Sunday, October 29, 2017

DIY SCANNING TUNNELING MICROSCOPE


DIY SCANNING TUNNELING ELECTRON MICROSCOPE 

[A WORK IN PROGRESS]



This is a build notebook for a scanning tunneling microscope (STM). I'll keep updating this project. I've got a few weeks until most of the parts arrive, hopefully before December 2017. Theory and design ideas will get pushed to the bottom as actual build notes are posted.




Pretty much the finished circuit-except i have to un-ground the potentiometers! They're supposed to be used as variable resistors (2 wires) not voltage dividers (3 wires). Probably the biggest thing I've learned so far.

I had some oscillation at one point.

Phase two:

  1. Learning KiCAD and inputting the schematic, then generating a nice printed circuit board layout to have manufactured: $10.
  2. Webcam or digital boroscope so I can view (magnified) the tip approaching to sample. $25 for Bluetooth boroscope. I ordered a $24 digital Blue Tooth boroscope from Amazon, I can tape a lens to it's front to add magnification. It has a built-in ring light.
  3. Rubber band to hold top and bottom tight.
  4. Silver conductive epoxy for disk wires? Solder worked, but isn't precise: $42. My soldering job on the piezo has actually turned out really nice; just in case I bought a cheap little jar of "Sciplus 400 Electrically Conductive Soldering Gun Wire in a Jar" that is an electrically conductive (low voltage) glue. You can thin it with water, or wait until it sets overnight and coat it with Super Glue to strengthen it! Apply with a toothpick.
  5. Stiffen the tip in its holder. I'll crush the IC female mount to make it tighter, and fill it up with a little solder or Sciplu 400.
  6. Beta radiation source: will it trigger a reading/tunnel?








Photo above is a hunk of pyrolytic graphite I had lying around. I think this was 2 volts, triangle wave, 0.1-3khz on both channels of the function generator. Probe tip not secure enough, bad angle, etc.


Due to weak coupling between adjacent layers, graphite has low sliding friction, which means its layers can be interfered with yielding a super lattice and hexagonal Moire pattern. The graphite actually warps during scanning.








Lunchtime scribbles: dual channel function signal generator with BNC tee-adapters feeding the X and Y of the oscilloscope and the microscope. The microscope outputs to the Z input on the back of the oscilloscope. Z gives brightness data and "paints" the picture as X and Y scan the tip. All three of my oscilloscopes have a Z and ground input on their backs! I was just given a fourth oscilloscope (a Tektronix D-10) which has a BNC input right on the front labeled "External Intensity"-nice!



PARTS AND TOOLS 
Part #s are from John Alexander's original 2003 design, numbers in brackets [xx] are my 2017 Digi-Key # substitutions:
  • C1   2.2pF       P4012A-ND [445-173540-1-ND]   1

  • C2,C3, C9-C12    100pF   P4925-ND  [399-13947-ND]  6

  • C4      1uF                      [478-5741-ND]          1

  • C5,C6,C7,C8    0.1uF        P4924-ND                   4

  • HD-DB15   Male   T815M-ND [1195-5793-ND] 1
  • HD-DB15  Female    T815F-ND    [1195-2295-ND]    1

  • J2,J3,J4    BNC   ARFX1064-ND  [ARFX1064-ND]  3

  • J5,J7    9V Bat Con         BS61-HD-ND [36-232-ND]  2

  • R3       20K      100KXTR-ND  [PPC20.0KXCT-ND]   1

  • R1,R2,R11,R12    100K      100KXTR-ND            4

  • R4                  10M           [PPCHF10MCT-ND]     1

  • R5                2.21k          [PPC2.21KXCT-ND]       1

  • R6-R9, R21-R24, R27-R30      10.0K                  10.0KXTR-ND [PPC10.0KXCT-ND]   12

  • R10,R18         4.99K               4.99KXTR-ND           2

  • R13,R19          220 Ohms         220XTR-ND         2

  • R25, R26, R31, R32      221 Ohms              [RNF14FTD221RCT-ND]    4

  • U1     LF411N      LF411CN-ND [LF411CN/NOPB-ND] 1

  • U2, U3     TL074         TL074CN [296-1777-5-ND]     2

  • SIP 0.1" Socket single connector strip for mounting U2, U3 & tip     1

  • VR1,VR2       10K                 [CT2205-ND]       2

  • VR4           20K (used 25k)        [CT2206-ND]  1

  • VR3,VR5         100K                  [CT2208-ND]   2

  • Unimorph Piezoelectric disk get cheap 20mm-24mm                     1

  • Prototyping board          Vector Proto board        1005-ND [V1005-ND]          1


