Creating Ultra-Fine Details in Titanium - 20 Micron Resolution
Вставка
- Опубліковано 21 лип 2024
- Need PCBs, CNC machined or 3D printed parts? Check out PCBWay: www.pcbway.com/
Today's video is a continuation of my previous one on photolithography. Using a custom-built apparatus that projects a mask onto the photoresist through a lens system, I can achieve structures in titanium on glass with a resolution of 20 micrometers through subsequent sputtering.
Join my Patreon and support my projects! Your contribution means the world to me and helps bring my ideas to life. I truly appreciate your support! / advancedtinkering
Optics Simulator: phydemo.app/ray-optics/simulator
Test Patterns used in the Video: www.bealecorner.org/red/test-...
Applied Science: @AppliedScience
Huygens Optics: @HuygensOptics
Breaking Taps: @BreakingTaps
Music:
Royalty Free Music from Tunetank.com
Track: Dancing Funky by TrendingAudio
tunetank.com/track/3619-danci... - Наука та технологія
Soon he will start making processors in his basement.
And it will be awesome.
And if he makes more res processors they might be much more resistant to emf, jamming etc if designed for it. Über Cool whichever way this goes.
@@dripwastaken7884 you’ll know you’re in another universe when he casually drops something like “I wanted to get a closer look so I fired up the Scanning Electron Microscope I made for an earlier project…”
Intaglio.
That was actually my first thought when I saw the thumbnail and took a quick look at the video. People make their own processors on breadboards and PCBs, why not like this? He could beat Intel and AMD to glass substrate processors, lol
Awesome! 20um puts you at about 1968 for the technology node, only a couple of shrinks away from the Z80 in 1974/6um 😄
Nice description! Still good enough for landing people on the moon…😅
Or he’s at 1870s microfilm and a few more shrinks away from 1925 microdot
Never thought of it that way. Nice comparison! :D
You and Applied Science have some of the most sophisticated machines I can ever imagine at home(Or studio maybe)!
Haha, thanks! But I'm sure Applied Science has more interesting gadgets in his workshop. And Breaking Taps is also quite impressive with his electron microscope. And his vacuum chamber is very fancy too!
Check out @HuygensOptics
@@AdvancedTinkering was just coming to mention Mr.Taps' SEM, and don't forget Nile's incredible lab.
@@AdvancedTinkering Thought Emporium could probably use your help in their project to make their own electrodes for their “rat brains play doom” project
@@JinKee I think he switched to commercially available electrodes. And Applied Science made some for him a while ago.
I'm a photographer and have used the lens you are currently using. Its cheap, but not great quality, and is known for soft images. Modern day lenses are a lot sharper. Also a bigger aperture (low number) is great, but most lenses suffer from edge sharpness in photo's. Getting a modern F2.8 and stopping down to F4 or F5.6 would fix most edge sharpness problems tho :)
Thanks for the advice! I will definitely try a different lens as long as it is not too expensive. Do you have any suggestions for a modern F2.8 lens with a manual focus? I noticed that many lenses made for digital cameras do not actually change the focus when they are not connected to the camera.
@@AdvancedTinkering Modern, not too expensive and F2.8 is a hard combo, especially if you want a zoom lens. as for manual focus, I personally shoot pentax and on there lenses physically move when turning the focus ring, so thats not a problem. A bigger problem is most modern lenses don't have a diaphragm ring anymore. You could try looking for a sigma 17-50 OS HSM F2.8. it doesn't have a diaphragm ring so thats a downside, although on at least the pentax version you can manually close up the diaphragm using the little lever on the back (not to exact numbers but that doesn't matter). Not sure if thats available on nikon/canon as I don't use that myself.
Alternatively you could go fixed focal distance, those tend to be a lot sharper even for older models, and relatively cheap. A manual focus pentax-m 50mm F1.7 could set you back as little as 15-25 euro.
Maybe try „Nikon AF-D 50mm f/1.8“ lens. With „bayonet F adapter“ it can fit in your m42 ring. This leans is modern with mechanical diaphragm. Although, you can try to go to local store and buy this gears, and when nothing happens or you achieve bad results, return in few days without big money losses.
