Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1225-7591(Print)
ISSN : 2287-8173(Online)
Journal of Korean Powder Metallurgy Institute Vol.27 No.3 pp.256-267

Multi-step Metals Additive Manufacturing Technologies

Ji-Won Oh, Jinsu Park, Hanshin Choi*
Advanced Materials and Process R&D Department, Korea Institute of Industrial Technology, Incheon 406-840, Republic of Korea

오지원·박진수: 학생, 최한신: 수석연구원

*Corresponding Author: Hanshin Choi, TEL: +82-32-850-0135, FAX: +82-32-850-0410, E-mail:
June 14, 2020 June 23, 2020 June 23, 2020


Metal additive manufacturing (AM) technologies are classified into two groups according to the consolidation mechanisms and densification degrees of the as-built parts. Densified parts are obtained via a single-step process such as powder bed fusion, directed energy deposition, and sheet lamination AM technologies. Conversely, green bodies are consolidated with the aid of binder phases in multi-step processes such as binder jetting and material extrusion AM. Green-body part shapes are sustained by binder phases, which are removed for the debinding process. Chemical and/or thermal debinding processes are usually devised to enhance debinding kinetics. The pathways to final densification of the green parts are sintering and/or molten metal infiltration. With respect to innovation types, the multistep metal AM process allows conventional powder metallurgy manufacturing to be innovated continuously. Eliminating cost/time-consuming molds, enlarged 3D design freedom, and wide material selectivity create opportunities for the industrial adoption of multi-step AM technologies. In addition, knowledge of powders and powder metallurgy fuel advances of multi-step AM technologies. In the present study, multi-step AM technologies are briefly introduced from the viewpoint of the entire manufacturing lifecycle.


    Korea Institute of Industrial Technology

    1. Introduction

    Additive manufacturing (AM) technology is to reshape traditional discrete manufacturing systems because of enlarged design freedom and corresponding changes in configuration of supply chains as well as manufacturing and assembly plans [1]. Innovations will be envisioned in either disruptive or continuous manners. Fuel nozzle of GE Aviation is representative for disruptive innovation. Monolithic fuel nozzle by powder bed fusion (PBF) AM reduces numbers of components and pertaining processes. As a consequence, supply chain is radically reconfigured by in-house adoption of AM technology. In addition, certified fuel nozzles allow business operation scope of OEM to be expanded toward after-sales spare parts markets. To the contrary, continuous innovations by customized production are achieved within the existing supply chain when binder jetting (BJT) AM technology is adopted by metal injection molding (MIM) parts providers. What makes AM technology reshape manufacturing landscape is originated from both principles of AM technologies and industrial adoption patterns.

    Additive manufacturing technology is defined and classified in ASTM standard [2]. Additionally, AM technologies are categorized into single-step processes and multistep processes according to consolidation degrees of asbuilt AM parts. With respects to metal AM, PBF, directed energy deposition (DED) and sheet lamination (SHL) AM result in consolidated 3D parts and they belong to single-step process. To the contrary, green-body parts are produced by multi-step processes such as BJT and material extrusion (MEX) AM. Until now, singlestep process has been main stream for metal AM because it shows opportunities in particular for less of more markets. Nevertheless, transition from early adoption to major adoption is retarded in the lifecycle of innovative technologies, which is ironically due to disruptive natures. As an alternative option, powder metallurgy based additive manufacturing technologies draw attentions and various researches and developments are conducted. Main driving forces of industrial adoption of BJT AM and MEX AM are cost reduction, material selectivity and continuous innovation. Quality and cost effectiveness of BJT AM and MEX AM are compared to laser powder bed fusion AM and metal injection molding in the literature [3]. Relative competitiveness of a technology against the other technologies was different according to part designs and production situations. For instance, costeffectiveness is expected to increase in the order of MEX, PBF, BJT and MIM as part production number increases. At the same production volume, the competency order can be changed by design complexity and readiness of post-AM processes.