  • Tungsten Wire  0.25 mm (0.0098") very thin wire for tip probe.        
  • Wire cutters with hardened jaws: snip-shattering the tungsten will ruin this "disposable" pair, get a $4 pair to sacrifice.
  • Soldering Iron and non-acid solder, flux, Sal ammoniac, sponge.
  • Conductive silver paint/syringe of silver solder if you're scared to solder to piezoelectric disk.
  • 3 fine adjusting screws and threaded jackets: violin tuners!
  • Super Glue (the gel kind in the blue bottle that doesn't run) to secure thin wires so they don't pull on piezoelectric disk, and to fill out and better secure the violin tuners..
  • 2 flat washers with holes bigger than only the ceramic portion of piezoelectric disk.
  • File or saw to groove wire relief trenches into washers.
  • Material for mechanical portion: steel, aluminum, wood?
  • Drill, hacksaw, file to machine mechanical portion.
  • Utility knife (not Xacto).

  • Oscilloscope with channels for X, Y and Z input.
  • Function signal generator (dual or two of them) for X and Y output.
  • BNC Tee connectors to split output of function generator to microscope and oscilloscope.
  • 5 cables (4 from signal generator to circuit & oscilloscope, 1 to oscilloscope z axis).
  • 2 Nine volt batteries.
  • Magnifying glass for tip adjustment.




Circuit bias power supply. Two 9 volt batteries are used to supply cleaner power which has less noise to swamp out the tiny quantum voltage signals. A red battery wire from one battery gets attached to a black wire from the other battery. Yes, that's weird, but later on in my posts you can see the same setup in the Chua Chaos Circuit that I made...and just replacing this with a single DC power supply doesn't work. Go the easy way and just use two 9v batteries. What's cool is that has capacitors right there--the original Chua circuit I built doesn't have that, but later versions did to smooth out the voltages.


This is the bottom layer: power and ground. Two more layers of circuitry get thrown on top of this!



Bottom of the microscope, featuring a sample holder. It uses one of the TL074CN op-amps (U2).




Someone needs to put the schematic into KiCAD or Eagle and generate. PCB layout! Laying three blobs of circuits onto each other will be fun.





Two more TL074CN op-amps, and an LF411N handling data from the tip. 100k resistors are anti-static, 220k resistors are anti-oscillation.




Point to point wiring! By the time it's done it'll look like a loosely wound ball of barbed wire. This is another layer of circuitry.



Now on to the actual unimorph scanner, stage mechanism, sample holder...

















The unimorph scanner: a piezoelectric disk scored into four, with a spike tip in the center. Simple! Until you try soldering all those wires to it.





I scored an "x" on the ceramic (white) part of the disk. You have to go deep and an Xacto knife isn't tough enough. Use a utility knife. Check for any continuity between quadrants: you want none!

The photo shows a test disk. Use a metal ruler to get straight cuts. It's probably best to take care and mark your "+" cut so it's as evenly split as possible: your quadrants should be equal in area.

I tried multiple times with an Xacto knife, and although it looked great, the multimeter showed conductivity between all four quadrants! Two heavy swipes with the utility knife and the quadrants were isolated from each other. Meaning: I set the multimeter to test for continuity ->|+))) and touch the probes together. The reading (or beep tone) tells you the probes are touching. If I touch the probes to a piece of metal: the same thing. If I touch it to the ceramic part of the disk: the same. If I cut the quadrants deep enough: touching two different quadrants no longer beeps. They no longer are electrically connected to each other!

The quadrants are now my -x, +x, -y, +y parts of the circuit. The metal part of the disk is (and always was) also electrically separate: it is another part of the circuit. The probe top with have yet another wire coming off it: the z-axis.