@@AdvancedTinkering maybe a nifty fifty would be good for you. canon 50mm f2.8 :) its really sharp, but its a prime stuck at 50mm so not sure if it'll work for you
@@AdvancedTinkering It isn't a 2.8, but the Mamiya 7 rangefinder lenses are still unbeaten for lines/mm.
Awesome results, congrats! The failed parameter graveyard at 9:19 is very relatable 💀 Can't wait to see the next version, 10um should totally be doable without any major changes I think!
(also the sponsor segue made me actually lol, very nice 😄)
There is one thing: if you use a photographic lens and want to focus violet light (I am not even talking about UV, here), the lenses will never be quite as well-corrected as they are for visible light. Most photo lenses will also rapidly become opaque below 400 nm, because the flint elements do not transmit much light anymore; also most optical cements will fluoresc below 350 nm.
I would suggest using a Schneider Kreuznach Componon-S 5.6/150, which is a symmetric repro lens. It has been designed with violet/UV applications in mind.
I had a look at some measurements and at 360 nm, the lens still has a tolerable transmission of 33%, but below 340 nm, less than 10% of light comes through.
as the light source is almost monochromatic, there will be practically no chromatic aberration. as for the spherical aberration, if you stop down reasonably, they will be practically gone. but these are thick lenses and are optimized at the focal plane. a microscope objective mounted reverse may give somewhat more uniform field focus.
@@janami-dharmam The chromatic aberrations are not a concern with a narrow band LED light source, but spherical aberration, astigmatism and, to a much lesser degree, distortion and field flatness are. However, even a single plano convex quartz lens from thorlabs at the geometry shown here may just do fine for a small field.
@@kaibroeking9968 you are right but getting a flat field of focus close to a thin lens may be demanding. many of the camera lenses work close to the theoretical limit, I am told. But they are optimized for the focal plane for a particular distance.
Can I recommend one video for the Advanced Tinkering to watch? "Poor Man's Clean Room" by Tech Ingredients goes through and talks about how to properly clean lenses (not saying you did anything wrong - just a lot of good information contained within the video). The Tech Ingredients guy goes in-depth on how to clean lenses properly.
I’m a big fan of Tech Ingredients channel. So, I have seen it. But thanks for the suggestion. I’m always hoping to discover new and great channels.
the reason is that the led is not an isotropic source, meaning it project most of its light forward, and less and less as the angle increase. In addition, mask plane is, well, a plane, while the surface of costant power is a sphere, meaning you have a drop in power intensity on the mask plane proportional to the square of the distance from the center. To have a homogeneous illumination the industrial standard is a couple of microlens arrays, coupled with a collimator.
Which is probably reasonably easy to get - along with a high-power 395nm/405nm multi-LED source - as parts from resin 3d printers. They have the same issue of needing collimated UV at completely uniform intensity across the LCD/resin tray plane, which is probably about the same size as the masks you're using. Putting a Fresnel lens above that, parallel, to focus the rays in towards the lens should work?
And, just to expand on the non-isotropy (and encourage UA-cam to show this comment more prominently!), this is why most old-school LEDs (like you might see on your TV remote, or a power indicator) have their distinctive plastic cylinder with hemispherical tip shape. The aggressive curvature of the hemisphere helps produce a bunch of internal reflections for the produced light, which is a pretty cheap way to spread it over a wider angle.
That is so cool you’re creating titanium, glass, microfiche
How good is Huygen 👌
Bildung mit dem Lorenz👍
Das ist das >_Lorenz? lol
7:52 i very highly recommend using a high quality coffee filter for this step and NOT a paper towel! high quality paper coffee filters are designed to absorb moisture and not leave behind any particles of lint or broken fibers like a paper towel might. coffee filters with 90% IPA is the go to for all computer builders for cleaning lapped surfaces since the dawn of time lol. i can get the nastiest, crustiest heat spreader to literally shine like a mirror with just a coffee filter and some 90% propanol
Unfortunately, I didn't have any lint-free cloths, but using coffee filters is a great idea. I'll give that a try. Thanks for the tip!
kimtech wipes %100. paper towels leave a lot of unwanted fibers
Can't believe I already lost the game just from the thumbnail. But now so have you.
