raster patterns, um, and a whole bunch of other things that just generally can improve your image quality.

Adam Mccombs: Yeah.

Adam McCombs: I mean, I guess do people need, is it even something that like people need or is it more of just a nicety? Unless you want to be pointing a camera at a, at a phosphor screen to record your images, you really do need digital beam control of some type. Uh, when the microscopes first came out, there is actually on all the, on almost every motor microscope, there's a Polaroid film holder that you put in front of a CRT to record the image. Yeah. Just like they did with old, uh, old tech scopes. Exactly. It's, it's, it's, it's usually the same Polaroid film holder too. So.

Adam Mccombs: Oh really?

Adam McCombs: Okay. Uh, yeah. Yeah. Uh, bit across over there. Cause I mean, it, it works a lot like a, uh, uh, oscilloscope phosphor and you scan the beam across the surface and you can change the intensity of the beam. That's what the old phosphors worked as. Um, that's not, you can, of course do that with the DSLR, but the far superior option is definitely a digital interface of some sort. And you can find systems such as, uh, I'm just going to list them off like 4pi or Windus or Orion or Sirius. Those are, those tend to be older systems, but those are the ones you kind of will generally get for free with a microscope. Um, if you don't have one, you can build your own. Um, at some, at some point I want to learn how to work with FPGAs. Um, and I know that's been talked about some, someone on this podcast before and build a really nice open source system, but I gotta, I gotta get some other projects done first, but that, that'll happen eventually. So.

Adam Mccombs: Okay. Well, maybe someone listening to was interested in looking for an FPGA project. I know there's all the people doing the ice 40 and EPC, ECP five stuff.

Adam McCombs: So I was talking to Peter about this somewhat too. So, um, hopefully this will happen at some point, just a lot of other SCM related projects in house right now. So. Yeah. That makes sense.

Adam Mccombs: Yeah. Well, and you do repairs and you do moving and stuff like that. So it seems like you're doing this as, as your business now, which seems crazy.

Adam McCombs: I am indeed. Yeah. Um, yeah, I, I kind of got my start and when I picked up my 35 C the lab I was in had a broken, uh, it's an Phillips EM 400 T or EM 410 TEM. Uh, that was built in the seventies, completely analog, just new enough to not have any tubes, just old enough to not have any computers. Um, and that, that was down and they're like, Hey, you seem to know what you're doing. Can you try to fix this? And I'm like, I've never touched a TEM before, but sure. We'll try. And that kind of slowly morphed into, into my first microscope job. And I ended up maintaining that company's microscopes for a while. They have six TEMs. Uh, when I started working there, they had five and then I purchased and rebuilt number six for them. That was a Joel 1200 EX Mark one. Uh, it's a 120 KV TEM. And it's one of my favorite microscopes out there and kind of, that was a really good start for me. Um, and since then I've kind of branched out on my own and, uh, I'll do, I'll do repairs on the microscopes, uh, consulting relocations. Um, and kind of, I like to do weird stuff with them too. So recreating old hardware, setting up weird in situ stuff. Um, we need to re-engineer an old system to make you do something special. That's always a lot of fun. What does that look like?

Adam Mccombs: So from a, from a design standpoint?

Adam McCombs: Yeah. So from, sometimes it can just be, um, electrical engineering type stuff where you're recreating an old interface card or, uh, breaking into an old or breaking into the computer system on old microscope. And other times that can be very CNC or that can be, that can be very manufacturing heavy where you're doing a lot of CNC work, uh, to produce new stages that can hold samples in weird ways, or you're, you're adapting parts of the microscope to enable like that static deflection stuff. Uh, I was talking about where you're moving the frame across different parts of the sensor. Uh, the company who makes that it's entirely custom hardware that's adapted to an old microscope. That makes sense. Yeah. Those, uh, those six TEMs that are doing that really custom, that, that really crazy serial, uh, serial brain section imaging stuff. That's entirely custom hardware, uh, adapted to a microscope built in 1984.

Adam Mccombs: So I can't imagine there's like a, much of a market for brain imaging of that very specific type. Like how many people are going to be able to get to do that? You'd be surprised. Well, I just mean like if you'd have to get all of them and all at once and then you're out of people.

