Design and Additive Manufacturing of Periodic Ceramic Architectures
Design and Additive Manufacturing of Periodic Ceramic Architectures
Abstract
Cellular ceramics are attracting material solutions for high temperature applications because of their
outstanding effective properties. Even though ceramic foams are already widely industrially
employed, they show scattered properties because of their randomness and fragility. Moreover there
are few parameters which can be varied in order to engineer their properties.
We first show how, through finite element simulations, periodic cellular ceramics can be designed
upon user’s requirements.
Then, thanks to additive manufacturing (AM), the aforementioned numerical domains can be
readily transformed into physical objects. Among many AM techniques nowadays available, we
developed an original method in which polymeric lattices templates are produced by 3D printing,
further coated by replica with ceramic slurries and finally heat treated. The advantage of this
technique is its flexibility. Practically any ceramic material already produced in bulk form can be
realized.
We then present some case study where we were able to design and produce components for the
concentrated volumetric solar receivers, thermal protections for aerospace entry vehicles, and heat
exchangers.
This very randomness could be, on the other hand, a limiting factor. The polyurethane foams used
as templates are deriving from a liquid phase in which, before its polymerization, gas bubbles of
different size are growing inside. Because of that, the struts, i.e. the skeleton of these foams, are
random in shape, length, distribution and orientation in the space (Figure 1). As a consequence, the
20
local properties of the porous body dramatically change along its volume. Foams’ properties are
thus calculated using homogenization techniques in which the selected volume has to be rather large
21-23
in terms of number of pores in order to be treated as a representative one . Moreover using
Figure 1 Stress distribution (blue low, red high) from the FEM analysis of a random foam under
compression (left)23 and bending (right) 26.
According to users’ requirements, to be fully reliable, macro porous ceramics components should
be realized as structures with a repeatable architecture.
With random foams this is impossible.
Another issue in random ceramic foams is the shape of their struts; this has huge consequences in
23
foams’ mechanical properties . Replica is accomplished employing polyurethane foams which,
because of their manufacturing process 5, have their struts with piercing corners. At the end of the
process each ceramic strut presents an hollow shape with sharp corners imposed by the former
polyurethane foam. Sharp corners act as notches from where, under mechanical or thermal loading,
a crack can easily propagate (Figure 2) 27.
All these issues could be solved if a periodic cellular architecture with smooth, regular and
interconnected struts would exist.
This need was met with the advent of additive manufacturing.
Additive manufacturing or 3D printing is a near to net shape production technique developed
several years ago which allows the assembly of a material starting from a CAD model. The main
advantage of this technique is that it can allow the direct production of components whose shape is
impossible to realize with the conventional manufacturing methods. This is especially true for
porous materials. Presently this technique is industrially mature for metallic and plastic materials
while it is still in a development phase for ceramics. Our idea was thus to combine plastic AM, the
most mature, with replica 28.
The following paragraphs of this paper will describe how CAD files of cellular architectures are
generated and how they can be converted into ceramic objects. Finally some application in different
industrial fields will be reviewed.
One can immediately see in Figure 3 that the first two structures already solved the problem of
having unconnected struts (Figure 1), this feature rises the mechanical properties of the resulting
20
structures . The next advantage is that the struts can be designed to overcome the problems
highlighted in the introduction 22, 27, 30. We tested several strut’s shapes, but the easiest solution is to
assign to the strut a cylindrical shape which will be, depending on the following transformation
technique, a tube or a solid bar. If for any reason the cell chosen doesn’t allow to have unconnected
external struts, we developed a software tool which cuts the numerical domain while keeping the
struts connected (Figure 4). To do this, solid lattices were made and cut in desired form (Figure 4,
Figure 4 Numerically solving the problem of unconnected struts in a rotated cube configuration
adapted to fit a cylindrical volume (Ehsan Rezaei’s PhD thesis - ongoin).
If the problem is to fit a given volume with cells allowing cells’ distortion, another MATLAB script
can solve it (Figure 5). This code receives a meshed structure as the input and converts the structure
into any of tetrakaidecahedron, cubic or octet lattice. This code basically can be further developed
for any cell type, which can fit into a hexahedron.