Danm
I'm not sure how I got here but thanks youtube, this was a great video.
I'm glad you liked it!
This is next level microfilm! Just imagine having Ti microfilm that could last longer than the 500 years at best that current microfilm could last
Seeing videos like this kind of excites me for the future. Who knows, maybe things will be better with people doing projects like this in home labs.
this is the kind of thing you'd fill a time capsule with, awesome
this work just got you a new patreon; well done!
Thank you very much! I greatly appreciate it!
Videos like this make me giddy with excitement, I can't wait for you to get some transistors on a wafer done!
Love the Shrek script text :). Also, it's amazing how much detail you can get out of the photoresist! I would totally expect it to peel off big chunks of the text!
About that hotspot, there are flashlights there you can set the focal point to see the chip surface perfectly projected to the wall^^
edit: but that is only one part of LED hotspot (look at the datasheet) they have 120° emit angle but only 70% of that light 70-90° of the beam and in center the most, you need a lens^^
Yes, during testing different light sources I also tried flashlights and noticed that phenomenon.
right, you can even make out the connecting wires (they are also micron sized); even a cheap plastic lens can do wonders
Well done my friend!! I’m excited to see where you take this 👌
Thanks! I really hope to be able to make it even smaller at some point.
@@AdvancedTinkering Bet that, dude!!😎 I’m sure you’ll be able to make it happen. Do you have anything in mind for that kind of fidelity? You aren’t doing the human neuron growth plate thing like “the thought emporium” is doing, are you?🤔
@@boiwithskillz My current goal is simply to create structures as small as possible. Producing a microdot would be fantastic. I've also thought about creating a rudimentary OLED display, but that's still far in the future. I'm not working on a neuron growth plate; I leave biology to Thought Emporium :D
@@AdvancedTinkering gotcha, I can get behind a goal as simple as going smaller 🫡 I thought maybe you were providing plates to TTE kinda like how “Applied Science” did a while back. You DO have some awesome equipment after all! 😁 I’m fascinated by microstructures myself. If you can pull of some sort of ‘Macro’ MEMS device, that would be freaking sweet. I love the idea of OLED as well, but as you said… a bit down the line 😊 I look forward to your endeavor in going smaller, bro!!! Keep at it my guy
All my preferred channels are referencing each other! Awesome! So much knowledge accessible to everyone!
Pretty cool stuff! Keep in mind that depth of field and sharpness are not the same, every lens has a sweet spot, often around 3-5 f-stops down from fully open aperture. To small aperture introduces diffraction, and with your flat samples deep focus is not needed. Paint focus points in the corners to check that your focal plane is level with your surface, you can add set-screws to your holder to level it. good luck!
Use the Raleigh idea, stop the central rays with a black stop and collect the marginal rays.
Just incredible, awesome work
Quite impressive!
this is so freaking cool man. the victory dance is well deserved! bravo on an awesome project.
Phenomenal work! Very impressive!
Thank you!
really amazing!
Very nice results. Commenting for the algorithm!
This is such an amazing process. The time and effort is outstanding ✌️
Thank you!
Very very cool!
Excellent choice of text. :D
Thanks!
Thank you very much for the generous gift!
Another great video, keep up the good work.
Thanks! I'm glad you liked it!
Nice Video!
Congratulations!
Thank you!
Fascinating 🙂. Thank you 👍🇳🇱
Wahnsinn, super cool! Greetings from Austria, keep doing what you're doing :)
Danke dir!
You're the man! 😎
dude, im speechless
TTArtisan makes a f0.95 35mm lens and a f0.095 50mm lens in case you want to try that out (they are both around 200€) if you want to try that out. Also zoom lenses are known to be less good with diffraction than fixed focal length lenses. Althought you might wanna test with a well known 50mm prime lens at f1.4 or f1.2 since they are known to be razor sharp.
Pretty cool, did you try using DLP displays? Now days they are more common available, since they are used in 3d printers, also the resolution is reasonable high
Awesome!
No way, exactly what I need for a project, you are a true hero!
