@@projectsinflight this thing is the second best project after clicksprings antikythera project on all of youtube :D and i know that sounds mean but thats high praise!
People might poo poo me for putting it like this, but it's pretty goddamned hype to see a second DIY semiconductor fab spinning up on RUclips. The more people doing stuff and communicating how they are doing the stuff, the more everyone can learn and that's just stellar.
hell yeah brother. I once kicked around the idea of reducing home depot play sand grade SiO2 into silicon, crystalizing it with a diy HHO Verneuil furnace, and then trying to dope my own wafers with stuff like borax or phosphorus bearing fertilizer, but in the end i realized there was no way id make something pure enough to exhibit the necessary properties . Maybe one day, but realistically thats probably outside the scope of mere mortals like me. @@projectsinflight
Ben from Applied Science addresses the problem of photomask density in his video on photomechanical machining. He ends up using a screenprinting ink in an inkjet and a transparency with a special emulsion. His techniques do stretch the limits of DIY.
For the printer DPI issue, you usually need to specifically go into some kind of control panel or print settings menu to enable the 1200dpi mode, as it's not the default printing mode for nearly any printer.
You can try to expose the film directly to UV without using a printed mask by using a DLP or SLA 3d resin printer's laser bed, it probably has greater precision than the paper printer
Why not just set up a basic LLC for $100? Then you can just directly order from these photoresist suppliers, plus if youtube is your business venture for the company you can probably even deduct these expenses.
This is a great idea for this specific project, but it would be great if someone knows of a supplier that can sell to individuals, that way this project can be replicated more easily.
Yes, you could do that. But IIRC Ben from Applied Science made his own, or found another solution. He also used a thinner lacquer with a spincoating process, which leads to a thinner layer and thus a finer resolution.
@@projectsinflight I wanna start a business that produces semiconductors and lets everyone design their own ICs and test whatever architecture they want. I plan on starting with mcirometer-sized transistors. Keep posting stuff and hit me up if you're into this kinda stuff, I plan on making my research as much as possible public too. Another goal would be designing a consumer-available litho machine. Looking to make connections right now so anyone can hit me up if they're interested
You might want to research PMMA (Poly Methyl Methacrylate). It's generally used for e-beam lithography, but I believe it can be used with UV exposure, though it may need a light source like low pressure mercury vapor. I've bought PMMA in the past without any problems (for e-beam lithography). Another really cool thing with PMMA is you can layer different percentages and wind up being able to create tailored undercuts allowing 3 dimensional structures instead of just 2D. I had marginal success with standard screen printing emulsion, though my research indicated there were emulsion forms that could get higher resolution. I thinned the material (a green gunk; I believe with water) and using a brush (I think; it was years ago) was able to get a somewhat reproducible layer. I'm sure spin coating would be way better. I believe I was getting feature sizes close to a micron (I was aiming for 100 -50 nm) when my investors ghosted me and I dropped the project. Feel free to reach out if you'd like me to try and dig up more details.
@@Gin-toki I'm sure it was just a few months ago I was looking for exactly this. Either my search skills have atrophied or these are relatively new... Thanks!
I've been looking into PMMA and it does seem to be a good prospect. I wasn't aware deep UV would expose it without an additional synthesizer chemical so that's interesting. I also am in the process of building my own e-beam source :)
I have no idea if it would work because it is for PCBs but Kontakt Chemie sells a product called Positiv 20 that is a UV reactive positive paint that comes in a spray paint can and it's specifically made for very high resolution circuit boards. It's like 20 bucks so it might be worth checking out if there are no other options
Have you looked into using using a resin 3d printer to exposing your photoresist? That would allow direct exposing of the photoresist. They use a UV light to cure the resin. Some of the more expensive printer claim over 1000dpi resolution. Also the resin itself could be evaluated for spin coating and use as a resist.
3D printer is the way to go. So much easier than the transparency method. I use an Anycubic Photon Mono 4k to expose dry film for PCBs. I can easily and reliably get 250 micron traces, and I had success with test patterns down to 70 micron with 70 micron gaps, or 2 pixels across. I used to do exactly what you did in the video, with that exact model printer even (except the acetone smoothing - wish I'd thought of that!). For your own sanity, ditch the transparency. You even get a 3D printer as a bonus!
I wonder if your plain old photography film would work as a precise mask. By printing the transparency on a regular printer then optically shrinking it down and exposing the film. Then developing it as if it was a normal photography negative.
I'm amazed at how quickly yet thoroughly this series is progressing, phenomenal work! Only found your channel recently but I'm pretty invested already.
Thanks man! I was doing work my final year project of my studies which it PCB fabricator! Currently we are working on PCB etching! Thanks to you i have a clear direction, now i can work more fastly and efficiently! ❤
In order to get a sharper exposure, you should print the mask mirrored and then place it upside down. That makes the ink come much closer to the photo resist so less light can sneak in from the sides. You should also weigh it down with a piece of glass. Also the farther away you place the light source, the more parallel the light rays, which will also help with sharpness.
Have one of the amazing chemistry RUclipsrs just create The GOOD Stuff for you. Sounds like a really cool project for them. One of the ingredients is Cyclopentanone which half of them have made w/ the Cubane wars.
I would LOVE to collaborate with one of them at some point. I'm just concerned because a lot of the chemicals you need for light-sensitive chemistry are pretty toxic, so i wouldn't want to put them at risk
You might be able to get a finer resolution on the mask by using a material that shrinks uniformly when heated (I've always heard them called shrinky-dinks) Some of them are transparent and can shrink to several times smaller than they started
Not a bad idea- though i don't know if it would be consistent enough for multiple mask layers. most semiconductors need several masks for different steps and they'd all need to be the same exact size
@@projectsinflight I believe shrinky-dink film is made by stretching the polystyrene film, so that it relaxes when heated. It reliably contracts the same amount. Whether it's viable is another question, but the darkening effect is a good point- ruclips.net/video/a15AqeJs3W8/видео.html
@@projectsinflight Actually I recall hearing about a Collage or University students that used something similar for making tiny flow channels. Originally they had issues making the markings precise at the small scale, however it was found that Shrinky-Dinks or at least the generic material would shrink in a predictable manner. Each test/set-up was a minimum of three layers. Such that they could make very precise designs, just needing to make sure that it was held flat during the Shrinking to avoid curling. It was clear for filming the internals.
it's incredible you made that connection between 3d printing acetone spreading to making your stencil. wow i'm blown away. thank you for making these video, they're incredibly informative.
Small piece of feedback, when you're talking about the mask at ~2:40 I think it would be helpful to animate/highlight the appropriate portion of the mask as you mention it, i.e. transparent vs opaque. When I initially saw the mask I assumed the black bars would indicate the transparent areas and that would have just made things a little more clear.
After a lot of trial and error with my laser printer at 600dpi that is great for toner transfer of pcbs, i found that an inkjet printer with 1200dpi is the best solution for a better resolution and black coverage for the transparency film, many of my pcb tracks are 10 mils and its very easy and reliable to get this result with the inkjet printer. After loosing a lot of time with bubles i gave up the dry film and i use only the paint with an oven for drying pcbs.
Yeah, I hate the dry film and I'm hoping to find a substitute soon. I also want to try an inkjet to see if it works better for me. I was concerned that the ink would be even less opaque than the toner
Have you considered replacing the mask all together with a set of UV/blue laser galvos? They use mirrors on precise motors to move a laser beam around. With that kind of laser power, you should be able to expose the photoresist really quickly. So, moving the laser slowly across the chip might work. As a bonus, you don't need to worry about the opacity of a mask. Though, the beam width would be a limiting factor.
I've considered galvos but i don't have any experience with them and i don't want to get too sidetracked right now. I'll come back at some point and improve the technique, either with galvos or something similar to improve resolution
Hey! Great video, I absolutely love youtubers who educate people about science. I have a tangent idea for you: PMMA, which is essentially acrylic, is used as a resist for electron beam lithography on silicon. Although this process is extremely challenging for a hobbyist, you can snatch high power electron gun, beam steering columns, and accompanying electronics from a CRT, construct a crude vacuum chamber, and try to perform maskless lithography. I believe, PMMA solvents and solutions are somewhat easier to come by as an individual. Small amounts of chemicals with pre-coated silicone chips are usually sold as a student training kits by SEM manufacturers, so you could try to pose as one, and maybe you'll get lucky.
