Additive Manufacturing of TiC/Steel Composites Using Electron Beam Melting and In Situ Infiltration

Cemented carbides containing titanium carbide (TiC), characterized by low specific weight, superior hardness, and excellent high‐temperature resistance, are widely used as wear‐resistant cutting tools and brake discs. Instead of using conventional powder metallurgical approaches, composites comprising TiC and stainless steel are fabricated in this study by means of an innovative method combining electron beam powder bed fusion (PBF‐EB) and in situ liquid metal infiltration. For PBF‐EB, a pure TiC powder as feedstock and a stainless steel plate as substrate are used, respectively. Owing to the high processing temperature in the hatching area, the TiC particles are inhomogeneously sintered. Simultaneously, an in situ infiltration of liquid steel into the pores between TiC particles takes place during PBF‐EB, forming a dense and crack‐free composite material containing ≈68% TiC. The electrical properties of the raw TiC powder are investigated at different temperatures to optimize the process parameters ensuring high process stability during PBF‐EB. The microstructure and mechanical properties (compression strength, hardness, and fracture toughness) of the PBF‐EB‐produced TiC/steel composite are analyzed.


Introduction
Titanium carbide (TiC) is an interstitial metal carbide that is formed by the incorporation of carbon atoms into the interstitials of titanium, crystallizing in a cubic B1 rock salt structure. [1]It exhibits relatively good electrical and thermal conductivity (≈10 6 S m À1 and ≈29 W m À1 K À1 , respectively, being comparable to titanium and its alloys), high hardness (>28 GPa), high ≈22.5 μm. [25,26]Maurya et al. utilized laser pulsed shaping techniques to control the energy delivery during L-PBF, and fabricated crack-free TiC-based composites using Fe as binders instead of Ni or Co, offering a safer, more cost-effective, and sustainable alternative. [27]Electron beam powder bed fusion (PBF-EB) is another powder bed-based AM approach for producing high-performance components, during which the energy input and absorption are very high. [28]Rock et al. successfully processed a mechanically alloyed powder composed of 60 vol% Mo and 40 vol% TiC using PBF-EB, demonstrating the versatility and potential of PBF-EB in processing TiC-based composites. [29]o far, to produce ceramic/metal composites using powder bed fusion AM approaches, a powder blend (ceramic powder þ metal powder) has always been used as feedstock. [14]Commonly, ceramic powders are produced via milling, while metal powders are derived by atomization, corresponding to a high fabrication price. [30,31]The utilization of powder blends presents significant recycling challenges that further increase manufacturing costs.In addition, possible demixing could take place during melting as well as raking owing to different specific weights and particle sizes of the powdered raw materials. [32,33]n this work, TiC powder was utilized as a single feedstock for PBF-EB.This approach is expected to greatly reduce material costs and significantly improve the efficiency of the manufacturing process compared to the use of powder blends.During PBF-EB under vacuum conditions, very high processing temperatures of the powder bed could easily be reached, [34,35] so that the start plate composed of stainless steel was molten and infiltrated into the porous printed TiC parts spontaneously.Based on this procedure combining PBF-EB and in situ melt infiltration, dense TiC/steel composite materials free of cracks were successfully produced in this work.So far, PBF-EB of pure carbide ceramics as feedstock has not yet been reported in the literature.In order to gain a deep insight into the process stability during PBF-EB that was used for the first time to process pure TiC, the temperaturedependent electrical resistance of TiC powder was analyzed and discussed.The microstructure and mechanical properties (compressive strength, hardness, and fracture toughness) of the TiC/steel composites produced by means of PBF-EB and in situ infiltration were investigated.