Adam McCombs: Yeah. Yeah. Uh, it's actually, it's, it's growing very rapidly though. Um, it's, yeah, that, that, that, that hopefully you'll see more about that eventually, but, um, yeah, it's actually is a shockingly big market for these used microscopes. Um, the used TEM market is insane, especially with how much you can charge these systems. It's, it's, uh, very surprising how valuable they still are considering how, considering how old they are. I mean, um, systems built in the eighties still sell for 120, $150,000. Right. But the guts are the same, right? I mean, like, it sounds like there's just so much hardware, you know, like there's, there's a massive amount of hardware. If you look at a Joel 1200 EX Mark one, which was built in 1984 and you look at a Joel 1400 flash, which was built this year, uh, the scan call or the, the column itself looks almost identical and inside it is almost identical. And so the optical performance is very similar in many ways. It's just your PC control instead of analog control or micro, the, the 1200 was, was one of the first microcontroller scopes out there. A more extreme example of that though, is the Phillips microscopes. So Phillips sold to FEI, sold to Thermo Fisher. So if I call it different things, just that's why.

Adam Mccombs: Phillips doesn't know, isn't known for changing their names over and over and over again. I mean, NXP, I mean, at Nexperia, I mean, you know. I'm like, was that them? Maybe that was a different one. Anyways. Yeah. They changed a lot.

Adam McCombs: Yeah. Phillips sold their microscope division to FEI, but the, the Phillips EM series of microscopes of the EM 400, 410, 420 was built in the seventies. And before they built the microscope, they spent several million dollars on solving electron optics. They literally solved it. Like they got it. Perfect. Wow. And perfect. Perfect. Like what does that define as good as you're going to get with a thermionic electron source. Okay. So they, they, they got the maximum theoretical attainable resolution out of that column. And so they then, they just built the same column. And so they built the same column when they went to the CM series of microscopes in the eighties. And they built the same column when they went to the techni series of microscopes in the nineties and two thousands. And they built the same column, I think all the way up to 2015 or 16. And it was the exact same optics manufacturing processes. This thing changed slightly, but the geometries of the pole pieces, the shape of the coils, where the coils are, that didn't change at all. And there was parts that were for an EM 400 that you could still put on a techni T12 and it would still work just fine.

Adam Mccombs: Oh, that's nice for interoperability and repairing stuff then, huh?

Adam McCombs: Yeah. Phillip, uh, Phillips FEI Thermo Fisher, uh, has a very big philosophy of reuse in all of their systems and you can find some really interesting presentations out there on this. And so you can like use like a scan generator card from a TEM on one of their SEMs 20 years apart sometimes. Yeah. That's crazy. It's, it's, it's really crazy. And the, so when you, when you get familiar with the, with those systems, you can work on a lot of them pretty effectively.

Adam Mccombs: You keep saying the, uh, the, the scan column as well. And so that's like the vacuums pulled in there and the electron beam is shooting through it. Is that right?

Adam McCombs: Yep. And then the, uh, electromagnetic optics surround it.

Adam Mccombs: Okay. But what is that, what is the material that actually instructs that thing? Is it like a big metal column or what does it look like?

Adam McCombs: Yeah. It's, uh, it's usually an iron column. Um, iron is really good at, uh, canceling out magnetic fields. Um, and you're, you're dealing with magnetic optics here. So materials you use are dependent on what the magnetic field in that area is doing. So you'll get a lot of different, uh, like, uh, vacuum alloys. So like, um, aluminum bronze is a really common vacuum alloy, uh, stainless steel, but if you're doing the magnetic work, then you need to use iron, iron or different iron based alloys. And it's just depending on what the shape of the field is in that area. But the, you'll get the pull pieces, which are magnetic fields are. They're usually the iron alloys and then the liners or the aperture holders or anything else will generally be stainless steel or the aluminum bronze.

Adam Mccombs: And what about the, um, how, how often are you modifying the electronics? Like, are you, so it sounds like you're making some retrofits and stuff, but do you have to actually go in and like we mentioned at the beginning, like recap stuff or like, are you, are you fixing up driver boards and stuff like that as well?

