Potential of Extrusion Based 3D-printed Hardmetal and Cermet Parts

Abstract

Hardmetal and cermet bodies were printed by fused-filament fabrication (FFF) and composite-extrusion modelling (CEM) in an SDS (shaping-debinding-sintering) process. For FFF the filaments were prepared from hardmetal (WC-10Co) and cermet powder (Ti(C,N)-Co/Ni-based) and organic binder. The CEM feedstock consisted of WC-Co MIM powder. A 3D filament printer as well as a 3D printer working with a granulate such as used in MIM were employed to fabricate printed bodies by FFF and CEM, respectively. The solvent debinding process was performed in cyclohexane (FFF-printed bodies) or water (CEM-printed bodies). Thermal debinding of all parts was performed in a tube furnace up to a temperature of 800℃. The pre-sintered parts were then subjected to vacuum sintering by application of conventional vacuum sintering profiles up to 1430℃ for hardmetals and up to 1460℃ for cermets. Dimensional and mass changes upon the various preparation steps as well as microstructure and porosity of the sintered bodies were investigated. While the microstructure is practically identical to that of conventionally prepared materials, some cavities are present from the printing process because of yet non-optimised printing strategy. The study shows that with the applied 3D printing techniques, hardmetal and cermet parts with innovative geometries are accessible.

Full text

·938· 2018 World Congress on Powder Metallurgy
Potential of Extrusion Based 3D-printed
Hardmetal and Cermet Parts
Kitzmantel Michael1,2*, Lengauer Walter2*, Duretek Ivica3, Schwarz Viktoria2, Kukla
Christian4, Lieberwirth Clemens5, Morrison Vincent5, Wilfinger Thomas1, Neubauer Erich1 
(1. RHP-Technology GmbH, A-2444 Seibersdorf; 2. Vienna University of Technology, A-1060 Vienna;
3. Montanuniversität Leoben, Dept. of Polymer Engineering Science, A-8700 Leoben; 4. Montanuniversität
Leoben, Industrial Liaison Dept., A-8700 Leoben; 5. AIM3D GmbH, D-18069 Rostock)
Abstract: Hardmetal and cermet bodies were printed by fused-filament fabrication (FFF) and composite- extrusion
modelling (CEM) in an SDS (shaping – debinding – sintering) process. For FFF the filaments were prepared from
hardmetal (WC-10Co) and cermet powder (Ti(C,N)-Co/Ni-based) and organic binder. The CEM feedstock consisted of
WC-Co MIM powder. A 3D filament printer as well as a 3D printer working with a granulate such as used in MIM were
employed to fabricate printed bodies by FFF and CEM, respectively. The solvent debinding process was performed in
cyclohexane (FFF-printed bodies) or water (CEM-printed bodies). Thermal debinding of all parts was performed in a tube
furnace up to a temperature of 800℃. The pre-sintered parts were then subjected to vacuum sintering by application of
conventional vacuum sintering profiles up to 1430℃ for hardmetals and up to 1460℃ for cermets. Dimensional and mass
changes upon the various preparation steps as well as microstructure and porosity of the sintered bodies were investigated.
While the microstructure is practically identical to that of conventionally prepared materials, some cavities are present from
the printing process because of yet non-optimised printing strategy. The study shows that with the applied 3D printing
techniques, hardmetal and cermet parts with innovative geometries are accessible.
1 Introduction
Additive manufacturing (AM) is a fast-developing technology in the field of powder metallurgy. Very
complicated part geometries in combination with reduction in raw-powder consumption are the main benefits.
A low cost and very popular AM technique is the extrusion-based printing technology such as fused-filament
printing. Since more than twenty years [1-3] feedstocks containing metal powders and various types of ceramic
powders are being developed and plenty of studies appeared in the meantime. Contrary to the metal and ceramic
powders studied so far, hardmetal and cermet formulations are mixtures of many different powders of grain sizes
often smaller than 1 µm.
For 3D manufacturing of hardmetal and cermet bodies extrusion-based techniques are very interesting because
of several advantages over other AM fabrication routes:
(1) There is practically no raw material loss of extrusion-based methods. This is a benefit in view of the
expensive hardmetal and cermet powder mixtures.
(2) Hardmetal and cermet powder mixtures contain many different powders such as hard-phase powders (WC,
Ti(C,N), VC, Cr3C2…) and binder powders (Co, Fe, Ni) which can separate if unbound powders incur in the
shaping process (e.g. compare to powder-bed laser sintering). Such powders are difficult to recycle for re-use in
the printing process.
(3) Generally, hardmetal and cermets must have a uniform microstructure. Because of the presence of volatile
Co, an SDS method (shaping - debinding - sintering) is a first choice. The whole part is sintered and no local
*Corresponding authors: michael.kitzmantel@tuwien.ac.at; walter.lengauer@tuwien.ac.at.