Experimental Section
A crushed titanium carbide powder (abcr GmbH, Karlsruhe, Germany) was used for PBF-EB in this study.The powder was reused after each build, with the recycling processes including blasting, sieving, and oxygen content testing.Stainless steel 1.4571 (AISI 316Ti) was used as the substrate (diameter: 100 mm; thickness: 20 mm).The particle morphology of the TiC powder was investigated utilizing a scanning electron microscope (SEM, NanoLab 600 DualBeam, FEI-Quanta 450, USA).Compositional analysis of the printed parts was performed using energy-dispersive X-Ray spectroscopy (EDS, integrated in the SEM system), whereas X-Ray fluorescent analyzer (XRF, S1 Titan, Bruker Corporation, Billerica, USA) was applied to determine the chemical composition of the stainless steel substrate.The oxygen content of the raw and reused powder was measured via carrier gas hot extraction method using an oxygen-nitrogen analyzer (EMGA-920, HORIBA Ltd, Kyoto, Japan).The particle size distribution was measured by means of laser diffraction granulometry (LDG, Mastersizer 3000, Malvern Panalytical, London, UK) and image analysis (MIPAR, MIPAR Software LLC, Columbus, USA). [36]Currently, PBF-EB is mainly applied to process metals with high electrical conductivity.Materials with poor electrical conductivity suffer an explosive powder scattering caused by Coulomb interaction during PBF-EB, also known as the "smoke phenomenon," which can lead to the termination of the entire PBF-EB process. [37]Based on the results reported by Chiba et al. the electrical conductivity of the powder bed significantly increases with temperature. [38,39]With increasing electrical conductivity of the powder layer, electrostatic charge can be immediately transferred to neighboring particles after the electron injection, suppressing the powder smoking. [40]Thus, to ensure good process stability and to avoid powder smoking, the direct current (DC) electrical resistance of the TiC powder was analyzed at various temperatures using an in-house developed experimental setup (Figure 1).The powdered TiC sample was loosely packed in an Al 2 O 3 crucible, whose bottom side is connected with two electrodes to determine the electrical resistance at different temperatures.The measurement of the electrical resistance was carried out in a vacuum furnace (Gero LHTM Carbolite, Gero GmbH, Germany).The powdered sample was heated under vacuum conditions (10 À5 mbar) up to 1000 °C with a heating rate of 5 K min À1 and a dwell time of 1 h.Subsequently, the sample was furnace-cooled to room temperature under vacuum conditions.After cooling down, the furnace chamber was flooded with air up to atmospheric pressure.After staying under normal atmospheric conditions and at room temperature for 12 h, the sample was then subjected to another heating cycle using the identical parameters, as described earlier.
PBF-EB was conducted on a Freemelt ONE machine (Freemelt AB, Mölndal, Sweden), where the chamber pressure was initially reduced to ≈10 À5 mbar and a helium supply at a partial pressure of ≈10 À3 mbar was maintained during the building process.The first layer of TiC powder was raked onto a preheated stainless steel substrate.The raked powder layer was then preheated using a defocused electron beam.The temperature of the bottom side of the substrate was measured by means of a thermocouple, which was displayed as the preheating temperature.The preheating temperature was maintained at ≈1200 °C during the whole PBF-EB process by preheating each newly raked powder layer.
During the subsequent selective melting procedure, a defocused electron beam with a diameter of ≈1000 μm was used, resulting in the inhomogeneous sintering of the TiC particles and a relatively high porosity of the printed TiC layer.The increased degree of sintering within the hatching area leads to improved thermal connectivity among the particles compared to the looser, less sintered powder. [41]This enhanced thermal conductivity facilitates heat transfer throughout the hatch area.Simultaneously, the upper area of the stainless steel substrate in contact with TiC was molten owing to the high processing temperature in the hatching area, leading to an in situ melt infiltration into the printed porous TiC parts.During selective melting, parallel hatching with a line offset of 100 μm was applied; the scan speed and the beam power were set to 2000 mm s À1 and 1200 W, respectively.Table 1 shows the processing parameters used for the fabrication of cuboid-shaped TiC/steel composites.After PBF-EB, cuboid-shaped TiC/steel composite specimens measuring 10 Â 25 Â 5 mm 3 (x Â y Â z, see the coordinate system in Figure 4b) were produced.The reproducibility of the microstructural features was verified by analyzing at least 6 samples manufactured using the same process parameters.For microstructure characterization, the cuboids were first cut parallel to the build direction (z-direction) using wire erosion followed by grinding and polishing.The so-prepared yz-cross sections were investigated using SEM.To determine the size and shape of the TiC phase, the SEM micrographs were investigated using MIPAR software.Furthermore, a SEM (Helios NanoLab600, FEI, USA) equipped with an electron back-scattered diffraction (EBSD) detector was employed to explore the local grain orientation of the printed parts.The local element distribution of the cross sections was investigated using electron probe micro analysis (EPMA, JXA 8100, Jeol, Tokyo, Japan).The area fraction of the pores was measured on binarized images of the cross sections, and the relative density was calculated.The true density of the samples was determined using a helium pycnometer (Pycnomatic ATC EVO, Porotec GmbH, Germany).
To evaluate the mechanical properties of the PBF-EB-built TiC/steel composites, the compression tests were conducted on samples with dimensions of Φ4.5 mm Â 4.5 mm, following the standards of ISO 4506:2021, utilizing a universal testing machine (Instron 1195, Instron GmbH, Norwood, USA).Vickers micro-indentation testing was conducted on the TiC and steel substrate using an LM300 hardness tester (LECO instrument GmbH, USA).A load of 0.2 kg and a dwell time of 12 s were used for the indentation tests.Based on the indentation results, the fracture toughness (K IC ) of the TiC phase at room temperature was calculated according to the following equation [42] K where E is the Young's modulus [GPa], H V denotes the hardness [GPa], P corresponds to the maximum applied force [mN], C represents the length of the crack path from the center of the indent [μm], and δ is an empirically determined calibration constant, taken as 0.016 for the Vickers tip. [43]

