Writing in Titanium on Glass - Lift-Off Photolithography
Вставка
- Опубліковано 24 тра 2024
- In today's video, I will be using photolithography to create fine structures and text in metal (titanium and silver) on glass. For this purpose, I will coat the glass with a photoresist, expose it to light, and then develop it. Using the mask created in this way, the glass will be coated with titanium in my vacuum chamber, and then the mask will be removed.
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
If you need custom circuit boards, 3D printed parts, or CNC-machined components, visit PCBWay: www.pcbway.com/
ProjectsInFlight: • Using Hydrofluoric Aci...
AppliedScience: • How to make precise sh...
SVGs for the masks:
www.freepik.com/free-vector/t...
www.freepik.com/free-vector/l...
www.freepik.com/free-vector/l... - Наука та технологія
"If you want to feel bad, you can always compare yourself to AppliedScience" - I agree a little too hard.
Missed opportunity to put the bee movie script onto the glass slide 😂
Jokes aside, awesome video!
Haha, true. Or the Shrek script.
Thanks!
Getting copyright striked with style! (and science)
I guess Ben (applied science) set the bar quite high. But don't feel bad, your content is top tier too (based on my years of experience binge watching science video).
That transition to the sponsor read was perfect.
Maybe you could use a mono LCD, like in MSLA resin 3d printers? Human hairs are about 80µm, 2k 5.5 inch LCDs have a pixel size of about 47µm and 4k 6.23" means a 35µm pixel size. You can get them with drivers *relatively* cheaply.
Yes, that is one of the methods I will try in the future :)
Some more advanced and older type resin printers used a UV laser, I wonder if using something like this and turning the laser power down might work better. Because using a laser is more like plotting a vector than using pixels.
The laser based printers were called SLA. They could produce better curves because there are no pixels, but they are slower than DLP printers that have taken over.
@@thoughtstream9591the mainstream printers are actually "Masked Stereo LithogrAphy" (yeah, no idea why they chose the A either...) or "MSLA" and not "DLP".
DLP is a slightly different technology where a projector shines an image onto a mirror, which exposes the layer on the vat. MSLA uses a mask to selectively expose a layer.
The technologies are very similar, because a projector is basically a light source with a masking element - but I guess when it comes to patent laws, these things matter...! 😅
Yeah, (non-mask)SLA would be pretty interesting too, especially since for a single 'layer' the longer exposure time isn't a problem, but getting a galvo stage with a laser would probably cost quite a lot. Then again, some may be available for a reasonable price on the used market or a DIY one could be made, but that's a whole other project.
"If you want to feel bad, you can always compare yourself to Applied Science"
Ain't that the truth, haha
Also, the idea of ol' Liz's face on that silver coin being converted to hot plasma is very funny to me
I ran into the same issue of expensive sputtering targets, so I bought an inexpensive manual metal rolling machine. It allows you to create thin, uniform sheets of metal out of anything that's fairly malleable: copper, tin, gold, platinum, silver, etc. Some metals may need to be annealed periodically with a torch since they work-harden, but it's surprising just how many different metals can be rolled. Since sputtering removes such a microscopic amount of metal, you don't need a thick target, and you can glue it to a thicker piece of metal to aid in mounting it in your sputtering system.
Try looking for CTF masks at small print shops, I can get them for just a few bucks at 4000+ dpi resolution, optical density of 4 and very crisp lines. That's what PCB fabs typically use
I never heard of CTF masks and couldn’t find a lot on google. What does “CTF” stand for?
But I will look into it, thanks!
@@AdvancedTinkering Computer-to-film, it's a type of a printing process, still popular for screen printing t-shirts and art I think. Masks look like laser printed transparencies except they are much blacker and higher quality. If you really can't find them I can mail you some test patterns I was about to throw out
Something to test: a lot of photo machines use a thin layer of fluid (generally di) between the mask (reticle) and wafer, this helps with focusing. Easy enough to test in your setup if this gives better results.
I didn’t know that. Thanks for the information!
