Introduction

Traditional manufacturing techniques like casting, injection molding, and forging are called formative processes that need expensive molds or die tools to form the desired shape. A secondary operation like machining gives the final shape and feature to the material. The whole manufacturing cycle is called a subtractive process as the material is removed to make the final product. However, in additive manufacturing, materials are added layer by layer either in liquid, powder, or solid form with or without application of heat to give the required shape to the final product [1].

The Additive Manufacturing (AM) technology was developed in the early 1980s by Hideo Kodama of NMIRI, Japan using photopolymers. A few years later, Charles Hull patented the Stereolithography (SLA) technique which used the principle of curing a liquid photopolymer resin using UV lasers [2]. Slowly, other techniques like Fused Deposition Modeling (FDM), Selective Laser Melting (SLM), Selective Laser Sintering (SLS), and Laminated object manufacturing emerged in the market which kickstarted the commercial adaptability of additive manufacturing techniques [3]. From the usage of polymeric compounds as the traditional material in the 1980s, the technology has grown to a level that metals [4], ceramics [5] and combinations of multi-metallic parts [6, 7] are being manufactured to make functional components with detailed features and marketing ready aesthetics.

Most of the AM techniques use a heat source (laser beam, electron beam, UV light or electrical resistance) which is focused onto a bed of raw material (plastic, metal or ceramic) to melt it from its initial state (powder, liquid or wire) to the final shape. Due to its wide variety of applications, the process is currently being used in automobiles [8], aerospace [9], biomedical [10], consumer electronics [11], and jewelry industries [12]. Functional components made of metals and alloys like aluminum [13, 14], steel [15], martensitic stainless steel [16], stainless steel [17, 18], maraging steel [19], titanium [20], Inconel [21, 22] and cobalt-chrome-based superalloys [23] are being manufactured by this technique.

Unlike other metals, tungsten carbide hardmetal parts are manufactured by the powder metallurgy technique that can only produce parts with limited geometrical complexity. The process involves different complex stages like powder processing, mixing and milling of powders, pressing in die tool to get the required shape, dewaxing, liquid phase sintering, and post-sintering operations like grinding and blasting. Process parameters at each stage are controlled carefully as any change in the variable would influence the subsequent processes and final quality. Especially, sintering is one of the critical steps with numerous control variables like temperature, pressure, time, atmosphere, and rate of heating and cooling which can alter the microstructure and mechanical properties of the final product. Therefore, until a few years ago it was assumed that either it is impossible to manufacture tungsten carbide parts by the AM process or the components cannot meet the stringent quality prerequisite to meet the challenging functional requirements. However, researchers have overcome the challenge and succeeded in finding techniques to manufacture tungsten carbide hardmetal components with properties close to their conventional counterparts.

Hence, the objective of the paper is to summarize the research findings in the area of additive manufacturing of tungsten carbide parts with a focus on SLM, SLS, and BJ3DP. The process parameters, metallurgical, and mechanical properties of the parts manufactured, and the challenges faced in each process are discussed in detail. The successful applications and scope for future studies are also discussed.

Tungsten Carbide (WC)

Tungsten carbide is one of the hardest materials known to man that is being used in cutting tool manufacturing, wear applications, mining, construction, and oil & gas exploration industries [24, 25]. With a hardness value of 9–9.5, WC lies just below the diamond in the Mohs hardness scale [24, 25]. The WC based materials are manufactured by cementing the hard tungsten carbide (WC) particles in a tough cobalt (Co) matrix (hence it is commonly referred to as “cemented carbide”) which makes it a metal matrix composite [26]. Cemented carbide with only tungsten and cobalt are called as the straight grade. Sometimes, carbides, nitrides, or carbonitrides of the transition metal group (Group IVB, Group VB and Group VIB metal carbides) are alloyed to make mixed grades to improve the properties and cater to application-specific requirements [27].

The superior hardness, density, thermal resistance and fracture toughness provided by tungsten carbide-cobalt combination makes the material an ideal candidate in challenging metal cutting and wear applications [28]. More than 65% of the material is consumed by cutting tool manufacturing industries, followed by oil and gas exploration and mining industries which consume around 15%, whereas construction and wood processing industries share the rest [29]. The very properties that make the material suitable for such demanding applications also make it difficult to manufacture using conventional techniques like casting, forging, or machining. Hence, powder metallurgy technology is used to manufacture WC-based parts that require expensive die-tooling and sintering setups. The various densification stages and microstructure evolution as observed by Oliver et al. [30] in WC-Co composite manufactured by conventional powder metallurgy technique is shown in Fig. 1.

Fig. 1
figure 1

Schematic overview of microstructure evolution and corresponding fracture images (SEM) during sintering of WC-Co cemented carbide [30] (reproduced from [30], with permission from Elsevier)

Full size image

Even with such advancements, the technique is time-consuming, expensive, subjected to considerable design restrictions, and challenging, especially for prototyping and small lot size requirements. So, the hardmetal manufacturing companies are constantly in search of a process that can minimize the manufacturing time and resources required to meet the application-specific tool requirements (also called as special tools).

Hence, industrial and academic researchers have tried additive manufacturing processes like Selective Laser Melting (SLM), Selective Laser Sintering (SLS), Laser Engineering Net Shaping (LENS), Direct Laser Fabrication (DLM), Binder jet 3D printing (BJ3DP) and 3D gel-printing (3DGP) to manufacture tungsten carbide-based hardmetal components. The technique gives exceptional benefits in manufacturing complex, lightweight tools, reduce manufacturing lead time for special tools, optimize coolant channel design, and chip flute, which is otherwise impossible with conventional techniques [31]. Consequently, AM may not completely replace the traditional method of manufacturing WC-based parts but can serve as a complementary process for specific requirements.