Results and Discussion
Figure 2a shows the particle morphology of the TiC powder.
According to the high-resolution SEM, nano-sized particulates are observed on the surface of the fresh and reused TiC powder (insert of Figure 2a).Oxygen measurements indicate the presence of oxygen in the samples, suggesting that these tiny particulates could be TiO 2 oxides with very high thermodynamic stability (see the following equations). [44,45]C The highly negative values of the Gibbs free energy of reaction (Δ r G) at low temperatures do not mean that the reactions are more favorable with decreasing temperature; at low temperatures, the reactions are mainly dominated by kinetics.According to Shabalin et al. a high reaction rate of the oxidation of TiC was observed at ≈800 °C, especially in the case of small TiC particles with a high specific surface area being more susceptible to oxidation. [46]The influence of the small TiO 2 particulates on the process stability and on the electrical conductivity at various temperatures is further discussed in the following paragraphs.The results of the particle size distribution of the TiC powder measured via LDG and image analysis are in good accordance with each other (Figure 2b).Thus, the image analysis  method was also used to characterize the size of different phases of the printed TiC/steel composites.According to Cordero et al. the undesired "smoke phenomenon" is mainly induced by the Coulomb repulsion forces between the particles caused by the electrostatic charge. [37]hus, a high electrical conductivity of the powder bed is required to enhance the process stability and to suppress the smoke effect.As reported in the literature, the electrical conductivity of TiC significantly decreases with increasing temperature. [47,48]herefore, the TiC powder was first processed at low temperatures using PBF-EB, during which powder smoking was always observed and could barely be avoided.Nevertheless, after some trials and errors, it was found that the process stability could be improved when increasing the preheating temperature.Moreover, Figure 3a shows that the DC electrical resistance of the raw TiC powder significantly decreases with increasing temperature, which is considered to be attributed to the presence of the tiny oxide particulates (inset of Figure 2a), whose electrical conductivity increases with temperature following the Arrhenius law. [49,50]In similarity to the observation gained by Chiba et al. the low electrical resistance remained after cooling (Figure 3a). [38,39]imultaneously, the TiC powder particles in the crucible were very slightly sintered after the electrical resistance measurement.The temperature-dependent DC electrical resistance of the reheated sample is presented in Figure 3b.After 12 h under normal atmospheric conditions and at room temperature, the good electrical conductivity of the powder cake was maintained.The electrical resistance of the reheated sample remains within the same order of magnitude as that of the powder bed after the first heating.The same experiments were repeated several times, showing high reproducibility.The origin of the remaining low electrical resistance at room temperature after the thermal treatments has been so far not well understood and might be explained by the slight sintering effect of the nano-sized oxides with a significantly lower melting point (melting point of TiO 2 : 1 843 °C) compared to TiC (melting point of TiC: 3 065 °C).
Figure 4a depicts a schematic drawing to show the fabrication of TiC/steel composites using PBF-EB combined with in situ melt infiltration.During the selective melting procedure, the raked TiC powder layer was inhomogeneous sintered, forming a porous structure comprising large elliptical and spherical TiC particles with a size of several hundreds of micrometers (dark phase in Figure 4c).At the same time, owing to the high processing temperature, the upper part of the stainless steel substrate was molten and infiltrated into the pores between the TiC particles through capillary effects.In the course of PBF-EB, the relatively good thermal conductivity of the printed TiC/steel structure ensures continuous in situ melt infiltration.According to Washburn's equation, [51] the theoretical maximum infiltration height (h max ) was found to be on the order of 1 m when infiltrating the steel melt into a porous TiC preform with pore sizes of several tens to a few hundred micrometers.Nevertheless, from the experimental point of view, h max also depends on the temperature of the steel substrate, which is affected by the thermal conduction of the printed TiC/steel structure and the powder bed during PBF-EB, i.e., the in situ infiltration procedure is terminated, when the heat of the layer being printed cannot be effectively conducted to the steel substrate due to the high thickness of the printed parts, and when the temperature of the upper area of the stainless steel substrate cannot reach the melting point of steel.
The as-built TiC/steel specimens are shown in Figure 4b.After PBF-EB, the surrounding TiC powder particles were barely or not sintered and could be very easily removed, sieved, as well as reused.Figure 4c presents the cross-sectional microstructures of the printed TiC/steel composite material.As can be observed, the TiC/steel parts are dense and free of cracks, which is further supported by true density measurements indicating a porosity of less than 1.5%.According to the image analysis, the fraction of the dark phase (TiC) is ≈68%.Based on XRF and EDS, the chemical compositions of the stainless steel substrate (Figure 4b) and the bright phase of the printed parts (Figure 4c) are identical owing to the in situ infiltration of molten steel substrate during PBF-EB.As depicted in Figure 4c, a high microstructural homogeneity is observed along the thickness direction (z-direction as shown in Figure 4b).In the micrographs with a higher magnification (lower images in Figure 4c), apart from the TiC particles with a size of several tens of micrometers being comparable to the size of the initial TiC powder, large spherical and elliptical TiC agglomerates with several hundreds of micrometers in size are detected.The formation of the large TiC agglomerates is  caused by the sintering of the original TiC powder particles, due to the inhomogeneous heat distribution and powder bed properties during PBF-EB.These microstructural features were found to be highly reproducible and were also observed in other samples produced using the same parameters.It is considered that the size and fraction of the TiC phase in the printed parts can be further adjusted by varying the process parameters.This approach, based on PBF-EB of refractory carbide powder as a single feedstock combined with in situ infiltration, offers potential applicability in the economical fabrication of various cemented carbide composites (i.e., the energy-consuming production of metal powder as a feedstock can be saved).
To confirm the formation of these large TiC agglomerates, which are believed to result from inhomogeneous sintering, EBSD analysis was conducted on TiC/steel composites fabricated using PBF-EB combined with in situ infiltration.Figure 5 shows the SEM image of the TiC/steel composites and its corresponding inverse pole figures of electron back-scattered diffraction (EBSD-IPF) images.While the SEM image (Figure 5a) shows the connection of TiC particles within the large agglomerate, the grain boundaries between these interconnected particles are distinctly illustrated in the EBSD-IPF image (Figure 5b).These features are not visible in the SEM image.Furthermore, the TiC particles within large TiC agglomerates exhibited a grain size consistent with the observed individual  TiC particles, confirming that these TiC agglomerates were formed through sintering.It has been reported that the dissolution of TiC occurs below its melting point (3065 °C) when the content of carbon (C) and titanium (Ti) is below the solubility limits of TiC in pure molten iron. [52]Additionally, according to Booker's analysis of the pseudo-binary TiC 0.9 -Ti 0 .015 Fe 0 .985 , TiC can dissolve in molten steel above 1475 °C, with its solubility increasing from ≈13 mol% at 1475 °C to 22 mol% at 1600 °C. [53]From the above discussion, it can be inferred that the dissolution of TiC occurs within the liquid AISI 316Ti steel, accompanied by element diffusion and reactions.For further phase analysis in the TiC/steel composite, the distribution of the elements was characterized by EPMA, and the results are displayed in Figure 6.An SEM image (Figure 6a) of the TiC/steel composite shows the presence of a dispersed TiC phase and unidentified phases consisting of variably sized blocks within the steel phase.These phases are characterized by high color intensities of Cr and C, contrasting with the lower color intensities of Ni and Fe elements (Figure 6a).Although AISI 316Ti steel is typically alloyed with strong carbide-forming elements such as titanium to prevent the formation of chromium carbides, carbon plays a fundamental role in carbide formation. [54]The high processing temperatures during the PBF-EB process caused the partial dissolution of TiC in the molten steel, with carbon from the TiC diffusing and reacting with the Cr in the liquid steel, resulting in the formation of Cr-rich carbides (Cr x C y ).This inference is drawn from the fact that the sites with the highest Cr concentrations are coincident with regions of high C and low Fe content.Based on thermodynamic calculations with Thermo-Calc software and HSC Chemistry software, these Cr-rich carbides might be of the M 7 C 3 and M 23 C 6 types, exhibiting high thermodynamic stability (see the following equations). [44,45] þ 7Cr ¼ Cr 7 C 3 Δ r Gðat 25 °C, 1 atmÞ ¼ À166.32 kJ mol À1 Δ r Gðat 1000 °C, 1 atmÞ ¼ À203.75 kJ mol À1  further evidenced by the morphological transition of TiC particles from angular shapes in the raw powder to smoother-edged forms in the TiC/steel composites, as depicted in Figure 5a.
The particle size distribution of the TiC phase in the cross sections of both the initial TiC powder and the printed composites, as determined by image analysis, reflects the behavior influenced by the sintering process.Specifically, the stable value of D (10)  suggests that dispersed particles retained their original size, and the increase in D(90) indicates the formation of larger TiC agglomerates during PBF-EB (Figure 7).It is considered that the size and fraction of the TiC phase in the printed parts are further adjustable when varying the process parameters.This approach, based on PBF-EB of refractory carbide powder as a single feedstock combined with in situ infiltration, can be extended to the fabrication of other cemented carbide composite materials in a relatively economical way (i.e., the energy-consuming production of metal powder as feedstock can be saved).
Figure 8 illustrates the microhardness of the PBF-EB-built TiC/steel composites and the corresponding indentation pictures.The Vickers hardness tests, conducted on individual large TiC particles, revealed a hardness value of 3177.8AE 183.9 HV0.2, which is comparable to the results reported in the literature. [55]y analyzing the crack length after the indentation tests, the fracture toughness (K IC ) was calculated to be 0.208 MPa m 1/2 (see also Equation ( 1)), which is in good accordance with the values of ≈0.277 MPa m 1/2 obtained for single crystalline TiC. [3]The Vickers hardness of the steel phase, measured on the steel substrate to avoid interference from surrounding TiC particles, was determined to be 411 AE 15.7 HV0.2.The compressive strength of the PBE-EBproduced TiC/steel composite is 1036.79AE 99.20 MPa, which is in good agreement with the Hall-Petch relation, [56,57] while Tebe reports that TiC cermets with a grain size of approximately 0.76 μm exhibit compressive strengths of around 2210 MPa. [58]

