J. Am. Ceram. Soc.
, 98 [7] 1983–2001 (2015)
Journal
Additive Manufacturing of Ceramics: Issues, Potentialities, and Opportunities
Andrea Zocca,‡ Paolo Colombo,‡,§ Cynthia M. Gomes,¶ and Jens G€
‡
Dipartimento di Ingegneria Industriale, University of Padova, via Marzolo 9, 35131 Padova, Italy
§
Department of Materials Science and Engineering, The Pennsylvania State University, University Park, Pennsylvania 16801
¶
Division of Ceramic Processing and Biomaterials, BAM Federal Institute for Materials Research and Testing, Unter den Eichen
Additive manufacturing (AM) is a technology which has the
potential not only to change the way of conventional industrial
manufacturing processes, adding material instead of subtract-
ing, but also to create entirely new production and business
strategies. Since about three decades, AM technologies have
been used to fabricate prototypes or models mostly from poly-
meric or metallic materials. Recently, products have been intro-
duced into the market that cannot be produced in another way
than additively. Ceramic materials are, however, not easy to
process by AM technologies, as their processing requirements
(in terms of feedstock and/or sintering) are very challenging.
On the other hand, it can be expected that AM technologies,
once successful, will have an extraordinary impact on the
industrial production of ceramic components and, moreover,
will open for ceramics new uses and new markets.
I. Introduction
A DDITIVE manufacturing (AM), also designated as solid
freeform fabrication (SFF), Rapid Prototyping (RP), or
3D printing, describes a class of technologies in which a part
is directly generated from a virtual model by adding material
to form the part. AM is defined by ASTM F2792-12a (Stan-
dard Terminology for Additive Manufacturing Technologies)
as the “process of joining materials to make objects from 3D
model data, usually layer upon layer, as opposed to subtrac-
tive manufacturing methodologies, such as traditional
machining”.1 According to the philosophy of AM, the choice
of a 3D computer graphic, 3D model, and material system
are sufficient to build a part. Consequently, it is possible to
generate parts with arbitrary geometries without the need of
adapting the typical manufacture process itself. Preferably,
parts ready to use, that is, possessing the final physical
properties of the desired object, are generated using AM
D. J. Green—contributing editor
Manuscript No. 36490. Received March 1, 2015; approved May 10, 2015.
†
Author to whom correspondence should be addressed. e-mail: jens.guenster@bam.
de
DOI: 10.1111/jace.13700
© 2015 The American Ceramic Society
unster¶,†
44-46, 12203 Berlin, Germany
technologies, but also the manufacture of semifinished parts,
requiring additional processing, has turned out to be techni-
cally and economically feasible. The material (feedstock) is
typically fed into the process as a powder/granulate, paste or
suspension, that is, the material is in a state optimized for
the layer deposition process. In the manufacturing process
itself, the material is used to build up the desired object and
it is simultaneously transferred into a state possessing its final
physical properties, or at least a mechanical strength suffi-
cient to transfer the built object to further processing steps.
According to the needs of the particular material used, the
desired geometry or the purpose of the object being built, an
entire set of AM technologies has been developed.2–5 Poly-
meric, metallic, and ceramic materials can be all processed
by these technologies.6–11 The initial motivation for the
development of RP technologies, that is, technologies for the
manufacturing of individual products without any require-
ment for dedicated tooling, was to reduce the time to market
of new products by shortening the time span between design,
testing and implementation. While in the ’80s and ’90s, the
basic technologies for AM were developed and flexibility in
design was the first priority, nowadays the physical proper-
ties of the parts generated are a major concern. Accordingly,
the terminology for this class of technologies shifted gradu-
ally from RP to AM. With minimum lead time and design
modifications by the click of a mouse, AM stimulated an
entire industry. AM technologies based on the processing of
polymeric and metallic materials have been, in this context,
the most successful. Even multimaterial textures and func-
tionally and compositionally graded structures can be syn-
chronously fabricated during the manufacturing process.12–14
To develop a more general understanding of the technical
challenges associated with AM of ceramics, it makes sense to
divide the following discussion into two main categories,
defined by the macroscopic physical properties and architec-
ture of the ceramics parts generated via AM, that is: (i) por-
ous components and (ii) fully dense monolithic components.
Accordingly, we will distinguish between three classes of AM
technologies:
(1) technologies not capable of producing dense structures
(which means, a microstructure without residual porosity),
(2) technologies capable of producing fully dense monolithic
Feature
1984 Journal of the American Ceramic Society—Zocca et al.
ceramic bodies with restriction in dimension, that is, dense
filaments or small structures including solid ceramic parts
not larger than a few mm in diameter, and (3) technologies
capable of producing fully dense monolithic ceramic bodies
without restriction in dimension.
It is well-known that the many AM technologies success-
fully applied to the manufacture of ceramics have been used
to fabricate porous structures, for example, scaffolds for bio-
logical applications, filters, lightweight structures etc. In these
cases, the success of the respective technology is not only
based on its ability to exactly generate the desired porous
structure, but on the fact that a certain amount of random
and residual porosity in the solid ceramic part of the scaffold
is tolerated or even desired, as it provides further functional-
ity to the component.
The fact that, so far, a major part of the research on AM
of ceramics has focused on porous structures, is due mainly
to two reasons:
1. AM enables the production of complex-shaped porous
architectures with a precise control of the dimension,
shape and amount of pores, which is not achievable by
any other technology currently used for the fabrication
of macroporous bodies. In these cases, AM is not in
competition with traditional manufacturing, but it
opens up completely new fields of research and appli-
cations.
2. Many of the AM technologies existing nowadays are
intrinsically particularly suited for the generation of
fine filigree structures. The specific reason for this
statement will be discussed later, when describing the
individual technologies, because it is different from
case to case. However, a general observation that can
be made here is that technologies that are suitable for
fine porous structures have shortcomings when scaling
up to large bulk parts, and, vice-versa, technologies
more suitable for the fabrication of dense parts may
have difficulties in producing highly porous and fine
geometries.
There are very few AM technologies on the market capa-
ble of generating fully dense monolithic ceramic bodies. AM
of monolithic ceramics, enabling the components to fully
retain their superior physico-chemical properties, is still a
challenge and remains the most important task that needs to
be solved to promote AM of ceramics to more than a niche
technology. AM of ceramics has the potential to open new
markets for ceramic products, as it is not strictly bound by
the paradigms of ceramic processing, such as high processing
cost for small lot productions, low flexibility, complex
machining, and limited molding shaping technologies. The
current situation, however, is not much different from what
Cawley stated about 15 yrs ago: “The need for ceramic pro-
cessing expertise to prepare feedstock and the fact that SFF
methods only yield partially processed parts means that their
introduction does not open this class of materials to new users.
(This is quite unlike the situation with their application in the
field of polymers. . .)”. Moreover, many of the processing
issues involved in AM of ceramics are the same as those that
characterize ceramic processing in general, and the field of
users of AM technologies for ceramics is practically limited
to those already engaged in ceramic production or
research.15
For the sake of completeness, it should be noted that the
preceramic polymer approach to obtain polymer-derived
ceramics (PDCs) offers a viable alternative to the powder
processing route, not only in terms of classical fabrication
technologies but also in the context of AM. The striking
advantage of PDCs arises from their relatively simple pro-
cessing starting from polymeric precursors, which enables the
components to be fabricated using technologies suited for
polymers. However, there are several technological challenges
associated with the pyrolysis of the polymeric structure into
Vol. 98, No. 7
a ceramic. The most critical, in this context, is the release of
gaseous species during the pyrolysis of the preceramic poly-
mer, with its associated shrinkage, which prevents the direct
production of bulk components. Strategies exist to reduce
these issues, such as adding inert or reactive fillers, but the
production of fully dense parts remains a challenge. More-
over, most preceramic polymers would lead only to silicon-
based ceramics, with boron-based systems being a notable
exception.16 Nonetheless, for AM of filigree, porous and
small structures, the PDCs are an attractive alternative to
conventional ceramic feedstocks.
This study aims to discuss the current status and perfor-
mance of AM of ceramics. An overview of technologies for
the fabrication of porous components will be given, and par-
ticular applications will be highlighted. Moreover, the tech-
nologies capable of providing dense monolithic ceramics will
be carefully analyzed, and general strategies for the produc-
tion of ceramic parts via AM will be formulated. Finally, a
comparison of the state of the art of AM of ceramics and
metals will help to predict what we can expect for AM of
ceramics in the future.
II. Overview of AM Processing of Ceramics
This paragraph gives an overview and a comparison of the
AM technologies currently available for ceramic materials,
see also Fig. 1.
For a detailed description of the various AM technologies,
we refer the readers to some recently published
reviews.4,5,17,18 The following description aims, rather, at
comparing the main characteristics of the AM techniques
with respect to both the properties of generated parts and
processing-related aspects, to provide a tool for roughly eval-
uating which AM technology may be more indicated for a
specific product and process chain.
Table I summarizes some of the general characteristics of
the most used AM technologies described in literature for
manufacturing ceramic components. Indications about the
quality and size of the produced parts (ability of making
dense and/or monolithic parts, precision, surface quality) and
cost of the production (cost of the feedstock preparation and
cost of the process) are included.
Before commenting on the individual technologies, it is
appropriate to qualify further the content of the columns in
Table I.
1. Dense struts: this term indicates technologies that can
provide a dense material (that is, without any residual
porosity), at least for some of the possible geometries
and features that they can produce (specifically, only
features possessing a small length scale or volume) and
some ceramic material systems. Importantly, it is rele-
vant to notice that there is a correlation between the
use of a liquid-based feedstock and the possibility of
achieving a dense microstructure in the struts (or small
printed features), as opposite to solid-based and, in
particular, powder-based technologies.
2. Monolithic: this term indicates technologies that are
able to generate parts without residual microporosity,
at least for some ceramic material systems, regardless
of its volume and thickness. A part with these charac-
teristics is what we define as a “ceramic monolithic
part”. Naturally, the possibility to sinter to full density
is related to the composition of the feedstock (i.e.,
presence of sintering aids, composition of the ceramic)
and on the sintering schedule and technique employed.
Here, however, we consider as given that the ceramic
powder could be sintered to full density (for instance if
it was shaped into a tablet by cold pressing), and are
interested in discussing whether it is the AM process
itself that prevents this full sintering to happen in the
final produced part.
July 2015 Additive Manufacturing of Ceramics
Fig. 1. Schematic of AM technologies.
3. Part dimension: this is the maximum part size that
can be produced. Based on the fact that the dimen-
sion of the process area (printing envelope) can be
potentially scaled-up for most technologies, it is diffi-
cult to provide precise data and trends. However, at
the current state of the art, it is often the case that,
for dimensions outside a certain range, it becomes
more convenient, for economic and technical reasons,
to switch from one technology to another. These
dimensional ranges are indicated as sizes from XS to
XL, corresponding to 100 lm or to 1 m, respectively,
as explained in the table caption.
4. We should also note that it is not only the dimension
of the printing envelope for each AM technology that
1985
imposes a limit to the maximum part size (area or
volume), but also the mechanical properties of the
parts during the printing process. For instance, struc-
tures constituted by thin filaments and not supported
by a liquid or powder bed can deform by their own
weight during fabrication, when they’re still not con-
solidated by sintering.
5. Precision and surface quality are partially related, and
are also correlated with the dimension and cost of the
part (printing speed) to be produced. On the other
hand, the surface of a part can become smoother due
to the minimization of its free surface energy, if suffi-
cient material transport can occur during fabrication
or sintering, or artifacts such as steps deriving from
deposited layers can be smoothed out by the presence
of a liquid phase during sintering. Here, only the best
state-of-the-art results achieved so far are listed. As the
dimension of the part increases, the precision in many
technologies can be reduced (e.g., by increasing the
diameter of extrusion nozzles or the thickness of
deposited powder layers) to obtain an economical
building rate. It is noteworthy to observe that surface
quality is important not only for aesthetical reasons
but also because it can influence the mechanical prop-
erties of the printed ceramic components. It is further-
more noteworthy to observe that filament-based
technologies can provide very smooth surfaces on a
microscale level, but by fusing multiple filaments a tex-
tured surface is generated, which can significantly devi-
ate from the surface contour of the 3D model.
6. The cost parameter is divided into two factors: cost
related to the feedstock and cost related to the pro-
cess. Cost for feedstock does not include the cost for
ceramic raw materials, as this would be similar for a
conventional fabrication process. A low feedstock
cost means that the material can be bought commer-
cially and is usable with none or minimal further pro-
cessing, irrespective of the cost of the specific ceramic
raw material selected; this is typically the case for
powders to be used in powder-based and slurry-based
technologies. Medium and high cost indicates that
the material requires extensive secondary processing,
and often has to be designed specifically for the appli-
cation. An example is the ink designed for extrusion
through very fine nozzles. Finally, the process cost
considers depreciation of capital expenditure and
operating expenses for the particular AM equipment,
and costs related to necessary processing and post-
processing of the final product. Machines which
employ lasers, for example, are usually quite expen-
sive and are not faster than equipment based on
rather cheap inkjet printing heads. On the other
hand, the extrusion of pastes through fine tips is typi-
cally based on a rather cheap setup, and the produc-
tion rate can be adapted by changing the diameter of
the tip, which makes this a low-cost process.
7. We can also point out that AM technologies can be
divided into two categories: direct and indirect. The
terms “direct”, “indirect” and “negative” have not
been used with consistent meaning in the history of
AM, as already pointed out by Tay et al. in a review
in 2003.19
In the following, we will use these terms with this spe-
cific meaning:
- Direct means that the material is directly depos-
ited only in the position giving the desired shape
of the final object.
- Indirect means that, first a layer of material is
deposited, subsequently the cross section (slice)
of the part is inscribed in the layer and, after
completing all layers to complete the process,
excess material surrounding the part is removed
1986 Journal of the American Ceramic Society—Zocca et al. Vol. 98, No. 7

Table I. Main Characteristics of Additive Manufacturing Technologies used for Ceramics

Part
dimension‡
Dense struts (size that Surface
(dense ceramic, can be (quality of parts, Cost of Direct
Feedstock but limited produced not of feedstock Cost of versus
†
Technology (liquid or solid) volume) Monolithic economically) single struts) Precision preparation process indirect

P-3DP Solid No No M-XL Medium 100 lm Low Medium Indirect

P-SLS Solid No No M-L Medium 100 lm Low High Indirect
P-SLM Solid No No M-L Medium 100 lm Low High Indirect
S-3DP Liquid Yes Yes M-XL High 100 lm Low Medium Indirect
S-SLS Liquid Yes Yes M-L High 100 lm Low High Indirect
SL Liquid Yes No XS-M High <1 lm Medium-High Medium Indirect
LOM Solid Yes Yes M-L Medium 100 lm Medium Medium Indirect
DIW/ Liquid Yes No S-XL Low 10 lm Low-Medium Low Direct
Robocasting
FDM Liquid Yes No S-M Low 100 lm Medium Low Direct
DIP Liquid Yes Yes S-M Medium 10 lm High Medium Direct
Direct Dep Liquid Low Direct
†
The definition of the acronyms for each technology is given later in the text.
‡
XS = 100 lm; S = 1 mm; M = 10 mm; L = 0.1 m; XL = 1 m.