Literature review shows that most of the past studies focus on SLM, SLS, and BJ3DP, with very few works on other processes. Essentially, SLM and SLS are two instantiations of the same concept (Powder Bed Fusion technique) with very little differences. The SLM process to print metallic parts uses a laser with higher intensity to achieve a full melt of the powder, whereas SLS uses a comparatively lower laser intensity to fuse or bind the particles together on a molecular level. However, in the case of AM of WC-based parts, SLM and SLS are fundamentally the same. Nevertheless, as many researchers have used the terms separately the original phrase stated by the researchers is used in the article. BJ3DP is a non-laser, non-heat-based process that uses a binder liquid to join the powder particles followed by a post-process sintering.

Additive Manufacturing of WC-Based Hardmetals

The different AM techniques used by the researchers to manufacture tungsten carbide-based hardmetal parts and the key process variables like laser power, spot size, scan speed, hatch spacing, layer thickness, and laser type used for the study are consolidated and shown in Table 1.

Table 1 Summary of literature on additive manufacturing of tungsten carbide
Full size table

Selective Laser Melting (SLM)

Selective laser melting (Fig. 2) is one of the most successful AM techniques in the manufacturing of high precision and high-quality functional components. Like other powder-based techniques, the SLM process has a build platform that is enclosed in a chamber of inert gas and a powder delivery platform. The movement of the recoating roller from the delivery platform deposits the powder mixture onto the build platform. By tracing the 3D CAD model, SLM uses a laser with higher intensity to melt the evenly distributed and newly deposited powder to solidify it. When the laser beam scans the powder, it absorbs the energy by bulk coupling mechanism and powder coupling mechanism. Initially, the absorption of heat energy is through a narrow layer of individual powder particles, which further flows to the center of the particle until a steady-state is reached. Finally, heat energy rapidly increases the operating temperature and melts the powder to form a molten pool [32]. Once the first layer is built, the build platform moves down to a predefined distance (which is called as layer thickness) for the next layer of powder to be deposited. The process continues for individual layers of the sliced CAD model and builds the part surfaces, one over the other until the final part is built.

Fig. 2
figure 2

Schematic diagram of the selective laser melting (SLM) process [62] (reproduced from [62], with permission from MDPI)

Full size image

Though the technique looks simpler, it is imperative to precisely control the process parameters to avoid possible defects like pores, micro-cracks, layer delamination, warping, and degradation in the final property of the manufactured part [33]. Especially, proper control of laser parameters is the key to achieve high-quality parts. For instance, the powder material which is directly below the center of the laser beam absorbs more energy than the powder which is below the periphery. This leads to the formation of an uneven melt pool and promotes spattering [63, 64].

Selective Laser Sintering (SLS)

According to the ISO/ASTM 52911–1:2019 standards both SLM and SLS are categorized as Laser-based powder bed fusion of metals (PBF-LB/M). The SLS technique uses a mixture of high (structural or base powder) and low melting point (binder powder) metal powders, whereas SLM can operate without the binder powder [45]. In general, SLM and SLS are similar techniques as both the processes use a laser source to heat and solidify the metal powders [65]. The real difference is in the binding mechanism as SLS uses a comparatively low energy beam to bind or fuse the particles on a molecular level to form a solid mass without melting it to the liquefaction state [66].

Sintering of powders is a four-stage process that takes place at the temperature at which the binder melts. In the first stage, the light energy from the laser source gets converted to heat energy. In the second stage, the heat energy is absorbed by the powder bed which melts (third stage) the binder metal powder. In the final stage, the cooling of sintered powder occurs [45]. The penetration of the laser beam depends on various parameters like the wavelength of the laser beam, the properties of metal powder, and the size and shape of the powder particles [45].

The laser energy used in the process is carefully adjusted so that only the powder with a low melting point is targeted without affecting the structural powder [54]. As the powder is not melted to its liquid state, the chance of the final product being porous is highly likely. Hence, some researchers have used the infiltration of low-melting-point metals like copper or bronze as a post-consolidation stage to reduce porosity and enhance the properties [50, 67, 68]. Generally, it is recommended to avoid the infiltration technique as it increases the cost, complexity, and may not effectively improve the properties of the final product [68]. However, the sintered density can be increased by using a binder with smaller particle size than the base material, as a binder with bigger particle size could result in partial melting [54]. HIPing (Hot Isostatic Pressing) is another post-consolidation technique to get parts with full density and controlled porosity [69].

Binder Jet 3D Printing (BJ3DP)

Binder Jet Printing also called Binder Jet 3D printing (BJ3DP) is a technique that uses a liquid binding agent to join the powder particles instead of a heat source (laser or electron beam). Hence, thermal-induced complexities like residual stress, deformation, decarburization of WC, undesirable grain growth, and formation of eta-phase carbides can be avoided [56, 70]. The schematic diagram of a binder jet 3D printing process is shown in Fig. 3.

Fig. 3
figure 3

Schematic of binder jet 3D printing process [71] (reproduced from [71], with permission from Elsevier)

Full size image

The various stages involved in the binder jetting process to manufacture tungsten carbide-based hardmetal parts are shown in Fig. 4.

