Ceramic Resin Testing


#1

Working on developing a custom alumina ceramic for some engineering projects. It’s a lot more fun than I expected it to be, learning about the finer details of SLA resins.

Testing


#2

Hi there,

Are you starting from scratch? We’re doing some work with a different ceramic material but using a Form 2 to print. There are some great resources from professor John Halloran and his students at the University of Michigan on development of ceramic 3D printing by SLA. Also, a great base resin to work with is Tethon 3D’s Genesis resin. It’s a formulation of monomers and a photoinitiator that you can mix with your ceramic powders. Works well with the Form 2.

Andrew


#4

Hello there,
kind of bumping this topic, because we are trying to develop a composite resin. Has anybody made some experience with fillers in Formlabs standard resins? Has anybody done some experimenting with adjusting the viscosity of the standard resins? The next step might be the PR48 formulation by Autodesk, where every component of the resin can be adjustet. I am very interested in some exchange about this topic.

Stefan


#5

Hello Stefan,

we did some tests with filler materials in white resin. If you have globular particles with a particle size smaller then your layer height it “could” work.

However: Particles may absorb the UV-laser light and hence prevent the resin from fully curing. Some materials like High Temp use a higher laser exposure, so you might be able to compensate by using those resin profiles for your prints.

Unfortunately Formlabs does not offer OpenFL on the Form 2. Therefore, we had to bury our project (and a new potential product line with it).


#6

Thanks for the reply, the plan so far is to use globular particles with a diameter around 100 nm and low absorption. My plan is to print a casting template for small test specimen from Flexible Resin. Then fill them with some differently filled resin samples and just cure them without the printer. Then heat treat them and see which mechanical and surface properties the results yield. Hopefully this can save some time and effort compared to printing every specimen. Of course the results will always be fully cured, if the respective samples will work with the Form 2 is another story.

One relatively big problem right now is to determine viscosity of the resins. Formlabs states 850 - 900 mPas for their Clear Resin, but my findings are more in the range of 1300 mPas, both at 25 °C. I’m leaning towards only comparing the viscosity of filled resins relative to the Clear Resin as standard, because I’m mostly interested in their behavior with the Form 2. I am looking for some ideas to change the viscosity of the Clear Resin. I am thinking about using the reactive diluent from the Autodesk PR48 formulation to compensate the change in viscosity caused by the fillers.

I am pretty much at the beginning of this project :sweat_smile:

I don’t know OpenFL, I am using Autodesk Tinkercad right now, but I am using fairly simple geometries.


#7

Ok, so using the Flexible Resin as a molding form for Clear Resin works. After two hours Form Cure the sample is hard enough to get it out of the form. Unfortunately, the form cracks very easily after curing the resin in it.

Homogenization of samples will be a problem, did anybody experiment with small samples (~ 1 mL) of filled resin? How to homogenize?

I decided I will keep updating this topic. Any questions, help, suggestions will be greatly appreciated.

I have, by the way, looked at OpenFL. That would have been a very valuable tool.


#8

Dear Seriously, at which point are you with your experiences? Any luck? It would be tricky to homogenize 1mL of resin. And it depends of course on particle load ratio. You may want to use an ultrasonic bath. On the other hand, I don’t know if the viscosity is a problem. Formlabs’ ceramic resin is pretty viscous compared to unfilled resins and it doesn’t seem to be an issue if proper process parameters are used (Formlabs defaults of course). Instead of making casted samples I suggest to print some small details directly. Of course, it needs much more resin at more time, but final properties of your material will depend not only on formulation but also on process parameters and you may (and certainly will) get very different results for the same material cured in mold and in printer.


#9

The homogenization is the main concern so far. I backed away from the small sample sizes of 1 mL, those would only have been useful for casting. As you pointed out, any casted samples are not comparable to printed ones, I have found that out by now.

I have tried the Autodesk PR48 resin without any particles with a ClearV4 setting in open mode, that printed flawlessly. I plan on filling PR48 with particles, because it has a relatively low viscosity to begin with (~400 mPa s vs. ~900 mPa s on Clear V4, according to technical data sheets). Also, visco can be slightly adjusted by the amount of reactive diluent, I hope.

Determining viscosity doesn’t work really well right now at our lab, that is a problem I want to solve before using self-made resin.

Homogenization and milling of particles, and precise determination of viscosity are the main obstacles right now. I reckon that viscosity is not THAT big of an issue in open mode, I will just try if it works. If not, I plan on cleaning a Formlabs CR cartridge and refilling it with self-made resin, to be able to use heating and the wiper. For that, precise determination of viscosity is absolutely mandatory, or I could ruin the printer, I know. But that is sort of a plan B.