Adam McCombs: Shockingly, I've had very few capacitor problems in microscopes. Um, maybe four or five times I've had a capacitor problem. Uh, they use really good capacitors in these things and usually have good cooling. So it's not like you're dealing with the cheapest of cheap capacitors that you are in older electronics from that era. Uh, resistors drifting a lot. That's a common, that's a common failure. Uh, there's a lot of really high current drivers in here. So those, those can go out and then you'll just get really, really weird stuff sometimes, especially in the old analog system. So it's a, or the old, the, the, the early digital systems where it's just boards and boards and boards of 74 series logics, uh, paired up with a bunch of op amps. Uh, those, those are fun to troubleshoot sometimes. And when you get into the newer computer systems, yeah, there's some adaptions happening to kind of modernize some of the computer interfaces.

Adam Mccombs: Yeah. Okay. And you said at the very beginning, I think you said there isn't much in terms of diagrams and, and manuals and stuff is, is that right?

Adam McCombs: Um, it varies. Okay. Philips, Philips FA thermal Fisher has simply incredible manuals. Uh, I've actually like using a stack of, uh, Philips, uh, CM 300 manuals for my, uh, stack for the microphone right now. Okay. And these things are thousands of pages thick is why I can use them for this. And I'm only using about half of half of one instrument sets. Uh, every page is fold out massive schematics, incredibly well laid out. And then half the document is just an explanation of how the circuit works. Oh, really? You could, that's nice. Yeah. You, you could almost build a Philips microscope off of the service manuals for the thing. Um, you get into Joel, Hitachi, ISI. Um, if it's in English, you're one step ahead already. Uh, the manuals are commonly non-existent. You, you will usually get schematics. They are very interesting in schematics. They are commonly wrong and they are commonly incomplete. So it can make, it can make working on certain parts of the microscope rather interesting sometimes, but you, you can always get by.

Adam Mccombs: Well, like what kind of, what's your process for that though? Like what do you have to do in order to, I mean, is it just like owning out different areas or, I mean, do you start to see patterns between different things?

Adam McCombs: How are you actually figuring these things out? Yeah, a microscope very rarely fails in such a way that the entire system is going to shut down. You can usually generate some type of, some type of beam, some type of image. It's just going to be wrong. And you kind of have to use a knowledge of the electron optics and knowledge of the physical interactions, um, to be able to figure out what it is. So let's, I'll just use one example of a common failure, which is a beam jumping around a whole bunch and a beam jumping around a whole bunch can be a number of different things. It can be an instability in electronics. It can be a failed op amp. It can be a, uh, bad resistor and RC circuit causing some crazy oscillation. And those are the electronic failures. The other thing that could be is a contamination in the scan column or, or in, or in the column itself somewhere. And that contamination, if it's an insulator, like a piece of dust gets hit with the electrons. It, what we talked about this earlier, it builds up that charge and then it discharges. And that charging and discharging can deflect the beam in the column and cause it to flicker around on screen. Oh, so yeah, there's, there's, there's, there's a, it's really, when you're troubleshooting one of these systems, it's not always electronic. It's a lot of times it's physical and you have to like know where to clean. So you got to, uh, the, the failure analysis and the troubleshooting. Yeah. Cleaning, cleaning parts of the microscopes is a topic entirely unto itself. It's a multi-stage ordeal on the surface has to be almost atomically clean when you're done with it sometimes.

Adam Mccombs: So what's the, uh, what's the highest, the highest power isopropyl you use?

Adam McCombs: 101%. A hundred, a hundred percent. I'll use that sometimes. Um, when you're using the isopropanol, it matters a lot less that it's a high concentration. It matters a lot more that it's not denatured. Uh, DNA, uh, the denaturing or the denaturing agents can always leave stuff behind. And when you leave stuff behind in a vacuum system, it's not super optimal.

Adam Mccombs: Yeah.

Adam McCombs: Um, but that, that matters. That matters a lot more and a lot less depending on where you're working on the microscope.

Adam Mccombs: Okay.

