This is not an artist’s rendering, nor a physics simulation. This device held together with hardware-store MDF and eyebolts and connected to a breadboard, is taking pictures of actual atomic structures using actual measurements. All via an 80¢ piezo buzzer? Madness.
This apparent wizardry is called a scanning tunneling microscope which takes advantage of quantum tunneling. The device brings a needle atomically close to the object to be measured (by hand), applying a small voltage (+-15V), and stopping when it starts to conduct. Depending on the distance between the tip and the target, the voltage varies and does so precisely enough to identify whether an atom is underneath or not, and by how much.
The “pictures” are not photographs like a camera might take from a standard optical microscope, however they are neither guesses nor averages. They are representations of real physical measurements of specific individual atoms as they exist on the infinitesimal area being probed. It “sees” by measuring small voltage changes. Another difference lies in the “scanning.” The probe examines atoms the way one would draw ASCII images – single pixels at a time until an entire atom was drawn. Note that the resolution – as shown in the pictures – is sub-atomic. Sizes of atoms are apparent as are the distances between them. In this they are closer related to the far more expensive Scanning Electron Microscope technology, but are 10-100x zoomier; resolving 0.00000000001m, or 0.00000000039″.
One would presume that dealing with actual atoms requires precision machining vast orders of magnitude beyond the home hobbyist but, no. Any one of us could make this at home or in our hackerspaces, for nearly free. Apparently even sharpening a tip to a single atom is, as [Dan] says “not as hard to achieve as you might think!” You take some tungsten wire and pull on it as you cut so that it shatters diagonally. There are better ways he suggests, but that method is good enough.
The ordinary piezo buzzer that is key to the measurement is chopped into quadrants with an ordinary X-Acto knife by hand. Carefully, because it is fragile, but, nothing more to it than that. There are two better and common methods but they cost hundreds of dollars, not 80 cents. It should be carefully glued since soldering heat will damage it, but, [Dan] soldered his anyway because it was easier.
The scan head uses nuts and bolts from McMaster and some scrap aluminum. It is cut no more precisely than with a hacksaw and a file, further machined by an ordinary drill press. The sample is taped down to a penny held in place with a magnet.
“Won’t vibrations utterly swamp the precision required?”, well, [Dan] lives next to both a highway and an airport and all he did was suspend it on springs and a stack of scrap hard drive magnets to clamp oscillations. This is the same technique used on the first STM built in the 1980s that reeled in a Nobel Prize in physics. What looks like the fancy and complicated machined parts are just off-cuts of steel separated by slices of o-ring to soak up some of the shakes.
Slap on a power supply and some other trivial electronics and apply some image processing from the open-source Gwyddion software and you are ready to casually inspect individual atoms.
[Dan] is not a complete slacker, he does someday plan on using a stepper motor to bring the probe within fractions of an atom to the material but for now he adjusts it by hand.
The most difficult and expensive part seems to be finding something you can actually measure. The particular type of graphite [Dan] used costs $20 for a playing card size. Gold also works if you can get it flat enough. Most metals are tricky because they require an oxygen-free (vacuum) environment. Most non-metals will not work at all because they are not conductive. And of course, it goes slowly. You are sweeping across the surface like a CRT beam on an old TV, moving and measuring fractions of an atom at a time.
This method is good enough for a couple of hours of scanning before aluminum’s thermal expansion might throw off your results.
Some perspective is in order. Half a century ago, who would have thought that viewing the fundamental building blocks of chemistry through the application of quantum mechanics was easily within the reach of any teenager with a soldering iron, basic hand tools, and a shoestring budget?
Check out [Dan]’s project page for diagrams, explanations, tips, tricks, and links.
144 thoughts on “Cheap DIY Microscope Sees Individual Atoms”
Damn. Now i want to build one.
Go do it. I kept thinking I was getting trolled, in somewhat denial that this is possible.
It’s not like it’ll sink a lot of cost and you’ll learn something and, y’know, show people pictures of your pet atoms.
Except you’re likely to have several mistakes, so be prepared to go through about ten piezo elements and a few different wire types, eh.