Part 5—Refractory Metals, Hard Alloys, Superhard Materials & Ceramics ·939·
overheating or unequal temperature distribution occurs which would result in binder loss and inhomogeneities in
composition and grain size.
(4) Extrusion-based techniques are the best-suited AM techniques for printing materials’ combinations.
For the preparation of hardmetal and cermets we have studied here two extrusion-based methods, (1) the
fused-filament fabrication (FFF) for which a feedstock filament was prepared and fed into a 3D printer and
(2) composite-extrusion modelling (CEM) which can use MIM feedstock without need to prepare a filament. In
both techniques the feedstock is fused and free-form printed through a nozzle. Our work seems to be the first
study of FFF and CEM printing of hardmetals and cermets.
2 Experimental
2.1 Filament preparation
2.1.1 Filament feedstock
For feedstock preparation, two powder batches, one hardmetal (WC-10%Co) and one cermet (Ti(C,N), WC,
(Ta,Nb)C, Cr3C2, Co, Ni) batch (each approx. 1 kg) was prepared by ball mixing the starting powders in a 3 L
polyurethane drum by use of hardmetal balls under cyclohexane for 24 h with an appropriate rotation speed to
break agglomerates. After this procedure, the mixtures were dried in a vacuum oven.
The binder system for printing these inorganic powders had two components [4,5]: the main binder component
and the backbone polymer. The main binder component was a soft and flexible thermoplastic elastomer – TPE
(Kraiburg TPE GmbH & Co. KG, Germany). For the backbone, PP functionalized with maleic anhydride (BYK
Chemie GmbH, Germany) was used. Some information of the prepared feedstocks is shown in Table 1.
Table 1 Description of FFF feedstock materials
Feedstock Main binder Backbone Backbone content in binder (vol%) Powder content(vol%)
WC-Co
Cermet
TPE PP 40 50
Feedstocks were compounded in a kneader with counter rotating rollers (Plasti-Corder PL 2000, Brabender
GmbH & Co. KG, Germany) at a rotational speed of 60 rpm. All the feedstock components were premixed in solid
state. The kneader temperature was 200℃ and the total kneading time was 50 min. After compounding, the
feedstocks were ground in a cutting mill (SM200, Retsch GmbH, Germany) with a sieve with square perforations
of 4 mm in length.
2.1.2 Filament preparation
Filaments were prepared using a high-pressure capillary rheometer (Rheograph 2002, Göttfert Werkstoff-
Prüfmaschinen GmbH, Germany). A round capillary with a diameter of 1.75 mm and length of 30 mm was used.
Extrusion temperature was set at 200℃ and the piston speed was 1 mm.s–1. The apparent shear rate was 215 s–1.
At the exit of the die, a Teflon conveyor belt was placed to pull the filament as it was extruded.
Fig.1 Scheme of the FFF technique
Fig.2 CAD drawing of printed indexable inserts
2.1.3 Fused-filament fabrication (FFF)
Printing trials were performed on a Wanhao Duplicator i3V2 FFF with a nozzle diameter of 0.6 mm. A scheme