Have you considered using a small fish tank pump for the developer solution and a some blunt needles for spraying. It wouldn't take much to put together and would be gentler than cotton swabs.
Should try adding a surfactant like Triton before anything complex like that, various developers have them added in to help with wetting. In my experience manual agitation should be fine without any sort of scrubbing (for development, liftoff is another story) but I've never worked with dry films before.
I’m a PhD chemist, and I wanted to really commend you for your thoughtful and well planned use of this chemical. This is the right way to go about choosing to bring a material into your lab. The one piece of advice I’d offer, is you did not address how you intend to dispose of the waste from these HF sources when it’s generated, and it may be useful to consider that going forward.
well, so far i have produced very little HF waste, and i have neutralized it to form NaF, which i figured was ok to just put down the drain since it's an ingredient in toothpaste and mouthwash. I'm hoping that is an acceptable solution.
@ again I think you’re really going about this in a great way. Like you said NaF is a common ingredient in household products, so a few ml here and there is suitable for the average American municipal sewage system. But your video demonstrated that if you were generating liters and liters every day, you’d reconsider your waste disposal process.
I had a project where I wanted to produce diffraction gratings using photoresist on glass - so I had the same problem finding a resist. I had quite good results using a commonly available positive PCB spray resist "Kontak Chemie Positiv20" which comes in a spray bottle and resolved 10µm lines quite well (dont know what resolution you are aiming for). It contains mainly acetone as a solvent wich evaporates a little too fast for Spincoating uniform films. The solution was to mix it with a solvent called "1-Methoxy-2-propanol", which is also part of the resist itself and is easely available on abay. If you are interested in more details let me know. I can also share some experience with building a simple exposure unit.
@@projectsinflight Our good 'ol German Company was bought by CRC (Belgium). Maybe this helps when searching. (RUclips censorBS, Splitting Post into parts, Part 1 here) Edit: PLEASE MOVE DOWN TO MY LAST 3 POSTS AND IGNORE THE OTHERS TO PREVENT EYE-CANCER!!! I am letting this mess here for everyone to see instead of erasing it, just that everyone can see the hell I went through ... My apologies!
@@projectsinflight Alternatives(separated by semicolon): ELECTROLUBE PRP - POSITIVE PHOTORESIST (RUclips censorBS, Splitting Post into parts, Part 2 here)
@@projectsinflight mungolux (company for photoetching supply): R1000 ~ 25µm. Shipping worldwide (RUclips censorBS, Splitting Post into parts, Part 4 here)
Regarding photoresist, i think you could try to use any novolac-based spray-on pcb photoresist, like Positiv-20 by Kontakt Chemie. It is very close to the novolac non-chemically amplified resists like AZ15xx, but maybe has a little less controlled quality. You can even try to spin-coat it.
Yeah, I posted that one in the comments of the last video too. It's datasheet specifically mentions it's ability to be used as a resist for glass (SiO2) etching with strong HF. I think you can't easily get a more ideal photoresist than that. Farnell sells it in a lot of countries, so it's easy to get.
@@projectsinflight If you can figure out where it's sold you have 2 solid options: 1. Find a company that is willing to buy and ship it for a fee 2. Find a Patreon member that lives there and ask them nicely to send some to you
@@projectsinflightSeems I cannot post links (my comment with links got discarded), but you should be able to find something in your area by searching for "digital film output" or "digital film output for offset printing" This is slightly outdated technology (in modern offset printing they expose photoresist on plates directly), but it is still present. The fact that it is purposely made for curing UV photoresist makes it appealing in my eyes. You should be able to negotiate submitting your files as 1-bit tiff at target machine resolution (typically people send vector graphics which are being rasterized and then sent to the machine; so you need to skip a step in order to get perfect result). There are also options on which side you would like to get photoemulsion and if you want a negative or positive output.
Nice to see that cheap PCB photoresist works at least somewhat! There are options for acquiring semicon photoresist as a hobbyist. You might have a pretty good chance of getting the stuff by contacting the manufacturers directly, *not* the distributors, or looking overseas(China, Korea, Japan). You can also try UV glue, clear SLA resin, or even PMMA(it works without an ebeam too!). I've emailed you with more info on where to start digging.
@@projectsinflight> You could also look into the UV setting nail polish that the trendier kids/tweenies paint their nails with, seems cheap enough for an experiment or two, and it looks like it could be almost the right consistency for spin coating given it's usually applied with a small paint brush.
perfect, thank you for demonstration, to stick dry film in pcb i saw they spray water on surface and rub the layer with plastic like knife to stick the file and remove the bubbles, and second laminator is little thick to pass through i think although is perfect for pcb this is why there is bubble and did not stick properly
Yeah, the water seems to not significantly affect film adhesion to copper, but it completely prevents the film from sticking to silicon for some reason
Applied Science's video on "Patterning aluminum with the photoresist "lift-off" method" seems to be relevant. Also relevant are the Breaking Taps channel, and Sam Zeloof, though both of those seem to be a little out of reach for the average viewer. I've heard that Cooking with Jeri, by Jeri Ellsworth is worth a look, but I can't attest to it myself.
I used the paint, you can use toluene to thin it and the smallest that work reasonably well is 0.02mm(on copper) as for glass it's 50/50 after cleaning 0.1mm detail. Any smaller detail, resist start to peel off.
@@projectsinflight it doesn't seem to have much or any affect after it cured, it work the same as the dry one other than the colour change. As for mask I make it on glass (etch then use silver mirror process on it so it can build a thicker layer to electroplated or other) it might be overkill for few testing uses.
While I never done lithography... I got some time in the darkroom doing analog printing. And there is clearly some overlap. Some tricks you might learn is to do a test strip for time. block 80% and then move the blocker to 60% 5 seconds later and again and again. Gives you a stepped gradient.
Hi, did you consider using an ultrasonic cleaner? It should be quite affordabl, accessible, and i guess should agitate the solution enough so the chemicals gets renewed where it shold be while moving the disolved particles.
@@projectsinflight They are sometimes sold as jewelry cleaners, and the price is mostly driven by the volume of the bath. Since you are working on a very small scale, it should be quite affordable!
An experiment I have been looking into doing is using Elegoo's photosensitive polymers for 3d printers as a negative resist. It should be possible to estimate the exposure dose and resolution if you know your way around a 3d resin printer
I love ur videos, please stick to ur methode of producing. It is super easy to understand even if you are new to the topic. I glad that RUclips brought us together ❤
about the lamination issue with dry film, it can be greatly improved by applying it wet, similar to vinyl stickers etc. Pass it through the laminator while still a bit wet and able to slide on surface.
@@projectsinflight Yeah, it can be interesting to compare multiple methods! I've seen someone use this for PCBs and it seems to work well, but there are a few drawbacks: - pixels are not totally opaque, so some light pass through - UV LED arrays on resin printers don't produce a very uniform light, I had a few problems when trying to print microfluidics, but it's probably not a problem here especially if you use your own light source - LCDs screens have a limited lifespan when exposed to UV light I can't wait to see what happens next in any case! Good luck with your projects.
Many have spoken of doing this, you seem to be the one making the most progress the fastest. I wish you all the success in the world and an following avidly!
im absolutely loving this project. the idea of taking something that seems like magic, only able to be done by large industry and working it out in a garage is just awesome. do you have any plans for what you want your first chips or circuits to be? or i guess the first thing would be to make loads of single transistors to test right?
yeah i feel you. i delayed this project for years out of fear of HF. i eventually decided that every hobby is risky to some extent, and that as long as i was using HF in a safe manner i was ok with the level of risk.
can a 3d laser printer not be used to edge in beautiful logic patterns, or a mask, that you can then fill with mist and wash again? like laser edging a pcb, then filling with thin. maybe print the design on the same translucent paper multiple times, for a more evenly, or thicker distribution?