It’s amazing how many really high-tech instruments you can make pretty crappy versions of relatively cheaply and easily, but in retrospect it makes perfect sense — these instruments and projects were mostly invented by small teams of scientists or engineers, if not by individuals, and in settings where funds were not unlimited. So, the very first STM, PCR, FET, etc. were all things that someone with more free time than sense could feasibly put together out of things lying around the house (if your house happens to have lots of miscellaneous hardware leftover from previous projects, and a rig to dope silicon rods with melted germanium).
You know the first diodes and transistors? They look really shitty, all bits of scraggly wire and chunks of crystal bolted to bits of wood. I think something about not knowing that you cannot do something seems to help. I am always amazed by what scientists investigating early electronics and electrical principles were able to do.
Make a diode! One of the simplest and most interesting projects I’de done was making a “cat’s whisker” AM radio receiver.
Flame diodes are still pretty awesome looking.
Not a hack.
Only kidding – this thing is incredible.
Please don’t post looping images like this just to make a point.
It’s really very annoying and doesn’t add to the discussion.
Please don’t complain about other’s posts.
It’s really very annoying and doesn’t add to the discussion.
Please don’t post comments criticizing someone’s post.
It’s really very annoying and doesn’t add to the discussion.
Maximum recursion level reached.
Please clear the stack now.
Please discuss someone’s criticizing comments.
Its making me dizzy.
IF you use firefox
It was a joke, it added to the funny. funny+funny=2funny
This thread of the discussion is the funny thread.
Most annoying part of that image is we are all still using GIFs.
The reluctance of web-browsers to adopt something better (Like APNG) annoys me far more then it should.
I agree wholeheartedly.twdarkflame.
Did you know there was an official format called MNG,for animated PNG but it was dropped from firefoxbecause it ‘took too much space’ in the form of a whopping 170KB, that gives you an idea how long ago that was’
And the weird thing is that the PNG support groups hates on APNG, and the W3C also doesn’t like to officially support anything but overly complicated or outdated stuff.
And now sites are trying to fake GIF through embedded WEBM video with the controls hidden.
I wonder if one day I’ll find out why they really all hate APNG/MNG.
yeah, I was aware of the APNG/MNG split.
Coding wise I understand MNG is “more complex”. But I dought its more complex then some other browser stuff we have right now. And, as you say, websites faking it in other ways is far worse for everyones bandwidth.
Animation Shop 3 used to save its animations as MNGs. Not sure it was the same MNG though. Was RGBA at least.
In either case its a sad state for the web right now bickering over a format that could have been in a decade ago, and could have made GIFs dead by now.
Very interesting project.
Do you think though that given the technical nature of Hackaday and the scientifically savvy nature of the readership that we might be entrusted with figures in standard form?
I’m not about to count those zeros. I get enough measurements in Olympic swimming pools or elephants from the regular press.
Well, given the technical and scientifically savy nature, one would hope you would be capable of things you could count on your fingers. Well, fingers plus maybe drop your pants if you need to count the one before the decimal too.
And you don’t have to count, that’s why I drew them out. Just look at all those zeroes!
Would you prefer 10picometers or 0.1 Angstroms?
And to be clear, that refers to the technology, not this one specifically. His accuracy varied and may be off by an order of magnitude or so by the time you account for his dollarstore budget. Still, pictures of atoms.
I would say that writing it as 0.00000000001m is non sensible but then again I guess that you wanted to make an impression of something really small. Perhaps 0.01 nanometers?
I know a small atom is about 1 angstrom. So 0.1 angstrom is the easiest to put into perspective…..
For someone suffering from vision loss due to glaucoma that long string of zeros can be VERY difficult to count. I’ve got a blind spot in my central vision. So 1960 can be easily misread as 1900. After counting 3-4 of anything it gets very tedious, because I can’t see the one I just counted. I use a bamboo fondue skewer to count pins on ICs and connectors. I don’t like it, but it’s my only option other than giving up.
I suggest you get your attitude checked.
It’s usual to write eg 10e-10 for very small numbers. They call it scientific notation.
It’s usual to write it that way when you are using the numbers to calculate. I’m not calculating anything, nor are any of you (I’d hope).
If you are using the number to communicate, there are more options from which to choose. Sometimes scientific notation may be the best, sometimes units, sometimes the raw number.