The op-amp wants input 1 and 2 to be the same. It will output the difference to try and make them the same. If you connect the output to one of the inputs with a resistor you make this device function as an amplifier. This provides feedback from the output to the input. To amplify you usually need more power! That's why there are "Added voltage" inputs at the top and bottom: to add even more power to let the op-amp operate: adding more power to try and balance out and make inputs 1 and 2 equal to each other. From my datasheet for the TL074CN I find that it's pins #4 (positive) and #11(negative) that get this "Added Voltage". Okay, now I'm finally feeling comfortable: I now know where to physically solder a resistor-what actual metal leg thingy on the op-amp. There are less than 60 components in the schematic, so this is like "2 down, 55 to go!"



So, with the op-amp trying to keep things the same the probe tip will jiggle around while responding to any difference in voltage/current: thereby "reading" the terrain of the sample-so to speak.







Sawing off the mechanism.


Almost ready for use.




Sawing off the excess tube, which will allow an extra quarter inch of tightening and travel.



Ready to be put into the holes I made in the steel plates I'm using for stages. I found a pile of these that are identical. I got about two dozen for a dollar at a Disabled American Vets (DAV) resale shop!




My three adjusting screws. The two slots were on all the plates: now I don't have to drill pass-throughs for the wiring.



I press fit them in to the holes I drilled! But I super glued them to be safe.




The camera lens is making it look out of square. I'm happy, but if I want to change it I can always use the other twenty plates I have.




I made countersunk "cones" in the bottom plate for the little rounded spike tips of the screws to fit into.




A rubber band will clamp the pieces together during operation.




Contrary to many stage designs, I'm mounting the disk scanner at the center of the top stage. Working together, the front and rear violin screws lower the scanner downwards. Used separately they lower the center of the stage the least amount possible. If that range of motion isn't enough to get the tip close to the sample-just lower using all three screws; and then lower slightly with one side's screw(s).




Line up two washers and drill two mounting screw-holes.



This is just a test piezoelectric disk, but a great way to solder them is to use an old aluminum heat sink from a computer as a table. These disks get too hot to touch within less than 2 seconds of solder! Have them on a heat sink and they stay nice and cool--and have a much smaller chance of depolarizing from the heat.




A DIP IC circuit connector mounting pin (female) is glued to the disk in a manner that it does not make any electrical contact with the disk--and does not accidentally join any of the four quadrants electrically either. Use non-conductive glue! There is the unimorph disk scanner!




Just a test disk without all wires or probe.

File a groove on the steel washer for wire pass through, and put electrical tape on both steel washers and clamp them onto the disk scanner: you just have to find two steel washers that have holes bigger than the ceramic part of the scanner (or drill them out).

The plastic washer is there for a guide to drill out (enlarge) the holes in the steel washers so they don't touch the ceramic part of the scanner.





The hole for the probe tip just has to be big enough for the tiny tip to pass through.

Test fit of scanner assembly! Bottom nuts will be removed.









Solder and Super Glue to keep the wires from moving.








Hold the wire with plyers and pull with the cutters held at a 45 degree angle to rip a sharp point.


SET UP

Signals are created by my dual channel KKMoon function generator which retails for well under a hundred dollars. It can output 20vDC peak to peak!



Both output BNCs are split with BNC Tee connectors.


These are the X and Y signals. They are both split: X to the oscilloscope channel 1 and X to the microscope circuit. Y to the oscilloscope channel 2 and Y to the microscope circuit. The "microscope circuit" in this case is basically the 4 ceramic quadrants of the piezoelectric disk. The 5th connection (the steel part of the disk) is an output to the z-axis input on the oscilloscope. This setup is called a unimorph disk scanner and it was the tremendously simple genius idea of John Alexander.




THEORY & DESIGN

In Newtonian physics light and electrons and photo-electrons travel as particles. When that hit a wall (barrier) they bounce off. In quantum physics they travel as waves. When a range hits a barrier there is a possibility (probability) that they'll tunnel through it! This microscope will measure the voltage of those electrons tunneling through the surface of the sample, giving a reading accurate within 1% of the true atomic diameter.