Adam McCombs: But, uh, yeah, troubleshooting them is definitely, uh, an ordeal unto itself just cause there's so many different things that can go wrong. I've had another fun, really, one really fun failure. I had was a Ram failure and a Joel 1200. Um, microscope would run fine for about 30 minutes and then just everything would change and every relay would start clicking. Every lens driver would start going crazy. The computer was just throwing, throwing up garbage on data rails, which is terrifying when that possessed machine is connected to 120 KB power supply. Right, right, right. Yeah. So the microscopes can damage themselves pretty badly, pretty often, but if you treat it right and you, and you, and you talk very gently to it and you do every single preventative maintenance item you can possibly do, it'll work most of the time. But yeah, they do fail. That was a Ram failure you said though? That was a Ram failure that caused that one. Yeah. Um, another common failure is just reseeding ICs, a whole lot of reseeding ICs. And they built everything with like very early cartridge connectors too, which if you know about early computing and cartridge connectors, uh, you took them out. It works so well all the time.

Adam Mccombs: You blow on the cartridge and then you, you know, you put it back in and see what happens, right? Just like old Nintendo's.

Adam McCombs: Yeah. Except blowing on the cartridge usually involves acetone, alcohol, and deoxid.

Adam Mccombs: That, that helps. Yeah. Yeah. Yeah. You did mention the high voltage a bunch of times. I guess we didn't talk about that. What are there, are there safety concerns that you have personally?

Adam McCombs: I mean, how do you get used to that kind of thing? Absolutely. The, the, the voltage can, can really hurt you. The saving grace here is that it's not super high current, but that doesn't mean you can't, you, you, you don't have to respect it. Um, it is something you get more comfortable with over time. Um, but it's not, it is one of the harder points to jump into at the microscopes. Uh, the other thing is. Even if you're working on the system to energize, there's a course of potential for stored voltage. And when you're testing the system, having troubleshoot it, you're having to probe thousand volt rail sometimes to look at waveforms. Yeah. And, um, I've lost many channels on oscilloscopes to microscopes. Oh no. Um, oscilloscopes in this field are kind of considered consumable sometimes. It feels really good to say, but when you're testing 120 kV systems with oscilloscopes, it's things, things, things blow up occasionally. So. And they're all, they're all handhelds or what? Uh, no, I, uh, I've got a DS, uh, 1040 Z that's, uh, been repaired a couple of times. Um, but the chassis grounding on those things doesn't matter as much. Uh, it depends on the, on the type of test ports, but it's usually when you get an arc over in a tank that you blow something up in your oscilloscope. Got it. Got it. Um, the tanks, the higher kV tanks are generally gas filled instead of oil insulated. The SEM tanks are almost always oil insulated or just completely potted. But the, the TEM tanks are such a high voltage that the general rule of thumb in high voltage land is that it will eventually kill itself and you have to be ready for that.

Adam Mccombs: Okay. I guess that's nice for having swappable components and knowing how to get things swapped out and stuff too.

Adam McCombs: Yeah. It gets really expensive when you, if you have to swap out a whole high voltage tank. So you, you really do try to repair the component that failed in the tank, which means you gotta be, you gotta be comfortable making the system safe to be able to do that. So know where to ground it, know how to short it out and know where to get rid of the stored energy. Um, you keep saying tank. What does it, what does that mean? Oh yeah. The, the high voltage tank. Um, it's literally a tank. Like it is very commonly a tank filled with oil that you dump all the high voltage electronics into. Cause at 30 KV, the Corona discharge from everything is catastrophic and things can arc over and you're just gonna blow stuff up. So you have to have a good insulator around it. So you can use oil for that, or you can use, um, SF six gas, sulfur hexafluoride. It's a really good insulator, especially at slightly above atmospheric pressure. So you pressurize the tank with that. And that's the one that, that makes you, uh, uh, speak real low, right? The, the, the, the, the, the helium, helium. And it does that, uh, it does that before it suffocates you. Right. So yeah, yeah, yeah.

Adam Mccombs: I, I know not to do it, but I'm just saying that's the one, that's the one that I think of. Yeah.

Adam McCombs: Um, and that's just the, and that's just the voltage for the electron gun. There's, um, 10 KV rails everywhere in the system. There's a lot of four or 500 volt rails and the four or 500 volt rails are just on circuits and on, on printed boards and they, they don't look like it necessarily. So you do have to be careful working around this and the very least become from working around, working around mains voltage. This isn't your five volt logic, uh, 15 volt op amp rails anymore. Right. This can really hurt you.

Adam Mccombs: We're not, we're not accelerating in electrons at 1.1.8 volts, huh? No, not really.