Everyone has a preference and I’m sure it varies wildly. I’m actually somewhat shocked that anyone cares enough to actually take the time to complain.
Ah crap, you know what I mean.
On the off chance that you are not just trolling. Have you ever thought of marking/selecting text as a matter of keeping track of the part already counted? Most of the time you can even advance the selection block with shift+arrowkey
Personally I get a better feeling for the magnitude of a number when I see it in scientific notation, but I guess I can see your point for the dramatic effect of many leading zeros.
I’m wondering at what point the critics will complain about my “playing card size” comment. I suppose I should have said “120mm x 90mm” so that everyone had to stop and think about how large that is.
Once upon a time I had to describe the size of a power supply I was going to give someone to finish a project. I didn’t have it handy and didn’t want to think of measurements, so I just said “about 3 toasts tall, 2 toasts in area”. This began the meme of “standard metric toast.”
I later measured it in near-toast equivalents:
The point being, despite perhaps being the first time toast was used as a unit, it was immediately effective communication. And people enjoyed it. I’m not writing a textbook, I’m writing a blog post.
This is starting to look like a challenge for you in your next few blog posts: create and use random units based upon objects you have next to you at the time of the post (bonus points for photos using your ‘rulers’ like your toast/bread ones above) :-)
Hmm, odd. That should have been in reply to Matt’s “post with the toast” :-/
And here I thought that was just a fluke. :-|
It is interesting that in school everything in labs was measured in cm and meters – when appropriate. mm were almost never used unless something was less than a cm. Example, length of a rod, 3.2cm. A small screw 3mm.
Now, I virtually never see cm. Is it from the camera and telescope people using mm for lenses? Even obvious numbers like a 10x10cm PCB is written as 1000x1000mm. Is it sticking to the factor of 1000 units? Then there is the liter, which is a cubic decimeter. Is this evolving to minimum number of named units? I see steel shafting on eBay that is 8mm diameter and length 6000mm. Why not 0.6m or 60cm? Does it imply higher precision in spec’ing the length? Curioser. Maybe the Chinese manufacturers have settled on just meters and millimeters to simplify and everyone else is conforming in order to communicate clearly. We all want to avoid the Spinal Tap effect.
Because when you do your design you don’t want to switch between units – adding them on dimensions can look messy and easily gets confusing. Most people tend to pick the smallest unit they need to not use more than one or two decimal places and stick to it.
Because when you do your design you don’t want to switch between units – adding them on dimensions can look messy and easily gets confusing. Most people tend to pick the smallest unit they need to not use more than one or two decimal places and stick to it.
I really tried to reply to Charlie Springer, For some reason all my posts land on the bottom :(
There is a maximum comment depth of 5.
For me it shows his comments to be both top level
“For me it shows his comments to be both top level”
Yeah, wordpress lost it’s mind when my editor purged one of my comments. The whole block of nested conversations below it got scattered to the winds. They’re sorted according to when they’re posted, but without any context.
In between when my post got sent to purgatory but before it was deleted outright, the thread still viewed properly for any editors. Now that it’s gone I’m finally seeing what the rest of you guys are.
For example, this was a reply to you. But wordpress plops it at top tier instead without context.
Meh, just abandon this article, I don’t think it’s going to get/can be fixed.
Got it. Good point. Still the use of thousands of mm in Chinese docs seems weird. My biggest gripe with the French Metric is all the syllables and quantities that sound like each. Each is unacceptable in today’s human interface considerations.
Toast reply here too.
I’ve always wondered how the hell something so silly/funny, and yet somehow beyond all logic and reason be useful, started. Kudos Matt, kudos. For your next trick, start a meme about… Actually, on second thought… ww.http://xkcd.com/1212/. We don’t need another incident.
This is turning into an episode of How It’s Made. Thus your faux pas. Area is measured in either city blocks, Rhode Islands, or Texases.
> You can’t please everyone.
Yea you can. Just write both. 0.00000000000000000001m to impress visual readers and 0.1 nm in brackets for those of us who ACTUALLY CARE about the resolution of this scanner! I mean clearly the numbers don’t mean anything to you besides being ‘impressive’ but some of us actually know how big individual atoms are and we want to know what this awesome home hack achieved in that context.