Nice work, with this pace, we will soon see a video from you with probably your first 100 micrometer working chip :) Great achievements always start from below so I am looking forward to it, Your video forced me to hit the like button :)
I did exactly this actually! The dry film dissolves nicely into acetone and spins on in a perfectly even coat. Unfortunately, something about the process ruins the ability to develop it, and it will just dissolve entirely when you try. Haven't come up with a solution yet to fix that.
Acetone vapour bath: As I recall, you want to lightly heat the acetone when smoothing 3D prints,. Easy to do when you're printing ABS with a heated bed: just put the acetone bath on the bed and warm it to like 40C. For developing: could you put the wafer and developer into an ultrasound bath? I imagine that might help with agitation in an inexpensive way. Also, some models have heaters in them, which might be good for faster developing times. You're setting up for spin coating for other steps, aren't you? Wouldn't spin coating the resist mask be easier than the troubles the film is giving you? Finally, I feel like you really need a small simple photo enlarger, and then set it up wrong to get an image reduction of 2x to 4x. Nothing too ambitious, but it would allow you get simpler prints *and* higher resolutions. It would also allow you to look at other printers and printing processes. Maybe an inkjet or pen plotter would give better contrast without fiddling with the acetone vapour smoothing. Oh! I know! you could print on the laser at 16x, expose onto copper-clad glass with a photo reduction of 4x, and then use that to expose the silicon! :D
I tried heating the acetone and it did indeed produce more vapor, but it condensed onto the mask and dissolved away the toner. I would probably need to heat the lid as well to keep that from happening
it's possible that would help. in industry they have pumps and they actually spray be developer directly onto the chip. honestly though i'm looking for a different photoresist, which will hopefully be a better fit for the substrate material
@@projectsinflight There is a youtube channel called Huygens Optics which is run by an Ex-ASML guy who is using some resin based photoresist. You might want to check him out.
@@projectsinflight There are resin based photoresists which is being used by another youtuber right now. If I write his name, youtube spam filter deletes my comment.
Photolithography. Where is this going? Well, its obvious where but this is super difficult and if you can make the cmos process cheap enough to do at home, you can guarantee that I will be doing this at home. I was wondering, how do you plan on doing doping? I've been looking into home cmos-ing and have been considering laser doping and was interested in your plan. You already have a furnace so laser doping may not be needed. Also, do you think that electron beam lithography would work here too? You can get smaller feature sizes.
CMOS will not be easy. Even for companies like Intel it wasn't easy. They took small steps from PMOS to NMOS to finaly CMOS. Any sort of metalic contaminants, especially sodium ond potassium, was causing troubles with CMOS. So for printed circuits boards sodium carbonate or sodium hydroxide is cheap and accessible. But CMOS would require nastier stuff like TMAH, which is even toxic.
You should add more "buffer" water in your washing solution. The photoresist material makes such goo instead of dissolving if the solution is close to saturation. Also I don't like the photoresist color, it appears to be partially exposed at the beginning. Or maybe just the camera changes the colors. Anyway, you can see the line across the material at 22:47; you should not use the exposed part. Don't forget that LED lamps may produce UV, because they are basically the UV lamps covered with luminescent stuff. I always work under the general incandescent lamp, and I have no exposing problems in like 10 mins I need to make everything. So you probably don't have to work with such dimmed light, just choose the proper source. Also I always stick the film to my boards before lamination by gently drawing the line across the center and then smoothing from that line to the edges to make air go away. But yeah, this may not work for you because my copper surface is much more rough than your "glass" one, and it's easier for air to escape~
The photoresist datasheet says you can go down to half-strength developer but says not to go below that. I tried stronger developer but didn't try weaker yet. I got my photoresist film from amazon, so it's probably a remnant and no guarantee it wasn't already partially exposed. part of the reason why I'd like a better resist. I tried sticking the resist to the board before rolling but I never was able to make that work. It always resulted in bubbles when I did it unfortunately, so i looked to other solutions.
@@projectsinflight I mean, use more washing solution, it should have space for the photoresist material to dissolve. I use Na2CO3 for partial dissolving and NaOH for cleaning after copper etching. There are liquid photoresists in form of spray. Haven't used them, but it's probably easier to make a thin film by this way. Also, photoresist and the chemicals are cheap, you may want to get some practice with thin glass :)
It is with great pleasure that I look forward to each of your videos, they are incredible and a real motivation for me. Thank you very much, may Jesus bless you.
Look for POSITIV 20 from Kontakt Chemie. Its a spray based positive photo resist, I've used it and it can stick in all kinds of surcafes, from copper PCBs to glass. Also you'd be far better off by using an big DPI Inkjet printer for masking at these scales, especially ones that are used for hi-res photos; some offer 5000+ DPI. Although you will need special transparencies for inkjet printers for these. Best of luck :)
Yeah, PMMA is an excellent E-beam resist and needs a special chemical additive to make it into photoresist. I don't have the means to make the additive though right now
@@projectsinflight for some reason I'm thinking about 3d printing photopolymer resins. They have photoinitiators out-of-box. I'm not chemist, but I think you can also buy photoinitiator instead of making it.
I use the dry film to make circuit boards for tht and smd and it works really well for that if you expose enough. As a mask i print on inkjet foil using a epson eco tank et 1810. You have to look through the settings a little to find the setting that uses the most ink but the masks work really well and the printer can print in a very high resolution. I also noticed that the dry film is sensitive to wavelength that get fully blocked by kapton tape. Something that might interest you… the dry film can be dissolved in a solvent to create a liquid film that can be spin coated but be aware that this doesn’t get close to commercial photo resists. You can get a better resolution though. With the dryfilm it’s very important to know that there are two different foils on it’ one is a bit stretchy the other is hard. The hard one is heat resistant the other is not. Also the laminator should be one with two heated rollers. I hope this helps a little for future experiments
Thanks- that is good info. Regarding dissolving the resist in acetone and spinning it on, I actually tried that. It spins on nicely but something about the dissolving step ruined the development process. It just dissolved off almost immediately, rather than making the pattern. Not sure why
Since the uncured Photoresist is soluble in aceton, can you not dissolve it/water down the other one and try to get a more even application with thinner films for better resolution? Then you don't need to worry about the laminator
I actually tried that and it does dissolve and does spin on in a nice coating. However, it somehow ruins the ability to develop it, and it just comes off all at once. not sure why
In the rainbow glass video, you used a hot plate to produce steam, and I can see a stir knob on it as well. You should be able to find a stir bar small enough to fit in that dish next to the chip, and it will circulate the developer for you.
It actually not as useful as one would hope. the used-up developer sticks to the surface so you really need high pressure jets and not just general stirring to remove it
@@projectsinflightwell, then what if we get silly and glue a little brush to the stirrer? (probably heat-welded, and two, and one of those 4+ limb or puck stirrers so it's stable)
Unfortunately you are on the wrong continent. At our place no one would even notice then 50-100ml of resist vanished 😅 I use SU-8, AZ 1518 and OIR 908 only a daily basis. Overall a very nice video, keep them coming!
Salivating over the thought of becoming a chem PhD for the express purpose of cornering the global consumer-grade lithography photo-resist liquid market 🦀 💰
Here is an idea so stupid it just might work: can you take the unprotected blue photo resist and dissolve it in acetone, then apply a thin layer of that to the wafer? Acetone might boil off too quickly which would leave a bad surface finish, but there are non-polar solvents you can use with a higher boiling point that should give a better finish. Also, I got the ball bonder moved into my garage. It is missing the part that actually touches the silicon though. Hoping to be able to fire it up sometime soon to see what can be done with it.
Actually I tried that. The acetone indeed dissolved the film and i was able to spin on a thin layer. The layer didn't develop properly for some reason.
@@projectsinflight I wonder if a catalyst got washed out or something. Could also have a much different polarity. Might be interesting to put a sample on a TLC plate to see if anything moves. If you are willing to mail me a 3x5" square I could play around with it to see if I can come up with something.
An LCD screen lets too much light through I think? But you can use a modified DLP projector for maskless photolitography. With no real limits in resolution because you can use lens to reduce or increase it.