to release the final object. Inscribing can be car-
ried out in several ways, each one with its own
characteristics.
- Negative means that an AM technology is used
to produce a mold, which is successively used to
produce the final part.
Already from this distinction, some general considerations
concerning the potentials and drawbacks of the two
approaches can be outlined:
1. The excess material in indirect technologies forms a sup-
port to the successive layers, enabling the generation of
parts with large overhangs. Direct techniques, on the other
hand, require support structures for overhangs above a
certain dimension, which may be often impractical or
time-consuming, and furthermore need to be removed to
obtain the precise desired shape of the final object.
2. Indirect technologies require the removal of the excess
material in the form of powder, liquid or solid pieces of
a sheet, which can be difficult or impossible for certain
geometries. For example, if the excess material is a
powder, it may be complicated to remove it completely
from small open pores and cells. In any case, it is not
possible to print any form of closed porosity, because
the material inside the pores that has not been inscribed
cannot be removed. The direct deposition of material,
instead, does not have such geometrical limitations.
3. Indirect technologies have the potential to be faster,
because the material is deposited in one complete layer
and the layer information, that is the cross section of
the part to be built, can be printed on the entire area,
selectively inscribing the desired profile. Moreover, the
simultaneous fabrication of multiple individual parts
can be achieved. In direct technologies, instead, the
material is usually deposited only in specific positions
and in a sequence.
4. Direct technologies offer straightforward possibilities
to use multimaterial systems (ceramic/ceramic with
functional grading, ceramic/plastic, ceramic/metal
interfaces,). Indirect technologies can also do this, but
with much more limitations, for example, the addition
of metallic materials acting as conductor in multilayer
ceramic systems is possible, but their concentration is
limited.20
Table II summarizes some of the relevant characteristics
of feedstocks used for AM of ceramics, which are categorized
as powder, paste or suspension.
III. AM Technologies for Ceramics
This section reviews the current state of the art of AM of
ceramics, describing the main technologies developed so far.
Specifically, the paragraph relative to each technology
includes a description of: (i) the fabrication of structures with
a designed macroporosity and considerations about the resid-
ual microporosity in the parts, which might be resulting from
the respective fabrication technology; (ii) the possibility of
producing monolithic ceramics, according to our previous
definition.
(1) Indirect AM Technologies
(A) Powder-Based 3D Printing (P-3DP): In powder-
based AM technologies, a solid structure is realized by the
successive deposition of layers of a flowable powder. A layer
of powdered material is first spread and subsequently the
corresponding layer information is selectively inscribed by,
for example, local compaction or gluing; these steps are itera-
tively replicated until the object is completed. The layer
information is defined by the corresponding cross section of
a sliced virtual 3D model of the object to be built. At the
end, the part is completely embedded in a powder bed, from
which it can be easily extracted and cleaned. The technology
used to inscribe the layer information depends on the specific
process considered: two of the most well-known and world-
wide spread processes are: “Three-dimensional printing”
(3DP) and “Selective laser sintering” (SLS).21 The milestone
patent “Three-dimensional printing techniques”22 by Sachs
et al. was filed in 1989, whereas Deckard’s patent “Method
and apparatus for producing parts by selective sintering”
dates back to 3 yrs earlier.23 Within powder-based AM tech-
nologies, powders are bound or fused to more or less dense
parts, often ready to use. The former method uses a printing
head to selectively spread out droplets of a liquid binder,
whereas the latter employs the energy of a focused laser
beam to selectively sinter/melt a powder.
In the P-3DP process, an inkjet printing head spits a bind-
ing liquid onto a powder bed, thus defining the cross section
of the object in that layer. Some P-3DP setups stabilize the
powder layer by misting with water droplets to avoid distur-
bance of the powder by the ballistic impact of the binder
droplets.24 Multiple nozzles can also be used, allowing to
cover a wide printing area. Printing machines with print
heads consisting of a few hundred of parallel working jets
and with a printing envelope as large as 6 m 9 6 m 9 6 m
are used to fabricate ceramic objects for architectural
July 2015 Additive Manufacturing of Ceramics 1987

Table II. Feedstocks for Additive Manufacturing of Ceramics

Notes about
Type of feedstock Habitus Rheological and physical properties processability Technology Notes about composition

Powder Particle Solid Poor flowability P-3DP Several possible

size > 20 lm P-SLS binder systems,
>20 lm Good flowability P-SLM organic or
80 lm> inorganic based.
<80 lm Excellent
flowability;
high powder
packing
density
Paste Filament Viscosity:107, 110, 146 Shear thinning DIW Ceramic solid
diameter 10–100 Pas @ 100 s
1 Robocasting loading 35–55
50 lm< G0 (eq) = 105–106 Pa vol%. Use of
<1000 lm Yield stress 102–103 Pa flocculated
suspensions or
of polymeric
binders
100 lm< FDM Typical ratio
<1000 lm ceramic/polymer
= 60/40 vol
Suspension Layer thickness Viscosity:11, 64 Shear thinning, SL Typical composition
10 lm< 100 mPas–110 Pas @ 100 s
1 suitable viscosity includes: monomer,
<50 lm highly depends photoinitiator,
on the recoating dispersants
system Ceramic solid loading
40–60 vol%
10 lm< Viscosity:97, 98, 147, 148 Shear thinning DIP Solids loading
<50 lm 5–15 mPas @ 1000 s
1 2–30 vol%
Surface energy: 20–70 mN/m
50 lm< Viscosity:1 mPas–1 Pas Shear thinning S-3DP Water basedCeramic
<200 lm S-SLS solid loading 30–50
vol%

applications based on the cold consolidation of inorganic
powders (such as marble) using a hydraulic or nonhydraulic
binder.25 In Fig. 2, we show, as an example, a large artwork
piece which is 1.28 m wide, but could be printed up to a
diameter of 3.5 m using such an equipment.
The polymer–polymer-derived approach has been applied
to P-3DP, and in this context it offers several advantages.
When using pure preceramic polymer powders, it is possible
to achieve a high polymeric density (~80% TD) because the
sprayed solvent locally dissolves the preceramic polymer
binding the particles together, filling also the interparticle
voids.26 The resulting ceramic part after heat treatment still
had a similar level of residual porosity, as the material does
not sinter during the process. Moreover, the preceramic poly-
mer can be used as a nonsacrificial, reactive binder mixed
with a glass powder to create bioceramic scaffolds based on
wollastonite-apatite.27
(a) Porous Components: The vast majority of the
research on powder-based 3D printed porous parts has been
devoted to the shaping of biomedical scaffolds for tissue
engineering (TE). Two major classes of material-binder sys-
tems have been used for this purpose: (1) the ceramic powder
is bound by an organic glue, which is dissolved in a water-
or solvent-based liquid and sprayed on the printing bed28;
alternatively, a polymer can be mixed with the ceramic pow-
der and then used to glue the powder when in contact with
the jetted liquid.29,30 The green part has to be successively
sintered at high temperatures to obtain the final ceramic
object. (2) a reactive liquid-powder system is employed,
which involves a setting reaction at low temperature. There
are several examples of systems based on calcium phosphates
using phosphoric acids as a binder.31–33
A review of the important scaffold parameters for this spe-
cific application, mostly focused on the P-3DP process, is

(a) (b)

