Method of W-Ni-Fe Composite Spherical Powder Production and the Possibility of Its Application in Selective Laser Melting Technology

Abstract

For the first time, a powder of W-5Ni-2Fe composition with spherical particles from 15 to 50 microns and a tungsten grain size from 0.5 to 3 microns was obtained using a new technological approach, developed by the authors, based on plasma spheroidization of powder granules made from nanoparticles synthesized in a plasma chemical process. The possibility of using the obtained spheroidized powder W-5Ni-2Fe in the process of selective laser melting (SLM) has been proved. The microstructure, physical, and mechanical characteristics of experimental samples made using SLM technology from the produced W-5Ni-2Fe powder have been studied. The results of the performed studies have shown that the microstructure of experimental samples is extremely dependent on the parameters of the SLM process. The precise choice of the SLM process mode made it possible to obtain a homogeneous structure of experimental samples of tungsten heavy alloy (WHA), with a tungsten grain size of about 1–2 microns, which is much smaller than the tungsten grain size in traditional heavy alloys. This creates prerequisites for increasing the strength characteristics of parts of complex shapes made by the SLM method from such powders. The maximum values of density and hardness of experimental samples obtained in the conducted studies are not worse than the values of samples obtained using traditional liquid-phase sintering technology. It is determined that the main problem of SLM powder W-5Ni-2Fe during investigation is the heterogeneity of the microstructure of massive samples and the formation of micropores and microcracks.

1. Introduction

In recent years, there has been a growing interest in additive manufacturing of tungsten-based materials that are hard to process [1,2,3,4,5,6,7]. However, such powder materials are still not mass produced, since there are no reliable scientific data to confirm the possibility of obtaining high-quality products on tungsten using additive 3D printing. The greatest practical interest is in the potential of making products from W-Ni-Fe heavy alloys, with limited mutual solubility of components. W-Ni-Fe heavy alloys serve mostly to manufacture products and structures for special purposes [8,9,10]. Currently, bulk W-Ni-Fe products can be made using the following powder metallurgy methods: liquid phase melting, which is sintering at temperatures higher than the melting temperatures of the most low-melting phase of an alloy without applied pressure [11,12,13], and solid phase spark plasma sintering [14,15,16,17]. Solid phase spark plasma sintering involves the high-speed heating of powders in a graphite mold by passing millisecond current pulses, with an amplitude of several thousand amperes, under external pressure.

A modern technology that is potentially effective in manufacturing W-Ni-Fe products is selective laser melting (SLM). This is a laser powder bed fusion (LPBF) technique that uses a scanning laser beam to melt the powder bed and to perform layer-by-layer synthesis of the product of a given shape, in compliance with a 3D drawing [18,19]. Unlike other powder technologies, SLM allows the production of geometrically complex products with high physical and mechanical characteristics [20,21].

Currently, the development of additive 3D alloy product printing is challenging since W-Ni-Fe powders for SLM are not mass-produced. Additive powders should meet the following summarized requirements: powders should be homogeneous in composition (a uniform distribution of chemical elements is required inside each particle of a powder), powders should consist of spherical particles, and powders should have high flowability.

The aim of this study is to obtain a W-Ni-Fe composite powder with spherical particles using a new technological approach developed by the authors, based on plasma spheroidization of powder granules made from nanoparticles synthesized in a plasma chemical process. An important part of the work is to study the possibility of using such a powder in selective laser melting by synthesizing test samples through physical and mechanical, as well as structural, studies.

2. Materials and Methods

2.1. Procedures for Studying the Structure and Properties of Powders and 3D Specimens of Materials

Structural studies of powder materials were carried out using various SEM and TEM microscopes: TEM Tecnai G² F20 20–200 kV, TEM Titan G2 60–300 kV, SEM Scios DualBeam, SEM Quanta 3D (FEG) DualBeam—all microscopes by FEI Company, Netherlands, Eindhoven, and SEM JSM-7600F (Jeol, Japan, Tokio). Each electron (SEM and TEM) microscope used had a wide-angle energy dispersive X-ray (EDX) detector for elemental microanalysis. All SEM microscopes (FEI) were also equipped with a focused FIB ion beam with Ga ion etching. The TriStar 3000 specific surface analyzer (Micromeritics, Norcross, GA, USA) was used to measure the specific surface area of powders, using the Brunauer–Emmett–Teller polymolecular adsorption method. Total oxygen and nitrogen were measured with the TC-600 analyzer (LECO, USA, St. Joseph, Michigan). Total hydrogen in powder materials was measured with the RHEN-60 analyzer (LECO, St. Joseph, MI, USA) during reduction melting in a graphite crucible. Particle size distribution of the micropowder was measured with the Mastersizer 2000 M (Malvern Panalytical, Malvern, Great Britain) particle size analyzer using laser diffraction. Nanoparticles in spheroidized micropowders were separated using fractional separation in liquid by suspension sedimentation, after treatment with the Sonopuls HD 3100 ultrasonic homogenizer (Bandelin, Berlin, Germany). Powder fluidity was determined using a calibrated funnel (Hall flowmeter, ISO 4490) and a stopwatch. Powder bulk density was determined by the gravimetric method using a funnel. X-Ray Diffraction (XRD) phase analysis was performed on the Bruker AXS D8 ADVANCE diffractometer (Karlsruhe, Germany) using filtered Cu Kα radiation, with multi-channel linear detector LynxEye. Phase analysis was performed using the EVA V5.0 program (Bruker, Karlsruhe, Germany) and the PDF Crystallography Database 2016.

