DIY Photolithography using 1980s Carl Zeiss S-Planar Lens (405nm)

In this video I demonstrate how I used a Carl Zeiss S-planar wafer stepper lens to do micron-size photolithography and used this pattern to etch a chromium layer.
More info on how to etch chromium layers can be found in the second half of the following video: • Photomasks Explained (...

[@delphicdescant]

There are far too few DIY photolithography resources on the internet. Thank you for taking such bold steps on this front.

[@denizaliylmaz9263]

"Screw it! I'm doing my own computer."

[@GermanTopGameTV]

Chip shortage? Allright, I'll do it myself then.

[@kevin-jm3qb]

Lmao

[@caiocc12]

You kid but in the next years I guarantee someone will do it in their garage

[@LL-01]

@@caiocc12 someone already have, forgot the name but he does have a semiconcunductor fab lab in his garage, rather awesome to see someone make chips without umbelivably clean rooms.

[@Drozerix]

@@LL-01 makes me wonder if clean rooms are only needed to get consistently good batches needed in a mass production type environment. Obviously dust and other particles will interfere, but if some once can get away without a clean room then it certainly suggests something.

[@Grak70]

As a professional lithographer, this is damn impressive. Well done!

[@weedmanwestvancouverbc9266]

As a side note, silicon wafers are annealed with an intense light source called a Vortek lamp. In a fused silica tube, water is pumped one way, argon gas the other and a strong AC current is passed through.

[@Alexander_Sannikov]

i'm doing a hobby project on computational wave optics and your series is very useful for me to get into the right mood

[@DanielRowe]

Wow that optical effect looks impressive.

[@craigking391]

I have a collection of old 100mm test Si wafers that also have the multi colored effect. The photo resist is somewhere around 40 to 100 nm thickness. They were exposed with EUV light where I used to work in Berkeley. I'm quite impressed by your ability to get results this good in a home lab setup. Nicely done.

[@ViceChief]

You're doing great work and these videos are really fantastic. Thank you.

[@martonlerant5672]

This channel is likely to become far more viewed and subbed. Keep up the good work!

[@chaudry123]

Really love and appreciate your work especially the curiosity

[@impooser]

Wow!!! You did all by yourself! Great respect.

[@patnutoris4054]

Nice video. Just a small recommendation. You could also bake the substrate at 115-120C before the photo-resist coating to drive out any possible water on the surface of the substrate, this improves adhesion. If possible, also apply a chemical called HMDS to further improve the adhesion of photo-resist. This small steps tend to improve the resolution somewhat.

[@craigking391]

Yes, a prebake is good for resists that are sensitive to moisture in the air. A 2 minute bake and on to a copper plate for 10 seconds then coat immediately. Usually gets me a good quality coating.

[@TalkingBook]

Loved loved loved this!

[@MrAndersonHimself]

This is incredible!

[@Kargoneth]

Structural colouration. Like bird feathers and buttfly wings. Nice results.

[@causeitso]

The color effect I think is due to the same reason some butterfly wing is blue. Micro structures reflect selectively, but the interesting thing is that this butterfly wing is blue even with different view angles so it's not just plain diffraction. This should be interesting if you could produce colors using micro structures with your setup. Great video.

[@yoppindia]

genius, i have some hope of making my own chip dream come true.

[@WhatAWondWorld]

incredible

[@Allan_aka_RocKITEman]

*Great work!*

[@extradimension7356]

Very cool unexpected result with the discrete color effects, and the way they overlay. Funny too dropping the sample "Jam side down" , there is a lot of good physics and statistics as to why that's almost an universal law.

[@HuygensOptics]

You are referring to Murphy I assume...

[@extradimension7356]

@@HuygensOptics Actually I did this with my young son at the time to flip toast off a counter top as part of a school science project. So the idea is you nudge the toast jam side up horizontally off the kitchen counter (simulate the normal kitchen blunder"), and generally as the toast slips over the edge a partial rotation is introduced, given the standard height of a counter top, 1G and the time taken to make half a rotation tends to be a very good fit for said toast to land jam side down. We did tests with higher counter tops and even from the top of the fridge to see if we could introduce a full rotation so it would land Jam side UP. I believe a University in the U.K. also did similar tests (Southampton or Exeter, can't remember ) did some very rigorous work on that also. There is some substantial research and statistics on this (now). I wonder how often stuff gets dropped at Intel lol.

[@daviperes]

Thanks for the explanation!!!

[@arkanjo7509]

thank you so much. no way to thank you. very good video and great quality.