0.00000000001m is 0.01nm, 0.00000000000000000001m is 0m. At least it is according to galculator.
Reply has broken for me too. Is it a time delay thing?
Don’t be a dick, Matt. Is this really the attitude you want to convey to your audience? Step away from the keyboard, go for a walk, and think about it.
You’re not a dick Matt. I liked your zeroes very much.
i agree with this guy, i myself sometimes have problems counting…
laser damage on my retina is so small i dont notice it…
but it DOES somewhat affect my ability to count.
funny how i can still solder okay (edit: the same)
I’m amazed at your attitude and that you are still one of the HaD staff. Good luck. You’ll need it.
You do realize my light-hearted comments are directed at a person who is literally nagging about the manner in which I wrote a number on an article read by thousands of people, because it was not suited perfectly for his medical disability, right?
If me making light of and poking a bit of fun at that makes you declare me to have an attitude problem so severe that I should be fired from Hackaday, I find you to be ridiculous.
Sometimes comments serve as textual Rorschach tests. Good-natured, humorous, friendly people will see them as light-hearted. Miserable, crotchety, cruel people take them to be mean. It reveals more about who you are than me.
In my experience: Those quick to take offense are those quick to intend it.
Matt, I can’t help but notice that your measurement toasts are not actually toast, they are just bread. How unscientific ;)
Read-again, I clearly stated they were “near-toast equivalents”. One goes to war with the army one has, not the army one wants.
At the time I may have lacked the actual conversion mechanism between near-toast equivalents and toasts.
Such is life.
Picometers, scientific notation, or even angstroms would be acceptable. Though one could argue against angstroms since they are antiquated,and not really used outside of crystallography thanks to historical reasons.
As for the accuracy of this particular instrument at this particular magnification. Someone with his skill and understanding could easily figure out the effective resolution through some simple calculations and trials. Perhaps that would be a wise next step?
Us STM jocks use angstroms and nanometers because those are the closest units to the size of atoms.
If you want to figure out resolution, HOPG’s lattice parameters are well known, you could extrapolate from there.
Angstrom is a historical measurement, convinient only at the atomic scale and not easily relatable beyond it – its not an SI.
Or perhaps “about one hundred-bilionth of a meter” gets the message actross?
Oh yeah. The metric american readership fails you, once again
I’d prefer 1×10^-10. It tells you exactly what the number is without having to count zeros. 100 picometers also good, not sure about Angstroms – sure it’s valid, but how many people know how big an Angstrom is?
My point is really that in an otherwise well written and easy to read article you have this hiccup where you have to stop, think “how small is that?” Count the zeros. Continue. It’s somewhat irksome.
What about “1e-10”?
Angström is not an ISO unit. Anyway, feel free to use it for insanely low speed units like “Angström per week”.
I love google unit conversions:
1 mile per hour = 2.70369792 × 10^15 angstroms per week.
1 attoParsec = 1.21483369 inches.
Which leads to:
1 (attoParsec per microfortnight) = 1.54283879 × 10^14 angstroms per week.
To many of us, “1×10^-10” is utterly meaningless. My education did not include advanced mathematics, so scientific notation is useless. I suspect that most people would find 1×10^-10 meaningless, but 0.00000000001m to be precise and imaginable. I might just be the only one on this forum to be willing to admit that my education was all practical, hands-on work instead of being made up of endless classroom work. With that in mind, sometimes it’s nice when the writer of a piece “dumbs it down” for people like me and the other 90% of the world who understand real numbers perfectly but don’t have the education (or even the desire) to deal with numbers that more like digital noise than useful figures.
Pretty brave of you to wade into this discussion with critics like these roaming about. You too are in a minority, but just as I wouldn’t placate anyone for a specific medical condition, I also wouldn’t write to suit your fringe illiteracy for scientific notation. This time you lucked out.
I didn’t see it as dumbing down in the slightest but you hit the nail on the head with your description of it being “imaginable.”
I see it as an extreme number having visual impact at a glance, without thought.
It’s less good at telling everyone exactly, but everyone at a glance knows roughly how many zeros there are, +- an order of magnitude. No one is using this for calculation (god rest your soul if you’d build something based on a blog post without confirming it).