@AdvancedTinkering yeah, honestly I'm not sure if it will help here tho, it's used near the diffraction limit to reduce fringe interference effects iirc. However it does help with focusing, so I still think it's worth a test haha.
@AdvancedTinkering yeah, honestly I'm not sure if it will help here tho, it's used near the diffraction limit to reduce fringe interference effects iirc. However it does help with focusing, so I still think it's worth a test haha.
What is this di?
@@MekazaBitrustyDeionized water.
Laser printing onto tracing paper works pretty well for photomasks - use heavy (>90gsm) tracing paper to avoid crinkling in the fuser, and set the printer DPI as high as possible.
I don't know if it's still easily available or if the process is now obsolete, but you used to be able to get super high-res transparencies (>1200dpi) from a print shop that has a phototypesetter. Some older PCB fabs also use photoplots onto a clear film. Photographic film might be another option but I don't know how well it transmits UV.
We used to use the transfer paper printed with laser printer and a iron to transfer the toner onto pcb. I don't see why that method can't be used here.. Skip the lithography altogether
Lithography and photomask are far more repeatable than toner transfer.
To get the best density with toner+velum I explored using heat transfer foils which have essentially 0% transmission, but this was for larger feature sizes and I'm not sure how well it scales down.
It made for very pretty metallic masks though.
I've done some experimentation with black and white film when I was doing my cyanotyping course and I did find /some/ uv transmissibility but it seems that a decent chunk of the optic chain that I used (which was a reversed camera lens) even without UV filter absorbs a lot of the UV, not that my UV source was very strong. It's been on my todo list to revisit (preferably before the remainder of my film expires) thou it is also on my todo list have a go at using a resin printer to expose the UV resist.
I haven't tried it with my B/W enlarger because it uses white styrofoam as a diffuser/reflector so the losses would be off the charts
I work in a semiconductor research lab, if we need to make a one off wafer with something really "big" like a photo in gold on a wafer we use transparency sheets for overhead projectors, they are available for laser printers.
Print on a black and white laser printer with the highest density and you get something quick but not the greatest resolution.
Another way to obtain optically dense masks is to use computer-to-film services (for offset printing); while it is not the most advanced technology, they still do it.
Some places accept input as 1-bit tiff files (in spatial resolution of imagesetter), and in that case you can obtain masks with resolution significantly better than a printer.
Pro: cheap silver target for the magnetron.
Con: never be able to enter England again.
:D
How could not being able to enter England be a "con". ;)
@@edgeeffect I like the people, and the culture, and the geography, but they have terrible taste in government..
It's actually an Australian coin.
@@sophietaylor9753 That makes sense, if it was English it wouldn't be silver it would be steel! ;)
@@sophietaylor9753 That makes sense, if it was English, it wouldn't be silver, it would be steel. ;)
I used this positive 20 spray for pcb and it worked best usin spin coating. But the photoresist has to be cold. And The RPM has to be low. For masks use silver coated film from a commercial painting fab that will produce them for fabrication of sieves for silkscreen printing.
Thanks for the advice!
This is so much harder without a lab, good job man! The results are incredible
Thank you!
One easy to increase resolution would be to costruct a diy litography stepper. Basically it's a pc projector that you focus down the plate using microstope. Huygens Optics channel has a great video about that. Another, easier option may be to buy resin 3d printer - they expose the resin with UV light through high density lcd screen.
You can get PCB way to print masks on PET plastic film. They do this for flexible PCBs. PET is transparent and has a workable UV transmission up to around 350nm(looking at transmission spectra on Google) but you can increase the exposure time to compensate. The mask will have metal features so you will have a perfect contrast ratio. PCB way can do a 60um line/space resolution. I'm sure you can negotiate a suitable set of parameters if you explain what you want to do.
Woah, the resist under UV light at 9:01 glows! That's cool, I wonder why it's mostly at the edges
Also, it's nice to see you using items from multiple previous videos to make this happen!
Nice catch! My interpretation: the resist reacts to uv, meaning necessarily something within it is ionized at those wavelengths, that ionization leads to crosslinking (negative resist) or bond breaking (positive resist), from there one of two things must be happening: either the chemical change only requires part of the photons energy and the rest is lost as a photon at a higher wavelength, or once the chemical change has occurred the energy cannot be fully absorbed and the electron releases what isn't converted to heat as a photon. These lead to phosphorescence.