Adam McCombs: Uh, and unfortunately, no, uh, everything'd be way easier if you were. Yeah. Uh, then the other, the other big thing is x-rays. Okay. Um, when an x-ray is formed when an electron falls down in an energy level in an atom. Um, it turns out that hitting an atom with a whole bunch of really high energy electrons from electron gun is a really good way to get a whole lot of. Yeah. Yeah. So x-ray safety is a big concern in these instruments. The SCMs less so, um, uh, uh, uh, an SCM that's from manufacturer and doesn't have really any weird modifications done to it. It's going to be fine. It's pretty easy to stop 30 KV x-rays when you're working on though, the 100, 120, 200, 300 KV TEMs. Uh, you work with a Geiger counter and when you're bringing a system up, you always verify that you're not getting any x-ray leakage everywhere because the system is covered in lead. Yeah. Um, at the higher energies, every single panel has lead on it. So you need to be very careful. You're not getting any, um, x-ray leakage anywhere. Hmm. And that's an invisible hazard that you have to be able to detect somehow if you want to do it safely. So you have multiple Geiger counters, you said? Yeah. You, uh, you want to use, um, thin window Geiger counters. So I'm forgetting, I think it's a Ludlam model three with a 76 probe on it. I, I know I'm getting those numbers wrong, but there's a really common Ludlam counter that has a mica window on it. There's a really thin window that can detect relatively low energy x-rays and low energy x-rays. They can't travel very far through materials, but they can travel just far enough in your skin to cause cancer. So. Yep. That's, that's the problem. So you need, you need to be able to detect them. Um, but if you're running a stock SEM, you've got all the manufacturing hardware in place, all the manufacturers hardware in place, and you haven't done something like replace a blanking port or a blank port on a vacuum with a piece of aluminum, you're going to be fine. Um, TEMs is where you need to be a lot more careful. Just because the energy is involved. Yeah.

Adam Mccombs: Yeah. Exactly. That's interesting too, that you said leads on everything in those two, because that, that's just another thing that adds more weight to the moving then, huh?

Adam McCombs: Yep. It definitely is. Um, another hazard is of course the chemicals that are used in this and the sample preparation is really common to use incredibly carcinogenic chemicals. So if you're working in a lab, deinstalling one, you kind of don't lick the bottles. And in some cases don't even touch the bottles. We don't know what they are because they can't have pretty hazardous chemicals in them. I mean, what don't lick the bottle. I mean, yeah. Don't lick the bottles ever. Ever. Ever. Yeah. Yeah. No, ever. But don't, but in many cases, the stuff is designed to penetrate tissue. Right. That is its job. Right. And you don't want it penetrating the tissue. You're trying very hard to keep alive on your hands. Right.

Adam Mccombs: Yeah.

Adam McCombs: So don't touch the bottles.

Adam Mccombs: Don't be near the bottle. Yeah. I get it.

Adam McCombs: Yeah. And the other big hazard is when you're moving stuff. Once again, it's very top heavy. They can tip over and they can get moving really quickly. So you do need to be comfortable moving larger things and using rigging equipment, know how to safely strap stuff down to trailers, know how to load a trailer properly so it doesn't start fish trailing on you at 60 miles per hour down the freeway. There's a lot of really practical considerations you got to take into account if you're getting a bigger system. Have you learned any of these the hard way? I mean, like. I really haven't. I've been. I was really fortunate that my dad was kind of in construction industry. So whenever I need to. When I started out, whenever I need to move something, we'd show up to one of the storage containers and grab the rigging shelf that day. And we would just have every tool imaginable to do everything properly. And so that's set me pretty. That set me up pretty well to go into the field. I know what equipment to rent. Knowing the names of the right pieces of equipment to rent is huge. Yeah. So because a lot of times you can't just use a paddle jack to move something. It's too unwieldy. You can't get the paddle jack around a corner or the paddle jack won't fit under the instrument. So you got to come up with some other way of moving it. I fortunately, I haven't had any major incidents in moving things. I've had straps break or mounts break before. But if you're doing if you're doing a big heavy rig correctly, if one thing breaks, you have you have something else that will take over for it. Right, right.

Adam Mccombs: At least long enough to hold it up so you can go get other supports and fix it up. Right.