·940· 2018 World Congress on Powder Metallurgy
of the FFF technique is shown in Fig.1. The printed parts were indexable cutting inserts with various geometries.
The software KISSlicer was used to prepare the G-code for printing. Fig.2 shows a CAD drawing of parts printed
by FFF.
In order to observe the powder wetting the filaments were investigated by means of scanning electron
microscopy (SEM). The filaments were cryogenically fractured under liquid nitrogen.
2.2 Composite-extrusion modelling (CEM)
Another series of samples was fabricated by CEM-printing green bodies [6,7] from a hardmetal granulate used
for MIM (PolyMIM, Bad Sobernheim, Germany). An ExAM 255 printer (AIM3D GmbH) with an extruder heated
to 190℃ was used. A scheme of this method is shown in Fig.3. After CAD modelling the program of Simplify 3D
Inc. (Blue Ash, USA) was used to create printer and extruder commands. Printing
was performed with a 0.3 mm nozzle and a layer height of 0.1 mm. The bed
temperature was 100℃. Printing time was between 15 and 20 min, depending on
the part size.
2.3 Debinding processes
2.3.1 Solvent debinding
FFF-printed parts of hardmetals and cermets were debound in cyclohexane at
60℃ for 48 h. All parts were dried at 70℃ in a vacuum oven.
CEM-printed green hardmetal bodies from PolyMIM feedstock were water
debound by use of a mixture of H2O with 2% inhibitor (Metamax I-15, Zschimmer
& Schwarz, Germany) at 60℃ for 48 h.
2.3.2 Thermal debinding – pre-sintering
All parts were thermally debound in a tube furnace under flowing H2 with a temperature ramp of 10 h to a
maximum temperature of 800℃.
2.4 Sintering and characterisation
After thermal debinding the bodies were subjected to vacuum sintering. Hardmetals were sintered within a
simple time-temperature profile with a dwell at 1150℃ and at 1430℃. A more complicated temperature profile
was applied to cermets with three temperature dwells and a maximum sintering temperature of 1460℃[8].
The parts were cut, diamond polished and inspected by light-optical (LOM) and scanning electron (SEM)
microscopy.
3 Results and Discussion
3.1 Filament quality
A constant geometry of the filament is important for its continuous transportation in an FFF machine. For this
reason, the diameter and the ovality of the produced filament were monitored during production.
The ovality distribution was more or less constant with an average of ±0.030 mm, which is sufficiently low to
consider the filament to be round.Fig.4 shows SEM images where the good compatibility between the particles
and the binder can be seen, since there are no gaps between the binder and the individual particles. One can see
that the powder particles are equally distributed in the filament, which is important to have an even distribution of
particles in the printed and sintered parts.
3.2 FFF-printed parts
In all printed parts the first layer had a good contact with the printing bed (Fig.5). The layer in contact with the
printing bed had a more or less smooth and continuous surface (Fig.6). The printed layers were bonded correctly
Fig.3 Scheme of the CEM
technique, 1-pellet container,
2-extruder, 3- printed part,
x-y-movable x-y table,
z-movable z axis

Part 5—Refractory Metals, Hard Alloys, Superhard Materials & Ceramics ·941·
(Fig.7). Some defects can also be observed as a result of errors occurred during printing, for example small voids
between the layers.
Fig.4 Overall view and detail of the filament filled with powder particles (left: hardmetal, right: cermet)
Fig.5 FFF printed parts on the printing bed
Fig.6 Bottom surface of FFF printed parts in contact with the bed
Fig.7 Layers of FFF printed parts
3.3 Comparison of printed and sintered parts
Fig.8 shows a comparison between green and sintered FFF- and CEM-printed parts of hardmetals and cermets
with a geometry similar to that of indexable inserts applied in metal cutting processes (turning and milling).
Fig.8 Printed and sintered bodies. Top row: FFF-printed (left) and sintered (right) hardmetal, middle row:
FFF-printed (left) and sintered (right) cermet, bottom row: CEM-printed (left) and sintered (right) hardmetal