Displays for smarphones nowaday have resolution around 400 ppi, which is great, but even better printer and transparencies could do better. It could be also possible to cheaply outsource the transparencies for some DTP studio or other company, which produces transparencies for screen printing, like for T-shirts, or printed circuit boards.
@@oni2inkyou can absorb most to all light with a LCD screen. In a automatic welding mask is a LCD display too and it can absorb most of the light. But with a transparent foil and a few lenses would be also a great idea. Do you know a overhead projector
There is and maskless photolithography with DLP projector with removed UV filter and microscope objective. On that way it is possible to achieve very high resolution.
How about using black and white photographic film in a copy camera setup for making the mask? You can print out your initial mask at a much larger size, then reduce it in size using the camera and 35mm film.
I thought about doing that. You can get special lithographic film for this process too! I went this route for now since it was the simplest possible solution with the fewest pieces of specialized equipment. I plan to improve the process soon!
@@projectsinflightI have a 35mm camera, what’s the name of the emulsion? I might be able to get some of the film and develop it. Then I could send the developed film to you.
This is an awesome series! ik sam zeloff did something similar and with greater presicion but he had $100,000+ worth of equipment. If you werent limited by the photoresist couldn't you use optics to create much more detailed results like you showed in the beginning?
Also i have an idea for masks, can you print a negative, silver it like a mirror and then use a solvent to remove the toner and thus the silver on top of it?
maybe a fish tank pump or a small fan could blow the developer around to keep fresh developer moving over the chip? electron beam lithography often uses acrylic or PDMS (silicone rubber) for resist, maybe they can be used for UV resists too. Or maybe the photopolymers used to make 3D printer resin could be bought online and used as a resist.
The used-up developer sticks pretty well to the surface, so you really have to use high pressure jets rather than gentle stirring. I want to try PMMA and e-beam lithography as soon as I have an E-beam source. I don't think SLA resin will work because I don't think it dissolves well but i don't know for certain
I wake up 4 hours early by accident, and what do you know, this video has been published. No chance I'm not watching this before trying to sleep again... [Edit:] Regarding the photoresist lifting due to water exposure, some colleagues use an HF vapour etcher. The vapour pressure of HF isn't much higher than water, but perhaps not being submerged could be critical. Also wondered about you making your own photoresist from 'near-enough' constituent components, or equivalent replacements, but I'll investigate how realistic this is. Another thing occurred to me was to use an ultrasonic bath when developing the resist, rather than scrubbing with a cotton bud. Could also use a high-sided dish, and spin the complete assembly to create a small vortex with the resultant flow of developer to achieve more consistent removal. Just throwing these ideas out there. Finally, are you aware of Sam Zeloof's work?
it's a cool idea but I'm gonna avoid HF vapors for the foreseeable future due to toxicity concerns. I want to make photoresist but i haven't yet found a suitable recipe that doesn't use toxic and carcinogenic chemicals. I've considered the ultrasonic cleaner but I think spray jets might be better (at least according to the datasheet of the resist). I am aware of sam's work and considered the DLP-maskless lithography source as a future project. I thought it was silly at first but it turns out masks are hard to make lol
@@projectsinflight Fair enough! I stay as far away from the stuff as possible tbh - vapour or otherwise. The one in work is basically the same as doing a normal etch, it's just not submerged. It looks like the Idonus model they use, but will ask them tomorrow. Sent you mail btw; thought I'd mention as I'm guessing the PIF address isn't your main one.
![I.N "HALLUCINATION" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/n5B5q1Hwt_U/mqdefault.jpg)
![BLACK BAG - Official Trailer [HD] - Only in Theaters March 14](http://i.ytimg.com/vi/Du0Xp8WX_7I/mqdefault.jpg)
I have not been this invested in a youtube series in years. Please continue this project, its incredibly interesting and motivating!!!
Thanks! i am doing my best!
@@projectsinflight this thing is the second best project after clicksprings antikythera project on all of youtube :D
and i know that sounds mean but thats high praise!
People might poo poo me for putting it like this, but it's pretty goddamned hype to see a second DIY semiconductor fab spinning up on RUclips. The more people doing stuff and communicating how they are doing the stuff, the more everyone can learn and that's just stellar.
My dream is to make this hobby more attainable for the average hobbyist. I hope I can do that :)
hell yeah brother. I once kicked around the idea of reducing home depot play sand grade SiO2 into silicon, crystalizing it with a diy HHO Verneuil furnace, and then trying to dope my own wafers with stuff like borax or phosphorus bearing fertilizer, but in the end i realized there was no way id make something pure enough to exhibit the necessary properties . Maybe one day, but realistically thats probably outside the scope of mere mortals like me. @@projectsinflight
@@arealhumanname4150 Yeah definitly recommend buying wafers. The process of making them involves a lot of flammable and toxic chemicals
Ben from Applied Science addresses the problem of photomask density in his video on photomechanical machining. He ends up using a screenprinting ink in an inkjet and a transparency with a special emulsion. His techniques do stretch the limits of DIY.
Which video is this? Not sure i've seen it
ruclips.net/video/bR9EN3kUlfg/видео.html this one?
ruclips.net/video/NS8Q9LUIKA8/видео.htmlsi=bCrGyb-nrxubePBe
@@projectsinflight "How to make precise sheet metal parts (photochemical machining)" from 11 months ago. watch?v=bR9EN3kUlfg
@@projectsinflight ruclips.net/video/YAPt_DcWAvw/видео.html
For the printer DPI issue, you usually need to specifically go into some kind of control panel or print settings menu to enable the 1200dpi mode, as it's not the default printing mode for nearly any printer.
You can try to expose the film directly to UV without using a printed mask by using a DLP or SLA 3d resin printer's laser bed, it probably has greater precision than the paper printer
Ya, a Mars 4 9K for example has an 18um (1410dpi) resolution.
I've definitely wanted to try this once i get my hands on a suitable printer
This is what I do when I etch PCB's at home. Works great.
@@projectsinflight reach out to your local college with a technology campus and they can help you out
@@projectsinflight Do the liquids used in resin 3d printers work as photoresists; I mean they are things that harden under UV exposure?
Why not just set up a basic LLC for $100? Then you can just directly order from these photoresist suppliers, plus if youtube is your business venture for the company you can probably even deduct these expenses.
This is a great idea for this specific project, but it would be great if someone knows of a supplier that can sell to individuals, that way this project can be replicated more easily.
@@josuelservin Perhaps whomever sets up that initial basic LLC might then sell on to individuals? 😉
My goal is to do stuff that anyone can replicate so i'm doing my best to keep from using stuff gated behind lab suppliers and commercial adresses
Yes, you could do that. But IIRC Ben from Applied Science made his own, or found another solution. He also used a thinner lacquer with a spincoating process, which leads to a thinner layer and thus a finer resolution.
@@projectsinflight I wanna start a business that produces semiconductors and lets everyone design their own ICs and test whatever architecture they want. I plan on starting with mcirometer-sized transistors. Keep posting stuff and hit me up if you're into this kinda stuff, I plan on making my research as much as possible public too. Another goal would be designing a consumer-available litho machine. Looking to make connections right now so anyone can hit me up if they're interested
An easy way to get a photomask with microscale features is to use a film camera.
Yeah i considered that. they even have lithographic film which is super high contrast B&W film for just this reason
You might want to research PMMA (Poly Methyl Methacrylate). It's generally used for e-beam lithography, but I believe it can be used with UV exposure, though it may need a light source like low pressure mercury vapor. I've bought PMMA in the past without any problems (for e-beam lithography). Another really cool thing with PMMA is you can layer different percentages and wind up being able to create tailored undercuts allowing 3 dimensional structures instead of just 2D.
I had marginal success with standard screen printing emulsion, though my research indicated there were emulsion forms that could get higher resolution. I thinned the material (a green gunk; I believe with water) and using a brush (I think; it was years ago) was able to get a somewhat reproducible layer. I'm sure spin coating would be way better. I believe I was getting feature sizes close to a micron (I was aiming for 100 -50 nm) when my investors ghosted me and I dropped the project.
Feel free to reach out if you'd like me to try and dig up more details.