[@emanuelnowak5809]

Applied Science have serious competition! thanks for interesting video, subbed and wait for more ;)

[@HuygensOptics]

Thanks for the huge compliment. Actualy I'm a big fan of the Applied Science channel and feel that it would be impossible for me to ever match what Ben Krasnow has achieved. There will be another photolithography video, but it will take me quite some more time to complete.

[@markkrutzmann6862]

@@HuygensOptics I think you are achieving even more here, as you actually go in the depths of making your solution a more permanent fixture than Ben does. He only shows a proof of concept mostly, but credit where it is due, he goes into many many projects. You both are highly appreciated though, and I like watching both channels.

[@AnkushGirotra]

Love both channels. Huge demand for more content like this!

[@MrSchubiduuu]

Ist's so amazing top see what I should be able to do after university. Thank you so much for this!

[@HuygensOptics]

You don't have to wait untill you finish university...

[@mubafaw]

Great explanation. Thank you Thank you Thank you!!!!!

[@plastodoido]

Amazing!!

[@warp188]

Thanks for the content, my dream is to one day have open source tools for fotolitography

[@madisonhanberry6019]

"I accidentally dropped the sample on the floor" 😂 Pretty great result, all things considered

[@glasslinger]

Looks good! I think I will try to make a pentium chip!

[@Drozerix]

It has been a life dream of mine to make a pentium chip. If you can make a 32-bit chip you are golden in my eyes. I'm very excited about learning this sort of complex stuff.

[@hanselda]

That is interesting. The first lithographic optics from Zeiss was indeed based on Planar.

[@user-iq1zn9lo7q]

Super !

[@theastroslav]

Great stuff! For the pattern, I wonder if you can instead place just the elements of a 4K LCD and draw patterns on your computer to be used. I’m assuming the polarization would cause issues to the focusing?

[@HuygensOptics]

kzbin.info/www/bejne/laiTi2WPatuUd7M You can also use the lcd of a 3d printer.

[@AFAMINE]

Nice work and well explanation , which photoresist did you use ? And where did you buy it ? Thanks

[@Ryzler13]

OH DIY, sure, next video we show how to make a DIY hadron collider.

[@douro20]

Fujifilm owns what used to be Olin's photochemical business.

[@ShopperPlug]

How do you make the photomask? can this be used on a vertical CNC bed and sensitize photoresist of PCBs with resolution of 1 micron?

[@rock3tcatU233]

ASML should watch out for some new competition here. :) That's a nice setup you got there, but with regards to the spin coater; How did you manage to keep the sample perfectly centered and now cause it to vibrate itself loose?

[@HuygensOptics]

The double side tape I use is actually way too strong by itself, so I have to decrease the sticking quite a lot before I can use it. The samples are small and so are the centrifugal forces. There is no need to center exactly.

[@honigbrai]

Nice project! Which photo resist did you use?

[@teresashinkansen9402]

Im attempting something similar with microscope objectives, however I am having a hard time achieving good focus, I tried to use another objective to focus the projected image but the thickness of the slide is too much and gives unusable images due spherical aberration.

[@HuygensOptics]

Correct, most microscope objectives require a 0.17mm cover slip to get to the diffraction limit. Also, you need a tube lens in case your objective is infinity corrected. I will actually be discussing this in the upcoming video when measuring the wavefront errors in various optics.

[@Grateful.For.Everything]

The image at 8:01 is wild, how is that even a thing lol, I'm a bit elevated currently and my mind is not comprehending this pattern but I am really intrigued now, who figured all that out, what is that, ahhhhh so many questions lol, sorry but I do really wanna know now.

[@juancarlosdominguezsolis.4665]

Thanks for making this video but I am doing something similar but with other material but I cannot find any image of any processor or integrated to do the photolithography, do you know where to find one?

[@HuygensOptics]

At zeptobars.com you can find a lot of IC designs. However finding masks for the individual layers is hard since this is proprietary information. I see that currently photomasks of Cray computer chips being sold on ebay. But you need a full set to make anything useful.

[@TiagoTiagoT]

Will be cool if you can figure out how to control the colors and make some images with structural color; or perhaps even holograms! :D

[@joshuatanase3718]

Very nice. Does the lens avoid barrel distortion or is it present?

[@HuygensOptics]

It's minimized as much as possible.

[@niteman555]

Those color patterns seem to be the result of regular, non-even lithography due to variations in feature shape caused by edge effects. The uneven etching would cause interference patterns in the reflected visible light, and the color regularity would be caused by similar features having similar interference patterns. What do you think?