The higher amplitude at the edges I'd guess is due to the different surfaces refracting the light and channeling it to the edges, similar to a fiber optic cable.
This is my guess (sorry for my messy English): The UV light passes through the film and enters the glass, and some of it undergoes total internal reflection (TIR) within the slide. The light bounces off the inner surfaces, traveling inside along the whole width of the glass. It eventually reaches the edges, where it can escape. The effect is more pronounced along the sides because of this, as the light interacts more this way with the glass triggering fluorescence more noticeably than on the flat surfaces.
Dye-sublimation printers block UV light a lot better than jets or lasers and also print nicely on transparencies. Worked well in the late 90´s when making pad printing plates. The UV exposure machine had a total number of 3 UV fluorescent tubes that each were about 30 cm long*25mm wide at 8cm distance from the glass, so at the very least, plenty of UV.
I've seen Ben's videos, and he IS SO smart. But you are too! You should feel so proud of what you've done so far because it is great! Great job!
Goodness, I need one of those text prints. Ozymandius on that? Perfect.
Thank you for sharing your learning adventure with us. The day we stop learning is the day we start dying!
You can improve contrast of that laser printed mask by exposing the print to acetone vapors.
I forgot to mention that but I have tried it (since ProjectsInFlight suggested it in his video) but the transparency was still a lot higher than the screen printing paper.
This is very interesting, with this level of detail you could probably create a simple semiconductor device. It for sure wouldn't be efficient, but it might be worth-it to try.
"If you want to feel bad you can always compare yourself to applied science" had me dying. 😂
Space age printing press! This was an incredible video to watch. I'm going to have to watch the build video now.
Dang this is good stuff. You can make some really interesting doohickeys with this setup!
Thank you! If you have any specific ideas, let me know.
@@AdvancedTinkering Miniature vacuum tubes come to mind
In offset printing, real black is formed by not only printing black, but duplicating the black channel onto the cyan. Probably on inkjet you can even try all other colors. If you use Illustrator or Photoshop you can view each color channel and copy paste them to one another.
Very good idea! I heard before, that some printers print an additional color besides black, but I didn't think of using it here. Overlaying different colors in Photoshop is a good idea! Although I didn't feel like the issue was that my mask was letting light through. Rather, the printer's resolution was reaching its limits.
@@AdvancedTinkering I good printer driver (at least on Mac, has the option for 'overprinting' in the settings too.
AMD better watch out. Semiconductors coming next? Using a cathode ray tube with a deflection coil in your vacuum setup would be very cool! Electron beam lithography would be a neat project.
kinda crazy to me that this way of metal covering even works. great job!
Old-school chip fabs used to use a type of heat-shrinking film. the design was layered together by hand (taping-out) then shurnk down to create a photolithography mask.
Although a big camera might be easier!
Here's a guy who has the best toys.
You should build a jig to project a larger mask first to a smaller point, then with a Fresnel lens into collimated light to then project it onto the actual slide. This way the resolution of the printer no longer becomes a limiting factor in the resolution of small parts.
A fresnel lens would do more harm than good, since their image forming quality is intrinsically inferior, especially with regards to high-frequency details and optical abberations.
Fantastic experiments, super interesting and amazing that you are able to do this at home.
For self made PCBs, i just used naked copper plated PCBs and applied a thin coat of black spray paint.
I then used a laser engrave to remove the paint where i wanted to etch the copper away.
After etching i removed the paint with thinner.
Hm- you could use a lens to reduce the size of your image. That way you could print at a much larger scale and then focus the image down to whatever size you need. The only difficulty would be the fact your UV light pretty much has to come from a point source. Depending on the photoresist you might be able to get away with a 405 nm laser diode, otherwise you could probably use a single-chip UV LED at longer exposures.
Using some sort of optical system/lenses is probably the best way to get a good resolution. I will see what I can do for a future video.