Can probably get a UV led short enough now to do the job without mucking about with the mercury lamps
@@zyeborm My research on UV LEDs seems to have a hard limit around 350 nm, which may not be enough for PMMA.
@@mitakeetFrom some quick searching on the web, 255nm UV LED lamps seems to be rather readily avaible.
@@Gin-toki I'm sure it was just a few months ago I was looking for exactly this. Either my search skills have atrophied or these are relatively new... Thanks!
I've been looking into PMMA and it does seem to be a good prospect. I wasn't aware deep UV would expose it without an additional synthesizer chemical so that's interesting. I also am in the process of building my own e-beam source :)
Been waiting weeks for this video. Absolutely made my day. Really loving the ground up approach on semiconductors in a home lab
Thanks! It means a lot to me :)
I have no idea if it would work because it is for PCBs but Kontakt Chemie sells a product called Positiv 20 that is a UV reactive positive paint that comes in a spray paint can and it's specifically made for very high resolution circuit boards. It's like 20 bucks so it might be worth checking out if there are no other options
I would love some but i was unable to find a source in the US
Have you looked into using using a resin 3d printer to exposing your photoresist? That would allow direct exposing of the photoresist. They use a UV light to cure the resin. Some of the more expensive printer claim over 1000dpi resolution. Also the resin itself could be evaluated for spin coating and use as a resist.
i actually had a conversation about the 3d printer as an exposure source yesterday, and i'm definitely interested in doing that if i can
3D printer is the way to go. So much easier than the transparency method. I use an Anycubic Photon Mono 4k to expose dry film for PCBs. I can easily and reliably get 250 micron traces, and I had success with test patterns down to 70 micron with 70 micron gaps, or 2 pixels across. I used to do exactly what you did in the video, with that exact model printer even (except the acetone smoothing - wish I'd thought of that!).
For your own sanity, ditch the transparency. You even get a 3D printer as a bonus!
I wonder if your plain old photography film would work as a precise mask. By printing the transparency on a regular printer then optically shrinking it down and exposing the film. Then developing it as if it was a normal photography negative.
Film will work- there is even a film specifically for this called lithography film. Its basically super high contrast B&W film.
A thousand little pitfalls, thanks for your effort on working trough them. Already excited about the next one.
Thank you :) It really did feel like it would never come together but it turns out if you just refuse to fail you eventually succeed lol
Really cool problem solving with the in holes man. You are making basic homemade semiconductors seem a lot more accessible.
Thanks! there were even more minor pitfalls that didn't make it into this video, but I think i covered the most consequential stuff
I'm amazed at how quickly yet thoroughly this series is progressing, phenomenal work!
Only found your channel recently but I'm pretty invested already.
Thank you! It's been a really fun and challenging project!
Thanks man!
I was doing work my final year project of my studies which it PCB fabricator!
Currently we are working on PCB etching!
Thanks to you i have a clear direction, now i can work more fastly and efficiently! ❤
In order to get a sharper exposure, you should print the mask mirrored and then place it upside down. That makes the ink come much closer to the photo resist so less light can sneak in from the sides.
You should also weigh it down with a piece of glass.
Also the farther away you place the light source, the more parallel the light rays, which will also help with sharpness.
Have one of the amazing chemistry RUclipsrs just create The GOOD Stuff for you. Sounds like a really cool project for them. One of the ingredients is Cyclopentanone which half of them have made w/ the Cubane wars.
I would LOVE to collaborate with one of them at some point. I'm just concerned because a lot of the chemicals you need for light-sensitive chemistry are pretty toxic, so i wouldn't want to put them at risk
You might be able to get a finer resolution on the mask by using a material that shrinks uniformly when heated (I've always heard them called shrinky-dinks)
Some of them are transparent and can shrink to several times smaller than they started
This also has the side effect of making pigments darker, which would mean no need to stack multiple transparencies.
Not a bad idea- though i don't know if it would be consistent enough for multiple mask layers. most semiconductors need several masks for different steps and they'd all need to be the same exact size
@@projectsinflight I believe shrinky-dink film is made by stretching the polystyrene film, so that it relaxes when heated. It reliably contracts the same amount. Whether it's viable is another question, but the darkening effect is a good point- ruclips.net/video/a15AqeJs3W8/видео.html
@@projectsinflight Actually I recall hearing about a Collage or University students that used something similar for making tiny flow channels. Originally they had issues making the markings precise at the small scale, however it was found that Shrinky-Dinks or at least the generic material would shrink in a predictable manner. Each test/set-up was a minimum of three layers. Such that they could make very precise designs, just needing to make sure that it was held flat during the Shrinking to avoid curling. It was clear for filming the internals.
You can use a resin 3d printer lcd display to project the pattern to the photoresists
I want to try that whenever i get access to a suitable printer
it's incredible you made that connection between 3d printing acetone spreading to making your stencil. wow i'm blown away. thank you for making these video, they're incredibly informative.
THANK YOU for all this tedious work !
The lack of photoresist for hobbyist is a plague.
Small piece of feedback, when you're talking about the mask at ~2:40 I think it would be helpful to animate/highlight the appropriate portion of the mask as you mention it, i.e. transparent vs opaque. When I initially saw the mask I assumed the black bars would indicate the transparent areas and that would have just made things a little more clear.
Whoops- sorry about that. I'm still learning how to do these animations
Nothing to apologize for!@@projectsinflight
I applaud your bravery for trying to use dry film. These are amazing results for such a low-tech process!
Thank you! I wanted to make sure this was as approachable as possible
Wow this is probably the most informative video on youtube... Trust me videos like this are very rare these days... Keep it up buddy
After a lot of trial and error with my laser printer at 600dpi that is great for toner transfer of pcbs, i found that an inkjet printer with 1200dpi is the best solution for a better resolution and black coverage for the transparency film, many of my pcb tracks are 10 mils and its very easy and reliable to get this result with the inkjet printer. After loosing a lot of time with bubles i gave up the dry film and i use only the paint with an oven for drying pcbs.
Yeah, I hate the dry film and I'm hoping to find a substitute soon. I also want to try an inkjet to see if it works better for me. I was concerned that the ink would be even less opaque than the toner
Have you considered replacing the mask all together with a set of UV/blue laser galvos? They use mirrors on precise motors to move a laser beam around. With that kind of laser power, you should be able to expose the photoresist really quickly. So, moving the laser slowly across the chip might work. As a bonus, you don't need to worry about the opacity of a mask. Though, the beam width would be a limiting factor.
I've considered galvos but i don't have any experience with them and i don't want to get too sidetracked right now. I'll come back at some point and improve the technique, either with galvos or something similar to improve resolution
Photolithography is much more repeatable and accurate, there's a reason photo transfers are still used.
I wonder if it also might work if you replaced the lamp of a DLP projector with a UV one?
i bet you could use galvos and photosensitive materials to *make* a mask though
Hey! Great video, I absolutely love youtubers who educate people about science. I have a tangent idea for you: PMMA, which is essentially acrylic, is used as a resist for electron beam lithography on silicon. Although this process is extremely challenging for a hobbyist, you can snatch high power electron gun, beam steering columns, and accompanying electronics from a CRT, construct a crude vacuum chamber, and try to perform maskless lithography. I believe, PMMA solvents and solutions are somewhat easier to come by as an individual. Small amounts of chemicals with pre-coated silicone chips are usually sold as a student training kits by SEM manufacturers, so you could try to pose as one, and maybe you'll get lucky.
Have you considered using a small fish tank pump for the developer solution and a some blunt needles for spraying. It wouldn't take much to put together and would be gentler than cotton swabs.
Should try adding a surfactant like Triton before anything complex like that, various developers have them added in to help with wetting. In my experience manual agitation should be fine without any sort of scrubbing (for development, liftoff is another story) but I've never worked with dry films before.
Photoresist datasheet calls for several bars of pressure. Do you think that would be attainable using one of these pumps?