Fig. 4
figure 4

Stages of binder jet printing [72] (reproduced from [72], with permission from Elsevier)

Full size image

The first step is the preparation of powder from raw material with the required characteristics. The morphology, bulk density, and flowability of the powder (hall flow) used in the process are critical in achieving a good quality part with the required mechanical properties. In the second step, the powder is deposited over a preheated powder-bed using rollers, and binder droplets are applied through the print head over the designated part of the powder bed which glues the powder particles together to form a given layer. The binder consists of a solvent that evaporates during curing [58]. Different organic binders that contains one or more polymeric materials like polyvinylpyrrolidone or polyethylene glycol are used. A new layer of powder is added over the existing layer, and the process continues until the required shape is printed. Even with a binder, the green part may not have enough strength and hence, the printed part is cured by heating it to a certain temperature (150–200 °C) for around 4 h to increase the green strength. Once the part is built, the loose powder around the part is removed. The printed preform further undergoes debinding and post-processing treatment like sintering or vacuum sintering under a hydrogen or argon atmosphere as per the conventional powder metallurgy process [57, 72]. The overview of the advantages and disadvantages of different AM processes used in the manufacturing of WC-based parts [73] are shown in Table 2.

Table 2 Advantages, disadvantages and major research groups working on different AM processes [73] (reproduced from [73], with permission from Springer)
Full size table

Besides, a few studies on Selective Electron Beam Melting (SEBM) [74], 3D Gel-Printing (3DGP) [75], Fused Filament Fabrication (FFF) [76], Laser Engineering Net Shaping (LENS) [77], and Thermoplastic 3D Printing (T3DP) [78] of WC-based hardmetal samples were also reported which are not considered in this review.

AM Process Parameters

In the AM process, a track (also called bead) is an elementary unit that represents a single scan line and a combination of more than one tracks forms a layer, whereas layers stacked one over the other forms the final complex volumetric part. Inspecting the cross-section of a single track is often enough to check and understand the quality of a printed part [37]. Controlling the process parameters to achieve the required geometrical characteristics of the track and good adhesion between the layers are the main criteria to control dimensional accuracy and mechanical properties. The effects of varying each process parameters as reported by different researchers are discussed in this section. As SLM/SLS is heat-based and BJ3DP is a non-heat-based technique, the parameters of both the processes are discussed separately.

Process Variables in SLM/SLS

Each variable like scanning speed, laser power, scan line spacing (or hatch spacing), laser energy density, and layer thickness has a specific effect on the final properties of the manufactured product. Generally, the process variables affect the amount of laser energy impinges on the powder. Reports claim that CO2 laser is most favorable in AM of oxides of ceramics, whereas Nd:YAG (Neodymium-doped yttrium aluminium garnet), Yb:YAG (Ytterbium-doped Yttrium Aluminum Garnet) or Nd:YVO4 (Neodymium Doped Yttrium Orthvanadate) lasers are suitable for metals and carbide ceramics due to the favorable wavelengths [79].

Low laser power, high-speed scanning or high layer thickness could lead to lack of energy to melt the powder which results in balling (formation of small beads due to surface tension and inadequate wetting of the previous layer) and formation of small melt pools, whereas, high laser energy could lead to vaporization of the powder material and results in pores due to bubble formation [80]. The vaporized material could often shield the laser window and prevent it from reaching the powder [81]. Researchers also claim that a uniform profile laser beam is beneficial over the Gaussian beam profile [82]. The high laser intensity at the center of the Gaussian beam could vaporize the metal resulting in energy imbalance [83].

Davydova et al. [84] investigated the effect of scanning speed (100 to 300 mm/s) on SLM of boron carbide particles coated by a cobalt-based metal layer. They have reported that at a laser power of 60 W and a layer thickness of 50 μm, the optimum scanning velocity to get tracks with constant dimensions is 100 mm/s as shown in Fig. 5. However, cracks at the interface between the carbide particles and metal matrix were seen and the porosity of the final printed part measured using the Archimedes method was close to 37%.

Fig. 5
figure 5

Cross-section of single-track at different laser scanning velocities in SLM of boron carbide particles coated by a cobalt-based metal layer (Laser power 60 W, Layer thickness 50 μm) [84] (reproduced from [84], with permission from Elsevier)

Full size image

Gu and Meiners [32] performed SLM technique on W-Ni-graphite powder to prepare in situ WC/Ni2W4C(M6C) cemented carbide parts manufactured at different scanning speeds (0.8 m/s, 1.0 m/s and 1.2 m/s) and studied the microstructure. Ni2W4C (eta phase carbide) was seen at all three speeds. However, the percentage of eta was higher at 1.2 m/s when compared to the lower speeds, but small-sized pores were formed at a lower scanning speed. Khmyrov et al. [35] have shown (Fig. 6) that the width of single-track increases when the scan speed decreases, as the laser energy per unit length increases.

Fig. 6
figure 6

Single re-melted beads at powder layer thickness 100 μm, laser power 80 W, and scan speeds 10 to 50 mm/s [35] (reproduced from [35], with permission from Elsevier)

Full size image

In the samples printed by SLM process using the commercially available T311 powder (86% WC, 10% Co, and 4% Cr), Campanelli et al. [39] varied the scanning speed in the range 40–100 mm/s to control the energy density of the laser beam. Reduction in scan speed increased the energy density, which decreased the surface roughness and porosity. Increasing the scan speed resulted in a linear decrease in the dimensions of the sintered zone (thickness and width). Similarly, an increase in laser power increased the thickness and width of the sintered zone [45]. However, in the study by Khmyrov et al. [35], they have reported that an increase in laser power improved the surface finish of WC-Co parts manufactured by the SLM process. The average surface roughness (Ra) at 30 W, 43 W, and 50 W were 2.1 μm, 1.5 μm, and 1.2 μm, respectively.