OP-AMPS 
There are two op-amps needed for this project: TL074CN & LF411CN. I think they'll be alright (rechecking the schematic I ordered another two more op-amps for X & Y feed outs). This will be the way I learn about op-amps, so who knows. They were cheap and the exact units originally used seemed obsoleted at first, but I think I tracked down the correct units. We'll see. STM builder Dan Berard recommended ADA4530 amps due to low input bias current (20pF vs the TL074CN 200pF)--but my whole project is based on a decades earlier design...and these new builders are using computers and not oscilloscopes. My gain will be -1. Parasitic capacitance? My whole setup will be fueled by parasitic capacitance! I'm using antiquated components, oscilloscopes and anything I find in the junk drawer. LOL!

Anyway, I've learned that part of my schematic contains an op-amp that is supposed to be inverted (along with feedback resistors running from input to output to force this component to actually amplify). I thought it was just drawn upside down...and then shorted out with itself for some reason. See: on Monday I didn't know what an op-amp (inverted or not) was, and now I'm slowly realizing what the circuits (hopefully) will do. There are like literally 19 spots on the schematic that show ground...so I'm going to have to assume (until I learn otherwise) that there will be 19 tiny little wires running to ground (the joined red and black wires of the 9V batteries). The purpose of the op-amps are to: amplify the teeny-tiny voltages/current produced by the quantum tunneling (Z axis) so that they're  strong enough to get "pushed" through the rest of the physical apparatus and large enough to be measured/displayed via the oscilloscope. The additional op-amps are to increase the drive range of the X & Y axes.

Piezoelectric disks: From my previous experiments in sonoluminescence (glowing water) I have a bunch of piezo electric buzzer discs at my disposal. The original design calls for a 4kHz 15v 24mm piezo—and I’ve got a bunch that are 20mm diameter. Just to be on the safe side I also ordered the “you might also be interested” item on Digi-Key’s website, which was a 6.3kHz (very close) but also the smaller 20mm diameter. The piezo gets scored on the ceramic side to make it into 4 quadrants—and thus 4 piezos that are fed by both sides of my dual signal function generator. So, the physical holder for this disk may be of an altered size—and the frequencies used to drive the piezo will be slightly different. Hopefully they’ll be similar enough in other properties (impedance, capacitance, etc.) to not ruin the project. The original disk had a travel of 0.16 mm/Volt in the Z axis (up and down); while I don't know the specs of most of the disks I have-the new one has a total thickness of 0.42 mm and plate thickness of 0.2 mm according to the manufacturer Murata. As you go up the scale in size the plate thickness just goes to 0.3 mm in most cases--so I'm probably OK. If not, buzzers are cheap and I can experiment (have fun) trying different ones.


TIP PROBE

The tip is made of .0098"(0.25mm) tungsten wire. There are two schools of thought for making tips: 1) Pull the wire taught and cut it. This, amazingly, gives a point that is fairly close to being only a single atom across; or 2) do the same, but then plunge the tip into sodium hydroxide with 5 volts current running through it to etch it at the meniscus, as it's eaten away at the meniscus the weight of the submerged portion of the wire will pull and break off and fall to the bottom, leaving a sharp point on the wire remaining above the liquid...then you have to remove the oxides from it with hydrofluoric acid. I don't feel like having fluoric acid in my home...it corrodes glass. sodium hydroxide is so exothermic that you need a bucket of ice around any container you're mixing it in, just like when my family makes soap: it heats up as you mix it into water and it's a big messy annoyance. So, it's the pull-and-cut method for me. Pull the wire tight, partially sever the wire and then pull it apart. You pull apart to get the point-and the cutters never touch the actual "tip" of the tip. I bought a cheap pair of cutters, and paid an extra dollar for heat treated steel ($3.99). I don't care if these get ruined, just as long as they cut/shatter the wire.



I wonder if you can also sharpen the tip the same way I sharpened some wire filaments: set up a Jacob's Ladder and blast a zillion volts through it...every spark blasting across the gap is vaporizing part of the tip, making it sharper and sharper...or making it shorter and shorter...and blacker and blacker. Probably not. I've seen etched tipped that haven't then been rinsed with hydrofluoric acid and they look like sandpaper: just atrocious. That, and melting tips during annealing have little blobs of metal at the tip.

