Two pots that were inspired
Two beamlike objects that change cross-section along
The shape to the left is a computer model of inter-
After objects are printed,
As the firing temperature increases, so does the
These simple cups have been infiltrated with colloidal silica, fired, then glazed and fired
with recipes for Printing Slips
In the Solheim Rapid Manufacturing Laboratory (located in the Mechanical Engineering Building at the University of Washington in Seattle), our research focuses on new and improved methods to describe complex shapes in a way that a computer can “understand” and to fabricate those shapes in ways that the computer can control (a.k.a., rapid prototyping). Three-dimensional printing (3DP) is our favorite method of rapid prototyping, because the required equipment is not outrageously expensive and you can use just about any material that can be obtained in powdered form. While our initial research aimed to address a biomedical application (digital fabrication of alumina dental implants), it was not long before discussions with a co-worker led to consideration of other kinds of ceramics. (The university setting is nice, because the head of your receiving department, like our own Ben Jones, just might turn out to have an M.F.A. and lots of good, challenging questions.) This article presents the basics of 3DP and everything you need to know to put together the materials for producing ceramic art objects on a 3D printer.
About ten years ago, we embarked on a project aimed at using a new type of geometric model to support the creation of some very interesting shapes involving lofts or variable section extrusions. This would be like starting an extrusion with one die and ending with another, with continuous connection between the two. Traditional commercial modelers include some lofting capabilities, but major changes in cross-section (e.g. changing the number of holes) can cause traditional modelers to break down. If the software system used to represent these shapes required unusual flexibility, then the manufacturing system to produce these shapes would need to be unusually flexible as well. Enter 3D printing.
3D printing was invented in Emanuel Sachs’ lab at MIT and first became available in the early 1990s. One of the companies to license the MIT-Sachs technology was Z-Corporation, and they continue to produce 3D printers in a variety of sizes and capabilities. They can sometimes be found on internet auction sites for four-figure sums.
How Three-Dimensional Printing Works
Three-dimensional printing takes a digital model and produces a real, three-dimensional object by adding powder in layers and selectively printing a binder on each layer that causes the powder to adhere in the area of the desired design. Typically, the digital model specifies a collection of triangular planes that wrap the surface of the object to be printed. A common file format is .STF which can be exported by most 3D computer-aided design (CAD) software packages.
The actual build process goes as follows. The 3DP system’s software slices the object into layers ranging from 0.003 inches to 0.013 inches thick. A layer of powder matching the thickness of the digital layer (in our case 0.005 inches) is spread onto a build platform, or print bed. An ink-jet printing system deposits binder into the powder layer corresponding to the image of the current layer. The print bed is lowered, another layer of powder is spread, another slice is printed, and the system continues until all layers are processed. When the 3D print is finished, our object composed of bound powder is supported in a bed of unbound powder. We now remove the unbound powder to reveal our finished object by a combination of manual brushing, vacuum removal and compressed air (see p. 38). At times, one feels a bit like an archeologist at a dig site—and often with just as much excitement.
After an object is removed from the bed and de-powdered, one of a variety of post-processing techniques may be employed to “finish” the object, depending on its final use. Post-processing options include wax infiltration, epoxy infiltration, CA (CyanoAcrylate) glue infiltration, elastomer infiltration, or painting. These final steps often enable the part to function as a true prototype rather than just a form-and-feel object. For our ceramic-slip parts, post-processing consisted of kiln-firing, glazing, and glaze firing.
Adapting 3DP to Ceramics
We recently had some students experimenting with casting metal into a low fire (cone 06) slip locally known as Xtra-White (from Seattle Pottery Supply). Since we had this slip on hand in powder form, it seemed like the obvious initial choice. We loaded the printer with Xtra-White slip powder, and used an existing alcohol-water solution as binder. It seemed like a good first test as simply mixing slip with water and letting it dry produces a functional greenware. Let’s just say our first 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 slip powder spread extremely well and had a very nice surface finish on the printing-bed surface.