Studies have proven that the density and porosity of AM parts depend on the hatch distance (or scan spacing), the composition of the powder, layer thickness, and laser power [85]. Scanning strategies like single-scan, repeated scan or cross-scan must be selected based on the material and it has a significant effect on the surface roughness and the mechanical properties of the printed part [79]. The distance between the two vectors or melt lines is the hatch distance and it is typically smaller than the spot size of laser. Researchers suggest that a 30% overlap between the hatch spacing and multidirectional scanning strategy (Fig. 7) can produce parts with minimal pores and avoid major deviations from the desired shape [37].

Fig. 7
figure 7

a Schematic diagram for hatch spacing and b scanning strategy [37] (reproduced from [37], with permission from MDPI)

Full size image

Other parameters like atmospheric control inside the build chamber and bed temperature also have a minor impact on the quality of the final product [86]. When compared to polymer powders, metallic powders are highly susceptible to contamination as they are highly reactive to moisture, absorbed gases, and formation of oxide and nitride layers which can affect the microstructure and degrade the properties [81].

Apart from the laser processing conditions, properties like purity, morphology, and size of the metal powders also play a critical role in defining the final properties of the manufactured parts [87, 88]. The size of powder particles affects the efficiency of melting. Generally, powders with large particle sizes require more laser energy to melt, whereas powders with small particle sizes can agglomerate. Besides, the powder particles with good flowability are essential to achieve a consistent layer thickness and uniform absorption of the laser energy. WC-Co powder with spherical granules is suitable to get high-density final products (as their apparent density is higher) when compared to the granules that are irregularly shaped [38] as shown in Fig. 8.

Fig. 8
figure 8

Morphology of WC-Co granules made by a spray drying (spherical shape), and b cold pressing and crushing (irregular shape), and microstructure of parts made by c spherical granules, and d irregular shaped granules at identical SLM parameters [38] (reproduced from [38], with permission from Elsevier)

Full size image

Some researchers have claimed that coating the powders with the binder element could improve the density and assist in uniform distribution of the binder [68, 84, 89]. Moreover, the absorptivity of the powder particle is different from the absorptivity of their bulk materials, which further increases the complexity of the process [90]. The scanning electron microscope (SEM) image of boron carbide particles coated by a cobalt-based layer used by Davydova et al. [84] is shown in Fig. 9.

Fig. 9
figure 9

SEM images of boron carbide particles coated by a cobalt-based layer a particle cross-section, b powder general view [84] (reproduced from [84], with permission from Elsevier)

Full size image

Researchers have shown that low laser power, high scanning speed, and high layer thickness result in a lack of energy to melt the powder, whereas high laser power could promote cracks and depletion of cobalt. Increasing the scanning speed also increases the chances of porosity, whereas high hatch spacing results in delamination and distortion of parts from the required shape [43].

The extensive literature review (Table 1) shows that each researcher has used a different set of process parameters which shows the infancy of the AM process in manufacturing WC-based parts. However, it is a challenge to have one set of optimized parameters for different WC and Co mixtures. Even in the traditional powder metallurgy technique, sintering is one of the critical stages that is carefully controlled to eliminate pores and get the required microstructure. Similar precision is required in optimizing and monitoring the laser parameters in manufacturing WC parts using AM techniques. Maintaining the right carbon balance throughout the process is important to avoid the formation of eta-phase or graphite [29]. Controlling the grain size in AM parts would be another challenge that is done by adding grain growth inhibitors like V, Ti, Cr, Ta, Mo, and Nb in the conventional powder metallurgy process [29].

Hence, a systematic design of experiments studies can be performed as a preliminary trial run to print individual tracks with different process parameters, and selecting the optimized parameter for the powder mixture that is being used could be a viable way to get defect-free parts [33, 49, 50].

Process Variables in BJ3DP

As BJ3DP is a non-heat-based AM technique, the process variables used are slightly different from SLM/SLS except the characteristics and quality requirement of the initial powder mixture used. The properties of binder (physical and chemical properties), layer thickness, binder saturation, the volume fraction of binder, drying time, densification process and post-process sintering are some of the critical variables that must be carefully controlled in BJ3DP process to manufacture parts with higher build quality.

Several studies have been performed to identify the effect of process parameters on the quality of parts made by BJ3DP process and they have concluded that binder saturation and layer thickness are critical to the strength and dimensional accuracy [91, 92]. Binder saturation is the ratio of binder used to the volume of voids in each layer of the powder bed [72]. Enneti and Prough [57] studied the effect of binder saturation and layer thickness of WC-12%Co hardmetal parts made by BJ3DP process. They have reported that the strength of the samples increased when the binder saturation ratio increases. The transverse rupture strength (TRS) of the parts printed with different binder saturation and layer thickness is shown in Fig. 10.

