3D Printing With Chemicals

From the horse’s mouth,

“In this lithography experiment light creates free radicals from phenylbis(2,4,6-trimethylbenzoyl)phosphine oxide which induce polymerization of 1,6-hexanediol diacrylate.”

Or for those without a Chemical Engineering degree, light from a (high resolution) projector interacts with a special liquid, producing a hard polymer on the surface. A platform within the liquid is lowered, taking the layer of polymer with it. Shine the projector again to produce another layer: lather, rinse, repeat. Long story short, an atypical 3D printer using light on a very small scale.

You get the chemicals and lab equipment, we’ll get the laptop and projector, and for goodness sake [Jimmie] stop bumping the table.

[Thanks Tomas]

45 thoughts on “3D Printing With Chemicals

  1. I think this is really neat because with the proper controller on that you could theoretically have a 3D prototyped part without visible (or possibly any) stepping.

  2. After I installed Quicktime, I got movies.

    I wonder if you could continuously lower the
    platform and play 60 frames-per-second of
    video at the upper surface and get much quicker
    building and smoother objects out of it.

    Somehow, I think you’d need the delay t get the
    liquid to fill-in, but I can’t tell from the
    videos how viscous it is.

  3. Either way, you’ll should be able to put a (step)motor on the platform, which is controlled by a computer.

    The thing I am thinking on, is it the white light or the heat from projector, that do the trick.

    If it is the white light, you’ll be able to cut out the magnifying glass and the projector, and replace it with a LCD screen. To Make it more portable.

    But I think this is really cool, and perhaps I’ll in some future i’ll try to build one at home :D

  4. what’s the difference with machines like Envisiontec Perfactory? Home or student made, great, atypical… debatable. Maybe they figured a new material or something, but the concept of “projecting into liquid to make it solid” is old.

  5. what’s the difference with machines like Envisiontec Perfactory? Home or student made, great, atypical… debatable. Maybe they figured a new material or something, but the concept of “projecting into liquid to make it solid” is old.

    This is exactly the same technology that EnvisionTEC / Zcorp (which licensed the tech and private labeled it from EnvisionTEC) use. It has been commercialized.

    The material they are using is a fairly low quality UV resin. Nothing special here. Neat implementation though and it echos what I have been saying for a while now about DIY 3d printing. This is lightyears beyond the current makerbots.

  6. so a uv laser, an lcd, and a couple glass fresnle lenses? I am REALLY amazed nobody is doing this. sources for UV epoxy anyone? 3d fabrication with all of 2 moving parts? sign me up….

  7. As everyone is saying this process is pretty common: Stereolithography is the term most commonly used in the 3D printing realm. FYI, 3d-prototyping solutions:

    SL/SLA: Stereolithography: A liquid bed is hit with precise amounts of UV light, causing the liquid polymer to cure where the light hits. The bed is then lowered and the next layer is cured. SLA allows pretty incredible surface finishes and one of the lowest wall thicknesses of the rapid prototyping solutions.

    LS/SLS: (Selective) Laser Sintering: A high powered laser hits a powdered metal or plastic causing it to sinter (fuse). After a layer is made a new layer of dust is cast onto the part

    FDM: Fused deposition modeling: A plastic is extruded through a heated head which moves around. The heated plastic is deposited to build the part. This creates an obviously layered part which loads poorly in shear.

    Then there are more proprietary solutions like polyjet and zcorp.

    Is anyone here working on a reprap? I’m hoping to make a reprap with an interchangeable head that would allow use for FDM, a mini-mill, a laser cutter or etcher, and possibly LS or SLA. Ideas?

  8. @PolyJetter
    lightyears ahead? if you look at the paper this particular method can’t even print a non-warped object more than two centimeters or so; repraps/makerbots most certainly can :P
    also I’d consider materials cost as a factor in comparison.

    I haven’t seen anyone put a mill or a laser on a mendel (current reprap design) but it’s an open design, no reason not to go for it.

  9. The underlying tech is lightyears ahead in terms of final, finished quality of parts. This particular implementation may not be (yet?) but commercial versions are lightyears ahead of repraps/makerbots/etc. They include such things as automatic support generation, automated recoating, better quality materials (rather than using only a difunctional monomer like they did here) etc.

    “why not to add a drop of liquid for each frame,
    instead of a lowering the platform ???”

    That, in essence, is polyjet technology. It has been done commercially since the early 2000’s

    “and where is the Hexanediol diacrylate used ?”

    It is the vat of slightly yellow colored difunctional acrylic resin they project light into. In essence, that plus a photoinitiator + UV light = cured resin. In reality, it is a bit more complicated because you need the right spectrum of light for the right photoinitiator, you have to deal with oxygen inhibition, etc.

  10. “The thing I am thinking on, is it the white light or the heat from projector, that do the trick.