I love this channel, why it has only 50k subs??? My dream is to be able to do such a things on daily basis
If you have to use a printer, yellow ink is usually the most uv-opaque (this is exploited when making digital negatives for alternative printing techniques), but for the best resolution one of the most accessible mediums is microfilm stock, you can find it packaged in 35mm or medium format rolls and expose it with a camera (using a highly resolving macro lens).
Thank you for the information!
The converging line pattern looked like the printer head moved across, not along the lines. Maybe you can get better definition of features along the axis of printer head movement?
It's been a few years, but when I did PCB etching I think laser printers had superior resolution.
Other than that, I think optical projection is the best way to increase the resolution. Also crazy expensive I imagine for large area patterns without warping dimensions.
Oh, that's a very good point! Thanks for pointing that out! I will try and see if the results are better if the lines are oriented 90°.
My laser printer and ink jet have the same dpi. But the ink jet prints were a lot darker.
Just mind-blowing 😮🎉
Instead of using a mask directly on top of the slide, you could try to use lenses to project the mask from near the light instead. You could then use a bigger printed mask and shrink the pattern using lenses. Maybe an old microscope and projector.
Fabulous content as always, thank you 😊
That's pretty awesome. Can't wait to see what interesting projects will come from this! 🤩
Thank you! If you have any suggestions, let me know. :)
Impressive!
Nice, photolythografie.
Some time ago, I researched screen printing. There was a piece of software which used a six-cartridge Epson printer with each cartridge being black to make the darkest masks.
That's cool, I really need to get back to building V2 of my sputter...
I used to use my 3d printer with a small cube being printed at 0c bed 0c nozel amd 0% flow. I "print" this cube wayyy up high off the bed..
Anywho, then I use the bed thats moving back and forth as a shake table. It makes even development nearly perfect every time. Was more effective with the copper etching portion with ferric chloride when I was doing this for PCB manufacturing though, but I used it for the developing too.
Using coins as magnetron sputter targets works really nice. The lab I did my Bachelors degree in did that for gold.
You really built up a nice setup there.
Amusing to hear, that even pros are doing it that way.
A gold coin is still very expensive. But probably cheaper than a gold target. If I ever have the money, I will buy a small piece of gold and roll/hammer it into a thin disk.
You could try cleaning the jets on the ink jet printer. Most printers have a clean function you can use.
I mean the slides are of about the same size as a 35mm film frame. UV Lightbox with your printed master in it, + Camera lens pointed at it from a fixed distance away to easy reducing optics with tall the annoying abberations to correct for like if you just made your own from raw lens elements. If that works, but you want to expose a larger area, then you can worry about finding a larger lens?
Yes, I think using optics is the best way to get a good resolution. I will try your suggestion using a camera lens and see if it works.
Very cool, and thanks for demonstrating all the steps and the learnings! I've tried laser exposure with my LDGraphy project a while back; the first result was mostly good for PCBs, but not this fine resolution. I think for high resolution, taking a picture with regular film camera e.g. from a screen to a high-contrast film is indeed a good idea, and would be a good next experiment.
Mungolux has photoresist that is quite thick and works well for spin coating, if you want a thinner layer you can dilute it, I used acetone but other solvents that are less volatile might be even better. The resist itself has some issues with being tricky to develop right if your setup has a low contrast ratio but the resolution can be quite good.
great video
Thanks! I'm glad you liked it!
You can use some lenses to project image from film to glass (like when you print photo from film), but instead of magnifying you can reduct size of projected picture. It will allow you to increase resolution comparing your printer can do. However, it will be tricky to find lences that are transparent to UV light.
Remember that a lot of photo resist have a top and bottom side. Usually the inside curl of a roll should be the side to laminate down onto the surface.
(I think you did it right in this video so no worries.) I've had adhesion issues when doing it the other way before.
Great video I can't get enough of this sort of thing ! I imagine your aware of the channel 'Breaking Taps' he plays with magnetron sputtering and deposition issues too.....cheers.
Thank you! Yes, Breaking Taps is one of my favorite channels.