To me, “1×10^-10” is an asinine way of communicating even to people who do understand it which, to be fair, is most of us. First, without superscript it’s awkward to read. Second, there’s 2 math operations built into it. Visually almost anything in scientific notation appears identical and you have to stop and think about its scale. It’s great for crunching numbers though.
This is not a proof written in math, it’s a narrative written in English. I wouldn’t say I’d always choose to display a value numerically but in this case I liked it best.
Yeah, I get 10 picometers is really really small, but it doesn’t have the impact that does drawing out what that number actually is.
Somehow this discussion is somewhat interesting to me despite the mind-numbing pedantry of some of our readers.
Brave? No, I was just in a pissy mood and felt like shooting off for no good reason. I do prefer it when people use real numbers, but I certainly don’t expect it, especially not here.
I am a writer also (not a very good one, mind you), and I know that you can’t please everyone. At first I was slightly miffed about your “fringe illiteracy” comment, then I got over myself and had a good laugh about it. Yes, you called it. I am skilled in many areas, but math is not one of them. 2+2 I can get. Heck, I even know that 1 + 1 = 10. But any time someone starts throwing in letters and funky symbols, I put my life jacket on and start doggy-paddling back to the shallow end.
Maybe I am being too pedantic. I spend a lot of time reading engineering/scientific material so it comes naturally to me. Perhaps I’ve misjudged how many people don’t understand it.
You’re the writer Matt – do whatever you feel is right for your audience. Having an excess of zeros isn’t going to stop me reading it, although I’d prefer a technical notation. Perhaps a compromise and put both?
How ’bout using a big number instead of such a small one? 6.187927354e+24 Planck Lengths = 1 Angstrom
Scientific notation or E-notaton isn’t too advanced. When you see 1*10^-10, visualize the decimal place after the 1: “1.”, then move it 10 places to the left: .͜͜0͜0͜0͜0͜0͜0͜0͜0͜0͜1
Dyslexics of the World, Untie!
Are you a child or a chauvinist? And make mine Angstroms with a twist.
Personally, I’d prefer 49.7 femtofurlongs
What’s that in light years?
Why not go for a perfectly sensible unit, e.g. 20 beard-milliseconds.
Sucks that HAD staff are no longer capable of rising above the bullshit and being mature and professional. It seems like the more HAD does to monetize the less I respect the people who deliver their content.
This is one of the greatest projects I have seen in a long time, Kudos I hope to see more like it. I think HAD should make a tag or even a section for cheap tools for your home lab or workshop. Any thoughts on a more compact name?
One thing I’ve seen on Hackaday is that, left to themselves, the trolls can really run amok. I won’t engage them head-on, because that’s what they want, but, since they kinda bring everyone down, I like to undo the damage by poking some fun back at them. I’m not a stoic, suit & tie, stone-faced expressionless kind of guy. I’m a jovial, light-hearted kind of person. I think most people take themselves too seriously and that the world is a better place if people relax a bit.
If this was a very serious topic I might have given criticism a serious reply but, this is literally someone complaining about the manner in which I wrote a number. I wouldn’t say he’s trolling, but he’s definitely being a bit smarmy and ribbing in his complaint, and he can clearly take a joke too so good on him.
I’m not sure what this has to do with Hackaday monetizing anything. Was there some decision made in writing the article that you think was dishonest or motivated by greed?
I regret that me joking around with someone insisting on “entrusting” him enough to use scientific notation has made you respect me and Hackaday less, but I’m not sure what I should do about that.
Now I want to know what the resolution of this thing is in Olympic swimming pool units.
I have no problem with all the zeros, they clearly convey the message of “freakin’ awesome resolving power” to the casual reader. If I’m going to build one of these things I’m not going to calibrate my expectations from a number given in an HaD post.
Good article, and *amazing* build!
10 pm = 1.0 E-11m an olympic pool is 50m so a TEM has a resolution of 2 E-13 Olympic pools. Or if you prefer, 0.0000000000002 olympic pools.
* STM woops!
ROFL. This is hilarious.
We are also all CLEARLY nerds for finding both significance and humor in this matter.