    In the present study, multi-step metal additive manufacturing technologies, BJT AM and MEX AM, are briefly reviewed to encourage powder metallurgy scientists, engineers and companies to exploit multi-step AM processes in PM innovations.

    2. Metals Additive Manufacturing

    Metal additive manufacturing is to fabricate metal parts by additive manufacturing technologies. In addition to standard classification, AM technologies are categorized into single-step process and multi-step process in ISO/ ASTM DIS 52900 [4]. Characteristic features are compared in Fig. 1.

    2.1 Single-step AM processes

    Consolidated bulk parts with specified geometric features/ shapes are fabricated via single-step AM processes. In metal AM, directed energy deposition AM, powder bed fusion AM and SHL AM belong to single-step AM process. In most cases, 3 dimensional parts are built by multiple-pass welding principles while particle impaction, friction-stir welding and ultrasonic welding are utilized for consolidation to bulk. Among them, laser PBF AM is regarded as main-stream technology.

    2.2 Multi-step AM processes

    Geometric shapes are embodied after AM processing. However, powders are bound to maintain the shape with aids of adhesives or polymers. Therefore, densification process is necessary for the as-built parts. BJT AM and MEX AM are representative for metal multi-step AM process.

    From the viewpoint of completeness of production, single- step processes are advantageous. Furnaces and facilities for debinding and densification are additionally installed at multi-step AM shop. Both debinding and densification processes are time-consuming though productivity depends on throughput and production volume of parts. At the current state of technology, PBF and DED AM technologies are not stand-alone ones. Localized melting-solidification results in sharp temperature gradient, thermal stress and complex phase distribution. In addition, support structures which is necessary but notvalue- added factor are required for thermal management and so-called overhang structure building. As a result, support removal and post-AM heat treatment (at least stress relief heat treatment) are practically necessary for the single-step processes.

    Material selectivity is of multi-dimensional aspects. Multi-step processes shows much wider availability of materials than single-step processes. Because single-step process adopts melting-solidification consolidation mechanism, selection of materials is filtered from the viewpoints of weldability of powder materials, reactivity with oxygen, vaporization of alloying elements, metallurgical reactions between different powder materials and physical properties. To the contrary, sinterability and wettability are main criteria for multi-step processes. On the other hand, powder selectivity of in-vessel processes such as BJT and PBF is practically limited. Powder layer is placed in situ. Therefore, powders should have properties to form powder bed with reproducibility. Changing one feedstock powder to different chemical composition powder causes cross-contamination issues. Practically, it seems that changing feedstock powders and utilizing mixed powders (multi-modal powders and dissimilar chemical compositions) are impossible for safety-critical parts fabrication of regulation-based industries. To the contrary, out-of-vessel processes have much more freedom to change chemical compositions of feedstock. Even elemental powders are utilized to fabricate alloyed 3D parts for DED AM and MEX AM.

    ISO/ASTM 52910:2018 provides a guideline for industrial AM adoption and lots of researches have been studied to evaluate quality, processing ability, productivity and cost effectiveness of AM technologies [4]. What is important is not the validation of competitiveness of an AM technology but the optimum selection among AM technologies. When it is extended to manufacturing business operation, manufacturing process plan according to market phases and reconfiguration of supply chains according to product lifecycles are key performance with sustaining quality, cost and lead-to-market managements. Functional parts are evaluated by MEX AM at the market generation stage. When market enters growth stage, BJT AM replaces MEX AM. Finally, MIM will be major manufacturing process at matured market phase while BJT AM and MEX AM are allocated for scheduled production and unscheduled production of spare parts, respectively.

    3. Binder Jetting Additive Manufacturing

    BJT AM has been widely applied to make figures, rapid prototyping and casting tools. Customer products are directly fabricated from gypsum powders with multicolor binders. Sand parts from BJT AM are used as molds or inserts in casting. When it comes to metal BJT AM, ExOne (USA), Digital Metal, Desktop Metal and HP are main vendors for BJT AM machines.