You might be able to increase the definition of your mask by overprinting it (printing the same pattern on top of the existing one) a few times.
I’ve done something similar with parallax barrier masks on transparencies.
I have tried it when I was still using the normal transparent paper but couldn't get them to align perfectly on the second print. I suppose you need a printer that supports this feature and does not eject the paper after the first print.
@@AdvancedTinkeringI believe the printer had guides that aligned the paper when printing; perhaps one can 3D model/print a similar guide for the slide mask?
I use tracing paper and a laser printer to create photomasks for PCBs and it works without problems. That it doesn't work well for you is very likely a combination of materials used and the specific laser printer, not the process itself. Make sure that your photoresist wavelength matches your lights. LEDs are usually really bad at this (unless you made sure they are the right wavelength), because their wavelength is often just slightly below 400nm, while many photoresists are more sensitive closer to 350nm. The dry film is also very thick, usually several tens of µm (compared to a few µm for other film types). That is important, because resists have an "aspect ratio", that is how narrow a feature can be compared to film thickness. E. g. an aspect ratio of 10:1 means for a film thickness of 50µm features can't get narrower than 5µm. It's essentially the same reason why you put your printouts face down onto the resist.
For reference:
• The paper I use is tracing paper 82g/m² (Schoellershammer Glama Microdraft hochtransparentes Zeichenpapier, DIN A4 Block mit 50 Blatt, 82 g/m²).
• The printer is an ancient Brother HL-2030 with unbranded refill toner, set to the highest resolution and no "toner saving". No double stacking or printing needed.
• The PCBs are Bungard presensitized boards. Obviously PCBs are not what you need, but Bungard offers a "Presensitizing Service" for your own material. Maybe they can do it on glass as well? Their photoresist has a very wide margin between removing the exposed resist and dissolving the non-exposed resist.
• The exposure is done by a repurposed face tanner. I put the 4 UV tubes and the back reflector of the tanner into an old flatbed scanner case.
you might want to try dlp 3d printer which maybe you can create some sort of undercut for easy lift off, or build a machine to focus a controlled beam of uv light directly to the photo resist..
Awesome
Thank you!
You could think of some kind of optical system for exposure. So that you can print the mask at a higher scale and fokus it down to the actual dimensions.
Have you considered using an MSLA 3d printer to expose the photoresist? Modern ones have nicely designed optical systems that project very detailed UV light in whatever pattern you want. It would add a bit of cost, but you could skip the special paper and printing and probably have higher detail as well.
I would experiment with a 4K monitor and a lens to project it on the glass. Project with blue only might be enough to develop the photo resist. If not you could remove the back panel and replace the white led backlight with ultraviolet leds.
4K pixels (horizontal) projected to 4cm glass surface would yield 1000 pixels per cm = 10,000 per meter resulting in a 0.00001 meter resolution which is 10micron.
If you ever want to feel bad just compare yourself to Applied Science. 😂
100um is suspiciously bad considering the process that you're following. You should really be able to get nice straight lines, closer to 60um resolution and without that 'dithered' or grainy appearance.
But just in case this is the best you can hope for, you can try something like multiple patterning. That is, breaking up your original mask design into 2 (or more) masks where each has only features that are spaced > 150um apart, then you just need to do accurate aligning once, since you can use the produced titanium part as a single mask for all subsequent production runs.
The only thing I can think of (apart from using a spin-on photoresist like you already said) would be to ensure you don't have any printer drivers/software still lurking somewhere in your printing data path, because those masks look very similar to what you get when the image you want to print gets pre-processed for normal photo use. Rather than reproduce sharp, high contrast lines, you just get a dithered mess.
If you're not able to get an inkjet printer setup where it prints a bit-perfect copy of your input data, you might be better off just hacking open a cheap/old 4k projector.
They can reliably produce perfectly planar images, down to 10um resolution, especially if you replace its stock light source with a 405nm laser. It doesn't need to be a particularly powerful laser, and even cheap multimode lasers can be made to work quite well if you insert a beam expander, an optical diffuser and a polariser (aligned to the axis of rotation of the DMD's mirrors) in before the collimator lens.