Fig. 2. (a) powder-based 3D printed “The Flower” artwork model (the part is 1.28 m in diameter); (b) image of a layer during the 3D printing
process (model designed by Yassi Mazandi; image courtesy of Antonino Italiano, Desamanera SRL, Italy).
1988 Journal of the American Ceramic Society—Zocca et al.
already present in the literature.34 A more general review
describes benefits and drawbacks of various AM technologies
in scaffold fabrication.35 Typically, TE scaffolds have pores
in the range 50–1000 lm and a porosity >60%, because these
characteristics are considered a requisite for successful bone
ingrowth and full vascularization of the implants. Residual
microporosity (< ~10 lm) in the printed part (i.e., not the
designed macroporosity) was also recognized to be important
for increasing the surface area, which leads to higher protein
absorption and ion exchange. It should be noted that small
pores (< ~500 lm) cannot be printed because of resolution
limits and, more importantly, because of the difficulty in
removing the excess powder.
Calcium phosphate scaffolds made by P-3DP tend to have
exactly this bimodal porosity: a designed porosity produced
by P-3DP, in the range ~0.5–2 mm, and microporosity in the
struts due to incomplete densification after binder/powder
reaction and/or after sintering. In particular, sintering of the
3D printed material to high density is hindered by the large
particle size (indicatively 30–100 lm) of the powder used for
printing, which is necessary to have a suitable level of flow-
ability during the deposition of the layers. Even though the
flowability can be improved by granulation of a fine
(<10 lm) powder, many authors recognized that during the
heat treatment, a good level of densification can be achieved
within the single granules, but only sintering necks are
formed between the granules.30,36 An alternative approach to
improve the flowability of fine powders is the surface treat-
ment of the particles, for example, by plasma-enhanced
chemical vapor deposition, as discussed in reference34 and
the literature cited therein, especially.37 However, the intro-
duction of fine silica on the surface of the particles resulting
from the plasma process may be considered as a contamina-
tion, especially in the field of bioceramics where a high purity
is a requisite.
Parts made by P-3DP generally have a high level of resid-
ual porosity and this limits their mechanical properties.
Chumnanklang et al. made HA samples by P-3DP which
had 20–40 vol% microporosity and a 3-point-bending
strength <1.5 MPa.30 Klammert et al. developed a P-3DP
process for Brushite (CaHPO42H2O) which had a micropo-
rosity of 38.8 vol% and a compressive strength of 23.4 MPa,
and similarly for Monetite (CaHPO4), obtaining 43.8 vol%
of microporosity and 15.3 MPa compressive strength.33 Cas-
tilho et al. measured the compressive strength of calcium
phosphate cylinders, which reached a maximum of 23.8 MPa
for a microporosity of 43.1 vol%.29
In addition to applications for TE, P-3DP of silicate
ceramics for different applications, including art and architec-
ture, and foundry sands are rapidly growing markets.38, 39
(b) Monolithic Components: P-3DP is very well suited
to generate porous ceramic structures. Still, a major issue,
hampering a completely autonomous production of parts
and even restricting the freedom of design by means of these
technologies, is the low density and stability of the parts
during the building process, which is a direct result of the
low powder packing in the powder bed [as low as 25% TD
(theoretical density)]. The powder bed surrounding the part
has an essential role, since it should support the structure
during building, until it’s ready for removal. Furthermore,
powders with appreciable flowability for the deposition of
thin layers are in a particle size range >20 lm which gener-
ally does not provide sufficient sintering activity to form
dense ceramics. One straightforward approach to increase
the density of the powder bed and to use more fine powders
is the deposition of ceramic slurries instead of dry powders.
This approach, which will be introduced in a following sec-
tion, results in powder beds with powder packing densities
>55% TD, and thus sufficient for a full densification in sub-
sequent sintering steps. Even though, as discussed above, the
microstructure of parts obtained by P-3DP usually contains
a high amount of residual porosity, some authors achieved
Vol. 98, No. 7
higher densities after the heat-treatment process. Three dif-
ferent approaches have been mainly investigated with this
aim.
The first approach consists in modifying the composition of
the ceramic material to enhance the sintering, by addition of
dopants or of a viscous liquid-forming phase. Fielding et al., for
example, obtained 95% density when adding ZnO and SiO2 to
tricalcium phosphate (TCP),28 whereas Suwanprateeb et al.
took advantage of liquid phase sintering producing a 97.5%
dense HA + HA/wollastonite glass-ceramic composite.40
The second approach is the infiltration of a porous pre-
form produced by P-3DP. Melcher et al. used P-3DP to print
alumina preforms which had a porosity of 36% after sinter-
ing, and infiltrated them with Cu/Cu2O.41 The resulting
Al2O3/Cu/CuO material had a flexural strength of 236 MPa
and fracture toughness of 5.5 MPam1/2 thanks to the crack
bridging mechanism of the metallic phase. A similar process
was followed to infiltrate alumina preforms with glass.42
Another interesting material system achievable with this
approach is the formation of reaction-bonded silicon carbide
(RBSiC) ceramics. In a first example, Fu et al. printed by P-
3DP a mixture of Si, SiC and dextrin.43 Successive infiltra-
tion with a liquid silicone resin and pyrolysis at 1000°C in
nitrogen yielded carbon from the dextrin and an amorphous
SiOC residue from the silicone resin. Infiltration of liquid sili-
con at 1500°C in vacuum finally led to the RBSiC material.
Moon et al., instead, printed by P-3DP a carbon preform
which had 48% open porosity, by using a binder based on
furfuryl resin.44 Following infiltration of Si at 1450°C in
nitrogen led to the desired RBSiC composite. It is interesting
to notice that in this latter example part of the carbon was
derived from the polymeric binder, and this allowed the gen-
eration of functionally graded parts by varying the dosage of
carbon-yielding binder and thus controlling the amount of
SiC formed.
The first two approaches may be however difficult to apply
to pure technical ceramics, such as alumina or silicon nitride.
A third approach has been developed which consists of cold
isostatic pressing or warm isostatic pressing (WIP) the green
body printed by P-3DP, before sintering. Yoo et al. obtained
almost dense parts (99.2% relative density) by applying WIP
to alumina parts (with MgO dopant) which had a flexural
strength of 324 MPa after sintering.45 This method has been
applied also to other ceramic systems, such as Ti3SiC2, yield-
ing almost dense microstructures, but it poses geometrical
limitations when applied to complex geometries, such as
parts with internal cavities.46
Concluding, P-3DP is very attractive because it has almost
no limitation regarding the parts geometry (except for closed
porosity) and dimension; however, it is potentially able to
produce ceramic monolithic parts only for specific material
systems, typically involving postprinting infiltration with met-
als of semimetals.
(B) Powder-Based Selective Laser sintering (P-
SLS): The P-SLS process works similarly to the P-3DP
one, with the difference that the cross sections of the part to
be built are inscribed by means of a laser beam. The local
densification of the powders can be obtained by directly sin-
tering the ceramic powder, or by mixing the powder with a
binder phase (inorganic or, more often, polymeric) which is
then melted by the laser light. Direct laser sintering of ceram-
ics is complicated by the poor resistance of this class of
materials to thermal shock. Furthermore, the short interac-
tion time between laser and powder limits material diffusion,
leading to poor sintering.17, 47–49 Binders within the powder
which can be thermally activated, on the other hand, have
been successfully used to shape porous ceramic parts, for
applications and with results similar to those produced by P-
3DP.50, 51 In the context of also this AM technology, prece-
ramic polymers can offer some interesting opportunities, as
highlighted by the work of Friedel et al., which used a laser
light to selectively melt and crosslink a solid silicone resin
July 2015 Additive Manufacturing of Ceramics
mixed with SiC powders, to generate complex structures later
sintered and infiltrated with a Si melt.52
(a) Porous Components: Kolan et al. produced bio-
glass scaffolds with 50% porosity and pores in the range
300–800 lm by using an acrylic binder, followed by sintering
at 675°C–695°C.53 Liu et al. deposited layers of a slurry
made of hydroxyapatite powder and silica sol, exploiting the
gelling of silica as binding mechanism, followed by sintering
at 1200°C.54 The produced scaffolds had pores in the range
750–1050 lm and a porosity of 25%–32%.
Another approach in the production of scaffolds for TE
applications is the use of a bio-compatible polymer as binder.
Simpson et al. reported the shaping of 95/5 Poly(L-lactide-
co-glycolide) and HA/TCP composites,55 whereas Tan et al.
employed a PEEK/HA composite.56 It is worth noting that,
in this case, the parts produced are a biopolymer-ceramic
composite and they do not necessitate a sintering posttreat-
ment, hence they are potentially ready to use as they come
out of the SLS machine.
(b) Monolithic Components: Similarly to P-3DP, there
are some rare examples in which dense monolithic ceramics
have been obtained by P-SLS by infiltration with metals of
semimetals after printing a porous ceramic preform. One of
the early results concerns the generation of Si/SiC parts.
Within the SLS process, SiC particles are glued together as a
result of the formation of SiO2 in oxidizing atmospheres, that
is, ambient air. The highly porous parts have been used as
preforms for the silicon infiltration.57 More recent work con-
centrates on selective laser melting (SLM), where the forma-
tion of a liquid phase ensures rapid densification.58 The use
of dense alumina and zirconia spherical particles (30 lm) as
feedstock results in a high packing density of the particles in
the powder bed. Preheating of the powder bed to a tempera-
ture of at least 1600°C reduces the thermal shock within the
part to be built. Alumina–Zirconia eutectic mixtures reduce
the melting temperature of the powder bed to about 1860°C.
With this approach, dense ceramic parts with remarkable
mechanical properties could be generated via SLM. Nonethe-
less, crack formation within the ceramic bodies could not be
avoided completely and an uncontrolled infiltration of the
preheated powder bed by the liquid phase formed within the
laser spot results in an ill-defined outer contour of the parts
generated. Even though there is currently no strategy in sight
which would allow the formation of dense monolithic cera-
mic parts via SLM, with precise contour and properties simi-
lar to conventional manufactured ceramic parts, the laser-
based technologies SLS and SLM are the only processes
which could potentially deliver ready to use ceramic parts
from ceramic materials that require sintering.
(C) Stereolithography (SL): The stereolithography
technique is based on the photopolymerization of a liquid
resin filled with ceramic particles, in a layer-by-layer fashion
similarly to the other indirect AM technologies.
Typical compositions for the slurry include a monomer
solution, a photoinitiator and additives for the dispersion
of the ceramic powder, with a concentration of 40–60
vol%.11, 59–61
Contrary to powder-based AM, in stereolithography the
material surrounding the part is liquid and therefore it is
unable to provide any support; for this reason, depending on
the specific geometry of the part, support structures are built
together with the part, requiring subsequent removal.59 Due
to the high green densities and the use of fine ceramic parti-
cles, stereolithography permits the production of almost
dense ceramic parts after a sintering posttreatment.11 There
are different technical solution for the SL process on the
market. Lithoz (Lithoz GmbH, Vienna, Austria) follows an
approach with the part built not fully immersed in the liquid
feedstock, but mounted upside down on the building plat-
form and dipped in a thin slurry layer deposited on a glass
plate. The striking advantage of this strategy is the small
amount of slurry required for building the part.60 Table III
1989
Table III. Density of an as-Printed Sample (cylindrical,
height ~11.8 mm, diameter ~10.0 mm) and of a Printed
Sample After Debinding, Made from Alumina Powder (purity
level 99.9%) by the Lithoz SL Process
As-printed green compact (g/cm3) Brown compact (g/cm3)
Bulk density† 2.561  0.020 2.211  0.006
True density‡ 2.568  0.002 3.971  0.003
†
Measured geometrically.
‡
Measured using a helium pycnometer.
provides densities of a cylindrical alumina part manufactured
by this process, measured before and after debinding. The
powder compact after debinding is sometimes denoted as
“brown” compact. According to Table III, the powder pack-
ing density in the compact after debinding is 55% TD. With
this material system, it is possible to achieve densities exceed-
ing 99% TD after sintering.
(a) Porous Components: The utilization of this tech-
nique was investigated by several authors for the generation
of porous ceramics for various applications, taking advan-
tage of the high precision that this technology provides and
the possibility of obtaining dense struts. Kirihara et al. pro-
duced various types of samples possessing an ordered poros-
ity, all with a total porosity value of 75 vol%.62 YSZ lattices
had about 100-lm-long and -thick struts with a lattice con-
stant of 250 lm, and they were designed for anode electrodes
in solid oxide fuel cells. Al2O3 parts were created based on
the diamond structure, and their properties for an applica-
tion as photonic crystals were simulated and measured.
Finally, hydroxyapatite could be shaped into a dendrite
structure with graded porosity, for TE applications. In all
cases, the struts density ranged between 95% and 98%. Chu
et al. also produced hydroxyapatite porous implants with a
designed porosity of 40% and orthogonal pores.63 The
authors measured their mechanical properties and in vivo
behavior, reporting a high compressive strength of 30 MPa.
Chartier et al. made Al2O3 parts with a highly concentrated
ceramic suspension (60 vol%) and recoating layers as thin as
25 lm.64 1 and 2 mm-2D porous meshes were produced to
demonstrate the process resolution.
Bian et al. produced bioinspired osteochondral scaffolds
by first generating a porous b-TCP scaffold by SL followed
by debinding and sintering and successively simulating a car-
tilage phase by freeze casting and cross-linking of a type I
collagen suspension.65 In parallel to the traditional stereoli-
thography process, a novel nano-stereolithography technique
employs a two-photon equipment to achieve the cross-linking
of a resin with a resolution down to 200 nm. Pham et al.
modified a polyvinylsilazane preceramic polymer and used it
to shape complex micrometer-sized structures.66 The polymer
was subsequently pyrolyzed at 600°C in nitrogen to obtain a
SiCN ceramic. Silica nanoparticles were added to obtain a
reduced and isotropic shrinkage. In particular, the generation
of a 9 lm 9 9 lm 99 lm scaffold with approximately 1 lm
cells was reported.
(b) Monolithic Components: In general, it is possible
to produce dense monolithic ceramic parts via SL, because
powder compacts with powder packing densities higher than
55% can be generated. The only restriction is imposed by the
debinding step, which is required to remove the binding addi-
tives from the green compacts before sintering. The thermally
activated debinding steadily transforms the binder into gas-
eous species, and is one of the most critical steps in the SL
process due to the possible formation of defects in the brown
compact. As thermal debinding takes place in the volume of
the part and dissociation and combustion products must
leave the bulk by diffusion, the heating process has to be car-
ried out very slowly keeping the pressure inside the part
below a critical threshold avoiding the formation of cracks
1990 Journal of the American Ceramic Society—Zocca et al.
and voids. On the other hand, there already exists a vast
expertise on the thermally and solvent-based debinding of
ceramic green compacts, from injection molding of ceramics.
Injection molding also uses pastes with high organic content,
as binders, plasticizers, dispersants and lubricants, to produce
parts with high complexity.67 Thus, regarding the production
of dense monolithic parts, a clear restriction of the SL pro-
cess is, similarly to injection molding, the maximum volume
or wall thickness of the ceramic green body at which debind-
ing can be achieved in a reasonable time.68 At wall thickness-
es of about 1 cm, debinding already requires a thermal
treatment over a time period of several days.
(D) Slurry-Based 3DP (S-3DP) and Slurry-Based SLS
(S-SLS): S-3DP and S-SLS are technologies which are a
logic extension of powder-based technologies. Instead of
spreading a dry powder into thin layers, a ceramic slurry or
slip is used to increase the powder packing density in the
powder bed. Already in the milestone patent “Three-dimen-
sional printing techniques” by Sachs et al., filed in 1989, the
dispersion of powders “in a liquid vehicle” is suggested “to
provide the desired uniform depositions of such powders at
the required rates and densities”.22 In a later patent by Sachs
et al., filed in 1998, the use of liquid dispersions of powdered
materials for the built up of a powder bed in a layer by layer
fashion is explicitly described.69
A major concern at that time was, however, the deposition
of uniform layers with a flat and smooth surface. Since then,
technologies capable of repeatedly depositing thin layers of
ceramic slurries to build up dense powder beds for S-3DP
and S-SLS have been developed.70–72 A survey of recently
published work reveals a great potential of this technology
for the manufacture of dense powder compacts, comparable
to those produced by classical powder processing, that is
with low organic content.
In S-3DP and S-SLS a liquid suspension of ceramic parti-
cles is spread as a thin layer comparable to tape casting,
whereas within this layer a cast is formed comparable to slip
casting. In the slip casting process, the slip is brought into
contact with a porous body, the casting mold. When a liquid
is wetting a porous body, capillary forces draw the liquid
phase of the suspension into the porous mold. The particles
start to form a powder compact, the cast, at the molds sur-
face. The cast formation kinetics obeys a square-root-of-time
law by depositing individual particles from the suspension to
the surface of the mold. In case of the layerwise slurry depo-
sition (LSD) process the mold is formed by the previously
deposited and dried layers.73 Slip casting results in the for-
mation of powder compacts with densities potentially exceed-
ing 60% TD. Furthermore, due to the free settling of the
powder particles from the suspension, no obvious interface is
formed between individual layers. The formation of a cast
starts as soon as the slurry is in contact with previously
deposited and dry layers. Hence, the speed of the doctor
(a)
Fig. 3. Release of two espresso cups made of porcelain by employing the Layerwise Slurry Deposition (LSD) technology. (a) powder bed,
formed by 132 layers with a thickness of 200 lm, containing the two laser-sintered cups; (b) cups partially released from the powder bed.
Vol. 98, No. 7
blade relative to the dry layers must be high enough to pre-
vent collision of the cast with the doctor blade. On the other
hand, when the speed of the doctor blade is too high, layer
deposition is not uniform, because shear stresses within the
suspension will result in an inhomogeneous deposition. For
commercial silicate ceramic slurries, a speed of the doctor
blade of 50 mm/s is typically chosen. For technical ceramics,
the cast grows significantly faster and the speed of the doctor
blade must be increased to 100 mm/s or higher. In the P-
3DP and P-SLS powder-based processes, the part built is
easily released from the powder bed, but this is not the case
for the slurry-based process. After deposition of all layers
required for building up the desired geometry, the laser sin-
tered or printed body is embedded in the powder bed which
now is a block of densely packed particles. For the release of
the part, the powder compact must be dissolved by a solvent.
In case water-based slurries are used for the layer deposition,
water can act as a solvent for the powder bed. Figure 3
shows an example for the release of two espresso cups fabri-
cated by the S-SLS process.
For the manufacture of powder compacts with properties
comparable to those produced by classical powder process-
ing, S-3DP has the highest potential. By the use of small
amounts of organic or inorganic binders provided by a print-
ing head, the particles in the densely packed powder bed are
locally glued together, and compacts generated by this tech-
nology don’t require any additional treatment prior to sinter-
ing. Thus, we can forecast that S-3DP will be more and
more used for the printing of ceramic compacts. The laser
treatment in S-SLS, however, will affect the microstructure
of the powder compact typically in a direction which is not
favorable for the manufacture of advanced ceramic.
(E) Laminated Object Manufacturing (LOM): Origi-
nally known as layered laminated manufacturing, later on
commercialized under the trade name LOM by Helysis (now
Cubic Technologies; Torrance, CA) or CerLAM and Met-
LAM by Javelin 3D (Draper, UT), this AM process deals with
the 3D building of parts by the lamination of paper, tapes, or
similar shapes at low-medium temperature and pressure. LOM
has been considered as an alternative for the lamination of
green ceramic tapes into 3D dense or semidense parts.74–81
The main advantage of this technique is the possibility of
producing laminates directly from green tapes, which in turn
are produced by extrusion,82 tape casting,74, 75 or preceramic
paper.80, 81 In addition, LOM processing: (1) does not intro-
duce an extra organic component that must be eliminated dur-
ing the thermal treatment process, such as for
stereolithography; (2) compared to low-temperature lamina-
tion process, the applied pressure and temperature are lower,
which prevents delamination caused by an inhomogeneous
pressure distribution in complex shapes; (3) does not require
higher capillary forces to promote the union between adjacent
tapes during thermal treatment; (4) allows the lamination of
(b)
July 2015 Additive Manufacturing of Ceramics
water-based green tapes, which is not successfully achieved by
thermo-compression, for example.
Compared to some conventional manufacturing processes
such as injection molding,83 extrusion,84 or roll pressing,85
LOM parts exhibit moderate values of flexural strength at
relatively high porosities.75, 86 Values of mechanical strength
for silicon nitride parts obtained by LOM of ~475  34 MPa
have been reported,87 which is comparable to silicon nitride
ceramics prepared by pressureless sintering88–90 and higher
than for the components fabricated by gel casting
(54.5  2.3 MPa).91, 92
The main drawbacks of this technology, as in most lami-
nation process, are related with the quality of the interfaces
between tapes and presence of defects. Common issues being
reported are delamination, interfacial porosities, and differen-
tial shrinkage.75, 86
(2) Direct AM Technologies
(A) Direct Inkjet Printing (DIP): Direct inkjet print-
ing is a technology which uses a ceramic suspension to build
up structures by the successive deposition of individual drops
provided by a printing head. The ceramic suspension, or ink,
consists of ceramic particles (typically <30 vol%) dispersed
in a liquid carrier containing different additives to stabilize
the suspension, adjusts its viscosity and surface tension, and
to control the spreading and drying of the deposited drop-
lets.93, 94 Derby et al. modeled the conditions for a stable
drop formation and proposed that the inverse Ohnesorge
number should be in the range 1 < Z < 10 (the Ohnesorge
number is the ratio of the square root of the Weber number
and the Reynolds number),95, 96 even though there is not
always a clear correlation between Ohnesorge number and
the characteristics of drop formation.97 As liquid carrier,
water or organic liquids can be used, depending on the spe-
cific requirements of the printing head. To ensure that cera-
mic particles can be fed as a suspension through the
individual deposition nozzles of the printing head, their size
(d50) is typically kept in the submicrometer range at a narrow
particle size distribution.98 Different to powder-based 3D
printing (P-3DP) DIP builds up structures directly from the
material provided by the ink. DIP technology is capable to
generate complex structures with a high powder packing den-
sity, as it uses ceramic suspensions. Together with a good
sintering activity of the generated green parts, ensured by the
use of submicrometer powders, DIP can provide parts from
oxide and non-oxide ceramics with appreciable mechanical
properties.99–101 Figure 4 presents a printed zirconia molar
crown made by DIP.
Furthermore, due to the small volume (picoliter) of the
individual droplets provided by the printing head, the gener-
Fig. 4. SEM micrograph showing a DIP 3D printed occlusal
surface of a dental crown (image courtesy of Prof. Rainer Telle,
University of Aachen, Germany).
1991
ation of structures with well-controllable material gradients
is possible in multimaterial DIP setups.93 The printing of
complex-shaped structures with overhangs and undercuts can
be realized by printing support structures with an ink provid-
ing a sacrificial material (e.g., carbon black), which is conve-
niently removed during firing in oxidizing conditions.
Currently, DIP of ceramic materials is industrially developed
for the fabrication of 2D components (e.g., printed electron-
ics)102 but not yet well exploited for making 3D parts.
An alternative approach is the direct printing of a prece-
ramic polymer solution, which takes advantage of the solu-
bility of many preceramic polymers in several solvents. Mott
et al. for example used polycarbosilane to produce silicon
carbide parts using DIP, with or without the use of SiC par-
ticles in the ink.103
(B) Filament Extrusion 3D Printing: Robocasting,
Direct Ink Writing (DIW), Fused Deposition Modeling
(FDM): The direct deposition of ceramic slurries is argu-
ably the most used AM technology for the generation of por-
ous structures. It consists in the extrusion through a nozzle
of a viscous ceramic paste in the form of a filament. The
control of the rheological properties of the filament is essen-
tial to prevent deformation of the part after extrusion and
sagging of the filaments, especially when the geometry
includes spanning features as in most porous components.
The technique was originally patented and developed by Ce-
sarano at the Sandia National Laboratories in the USA
under the name of “Robocasting”.104 In Robocasting, a cera-
mic suspension with a high solid loading undergoes a trans-
formation from a pseudoplastic to a dilatant behavior when
extruded in air, triggered by minimal drying of the slurry.
Drying in air, however, limits the minimum diameter of the
printing nozzle to about 500 lm, otherwise clogging is expe-
rienced. To overcome these issues, inks have been developed
which have a reversible gel transformation and are deposited
in a nonwetting (often oil) bath. Briefly, they behave like a
viscous gel when loaded in the printing head, but as they are
extruded the shear stress breaks the gel structure and the vis-
cosity decreases of several orders of magnitude. After extru-
sion, the fluid undergoes a quick gelation increasing again
the viscosity to a level before shearing which prevents defor-
mation of the filament. This behavior can be achieved by
controlled flocculation of a ceramic suspension (e.g., by a
change in pH, ionic strength of the solvent, addition of poly-
electrolytes) to form a gel,105 or by using gelling additives,
for example, by using an inverse thermo-reversible gel.106
Recently, Eqtesadi et al. showed that the use of polyelectro-
lytes and the control of the pH may be a difficult approach
for bioglass inks, because of the glass dissolution and leach-
ing of ions in water. In this case, carboxymethyl cellulose
was successfully used as an additive to disperse the glass
powder and to obtain a suitable rheological behavior.107
Another possibility is the formulation of a ceramic ink con-
taining a polymeric binder and plasticizer (e.g., PVB and
PEG), which can be added up to 23 wt% (relative to the
ceramic phase).108 For fine lattice structures, defects gener-
ated by debinding were not reported. These variations in
Robocasting have been mostly denominated DIW technolo-
gies.
Typical nozzle sizes are in the range 100–1000 lm and,
after drying, the material has a high green density (up to
60%), which allows to achieve an almost complete densifica-
tion upon sintering. The combination of fine and dense fila-
ments makes the technology particularly attractive for the
production of components suitable for several applications,
based on structural and functional ceramics.
(a) Porous Components: Stuecker et al. reported the
production of mullite porous meshes with 100–1000 lm pore
sizes and 225–1000 lm filament diameter, obtaining a rela-
tive green strut density of 55% and a sintered strut density
of 96%.109 Smay et al. produced PZT periodic lattices with a
diameter of 200–400 lm, consisting of linear or radial
1992 Journal of the American Ceramic Society—Zocca et al.
arrays.110 The rod spacing varied from 300 to 1200 lm, thus
determining spanning features. The same material system has
been used more recently to produce PZT arrays for ultra-
sonic sensing applications in the 2–30 MHz frequency
range.111 These arrays were densified and infiltrated with an
epoxy resin to fabricate PZT–polymer composites with 2–2
connectivity. Smay et al. also demonstrated the DIW of ter-
nary mixtures of BaTiO3, BaZrO3, and SrTiO3 by using a
nozzle mixer.112 Furthermore, they produced Ni-BaTiO3
metal-ceramic multilayer structures by using an array of sin-
gle nozzles, demonstrating that by this technique it is possi-
ble to fabricate graded and multimaterial porous structures.
Also for this fabrication method, a major part of the
research has focused on the shaping of bioceramics into por-
ous scaffolds. Genet et al. modeled the compressive strength
of hydroxyapatite porous scaffolds based on a Weibull
approach.113 They could fit well-experimental data showing a
decrease in compressive strength upon increasing rod spacing
(i.e., porosity), keeping a constant rod diameter. Interest-
ingly, even when the rod spacing (400–985 lm) and diameter
(200–610 lm) were varied to keep the porosity constant
(45 vol%), there was almost a four folds decrease in strength
as the rod diameter increased. Franco et al. formulated an
ink containing a thermo-reversible gel which had low viscos-
ity at low extrusion temperature and underwent a quick gela-
tion following deposition in a warm (60°C) oil bath.114 They
produced HA, TCP and biphasic calcium phosphate scaffolds
and found that the microporosity in the struts varied in a
wide range 5%–40% as the content of thermo-reversible gel
was increased, and the bending strength decreased accord-
ingly from ~25 to ~2 MPa.
Miranda et al. produced HA scaffolds which had 39%
total porosity and 15% strut microporosity, and studied the
fracture modes under uniaxial compression.115 These scaf-
folds had a compressive strength of ~50 MPa, and in a simi-
lar work it was shown that this value was doubled after
2 weeks immersion in simulated body fluid.116 Fu et al. pro-
duced particularly strong bioglass scaffolds with completely
dense rods of 100 lm diameter and a total porosity of
60%.117 These scaffolds had unidirectional pores and, as
expected, anisotropic mechanical properties, reaching
136 MPa in the channels direction and 55 MPa in perpendic-
ular direction. The dependence of the compressive strength
on the porosity in the 60%–80% range was also in good
agreement with the Gibson-Ashby model for closed-cells cel-
lular ceramics.118
It is worth noting, however, that by DIW it is also possi-
ble to obtain scaffolds with porous struts by adding a poro-
gen to the ceramic paste.119, 120 Dellinger et al. added Poly
(methyl methacrylate) (PMMA) microspheres to the ink and
produced scaffolds with three levels of porosity: macropores
(100–600 lm) were designed by controlling the arrangement
and spacing between rods of HA; micropores (1–30 lm) were
produced within the rods by including PMMA microspheres;
submicrometer pores (less than 1 lm) were the result of
incomplete sintering.119
A further degree of versatility for this 3D printing tech-
nique can be achieved when using extrusion nozzles which
enable the fabrication of cross sections different from the
standard circular one, such as square, hexagonal and other
complex micrometer-sized geometries,121 or even hollow fila-
ments.122
In parallel with the extrusion of ceramic pastes, and in
analogy with the most popular process used for the produc-
tion of polymeric components, another approach has been
developed: the FDM of ceramics. This techniques is based
on the extrusion of a mixture of a ceramic powder mixed
with a significant amount of a polymer, which is melted dur-
ing the extrusion because the printing head is heated, and
solidifies by cooling immediately after to prevent deformation
of the filament. Ideally, in analogy with FDM of conven-
tional polymers, a continuous filament should be produced
Vol. 98, No. 7
and fed to the printing head, but it is very difficult to prepare
a filament with a high loading of ceramic particles that is not
brittle and can be collected in spools to be continuously sup-
plied to the printing head. After printing, the polymeric bin-
der is burnt out and the ceramic phase is sintered. Grida
et al. extruded a mixture of 55 vol% zirconia in wax through
nozzles with diameter between 76 and 510 lm.123 The mini-
mum diameter of the nozzles is usually larger in FDM than
in DIW. The same authors observed that fine log-pile struc-
tures could be produced with a 100 lm diameter filament
but not with a 76 lm filament. The reason was that the latter
filament solidified too quickly in ambient air. Park et al.
instead extruded 40 wt% HA mixed with PCL at 100°C
through nozzles with diameter 400–600 lm.124 They also
measured the influence of the deposition pattern on the com-
pressive modulus and they studied the attachment and prolif-
eration of osteoblast-like cells.
Kalita et al. similarly produced composites polypropylene-
TCP composites by FDM, and obtained scaffolds with a
compressive strength of 12.7 MPa for 36% total porosity
and a pore size of 160 lm.125 Bandyopadhyay et al. investi-
gated the processing of piezocomposites by FDM, both using
a negative approach (producing a polymeric mold and filling
it with a ceramic slurry), and a positive approach, extruding
a polymeric filament loaded with 50–55 vol% PZT powder.
A spool of filament (1.75 mm in diameter) was first produced
and extruded through the nozzle of a commercial FDM 3D-
printer, at a temperature of 140°C–200°C, to produce a 3-3
piezocomposite ladder structure after sintering and impregna-
tion with epoxy resin.126 At Rutgers University (New Jersey),
an FDM apparatus has been developed which is able to
deposit up to four different materials in the same layers,
employing four deposition units.127 Jafari et al., in the same
work also produced a multilayer piezoelectric component
transducer composed of soft and hard PZT, and they sug-
gested the fabrication of components based even more com-
plex composite materials, for example, through the
simultaneous deposition of ceramic, metal, electrode, and
fugitive phases.
Scaffolds produced by DIW generally possess better
mechanical properties compared to scaffolds produced by
powder-based technologies. Figure 5 shows the dependence
of the compressive strength on the porosity, for various
porous scaffolds fabricated using different AM technologies
and having a total porosity in a wide range (20–80 vol%).
It is not easy to directly compare scaffolds with different
geometries and made of different materials, but the exam-
ples reported in the plot include only biomaterials, in par-
ticular based mainly on calcium phosphates. Scaffolds
produced by DIW generally had a cubic geometry with
orthogonal pores, whereas scaffolds produced by indirect
AM techniques mostly had a cylindrical geometry with
orthogonal or radial pores resembling the structure of bone
implants. Both geometries are possible to be replicated by
direct and indirect AM, but a cubic geometry is used in the
case of DIW because it simplifies the deposition path,
whereas indirect technologies are almost geometry indepen-
dent.
Plotting the y-axis with a logarithmic scale allows us to
observe that the strength data tend to follow a linear depen-
dence with the porosity, corresponding to a well-known
exponential law: r (P) = r0 exp (
bP), where r (P) is the
compressive strength, r0 a constant related to the strength of
the zero-porosity material, b another constant and P the
porosity.128
The absolute values of compressive strength are distrib-
uted in a wide range of values, from 0.5 to over 100 MPa;
however, there is a clear trend. Scaffolds produced by P-3DP
have, in particular, a strength which is up to one order of
magnitude lower than the strength of scaffolds made by
DIW. It seems reasonable to correlate this observation with
the microporosity in the scaffold struts: even though this
July 2015 Additive Manufacturing of Ceramics 1993