The Sartorius CPA225D balance (Sartorius, Helsinki, Finland) was used to measure the density of test specimens by hydrostatic weighing. Structural studies of the test specimens were carried out using the Leica IM DRM optical microscope (Leica microsystems, Wetzlar, Germany) and the Tescan Vega 2 (Tescan, Brno, Czech Republic) scanning electron microscope. Vickers hardness of the test specimens was measured in compliance with ISO 6507-1:2018 using the QNESS Q60A+ automatic hardness tester (ATM Qness GmbH, Mammelzen, Germany) at a load of 10 kg. Microhardness of the test specimens was measured in compliance with ISO 14577-1:2015 using the Struers Duramin-5 microhardness (Struers, Ballerup, Denmark) tester and a tetrahedral pyramid at a load of 200 g.

2.2. Powder Processing Procedures

W-5Ni-2Fe heavy alloy powder was processed following a procedure similar to [22], the main stages of which include the plasma chemical synthesis of composite W-Ni-Fe nanopowders, the granulation of nanopowders to obtain nanopowder microgranules, and their spheroidization and compaction in a thermal plasma flow.

Powders of WO₃, Fe₂O₃, and NiO, consisting of particles less than 25 μm in size, were mixed in the C2.0 Turbula mechanical mixer at a ratio of metal oxides corresponding to the content of metals in W-Ni-Fe heavy alloy.

A mixture of oxides was reduced to metals in a flow of hydrogen–nitrogen thermal plasma, generated in an electric arc plasma torch with a nominal power of 30 kW. For a detailed description of the device, see [23].

The process in a plasma flow ensures reduction reactions in the gas phase, leading to the formation of target metals in the form of nanosized particles. The resulting composite W-Ni-Fe nanopowders consist of nanosized particles of predominantly spherical shape (Figure 1—BF TEM images obtained under 150 kV/43,000× and 200 kV/145,000× respectively), with the specific surface area of the nanopowder being 4.8 m²/g.

Previously [24], it has been established that nanoparticles have a “core–shell” structure, where Ni and Fe form an alloy in near-surface layers of nanoparticles, while the core of nanoparticles consists of W. It has been proved that a uniform distribution of Ni and Fe is achieved over the surface of all tungsten nanoparticles (Figure 2). Elemental maps were obtained in the scanning mode of TEM (STEM, 200 kV and 320,000×). The oxygen content in the nanopowders obtained ranged between 1.3 and 2.5 wt. %. For a detailed study of the dispersity and chemical composition of the resulting nanoparticles, as well as their internal structure and surface state, see [24].

The resulting nanopowder was subjected to “dry” granulation by pressing, followed by grinding and sieve classification. Such a nanopowder granulation scheme allows for a 40% yield of the nanopowder microgranules fraction (Figure 3a). Nanopowder granules have a homogeneous internal structure, no cavities (Figure 3b), and a uniform distribution of elements.

As a result of processing nanopowder microgranules in an argon–hydrogen plasma flow generated in an electric arc plasma torch, spherical microparticles of the W-Ni-Fe composite were obtained (Figure 4a). Along with microparticles, plasma-treated products contain nanoparticles, formed after condensation of partially evaporated Ni and Fe (Figure 4b). For more details about plasma spheroidization and its capabilities, see [25].

Sedimentation in liquid (distilled water) was used to separate spherical microparticles from nanoparticles, since microparticles of a plasma-treated powder precipitated while nanoparticles stayed in liquid.

3. Results and Discussion

Particle size distribution tests of a plasma-treated powder have shown that the average particle size is 30 µm, while 65% of particles range from 15 to 35 µm (Figure 5). Particle size distribution is close to lognormal.

Spherical particles are predominantly dense and have no internal cavities (Figure 6a). Analysis of the microstructure of plasma-treated powder particles has shown that the size of tungsten grains corresponds to the micron range (Figure 6b,c).

The W-Ni-Fe micropowder, obtained by processing nanopowder microgranules in a hydrogen-containing plasma jet, contains 0.7 wt. % of oxygen, 0.005 wt. % of nitrogen, and 0.0001 wt. % of hydrogen. The powder bulk density is 8.9 g/cm³ and the powder flow rate is 10 s (Hall flowmeter, 50 g). The phase composition of W-Ni-Fe micropowder is characterized by the presence of the main phase W and slight phases of Fe_0.95W_0.05 and Ni_8.11W_1.89 (Figure 7).