    If it is the white light, you’ll be able to cut out the magnifying glass and the projector, and replace it with a LCD screen. To Make it more portable.”

    It is the UV light actually. So, in a way, it is the white light. You *CAN* polymerize UV reactive chemistries with heat alone but it is markedly easier to use, you know, UV light. You need a lot of concentrated UV light though – if I put my hand in front of the light on my objet printer, I would very quickly get a deep sunburn. It is many times more intense than the sun. Well over 100 fold.

    An LCD screen has nowhere near that level of UV light and I suspect that you would have a difficult time modifying one – my understand (which may be incorrect) is that LCD screens don’t take well to UV light. Plus, most backlights don’t produce much, if any, UV light in the first place.

  11. What I *AM* arguing however is that there is no compelling reason why, with some proper DIY ingenuity, Repraps (or something akin to them) using photopolymer and projector technology could make parts as good as the best commercially available technology but at a very low cost – and be highly useful parts. Not sintered metal but a quality plastic with an exceptional surface finish and amazing detail. Think about it. Z axis stepper, some decent software and a projector. There isn’t much to it and the resolution and build times would be superb. As good as a $40,000 machine (which is what a EnvisionTEC costs now) but for perhaps $1500 in parts and starter material. This is something you could easily build as a kit using an off the shelf USB powered servo Z axis and VGA or DVI or HDMI or possibly even USB projector since it is just still images.

    Bonus points for linux capability!

  12. While eye and skin protection is very important when working with any UV source, they are using an off the shelf (as far as I can tell) projector. I believe the risk to yourself or others to be very low from such a source.

    Honestly, I think they may even have modified the bulb to get enough UV density to cure their difunctional monomer. That or they have the photoinitiator levels at absurd amounts or they removed the UV filter.

  13. As mentioned it’s all old stuff but nice to see explained for people and to hopefully give people ideas. I’m not a big fan of such things though, for anything I do I don’t need plastic protos, I need the real thing made from a lump of cold hard metal, all a plastic copy would do is confirm what I have on screen and cost me a fortune :)

  14. Nice! A telecentric lens mounted on the projector could dramatically help getting precise parts as the projected image does not change size while moving the platform.

  15. I’m surprised that this works so well with a stock projector, I’d have thought a DLP + retrofitted YV source would be a prerequisite. Looks like all you need is to build the simple stand rig, and prepare the required chemicals. Pity the paper describing that critical process is paywalled.

  16. I looked up the chemical cost on Aldrich and its not cheap for small amounts. ONE of the 3 needed is £27 for 10g ! The other is £24 for 100g.
    I didnt see any detailed mixing information to know what ratio these go in, but if the more expensive one is the main ingredient, it would cost far too much to make anything with this.
    On the other hand, the resolution available is great!
    I may just order the chemicals and see how it goes!

  17. @Pyrofer:

    “After the object has been made, exposing it to UV light will cause any unreacted chemical to polymerize. This can be achieved by taking the object out into the sunlight. After this occurs, the objects are safe for students to take home.”

    Also, as to scale: They run the output of the UNMODIFIED projector through a reducing lens to make the image smaller. This gives them enough UV light to make objects, with the disadvantage of those objects being small (they have a picture of a printed Lincoln Memorial sitting on top of a penny). However, with an adequate UV projection system, this could easily scale up.

  18. what’s with the comparison to commercial machines? this setup is really simple and looks like it can be replicated pretty easily using stuff we can get hold of (o.k, the chemicals might be tricky). also looks like automation might be possible with motors etc. good, informative hack, we can use this.

  19. “on the order of 400 nm” layer thickness. That’s pretty amazing. Slow but amazing. 1 nm = 1000 microns. Human hair is 100+ microns thick. So 400 nm is .4 microns.

    The best Objet or EnvisionTEC machines do about 15 microns per layer in quality and 30-50 microns in high speed modes.

    Looks like they are using an off the shelf projector, just concentrating the output to a small size to capture enough UV.

    These chemicals are expensive because they are being bought from Sigma. EVERYTHING from Sigma is overpriced.

    You do need to either accurately! move the projector when you move the Z axis – or – refocus before exposure – or – drop the stage down into the resin so your “exposure” is always on the same plane and you need to refocus at the start of the next build when you take the model out (and thus change the volume of resin).

  20. the reason that commercial versions of 3d printers are spendy is you’re paying for the r&d that went into them.

    maker bots while cute, can’t get near the resolution of commercial machines, and if you add in the hours an engineer might spend on trouble shooting, or even just a decently paid tech, we’re prolly put in close to 3K or more and still aren’t getting a great print because of inherent slop in the machine.

    repraps are in theory better, but again you’ve got hours of personalized calibration, and if you’re geekly enough to be able to assemble one from the (last i checked) spotty documentation, you’d probably be billed out at $100 an hour, meaning that in a used commercial model ready to go and calibrated is available for less time than you spent building it…. not that the build isn’t the fun of it.

    this is the first home brew sort of uv cured printing i’ve seen yet. pretty slick for a hack.