Currently making first experiences with Formlabs Ceramic Resin, took a while to arrive. First print is running right now. I am planning on determining particle size of the Formlabs CR with a zetasizer and take that as a reference value for particle size.


#10

I think I saw somewhere (ebay?) a universal cartridge. It comes with a kind of programmable module so you can change the resin type recognized by printer. Costs something less than 60 euros. By the way, do you have any idea of particle load ratio in ceramic resin? I think since it shrinks 15% in all directions it should be something about 85% of particles?


#11

I don’t know if I am allowed to link it here, but I think you mean the one you find by googling “3rd party universal cartridge for Formlabs Form 2”? Thinking about buying one. I would have guessed about 80-90 % particle load by shrinkage, too. I am planning on weighing the green and the sintered part I am printing today, and maybe use thermogravimetric analysis in the coming weeks on the Formlabs CR.

I will occasionaly report findings about CR properties here when I have some. Although I am not sure if Formlabs is OK with posting properties of their proprietary formulations here.


#12

The first print I finished with Formlabs Ceramic Resin had a weight loss of 37 % from green to sintered, so apparently particle load is somewhat lower than expected. Going to confirm with TG analysis in the coming days.


#13

I picked up a “universal cartridge”. Haven’t gotten around to playing with it yet but I can attest that the seller is a stand-up guy - the process of buying, shipping, etc. was a pleasure.

Not sure how much value this adds to the conversation as my testing wasn’t very scientific, but in terms of mixing I’ve had great success with “nano-dispersions” from Peacock Colors and FP-600 (flame retardant) from Adeka. Got some photos buried in the forums somewhere.


#14

Yes, I’ve got 35% on 10 samples. However, the shrinking is not homogeneous by Z and in-plane. I’ve got 20-24% by Z and only 17% in-plane.


#15

The Formlabs guide to ceramic resin explains this as extra shrinkage due to gravity in z direction. The new PreForm version accounts for it, you can scale the model in z direction independently.


#16

Yes, but the problem is that even in-plane shrinkage is higher than expected. 5% of dimensional deviation (17-18% compared to correction factor 1.13) is too much for our application. By the way, do you have any idea of particle size? I am trying to load the standard resin with particles but the problem is that laser beam does not reach the platform even at low concentrations. I guess that is because my particles are too small and stick to the vat and disabled wiper doesn’t make it any easier.


#17

I tried to determine the CR particle size by diluting 1:100 in IPA, but the particles seem to agglomerate. My understanding is that a particle size smaller than the wavelength of the laser is needed, otherwise the particles act as a photoblocker.
Initial particle size was 700 nm, after 1,5 h I found 3000 nm. I guess the primary particles as dispersed in resin are smaller.

It is a real problem that particle size can not be determined while dispersed in resin only. Any diluent might change particle size.


#18

700 nm is already nearly twice laser wavelength. My particles are 1-2 µm (according to datasheet). What method do you use to determine the size? SLS/DLS? If so, you might need to dilute CR in something less polar than IPA, toluene for example. It should dissolve well liquid resin. And probably you may want to use smaller concentrations, say 1:1000. As an ultimate solution and if the particle size is really important, you may use TEM. I have some experience with colloid systems, but at my current workplace I don’t have access to any analytic equipment :frowning:


#19

HI Photoengineer,

Any luck? I am looking to work with industrial ceramics like alumina and zirconia. Any tips would be greatly appreciated.

Or better yet a source for materials, printing and firing parameters would be great also. I will buy the materials from anyone that can produce a quality raw material to work with.


#20

@azaitsev I am using DLS for particle size. 1:500 and 1:1000 in toluene are inconclusive. Average size is 4 µm with a very wide size distribution, so most likely agglomerates. There is one specific peak that correlates to 240 nm particles. I would take it as a hint, not a conclusion.

We don’t have a TEM available, but SEM is an option in the next weeks. Thinking about a way to prepare a sample, maybe applying a very small sample size to an object stage and then carefully burn the organic phase.

@tkimball Tips: Use nanoparticles. Think about your possibilities to disperse them in resin (milling, ultrasound, …) before you start experimenting. The rest of your questions are basically my questions, too.


#21

I read a paper that concluded random particle size produced stronger parts than uniform in conventional manufacturing of alumina ceramics.