Fig. 10
figure 10

Green strength of printed samples at different saturation levels and powder layer thickness [57] (reproduced from [57], with permission from Elsevier)

Full size image

However, too much binder can result in the collapse and deformation of the parts, which could affect the dimensional accuracy and strength [72]. Researchers have also reported that when the binder saturation increases, the accuracy of the part decreases [93, 94]. Hence, it is critical to find the optimal binder saturation ratio to get the required strength and dimensional accuracy. Setting time of binder is also important to improve the green strength. Increasing the binder set time from 7 s to 14 s assisted in increasing the green strength of WC-12%Co samples by more than 50% [57].

Hot isostatic pressing (HIP) can be applied after post-sintering or during vacuum sintering to increase the density. The green parts that were vacuum sintered or HIP at 1450–1500 °C for 0.5–1 h produced parts with 99.3% of theoretical density and transverse rupture strength of around 3040 MPa [95]. Besides, the green strength of the printed parts can be enhanced by using particles with a smaller size which increases the contact proximity, contact points with adjacent powder particles, and packing efficiency. The higher green density improves the density of the sintered part which helps in reducing the sintering temperature (which in turn prevents grain growth) and avoids additional post-processing techniques like HIPing. Pores and cracks that are formed due to gas expansion and gas pocket formation during thermal debinding can be avoided by applying binder only to the outer surface [96].

The current knowledge on AM of WC-based hardmetal parts by BJ3DP and the effect of process parameters on the mechanical properties and microstructural changes are scarce due to the limited published literature on the subject. However, the similarity of BJ3DP process to the conventional powder metallurgy technique makes the process a potential candidate for future researchers.

Properties of Additive Manufactured Tungsten Carbide

Formation of cracks and pores, lower density, changes in microstructure, dimensional inaccuracies and poor mechanical properties are the key challenges irrespective of the AM technique (SLM, SLS, or BJ3DP) used. As WC-based parts are highly sensitive to minuscule microstructural and dimensional irregularities [97, 98], it is unlikely that the AM parts can meet the quality required for functional and commercial applications with the current technology. Although the process parameters have a significant effect on the properties of the final product, currently there is a lack of systematic research to relate all the factors to arrive at an optimized processing condition. The microstructural and mechanical properties of parts made by AM technique reported by researchers are discussed in this section.

Cracks and Pores

Literature shows that cracks and pores are the most commonly seen quality concerns in AM of WC-based parts [32,33,34,35, 37,38,39,40,41,42,43, 48, 54]. Cracks are primarily formed due to the evaporation of binder, formation of ternary phases and residual stress [34, 41, 52, 99], whereas pores are formed due to improper energy density, poor distribution of powder, moisture, and evaporation of binder [38, 73, 100]. A large amount of microscopic and elongated pores and poor bonding between the layers were seen in 85WC-5C-10Ni cemented carbides as seen in Fig. 11. Increasing the scanning speed increased the size and distribution of pores and reduced the relative density [32].

Fig. 11
figure 11

Size and distribution of pores in 85 W-10Ni -5C cemented carbide for increasing scan speed [32] (reproduced from [32], with permission from Elsevier)

Full size image

Increasing the layer density and powder bed density by using rollers to apply compressive force can decrease the chances of cracks and pores [101]. Ku et al. [43] have seen uniformly distributed pores as big as 0.5 mm which are connected by cracks and lower density in WC-10Fe-Ni-Zr cemented carbide parts. They have correlated the regularity in the distribution of pores to the layer thickness and reported that adjusting the layer thickness could reduce porosity. Chen et al. [38] have reported that using a spherical shaped powder granule can significantly increase the relative density of the final product and reduce the chances of formation of pores, cracks, and voids when compared to irregularly shaped granules. The poor powder flowability of submicron size powder granules may reduce the reproducibility of powder layers and increase the chances of porosity [43]. Circular pores as seen in Fig. 12a are indications of entrapped gases, which are again attributed to the poor layering [43, 102]. The porosity within the WC-grains as shown in Fig. 12b could be due to the sintering mechanism and high grain boundary mobility [43].

Fig. 12
figure 12

SEM cross-section images showing a circular porosity and b WC grains within the microstructure [43] reproduced from [43], with permission from Springer)

Full size image

In the case of BJ3DP process, the particle size affects the size and distribution of voids. Fine powders tend to form lumps that leave macro-voids inside the layer. The void could either alter the direction of the flow of binder droplets or completely stop it from flowing into the pores [103] as shown in Fig. 13.

Fig. 13
figure 13

Droplet penetration in homogeneously-distributed and heterogeneously-distributed powder beds [103] (reproduced from [103], with permission from Elsevier)

Full size image

Hence, sometimes powders with a mixture of fine and coarse particles are used so that the fine powder particles can fill the gaps formed by coarse particles [72]. Flon [58] suggests that fine powder particles of <10 μm diameter with 45–65 Vol.% increase the locking of binder liquid that is critical to get a high green strength. Studies have also shown that the printability of the binder also depends on the droplet-formation mechanism [104, 105].

According to past studies, the large amount of temperature gradient formed in the AM process leads to residual stresses and induce cracks [34]. Cracks are formed predominantly due to thermal shock/stress brought by the localized heating nature of the laser source and repeated heating and cooling due to layering and are very difficult to control [35, 43]. The effect of thermal gradient and stress can be reduced by preheating the powder bed [35], but preheating may reduce the flowability of the powder.

Nonuniform thermal fields result in thermomechanical stresses and if the stresses exceed the mechanical strength of the material it forms cracks [37]. The cracks formed on a single track and monolayer are shown in Fig. 14.