More than 3 zeroes (or anything else) and I’m lost. Have to count them with a pointer. I think I was in college before I found out everyone doesn’t have to do that. Angstrom is good. Just because SI didn’t like to because it didn’t fit their arbitrary pattern of factors of 1000, it fits the problem. But then I still use cycles per second (cps) instead of Hz half the time and have no feel at all for pressure in Pascals. It would be helpful to the average person if you were to calculate how long it takes the typical person to drink an Olympic swimming pool. In fact, you could call it one “Lethoa” and send to news editors on TV, Papers, and net sites so they can put time periods into more understandable units. “Wow, it took me nearly three Lethoas to finally build that house”.
old stuff…. some german guys did similar 14years ago…
Not even close.
Lots of custom machined stuff in that project. I don’t think it was ever in the same building as a hacksaw or MDF.
Seriously, what’s with all the trolls?
Yes, a team from an university, probably with more money. Can’t you get amazed sometimes by what people can do in their garage ?
And some guys did it 35 years ago and won the Nobel prize! This is still fucking awesome!
Actually, only ~25 years ago. They won in 1986. Source: http://en.wikipedia.org/wiki/List_of_Nobel_laureates_in_Physics
Used basically the same methods too.
And wow, something that was so outrageous that it won a Nobel prize 25 years ago can be made by a teenager for pennies today. SCIENCE!
To be fair I don’t think Hans was trolling. Keep in mind:
1 – Likely language barrier.
2 – Time of day (afternoon in Germany) means likely language barrier.
3 – He doesn’t say anything negative. Is similar, is old.
Scarily, thats actually more like ~30 years ago
Time continues it’s relentless advance
Derp. Yeah let’s split the difference and call it 30.
“Time continues it’s relentless advance”
If The Matrix was a person, they’d be old enough to have a driver’s license now.
Pulp Fiction would have a 4-year degree and is now allowed to drink in a bar.
The Matrix is in a sanatorium arguing with 12 Monkeys as to exactly why escape is pointless. Reloaded and Revolutions as they are now known have cast off the attitude they never really felt anyway and now hope to pass exams at a grade that will grant them work experience at an investment bank. Pulp fiction is still alive and writes poetry to Jazz vinyl played at reduced speed, returning to rehab every anniversary of the day the Kill Bill twins murdered Natural Born Killers for being illegitimate.
Long threaded bolts, check, heavy weight, check, piezo elements, check, magnets, check… drill press, check… tungsten wire, check… I think this is going to go pretty high up my TO DO list… as usual, the only missing part is “spare time”.
This is so awesome, there are almost no words to praise it.
Will he be able to also draw by flipping atoms from their lattice, through a higher voltage spike? (like Professor Moriarty (no joking on the name!) from “sixty symbols” youtube channel.
Unfortunately I don’t think this would be possible without cooling to very low temperatures in UHV, but it might be possible to draw by very lightly poking the tip into the sample surface and dragging it around…
There have been some hacks for DIY UHV chambers and extreme cooling chambers if I remember correctly. Maybe if someone combines these hacks, they could pull it off?
making an electron microscope has been stirring around in the back of my mind for a few years now. this is an entirely different approach than i was thinking but you actually have a working model. BRAVO this is utterly amazing!!
What do you call the weird optical illusion of the atoms dancing around in an oscillatory manner. Or am I the only one that notices it? I’m viewing on a Zenith 921Z LCD at 1280×1024.
Well, I wrote the thing and I was like “Huh? What optical illusion?” then scrolled up and they started dancing on me.
What is this cybermushroom you’ve delivered to me and can I go into business with you? Riches and fame.
Gaah! Now I have them too!
Why have you done this?!
Must have been the quantum effect.
Probably http://en.wikipedia.org/wiki/Grid_illusion or http://en.wikipedia.org/wiki/Illusory_motion#Gallery
Yeah, I could do this. /s
Wait… I could do this?!?
Don’t forget this one:
Atomic force microscopes can measure a bit different set of materials than STMs.
don’t forget, for finding other scientific instrument ideas
Now if you find some other sources for fine tips, you could make some of these too…
i jumped strait into the “like vacuum tubes?” mindset… now i have to go read that article…
This is a truly amazing article. A good read for anyone who thinks there’s undiscovered secrets hidden in abandoned technology. This + graphene + coherer = how to tame a horse in Minecraft.