One slightly less toxic alternative to using KOH+IPA for cleaning your glass might be to use "whiting," or calcium carbonate. It is used for cleaning flux and cement in creating stained glass. It was recommended to me in a video by HISglassworks for preparing glass for gluing with HXTAL- an archival grade polyester epoxy.
Glass artists use a titanium "pen" to indelibly sign their artwork. The titanium marking is improved by marking the glass while it is wet. I have had good results creating images on glass using a TiPen in a desktop CNC. I have not had good results attempting to electroplate the titanium images with copper- the plating achieves a certain thickness and peels back away from the glass in the plating bath, and further plating creates more "images" which also lift up. The result was like pages of an open book, lifting slightly due to their own stiffness. The copper doesn't bond well to the titanium.
Something tells me the anti-aliasing in your computer's fonts or OS may be reducing the quality of your masks. Hence the single side fuzzies perhaps.
If not that, an image reducing setup may be necessary to get higher quality. Collimated lighting followed by reflection off of a glass backed first surface mirror mask (created by magnetron sputtering) and then shrinking lenses or mirrors should be adequate.
Maybe use a Fresnel lens and a "light tunnel" to increase the intensity of the lighting and the "directness" of it..
Also maybe you can make a bigger pattern on a glass slide, and place it behind a focusing lens so its shadow mask of the area of the real target.
Awesome!
Rather than improving the mask resolution, would it be easier to focus the UV optically? You mention wanting the lamp "as far away as possible", presumably so that the rays are _approximately_ parallel at the shadow mask, but it should be possible to put a much larger mask directly on the lamp and use a concave lens to focus it down to slide-sized parallel rays. Similarly, you could then use an LCD (e.g. from a resin printer) - or a DLP and mirror - to make it an active mask.
Riston is fine under LED's as they have no UV, get an old UV eprom eraser for uber results or use a SLA 3d printer to do the exposure without all the printing drama as it prints directly to the Riston don't forget to reverse stuff, we use it to make PCB's
man.. i really wanna start doing this kinda stuff again..
You can also try your inkjet in photo mode which may increase resolution a lot.
LED household lights don't produce any UV.
I work with photoresist under white LED light regularly without issue.
As for increasing resolution, is it possible to use multiple masks layered like they do for silicon chip production?
I wonder if you used polarized light for the uv step if it would clean up the lettering a bit
Maybe you can find a way to use a scaling factor on the mask. In other words, print the mask x2 as big, then scale it down to the desired size using lenses like they do in the semiconductor industry.
In my experience you should use a point light source. Doing printed circuit boards, a light source that wide as the one you use severely limits the achievable resolution. I actually prefer a metal halide lamp without reflector.
I don't think there's really a better way to improve the mask than to make it a lot bigger. A regular old camera lens and something to mount stuff to should be pretty easy to cobble together into basically a backwards photographic enlarger.
I have one idea, how you could get finer details. But this way would involve Optics. You can Print your stencil way bigger, like A4 Paper Size biger and than project it with lenses smaler on your resist. Like a Overhead Projector but making everything smaler.
Or maybe with an Resin SLA Printer.
And your Ad was perfect. normaly is skip this ads but you made it intresting and not disrupting your video an i watched it.
is it not possible to use the screen of a resin 3d printer instead of a uv masking method as it kinda functions as a uv mask itself, maybe a option for testing.
You vould try printing the image 200% larger and using a lens to focus it and scale it back down. Effectivly doubling the resolution. I believe this technique is used in cpu manufacturing.
You should try ordering a metal mask with your text or designs laser cut. PCB Way should be able to accommodate that. They make solder paste masks for custom pcbs.
The issue with this is unsupported lands, such as the inside of a lowercase e for example. Those will just fall out. With a PCB paste mask, there are no unsupported islands like this.
Is it possible to print a larger template, to make a smaller one?
I'm thinking along the lines of a hot halogen bulb, a magnifying glass and some thermal printing paper?
I don't know how transparent the black would be, but it should make crisp edges at minute scales.