    Green body parts are fabricated by BJT AM, thermal curing and depowdering. 2D pattern is translated by powder placement, selective BJT and infra-red curing [5]. In the case of powder placement, powder is dispersed on build surface firstly and counter-clock-wise rotating roller flattens dispersed powders during traveling. Adhesive ink is jetted by either continuous mode or drop-on-demand mode. Aqueous adhesive is spread on powder surface and it permeates through pore network. Balance between imbibition and drainage results in footprint in powder layer in which powders are covered with aqueous adhesive. 2D pattern is partially cured by IR irradiation to achieve pattern stiffness during subsequent powder placement. After completion of final 2D pattern, parts are embedded in powder bed of job box. They are thermally cured in oven to develop green part strength for handling. Parts are extracted by depowdering process. Unused powders are recycled by mechanical sieving.

    Key-process variables which are affecting processing abilities and quality of green-body parts are summarized in Table 1.

    Metal BJT AM is contrasted to metal PBF AM. Thank to low-temperature process in arbitrary environment, anchoring to build platform and support structures are not necessary. Powders are basically inert to process environment while oxidation and vaporization of powders are considered for PBF AM processing. With respects to productivity, so-called bin packing density is much improved for BJT AM (Fig. 2(b)) as compared in PBF AM (Fig. 2(a)). At the current state of PBF AM technology, thermal management is the primary issue and thus parts are placed directly on build platform as shown in Fig. 2(a). Though support-free PBF AM technology (Velo3D) is introduced [6], part packing in building volume is still limited. To the contrary, BJT AM allows parts to be freely placed within the building volume of machine. Bin packing density emphasizes flexibility of manufacturing scheduling.

    Productivity of PBF AM has been improved by adopting multiple energy sources in a single machine. In the case of BJT AM, simplified toolpath for 2D pattern translation is focused. As schematically represented in Fig. 3(a), complex tool paths are required for the conventional BJT machine. From the viewpoint of lean manufacturing, unnecessary movement/motion is wasteful. As a matter of fact, Desktop Metal adopts bidirectional lay- ering system in Fig. 3(b). It means that powder dispersion, powder compaction, BJT and curing are continuously conducted for a 2D pattern formation on one-way travel direction. Subsequent 2D pattern is generated during travel in the opposite direction. It reduces unnecessary movements and accordingly productivity is much improved.

    In these contexts, BJT AM is a potent option for mass customization with compatible cost effectiveness, large design freedom and flexible manufacturing plan.

    4. Material Extrusion Additive Manufacturing

    In plastic injection molding, serial processes of melting, forced flow and solidification of polymer feedstock are utilized with aids of formative molds. On the other hand, polymer MEX AM, extrudate of molten polymer is selectively placing to build 3 dimensional part, which eliminate necessity of molds. In the same way, greenbody parts in which metal powders are bounded by binder phase are manufactured by either metal injection molding or metallic feedstock extrusion AM. Metallic feedstock extrusion AM can be divided into three groups according to feedstock types: filament-type feedstock, rod-type feedstock and pellet-type feedstock.

    Volumetric loading of metallic powder in feedstock is of duality. As it is increased, higher density of greenbody parts are produced, which is advantageous from the viewpoint of sintering. However, filament becomes toobrittle to be wound for continuous feeding. Additionally, increased viscosity of molten feedstock deteriorates extrudability [7]. In this context, researches have been conducted to develop binder phase systems for flexible filament with high volumetric loading of metallic powder. BASF commercializes metallic feedstock filaments which are adaptable for commercial polymer extrusion AM machines. In filament-type feedstock, filament from spool is fed into hot end to melt while unmelted upper filament acts as a plunger. Rod-type feedstock seems to have similar function with filament. However, discrete rod do not have to be bendable so higher solid loading can be achieved. Rod feedstock is periodically charged before complete consumption of previous one. Desktop Metal adopts the rod-type feedstock [8]. Pellet is typical feedstock for MIM and it is adopted by AIM3D [9]. Pellet- type feedstock of which dimension is not specified is fed into screw-type extruder. Pellets in reservoir seem to be periodically charged into extruder.