This is pretty awesome. I’d have no use for one, but I want to build one just for the sake of seeing it running!
This has to be one of the most impressive builds I’ve ever seen, hands down. Unbelievable, or nearly so. I found myself wondering if this was a hoax, but that doesn’t appear to be the case.
Incredible job, Dan! Thank you for sharing this amazing device with the world.
Amazing work, I am guessing you used pyrolitic graphite for the top right sample pic? BRAVO!
*scoffs*. What do you think this is, kindergarten? It’s clearly not pyrolytic graphite. It’s HIGHLY-ORIENTED pyrolytic graphite (HOPG).
The diversity of our readership is neat. “Here are some atoms”, “Oh, I recognize those.”
Haha, yup, HOPG. I was quite surprised to find it digikey actually. The “scientific grade” stuff is pretty expensive.
I wonder if nonmetallic samples could be measured by coating with graphite, much like a recent post noted for electroplating 3D prints? The only problem I would worry about is how closely the graphite would bond to the substrate. If it end up acting like a filler and smooths out the interesting depth changes that wouldn’t make for as interesting a result.
Coating with metal certainly would work, but not for resolving anything remotely close to atomic scale. However, it’s possible to image thin insulators (i.e. proteins, DNA, insulating monolayers…) coated on conductive substrates, since tunneling can occur through the insulator.
This is interesting and impressive, but I can’t think of where I would need to see atoms, and if I did I don’t think I’d manage to point the thing at the right spot because where seeing atoms would be interesting is at crossovers in materials and maybe doping.
I’m impressed by the method to avoid vibration too btw, clever.
My dad was a physics professor at UCSD a few decades back. One year he decided an interesting project would be a homegrown tunneling microscope. Now, they DID do it university style with vacuum chambers and everything. But the key were these super tiny grown silicon crystals that looked like tiny phonograph needles. He had them custom made of course.
Anyway…not cheap. But this reminded me of that cool project.
actually, needs more actual actuators.
That’s what I was thinking, if he used a piezo linear actuator those things get sub-nanometer resolution which I would hope could be an improvement on his hand-adjusted method
The scanner is a piezo actuator. The coarse approach screw only needs to bring the tip and sample to within the travel range of the scanner’s Z-axis. This is around 3 um, so it’s not particularly difficult to do by hand. As soon as they come close enough together, the feedback loop kicks in and holds the tip and sample within tunneling distance.
I think he was making a joke at my over-use of the word “actual” in the lede.
Very cool design. I’ve been wanting to make one of these for a while and this might just push me over the edge to finish it off.
I also like his vibration mount, the long springs and eddy current damping seems a bit better than the ‘slate tyre on an innertube’ setup I’ve seen previously.
The probes for the first STM’s were made by smashing diamonds between metal plates then picking through the wreckage with an optical microscope for shards that looked pointy enough.
Might be how it was done for AFM’s but for STM it needs to be conductive.
sooo can you scan a chip with this? ooh ooh do a 555
Presumably to big and after all also boring, for SPM most samples are boring because feature sizes are too big.
When sharpening 0.015″ brass wire to a fine point to thread through a die on a machine I run, the technique is to put light tension on the wire whilst warming it slightly, this draws the wire out to a thin point as it fails. I’ve never had any luck cutting the wires to a point, but I can draw/pull one fairly easily. Cut wires tend to end up with burrs and won’t thread easily.
I don’t know how thin the point ends up or if its suitable or fine enough for this, but it might.
From what I know of people who’ve tried to work with Tungsten wire, it has a “feature” that normally sucks but probably helps in this situation.
Tungsten shatters into splinters.
It’s the complete opposite of brass which is soft and goopy like silly putty in comparison.
I think the reason that pulling on tungsten works is simply so that when you cut you let it shatter properly, not mash up the tip by rubbing it against the cutters through the rest of the cut.
Just a guess.
I’ll bear that in mind, so far I’ve just ran hard brass and zinc coated, but I’m working myself up to moly and tungsten wire because the brass is expensive enough as it is, though the “tungsten” wire is a alloy of copper and tungsten in my case (30/70).