I’m a PhD chemist, and I wanted to really commend you for your thoughtful and well planned use of this chemical. This is the right way to go about choosing to bring a material into your lab. The one piece of advice I’d offer, is you did not address how you intend to dispose of the waste from these HF sources when it’s generated, and it may be useful to consider that going forward.
well, so far i have produced very little HF waste, and i have neutralized it to form NaF, which i figured was ok to just put down the drain since it's an ingredient in toothpaste and mouthwash. I'm hoping that is an acceptable solution.
@ again I think you’re really going about this in a great way. Like you said NaF is a common ingredient in household products, so a few ml here and there is suitable for the average American municipal sewage system. But your video demonstrated that if you were generating liters and liters every day, you’d reconsider your waste disposal process.
I had a project where I wanted to produce diffraction gratings using photoresist on glass - so I had the same problem finding a resist. I had quite good results using a commonly available positive PCB spray resist "Kontak Chemie Positiv20" which comes in a spray bottle and resolved 10µm lines quite well (dont know what resolution you are aiming for). It contains mainly acetone as a solvent wich evaporates a little too fast for Spincoating uniform films. The solution was to mix it with a solvent called "1-Methoxy-2-propanol", which is also part of the resist itself and is easely available on abay. If you are interested in more details let me know. I can also share some experience with building a simple exposure unit.
I am interested, but I was not able to find any of the positiv 20 available for sale in the US
@@projectsinflight Is farnell an option for you? Maybe they sell and ship to US ?
@@projectsinflight Our good 'ol German Company was bought by CRC (Belgium). Maybe this helps when searching.
(RUclips censorBS, Splitting Post into parts, Part 1 here)
Edit: PLEASE MOVE DOWN TO MY LAST 3 POSTS AND IGNORE THE OTHERS TO PREVENT EYE-CANCER!!!
I am letting this mess here for everyone to see instead of erasing it, just that everyone can see the hell I went through ... My apologies!
@@projectsinflight Alternatives(separated by semicolon): ELECTROLUBE PRP - POSITIVE PHOTORESIST
(RUclips censorBS, Splitting Post into parts, Part 2 here)
@@projectsinflight mungolux (company for photoetching supply): R1000 ~ 25µm. Shipping worldwide
(RUclips censorBS, Splitting Post into parts, Part 4 here)
Honestly, this is pretty incredible work. It is for sure the most interesting series I've seen in ages.
Regarding photoresist, i think you could try to use any novolac-based spray-on pcb photoresist, like Positiv-20 by Kontakt Chemie. It is very close to the novolac non-chemically amplified resists like AZ15xx, but maybe has a little less controlled quality. You can even try to spin-coat it.
Yeah, I posted that one in the comments of the last video too. It's datasheet specifically mentions it's ability to be used as a resist for glass (SiO2) etching with strong HF. I think you can't easily get a more ideal photoresist than that. Farnell sells it in a lot of countries, so it's easy to get.
I considered that one, but was unable to find it for sale anywhere
@@projectsinflight If you can figure out where it's sold you have 2 solid options:
1. Find a company that is willing to buy and ship it for a fee
2. Find a Patreon member that lives there and ask them nicely to send some to you
You can get dense masks with high resolution from places that run film imagesetters (for Computer-to-Film typography process)
Never heard of this- do you have a link?
@@projectsinflightSeems I cannot post links (my comment with links got discarded), but you should be able to find something in your area by searching for "digital film output" or "digital film output for offset printing"
This is slightly outdated technology (in modern offset printing they expose photoresist on plates directly), but it is still present.
The fact that it is purposely made for curing UV photoresist makes it appealing in my eyes.
You should be able to negotiate submitting your files as 1-bit tiff at target machine resolution (typically people send vector graphics which are being rasterized and then sent to the machine; so you need to skip a step in order to get perfect result).
There are also options on which side you would like to get photoemulsion and if you want a negative or positive output.
Nice to see that cheap PCB photoresist works at least somewhat! There are options for acquiring semicon photoresist as a hobbyist.
You might have a pretty good chance of getting the stuff by contacting the manufacturers directly, *not* the distributors, or looking overseas(China, Korea, Japan). You can also try UV glue, clear SLA resin, or even PMMA(it works without an ebeam too!). I've emailed you with more info on where to start digging.
Thank you so much for providing this information- I'll check out the email soon!
@@projectsinflight> You could also look into the UV setting nail polish that the trendier kids/tweenies paint their nails with, seems cheap enough for an experiment or two, and it looks like it could be almost the right consistency for spin coating given it's usually applied with a small paint brush.
Please forward me this email, I'd like to know how to as well, as this process is very confusing to me!
perfect, thank you for demonstration, to stick dry film in pcb i saw they spray water on surface and rub the layer with plastic like knife to stick the file and remove the bubbles, and second laminator is little thick to pass through i think although is perfect for pcb this is why there is bubble and did not stick properly
Yeah, the water seems to not significantly affect film adhesion to copper, but it completely prevents the film from sticking to silicon for some reason
Amazing !!!
You are the only one who is able to reproduce the magical world of ICs. Congratulations.
Applied Science's video on "Patterning aluminum with the photoresist "lift-off" method" seems to be relevant. Also relevant are the Breaking Taps channel, and Sam Zeloof, though both of those seem to be a little out of reach for the average viewer. I've heard that Cooking with Jeri, by Jeri Ellsworth is worth a look, but I can't attest to it myself.
Yeah, I'd love to do some kind of collaboration in the future!
I used the paint, you can use toluene to thin it and the smallest that work reasonably well is 0.02mm(on copper) as for glass it's 50/50 after cleaning 0.1mm detail. Any smaller detail, resist start to peel off.
i'll have to try toluene then. does that work to remove the paint once it's cured?
@@projectsinflight it doesn't seem to have much or any affect after it cured, it work the same as the dry one other than the colour change. As for mask I make it on glass (etch then use silver mirror process on it so it can build a thicker layer to electroplated or other) it might be overkill for few testing uses.
While I never done lithography... I got some time in the darkroom doing analog printing. And there is clearly some overlap.
Some tricks you might learn is to do a test strip for time. block 80% and then move the blocker to 60% 5 seconds later and again and again. Gives you a stepped gradient.
Hi, did you consider using an ultrasonic cleaner? It should be quite affordabl, accessible, and i guess should agitate the solution enough so the chemicals gets renewed where it shold be while moving the disolved particles.
I would definitely like to try the ulrasonic cleaner whenever i get my hands on one
@@projectsinflight
They are sometimes sold as jewelry cleaners, and the price is mostly driven by the volume of the bath. Since you are working on a very small scale, it should be quite affordable!
An experiment I have been looking into doing is using Elegoo's photosensitive polymers for 3d printers as a negative resist. It should be possible to estimate the exposure dose and resolution if you know your way around a 3d resin printer
I love ur videos, please stick to ur methode of producing. It is super easy to understand even if you are new to the topic. I glad that RUclips brought us together ❤
This is premium content right here!
Wonderful! As someone very interested in semiconductors this series is fantastic!
Thank you! I'm doing my best
acetone vapour trick on toner was crazy
about the lamination issue with dry film, it can be greatly improved by applying it wet, similar to vinyl stickers etc. Pass it through the laminator while still a bit wet and able to slide on surface.
unfortunately that didn't seem to work for me. the silicon is so perfectly smooth that any moisture underneath causes a total lack of adhesion.
Can't wait for the chemists to weigh in and tell you how to make your own photoresist
Oh my god i really hope so. The only recipes i found were pretty poor and used toxic chromium compounds
You're tenacious. Nicely done. Can't wait for the next chapter.
Thank you! i'm working my hardest on it
Hi, great video! I love this serie!
You can also use a modified DLP projector to replace the masks. Sam Zeloof tried this some time ago.
Yep i saw that video. I'm gonna try it at some point I think. I suspect I might also try a resin printer LCD as the pixel density is higher
@@projectsinflight Yeah, it can be interesting to compare multiple methods!
I've seen someone use this for PCBs and it seems to work well, but there are a few drawbacks:
- pixels are not totally opaque, so some light pass through
- UV LED arrays on resin printers don't produce a very uniform light, I had a few problems when trying to print microfluidics, but it's probably not a problem here especially if you use your own light source
- LCDs screens have a limited lifespan when exposed to UV light
I can't wait to see what happens next in any case!