·942· 2018 World Congress on Powder Metallurgy
The printing lines which are clearly visible in the green bodies (Fig.8, left column), prevail in the sintered parts
(Fig.8, right column). This is also true for fine details and printing errors.
3.4 Mass change and geometries
Table 2 gives details on mass change upon the various processing steps. After solvent debinding only the
backbone remains in the body which serves for mechanical stability upon handling. Because of the substantially
different density of hardmetals (ca. 15 gcm-3) and cermets (ca. 6.5 gcm-3) the mass loss of the latter is much
higher.
Table 2 Mass change after the various steps relative to the mass of printed body
Method Part ∆m, %solvent debinding ∆m, %thermal debinding ∆m, %total, after sintering
FFF Hardmetal -3.55 -2.77 -6.31
FFF Cermet -7.08 -6.71 -14.62
CEM Hardmetal -3.40 -4.26 -7.82
Table 3 gives details on the dimensional changes. These have been determined by measuring the parts parallel
and perpendicular to the printing plane.
Table 3 Dimensional changes before (3D-printed body) and after sintering. Theoretical shrinkage of FFF parts from
powder/binder ratio, for CEM parts from PolyMIM data sheet
Method Part Shrinkage x-y% Shrinkage z% Theoretical%
FFF Hardmetal (coarse WC powder) 21.0 22.0 20.6
FFF Cermet 21.6 22.3 20.6
CEM Hardmetal (fine WC powder) 22.6 23.4 23.1
For all parts the shrinkage in z direction is larger than in x-y direction. This is a consequence of denser packing
density in x-y direction. Obviously, the printed lines within one layer are fused closer together than the one layer
to the next. This could be due to solidification of the x-y printed layer before the next one is printed.
3.5 Microstructure and properties of sintered bodies
Fig.9 shows examples of the microstructure of an FFF-printed coarse-grained hardmetal. In Fig.9(a)and (b)
cavities are visible which are aligned in the z direction (vertical). These cavities are due to some misalignment in
x-y positioning upon printing so that a wedge-type gaps exist between the fused deposited filament lines within
one layer. Similar observations of occurring cavities were reported on FFF-printed 361L steel parts [9]. The SEM
image in Fig.9(c) shows the typical microstructure of hardmetals.
Fig.9 Cross sections (LOM) and microstructures (SEM) of FFF-printed coarse-grained WC-Co hardmetal. (a) one half of the
sample shown in Fig.8, top (arrow: inner wall); (b) magnification of a (Murakami etched), (c)SEM microstructure
(a) (b) (c)

Part 5—Refractory Metals, Hard Alloys, Superhard Materials & Ceramics ·943·
In Fig.10 microstructures of CEM-printed parts of fine-grained hardmetals are shown. These parts were
fabricated with a finer nozzle than the FFF-printed parts. The cavities in z direction are more frequent but smaller.
Here, another type of porosity is identified along x-y direction. These pores are quite small and are separated in z
direction by the width of the fused filaments diameter. They stem from incomplete fusing of the fresh and liquid
top line with the bottom plane composed of already solidified lines.
The SEM image of higher magnification (Fig.10c) shows again the typical hardmetal microstructure. Both
hardmetal microstructures are quite uniform such as known from hardmetals fabricated by conventional
processing techniques. They are located in the two-phase region WC-Co and do not show any eta phases (if
carbon loss had occurred) or free C, which could be formed by carbon uptake upon thermal de-bindering from
cracked organics.
Similar observations than with FFF-printed hardmetals were made for cermet printed parts (Fig.11). Again,
there is mainly the cavities from misalignment in x-y direction (Fig.11a) because optimisation of printing strategy
was not yet performed. Hardly any fine porosity from the sequence in the layer in z direction is observed. The
SEM microstructure (Fig.11) shows the typical core-rime type character of hard-phase grains [10] not deviating
from a conventionally fabricated cermet.
Fig.10 Cross sections (LOM) and microstructures (SEM) of CEM-printed fine-grained WC-Co hardmetal. (a) one half of the
sample in Fig.8, bottom (arrow: inner wall); (b) magnification of a (Murakami etched), (c) SEM microstructure
Fig.11 Cross sections (LOM) and microstructure (SEM) of an FFF-printed Ti(C,N)-based cermet.
(a) one half of the part shown in Fig.8, centre (arrow: inner wall); (b) SEM microstructure
3.6 Modifications
3D printing techniques cover the potential of fabricating geometrical details
which are not accessible by other techniques. In case of the above-presented parts
which are related to indexable inserts for metal cutting applications this means that
hollow parts could be fabricated which can save material consumption or which
cover cooling channels. Fig.12 shows such a printed insert (with W-type geometry)
with a central hole and three cooling channels near the tips.
At the present stage of printer technology the nozzles have a diameter of ca. 0.3
– 0.6 mm. This is too large if very fine structures and reliefs are to be manufactured.
Fig.12 A prototype of a
CEM-printed indexable insert
(WC-Co hardmetal) with cooling
channels
(a) (b) (c)
(a) (b)