I have noticed the 99% pure tungstens I have around to use with ac welding on a older transformer tig welder are brittle and shatter when cut with a shear, so odds on your correct.
I have a relative with a electron microscope, I’ll get him to look at the finish on a drawn wire one day, maybe even measure the rugosity and thickness of it. Of course I’ll ask him to specify in familial units, one of these may be handy.
And it looks like the internet is broken again… MOOOOOMMMMMM!!!!!
I don’t think anyone has mentioned the kinematic mount. Nicely done. Anyone know how the groove was cut? The hole should be a corner cube, but a circle will do in a pinch.
I cut the groove with a file.
Although a little bit more hazzle to build, three vee grooves with 120° spacing allow the additional advantage of thermal symmetry (for a makr 2 version ;-)).
BTW really impressive build!
I’m trying to picture it. Don’t you need accurate spacing for the legs? The traditional kinematic mount has the groove, hole (constained hole) and a ball on an elevation screw that rides the surface. You can be really sloppy about where you pit each leg on a device and it will fit that mount by rotation. I guess for the three grooves you can fit any triangle, so yeah. Cool!
The stiffness changes (especially the rotational stiffness) and the thermal point shifts. As long as the three vee grooves intersect each other the kinematic mount is good and the thermal center is exactly there.
There is a vast knowledge at MIT and they have a great (retro) page describing everything in detail:
Okay, one of my comments is missing. Hey Matt, What;s up with that?
Something quantum is going on in this post and it’s not those pictured atoms.
“Okay, one of my comments is missing. Hey Matt, What;s up with that?”
“Something quantum is going on in this post and it’s not those pictured atoms.”
*shrugs*. I write articles. I’ve made suggestions, but have less than zero to do with any web stuff.
My guess, is, since I know my editor didn’t like me poking fun at some of the trolls and that he deleted at least one comment I made, that all the replies that replied to it (which was like, ~30-40) are now homeless and adrift at the top level. And the wordpress comment plugin is shitting the bed trying to interpret the aftermath when things get replied to.
Abandon ship, she’s full of holes and sinking.
One additional comment, nice samples (because everyone has some) are CDs or even Bluerays, this works at least with AFMs but I’m not sure if they are conductive enough. It is very simple to get to the pit pattern, with adhesive tape.
BTW what kind of controler are you using? PI(D)? What is your line scan rate?
CDs use aluminum, which oxidizes rapidly in air. There aren’t too many metals that can STM’d in air. However, tunneling can often occur through insulating layers that are one molecule thick. A few interesting samples could be: organic-thiol self-assembled monolayers on gold, DNA, proteins, carbon nanotubes, quantum dots, etc. as long as they are deposited on an unoxidized conductive substrate. I work in a biophysics lab and have access to many of these things, and I definitely plan on trying them out.
At the moment I’m just using integral feedback. Line scan rate depends on scan size, flatness of the scanned area, number of pixels, feedback loop speed, preamp bandwidth… For micron-sized scans, ~1 Hz works well. For small atomic-resolution scans like the one above, it’s important to go fast enough (~50-100 Hz) to avoid the effects of thermal drift.
You aught to be able to use any samples that have been vacuum deposition coated carbon or gold for an electron microscope. Got one in the lab and people throwing away old samples?
Regardless , the entire comments section has made my morning , spewed coffee and all !
Oh , and hey , nice microscope , well done .
>> The sample is taped down to a penny held in place with a magnet.
The point is that coins are small, magnetic (usually) and conductive. Mounting samples to them makes it quick and easy to exchange samples.
Yeah, I’m not sure either. Maybe it was from 1943.
It’s Canadian. Our pennies are magnetic.
Almost all our (Canadian) coins are magnetic.
I forgot you were Canadian and was too lazy to scroll up. I was thinking that pennies are copper clad mild steel and then had to google it and discovered only the 1943 US penny was magnetic, all their other coins aren’t.
this thread was a reeeeaaaall, killer ! i almost peed myself while reading it to the bottom. But, hell ! who cares about angströms, picos, or even olympic swimming pools as soon as one realizes the true meaning of life is 42 ?
and this, too…
Please be kind and respectful to help make the comments section excellent. (Comment Policy)