Fig. 14
figure 14

Cracks on the surface of a a single track and b the monolayer of WC-6%Co at Laser power = 120 W and scan speed = 300 mm/s [37] (reproduced from [37], with permission from MDPI)

Full size image

A high amount of hardmetal phase like WC is also responsible for the formation of cracks in parts manufactured by the AM process [40]. Khmyrov et al. [40] have succeeded in fabricating crack-free parts with a powder mixture that contains 25wt%WC. However, increasing the WC content to 50% resulted in cracks. Researchers have reported that optimizing the scanning technique, reducing the laser energy density, and preheating the build platform are a few of the strategies that can reduce the formation of cracks [33, 40, 43]. The high solidification rate of the molten pool and difference in the thermal expansion coefficient of base metal (WC) and binder (Co, Ni, or Fe) results in the formation of heat-induced stresses which initiate the formation of cracks when the part is subsequently cooled [41]. Hence, lowering the laser intensity could reduce the formation of thermal cracks, but it could result in residual porosity [33].

To get mastery over any manufacturing technique (be it conventional or AM), avoiding macroscopic defects like cracks and pores is the first and foremost step. The next step is to optimize the process to address the microstructural and metallurgical requirements. But up to now, researchers have not succeeded in manufacturing a pore and crack-free WC-based part by any of the AM processes whose mechanical properties are comparable to that of parts made by conventional techniques [37, 41].

Density

The ratio of the density of part manufactured by AM technique to the theoretical density of the bulk material is the relative density and it has a direct relationship with porosity. Insufficient bonding between the layers along with microscopic and macroscopic pores results in a considerable reduction in density [32]. Gu et al. [32] have achieved a high relative density by adjusting the scanning speed. A density of 81.7% was achieved at a laser scanning speed of 1.2 m/s, whereas a comparatively higher value of 94.2% and 96.3% of the theoretical density was achieved by reducing the scanning speed to 1 m/s and 0.8 m/s respectively. Uhlmann et al. [33] have reported that using high laser power, low scanning speed, low hatch spacing, and low layer thickness can increase the density. They have shown that increasing the laser power improves the continuity of the melt pool and lower scanning speed expands the area of melt pool which improves the density. Similarly, reducing the hatch spacing and layer thickness reduces the overlaps between the scan lines, which reduces porosity and increases the density [33].

The size and shape of the powder granules also contribute to the density of the final product. A relative density of around 86% (11.74 g/cm3) was achieved on parts printed using irregularly shaped powders and the density increased to 96% (13.14 g/cm3) when the spherical powder was used [38]. Researchers have shown that parts that have lower density can be re-densified by post-processing techniques like secondary metal infiltration, sintering, or hot isostatic pressing (HIP) [33, 43, 45]. The density can be further increased by partially melting the tungsten carbide grains with a high-power laser [50]. However, some reports suggest that post-process sintering performed above the melting point of binder cobalt neither increased the density nor the strength of the parts [45]. However, Ku et al. [43] have shown that the laser energy density can be adjusted by varying the laser power, scanning speed, and hatch spacing which can influence the density of the final part. Parts with high density are obtained at high laser energy density and a slight increase in density was seen after post-process HIPing as shown in Table 3.

Table 3 Part density variation for different laser energy before and after HIPing [43] (reproduced from [43], with permission from Springer)
Full size table

Uhlmann et al. [33] have reported that though higher energy density increases the part density, it could result in cracking and delamination. Chen et al. [38] have also shown that laser power input has the highest contribution to the final density. Powder mixture with high cobalt content can reduce the tendency to crack, but the parts may not meet the hardness and wear resistance requirements [33]. The infiltration technique to increase the porosity was performed by Cramer et al. [59,60,61] in parts printed by BJ3DP process. After printing, debonding, and sintering the infiltration process was carried out to improve the density, hardness, and fracture toughness. But even after infiltration the cracks and pores could not be eliminated. Even with a lower density (60%) the SLS parts made with nano-grains showed higher hardness and wear resistance than the micro-grained conventional parts with 95% density [55].

Enneti et al. [106] printed WC-12%Co samples by BJ3DP process using WC (1.1–1.4 μm) and Co powder (1.2 μm). With a binder saturation of 45%, they have achieved a green density of 42% of the theoretical density. After debinding, some of the samples were vacuum sintered, whereas, the rest were pressure sintered. They observed low hardness in the vacuum sintered parts, whereas very high hardness and near theoretical density were achieved in the pressure sintered samples. External pressure applied during the sintering reduced the porosity significantly but could not be eliminated as shown in Fig. 15.

Fig. 15
figure 15

Unetched microstructures of samples a vacuum sintered at various temperatures b pressure sintered at 1485 °C [106] (reproduced from [106], with permission from Elsevier)

Full size image

Literature shows that researchers have achieved comparatively high density in WC-based hardmetal parts manufactured by additive techniques, but the density is still lower than the traditional liquid phase sintered parts [38]. Even the part manufactured with 98% relative density is affected by deep cracks [33].

Microstructure

The microstructure of cemented carbide parts is critical in getting the required mechanical properties which can be achieved by controlling different variables like hard phase, metallic binder phase, alloying elements, and processing parameters [29]. Past studies show that the two critical non-destructive quality control techniques like magnetic saturation and coercive force measurement [26, 107,108,109] were not investigated in WC-based parts manufactured by AM technique.