·944· 2018 World Congress on Powder Metallurgy
One solution of making fine structures is the post-processing of the green printed parts. This is easy as these parts
contain about 50 vol% organic material and hence simple grinding, blasting and mechanical machining is possible
and inexpensive. Other methods could be laser surface treatment [9] or gas-phase treatment with solvents to
smoothen the surface is another possibility. In this work we have investigated a simple imprinting for surface
conditioning.
Fig.13 shows an FFF-printed part (after printing and after sintering) into which a relief was imprinted simply
by pressing a hot counterpart (of defined temperature for a defined time) onto the surface of the printed green
body.
Fig.13 Printed parts with an imprint of an indexable insert, (a) green part, (b) sintered part
The results show that even very fine structures of the green body persist the
whole preparation process and are present in the sintered part. Such an
imprinting, together with other techniques mentioned above, can circumvent
the present lack of accuracy of extrusion-based printed bodies. A cross section
of the part shown in Fig.13 is given in Fig.14.
The surface does not show any cavities from the misalignment of deposited
lines in x-y direction but some inner cavities are enlarged due to the impact of
static load.
4 Conclusion
In this study we have shown that extrusion-based 3D printing by FFF and
CEM is possible by employing powder mixtures of hardmetals and cermets.
Parts of these materials were sintered and showed the typical microstructures.
Complete density was not yet achieved because the printing strategy was not
yet optimised and cavities occur from the printing process. Hence, the organic
binder system together with the printing procedure and hardware (printer nozzles) need to be further optimised.
However, our work is the first study on extrusion-based 3D printing of hardmetal and cermet parts by CEM
(composite-extrusion modelling) and FFF (fused filament fabrication) and is was proven that these printing
techniques are well-suited for an SDS (shaping – debinding – sintering) type of part fabrication. By further
development of this technology hardmetal and cermet parts with complicated geometries and new properties not
accessible by other techniques can be expected.
References
[1] M.K.Agarwala, Structural ceramics by fused deposition of ceramics, Proc. Solid Freeform Fabrication Symposium, p.1-8
(1995).
[2] M.A.Yardimci, S.I.Guceri, S.C.Danforth, Process modelling for fused deposition of ceramics, Ceram.Engineer.Sci.Proc. 17(3),
78-82 (1996).
[3] G.Wu et al. Fabrication of Metal Components using FDMet: Fused Deposition of Metals, Proc. Solid Freeform Fabrication
Symposium 1996, p.775-782 (1996).
[4] C.Kukla, I.Duretek, S.Schuschnigg, J. Gonzalez-Gutierrez, C.Holzer, Properties for PIM Feedstocks Used in Fused Filament
Fabrication, Proc. WorldPM (2016).
[5] C.Kukla, J.Gonzalez-Gutierrez, C.Burkhardt, C.Holzer, The production of Magnets by FFF – Fused Filament Fabrication, Proc.
EuroPM (2017).
Fig.14 Cross section (LOM) of one
half of the part shown in Fig.11
(arrow: inner wall) with an imprint of
a hot counterpart on top
(a) (b)

Part 5—Refractory Metals, Hard Alloys, Superhard Materials & Ceramics ·945·
[6] C.Lieberwirth, A.Harder, H.Seitz, Additive manufacturing with metal injection molding granules, Rapid. Tech.- Intl. Trade
Show & Conf. for Additive Manufacturing (2016).
[7] C.Lieberwirth, A.Harder, H.Seitz, Extrusion Based Additive Manufacturing of Metal Parts, J.Mech.Engineer. Autom. 7, 79-83
(2017).
[8] A.Demoly, C.Veitsch, W.Lengauer, K.Rabitsch, Effect of submicron Ti(C,N) on the microstructure and the mechanical
properties of Ti(C,N)-based cermets, Int.J.Refr.Met.&Hard Mater. 29(6), 716-723 (2011).
[9] C.Burkhardt, P.Freigassner, O.Weber, P.Imgrund, S.Hampel, Fused filament fabrication (FFF) of 316L Green Parts for the MIM
process, Proc. WorldPM (2016).
[10] P.Ettmayer, H.Kolaska, W.Lengauer, K.Dreyer, Ti(C,N)-based Cermets – Metallurgy and Properties, Int.J.Refr.Met.&Hard
Mater. 13, 343-351 (1995).