    In principle, solid feedstock is melted and extruded through nozzle orifice. The extrudate is selectively overlaid to form bead. During overlaying, settled bead undergoes phase transformation [10]. Overlaid bead shape is determined by volume balance between volumetric extrusion rate and bead volume generation rate as shown in Fig. 4. The volume extrusion rate (b)-1 can be a function of bead width (b)-1 with working distance (a)-3, unit length (c)-1 and traveling speed (b)-2. Bead height (a)-3 represents the height of the bead that is actually formed. This relationship can be expressed as a function of (d).

    Volumetric extrusion rate is proportional to bead volume generation when feedstock materials have proper extrudability. At a given bead volume generation rate, cross-sectional bead shape is inversely proportional to linear travel speed. It means that bead shape can be controlled by changing travel speed.

    Bead shape is the critical factor affecting geometric capability and surface finish as well as productivity. In MEX AM, building time is estimated by total length for building and mean travel speed. As cross-section area of bead is increased, total length is shorten and therefore building time is reduced. However, geometric capability and surface roughness are sacrificed as exemplified in Fig. 5. In addition, volume shrinkage of bead and building part by cooling to room temperature is also considered owing to large coefficient of thermal expansion of polymeric binder materials.

    Parts are built on build platform in the manner of free standing. Therefore, parts which have overhang features need support structures: so-called 45 degree rule. In addition to mechanical supporting, part and support interfaces should be metallurgically separated during sintering. Desktop Metal devised ceramic interfacial layer. Narrow overhangs between features and features and interfaces between features and supports adopt ceramic layers. Ceramic layers sustain tolerance during post-AM processes. They are hard-to-sinter and thus they can be removed with facility after sintering.

    Key-process variables which are affecting processing abilities and quality of green-body parts are summarized in Table 2. The parameters are optimized from the viewpoints of extrudability, shape forming ability, bondability and buildability. Volumetric extrusion rate of feedstock is controlled on demand (extrudability). Overlaid bead is conformable to specified bead shape in design (shape forming ability). Beads which are directly contacted horizontally and vertically should be mixed to form continuous and strong interfaces (bondability). Vertically stacked beads have ability to sustain macroscopic design until completion of building (buildability). Parametric control is achieved by arranging relationships between parameters and establishing logical algorithm. Furthermore, building part healthiness should be managed by implementing real-time diagnostics and closed-loop control. It is challenging to establish sensing strategies owing to limitations of available space and complex phenomena.

    At the present state, flexibility of feedstock selection is main advantage of MEX AM in spite of design limitation, geometric capability limitation and low productivity. Less of More production is a good market penetration solution of MEX AM.

    5. Post-AM Densification Processes

    Shape of green-body parts from BJT AM and MEX AM is retained by binder phases. However, the binder phases should be removed before full density parts manufacturing. The binder removing process is called debinding process and green-body parts are transformed to brown-body parts. After that, brown-body parts are densified by either sintering process or molten-metal infiltration process. In the case of BJT AM, water of aqueous binder evaporates by curing and open-structured pore networks are well developed through green-body parts. Therefore, so-called thickness issues in debinding process kinetics can be neglected.