It's sort of impossible to make nano-sized things out of bigger things, you have to grow them from smaller things, but I've seen some great results at getting 1 atom width wire tips just through pull and cut methods. The annoying thing is that some references state that once you cut the wire, it will start oxidizing anyway--if it's tungsten or copper. If all else fails platinum-iridium wire may be used, although it costs $460 for two feet of it. Those little loops we'd heat and inoculate agar in petri dishes with are almost $200 each! So, I've ordered tungsten. I'm trying to have fun. Part of the fun is fiddling around with stuff. The University of Sydney's Physics Department has some great photos and instructions of tip production and mounting. I also got cheap (yet heat treated) sacrificial wire cutters to ruin on the hard tungsten.

Now, think about this paradox: we might take special care in making sure that the tip of the probe is only a single atom, but then we expect the sample to be super-flat? No, and yet the system still works. Until proven otherwise I'll think of quantum tunneling as this: the closest atom of the probe tip will interact with the closest atom of the (flat) sample. Let's not add any make-believe problems in worrying about the "single atom probe" tip. Some commercial STMs have the sample stage provide the voltage in a way that the stage is acts as the tip electrically speaking (reverse tunneling)! 

Funny enough, one of the two men who won the Nobel Prize for inventing the first STM even published their finding that the tunneling current will just flow through the few atoms sticking out the farthest from the tip--and that the sharpest tips don't necessarily give the best results! You can get atomic resolution using a tip that is up to 10 nm in radius.

Well, I guess I could just break open a light bulb and use the tungsten filament inside.

MECHANICAL STAGES

Other changes will include the moving stage and screws used to move the stage and the tip close to the sample. Also, I don't feel the need for any complicated anti-vibration components. If vibrations become a problem I'll just clamp the whole stage/tip device into a heavy vice: mass = resistance to motion. Others have gone through some hoops using long springs and magnets, etc. to keep reduce outside vibrations. I also have a 300 lbs. home-made lead container for my gamma spectrometry setup that won't allow vibrations to reach the scanning microscope. However, with a cantilever: the longer it is, the finer the resolution of movement; and the lighter the material, the higher the resonant frequency. Annoying noises tend to be lower frequency, like trucks driving by.

Many STM builders opine about finding the "correct" material for their mechanical workings so as to match the thermal expansion coefficient of the ceramic piezoelectric disk (somewhere between 2-10). Then...they go right ahead and use aluminum, which has a coefficient of 23! Heck, steel is between 10-13, although some stainless varieties can get over 17...which is still way better than aluminum. Also, the ceramic of the piezo is mounted to a metal disk which is either brass (19) or steel! Glass is great, but plastics start at over 50 and go over 150 upwards. I've read about worries of nickel contamination and using tantalum for parts of the STM. Hmm...DIY builders sometimes focus on "fancy" problems. I'm not using a cryogenic system, I'm not even using vacuum, so I'm going to pretend I didn't read that and carry on. This is just for "fun" remember? That worry sound just like the worry about high voltage static generation that was (not) needed to build my nuclear cloud chamber. Commercially available STM tip probes have been made of nickel!

However, this doesn't mean that precision isn't necessary: Tunneling current gets cut in half for every 0.2nm the gap grows.

Kinematic Stage Mounting: From telescope parts and stat cameras to optical filters, many things that I've used before are mounted kinematically (where you get to fiddle around with thumbscrews that adjust things in 6 different directions and 5 different variables). The kinematic variables are: time, start velocity, end velocity, final velocity and acceleration...but I just want my STM to sit there nicely. A tripod is the most stable thing you can have, and having the tripod "legs" of the top part of the microscope touch the bottom sample plate with a groove, a scooped out cone and a very flat area are sometimes used as kinematic mounts.

It's weird that most optical microscopes use a geared stage to bring the sample upward to the lens...then again, you're not trying to get 1 angstrom away from the lens, so more "resolution" in movement is needed. Couldn't you just have the sample on a plate, mounted on an inverted 80-pitch thread fine adjustment screw and bring it up slowly?