Gu and Meiners [32] have reported that scanning speed has a significant impact on the morphology and microstructure of W-10%Ni-5%C cemented carbide manufactured by SLM. Block shaped and elongated tungsten carbide grains (Fig. 16 (ii)(a)) were seen at a low scanning speed of 0.8 m/s, whereas, triangular-shaped grains (Fig. 16 (ii)(d)) was seen at 1 m/s. A ring-shaped interfacial layer was seen as shown in Fig. 16 (ii)(f) when the scanning speed was further increased to 1.2 m/s [32]. The changes in the interfacial energies of {101 ̅0} and {011 ̅0} facets and the resultant variation of truncation factor obtained at different scanning speeds could be the driving factor for the variation in morphology. WC crystal grows layer-by-layer by stacking of (0 0 0 1) basal along <0 0 0 1 > orientation to get the final 3D form as shown in Fig. 17 [32].

Fig. 16
figure 16

(i) XRD spectra and (ii) microstructure of 85 W-10Ni -5C cemented carbide for increasing scan speed [32] (reproduced from [32], with permission from Elsevier)

Full size image
Fig. 17
figure 17

Structure of WC grains showing a (0 0 0 1) basal, {10\( \overline{1} \)0}, and {01\( \overline{1} \)0} prismatic facets and b Growth of polygonal and triangular-shaped grains via different mechanisms [32] (reproduced from [32], with permission from Elsevier)

Full size image

The XRD-EDAX analysis performed on the samples identified the ring-shaped layer (Fig. 16 (ii)(f)) as M6C which is η-phase carbide (Ni2W4C). Carbon deficient regions attract cobalt and result in the formation of η-phase carbide as shown in Fig. 18. Based on the amount of carbon deficiency it may either appear as a uniformly distributed fine particles are as clusters [107, 108].

Fig. 18
figure 18

α, β, and ƞ-phase in sintered WC-Co hardmetal part [29]

Full size image

The volumetric fraction of η-carbide increased when the scanning speed was increased [25]. The XRD spectrum (Fig. 16(i)) detected around 1.9%, 3.5%, and 6.0% of Ni2W4C in the parts made at scanning speeds of 0.8 m/s, 1.0 m/s and 1.2 m/s, respectively [32]. The variations in carbon activity in the melt pool during heating and subsequent cooling [32] and variations in sintering temperature [110] could be the underlying reason for the formation of η-carbides (W2C, Co3W, Co6W6C, and Co3W3C) [32]. η-carbide was seen by other researchers as well [34, 36, 40, 44, 54].

Li et al. [44] have attributed the formation of η-phase to the diffusion of Fe and Ni from the base plate. They have also seen that the chemical composition and the distribution of η-phase varied along the building direction. Hence, the presence of the phase could be due to the temperature distribution within the single track and variation in temperature between the layers during the layering process [36]. The temperature distribution across the thickness of a single track and during subsequent layering is shown in Fig. 19.

Fig. 19
figure 19

Temperature distribution over a single track and b layer [38] (reproduced from [38], with permission from Elsevier)

Full size image

The interaction of laser power between the bottom and adjacent layers (arising out of layer thickness, hatch distance, laser power, and scan speed) results in reheating of neighboring layers. The repeated heating and cooling of neighboring layers could result in grain growth and formation of η-carbides [73]. The ternary η phase (Co3W3C) was also seen in WC-Co components manufactured using binder jet printing of WC infiltrated with Co [61]. However, η-phase was not seen in WC-13%Co samples printed by the SEBM process [74]. The rapid heating and cooling of layers in SEBM could be the reason for the clear tungsten carbide-cobalt interface without the inclusions of the third phase, as the growth of a reaction product (diffusion-controlled process) needs enough time to occur [111,112,113]. Studies have shown that η-phase acts like a hard spot and deteriorates the performance of tungsten carbide tools and hence must be avoided [29, 107, 114].

The phenomenon of grain growth between the layers was seen by Ku et al. [43] as shown in Fig. 20. They have observed a fine-grained structure on one side of the layer and a coarser structure on the other side.

Fig. 20
figure 20

Cross-section SEM images of layers at a low magnification and b zoomed-in to highlight evidence of layering in the microstructure [43] (reproduced from [43], with permission from Springer)

Full size image

A significant increase in the mean grain size was also seen when the scanning speed is increased. From the initial mean grain size of 2.5 μm, a linear increase to around 17.5 μm, 19.6 μm, and 40.8 μm was seen at scanning speeds of 0.8 m/s, 1.0 m/s and 1.2 m/s, respectively [32]. Enneti et al. [106] have achieved uniformly distributed grains with diameter 1.4–2 μm in the WC-12%Co part printed using BJ3DP process. However, clusters of coarse grains with around 20 μm size were seen due to some unknown phenomenon as shown in Fig. 21.

Fig. 21
figure 21

Clusters of coarse grains seen in BJ3DP process of WC-12%Co [106] (reproduced from [106], with permission from Elsevier)

Full size image

Manufacturing of WC-Co parts using nano or submicron grained structure is difficult in conventional powder metallurgy technique as sintering is a 15 to 20-h long process which is favorable for grain growth [35]. Khmyrov et al. [36] have reported that the typical high cooling rate of AM techniques can be used to manufacture parts with a submicron and nano-grain structure that can give superior performance when compared to parts manufactured by conventional liquid phase sintering using micro-sized grains. The parts manufactured by the SLM technique using nano-powder of 10 nm size have seen a grain growth to around 180 nm, and a combination of fine- and coarse-grained structure was seen between the layers [34]. In general, the microstructure of cemented carbide parts made of AM techniques like SLM and SLS is finer when compared to parts made of conventional techniques [35].