Fig. 5. Compressive strength versus porosity for bioceramic porous components fabricated using different AM technologies.

value depends also on the specific material system used, there
is a clear influence of the process, which can be found when
examining the properties of parts shaped without designed
porosity. Parts produced by P-3DP generally have a high
level of residual microporosity, whereas parts produced by
liquid-based processes (such as DIW and SL) can be almost
dense after sintering. A clear example is given in Fig. 6,
which shows HA scaffolds produced by P-3DP (a,b) and
DIW (c,d).
The scaffold produced by P-3DP had only the presence of
sintering necks between granules, leading to a high residual
porosity,129 whereas the struts produced by DIW could be
sintered to much higher density, although complete densifica-
tion was not achieved even in this case, as there was ~15%
residual microporosity in the struts.115 As mentioned before,
for specific applications such as TE, this strut microporosity
may be desirable, despite detrimentally affecting the mechani-
cal properties of the scaffold.
(b) Monolithic Components: Previously, we mentioned
that extrusion-based AM produces potentially dense struts;
however, there are issues when this process is used to pro-
duce monolithic ceramics, according to our definition. First
of all, in technologies which make use of a high amount of
polymeric binder, such as FDM, the debinding is potentially
problematic for monolithic components. Even for DIW,
which uses a feedstock which is almost free of organics, the
stacking of filaments creates a textured structure which is far
from that of a homogeneous ceramic monolith produced by
traditional ceramic technologies and this, as a consequence,
affects also the surface quality and the mechanical properties
of the printed part. As an example, which also illustrates the
difference between AM technologies for ceramics in terms of
surface finish of the printed part, we show in Fig. 7 images
of a zirconia impeller produced by SL [Fig. 7(a)] and of alu-
mina crucibles [Fig. 7(b)] produced by Robocasting. Note
that, with SL, a much smoother surface can be obtained,

(a) (b)

(c) (d)

Fig. 6. SEM micrographs of hydroxyapatite scaffolds produced by: (a) and (b) powder-based 3D printing; (c) and (d) by direct ink writing.
1994 Journal of the American Ceramic Society—Zocca et al.
(a)
Fig. 7. Optical photographs of: (a) zirconia impeller produced by stereolithography (image courtesy of Lithoz Gmbh, Austria); (b) alumina
crucibles and labware produced by Robocasting (image courtesy of Joe Cesarano, Robocasting Enterprises, USA).
leading to parts with excellent strength. In the case of Robo-
casting (and similarly for DIW and FDM), the surface qual-
ity depends on the diameter of the filament and diameters
below 100 lm can be used.130, 131 The result is generally a
compromise between the desired quality and processing time.
Figure 7(b) also illustrates the surface texture that is created
by the deposition of filaments with a diameter ~200 lm.
We should also mention that it is possible to couple fila-
ment-based AM technologies with computer numerical con-
trol machining that, at the same time, shapes the object and
produces a good surface finish.5 In any case, the surface tex-
ture is not always problematic, but since it is a specific char-
acteristic of these extrusion-based technologies, it has to be
taken into account when designing a component. Conclud-
ing, extrusion-based AM technologies are particularly suit-
able for producing arrays of separate struts, but are not able
to produce proper monolithic ceramics according to a most
general definition.
(3) Negative AM Technologies
While most current research focuses on the use of AM tech-
nologies that can directly produce ceramic components, nega-
tive replica methods are attractive because they can
overcome at least some of the limitations in the physical,
functional, and geometrical characteristics of components
produced. In this approach, a sacrificial polymeric mold is
fabricated by AM and then, generally, impregnated with a
ceramic slurry. After polymer removal (by decomposition,
burn-out or dissolution) and sintering, a ceramic component
is obtained; however, infiltration is by itself a delicate process
especially for complex and lattice structures which are the
ones taking the most advantage of AM technologies. The
fabrication of molds by AM adds a step to the processing
route, but it allows the use of AM technologies for polymeric
materials, some of which are at a greater stage of develop-
ment compared to AM of ceramics, implying an easier pro-
cess and often better resolution and surface quality. We
should also note that negative replica methods have been
long used for the casting of metals, by producing using AM
either a ceramic mold or a wax pattern for the investment
casting process.132
Almost every AM technology can be used to fabricate the
mold. Detsch et al. used a printer which deposits dropwise a
wax, and impregnated the wax mold with a HA slurry.133
They also replicated a similar geometry by Robocasting,
obtaining scaffolds with 44% porosity in the case of negative
AM and 37% for the sample produced by Robocasting. In
both cases, the pore size and strut diameter were 300 and
500 lm, respectively, and the strut relative density reached
96% after sintering, indicating the two approaches are fully
interchangeable.
A similar fabrication route was followed to produce by SL
epoxy molds to generate TiO2/SiO2 diamond structures for
application as photonic crystals.134 SL has also been used to
produce resin molds for HA scaffolds with 50% geometrical
Vol. 98, No. 7
(b)
porosity; in this case, a gel-casting technique was employed
to improve the strength of the ceramic green body.135
Bose et al. produced molds by FDM of a commercial
thermoplastic polymer, which were subsequently impregnated
with TCP or Al2O3 slurries to produce porous scaffolds, and
investigated the influence of the pore size (300–500 lm) and
porosity (25%–45%) on in vitro cell tests.136 Guo et al.
instead produced polystyrene molds by SLS, which were then
used as molds for PZT ceramic components obtained by gel-
casting; the selection of the polymer and of the decomposi-
tion thermal treatment is essential to avoid deformation of
the parts, which was instead observed when using molds pro-
duced by FDM of ABS polymer.137 Ortona et al. applied a
technology (“Polyjet”) which dropwise deposits a photo-cur-
able resin to produce a polymeric template possessing a com-
plex, regular three-dimensional cellular structure.138 In this
case, the polymeric template was coated with a SiC slurry
and after pyrolysis it was reactively infiltrated with molten Si
to generate RBSiC cellular structures with hollow struts.
Finally, Franchin et al. used FDM to fabricate polymeric
scaffolds which were infiltrated by a geopolymer slurry,
resulting in parts with different pore geometries.139
IV. Feasibility of AM Processing of Monolithic Ceramics
As already mentioned, AM of dense monolithic ceramics is
still a major challenge. The problem can be reduced down to
a few issues:
1. The superior physicochemical properties of advanced
ceramics critically depend on their microstructure,
which is intrinsically linked to the particle packing of
the powder compact and the sintering schedule
applied, as well as to their composition (e.g., presence
of sintering aids). Obtaining a green compact with the
desired microstructure is far from trivial when using
AM as an advanced shaping technology. From an eco-
nomical perspective, this statement should be under-
stood as a well-meant advice: the integration of AM
technologies into existing industrial ceramic production
is a prerequisite for making AM of ceramics reliable,
high quality, and economically viable.
2. Advanced ceramics are generally produced from very
fine powders showing a poor flowability. A good flow-
ability is, however, required for the deposition of lay-
ers in powder-based AM, such as P-3DP or P-SLS.
Spreading fine, submicrometer powders into thin layers
is very difficult to impossible. Furthermore, fine cera-
mic powders are prone to agglomeration and electro-
static charging and, thus, do not pack well. Packing
densities of dry ceramic powders in a particle size
range from 20 to 100 lm are generally below 50% of
the theoretical density (typically between 25% and
45%).140 The density also depends on the particular
layer spreading technology used, that is, blade, roller
etc. The use of ceramic suspensions or pastes, on the
other hand, enables deposition of very thin layers,
July 2015 Additive Manufacturing of Ceramics 1995