    5.1 Debinding

    Binder system has great influences on quality, processing ability, productivity and cost effectiveness. Mechanical supporting before achieving self-supporting ability via sintering as well as controllable rheology for viscous flow of multiple components feedstock are required for binder system for high solid powder volume fraction. With respects to processing ability, the binder system should be removed fast without remaining residues. Practically, binder phase is necessary but not-value-added one in multi-step AM process and MIM process. Binder removal strategy shall be considered in designing binder system. Thermal debinding, solvent debining, catalytic debinding and supercritical debinding are options for binder removal process. It is determined according to binder material systems. Thermal debinding process is conducted in air, vacuum, inert, reducing and wicking environment. In general, thermal debinding process is time-consuming because slow heating rate is applied to avoid cracking, distortion, blistering, slumping and peeling of parts [11]. Therefore, multiple debinding mechanisms are adopted to accomplish practical debinding process: solvent debinding and thermal debinding system or step-wise thermal debinding system. Regardless of binder system design, firstly removed binder phase results in open pore network while backbone binder still supports debinding part shape. At the higher temperature, backbone is thermally decomposed and transported through the pore channels.

    Pre-sintering is considered to sustain shape after complete removal of binder phase which binds metallic powders. Pre-sintering allows powders to form sintering neck at contacting surfaces. Sintering neck replaces binder phase to retain shape for brown-body parts during handling and sintering in particular for parts with overhung features. Otherwise, debinding process is conducted in wicking environment. Green-body parts are embedded in granule powders which have no reactions with greenbody parts and act as support for debinding parts and brown-body parts.

    5.2 Sintering

    Sintering is well-established phenomena for powder densification so anisotropy of linear shrinkage of AM parts and dependency of sintering mechanism of metallic powders are highlighted in the present study. Sintering is a thermally activated process and it reduces surface energy by consolidating powders to bulk. Therefore, densification accompanies volumetric shrinkage. In order to achieve near-net shaping or net shaping by sintering, brown parts show isotropic shrinkage ideally or anisotropic shrinkage shall be quantified reproducibly. In most cases for both BJT [12-14] and MEX [15] AM, shrinkage ratio along building direction (z direction) is higher than that of planar direction (x and y directions). It may be due to non-uniform distribution of powders and discontinuous defects owing to AM principles and building strategies. Quantified directional shrinkage ratio is fed back to part 3D modeling.

    Higher green-body density is advantageous for sintering. It is determined by powder properties and powder placement parameters for BJT AM. Spherical powders with bimodal size distribution are known to obtain higher powder packing density. However, utilization of bimodal powder poses quality management issues: uniform distribution during powder placement and reproducibility for reuse powder from recycling. On the other hand, powder selectivity for MEX feedstock is much large. When a single modal powder is used, finer powder size with high powder layer density is beneficial for enhanced sintering and density [16].

    Activated sintering has been explored by design chemical composition in powder metallurgy. Liquid phase sintering is more effective to achieve full density. IN718 green parts by BJT AM is densified above relative density of 99% with introduction of liquid phase sintering [17]. Tungsten is difficult to sinter but addition of transition metal sintering activators such as Ni and Fe improves sinterability of W. Slightly higher sintering temperature is required for BJT AM W heavy alloy than traditional powder metallurgy counterpart, which is due to coarse powder size of BJT AM [18]. Same to adoption of multi-modal powders, reproducibility of mixed powder uniformity shall be demonstrated for recycled feedstock powder in BJT AM.

    In the MEX AM, feedstock is fed to building part externally. Therefore, dissimilar feedstock can be selectively overlaid for a single part when multiple feeding system is available. In this case, dissimilar-material part shall be simultaneously sintered and different shrinkage behaviors result in deformation and cracking. In this regard, combination of retarded sintering and activated sintering is emphasized to manipulate co-sintering of a part with dissimilar powders.