An impressive amount of text about DIY STMs are devoted to the 3 screws used to raise and lower the tip stage onto the base. People have run calculations to determine what ratio their #2-56 or 80 thread bolts give them as microns lowered per screw turn, etc. I have a nice piece of furniture that actually uses #2-56 threaded screws for drawer pull attachments. These are not exotic things...unlike the trio of $9 bolts for my telescope's secondary mirror mounting! Of course, plain old ceramic light bulb fixtures use 80 tpi finely threaded screws...and some wall mounted outlets and light switches.


In the end I decided on violin tuning screws. You get four for $7 and they come with a capture nut and mechanism for mounting. Cello versions might have longer screws--but sometimes they don't.


SAMPLE

Speaking of oxidation: the sample material may oxidize too.  Highly oriented Pyrolytic Graphite (HOPG) is a great test sample to see if you have atomic resolution, but HOPG oxidizes! To clean the surface you can just stick and peel a piece of Scotch tape. This will remove the oxidized surface layer of HOPG. I've got tons of pyrolytic graphite laying around from my levitating graphite project. Graphite also has an atomic spacing of 0.255 nm, so you can use it as a calibration gauge.

Atoms, which are in the realm of seeing with an STM are about 0.3 nm in size.

Galena is a crystal mineral lead ore. I know this as the "cat's whiskers" from trench radio building, it is (one of) the crystals used in crystal radio sets. A quarter pound of this stuff goes for about $5 on eBay (2017 pricing). Drop it on the floor a couple times and image any tiny crystals that flake off! I already have some in my mineral collection. It was one of the first minerals I owned--way back in the late 1970s.

I probably won't be putting this in my vacuum chamber-although it'd be easy to convert my demo fusor to this usage. If want to image conducting metal samples you supposedly need to do it in a vacuum. I was reading about an Atomic Force Microscope, and the test sample was fragments of a DVD and Blu Ray discs. DVDs have track spacing of 740 nm and Blu Rays have spaces of 320 nm.

Of course I already knew this from some of my home built spectrometers, using discs as diffraction gratings. As I mentioned bees earlier I thought finding some conductive pollen might be interesting to try and test at some point, but very flat things are needed to set up and fine tune an STM, cool images of pollen are the realm of scanning electron microscopes.

Basically we need conducting or at least semi-conducting things (samples) to look at.

Sample Mounting: fancy store-bought STMs use silver paint to mount conductive samples to a metal sample rest. I think it's just colloidal silver...which I've made using a 9 volt battery with wires leading to water with a hunk of silver in it. Hmm...maybe I'll just try and use a thimble full of colloidal silver as my sample! Problem solved. I've also seen commercial STMs that just use a springy bent wire, like a super-cheap child's microscope slide holder to clamp the sample to the sample holder stage.


VIBRATION DAMPENING

Some small, commercial STMs are sold with stone tablets to rest on. Then the only vibrations are the eigenmodes produced by the dual signal generation and anything naturally occurring to the entire mechanical system. Eigenmode is just a term meaning: every part in a system vibrating at the same frequency: if you put a bunch of stuff in a box and throw it down a flight of stairs it'll all be vibrating at the same frequency (more or less). People who speak of "eigenmodes" speak of their solution: bungee cords and dampening the system with cinder blocks. LOL! Basically, don't try to use your STM on a moving train. Keep in mind I'm building one of these for fun, not for mass production, so eigenmodes may pose a huge design/manufacturing concern for actual commercial scientific instruments.

The lighter something is, the higher its resonant frequency. A light microscope on a heavy base.


In brief: one probe (out of two) of an oscilloscope is replaced with an extremely fine tungsten wire spike that is mounted on a modified piezoelectric disk. The ceramic of the disk is scored into 4 quadrants and wires are soldered to those four quadrants. Two signals are fed to those quadrants via a dual channel function signal generator (X and Y axes), and the results are fed back to the oscilloscope which outputs a visual image of the item being scanned. The Z axis is the height/rise and fall of the piezoelectric disk and the tungsten spike create current since the piezoelectric disk creates/responds to current. The rise and fall current involve quantum tunneling of electrons between the sample and the tungsten spike (referred to from here on as the "tip") as the tip raises and lowers (constant current mode). Locking in the Z axis in a constant height mode is also possible in some STMs by lowering the gain but increasing the scan frequency. The produced data (current) will be a function of X and Y.