down to 10 lm thickness, and the use of finer particu- (a)

lates. After removal of the liquid phase or plastifier,
appreciably high densities can be achieved. However,
according to the state of the art, the majority of cera-
mic suspensions or pastes used as feedstock for AM
are based on a significant amount of organic binders
40 to 60 vol%.11, 60 The removal of organics remain-
ing as a residue in the produced compact is a well-
established process, but limitations are still existing
when it comes to monolithic structures exceeding cer-
tain wall thicknesses or volumes. Nonetheless, organic
plastifiers are favored because their rate of evaporation
can be very well adapted to the specific needs of the
particular AM process. In this context, one of the
most successful technologies is stereolithography (SL),
using ceramic slurries containing photosensitive resins
for the layerwise buildup of ceramic compacts. So far,
little work has been devoted to the development of (b)
AM processes for the deposition of thin layers using
water-based ceramic suspensions or pastes with low
organic content.70–72Slurries facilitate the use of pow-
ders with a size ranging from <100 nm to >100 lm,
and classical ceramic processing know-how can be
applied for their preparation. Moreover, interparticle
interactions can be very well controlled and the mobil-
ity of particles within the slurry enables their free set-
tling, both a prerequisite for an effective powder
packing. The high powder packing obtained by the
LSD process73 can be understood within the frame-
work of a model taking into account the forces acting
on each individual particle, as represented in the draw-
ing of Fig. 8. In the LSD process, a thin (typically
100 lm) layer of a water-based slurry is spread on pre-
viously deposited and dried layers forming the powder
bed.73 As soon as the slurry is brought into contact Fig. 8. Illustration of the forces acting on each individual particle
during layer deposition in: (a) slurry-based 3DP and SLS, (b) gas
with the surface of the powder bed, capillary forces
flow assisted powder deposition. Fg = mg, with m the mass of the
will draw the liquid phase of the slurry into its net- individual particle and g the gravitational constant, Fpp is the
work of pores [see Fig. 8(a)]. The concomitant mate- interparticle force, Fd the force induced by the flow field of the
rial flow drags the particles of the freshly deposited liquid or gaseous medium into the porous support, that is, the
layer into locations at the surface of the powder bed powder bed, and Ff the buoyancy in the suspension.
which are characterized by a maximum contact with
the already settled nearest-neighbor particles, accord-
ing to a random close packing assemblage. Further- powders (d50 = 17 lm).140 Compared to slurry-based
more, each individual particle has the chance to freely technologies, the low mobility of individual particles in
settle into this favored position, whereas the rest of the the dry-state is hindering their free settling to a closed
not yet settled adjacent particles, which are still float- packed configuration in the recoating process, irrespec-
ing in the suspension, do not hamper the settling pro- tive of the type of recoating unit used.
cess. The particle settling is basically dominated by the 3. Concerning ready-to-use parts, the first question must
three forces acting on each individual particle: be: can the sintering step be directly integrated into an
Fg = mg, with m the mass of the individual particle AM process? This would imply that sintering is
and g the gravitational constant, Fpp representing the achieved simultaneously to adding material. Sintering
interparticle force and Fd representing the force is a very delicate process step, as it ultimately defines
induced by the flow of the liquid medium into the por- the microstructure of the material, its final physico-
ous support, that is, the powder bed. In suspensions, chemical properties and also has a strong influence on
the buoyancy Ff must also be taken into consider- the parts geometry. Therefore, a technology allowing
ation. In case of AM processes using dry powders, Ff sintering during additive buildup of a ceramic part
and Fd do not exist, and flowing of particles in the with physical properties comparable to a convention-
layer deposition process is dominated by gravitational ally processed part is hard to envision. Based on the
and interparticle force. It is intuitively clear that for SLS technology, ceramic powders have been sintered
smaller particles the interparticle force will dominate, layer by layer by the local application of laser energy.
as the gravitational force scales as the third power of Different to conventional sintering, which works with
the particles diameter, whereas the particles surface, heating rates of some K at most up to some 10 K/min
responsible for the interparticle force, as the second and dwell times at maximum temperature in the range
power of the diameter. This, in turn, explains why of hours, the so called laser sintering is a shock ther-
small particles do not pack well when freely set- apy for the material which by far exceeds thermal gra-
tled.Figure 8(b) illustrates an alternative approach in dients that ceramic materials, with their typically poor
which a gas flow is established throughout the powder thermal shock resistance, can withstand. Accordingly,
bed. The effect of this gas flow is comparable to the densification of the powder to a nonporous material is
slip casting process, and has shown its potential for hardly achieved. From all the experimental work done
the deposition of fine powders (<20 lm particle size) in this direction, it can be concluded that short anneal-
with appreciable packing densities which, however, still ing times associated with a local laser irradiation
not exceed 45% TD when using commercial alumina can result in the bonding of particles by the in situ
1996 Journal of the American Ceramic Society—Zocca et al.
formation of a liquid phase. In this context, two differ-
ent laser annealing strategies are applied on the basis
of existing laser systems. The first strategy maximizes
the time the material is annealed at a certain position,
and laser systems providing continuous laser radiation
are used. In the second strategy, lasers providing short
pulses, typically of the order of some nanoseconds, are
used to apply extremely high radiation intensities. The
first approach aims to initiate, similarly to classical sin-
tering, diffusion driven material transport and the for-
mation of a liquid phase for consolidation of powder
particles.47 The second approach aims to produce a
liquid phase for a short time interval and, by the for-
mation of an ablation plume, an additional force par-
allel to the surface normal resulting in a compression
and fusion of powder particles.49
The formation of a liquid phase is the only way to achieve
a rapid densification, which is fast enough to meet the
requirements of SLS technologies. In case an appreciable
densification is achieved, cracks are formed due to the high
thermal gradients introduced into the material by the local
laser treatment.48, 58 To the best of our knowledge no dense
and crack free ceramic parts comprising multiple layers of
deposited material could be produced by SLS or melting up
to now. It is noteworthy to observe that postsintering of
laser-sintered and not fully densified ceramic bodies not
always leads to a significant additional densification. The
local formation of a liquid phase during laser sintering
results in extended structural defects in the sintered body,
such as larger pores and cracks, which cannot be removed
during sintering even though the initial powders used provide
a high sintering activity. Figure 9 compares the pore size dis-
tribution of a laser-sintered porcelain part with that of the
powder bed from which it was generated through slurry-
based selective laser sintering (S-SLS). In S-SLS, the powder
bed has a density comparable to conventionally produced
green parts, that is, ~60%TD. However, it is clearly notice-
able in Fig. 9 that the laser treated material possesses large
residual pores, which would significantly reduce its sintering
activity during a post-laser-sintering heat treatment.141
Furthermore, one must bear in mind, that sintering does
not only result in the densification of a powder compact, but
also determines the phase composition of the material. In
this context, liquid phase formation might provide a route
for the rapid densification of a powder compact, but not nec-
essarily result in the desired phase composition and, thus,
does not result in the desired properties of the part gener-
ated. Alternatively, sintering could be a final treatment
embedded in an AM process. This scenario, however, is basi-
cally a direct replication of the powder process chain for the
manufacture of ceramics, in which shaping of a powder com-
pact precedes sintering and, as already mentioned, it must be
Fig. 9. Pore size distribution of a laser-sintered part and the
powder bed from which the part was generated through the S-SLS
process, measured by mercury intrusion porosimetry.
Vol. 98, No. 7
ensured that the sintering activity of the powders used is not
reduced in the AM process prior to the final sintering.
We conclude this section, by discussing examples for clas-
sical manufacture and industrialized AM of ceramic parts.
The first example illustrates how small series and single piece
production of ceramic parts using conventional technologies
is a time-consuming and complex process, which, depending
on the particular geometry of the part, also consumes a lot
of material. In Fig. 10, a housing structure of an avionic pod
camera, made from silicon nitride, is shown. The housing
green body was machined from a green powder compact
weighing ~130 kg, prepared by cold isostatic compaction of
a silicon nitride powder. About 80% of the powder compacts
material has been removed in a subtractive manufacturing
process, employing classical CAM/CAM machining technol-
ogy. The machining took seven days working round the
clock. In addition to extensive knowledge and experience in
ceramics manufacture and processing, a powder compact
with an extremely homogeneous microstructure and powder
packing density is a prerequisite to obtain a net-shape-sin-
tered body. This example is not uncommon for the ceramic
manufacturing industry. Companies specialized on the manu-
facture of individual complex ceramic parts receive hundreds
of orders per month with about 50% of them requesting new
designs at an average lot size of 10–20. As already discussed,
the manufacture of suitable green powder compacts by AM
technologies is still a challenge. Stereolithography might be
capable to provide homogeneous green bodies with a powder
packing density sufficient for a complete densification during
sintering, but binder burn out would be extremely time-con-
suming to impossible at such voluminous and compact struc-
tures and, even worse, lower packing densities, as compared
to conventionally manufactured powder compacts, result in
an enhanced sintering shrinkage, which significantly raises
the challenge to obtain net shape parts after sintering.
One the other hand, an example of a successful integration
of AM into an existing processing chain has been announced
very recently.142 Schunk (Schunk GmbH, Lauffen/Neckar,
Germany) has advanced its shaping technology for the produc-
tion of silicon-infiltrated, RBSiC components by using pow-
der-based 3D printing (P-3DP). With this new approach,
labeled IntrinSiCÒ (Schunk GmbH), large, complex, mono-
lithic ceramic components can be produced with a high degree
of design flexibility and short production lead times (see
Fig. 11). Additional advantages of this process, which derive
directly from the use of an AM technology, are: no time-con-
suming manufacturing going from patterns to molds (CAD-
data means parts immediately ready for manufacturing) and
design modification by the click of a mouse. From a ceramist
point of view, the silicon-infiltrated silicon carbide material
system is certainly very well suited for 3D printing, as the
residual porosity of the printed parts is of no concern, the par-
ticle size of the SiC powders used ensures excellent flowability
and formation of defect-free layers, and there is almost no
shrinkage associated with sintering and silicon infiltration of
the printed powder compacts. Moreover, the binder infiltrating
the SiC powder in the printing process may add just the right
amount of reactive carbon required to form secondary SiC
after its pyrolysis during sintering and Si infiltration. Aside
from these obvious synergies between material system and AM
process, these green SiC powder compacts formed by 3D print-
ing do not require a special treatment before and during sinter-
ing. Printed parts can be processed the same way as compacts
formed by conventional technologies, such as slip casting and
isostatic compaction. Thus, not only the material system per-
fectly fits to the specific needs of 3D printing but, on top of
that, all criteria for a successful implementation of AM in an
existing processing chain are fulfilled. The striking advantage
of the implementation of 3D printing in a conventional pro-
cessing chain is the combination of mass production of stan-
dardized products with exclusive small series production, as
well as single piece production. Except for 3D printing, all
July 2015 Additive Manufacturing of Ceramics
Fig. 10. Complex housing structure of an avionic pod camera made from silicon nitride (image courtesy of FCT Ingenieurkeramik GmbH,
Germany). The camera housing was machined from a ~130 kg powder compact by employing classical CAD/CAM technology (lot size 25).
Fig. 11. Tube of a recuperative burner made by 3D printing of SiC
(image courtesy of Schunk Ingenieurkeramik GmbH, Germany). The
tube shown has a length of 900 mm, and its larger diameter is
320 mm. The material is silicon-infiltrated, reaction-bonded silicon
carbide (RBSiC), with a density of 3.05 g/cm3.
equipment and know-how derives from existing and well-
established processes.
V. State of the Art of AM Technologies: Comparison
Between Ceramics and Metals
As AM for metals has graduated to a rather mature technol-
ogy with an ever increasing industrial support and expanding
applications, it is worth having a closer look into this tech-
nology. The hope is that we, as ceramists, would benefit from
these successful experiences, and that a similar trend would
occur also for ceramic AM technologies.
Economically relevant technologies for metals are mostly
based on powders, with a particle size range 5–50 lm. Metal
wires are also used as feedstock in industrial equipment, with
pastes or slurries currently of interest mainly to research lab-
oratories.143 There are two basic AM processes which differ
in the ways to apply the powder during the build process:
the powder bed fusion process, which includes the SLS/SLM
processes, and directed energy deposition (DED), which
includes the LENS process.143 In terms of the powder bed
processes, their widespread use for metals derives from wel-
dability of the materials systems. Whether they are called
SLS or SLM processes, the metal feedstock is melted and
completely fused, resulting in a theoretically dense compo-
nent, or at least higher than 99% of the reference density.
The energy for fusing the powder can be provided by two
different types of energy sources, laser or electron beam, with
1997
the latter not applicable to ceramics, as most ceramic materi-
als do not possess sufficient electron conductivity. The suc-
cess of AM of metals is also based on a broad base of
existing knowledge concerning the handling of metal powders
and laser material interactions for the important alloy sys-
tems in use in AM. For example, the study of laser material
interactions for similar processes, such as laser cladding, has
been long-standing, and there is a strong foundational and
multidisciplinary knowledge base from which to draw. The
alloys selected can be easily welded, making them very much
compatible with achieving defect-free microstructures during
AM. It is also important not to discount the commercial
availability of high quality metal powders in the appropriate
size ranges. With the wide use of powder metallurgy, laser
cladding, and coating technologies, there is a built-in supplier
base that can produce consistent powder sizes and chemis-
tries. This wide availability of feedstock material makes the
powder-based AM for specific alloy systems more economical
than for different materials systems which are not as widely
used. From an economic relevancy, DED processes have a
stronger economic argument, as they can provide product
repair and material grading capabilities that the powder bed
systems cannot get anywhere near close to. The DED process
has also much higher deposition rates, but with high deposi-
tion rates, the ability to produce fine features and internal