    5.3 Molten-metal infiltration

    Molten metal infiltration is of similarities with liquid phase sintering except that molten liquid is externally supplied toward porous skeleton bodies. Primary advantage of molten metal infiltration over sintering is reduced volume shrinkage during densification. Pressureless infiltration and pressurized infiltration are selectively applied according to wettability and capillary force of powder and infiltrant systems [19, 20]. In addition, reactions between powder and molten metal are considered for infiltration processing ability and infiltrated part properties. As current state of the art, bronze is generally used for low-melting infiltrant material for ferrous green-body parts [21]. In the molten metal infiltration, metallurgical interactions between powder materials and infiltration materials need to consider. The material systems are either reactive or non-reactive. For the reactive systems, dissolution and dissolution-precipitation can be utilized to design powder-matrix interfacial morphology, phase compositions and part properties after infiltration. In the literature, TiCx-steel system in which carbon activity and corresponding carbon transportation are dependent on carbon contents in TiCx and steel. TiCx feedstock powders were prepared by milling of mixed TiC and Ti with changing mixing ratio. Green-body parts from different TiCx powders were manufactured by BJT AM, separately. After stacking TiCx green parts in a whole structure, it was infiltrated with molten steel. Multi-layered parts with different phase compositions are reported. It is highlighted that final part properties are tailored by designing material interaction pathways and interactionpathway based 3D architectures. In other literature, WC skeleton parts from BJT and pre-sintering was infiltrated with Co molten metal. Pre-sintering of WC skeleton affects phase composition and shape retention ability of WC-Co parts in free-standing infiltration condition. With respects to defects, volumetric changes of solidification and cooling of infiltrated metals are considered. Voids/ pores can be generated when density of liquid is lower than that of solid [22]. To the contrary, volume expansion occurs when molten Si solidifies, which results in excess volume on surfaces of part [23].

    5.4 Hot isostatic pressing

    Thermo-mechanical sintering is more effective to achieve higher sintering density at lower sintering temperature. Though external pressure can be applied uniaxially to powders in mold, isostatic pressuring is inevitable for the near-net shape parts, hot isostatic pressing is effective to collapse closed pores which affect mechanical properties such as fatigue. In this regard, HIP process is generally specified for aerospace parts from PBF AM and casting. As a matter of fact, pressureless-sintered BJT parts are further densified above relative density of 99% via HIP when high-end applications are required [24]. To do this, sintered parts shall have only closed pores. It is noted that collapse of closed pores is dependent on internal gas pressure. As pore is shrunk, internal gas pressure is increased. Therefore, volumetric reduction of pores reaches equilibrium state by balancing external pressure and internal pressure at the given working temperature. X-ray CT study on collapsed pores after HIP process reveals that the pores are regrown with temperature increasing in ambient environment pressure [25]. It implies that effects of sintering environments and internal gas pressures of closed pores on pore collapsing ability need to be considered. Recently, rapid quenching/ cooling cycle is achieved in HIP [26]. It is beneficial for simplification of post-AM heat treatments pertaining to phase composition manipulation.

    6. Summary

    Multi-step metal additive manufacturing technologies are actively adopted in industries. Eliminating molds in green-body parts manufacturing is advantageous against metal injection molding. Continuous innovation from availability of conventional materials and manufacturing technologies contrasts them from single-step metal AM technologies. Nevertheless, there are still uncertainties pertaining to qualities of parts and process reliability. AM technologies shall demonstrate not only opportunities against conventional technologies but also compatibility to conventional quality control and quality assurance in industries. To do this, rate-limiting steps and cost-consuming factors shall be replaced. Simultaneously, geometric capabilities, defects control, properties control and processes control shall be arranged in the whole life cycle of manufacturing. Knowledges of powder metallurgy will fuel advances of the multi-step metal AM technologies.


    This research was financially supported by the Korea Institute of Industrial Technology (KITECH) through the Research and Development [Development of root technology for multi-product flexible production (EO200015)]


    Categorization of metal additive manufacturing technologies.
    Comparison of bin packing density: (a) typical placement in PBF AM and (b) typical placement of BJT AM.
    Comparison of tool-paths for a 2D pattern generation: (a) conventional binder jetting machine system (ExOne) and (b) bi-directional binder jetting machine system (Desktop Metal).
    Balance between volumetric extrusion rate and overlaying bead volume generation rate, (a)-1 : bead width, (a)-2 : bead height, (a)-3 : working distance, (b)-1 : volume extrusion rate, (b)-2 : traveling speed, (c)-1 : unit length, (c)-2 : Overlaying bead volume generation rate and (d) relationship between volumetric extrusion rate and bead shape evolution.
    Estimation of building time and surface roughness by varying cross-sectional area of bead: for a 100 × 100 × 100 mm3 cube at mean travel speed of 100 mm/sec.


