Low-Temperature Additive Manufacturing of Glass
11 comments
·June 18, 2025gsf_emergency_2
https://patents.google.com/patent/US11499234B2/en
I wonder what the other inorganics ("functional additive, embodied in some cases by Silver flakes") are, exactly. In their flagship product.
And the size of the nozzle, why not the easier to remember 400 micrometers?
giantg2
Interesting. I wonder if this could be used to print telescope mirrors if you could get an "ink" that has high thermal stability, or if that would be in conflict with the low temperature printing.
sheepscreek
Maybe with further augmentation to the process. The 3D printing seems to leave pretty big grooves on the glass - which can distort the passage of light. For any real optical application, that will have to be removed somehow. Such as baking the glass structure in a high heat oven with a centrifuge/fast spinning base and some kind of mould to put the initial piece on.
The high heat and spin action could, in theory, fuse the individual grooves for a more uniform dispersion of light, and the mould shape would further press on the glass to maintain or alter its shape to be closer to a lens.
Finish that off with fine “sanding” and/or polish, and that might do the trick?
Honestly, experimentation might be the only way to find out.
hbrav
That seems like it would be very difficult. For a telescope mirror you want a very good surface finish, and that's one thing that 3D printing does very poorly.
giantg2
The polishing process could be done differently. If you could print the specific parabolic shape, that could save a lot of time hogging out. Although slumping might work better if surface ends up smoother. Im not sure of the annealing process might go.
bhickey
The parabola of a telescope mirror generally deviates from a sphere by tens of micrometers at best. Figuring happens after grinding and after polishing. Slumping is neat, but I've heard from the grognards that they can shatter if the temperature changes too quickly.
taneq
There's a few combined-process setups coming out these days that use 3D printing as a first pass and then machine the print in-place before sintering etc. which saves a ton of setup and fixturing time.
metalman
this is a highly missleading "anouncement" about sodium silicate "glass"(ish),stuff
https://en.wikipedia.org/wiki/Water_glass
but if you want to plqy with some, it is readily avail8ble as a "cement" used for attaching seals to wood stove doors, and countless other low teck applications the comon name of "water glass" refers to it's solubility, so.....
jiehong
Very cool indeed!
So it sounds like a single-component process, no powder bed or vat of hardener liquid. "Silicate solution" presumably means waterglass rather than something like tetraethyl orthosilicate, and if it hardens rather than foaming up at 250°, it's either because they're doing the process under pressure or, more likely, the other inorganics they mention are donating polyvalent cations such as Ca⁺⁺, Mg⁺⁺, Al³⁺, or Fe⁺⁺ to crosslink the silicate residues.
Probably the polyvalent cations are bound up in a salt that has negligible solubility at room temperature, so the ink doesn't harden in the reservoir, but which can react with the silicate at 250°, maybe a hydroxide or carbonate. Calcium sulfate is probably too soluble.
Presumably the silica filler is amorphous and serves to strengthen the glass and reduce the TCE, though it probably raises the cost. Maybe it also makes the paste thixotropic without needing to include organics like carboxymethylcellulose which would be hard to remove later. And I guess maybe it could react directly with the waterglass to solidify it, in effect raising the waterglass's modulus out of the water-soluble region, without requiring any polyvalent cations. Silica fume is the most likely form of silica here.
If anyone digs up more details, I'd love to see them.
Hmm, I see gsf_emergency_2 found this patent from 02020: https://patents.google.com/patent/US11499234B2/en which is later than Dercuano but earlier than Derctuo and Dernocua. But its priority date is from a provisional patent application from 02019, so even Dercuano doesn't count as prior art, even if it anticipates some of the claims.