structures is not available. However, the most stringent limi-
tation to the powder bed fusion processes is that the designs
are limited to a single material. The future in metallic sys-
tems will be the ability to use multimaterial systems (metal/
metal, functional grading, metal/plastic, metal/ceramic inter-
faces), and the powder bed fusion systems will not be able to
produce these different material combinations.
From an historical perspective, the major advances that
have helped to mature AM of metals involve the develop-
ment of improved enabling technologies. Energy sources, pri-
marily lasers but also e-beam devices, have been greatly
improved and made more reliable. The development and
fielding of fiber laser systems and the integration of 1 lm
wavelength lasers has probably done more to advance the
technology than anything else. In addition, process controls,
motion control systems, and others, such as computer sys-
tems, have finally caught up with the process.
The primary market currently driving AM for metals is
the aerospace market. After aerospace, biomedical and
energy related applications have, at this stage, the greatest
potential. As for the automotive industry, most of the mar-
ket demand will be for relatively low-scale applications,
where complexity or cost is driving innovation. There is also
1998 Journal of the American Ceramic Society—Zocca et al.
a market driver for other uncommon materials. One in par-
ticular that is receiving attention is refractory metals. The
selection of AM as a process for fabricating aerospace com-
ponents was primarily based on the weight ratio between the
raw material used for a component and the weight of the
component itself. This so called “buy-to-fly” ratio can be as
high as 15–20 for flying components, adding a lot of cost to
the component for material and machining. Most of the
materials selected are expensive feedstock materials, and their
development for AM was primarily based on the need to
replace expensive forgings and castings that were largely
machined away before being put into service. AM enables
“buy-to-fly” ratios of nearly 1. On the other hand, the design
freedom granted by AM processes has largely not yet been
exploited.
The choice of technology to directly fabricate parts
depends on the application and the material. Compared to
AM of metals, the current limitations in AM of ceramics are
not the availability of starting materials for the production
of the feedstock or motion control and computer systems.
Ceramic powders with a wide range of characteristics are
commercially available for almost all ceramic systems. There
is also a great interest from the market for ceramic parts that
are additively manufactured. Molds used for forming ceramic
bodies are expensive and require thousands of parts to be
produced before becoming competitive and cost-efficient. The
marked demands single piece and small series production
with a lot size typically in the range between 10 and 20 parts,
which requires CAD/CAM processing and machining of
ceramic green bodies. Single piece and small series produc-
tion perfectly suited AM, which, moreover, provides a much
higher degree of freedom in design.
Despite the larger number of AM technologies suitable for
processing ceramics, each with its strengths and limitations,
the main problem seems to lie in the coupling of a specific
AM technology with the formulation of a proper feedstock
that would enable, in particular, the fabrication of dense
ceramic parts with optimal properties (density, mechanical
strength, surface finish, etc.).
VI. Future Developments
Similarly to what occurred for AM of metals, it can be
expected that powder-based technologies, in which a powder
bed is built-up layer by layer, will become the most economi-
cally relevant class of additive technologies for ceramics, as
their specific merits are: low cost of raw materials, easy scale
up to very large parts or parallel processing of multiple parts,
high flexibility in design, and direct manufacture of ceramic
green bodies with low content of organic additives. Even
multimaterial approaches have been envisioned and are
under development.144, 145 However, a very important issue
that nowadays remains to be satisfactorily solved is the low
powder packing density and stability of the powder bed,
when dry powders are used as feedstock, which implies the
need of support structures.
Instead of dry powders, ceramic slurries can be used for
the layer-by-layer buildup of powder beds with high powder
packing densities. Moreover, slurries facilitate the use of
powders with a size ranging from <100 nm to >100 lm, and
classical ceramic processing know-how can be applied for
preparing slurries. This concept is already realized in SL,
which at present is the only technology available on the
market for AM of advanced ceramics and which will likely
dominate markets for the manufacture of high precision
small components in the future, also. Based on these consid-
erations, the following strategies and outlook for the pro-
duction of high quality ceramic parts via AM can be
formulated:
1. AM of high-quality ceramic parts is inevitably linked
to the sintering of powder compacts with appropriate
Vol. 98, No. 7
particle size distribution and homogeneous particle
packing density;
2. cost-effective AM of high quality dense monolithic
ceramic parts must follow the powder processing route
of classical ceramic manufacturing;
3. the more classical powder processing know-how can
be implemented into AM of ceramics, the higher the
quality of parts and more cost-effective the manufac-
turing process will be;
4. AM technologies capable of providing powder com-
pacts with appropriate particle size distribution and
homogeneous, high particle packing density and with
maximum freedom in design will be economically the
most successful. Until this goal is reached, the success-
ful implementation of AM technologies for ceramics at
an industrial scale will be limited to a few specific
material systems and component geometries.
5. AM technologies capable of producing complex cera-
mic parts with very fine (from micro down to the
nano-size scale) filigree but dense lattice structures
(struts) will hold an outstanding position within all
AM technologies, as ceramic materials are thermody-
namically more stable than metals and polymers. In
this context, PDCs can be considered as an enabling
technology.
6. the bioceramic market is currently driving, and will
continue to drive, the development of AM of ceramics,
followed by the advanced ceramics and silicate ceram-
ics markets.
7. AM of ready-to-use ceramic parts will remain a chal-
lenge, and technical solutions for the manufacture of
dense monolithic ceramic parts will likely significantly
deviate from the powder processing route of classical
ceramic manufacturing.
Acknowledgments
The authors gratefully acknowledge the generous assistance and valuable
information provided by Todd Palmer (Pennsylvania State University) on
additive manufacturing of metals and of Thomas Mühler (Clausthal University
of Technology) on additive manufacturing of ceramics.
References
1
ASTM Standard F2792, Standard Terminology for Additive Manufactur-
ing Technologies. ASTM International, West Conshohocken, Pennsylvania,
2012.
2
I. Gibson, D. W. Rosen, B. Stucker, Additive Manufacturing Technologies.
Springer, New York, NY, 2010.
3
T. Wohlers and T. Caffrey, “Additive Manufacturing and 3D Printing State
of the Industry Annual Worldwide Progress Report. 2014”; Wohlers Associ-
ates Inc., Fort Collins, CO, 2013.
4
J. D. Mcguffin-Cawley, “Manufacturing Technology: Rapid Prototyping”;
pp. 415–37 in Ceramic Science Technology. Edited by R. Riedel and I.-W.
Chen. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2011.
5
N. Travitzky, et al., “Additive Manufacturing of Ceramic-Based Materi-
als,” Adv. Eng. Mater., 16 [6] 729–54 (2014).
6
P. F. Jacobs, Rapid Prototyping & Manufacturing: Fundamentals of Stereoli-
thography. Society of Manufacturing Engineers, Dearborn, MI, 1992.
7
J. C. Nelson, S. Xue, J. W. Barlow, J. J. Beaman, H. L. Marcus, and D. L.
Bourell, “Model of the Selective Laser Sintering of Bisphenol-A Polycarbon-
ate,” Ind. Eng. Chem. Res., 32 [10] 2305–17 (1993).
8
M. Atif Yardimci and S. G€ ußceri, “Conceptual Framework for the Thermal
Process Modelling of Fused Deposition,” Rapid Prototyp. J., 2 [2] 26–31
(1996).
9
S. Das, M. Wohlert, J. J. Beaman, and D. L. Bourell, “Producing Metal
Parts with Selective Laser Sintering/Hot Isostatic Pressing,” JOM, 50 [12] 17–
20 (1998).
10
J. Mazumder, J. Choi, K. Nagarathnam, J. Koch, and D. Hetzner, “The
Direct Metal Deposition of H13 Tool Steel for 3-D Components,” JOM, 49
[5] 55–60 (1997).
11
M. L. Griffith and J. W. Halloran, “Freeform Fabrication of Ceramics via
Stereolithography,” J. Am. Ceram. Soc., 79 [10] 2601–8 (1996).
12
http://www.stratasys.com/3d-printers/design-series/connex-systems,
Accessed on 27/02/2015, (2015). Stratasys, Eden Prairie, MN.
13
M. Vaezi, S. Chianrabutra, B. Mellor, and S. Yang, “Multiple Material
Additive Manufacturing – Part 1: A Review,” Virtual Phys. Prototyp., 8 [1]
19–50 (2013).
July 2015 Additive Manufacturing of Ceramics
14
D. C. Hofmann, et al., “Compositionally Graded Metals: A New
Frontier of Additive Manufacturing,” J. Mater. Res., 29 [17] 1899–910
(2014).
15
J. D. Cawley, “Solid Freeform Fabrication of Ceramics,” Curr. Opin. Solid
State Mater. Sci., 4 [5] 483–9 (1999).
16
P. Colombo, G. Mera, R. Riedel, and G. D. Sorar u, “Polymer-Derived
Ceramics: 40 Years of Research and Innovation in Advanced Ceramics,” J.
Am. Ceram. Soc., 93 [7] 1805–37 (2010).
17
J. Deckers, J. Vleugels, and J.-P. Kruth, “Additive Manufacturing of
Ceramics: A Review,” J. Ceram. Sci. and Tech., 5 [4] 245–60 (2014).
18
W. M. Sigmund, N. S. Bell, and L. Bergstr€ om, “Novel Powder-Processing
Methods for Advanced Ceramics,” J. Am. Ceram. Soc., 83 [7] 1557–74 (2000).
19
B. Y. Tay, J. R. G. Evans, and M. J. Edirisinghe, “Solid Freeform Fabri-
cation of Ceramics,” Int. Mater. Rev., 48 [6] 341–70 (2003).
20
W. Kollenberg, “Keramik und Multi-Material 3D-Druck (Ceramic and
Multi-Material 3D-Printing),” Keram. Z., 66 [4] 233–6 (2014).
21
D. L. Bourell, H. L. Marcus, J. W. Barlow, and J. J. Beaman, “Selective
Laser Sintering of Metals and Ceramics,” Int. J. Powder Metall., 28 [4] 369–81
(1992).
22
E. M. Sachs, J. S. Haggerty, M. J. Cima, P. A. Williams, and Inventors:
Massaschussets Institute of Technology, Assignee, “Three-Dimensional
Printing Techniques”; US Patent 5,204,055; April 20, 1993.
23
C. R. Deckard and Inventor: Board of Regents, The University of Texas
System,Assignee, “Method and Apparatus for Producing Parts by Selective
Sintering”; US Patent 4,863,538; September 5, 1989.
24
A. Butscher, M. Bohner, N. Doebelin, L. Galea, O. Loeffel, and R.
M€ uller, “Moisture Based Three-Dimensional Printing of Calcium Phosphate
Structures for Scaffold Engineering,” Acta Biomater., 9 [2] 5369–78 (2013).
25
G. Cesaretti, E. Dini, X. De Kestelier, V. Colla, and L. Pambaguian,
“Building Components for an Outpost on the Lunar Soil by Means of a Novel
3D Printing Technology,” Acta Astronaut., 93, 430–50 (2014).
26
A. Zocca, C. M. Gomes, A. Staude, E. Bernardo, J. G€ unster, and P. Co-
lombo, “SiOC Ceramics with Ordered Porosity by 3D-Printing of a Preceramic
Polymer,” J. Mater. Res., 28, 2243–52 (2013).
27
A. Zocca, et al., “3D-Printed Silicate Porous Bioceramics Using a Non-
Sacrificial Preceramic Polymer Binder,” Biofabrication, 7 (2015). doi: 10.1088/
1758-5090/7/2/025008.
28
G. A. Fielding, A. Bandyopadhyay, and S. Bose, “Effects of Silica and
Zinc Oxide Doping on Mechanical and Biological Properties of 3D Printed
Tricalcium Phosphate Tissue Engineering Scaffolds,” Dent. Mater., 28 [2] 113–
22 (2012).
29
M. Castilho, B. Gouveia, I. Pires, J. Rodrigues, and M. Pereira, “The
Role of Shell/Core Saturation Level on the Accuracy and Mechanical Charac-
teristics of Porous Calcium Phosphate Models Produced by 3Dprinting,”
Rapid Prototyp. J., 21 [1] 43–56 (2015).
30
R. Chumnanklang, T. Panyathanmaporn, K. Sitthiseripratip, and J. Su-
wanprateeb, “3D Printing of Hydroxyapatite: Effect of Binder Concentration
in Pre-Coated Particle on Part Strength,” Mater. Sci. Eng., C, 27 [4] 914–21
(2007).
31
C. Bergmann, et al., “3D Printing of Bone Substitute Implants Using Cal-
cium Phosphate and Bioactive Glasses,” J. Eur. Ceram. Soc., 30 [12] 2563–7
(2010).
32
M. Castilho, et al., “Direct 3D Powder Printing of Biphasic Calcium Phos-
phate Scaffolds for Substitution of Complex Bone Defects,” Biofabrication, 6
(2014). doi: 10.1088/1758-5082/6/1/015006.
33
U. Klammert, T. Reuther, C. Jahn, B. Kraski, A. C. K€ ubler, and U. Gbu-
reck, “Cytocompatibility of Brushite and Monetite Cell Culture Scaffolds
Made by Three-Dimensional Powder Printing,” Acta Biomater., 5 [2] 727–34
(2009).
34
A. Butscher, M. Bohner, S. Hofmann, L. Gauckler, and R. M€ uller,
“Structural and Material Approaches to Bone Tissue Engineering in Powder-
Based Three-Dimensional Printing,” Acta Biomater., 7 [3] 907–20 (2011).
35
S. Yang, K.-F. Leong, Z. Du, and C.-K. Chua, “The Design of Scaffolds
for Use in Tissue Engineering. Part II. Rapid Prototyping Techniques,” Tissue
Eng., 8 [1] 1–11 (2002).
36
B. Leukers, et al., “Hydroxyapatite Scaffolds for Bone Tissue Engineering
Made by 3D Printing,” J. Mater. Sci. Mater. Med., 16 [12] 1121–4 (2005).
37
A. Spillmann, A. Sonnenfeld, and P. R. von Rohr, “Flowability Modifica-
tion of Lactose Powder by Plasma Enhanced Chemical Vapor Deposition,”
Plasma Process. Polym., 4 [S1] S16–20 (2007).
38
D. Huson and S. Hoskins, “3D Printed Ceramics for Tableware, Artists/
Designers and Specialist Applications,” Key Eng. Mater., 608, 351–7 (2014).
39
http://www.voxeljet.de/en/services/castings/, accessed on 28/02/2015,
(2015). Voxeljet, Augsburg, Germany.
40
J. Suwanprateeb, R. Sanngam, W. Suvannapruk, and T. Panyathanmap-
orn, “Mechanical and In Vitro Performance of Apatite–Wollastonite Glass
Ceramic Reinforced Hydroxyapatite Composite Fabricated by 3D-Printing,”
J. Mater. Sci. Mater. Med., 20 [6] 1281–9 (2009).
41
R. Melcher, S. Martins, N. Travitzky, and P. Greil, “Fabrication of Al2O3-
Based Composites by Indirect 3D-Printing,” Mater. Lett., 60 [4] 572–5 (2006).
42
W. Zhang, R. Melcher, N. Travitzky, R. K. Bordia, and P. Greil, “Three-
Dimensional Printing of Complex-Shaped Alumina/Glass Composites,” Adv.
Eng. Mater., 11 [12] 1039–43 (2009).
43
Z. Fu, L. Schlier, N. Travitzky, and P. Greil, “Three-Dimensional Printing
of SiSiC Lattice Truss Structures,” Mater. Sci. Eng., A, 560, 851–6 (2013).
44
J. Moon, A. C. Caballero, L. Hozer, Y. M. Chiang, and M. J. Cima,
“Fabrication of Functionally Graded Reaction Infiltrated SiC–Si Composite
by Three-Dimensional Printing (3DPTM) Process,” Mater. Sci. Eng., A, 298
[1–2] 110–9 (2001).
1999
45
J. Yoo, M. J. Cima, S. Khanuja, and E. M. Sachs, “Structural Ceramic
Components by 3D Printing”; pp. 40–50 in Solid Free. Fabr. Symp. Edited by
H. L. Marcus, J. J. Beaman, J. W. Barlow, D. L. Bourell, and R. H. Craw-
ford. The University of Texas, Austin, TX, 1993.
46
W. Sun, D. J. Dcosta, F. Lin, and T. El-Raghy, “Freeform Fabrication of
Ti3SiC2 Powder-Based Structures: Part I—Integrated Fabrication Process,”
J. Mater. Process. Technol., 127 [3] 343–51 (2002).
47
J.-P. Kruth, G. Levy, F. Klocke, and T. H. C. Childs, “Consolidation
Phenomena in Laser and Powder-Bed Based Layered Manufacturing,” CIRP
Ann.-Manuf. Technol., 56 [2] 730–59 (2007).
48
E. Juste, F. Petit, V. Lardot, and F. Cambier, “Shaping of Ceramic Parts
by Selective Laser Melting of Powder Bed,” J. Mater. Res., 29 [17] 2086–94
(2014).
49
P. Regenfuss, et al., “Principles of Laser Micro Sintering,” Rapid Prototyp.
J., 13 [4] 204–12 (2007).
50
K. Subramanian, N. Vail, J. Barlow, and H. Marcus, “Selective Laser Sin-
tering of Alumina with Polymer Binders,” Rapid Prototyp. J., 1 [2] 24–35
(1995).
51
J.-P. Kruth, P. Mercelis, J. Van Vaerenbergh, L. Froyen, and M. Romb-
outs, “Binding Mechanisms in Selective Laser Sintering and Selective Laser
Melting,” Rapid Prototyp. J., 11 [1] 26–36 (2005).
52
T. Friedel, N. Travitzky, F. Niebling, M. Scheffler, and P. Greil, “Fabrica-
tion of Polymer Derived Ceramic Parts by Selective Laser Curing,” J. Eur.
Ceram. Soc., 25 [2–3] 193–7 (2005).
53
K. C. R. Kolan, M. C. Leu, G. E. Hilmas, R. F. Brown, and M. Velez,
“Fabrication of 13-93 Bioactive Glass Scaffolds for Bone Tissue Engineering
Using Indirect Selective Laser Sintering,” Biofabrication, 3 (2011). doi:
10.1088/1758-5082/3/2/025004.
54
F.-H. Liu, Y.-K. Shen, and J.-L. Lee, “Selective Laser Sintering of a
Hydroxyapatite-Silica Scaffold on Cultured MG63 Osteoblasts In Vitro,” Int.
J. Precis. Eng. Manuf., 13 [3] 439–44 (2012).
55
R. L. Simpson, et al., “Development of a 95/5 Poly(L-Lactide-co-Glyco-
lide)/Hydroxylapatite and b-Tricalcium Phosphate Scaffold as Bone Replace-
ment Material Via Selective Laser Sintering,” J. Biomed. Mater. Res. B Appl.
Biomater., 84B [1] 17–25 (2008).
56
K. H. Tan, et al., “Scaffold Development Using Selective Laser Sintering