I hope you are succesful with your project!
@@AdvancedTinkering Thank you very much, wish you lots of success too!
Printed circuits on glass has infinite possibilities.
Great video. For the light source, you could try and get your hands on a deuterium lamp. Maybe those would work better as a point source.
Just subbed btw 😁
Have you yet considered making printed digitizer screens for custom fitting any non-touch display?
The transparent tracings on said digitizers use a very thin lithographic layer of Indium Nitride.
Incidentally, that same material is used in IR blocking filters.
In other words, if you view a digitizer and/or IR cutoff filter with a wideband digital camera and an IR illuminator, you can see the etchings in the digitizer. The IR blocking filters look opaque as well.
Footnote: removing the IR blocking filter from any electronic camera allows the detector to also image IR.
In other words, any electronic camera can be easily modified into a wider band detector. All regular digital cameras have this InN filter built in stock OEM.
Very simple to modify.
It also works rather well within the UV regime below 400nm as well.
Happy tinkering.
You can substitute the fresnel lens with a diffuser from a screen. Since you are using a camera lens, you shouldn't need the collimated light.
Though you need to have the transparency and the last diffuser layer at the same plane, ideally at the same surface
nice project - have you looked at the light path in an old-style overhead projector - because this is essentially what you created here. There also the Fresnel lens projects the "mask" down into the objective lens. But there you will also find the condenser optics which was used to create the equivalent to a point source from the filament of the halogen light bulb - the same type of optics should be suitable for an LED of the type you have as well.
At this point you might want to consider that even with a perfect masking process, light projection will eventually leave aberrations on very sharp and tiny details such as those of typography. They will appear kind of rounded. Lithography processes that relied on light projection eventually encountered this very problem, and their solution was to create special typographic fonts that would mitigate light aberration to increase sharpness in very minute details.
I find your project fascinating, similar to the process that was created for microfilm documents which catches my interest for...reasons.
I have read many microfiche documents and I suspect that final quality is close to the original.
@@janami-dharmam It seems so! That on itself is incredible!
Probably surface tension from photoresist+titanium. Great video. Make a pocket version. I believe in you.
you should look at how optics are utilized for silicon wafer exposure in chip manufacturing..
start with a collimated light source, pass that through your image, and then rather than directly focus it onto the target, bounce it through a series of concave first surface mirrors. This will further condense the light, and therefore your pattern. The real bonuses here, is you save space because you can bounce the light back, and fourth, getting smaller, and more concentrated with each bounce, and therefor can build an enclosure to completely isolate the light source from all other sources, and any non-collumated light can be absorbed by the mat black interior of the enclosure. The other additional bonuses are that your developing stage can also be completely isolated from other light sources, and rather than adjust the target, you can adjust the mirrors to really dial the image in.
Another point i would like to make, is using a single LED like that isn't the best approach, as they are made of several smaller LED's, and a phoshorus gel is applied over them. The best way is to start with a clean source of light, but if you can't, multiple sources of light, so many, as to swamp the the collumator, and even out the overall intensity. This will also have an added bonus of cutting down the exposure time, as more light is able to reach the target at a higher intensity.
You may also look into photoresist that work with higher wavelengths, but that may be beyond the scope of this video.
Awesome results. I wonder though if the contrast in your setup is actually limited by diffraction. Maybe doing some tests on the optics would be interesting.
Thanks! Yes, I will definitely do more tests.
Neat!
I would have gone for an M42 or T-mount for the lens so I could try different lenses, but have a lens mount that was relatively easy to cobble-together. Zooms from the manual-focus era aren't that great for sharpness, so the ability to just swap-in a prime of a different focal length instead of zooming might improve performance.
Also you need a 'ground-glass' (which might actually be acid etched) for focus. But what you've done is the standard way to focus a large-format camera (which tend to be the most primitive like what you have here.) If you can find new replacement ground glass for something like a 4x5 view camera you can probably get a better focus screen.
Hint
Use a optical microspe do print the masks
They come with pretty good light condensers. All you have to do is use the Primostar the other way around.
For a little more price, you can get a 395nm laser (instead of LED). It's diameter is lot smaller than the LED's, and might obviate the "double diffuser" part.