The changes in metallurgical and mechanical properties observed by researchers are consolidated and shown in Table 4.

Table 4 Summary of metallurgical and mechanical properties
Full size table

Hardness, Toughness and Wear Resistance

The hardness, toughness, and wear resistance of WC parts are dependent on the microstructure and amount of binder used in the manufacturing. The hardness of cemented carbide parts is usually measured by the Vickers hardness technique as per the ISO 3878 procedure [108].

Grigoriev et al. [37] have compared the properties of WC-6%Co parts made with an average tungsten carbide particle radius of 400 nm and cobalt particle radius of 40 nm against the parts made by conventional technique. They have reported a 1.6 times higher hardness in the parts made by the SLM technique. However, the fracture toughness and wear rate reduced by around 22% and 23%, respectively as shown in Table 5. The increase in hardness could be due to the nanograin sized initial powder particles used.

Table 5 Properties of SLM and conventional WC-6%Co parts [37] (reproduced from [37], with permission from MDPI)
Full size table

However, a slightly lower hardness was seen in WC-20%Co made by the SLM process when compared to the conventional hot-pressed samples, which was attributed to the low density and coarser WC grains in the samples manufactured by SLM [38]. Kumar [52] attributed the increase in hardness to the formation of M6C and M12C eta phases. Heat treating WC-17%Co SLM samples to 600 °C increased the hardness from 55.2 HRC to 64.4 HRC (non-heat treated SLM part). However, increasing the heat treatment temperature beyond 600 °C resulted in a slight decrease in hardness. A slight improvement in wear resistance and fracture toughness was also seen.

It has been reported that hardness depends on the size of carbide particles used in manufacturing. The intermittent grain growth between the layers resulted in higher hardness in the region with fine-grained structure and lower hardness in the coarser zone [34]. Additionally, the nanophase samples showed significantly higher hardness than the standard samples in both zones as shown in Table 6.

Table 6 Microhardness of nano and conventional WC-12%Co [34]
Full size table

Cottle et al. [55] have proved that WC-15%Co parts made using nano-grains (20–50 nm) by SLS showed higher hardness and wear resistance than the WC-15%Co parts made using micro-grains (1.3 μm) by conventional technique. Ghosh et al. [54] used the Taguchi optimization technique to evaluate the effect of SLS process parameters on microhardness. They have reported that cobalt content has the highest significance of around 40%, followed by pulse energy (21%) and layer distance below the focal plane (10%).

There is very few literature that compares the mechanical properties (hardness and toughness) of WC hardmetal parts manufactured by AM and conventional techniques. But as the process is not fully established to eliminate cracks and pores, it is obvious that the parts made of the AM process could show inferior mechanical properties. However, comparing the properties of nano-grained AM parts against micro-grained conventional parts could show a comparable and meaningful benefit in making functional prototypes.

Conclusions

The additive manufacturing technique has no doubt received the attention of both academic and industrial researchers together in the last few years. Industries have succeeded in manufacturing polymeric and metallic functional parts that go into critical engineering applications like complex valve cages for power plant [115]. Besides, though steel parts made of additive techniques are mechanically stronger than those made by casting [116], studies have shown that additively manufactured WC-based parts are always weaker than their conventional counterparts. Moreover, identifying a suitable process and establishing the process parameters for manufacturing tungsten carbide (WC-Co) is still in the nascent stage. However, in terms of process, most of the research is concentrated around SLM/SLS.

Different powder preparation techniques, process parameters, laser type, and scanning techniques were used by researchers and most of them have succeeded in manufacturing final parts with properties close to their conventional counterparts. Despite their efforts, the parts had microscopic and macroscopic level defects like cracks, pores, and microstructural inconsistencies. Eliminating pores and cracks would be the primary objective of future researchers to claim understanding and supremacy over the process. However, the flexibility in using nano-grain powders that show lesser grain growth when compared to the conventional technique could be a real gamechanger. Further comprehensive studies are required to check if the properties of nano-grained parts even with microscopic defects could match the performance of defect-free micro-grained parts manufactured by conventional techniques.

Though some researchers have succeeded in reducing the effect of pores and cracks by using post-processing techniques like secondary metal infiltration, it reduces the strength of the final part and compounds the complexity of the process. Increasing the binder content may reduce cracks, but the real-life application of such highbinder parts would be limited. Unlike conventional liquid phase sintered WC-Co parts (by powder metallurgy technique), the relationship between the process parameters, material behavior, and microstructural evolution is yet to be established. Controlling the irregular grain growth (fine on one side and coarse structure on the other side) between the layers, formation of ƞ-phase carbide, and porosity within the grains are some critical challenges that must be addressed. Moreover, there is very limited literature that compares the mechanical properties of parts made by AM and conventional techniques as the final product is intended for highly demanding wear applications.

Due to the various challenges involved in the manufacturing of WC-based hardmetal parts and limited process knowledge, so far researchers have not succeeded in fabricating a functional AM part that can replace their conventional counterparts. However, the similarities of the BJ3DP process with the conventional powder metallurgy process (non-heat-based preprocessing of initial shape and postprocessing sintering) make the technique a potential candidate for future studies. The extensive literature survey shows that the manufacturing of cemented carbide hardmetal parts by AM technique is undoubtedly a challenge and has a lot of scope for researchers to understand and establish the process to meet the functional requirements.