Once the numerical model has been designed, it can be utilized to simulate its behaviour in different
applications. Mechanical, thermal, fluid dynamic, electromagnetic simulations have been used to
optimize the performance of a porous component 17, 20, 22, 23, 25, 30-32.
33, 34
There are several methods to build a porous ceramic architecture by AM . Two different
techniques were utilized in this work to transform a CAD file into a ceramic periodic cellular
architecture . They can be both considered as two step manufacturing methods, in the sense that 3D
printing was utilized to produce green bodies which were subsequently thermally treated to
transform them into ceramics. We are studying these two techniques because they are
complimentary in terms of printing volume capability, resolution and ceramic powder choice.
3.2. Stereolithography
The slurry receipt comprises photo polymeric resin TPGDA (Allnex, Luxenburg, Luxenburg)
mixed with ceramic powders (av. Diameter 2 m) 42 vol% (Nabaltec, Schwandorf, Germany) and
UV Photoinitiator: Irgacure 819 1.2 wt% (BASF, Ludwigshafen, Germany). Slurry was thoroughly
36
mixed by ball mixing for 24h with zirconia balls D=8.5mm, 30 vol%. . The green cellular
architectures were then built layer by layer (Figure 6 right) projecting with UV light a sequence of
images with a dark background obtained by slicing the CAD file of the object. Printing was
performed with a 3DL Printer – HD 2.0 (Robot Factory, Mirano, Italy). Parameters were set as
following: slice thickness: 0.05 mm, slice exposure time: 1.1 s. This technique is well suited to
produce porous bodies because, when slicing them, the projected area to be cured at each shot is
confined in small regions. A well cured, monomer-free slurry with highly packed and well
37-39
dispersed powders will result in a more dense layer-less bulk ceramic after sintering . 3D
printing large bulk ceramics with low surface-area-to-volume ratio (e.g. a sphere) is indeed the most
4. Applications
The following paragraphs shortly present some application of periodic cellular architectures for
high temperature, harsh working conditions. They are the outcome of research projects whose
results can be disclosed. Other applications, developed under an agreement of non-disclosure, will
be omitted.
In industrial application with SiSiC ceramic foams the power is mostly limited to 1 MW/m2 , with
SiSiC cellular ceramics made of periodic architectures output power can be increased. Furthermore
the major improvement of this solution over the random foams is during the transient regime of
switching ON and OFF the burner. Rapid heating and cooling will lead to lower local failures
because of thermal shock. As a consequence, the problem of local stress concentration because of
the weaker struts into a foam (Figure 1) is clearly reduced.
4.2. TPS
Innovative thermal protection solutions (TPS) for space vehicles entering planets’ atmospheres
43
were developed within the FP7 European research project THOR (www. http://thor-space.eu/) .
44
One of them was exploiting a sandwich structured ceramic matrix composite with its core
working as active cooling system (Figure 8). A prototypic leading edge for plasma wind tunnel
tests was designed, its core, made of ad hoc cellular ceramics, was realized by AM and joined to
CMC skins 45.
Following an extensive test campaign active cooling TPS proved to be an effective way to lower
outside temperatures of re-entry vehicles’ hottest parts31.
The BioRobur project focused on green hydrogen production by direct biogas auto thermal
reforming (ATR). The novelty is based on the use of period architectures as catalyst support for the
ATR reaction. ATR was favored by the ability of SiSiC periodic cellular materials to disperse the
heat axially in the reactor 46, 47.
Figure 10 Top: the solution adopted enhance heat exchange with engineered porous architectures.
Bottom: Camera view of the inside of the tube during preliminary tests in a high temperature
6. Acknowledgements
This work is the result of the effort of several people who were hard working on the topic in the last
years. We would like to thank them one by one:
Simone Pusterla , Claudio D’Angelo and Ehsan Rezaei for CAD models generation and thermo
mechanical simulations, prof. Maurizio Barbato and Luca Ferrari for thermo fluid dynamic
simulations, Oscar Santoliquido for the Al2O3 lattices manufacturing.