Conclusion
So far, PBF-EB has been mainly used for processing pure metals.In this work, pure TiC powder was used as a single feedstock for PBF-EB.It was demonstrated that PBF-EB shows a high potential for processing pure carbides.After PBF-EB, owing to the in situ infiltration of the steel substrate, dense and crack-free TiC/steel composites with a high microstructural homogeneity were fabricated.As a single feedstock for PBF-EB, the TiC powder could be simply recycled and reused for the next PBF-EB jobs.The process stability during PBF-EB of the TiC powder was significantly enhanced using high preheating temperatures.Due to the high processing temperature during PBF-EB, the TiC particles were sintered, resulting in an in situ infiltration of the molten steel into the pores between TiC particles.Simultaneously, TiC was partially dissolved in the molten steel, with carbon from the TiC diffusing and reacting with the Cr in the molten steel to form Cr-rich carbides (Cr x C y ).In the microstructure of the additively manufactured TiC/steel composite materials, Cr-rich carbides phase, small unmelted TiC particles with a size of several tens of micrometers, and large spherical and elliptical TiC agglomerates with several hundreds of micrometers in size were detected.The TiC phase showed a Vickers hardness of 3177.8AE 183.9 HV0.2 and a fracture toughness (K IC ) of 0.208 MPa m 1/2 being comparable to that of crystalline TiC reported in the literature.The as-build composites exhibit a compressive yield strength of 1036.79AE 99.20 MPa.Furthermore, the approach introduced in this study shows a high potential to produce cemented carbides, in which the size and fraction of the ceramic phase can be tailored in a targeted way when using appropriate process parameters.The approach reported in this study, combing PBF-EB and in situ metal substrate infiltration, can be extended to the fabrication of other ceramic/metal composites.