You can also use an ASML machine, I think you can get a good DUV for 60M
Should you not remove the 2nd protective film layer first? This would get your mask directly in contact with the resist film, instead of being separated by the protective layer. Also, there are various thickness of photoresist films and I think you would want the thinnest one you can find.
I'm not sure if another comment has mentioned it already but I'd be interested to see whether an LCD resin printer with a high enough resolutions could actually improve your resolution. I imagine you just place the slide on the print bed instead of the VAT of resin and do a manual exposure for as long as you need.
nice ! is it possible that you contaminate the gloves with the adhesive paper?
Can you sputter onto borosilicate and would the titanium survive being encased in more borosilicate? Or survive a trip through the flame in the 3 to 4000 thousand degree mark? If you could sputter some text or patterns onto a borosilicate glass slide I could encase it in glass the qay we encase dichroic or silver or gold in glass.
In borosilicate glassblowing we do a sort of sputtering with silver, gold, platinum, palladium and electrum, we call it fuming and it's like vaporizing the metal onto the target glass from a molten ball of metal held in the end of a glass rod. I'm trying to find some sort of stencil that can survive this process. People already use lasers to burn patterns on already fumed glass but I think this method could work as well to make some interesting art.
Awesome work, keep it up!
What about using a lcd uv resin bed to photo etch the material?
it looks like the imperfections in your print are in one axis, rotating your mask may allow you to squeeze out a bit more resolution, past that maybe try using optics to scale down a larger mask to a smaller area
I wonder if you could expose the glass in a 12K resin 3D printer to get a much finer resolution?
In effect you are doing what in photographic terms would be called a contact print, where the negative is in contact with the emulsion of the print. Photolithography would give higher resolution. It is the photographic equivalent of using the enlarger in reverse.
How about printing on paper with a laser printer and transfer it to the glass with acetone? Works well for etching pcbs, wonder what the resolution limit is though...
Could you use lenses somehow during the uv treatment of your photoresist?
You might try spin-coating with the liquid photo-resist. Based on long experience (watching UA-cam videos) that makes for a consistent thickness.
What wavelength are you using for your LED UV light? I’ve been looking for a good one that emits at 254 nm.
Btw i think would be easy for you to integrate in the setup a lens to focus the uv light better. Cant wait to see how youll improve the mask
Using lenses to project a smaller image of the mask onto the photoresist is the next step :)
You could probably easily increase the resolution by having a lens setup from some old TV fresnel lenses and a few normal ones so the image would get converged into a very small area on the slide as long as you had a sufficiently even light scattering and a good set up. It could easily take the resolution of two A4 papers and put it on the slide.
You can simplify it even further with a photography lens that has a small F stop (like a Helios 44m4, which is like 20-30 bux if you know where to look or even less) and a big box, you could straight up just use the photoresist as a photography film.
I bet you could even shove it in a camera and take UV pictures on glass in titanium.
There is also a BIG possibility of making art here, imagine taking the same "photo" with 5 slides but with a varied exposure, removing more and more film that way, then shoving it in and adding them in a series. You'd have a 3D titanium image on glass slides. Or just printing an image, making it PURE black and white with varying thresholds and doing it that way, would be insane for nighttime panorama pictures.
Have you also experimented with titanium oxide? It makes some funky colours, There's a few nice videos on titanium electrolisis you could basically add colour to the titanium and you could vary it with a great degree.
Maybe use laser light to deposit the mask on the photoresist layer. The laser light has a more directed light beam.
Maybe for the section that you what to deposit on the printer can be scaled down by the lenses? somehow...
Argon pressure during sputtering is ok for magnetron and too high for the turbopump, providing excessive heating of blades and motor overload. You could have used another input of this two-port turbopump to lighter the load, or provide flow restriction at its input.
I wonder if you could use a fiber laser on very, very low settings to "draw" on the photoresist. Of course that requires a fiber laser at the right wave length so it would be expensive and possible not work.
I was also thinking about that. But a fiber laser is prohibitively expensive for me. Otherwise it would definitely be worth a try. Maybe not even using photoresist but just burning away a coating (layer of paint) which can be dissolved after sputtering.