Currently, 3DP suppliers provide a limited selection of materials. We decided to knock down that barrier by experimenting with a wide variety of materials including glass. To do so, we would need to obtain a glass powder (or possibly glass microspheres), decide on appropriate binding/sintering additives, and design the appropriate printing fluid and develop a kiln firing schedule.

We loaded up the 3D printer with a glass powder (initially we used recycled glass), and used an existing binder solution. It seemed like a good first test as simply mixing ceramics with water and letting them dry produces functional greenware. Let’s just say our first 3D printing tests were not terribly successful. The parts were so weak that any contact caused crumbling, and we could not remove the parts from the powder bed. However, the glass powder spread extremely well, produced a nice surface finish on the printing-bed surface, and resulted in a densely compacted build bed (as compared to our previous work in ceramics). Thus, we needed to find a water-soluble adhesive to add to the glass powder to give strength to the printed parts. (The adhesive could also be added to the printing fluid, but keeping the adhesive in the powder bed avoids problems with clogged print heads.) {We favor the binder being simply a solvent.} The parts needed to be strong enough to survive general human handing, depowdering, and kiln firing. Again, the number of choices we have for in-powder-bed adhesive is quite large – any water or organic soluble adhesive with appropriate properties would be applicable. It should be noted that the binder and/or glass powder may contain chemical colorants that induce a color change both in the pre-sintered parts as well as a different color in the post-sintered parts.


Early test prints in recycled glass.

Previously, we had run hundreds of ceramic test mixtures along with various water-alcohol-polymer binder setups. After more test runs (each one with different printing parameters), we finally succeeded in printing glass parts that could be removed from the 3D printer bed and depowdered. Our big realization was that unlike all other 3DP materials glass did not absorb moisture and thus a different protocall was needed. As with our previous work in ceramics, we focused on test bars which were 10x10x100 mm as they printed quickly and didn’t require large quantities of glass powder to be mixed. Now, it was time to test fire the bars. Since glass has unusual material properties such as no defined melting temperature and its amorphous structure, we knew that success in glass for 3DP would depend on proper sintering/kiln firing schedules. Our initial sintering tests resulted in completing melting the 3DP glass shapes. We continued firing more test bars (and gathering additional data) at a variety of firing schedules until the test bars “crumbled to the touch” and thus exhibited no signs of sintering. We then explored a series of interim schedules until we arrived at the minimum amount of sintering (and thus the least shrinkage). Having determined some acceptable parameters for printing and firing, we were able to move on to fabricating simple functional shapes (a cube, a UW puck, and a castle turret).

Glass Samples

Various sample shapes in recycled glass.

Additional Steps and Results

With success and a bit better understanding of using the 3DP process with glass powders, we continued to explore more interesting object geometries (see work by artist Meghan Trainor).

M. Trainor textured cylinder.

M. Trainor’s textured cylinder.

The resulting fired objects are light and quite porous (they are essentially glass sponges), we discussed various ideas for infiltration processes to reduce the porosity (if desired) – infiltration with various polymers (such as polyester, polyurethane, epoxies, and CA), ultra-low fire ceramic glazes, and infiltration with colloidal suspension silica (or other glass formulations) plus any number of organic or inorganic compounds appropriate to a specific use (e.g. post-sintered parts are porous enough to act as a oil lamp wick). Lastly, we note that porous glass is among the materials that provide an environment which is conducive to cell growth and thus, post-processing may include introduction of living cells.

8 Comments on Glass 3DP…

  1. glen gardner says:

    What keeps the glass shapes from collapsing while sintering?

  2. ganter says:

    I tell students that either magic or little tiny fairies hold the particles together during sintering. After the organics burn off, there is some carbon reside left. Depending on the shape, sometimes an object fires fine without needing support. For others, you need to provide support in the form of an inert powder bed (basically anything that doesn’t melt during firing and doesn’t interact to the object). Think oxides and/or salts.

  3. Jeff Crowe says:

    I have been reading your blog and have a few questions. What temperature range are you using with the glass? How much does the object shrink within this temperature range? You mentioned that the objects are porous like sponges, can light shine through the object? The objects you modelled are solid, can they be altered prior to firing without breaking the object? For example, could you drill a hole through it, in the green state and then fire the object?

    I work with art glass frit. I am interested where your parameters are in relation to the what a studio glass artist uses to make objects.

    Thanks. Jeff

  4. admin says:

    Jeff, We have been sintering in the 1200 deg F range (but this depends on the type of glass that you are using). Also, the length of time effects end results. Shrinkage is about 8-14% depending on the temp and time.

    Shining a light works for thin pieces. We could generate more porosity by including more organics (think corn starch, etc) into the powder bed as they would just burn away to produce more empty space.

    Yes, you could machine (if your careful) and drill holes (if your careful). We sand, file, cut these objects in their green state.

  5. Jeff Crowe says:

    It seems to me from what you are saying , that the voids are contibuting to the light penetration rather than the glass material. In kiln formed glass, 1200 degrees is a cool temperature to fuse. Some high lead glass would melt more at that temperature than say a Bullseye soda lime glass. From what was posted in another blog post, I gathered the recycled glass you are using is closer to a soda lime type glass rather than a lead crystal type glass. Artist using the pate de verre technique can experence a shrinkage range closer to 50%. The glass could be fired in the 1400 degree range. At that temperature, the glass would contribute more to the light penetration than the voids. Also higher shrinkage. I suppose at that range a unsupported shaped sample would collapse.

    If I remember correctly, it was stated that the brownish color of the test pieces, was due to the organic fillers. Is a sample of the glass ( without fillers) fired in the same fusion program more of a white color? Is the pinkish color on Trainor’s sculpture, due to a oxide addition?


  6. admin says:

    Jeff, how about a do a firing run at about 1375 def F. We found a random green part in the lab. We will give it a try.

    The best we have gotten is a beautiful white (opaque) using the very best glass that we can obtain.

    The pink color is indication that the part is a green part. The pink is really red ink into a white powder bed. Thanks for checking in. We will report.

  7. lornetki sklep says:

    What blog script do you use on your site ?

  8. admin says:

    We are using a Turnkey Appliance Word Press (V10/09) which has been upgraded to WP 2.9 with Amazing Grace as the template.

Leave a Reply

WordPress Appliance - Powered by TurnKey Linux