The hotspot seems to be mostly the LED array, the type of LED you have there has a lens that directs the light at a larger angle. If you want more of a point source you need one without a lens.
The LED in this case is made up of several LEDs not one, which is what you are seeing when you turn it down. Though I think you main issue is the lens not the array, the array is very dense and should not make a difference.
of you want high uniform intensity it might make sense to get some WCLSP packaged LEDs and make a PCB for a dense array.
You also want to make sure to use a constant current driver for higher uniformity between the individual LED chips. A series with tightly controlled binning would also further help keep it consistent. Cree and Osram are good sources there.
Gut gemacht 👍! Mach bitte weiter so! Subscribed.
breaking taps has some awesome content.
Definitely!
Can't you use mirrors for the enlarger, instead of lenses?
Yes, a collimation or focusing mirror could be used. But in that size they are also quite expensive.
Could you adapt a Köhler illumination setup as used in microscopes for your light source? You use a diaphragm and a movable lens to produce a uniform flat illumination, then project that with a better lens than the Fresnel lens?
I was just about to comment the same thing. :). That was the big deal with his set up, "Köhler illumination acts to generate an even illumination of the sample and ensures that an image of the illumination source (for example a halogen lamp filament) is not visible in the resulting image" (wikipedia). The other option might be to forego this whole set up and stick the photoresist part on a resin printer's display (many people do this already to make PCB's). The pixel size is usually about 50 microns so not quite as good as the resolution shown here but way easier if you have the printer.
you can use a light tunnel from a projector assembly, it will make light uniform
Amazing content! The only comment I could make is that the brush in the lift off might be quite rough? Have you had issues with broken routes (not that important for just image printing, but if you start doing circuits with small scales it might be an issue)?
In our process we actually use a plastic pipette and continually blow air bubbles onto the surface while lifting off (we do gold and aluminium on silicon, so maybe the adhesion is a bit different, I have never worked with a glass surface)
that is a MONSTER kernel of corn, holy shit
Have you thought about trying direct laser writing?
I did not use it for photolithography, but got good results processing polymers with an monomode laser diode + collimation and a 10x microscope lense.
Monomode laser diodes would also make for a good point source of light for your current setup.
I wonder if the rounded characters is because the corners are being mechanically removed with the excess material when brushed off. Even a perfect mask, I would think, would have some tear out on more fragile features like corners.
Fucking nice, glad you finally managed to get this working well, dunno if you took inspiration from my comment or more people but that doesn't matter, honestly happy this project is moving forward. You could probably use some sort of a UV Night vision specialized optic (for UV tubes), really expensive but will probably work the best, also getting a more specific single wavelength source of UV will also be better, or at least a filter for that specific wavelength. LEDs usually give out a few different wavelengths instead of a single one and will cause some abberations since the lens definitely isn't made for UV, I know that the transmission probably isn't that important as contrast but my guess is that some of the non-transmissed light is also reflected instead of just absorbed, thus leaving you with lower contrast and stray light. Try to look for quartz glass lenses instead of borosillicate glass lenses, quartz glass has a much higher transmission of UV than borosillicate, it's just more fragile and harder to work with so not used as much.
Also, you probably don't even need a photography lens, my guess is that with your set up and a single wavelength of light, literally ANY high definition optical lens will work since abberation won't be a problem, you could probably just buy a cheap large sensor lens (olympus? maybe just a better magnifying glass lens?) and be able to focus the image on the entire slide.
Still waiting for some fine art to be made with this, world of possibilites with anodizing titanium.
Victory dance!!
Yay we get a happy dance!
Congratulations!
I'm curious though why don't you use an ultrasonic cleaner?
Me when the professor says we can use one 3x5 card on the final
My allowed cheat sheet in the physics exams was written almost as small :D
You should be able to use a laser pointer and lens to project the slides onto the wall
I suggest editing out the timer alarm noise (or make it quieter)
According to my prediction based on the speed of improvements, we will see a working silicone IC in 3 videos
Hello, you are making amazing content, I have been fascinated by vacuum technology for a long time. Your channel and everything on it really interested me. Could you please share with me some information about your vacuum chamber? I'm interested in as much as possible, pipe size, wall thickness...