Adam McCombs: You don't even want long enough to hold it up. If something breaks, you want your rig to stop right then and there and just freeze and stop itself. So if you're moving up a hill or something, multiple straps, chalks, things like that, then you never want to stand on the downhill side of something.

Adam Mccombs: Yes, that's right. That's right. Yeah. Did you apprentice with anyone while you were doing this stuff or is it kind of more figured out as you go?

Adam McCombs: A lot of it's figured out as I go. I've made some friends who have been in industry longer and I can call them up and ask them questions sometimes. And we kind of partnered up on some on some larger jobs before. But this was a lot of hours sitting in very dark labs staring at phosphorus screens wondering what was going on. That's great. That's great.

Adam Mccombs: Well, Adam, this is this is quite some work you've chosen for yourself. I mean, it's it sounds like it started as an interest, but it's it's it sure as hell is interesting to me, at least as an outsider. It's crazy.

Adam McCombs: I really enjoy it. And it's it's it's it's a diverse field and the stuff you get to work on.

Adam Mccombs: Yeah.

Adam McCombs: I mean, it's not every day you get to try and troubleshoot something electronic, lay out a board and then go move several tons of equipment. That's right. That's right. It's a lot of it's a lot of different skills involved. And that's kind of enjoyable to me.

Adam Mccombs: So where can people find out more if they want to follow along with your crazy experiments and all the things you're working on?

Adam McCombs: Yeah. So the place I post most commonly is Twitter. It's at nanographs. If I'm working on a microscope, I'm probably posting what I'm doing there. So like if you go on there, you'll see last week I moved ISID S one 30 for Connor. And the week before that, it's moving on TEM. And the week for that, I was fixing some electronics on my SCM. And I just like to post progress updates, post what's going on and post a lot of cool photos of electron microscopes. And then the Discord server we set up to vacuum hackers. If you go to vacuum hackers dot com, you can join up there. We've got a lot of really active members and a lot of really interesting discussions. We do everything from electron microscopes, nuclear fusion reactors, general vacuum systems, a lot of kind of radiation work, a lot of DIY, high energy physics work. And there's always a good discussion going on there. If you've got questions about these systems, about how to repair them or about how to move them, go there. We're always willing to help. And we want to see more people have fun with these really fun toys is what they kind of come down to at the end of the day. Really fun toys that the NSA keeps an eye on. That's right. I'm not sure. Maybe. I mean, there's enough microscopes out there.

Adam Mccombs: It's probably not that big of a risk.

Adam Mccombs: You did say fusion in there. You said fusion. So that's the.

Adam McCombs: Oh, yeah. The nuclear fusion stuff. That's actually that's actually where I got my start in the high vacuum, high voltage world was on nuclear fusion stuff. There is a this Microsoft manager that set up this program for high school students in his basement to go, you know, to come learn how to do this type of stuff. So it's a vacuum chamber with a hundred KV feed through and we generate we feed it deuterium gas, ionize deuterium gas, slam it together and we do nuclear fusion and we produce neutrons and x-rays and all sorts of fun radiation. It is interesting to play with. That's awesome. Yeah. If you want to learn more about that program, that's nwnclabs.org. It's a great program for students kind of in the Washington area. So and it was a great start for me. It got me hands on with electronics and high voltage and high vacuum in a in a really real way that made it way easier to actually get started hands on on the microscopes I was working on. Yeah.

Adam Mccombs: Yeah. That's great. Well, Adam, thank you for being on. Well, I hope people check this out and take on the hobby as well. I think the hobby and the profession in your case, too. It's it's your your gig now. So that's great. Yeah.

Adam McCombs: No, it's it's it's a time for working on the microscopes. If you want to get one of these things, there's there is a lot of barriers, but it's entirely doable. And then it's entirely overcomable. So and you can start out a variety of different scales. You don't have to go for the big, massive 300 KV machine. And I recommend you don't. Right. You can go for the little desktop units or maybe even the kind of the mid range, the thousand, two thousand pound SEM. So and they're always out there. If you want one right now, you can go buy one eBay for a few thousand bucks. If you wait long enough and you touch enough people, you'll probably find one for free eventually.

Adam Mccombs: Nice. Nice.

Adam McCombs: All right. Well, thank you for joining us. Yeah. Thank you very much for having me. It's been a pleasure. Thank you.