  21. Makerbots / RepRaps are an inherently limited tech. You can only make the “glue gun” so small. And – you can’t perfectly fill in pixels with circles.

    Same goes to commercial Dimension machines as well, in my opinion.

  22. Hello, I am the author of the process described in this post. I just came across it thanks to a reader.

    Hello, I am the author of the process described in this post. I work for an NSF funded center and my job it to take the research from our center faculty and make it accessible to students in secondary schools. The primary motivation is education and to inspire future scientists and engineers. This process is based on some research from Dr. Nicolas Fang here at the University of Illinois. He is making very small objects – on the order of microns to a few millimeters. Our method is a distillation and simplification of his method scaled up so students can see the results.

    We have the procedure described on our website. On that site we have video of the process as well as the files we used to make several objects. These are PowerPoint files that will also work in Open Office. The site is:

    One thing we have modified since our article was published is the solution we use. The new solution works much better and has resulted in much more success. I would suggest using this instead of the one described in the article until you are more comfortable with the procedure. The website I reference above uses the OLD solution and not this new recipe. The recipe for the solution I would suggest is below:

    98 ml Poly(ethylene glycol) diacrylate CAS 26570-48-9 Sigma Aldrich # 437441

    2 g Phenylbis (2,4,6-trimethylbenzoyl)phosphine oxide (Irgacure 819) CAS 162881-26-7 Sigma Aldrich #511447

    0.02 g Sudan I CAS 842-07-9 Sigma Aldrich # 103624

    The solution requires several days to dissolve – make at least 3 days before use, longer is better.
    Solution should be stored in an amber bottle.
    Solution will store for at least a year with no degradation.

    Several people have asked what lens we use. We use a Plano-Convex Lens 50.0mm Dia. x 150.0mm focal length, MgF2 Coating that we get from Edmund Optics, NT32-975. This is just the lens, if you want a holder as well; we use their holder M54-986 with a mounting post M54-940.

  23. Hi Joe
    Great technique, good work coming up with that, but I noticed one of the chemicals is marked as a possible carcinogen.
    If introduced into the classroom, its probably something some irresponsible teenagers perhaps shouldn’t be exposed to. I dont know.
    It would be nice to work towards using safer chemicals though, but the technique is still great.
    Any reason why this couldn’t be scaled up to larger object sizes, maybe with a different lens as you could make objects up to sizes fractionally smaller than the beaker.

  24. The 0.02 g of Sudan I being used is labeled (among other things):

    Limited evidence of a carcinogenic effect
    Possible risk of irreversible effects

    The Sudan is the dye being used to prevent over exposure. It is possible to use another dye.

    Also keep in mind though (not that this justifies using Sudan I) that Sudan I, aka 1-Phenylazo-2-naphthol, is chemically similar to the artificial dyes used in things like Tang. Not that I am arguing that artificial dyes are healthy by any means. But as long as you don’t plan on drinking the dye, avoid skin contact and you take reasonable safety precautions, you are probably ok. Most UV chemicals have a fairly low vapor pressure, something on the order of 0.1 mmHg or so, which means they don’t really evaporate (at room temperature) much compared to say, water.

    To get an idea of what vapor pressure is, if we were to place a substance in an evacuated, closed container, some of it would vaporize. The pressure in the space above the liquid would increase from zero and eventually stabilize at a constant value, the vapor pressure. The point at which most UV resins “stop evaporating” so to speak means they tend to not give evaporate huge amounts into the air.

    Both the artificial colors known as Yellow 5 (Tartrazine) and Yellow 6 (Sunset Yellow FCF) are azo compounds. Essentially they are sulfonated versions of Sudan I (the dye being used in this experiment). DayQuil uses Yellow 6 for its color.

    Kinda makes you want to avoid artificial dyes in your food, doesn’t it?

  25. Joe:

    What is the purpose of the Sudan I dye? I used for formulate UV/EB coatings (15 yrs) so this intrigues me. I’ve seen UV stereolithography demonstrated at past RadTech conferences and from a couple suppliers a couple years ago.

    When I read the initial article, I was wondering to myself “why is he using HDODA???” My experience has been that it’s pretty hard/brittle (for a diacrylate) and that there are other (better) monomers to use than HDODA. I would have used TRPGDA, DPGDA, or PEG(400-1600)DA or thrown a resin like Ebercryl 3211 in to give it some “guts”. Irgacure 819 is a pretty good choice but if you’re seeing a solubility problem with the photoinitiator, you could also try TPO-L (Ethyl – 2,4,6 – Trimethylbenzoylphenylphosphinate which is a yellow liquid TPO). You should just be able to stir it in.


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.