of Polyetheretherketone–Hydroxyapatite Biocomposite Blends,” Biomaterials,
24 [18] 3115–23 (2003).
57
W. L€ oschau, R. Lenk, S. Scharek, M. Teichgraber, S. Nowotny, and C.
Richter, “Prototyping of Complex-Shaped Parts and Tools of Si/SiC-Ceramics
by Selective Laser Sintering,” Ind. Ceram., 20 [2] 95–6 (2000).
58
J. Wilkes, Y.-C. Hagedorn, W. Meiners, and K. Wissenbach, “Additive
Manufacturing of ZrO2-Al2O3 Ceramic Components by Selective Laser Melt-
ing,” Rapid Prototyp. J., 19 [1] 51–7 (2013).
59
W. Zhou, D. Li, and H. Wang, “A Novel Aqueous Ceramic Suspension
for Ceramic Stereolithography,” Rapid Prototyp. J., 16 [1] 29–35 (2010).
60
J. Homa and M. Schwentenwein, “A Novel Additive Manufacturing Tech-
nology for High-Performance Ceramics”; pp. 33–40 in Advanced Processing
and Manufacturing Technologies for Nanostructured and Multifunctional Mate-
rials. Ceramic Engineering and Science Proceedings 35 [6]. Edited by T. Ohji,
M. Singh and S. Mathur. John Wiley & Sons, Inc., Hoboken, NJ, 2014.
61
T. Chartier, et al., “Fabrication of Millimeter Wave Components Via
Ceramic Stereo-and Microstereolithography Processes,” J. Am. Ceram. Soc.,
91 [8] 2469–74 (2008).
62
S. Kirihara, “Creation of Functional Ceramics Structures by Using Stere-
olitographic 3D Printing,” Trans. JWRI, 43 [1] 5–10 (2014).
63
T. M. G. Chu, D. G. Orton, S. J. Hollister, S. E. Feinberg, and J. W.
Halloran, “Mechanical and In Vivo Performance of Hydroxyapatite Implants
with Controlled Architectures,” Biomaterials, 23 [5] 1283–93 (2002).
64
T. Chartier, C. Chaput, F. Doreau, and M. Loiseau, “Stereolithography
of Structural Complex Ceramic Parts,” J. Mater. Sci., 37 [15] 3141–7
(2002).
65
W. Bian, et al., “Fabrication of a Bio-Inspired Beta-Tricalcium Phosphate/
Collagen Scaffold Based on Ceramic Stereolithography and Gel Casting for
Osteochondral Tissue Engineering,” Rapid Prototyp. J., 18 [1] 68–80 (2012).
66
T. A. Pham, D.-P. Kim, T.-W. Lim, S.-H. Park, D.-Y. Yang, and K.-S.
Lee, “Three-Dimensional SiCN Ceramic Microstructures Via Nano-Stereoli-
thography of Inorganic Polymer Photoresists,” Adv. Funct. Mater., 16 [9]
1235–41 (2006).
67
T. Chartier, E. Delhomme, J. F. Baumard, G. Veltl, and F. Ducloux,
“Injection Moulding of Hollow Silicon Nitride Parts Using Fusible Alloy
Cores,” Ceram. Int., 27 [7] 821–7 (2001).
68
G. H. Wroblewska, “Structural Ceramics with Complex Shape-Forming
Methods,” 25th Annu. Conf. Compos. Adv. Ceram. Mater. Struct. Ceram. Eng.
Sci. Proc., Wiley Online Library, 22 [3] 43–50 (2001).
69
E. M. Sachs, et al., “Jetting Layers of Powder and the Formation of Fine
Powder Beds Thereby”; US Patent 6,596,224; July 22; 2003.
70
H. C. Yen, “Experimental Studying on Development of Slurry-Layer Cast-
ing System for Additive Manufacturing of Ceramics,” Int. J. Adv. Manuf.
Technol., 77 [5–8] 915–25 (2014).
71
H.-H. Tang and H.-C. Yen, “Slurry-Based Additive Manufacturing of
Ceramic Parts by Selective Laser Burn-Out,” J. Eur. Ceram. Soc., 35 [3] 981–7
(2015).
72
T. M€ uhler, C. M. Gomes, J. Heinrich, and J. G€ unster, “Slurry-Based
Additive Manufacturing of Ceramics,” Int. J. Appl. Ceram. Technol., 12 [1]
18–25 (2015).
73
X. Tian, T. Muhler, C. Gomes, J. Gunster, and J. G. Heinrich, “Feasibil-
ity Study on Rapid Prototyping of Porcelain Products,” J. Ceram. Sci. Tech-
nol., 2 [4] 217–25 (2011).
2000 Journal of the American Ceramic Society—Zocca et al.
74
X. Cui, S. Ouyang, Z. Yu, C. Wang, and Y. Huang, “A Study on Green
Tapes for LOM with Water-Based Tape Casting Processing,” Mater. Lett., 57
[7] 1300–4 (2003).
75
C. M. Gomes, A. P. N. Oliveira, D. Hotza, N. Travitzky, and P. Greil,
“LZSA Glass-Ceramic Laminates: Fabrication and Mechanical Properties,” J.
Mater. Process. Technol., 206 [1–3] 194–201 (2008).
76
J. Kechagias, “An Experimental Investigation of the Surface Roughness of
Parts Produced by LOM Process,” Rapid Prototyp. J., 13 [1] 17–22 (2007).
77
D. Klosterman, R. Chartoff, N. Osborne, and G. Graves, “Automated
Fabrication of Monolithic and Ceramic Matrix Composites Via Laminated
Object Manufacturing (LOM)”; pp. 537–49 in Solid Free. Fabr. Symp. Proc.,
Edited by D. L. Bourell, J. J. Beaman, H. L. Marcus, R. H. Crawford, and J.
W Barlow. The University of Texas, Austin, TX, 1997.
78
D. Klosterman, R. Chartoff, G. Graves, N. Osborne, and B. Priore,
“Interfacial Characteristics of Composites Fabricated by Laminated Object
Manufacturing,” Compos. Part Appl. Sci. Manuf., 29 [9–10] 1165–74 (1998).
79
S. J. Rodrigues, R. P. Chartoff, D. A. Klosterman, M. Agarwala, and N.
Hecht, “Solid Freeform Fabrication of Functional Silicon Nitride Ceramics by
Laminated Object Manufacturing,”; Proc. Solid Free. Fabr. Symp., Edited by
D. L. Bourell, J. J. Beaman, R. H. Crawford, H. L. Marcus, and J. W Barlow.
The University of Texas, Austin, TX, 2000.
80
L. Weisensel, N. Travitzky, H. Sieber, and P. Greil, “Laminated Object
Manufacturing (LOM) of SiSiC Composites,” Adv. Eng. Mater., 6 [11] 899–
903 (2004).
81
H. Windsheimer, N. Travitzky, A. Hofenauer, and P. Greil, “Laminated
Object Manufacturing of Preceramic-Paper-Derived SiSiC Composites,” Adv.
Mater., 19 [24] 4515–9 (2007).
82
Y. Zhang, X. He, J. Han, and S. Du, “Ceramic Green Tape Extrusion for
Laminated Object Manufacturing,” Mater. Lett., 40 [6] 275–9 (1999).
83
L. Giassi, D. Hotza, O. E. Alarcon, M. C. Fredel, and A. P. Novaes de
Oliveira, “Sintering and Crystallization of LZSA Glass Powder Compacts
Formed by Injection Moulding,” Am. Ceram. Soc. Bull., 84 [6] 9301–6
(2005).
84
F. M. Bertan, O. R. K. Montedo, C. R. Rambo, D. Hotza, and A. P. N.
de Oliveira, “Extruded ZrSiO4 Particulate-Reinforced LZSA Glass–Ceramics
Matrix Composite,” J. Mater. Process. Technol., 209 [3] 1134–42 (2009).
85
G. M. Reitz, O. R. K. Montedo, O. E. Alarcon, D. Hotza, and A. P.
Novaes de Oliveira, “Roll Pressed LZSA Glass-Ceramics,” Adv. Sci. Technol.,
45, 442–6 (2006).
86
C. Gomes, et al., “Laminated Object Manufacturing of LZSA Glass-
Ceramics,” Rapid Prototyp. J., 17 [6] 424–8 (2011).
87
S. Liu, F. Ye, L. Liu, and Q. Liu, “Feasibility of Preparing of Silicon
Nitride Ceramics Components by Aqueous Tape Casting in Combination
with Laminated Object Manufacturing,” Mater. Des., 66, Part A, 331–5
(2015).
88
O. Penas, R. Zenati, J. Dubois, and G. Fantozzi, “Processing, Microstruc-
ture, Mechanical Properties of Si3N4 Obtained by Slip Casting and Pressure-
less Sintering,” Ceram. Int., 27 [5] 591–6 (2001).
89
G. Ling and H. Yang, “Pressureless Sintering of Silicon Nitride with Mag-
nesia and Yttria,” Mater. Chem. Phys., 90 [1] 31–4 (2005).
90
F. Ye, L. Liu, J. Zhang, M. Iwasa, and C.-L. Su, “Synthesis of Silicon
Nitride-Barium Aluminosilicate Self-Reinforced Ceramic Composite by a
Two-Step Pressureless Sintering,” Compos. Sci. Technol., 65 [14] 2233–9
(2005).
91
S. Wang, D. Jia, Z. Yang, X. Duan, Z. Tian, and Y. Zhou, “Effect of BN
Content on Microstructures, Mechanical and Dielectric Properties of Porous
BN/Si3N4 Composite Ceramics Prepared by Gel Casting,” Ceram. Int., 39 [4]
4231–7 (2013).
92
J. Yu, H. Wang, and J. Zhang, “Neural Network Modeling and Analysis
of Gel Casting Preparation of Porous Si3N4 Ceramics,” Ceram. Int., 35 [7]
2943–50 (2009).
93 €
E. Ozkol, J. Ebert, and R. Telle, “An Experimental Analysis of the Influ-
ence of the Ink Properties on the Drop Formation for Direct Thermal Inkjet
Printing of High Solid Content Aqueous 3Y-TZP Suspensions,” J. Eur.
Ceram. Soc., 30 [7] 1669–78 (2010).
94
W. D. Teng and M. J. Edirisinghe, “Development of Ceramic Inks for
Direct Continuous Jet Printing,” J. Am. Ceram. Soc., 81 [4] 1033–6 (1998).
95
B. Derby, “Inkjet Printing Ceramics: From Drops to Solid,” J. Eur.
Ceram. Soc., 31 [14] 2543–50 (2011).
96
B. Derby, “Inkjet Printing of Functional and Structural Materials: Fluid
Property Requirements, Feature Stability, and Resolution,” Annu. Rev. Mater.
Res., 40 [1] 395–414 (2010).
97
B. Cappi, J. Ebert, and R. Telle, “Rheological Properties of Aqueous
Si3N4 and MoSi2 Suspensions Tailor-Made for Direct Inkjet Printing,” J. Am.
Ceram. Soc., 94 [1] 111–6 (2011).
98
A. M. W€ atjen, P. Gingter, M. Kramer, and R. Telle, “Novel Prospects
and Possibilities in Additive Manufacturing of Ceramics by Means of Direct
Inkjet Printing,” Adv. Mech. Eng., 2014, e141346 (2014).
99 €
E. Ozkol, W. Zhang, J. Ebert, and R. Telle, “Potentials of the ‘Direct Ink-
jet Printing’ Method for Manufacturing 3Y-TZP Based Dental Restorations,”
J. Eur. Ceram. Soc., 32 [10] 2193–201 (2012).
100
R. I. Tomov, et al., “Direct Ceramic Inkjet Printing of Yttria-Stabilized
Zirconia Electrolyte Layers for Anode-Supported Solid Oxide Fuel Cells,” J.
Power Sources, 195 [21] 7160–7 (2010).
101 €
B. Cappi, E. Ozkol, J. Ebert, and R. Telle, “Direct Inkjet Printing of
Si3N4: Characterization of Ink, Green Bodies and Microstructure,” J. Eur.
Ceram. Soc., 28 [13] 2625–8 (2008).
102
http://www.ceradrop.fr/en/, accessed on 28/02/2015, (2015). Ceradrop,
Limoges, France.
Vol. 98, No. 7
103
M. Mott and J. R. G. Evans, “Solid Freeforming of Silicon Carbide by
Inkjet Printing Using a Polymeric Precursor,” J. Am. Ceram. Soc., 84 [2] 307–
13 (2001).
104
J. Cesarano, P. D. Calvert, and Inventor: Sandia Corporation, Assignee,
“Freeforming Objects with Low-Binder Slurry”; US Patent 6027326 A; Febru-
ary 22; US6027326 A, 2000.
105
J. A. Lewis, J. E. Smay, J. Stuecker, and J. Cesarano, “Direct Ink Writ-
ing of Three-Dimensional Ceramic Structures,” J. Am. Ceram. Soc., 89 [12]
3599–609 (2006).
106
Q. Fu, E. Saiz, and A. P. Tomsia, “Direct Ink Writing of Highly Porous
and Strong Glass Scaffolds for Load-Bearing Bone Defects Repair and Regen-
eration,” Acta Biomater., 7 [10] 3547–54 (2011).
107
S. Eqtesadi, A. Motealleh, P. Miranda, A. Pajares, A. Lemos, and J. M.
F. Ferreira, “Robocasting of 45S5 Bioactive Glass Scaffolds for Bone Tissue
Engineering,” J. Eur. Ceram. Soc., 34 [1] 107–18 (2014).
108
F. C. G. de Sousa and J. R. G. Evans, “Sintered Hydroxyapatite Lattice-
work for Bone Substitute,” J. Am. Ceram. Soc., 86 [3] 517–9 (2003).
109
J. N. Stuecker, J. Cesarano III, and D. A. Hirschfeld, “Control of the
Viscous Behavior of Highly Concentrated Mullite Suspensions for Robocast-
ing,” J. Mater. Process. Technol., 142 [2] 318–25 (2003).
110
J. E. Smay, J. Cesarano, and J. A. Lewis, “Colloidal Inks for Directed
Assembly of 3-D Periodic Structures,” Langmuir, 18 [14] 5429–37 (2002).
111
J. E. Smay, J. Cesarano, B. A. Tuttle, and J. A. Lewis, “Directed Colloi-
dal Assembly of Linear and Annular Lead Zirconate Titanate Arrays,” J. Am.
Ceram. Soc., 87 [2] 293–5 (2004).
112
J. E. Smay, S. S. Nadkarni, and J. Xu, “Direct Writing of Dielectric
Ceramics and Base Metal Electrodes,” Int. J. Appl. Ceram. Technol., 4 [1] 47–
52 (2007).
113
M. Genet, M. Houmard, S. Eslava, E. Saiz, and A. P. Tomsia, “A Two-
Scale Weibull Approach to the Failure of Porous Ceramic Structures Made by
Robocasting: Possibilities and Limits,” J. Eur. Ceram. Soc., 33 [4] 679–88
(2013).
114
J. Franco, P. Hunger, M. E. Launey, A. P. Tomsia, and E. Saiz, “Direct-
Write Assembly of Calcium Phosphate Scaffolds Using a Water-Based Hydro-
gel,” Acta Biomater., 6 [1] 218–28 (2010).
115
P. Miranda, A. Pajares, E. Saiz, A. P. Tomsia, and F. Guiberteau,
“Fracture Modes Under Uniaxial Compression in Hydroxyapatite Scaffolds
Fabricated by Robocasting,” J. Biomed. Mater. Res. A, 83A [3] 646–55
(2007).
116
P. Miranda, A. Pajares, E. Saiz, A. P. Tomsia, and F. Guiberteau,
“Mechanical Properties of Calcium Phosphate Scaffolds Fabricated by Robo-
casting,” J. Biomed. Mater. Res. A, 85A [1] 218–27 (2008).
117
Q. Fu, E. Saiz, and A. P. Tomsia, “Bioinspired Strong and Highly Por-
ous Glass Scaffolds,” Adv. Funct. Mater., 21 [6] 1058–63 (2011).
118
L. J. Gibson and M. F. Ashby, “Cellular Solids: Structure and Proper-
ties”. Cambridge University Press, Cambridge, 1997.
119
J. G. Dellinger, J. Cesarano, and R. D. Jamison, “Robotic Deposition of
Model Hydroxyapatite Scaffolds with Multiple Architectures and Multiscale
Porosity for Bone Tissue Engineering,” J. Biomed. Mater. Res. A, 82A [2]
383–94 (2007).
120
J. Cesarano, et al., “Customization of Load-Bearing Hydroxyapatite Lat-
tice Scaffolds,” Int. J. Appl. Ceram. Technol., 2 [3] 212–20 (2005).
121
J. A. Lewis, “Direct Ink Writing of 3D Functional Materials,” Adv.
Funct. Mater., 16 [17] 2193–204 (2006).
122
T. Schlordt, S. Schwanke, F. Keppner, T. Fey, N. Travitzky, and P.
Greil, “Robocasting of Alumina Hollow Filament Lattice Structures,” J. Eur.
Ceram. Soc., 33 [15–16] 3243–8 (2013).
123
I. Grida and J. R. Evans, “Extrusion Freeforming of Ceramics Through
Fine Nozzles,” J. Eur. Ceram. Soc., 23 [5] 629–35 (2003).
124
S. A. Park, S. H. Lee, and W. D. Kim, “Fabrication of Porous Polycap-
rolactone/Hydroxyapatite (PCL/HA) Blend Scaffolds Using a 3D Plotting Sys-
tem for Bone Tissue Engineering,” Bioprocess Biosyst. Eng., 34 [4] 505–13
(2011).
125
S. J. Kalita, S. Bose, H. L. Hosick, and A. Bandyopadhyay, “Develop-
ment of Controlled Porosity Polymer-Ceramic Composite Scaffolds Via Fused
Deposition Modeling,” Mater. Sci. Eng., C, 23 [5] 611–20 (2003).
126
A. Bandyopadhyay, R. K. Panda, V. F. Janas, M. K. Agarwala, S. C.
Danforth, and A. Safari, “Processing of Piezocomposites by Fused Deposition
Technique,” J. Am. Ceram. Soc., 80 [6] 1366–72 (1997).
127
M. A. Jafari, W. Han, F. Mohammadi, A. Safari, S. C. Danforth, and N.
Langrana, “A Novel System for Fused Deposition of Advanced Multiple
Ceramics,” Rapid Prototyp. J., 6 [3] 161–75 (2000).
128
D.-M. Liu, “Influence of Porosity and Pore Size on the Compressive
Strength of Porous Hydroxyapatite Ceramic,” Ceram. Int., 23 [2] 135–9
(1997).
129
B. Leukers, et al., “Biocompatibility of Ceramic Scaffolds for Bone
Replacement Made by 3D Printing,” Mater. Werkst, 36 [12] 781–7 (2005).
130
H. Yang, S. Yang, X. Chi, and J. R. G. Evans, “Fine Ceramic Lattices
Prepared by Extrusion Freeforming,” J. Biomed. Mater. Res. B Appl. Bioma-
ter., 79B [1] 116–21 (2006).
131
X. Lu, et al., “Fabrication of Millimeter-Wave Electromagnetic Bandgap
Crystals Using Microwave Dielectric Powders,” J. Am. Ceram. Soc., 92 [2]
371–8 (2009).
132
M. Chhabra and R. Singh, “Rapid Casting Solutions: A Review,” Rapid
Prototyp. J., 17 [5] 328–50 (2011).
133
R. Detsch, F. Uhl, U. Deisinger, and G. Ziegler, “3D-Cultivation of Bone
Marrow Stromal Cells on Hydroxyapatite Scaffolds Fabricated by Dispense-
Plotting and Negative Mould Technique,” J. Mater. Sci. Mater. Med., 19 [4]
1491–6 (2008).
July 2015 Additive Manufacturing of Ceramics 2001
134 141
H. Yin, S. Kirihara, and Y. Miyamoto, “Fabrication of Ceramic Pho- J. G€unster, S. Engler, and J. G. Heinrich, “Forming of Complex Shaped Cera-
tonic Crystals with Diamond Structure for Microwave Applications,” J. Am. mic Products Via Layer-Wise Slurry Deposition,” Bull. ECERS, 1, 25–8 (2003).
142
Ceram. Soc., 87 [4] 598–601 (2004). http://www.schunk-group.com/en/applications/show/Category.1538/
135
A. Woesz, et al., “Towards Bone Replacement Materials from Calcium action/feList/controller/Product/, accessed on 27/02/2015, (2015). Schunk Inge-
Phosphates Via Rapid Prototyping and Ceramic Gelcasting,” Mater. Sci. Eng., nieurkeramik GmbH, Lauffer/Neckar, Germany.
143
C, 25 [2] 181–6 (2005). W. E. Frazier, “Metal Additive Manufacturing: A Review,” J. Mater.
136
S. Bose, J. Darsell, M. Kintner, H. Hosick, and A. Bandyopadhyay, Eng. Perform., 23 [6] 1917–28 (2014).
144
“Pore Size and Pore Volume Effects on Alumina and TCP Ceramic Scaffolds,” S. Chianrabutra, B. G. Mellor, and S. Yang, “A Dry Powder Material
Mater. Sci. Eng., C, 23 [4] 479–86 (2003). Delivery Device for Multiple Material Additive Manufacturing”; Solid Free-
137
D. Guo, L. Li, K. Cai, Z. Gui, and C. Nan, “Rapid Prototyping of Pie- form Fabrication Symposium Proceedings The University of Texas at Austin,
zoelectric Ceramics Via Selective Laser Sintering and Gelcasting,” J. Am. Austin, Texas, 2014.
145
Ceram. Soc., 87 [1] 17–22 (2004). S. Yang and J. R. G. Evans, “A Multi-Component Powder Dispensing
138
A. Ortona, C. D’Angelo, S. Gianella, and D. Gaia, “Cellular Ceramics System for Three Dimensional Functional Gradients,” Mater. Sci. Eng., A,
Produced by Rapid Prototyping and Replication,” Mater. Lett., 80, 95–8 379 [1–2] 351–9 (2004).
146
(2012). T. Schlordtil and P. Greil, “Robocasting of Alumina Lattice Truss Struc-
139
G. Franchin and P. Colombo, “Porous Geopolymer Components tures,” J. Ceram. Sci. Technol., 3 [2] 81–8 (2012).
147
Through Inverse Replica of 3D Printed Sacrificial Templates,” J. Ceram. Sci. M. Lejeune, T. Chartier, C. Dossou-Yovo, and R. Noguera, “Ink-Jet Print-
Technol., 6 [2] 105–12 (2015). ing of Ceramic Micro-Pillar Arrays,” J. Eur. Ceram. Soc., 29 [5] 905–11 (2009).
140
A. Zocca, C. M. Gomes, T. M€ uhler, and J. G€
unster, “Powder-Bed Stabil- 148 €
E. Ozkol, J. Ebert, K. Uibel, A. M. W€atjen, and R. Telle, “Development of
ization for Powder-Based Additive Manufacturing,” Adv. Mech. Eng., 6, High Solid Content Aqueous 3Y-TZP Suspensions for Direct Inkjet Printing
(2014) doi: 10.1155/2014/491581. Using a Thermal Inkjet Printer,” J. Eur. Ceram. Soc., 29 [3] 403–9 (2009). h