Good luck with your projects.
Many have spoken of doing this, you seem to be the one making the most progress the fastest. I wish you all the success in the world and an following avidly!
Thanks! I'm doing my best
im absolutely loving this project. the idea of taking something that seems like magic, only able to be done by large industry and working it out in a garage is just awesome. do you have any plans for what you want your first chips or circuits to be? or i guess the first thing would be to make loads of single transistors to test right?
Diodes -> Transistors -> 2-transistor IC -> 1st stage of op amp
that's the tentatve plan i think
You sir are a wizard! I love following this project
thanks! doing my best
Just what I was waiting for! Love all the historical info as well, very interesting!
Thank you! I appreciate the compliment
Dude, good job. Last year I did think to dive in this hobby. My big problem was to get the silicon wafer and the HF safety problem.
yeah i feel you. i delayed this project for years out of fear of HF. i eventually decided that every hobby is risky to some extent, and that as long as i was using HF in a safe manner i was ok with the level of risk.
Possibly one of the most interesting youtube channels, keep up the great work!
Thank you so much for the compliment!
Let's go!!!! That's a great video, can't wait for those logic gates to appear 😅
You might find that putting a drop of the liquid resist on a chip then covering with transparency and spreading using a soft scraper works.
Awesome channel, clear communication, pushing the boundaries of home brew. Got my sub.
Thank you! I do my best
A small SLA printer like the Phrozen Mini 8K with 7" screen may get you close to 20 um and has a well columnated UV light. This is such amazing work!
can a 3d laser printer not be used to edge in beautiful logic patterns, or a mask, that you can then fill with mist and wash again? like laser edging a pcb, then filling with thin.
maybe print the design on the same translucent paper multiple times, for a more evenly, or thicker distribution?
Nice work, with this pace, we will soon see a video from you with probably your first 100 micrometer working chip :) Great achievements always start from below so I am looking forward to it, Your video forced me to hit the like button :)
is it possible to use just film camera and lenses?
Yes, in fact there is a special film type for this called lithography film that is basically super high-contrast B&W film
if the dry resist is soluble in acetone, have you considered making a solution of it and evaporating off the acetone after spin coating?
I did exactly this actually! The dry film dissolves nicely into acetone and spins on in a perfectly even coat. Unfortunately, something about the process ruins the ability to develop it, and it will just dissolve entirely when you try. Haven't come up with a solution yet to fix that.
you can put it under uv light after developing to get stronger photoresist
Man found this channel about a week ago and I absolutely love the content you are making
Thank you! I'm doing my best
Acetone vapour bath: As I recall, you want to lightly heat the acetone when smoothing 3D prints,. Easy to do when you're printing ABS with a heated bed: just put the acetone bath on the bed and warm it to like 40C.
For developing: could you put the wafer and developer into an ultrasound bath? I imagine that might help with agitation in an inexpensive way. Also, some models have heaters in them, which might be good for faster developing times.
You're setting up for spin coating for other steps, aren't you? Wouldn't spin coating the resist mask be easier than the troubles the film is giving you?
Finally, I feel like you really need a small simple photo enlarger, and then set it up wrong to get an image reduction of 2x to 4x. Nothing too ambitious, but it would allow you get simpler prints *and* higher resolutions. It would also allow you to look at other printers and printing processes. Maybe an inkjet or pen plotter would give better contrast without fiddling with the acetone vapour smoothing. Oh! I know! you could print on the laser at 16x, expose onto copper-clad glass with a photo reduction of 4x, and then use that to expose the silicon! :D
I tried heating the acetone and it did indeed produce more vapor, but it condensed onto the mask and dissolved away the toner. I would probably need to heat the lid as well to keep that from happening
@@projectsinflight Ew. Yeah, I can understand why you didn't do it then...
20:00 I remember seeing some beakers on hotplates with automated stirring. Maybe that would help?
it's possible that would help. in industry they have pumps and they actually spray be developer directly onto the chip. honestly though i'm looking for a different photoresist, which will hopefully be a better fit for the substrate material
@@projectsinflight There is a youtube channel called Huygens Optics which is run by an Ex-ASML guy who is using some resin based photoresist. You might want to check him out.
@@projectsinflight There is a youtuber called HuygensOptics who is using resin based photoresist, you might want to check that out.
@@projectsinflight There is a youtuber called [REMOVED] who is using resin based photoresist, you might want to check that out.
@@projectsinflight There are resin based photoresists which is being used by another youtuber right now. If I write his name, youtube spam filter deletes my comment.
Photolithography. Where is this going? Well, its obvious where but this is super difficult and if you can make the cmos process cheap enough to do at home, you can guarantee that I will be doing this at home.
I was wondering, how do you plan on doing doping? I've been looking into home cmos-ing and have been considering laser doping and was interested in your plan. You already have a furnace so laser doping may not be needed.
Also, do you think that electron beam lithography would work here too? You can get smaller feature sizes.
CMOS will not be easy. Even for companies like Intel it wasn't easy. They took small steps from PMOS to NMOS to finaly CMOS. Any sort of metalic contaminants, especially sodium ond potassium, was causing troubles with CMOS. So for printed circuits boards sodium carbonate or sodium hydroxide is cheap and accessible. But CMOS would require nastier stuff like TMAH, which is even toxic.
I was able to get some success using diffusion of dopants like phosphorous and boron, but i'm still working out bugs in the process.
@@projectsinflight I am excitedly waiting for each episode. Good luck!
You should add more "buffer" water in your washing solution. The photoresist material makes such goo instead of dissolving if the solution is close to saturation.
Also I don't like the photoresist color, it appears to be partially exposed at the beginning. Or maybe just the camera changes the colors. Anyway, you can see the line across the material at 22:47; you should not use the exposed part.
Don't forget that LED lamps may produce UV, because they are basically the UV lamps covered with luminescent stuff. I always work under the general incandescent lamp, and I have no exposing problems in like 10 mins I need to make everything. So you probably don't have to work with such dimmed light, just choose the proper source.
Also I always stick the film to my boards before lamination by gently drawing the line across the center and then smoothing from that line to the edges to make air go away. But yeah, this may not work for you because my copper surface is much more rough than your "glass" one, and it's easier for air to escape~
The photoresist datasheet says you can go down to half-strength developer but says not to go below that. I tried stronger developer but didn't try weaker yet. I got my photoresist film from amazon, so it's probably a remnant and no guarantee it wasn't already partially exposed. part of the reason why I'd like a better resist.
I tried sticking the resist to the board before rolling but I never was able to make that work. It always resulted in bubbles when I did it unfortunately, so i looked to other solutions.
@@projectsinflight I mean, use more washing solution, it should have space for the photoresist material to dissolve. I use Na2CO3 for partial dissolving and NaOH for cleaning after copper etching.
There are liquid photoresists in form of spray. Haven't used them, but it's probably easier to make a thin film by this way.
Also, photoresist and the chemicals are cheap, you may want to get some practice with thin glass :)
It is with great pleasure that I look forward to each of your videos, they are incredible and a real motivation for me. Thank you very much, may Jesus bless you.
thanks! I appreciate it
Wow this was incredible! I’m at the edge of my seat! Can’t wait for the next one!!!!
Next up is probably diffusion!
@@projectsinflight AWESOME! :D Really looking forward to it!