Regards. @Advanced Tinkering
Would this technique work to create glass pcbs? Would i still need a hi vacuum sputter if all i wanted was copper on the glass?
Neat
Curious about dispersion -- have you verified that your optical path is in focus at the UV being exposed? Setting up with a frosted (etched) glass slide, stained with fluorescent dye, might be a good way to test that.
@12:50 I'm a little confused...rotary vane pumps can get down to 10^-3 or 10^-4 mbar, if you are at 10^-2 mbar for this experiment (or is that just because of introducing the Argon? and otherwise it'd be far lower?), why bother with the turbomolecular pump? That's more to get down to 10^-5 mbar and below.
Regarding the problem of the LED(4:00) thoes LEDs have built in lenses, and this tresults in the hot spot observed. Try using LED's without lenses, eighter remove them, or buy ones without them, or maby you could even salvage some from a broken(or working) LED light bulb if you were to be in a hurry
You should do halftones for photos
Do a mask without the diffusers, invert that mask. Put the inverted mask infront of the next exposure. It will now balance out.
You can make lenses for telescopes using titanium sputtering
Dude, you are insane...
Here I am just trying to anodize this damn shit, and make it all pretty like
I should probably give up, nah?
Would this work with 3d printing resin? You could make some very small objects.
Could you put something on the slide to focus then take it off?
Is this a home lab? That’s easily over $200k worth of equipment. I’m kind of jealous.
Yes, it's my home lab. But it's nowhere near 200k. Almost all of the equipment was bought used.
Dein selbstgebauter Overheadprojektor ist wahrscheinlich das deutscheste was ich diese Woche gesehen habe lol
the potassium video will be uploaded here or elias? thank you!
We will both upload a video about it.
I am confused, as this problem had been solved long time ago with photography (see millions of microfiche archives in local libraries)
So, you needed to use an overhead projector in reverse?
Ich arbeite bei Pfeiffer Vacuum wie kommt es das man sich für Zuhause eine splitflow und eine duo leistet günstig ist das jedenfalls nicht.
Aber cool was du machst
Ist wie bei jedem Hobby, das man mit Leidenschaft verfolgt. Die Pumpen habe ich relativ günstig gebraucht gekauft. Die Splitflow ist aus einem alten Massenspektrometer und ich habe eine Adapterplatte entworfen, mit der ich sie an die Kammer anschließen kann. Falls es dich interessiert, gibt es dazu ein Video auf meinem Kanal.
he literaly canibalised a meopta...do you have the head? I really need the head for mine (the light source) as I wanna do some enlarging
Yes, I still have the head. Those enlargers are not hard to find on ebay für 30-100 €.
@@AdvancedTinkering the german autocorrect speaking :D I am really looking for the Meopta color head. Trying to do darkroom prints.
Hey for your information, get lenses that have a HUGE fstop value (maximum Fstop) and the stop it down.
Focus, stop down to F5.6-8(keep within this range so you avoid difraction losses and lens distoetion
Keep in mind CHROMATIC ABERATION. Focus using the UV light not visible light. If you focused with visible light then your focusing plane will be shifted for that speciffic wavelenght.
Ypu need to focus using your main lightsource.
In the meopta head you can easily change the lens eithet by unthreading it or changing the whole plate out. Buy a set of M42 lenses, ideally 50 85 105-135MM lenses. F1.8 50mm lenses F2.8 lenses too (large taking lenses)
Chromatic aberation becomes WORSE the lens is further away from the projection surface (your taking surface that is exposed with the light)
The more the lens is extended the more you have to correct from chroma, so best thing to do is to focus using your ACTUAL light source.
would you be willing to make a mandelbrot set?
Wait isn't photolithography the technique used to etch silicon wafers?
You should make mechanical watch faces !
"WE already have microfiche at home. Microfiche at home:"
If you set the lens at F8 you don't deal with diffraction or softness of F2.8 or F2.0
Wow, how small you made the text is just ridiculous! lol 😂
It would be even cooler if you could go even smaller and spell out words using individual atoms!