We needed to find a water-soluble “glue” to add to the slip powder to give strength to the printed parts. The “glue” could also be added to the printing fluid, but keeping the glue in the powder bed avoids problems with clogged print heads. Again, the number of choices we have for in-powder “glues” is quite large, with PVA (PolyVinyl Alcohol), PVAc (PolyVinyl Acetate), SCMC (Sodium CarboxyMethyl Cellulose), PolyOx (Polyethylene Oxide) and various carbohydrates being high on the list of choices. Previously, we had run hundreds of test powder mixtures along with various water/alcohol binder setups. The process is not too different than glaze or clay formulation experiments (choose a test shape that is significant and keep good notes.) Based on experience, experimentation and a strong desire to produce a low-cost powder, we settled on a combination of PVA (PolyVinyl Alcohol) and extra-fine sugar (from the baking-supply aisle of the grocery store) as powder additives with the Xtra-White slip powder, along with an alcohol-water binder. After many more test runs (each one with different printing parameters), we finally succeeded in printing parts that could be removed from the 3D printer bed and depowdered. We focused on test bars that were 10×10×100 mm as they printed quickly and didn’t require large quantities of powder to be mixed. Now, it was time to test fire the bars (in lots of five). Since the Xtra-White was a cone 06 slip, it seemed that a cone 06 firing (1828°F) was in order. When the test bars were examined after the firing, they crumbled to the touch and exhibited minimal strength. We continued firing more test bars (and gathering additional data) at increasing temperatures until the test bars melted to the kiln shelves. Having determined proper parameters for printing and firing, we were able to move on to fabricating simple functional shapes.
Lastly, we arranged for engineering testing of about one hundred test bars to determine how the firing temperature affects both shrinkage and flexural strength. We thank our staff engineer Bill Kuykendall for his assistance. The results, presented graphically in the graphs on p. 38, allow for determination of a design point in the space of shrinkage, strength and firing temperature.
Additional Post-Processing Steps
With success and a bit more understanding of using the 3DP process with art-ceramic powders, we continued to explore more interesting object geometries. As the resulting fired objects are light and quite porous (they are essentially ceramic sponges), we discussed various ideas for infiltration processes to reduce the porosity: infiltration with original base material slip, infiltration with colloidal suspension silica, and direct application of glaze. We chose to try infiltration with colloidal suspension silica followed by an application of glaze and then a glaze firing with quite good results. Lastly, we tried direct application of glaze. The results improved with either thicker glaze application or multiple coats.
Results and Conclusions
It is clear that 3DP can be used to create ceramic-art objects, out of three different types of slip bodies, and can be finished using standard ceramic equipment and processes.
We hope that at least some readers will be excited enough about the possibilities of this new approach to creating ceramic art to give 3DP a try. You should find that your 3DP objects readily accept both infiltration and glazes, and 3DP offers a variety of advantages to ceramic artists. You can print many copies of the same object; you can print many different objects at the same time; you can print interlocking/interconnected geometries; you can print objects in different sizes within a given print; you can print objects in different materials; and 3DP can provide increased access to ceramics for a broader practitioner base.
Our future work includes infiltration of the post-fired bodies with liquid slip (or possibly terra sigillata), continued adaptations of other slip bodies and additional engineering testing to determine how shrinkage, strength and porosity depend on firing temperature data. Finally, we wish to express our appreciation to the National Science Foundation for supporting the research that led us to ceramic 3DP. Now, find someone with a 3D printer and print something with clay!
the authors Mark Ganter, Duane Storti and Ben Utela work in the University of Washington Department of Mechanical Engineering in Seattle, Washington.
PVA Printing Slip
Xtra-White, Redart TerraCotta
MaltoDextrin Printing Slip
Xtra-White, Redart TerraCotta
The PVA Printing Slip mixture produced quite acceptable results (with all slips) but the PVA is a little more costly when compared to MaltoDextrin (which is available at the grocery store under the brand name Benefiber). The MaltoDextrin Printing Slip was also stronger in greenware form.