[@HuygensOptics]

Very good suggestion. Yes, I think it could be this kind of effect. As was mentioned in the video I think it is the thickness variations in the resist causing the effect. It can be either refraction, or an interference effect, for example between the glass/resist interface and the resist/air interface.

[@extradimension7356]

@@HuygensOptics Can you use your set up to replicate miniature holographic plates ?

[@HuygensOptics]

No I don't think so , due to the 1:5 demagnification optimization for the lens. So making 1:1 copies is actually not feasible with the resolution required for holography.

[@extradimension7356]

@@HuygensOptics Ahhh I see what you mean, the diffraction pattern is locked to the wavelength(s) of light in the original holographic plate and hence reducing that by a factor 5, (If that was possible) would create a diffraction pattern (theoretically) that is not even in the visible spectrum. So you would need a synthetically generated holographic diffraction pattern (to start off with) that would be 5x the size to create a smaller holographic plate? [Which I guess would be pointless maybe?]. As an aside I've been wondering about aliasing artifacts (sampling theory) with the lens you are using is there are any useful effects or accidental byproducts from your rig (as well as the colors) ? I'll keep watching with interest and anticipation.

[@ARVash]

It would be cool to see how much text you could fit on a small piece of glass say 2.5cm, and then if it were possible to read it with the naked eye using reflection or something

[@ARVash]

Have you ever seen that effect where when you hold something very close to your eye if you allow the eye to defocus you can see microscopic features in the reflection, that's what I'm thinking

[@HAWXLEADER]

What do you mean by the maximum resolution of the lens? how can a lens have resolution? Is this due to lens imperfections in the polishing process?

[@hexane360]

Watch the previous video in the series. He's referring to diffraction limited resolution.

[@user-hv9rt7ye6h]

Could you please answer what the type of microscope did you use?

[@juancarlosdominguezsolis.4665]

Thanks for do this vídeo but what Is the chemical name of fotoresistor?

[@HuygensOptics]

The resist is of the dnq type: www.google.com/url?sa=t&source=web&rct=j&url=en.m.wikipedia.org/wiki/Photoresist&ved=2ahUKEwiCh4OTj-_pAhVE4aQKHVGyDhoQFjABegQICBAB&usg=AOvVaw2s2opYilYTx00tImmKTdh1

[@ShihWeiChieh]

hi Huygens, i am trying to cure wet photoresist film with 405nm laser, however i never get a clear develop image. Is it because of the scattered light?

[@julkiewicz]

Maybe a stupid question, but wouldn't the layer of photo resist be slightly thicker at the center of the spinning disk as the centrifugal forces are higher towards the edges?

[@HuygensOptics]

It's a very good question and actually I do not know a good answer to this. It is far from trivial why the film would end up uniformly, especially on large wafers that are used in chip manufacture. I guess it is all about the right combination of solvent (so the evaporation rate), the solid content, viscosity and spinning speed.

[@richardspectacular5327]

Not bad, but can you hack a CRT to do e-beam lithography?

[@woolfoma]

This is sick- how did you even get your hands on that lens?

[@HuygensOptics]

ebay, just search for Zeiss S-planar.

[@intexclouds9607]

Can you please tell me about the light source?? What is the make??...is it 4-5 nm UV?? Thanks in advance.

[@HuygensOptics]

I'm not sure about the make, but I bought the LEDs on Ebay for only a few dollars from China. If you search on Ebay for "405nm LED" you will find many hits, some even have a power rating of 50 to100W. Quality is quite variable but generally they work (maybe not at the specified rating) and prices are pretty low. Best regards, Jeroen

[@intexclouds9607]

@@HuygensOptics thank you for the reply, Sir... I mis-heard you saying 4 or 5 nm....so I was surprised, as it would come under extreme UV and its very difficult to even achieve such low wavelength.... Thank you once again for clarifying.

[@HuygensOptics]

Ah ok, sorry for my bad pronunciation. EUV is 13 nm and as you mentioned it is extremely difficult to make. If only EUV LEDs were available. That would make the people at ASML very happy... ;-).

[@intexclouds9607]

@@HuygensOptics 😄 true that.... thank you for the reply, Sir.

[@Axelito4201]

How did you do the masks?

[@extradimension7356]

I think it would be very interesting to shoot a Siemens star as an input image (mask) and then see the photoresist and etch of the 5:1 reduction ratio. Might be another way to visualize various limiting resolutions / MTF and potential aliasing characteristics of the whole system. Wondering if you get chromatic effects there as well ? Not just film thickness or regular diffraction effects / different colors ?