So, basically you have a +9v and -9v circuit that sets up a potential, like a Geiger counter. You also have two channels of signal frequency vibrating the disk at a set rate which do something similar--but also actually help scan the microscopic tip from side to side (X and Y) and are also hooked to the oscilloscope. The Z axis then informs the oscilloscope of the "brightness" the image must be on the screen. The tuning fork of the X and Y, plus the actual current readings of the Z's up and down set up a force--and then sense any disturbance in the force. The signal function generator(s) act as a trigger. Interlaced images have drift, which is lessened when scan lines are imposed over one another, and some fancier STMs use a lock-in amplifier on the X axis for this purpose-although they output to a computer instead of a simpler oscilloscope. The exponential relationship between tip distance and tunneling current lets you infer tip to sample distance.

The function signal generator provides the bias supply source for tunneling. The image is built up in one of two ways depending on if your oscilloscope is an old "live" view one or one that has "storage" abilities. Live view (old) scopes will provide a fast scan. Storage scopes have the ability to build up scan after scan and get a finer image. I have a few oscilloscopes and am doing this for fun, so I'm starting with the older, live view scopes. I think I have an HP Agilent that has storage. Plus, I've been desperately searching for an excuse to buy a digital Rigel or Siglent digital storage oscilloscope--maybe for Christmas! Modern scopes can have different colors assigned to different variables: maybe a false-color STM? The oscilloscope must have a Z input (usually on the back with its ground screw too). Luckily I have 3 oscilloscopes with Z inputs.

No Z input? Supposedly you can connect the Z-axis output to the second channel of a dual channel oscilloscope.  Set to chop or alternate mode dual trace display options,  use alternate if available. Set the scope to display both channels. I guess you'd need a 4 channel oscilloscope for this project? Signal generator 1, Signal generator 2 and Z?

How I mentally visualize the STM (before building it at least) is like this: the opposite voltages and opposite frequencies set up electric fields that are basically static. Anything "bumping into" that field changes it slightly, which is what the output of the STM is. Just like a bat's screeching sonar: anything interfering with the sound field is "imaged" by it.

STM vs AFM vs SEC: Scanning Tunneling Microscopes (STM) use an electrical potential at the tip to enable (via close proximity with the sample) quantum tunneling. Normally, when behaving as a particle it would be impossible for electrons to move through a barrier--it would bounce off, like a particle should. However, if a particle is behaving as a wave then there is a certain probability it would tunnel through a barrier. Recently (2013) Clark & Whitney showed that as bees fly through the air they build up a static charge (electrons) that is greater than the flowers (which are literally electrically grounded) and not only do grains of pollen jump up and stick to some bees before they even land on the flower--bees can sense these electrical differences/fields and use them to extrapolate information about the flowers!

So, an STM uses an electrified tip to quantum tunnel an electron. For a clarification, here are some similar devices, of which the STM is not. A scanning electron microscope (SEC) uses a higher voltage/current to sort of spray out electrons from their tip. In an atomic force microscope (AFM) the tip actually makes physical contact with the sample, and because it does this it can also measure the firmness, conductivity, etc. of the surface--and even use its probe to move molecules around! You can scribble your name in a pile of atoms with it! With the STM you can also move atoms around via the van der Waals force, which is strong enough to make atoms cling to other atoms stuck at the end of the STM's tip! Your very own Atomic Etch-A-Sketch!

In heated metal thermal energy is enough to free electrons. However according to the laws of classical physics says that you can't do this at room temperature. Quantum physics makes a different prediction: get things close enough and with a teeny-tiny amount of current you can tunnel electrons across the gap or through a barrier. Looks like Isaac Newton had it wrong again ;)


PERCEIVED DIFFICULTIES

Simply the wiring schematic! I used to solder circuits together when I was 10 or 12 years old, but, just like my ability to read sheet music: I've gotten very rusty. Luckily, there are only 49 parts to the electrical side of this project, and a bunch of those parts are just BNC connectors, plugs and the scanning disk. I already own the required dual signal generator and a pile of oscilloscopes.