    Key-process variables in BJT AM
    Key-process variables in MEX AM


    1. J. Oh, H. Na and H. Choi: J. Korean Powder Metall. Inst., 24 (2017) 494.
    2. ASTM, F 2792-12a, Standard Terminology for Additive Manufacturing Technologies.
    3. M. M. Matthias: Metal AM, 4 (2018) 159.
    4. ISO/ASTM DIS 52900:2018(E), Additive manufacturing - General principles - Terminology.
    5. J. Oh, S. Nahm, B. Kim and H. Choi: Korean J. Met. Mater., 57 (2019) 227.
    7. Q. Sun, G. Rizvi, C. T. Bellehumeur and G. Peihua: Rapid Prototyp. J., 14 (2008) 72.
    8. R. Burnham, J. LaPlante and A. Preston: USA US 10,464,260 (2019).
    9. D. M. Nieto, V. C. López and S. I. Molina: Addit. Manuf., 23 (2018) 79.
    10. J. Park, H. Lee and H. Choi: Arch. Metall. Mater., 65 (2020) 1069.
    11. R. K. Enneti, S. J. Park, R. M. German and S. V. Atre: Mater. Manuf. Process., 27 (2012) 103.
    12. Y. Wang and Y. F. Zhao: Procedia Manuf., 10 (2017) 779.
    13. H. Chen and Y. F. Zhao: Rapid Prototyp. J., 22 (2016) 527.
    14. I. Gibson, D. Rosen and B. Stucker: Additive Manufac turing Technologies, Springer, New York, (2015) 205.
    15. P. K. Gurrala and S. P. Regalla: Virtual Phys. Prototyp., 9 (2014) 127.
    16. S. Mirazababaei and S. Pasebani: J. Manuf. Mater. Process., 3 (2019) 1.
    17. P. Nandwana, A. M. Elliott, D. Siddel, A. Merriman, W. H. Peter and S. S. Babu: Curr. Opin. Solid State. Mater. Sci., 21(4) (2017) 207.
    18. M. T. Stawovy, K. Myers and S. Ohm: Int. J. Refract. Met. Hard Mater., 83 (2019) 104981.
    19. C. L. Cramer, P. Nandwana, R. A. Lowden and A. M. Elliott: Addit. Manuf., 28 (2019) 333.
    20. J. M. Molina, R. A. Saravanan, R. Arpón, C. Garcıa-Cordovilla, E. Louis and J. Narciso: Acta Mater., 50 (2002) 247.
    21. M. Doyle, K. Agarwal, W. Sealy and K. Schull: Procedia Manuf., 1 (2015) 251.
    22. M. Schöbel, G. Requena, G. Fiedler, D. Tolnai, S. Vaucher and H. P. Degischer: Compos. Part A Appl. Sci. Manuf., 66 (2014) 103.
    23. J. Roger, L. Guesnet, A. Marchais and Y. Le Petitcorps: J. Alloys Compd., 747 (2018) 484.
    24. V. Kumar, E. N. Chibuzo, J. A. Garza-Reyes, A. Kumari, L. Rocha-Lona and G. C. Lopez-Torres: Procedia Manuf., 10 (2017) 935.
    25. A. du Plessis and P. Rossouw: J. Mater. Sci. Technol., 24 (2015) 3137.
    26. Y. Y. Kaplanskii, Z. A. Sentyurina, P. A. Loginov, E. A. Levashov, A. V. Korotitskiy, A. Y. Travyanov and P. V. Petrovskii: Mater. Sci. Eng. A, 743 (2019) 567.