Andrea Zocca graduated from the waste. He is a member of the Engineering Ceramics Division
University of Padova, where he since 1992.
obtained his Bachelor’s and Master’s
degrees in 2009 and 2011, respec-
tively. Since then, he worked at
BAM, Federal institute for materials Cynthia M. Gomes is research scien-
research and testing in Berlin, and at tist at the BAM, Federal Institute for
the University of Padova in the Materials Research and Testing in
frame of a co-tutored doctoral degree Berlin, Germany. She received her
between the University of Padova Diploma in Materials Science and
and Clausthal University of Technology. After his expected Engineering from the Federal Univer-
defence in June 2015, he will obtain an Adolf Martens post- sity of Paraiba, Brazil, with M.Sc.
doctoral fellowship at BAM. His main research interests are and Ph.D. in Materials Science and
in the field of additive manufacturing of ceramic materials, Engineering both from the Federal
particularly technical ceramics, and the application of addi- University Santa Catarina, Brazil.
tive manufacturing to preceramic polymers. After working as guest visitor scientist and a two-year post
doctorate at the University of Erlangen-Nuremberg she has
Paolo Colombo is a professor at the joined the BAM. At the Division for Ceramic Processing and
University of Padova. He graduated Biomaterials she has been working mainly in the field of
from the University of Padova with additive manufacturing of ceramic materials, co-working also
a degree in chemical engineering in on national and international (DIN, ISO) groups for stan-
1985. Previously, he was an assistant dardization of these technologies.
professor at the University of Padova
and an associate professor at the
University of Bologna. He is also an Jens G€ unster is head of division
adjunct professor in the Department “Ceramic Processing and Biomater-
of Materials Science and Engineer- ials” at the Federal Institute for Mate-
ing, Pennsylvania State University, a rials Research and Testing (BAM) in
visiting professor in the Department Berlin and a professor at Clausthal
of Mechanical Engineering of University College London, a University of Technology. He grad-
member of the World Academy of Ceramics, a member of uated from Clausthal University of
the EPSRC Peer Review College and has been elected to Technology with a degree in physics
Fellow of the American Ceramic Society and of the Institute 1995. He has worked as an Alexander
of Materials, Minerals and Mining. In 2015, he is a Foreign von Humboldt Fellow with Ryutaro
Scientist at INSA, Lyon, and in 2016 he will be a DGF Souda at NIMS in Tsukuba, Japan, and as a postdoctoral fel-
Mercator professor at the Technical University Freiberg, low with Wayne D. Goodman at Texas A&M, USA. As a
Germany. He was awarded a Fulbright Scholarship for the research assistant of J€
urgen G. Heinrich he received his Venia
Pennsylvania State University in 1991, the 2007 Pfeil Award, Legendi in Materials Physics from Clausthal University of
the 2010 Global Star Award, the 2011 Edward C. Henry Technology in 2002 and joined 2003 the International Center
Award and the 2013 Verulam Medal & Prize. He is in the for Young Scientists ICYS at NIMS as visiting member. From
editorial board of 11 international scientific journals, and an 2008 to 2010 he worked as laser and glass specialist for Oerli-
Associate Editor of JACS. His current research interests kon Solar in Switzerland. He is board member of the German
include novel processing routes to porous glasses and ceram- Ceramic Society (DKG) and Associate Editor of the Journal
ics (including 3D printing), the development of ceramic com- of Ceramic Science and Technology. His current research
ponents from preceramic polymers and geopolymers, and the interests include functional materials, laser processing and
vitrification and reuse of hazardous industrial and natural additive manufacturing of ceramics.