The rest will just be fine-tuning. I know I can fine tune things. My sonoluminescence project involved fine tuning the signal generator to the piezoelectric transducer(s) while blowing bubbles with a tube with my mouth while surrounded by water and a rats nest of wires leading to a power supply in complete darkness! I think this project, once the parts arrive, will be two days of building and a month of tuning and fiddling.

According to "the internet" bar and rod shapes piezoelectric elements are rare and expensive...according to the hardware store  up the street they're cheap and available for oven and bbq grill ignitors. Cheap cigarette lighters are also starting to use piezoelectric elements which can be ripped out with a little effort.

Grow my own piezo crystals: Plus, I have Potassium Sodium Tartrate aka "Rochelle Salt", so I can easily grow my very own piezo crystals (just add to warm but not boiling water). I wonder if I could grow a piezoelectric crystal around a piece of tungsten wire? An integrated piezo tip! I know sort of the reverse has worked for STM: metal spike with a tiny diamond shard glued to it, like a hi-fi stereo turntable needle. The crystal could be the spike and the current carrier and the piezo!




When doing the sonoluminescence project I learned, yet again, that you don’t have to follow the recipe online as gospel. I used ZERO inductance and it worked fine. When I started researching a nuclear cloud chamber everything online had these high frequency (static) electricity generators running around them. Then someone used a balloon rubbed on a piece of wool for static. Then I decided that was all stupid and just filled an aquarium with 99% alcohol, dry ice and a radioactive sample and got the same exact results. Keep in mind:  this is while other people were online trying to design super expensive power supplies, thousand dollar Van de Graff generators and ultra-high voltage step-up boosters to create static fields around their chambers—all totally unnecessary! Then everyone was running around trying to buy old film projectors...because some genius posted on the internet that he used an old film projector to provide a bright light to see into the alcohol fog in the chamber! I JUST USED A BRIGHT FLASHLIGHT!!! 

Think of that: internet post after post about searching for old film projectors...because some dude did it once and posted it--when all that was needed was LIGHT. Heck, I've used my cloud chamber with no flashlight in a normally lighted room and it was fine. The conclusion: not every part in every project is necessarily needed and people online get caught up in following directions that have extra, unneeded steps sometimes. If you ask, “what does this part actually do”, and the answer is “nothing I really need”, then you can design around it. Also, if you scrape things together cheaply like I do, some parts are used just because "I found it in the junk drawer and it was free and faster than driving to the hardware store for a normal, easily obtained item." Why did I use a fancy, heavy stereo speaker magnet to weigh down that item while I was gluing it together? Because the only heavy brick I had was outside in the rain...use anything that's heavy to weigh it down--don't go and buy a stereo speaker magnet...it's just a weight! Yet, gillions of people reading online will order the stereo speaker magnet...to use as a simple weight to hold something down while the glue dries.


INSPIRATION

Most of this is inspired by an old project by John Alexander. When referring to the "original" design--it's Mr. Alexander's design I'm speaking of. His finalized design schematic (at least the one I'm using) was being worked on in May of 2000, and completed in 2004 and posted on a now defunct GeoCities personal website. There are points where I diverge from his design, but only due to practical reasons (obsolete parts that aren't available anymore, etc.). Also, so many of my jerky little projects have used theories and even physical parts of an STM already it was neat that I was familiar with the concepts. Mr. Alexander created his "free simple STM" project to inspire students to make their own. His project is only available on internet archives-hopefully my attempt ends in success, and/or inspires even more people to build their own.





DESIGN CHANGES SO FAR

Here are some divergences that already have happened in the early planning/ordering stage:

Some of the components, such as the 100k Ohm and 2M Ohm resistors, are only available now in a 2% tolerance instead of 1%. Also, if a 2% resistor had a minimum order of 1, and the 1% had a minimum order of 4,000 units, I went with the 2%. I don’t need to pay $150 for an extra 3,999 resistors.

Some of the resistors that were originally used were flat SMD (surface mount) tiny chips instead of regular cylindrical (axial) resistors—I went with the larger axial versions.