[@HuygensOptics]

It is possible to use a Siemens star, but generally it will show the limitations in resolution rather than show color effects. I am currently builing something with which it will be possible to engineer the observed color effects as desired. It will be discussed in an upcoming video, however it will take me quite some time to complete, since it will be my most complex project so far. cheers.

[@extradimension7356]

@@HuygensOptics That sounds incredible / superb ! I wonder with thin film interference you can predictively make different colors deliberately ? And Maybe devise methods of measurement. [I know some processes in micro machining one can machine 'rainbows" into a surface or deliberate corresponding colors/ wavelengths by control.]. Depending of fine-ness of features you could even do a color dithered plate of an image ? very intriguing also interested in how your photoresists interact with polarized light ? I think you could have some really cool Fritz Goro type imagery there. Good Luck / sounds cool what ever you are planning to do :-)

[@Dukey8668]

@@extradimension7356 Lippmann plates are somewhat similar to what you are describing.

[@extradimension7356]

@@Dukey8668 Thanks for mentioning that ! , Did not know of their existence. I'm reading about the Lippmann plates right now, (0.01 - 0.04 micron grain size , incredibly small) ~ Amazing interference effect with sandwiched reflective surface to replicate actual colors (from reflection) + "The developed and fixated diffraction patterns constitute a Bragg condition …" Lippmann won a Nobel prize for that in 1908 (apparently) , Interesting that they can't be replicated (has security applications) , would love to see one in person, (Wonder if people still make Lippmann plates today ? ). Seems a bit like good Daguerreotypes, hard to replicate on a computer screen but truly amazing to view in real life.

[@Dukey8668]

@@extradimension7356 There are still a few people making Lippmanns, mainly in this facebook group: facebook.com/groups/1422495841341822/ I'm no fan of facebook and you need to ask to join the group, but this is probably the best collection of Lippmanns on the internet.

[@RandomerFellow]

That lens is not easy to get hold of... Can others be used? If the quality requirements are slightly lower?

[@hypoallergeni]

What kind of photoresist did you use?

[@HuygensOptics]

The photoresist I used was Olin 00835 from Ciba Geigy. It was a sample given to me so I don't have any details. I also could not find any info on the Internet.

[@nicktohzyu]

why such a large distance between the mask and lens?

[@pizzablender]

The lens needs that, it has a long focal length.

[@weirdsciencetv4999]

Where did you order the photoresist?

[@HuygensOptics]

With MicroChemicals in Germany

[@weirdsciencetv4999]

@@HuygensOptics thank you! I will see if I can get some for my students

[@ramjetross]

Why not use a 405nm laser? You have some chromatic aberration with the LED. I suppose you'll have polarization and interference issues but that could be solved and averaged out by vibrating the laser slightly. Alternatively, you could use a band pass filter to narrow the LED spectrum or if your optics in the beam path sharply absorb below 405nm, like some of mine do, then you could use a short pass filter around 410nm to get a band pass effect. I'm also wondering how hard it was to focus the image onto the diffuser glass. Did you essentially have to level all for corners precisely in focus to get the whole pattern? Over all, awesome video and good methods.

[@HuygensOptics]

I tried using a 405nm laser and it gave quite a lot of speckles in the image. I guess, as you said, slightly vibrating the laser could solve that. But more importantly, it is not easy to blow up the laser beam to a size where it would illuminate the full surface of the condensing mirror. especially not if the laser is so close to the mirror surface. As for focussing: I set the diffuser as accurately perpendicular to the optical axis of the lens, then tweaked the angles of the mask a bit in order to get the image sharp in all corners in the image. Tweaking the mask is easier since it is 5 times less sensitive to movement, due to the demagnification.

[@Grak70]

Lasers have too much spatial coherence to use for DIY lithography. In industry excimer lasers, they add special optics to decouple spatial and temporal coherence.

[@sugar-alisportstv]

How do one 100% know that a resistor, inductor, capacitor, transistor, mosfet, and gate, or gate, or exclusive or gate has been etched to produce an actual working circuit?

[@montesumma4737]

you can make cotton candy with this thing

[@ikeetkreeft1973]

4:40 made me laugh

[@ONRIPRESENCE]

4:41 oops

[@fzigunov]

It's not DIY if you don't drop it on the floor!!

[@johnarizona3820]

^5

[@mohamedaboelenein7727]

Please, What light source did you exactly use?