Look for POSITIV 20 from Kontakt Chemie. Its a spray based positive photo resist, I've used it and it can stick in all kinds of surcafes, from copper PCBs to glass. Also you'd be far better off by using an big DPI Inkjet printer for masking at these scales, especially ones that are used for hi-res photos; some offer 5000+ DPI. Although you will need special transparencies for inkjet printers for these. Best of luck :)
I heard about PMMA being used as photoresist
Yeah, PMMA is an excellent E-beam resist and needs a special chemical additive to make it into photoresist. I don't have the means to make the additive though right now
@@projectsinflight for some reason I'm thinking about 3d printing photopolymer resins. They have photoinitiators out-of-box. I'm not chemist, but I think you can also buy photoinitiator instead of making it.
another great video as always
I use the dry film to make circuit boards for tht and smd and it works really well for that if you expose enough. As a mask i print on inkjet foil using a epson eco tank et 1810. You have to look through the settings a little to find the setting that uses the most ink but the masks work really well and the printer can print in a very high resolution. I also noticed that the dry film is sensitive to wavelength that get fully blocked by kapton tape. Something that might interest you… the dry film can be dissolved in a solvent to create a liquid film that can be spin coated but be aware that this doesn’t get close to commercial photo resists. You can get a better resolution though. With the dryfilm it’s very important to know that there are two different foils on it’ one is a bit stretchy the other is hard. The hard one is heat resistant the other is not. Also the laminator should be one with two heated rollers. I hope this helps a little for future experiments
Thanks- that is good info. Regarding dissolving the resist in acetone and spinning it on, I actually tried that. It spins on nicely but something about the dissolving step ruined the development process. It just dissolved off almost immediately, rather than making the pattern. Not sure why
Since the uncured Photoresist is soluble in aceton, can you not dissolve it/water down the other one and try to get a more even application with thinner films for better resolution? Then you don't need to worry about the laminator
I actually tried that and it does dissolve and does spin on in a nice coating. However, it somehow ruins the ability to develop it, and it just comes off all at once. not sure why
I'm very impressed by this video. Great content, instant subscribe.
Thank you! Glad to produce something interesting :)
Amazing! Thank you for make all these videos.
In the rainbow glass video, you used a hot plate to produce steam, and I can see a stir knob on it as well. You should be able to find a stir bar small enough to fit in that dish next to the chip, and it will circulate the developer for you.
It actually not as useful as one would hope. the used-up developer sticks to the surface so you really need high pressure jets and not just general stirring to remove it
@@projectsinflightwell, then what if we get silly and glue a little brush to the stirrer? (probably heat-welded, and two, and one of those 4+ limb or puck stirrers so it's stable)
Oh and looking forward to the next part :)
Unfortunately you are on the wrong continent.
At our place no one would even notice then 50-100ml of resist vanished 😅
I use SU-8, AZ 1518 and OIR 908 only a daily basis.
Overall a very nice video, keep them coming!
Salivating over the thought of becoming a chem PhD for the express purpose of cornering the global consumer-grade lithography photo-resist liquid market 🦀 💰
haha I know right. I looked into the chemistry but got scared off by too many carcinogenic and toxic chemicals
“This allows us to walk around with devices in our pocket that can do more calculations in a second than we can do in our entire lifetimes.” 😮
Very cool series! Excited to see more :3
Here is an idea so stupid it just might work: can you take the unprotected blue photo resist and dissolve it in acetone, then apply a thin layer of that to the wafer? Acetone might boil off too quickly which would leave a bad surface finish, but there are non-polar solvents you can use with a higher boiling point that should give a better finish.
Also, I got the ball bonder moved into my garage. It is missing the part that actually touches the silicon though. Hoping to be able to fire it up sometime soon to see what can be done with it.
Actually I tried that. The acetone indeed dissolved the film and i was able to spin on a thin layer. The layer didn't develop properly for some reason.
@@projectsinflight I wonder if a catalyst got washed out or something. Could also have a much different polarity. Might be interesting to put a sample on a TLC plate to see if anything moves. If you are willing to mail me a 3x5" square I could play around with it to see if I can come up with something.
@@seeigecannon feel free to email me and we can get the ball rolling
Can you use a lcd display for the mask. I dont know if it work but if it hase a quite high resolution it would work i guess
Or something more crazy you could use the Elektron beam out of a would tv
An LCD screen lets too much light through I think?
But you can use a modified DLP projector for maskless photolitography. With no real limits in resolution because you can use lens to reduce or increase it.
Displays for smarphones nowaday have resolution around 400 ppi, which is great, but even better printer and transparencies could do better. It could be also possible to cheaply outsource the transparencies for some DTP studio or other company, which produces transparencies for screen printing, like for T-shirts, or printed circuit boards.
Yeah- an SLA printer LCD would be perfect i think
@@oni2inkyou can absorb most to all light with a LCD screen. In a automatic welding mask is a LCD display too and it can absorb most of the light. But with a transparent foil and a few lenses would be also a great idea. Do you know a overhead projector
Super impressive 👍
amazing video, i'm always so excited to see your progress with the project
Thank you! I am doing my best
There is and maskless photolithography with DLP projector with removed UV filter and microscope objective. On that way it is possible to achieve very high resolution.
How about using black and white photographic film in a copy camera setup for making the mask?
You can print out your initial mask at a much larger size, then reduce it in size using the camera and 35mm film.
I thought about doing that. You can get special lithographic film for this process too! I went this route for now since it was the simplest possible solution with the fewest pieces of specialized equipment. I plan to improve the process soon!
@@projectsinflightI have a 35mm camera, what’s the name of the emulsion?
I might be able to get some of the film and develop it. Then I could send the developed film to you.
@@wackyvorlon here's the film www.bhphotovideo.com/c/product/1435200-REG/arista_531451_ortho_litho_film_3_0.html
Good for making higher voltage circuits.. that can control a lot of VA
dude you're videos are incredible good, i have no idea how you aren't any bigger
I'm pretty surprised how many people are already here, and I hope that I can continue to grow larger as I make more videos :)
This is an awesome series! ik sam zeloff did something similar and with greater presicion but he had $100,000+ worth of equipment. If you werent limited by the photoresist couldn't you use optics to create much more detailed results like you showed in the beginning?
I've actually received some good tips on getting a better resist so who knows- maybe i'll do some micron work in the future :)
@@projectsinflight make sure to share them in the next video
Also i have an idea for masks, can you print a negative, silver it like a mirror and then use a solvent to remove the toner and thus the silver on top of it?
I actually know about this one. My classmates thesis on this was using non-linear optics to allow for finer resolution
I think you can use 3d printer uv resin with spin coting machine . But i don't know if this resin can anti chemical ething
I'm confused, when you say mils, as a unit of measurement, what are you referring too?
a mil is 0.001 inches in imperial units (for some reason)
@@projectsinflight thanks, I'm not an imperial person so a mill is one tenth a CM
maybe a fish tank pump or a small fan could blow the developer around to keep fresh developer moving over the chip?
electron beam lithography often uses acrylic or PDMS (silicone rubber) for resist, maybe they can be used for UV resists too. Or maybe the photopolymers used to make 3D printer resin could be bought online and used as a resist.
The used-up developer sticks pretty well to the surface, so you really have to use high pressure jets rather than gentle stirring. I want to try PMMA and e-beam lithography as soon as I have an E-beam source. I don't think SLA resin will work because I don't think it dissolves well but i don't know for certain
I wake up 4 hours early by accident, and what do you know, this video has been published. No chance I'm not watching this before trying to sleep again...
[Edit:] Regarding the photoresist lifting due to water exposure, some colleagues use an HF vapour etcher. The vapour pressure of HF isn't much higher than water, but perhaps not being submerged could be critical.
Also wondered about you making your own photoresist from 'near-enough' constituent components, or equivalent replacements, but I'll investigate how realistic this is.
Another thing occurred to me was to use an ultrasonic bath when developing the resist, rather than scrubbing with a cotton bud. Could also use a high-sided dish, and spin the complete assembly to create a small vortex with the resultant flow of developer to achieve more consistent removal. Just throwing these ideas out there.
Finally, are you aware of Sam Zeloof's work?
it's a cool idea but I'm gonna avoid HF vapors for the foreseeable future due to toxicity concerns. I want to make photoresist but i haven't yet found a suitable recipe that doesn't use toxic and carcinogenic chemicals. I've considered the ultrasonic cleaner but I think spray jets might be better (at least according to the datasheet of the resist). I am aware of sam's work and considered the DLP-maskless lithography source as a future project. I thought it was silly at first but it turns out masks are hard to make lol
@@projectsinflight Fair enough! I stay as far away from the stuff as possible tbh - vapour or otherwise. The one in work is basically the same as doing a normal etch, it's just not submerged. It looks like the Idonus model they use, but will ask them tomorrow.
Sent you mail btw; thought I'd mention as I'm guessing the PIF address isn't your main one.