Figure 1 .
Figure 1.Schematic of the experimental setup to determine the electrical resistance of the TiC powder at various temperatures in a vacuum furnace.

Figure 2 .
Figure 2. a) SEM images of TiC powder with tiny oxide particulates on the surface.b) Particle size distribution of the TiC powder determined by means of laser diffraction granulometry (LDG) and image analysis, respectively.

Figure 3 .
Figure 3. a) Electrical resistance of raw TiC powder at various temperatures under vacuum conditions (vacuum pressure: 10 À5 mbar).b) Electrical resistance of reheated TiC powder at various temperatures under vacuum conditions (vacuum pressure: 10 À5 mbar).

Figure 4 .Figure 5 .
Figure 4. a) Schematic drawing of the fabrication of TiC/steel composites based on PBF-EB and in situ melt infiltration.b) Additively manufactured TiC/steel bulk material with dimensions of 25 Â 10 Â 5 mm 3 (x Â y Â z).c) Microstructure of the dense as-built specimen composed of TiC (dark phase) and steel (bright phase).

( 4 )Figure 6 .
Figure 6.a,b) Electron probe micro-analysis (EPMA) element mapping images of fabricated TiC/steel composites in the steel phase and the TiC phase.

Table 1 .
PBF-EB processing parameters to produce TiC materials.