Test specimens for studying the structure and physical and mechanical properties were processed from W-Ni-Fe powder by selective laser melting using the Realizer SLM 100 machine with a modernized 1000 W laser system. Basic parameters of SLM for W-Ni-Fe powder are provided in

Table 1. Operating parameters of selective laser melting of W-5Ni-2Fe powder.

Parameter	Value
Laser	Ytterbium fiber laser
Mode	Continuous
Laser wavelength, nm	1070 ± 10
Layer thickness, μm	65
Laser power, W	from 500 to 1000
Laser scanning velocity, mm/s	from 100 to 1000
Diameter of a laser beam, μm	75
Hatch spacing, μm	140
Hatch angle rotation between adjacent layers, °	80
Shielding gas	Argon (99.998%)

.

Specimens had a disk shape, a diameter of 9 mm, and a thickness of 2 mm. The effect of laser scanning velocity (at fixed laser power (P) of 700 W) and laser power (at fixed laser scanning velocity (V) of 500 mm/s) on density, hardness, microhardness, and structure of test specimens was studied.

Figure 8 shows a diagram of the W-Ni-Fe SLM test specimens’ density, depending on laser power and laser scanning velocity. The optimal fusion range conducive to high density is shown to be at a laser power of 550–750 W and a laser scanning velocity of 200–500 mm/s. Maximum density is observed at 650 ± 50 W and 250 ± 50 mm/s.

Figure 9 shows dependences of microhardness on laser power and laser scanning velocity.

An Increase in laser scanning velocity from 100 to 500 mm/s is shown to have practically no effect on microhardness. Further increase in laser scanning velocity from 500 to 1000 mm/s leads to a decrease in microhardness from 5.4 to 3.5 GPa. It should be noted that with an increase in laser scanning velocity, there is a significant increase in the dispersion of microhardness (mean-square deviation of microhardness values), which may indicate a high degree of inhomogeneity of the material structure. Any changes in laser power within the range of 600–900 W have practically no effect on microhardness, which is approximately 5.5 GPa within the measurement accuracy. A decrease in laser power from 600 to 400 W leads to a sharp decrease in microhardness from 5.5 to 3.8 GPa.

Figure 10 shows optical images (×4 and ×10 magnifications) of the structure of W-Ni-Fe specimens. At a laser scanning velocity of 300 mm/s, a network of microcracks is observed in the structure (Figure 10a). With laser scanning velocity increasing to 500 mm/s, the number of microcracks decreases, followed by growing porosity. Individual pores reach 200 µm in size (Figure 10b). A further increase in laser scanning velocity brings a proportion of microcracks down and a proportion of pores up (Figure 10c). At a laser scanning velocity of 900 mm/s, microcracks are hardly observed in a metallographic specimen; however, the porosity reaches 20–30% (Figure 10d).

Figure 11 shows the structures of SLM specimens at different laser scanning velocities. The images have been obtained using scanning electron microscopy methods. At a laser scanning velocity of 300 mm/s, the structure is predominantly “microdendritic”, with a microdendrite of about 10 μm in size and a step of microdendrite branches of 2 μm (Figure 11b). On the contrary, at a laser scanning velocity of 900 mm/s, a mixed-type microstructure is formed: there are areas with a pronounced dendritic structure, areas with an equiaxed microstructure with round tungsten grains of about 1 μm in size (Figure 11d), and individual particles of unmelted powder of about 5–10 μm in size (Figure 11c). It should be noted that microdendrites of about 10 μm in size were present in the initial structure of powder particles (see Figure 6). The microstructure observed in SLM materials can be formed during Ni and Fe melting while maintaining W dendrites of the original powder, rather than during complete W fusion.

Figure 12 (×4 magnification) shows typical images of the W-Ni-Fe SLM specimens’ structure at various laser powers. Any changes in laser power within the range of 600–900 W are shown to have practically no effect on the microstructure.

Figure 13 shows the structure of SLM specimens at varying laser power. The images have been obtained using scanning electron microscopy methods. At a laser power of 600 W, the structure is predominantly mixed: there are areas with a pronounced dendritic structure and areas with an equiaxed microstructure, with round tungsten grains of about 1 μm in size (Figure 13a). At a laser power of 900 W, a more uniform and coarse-grained microstructure is formed in SLM specimens, unlike specimens processed at 600 W (Figure 13b).

Figure 14 shows typical microstructures of W-Ni-Fe SLM specimens at various laser power and laser scanning velocities. An increase in laser scanning velocity is shown to lead to an increase in porosity of SLM specimens, with an excessive increase in laser power resulting in micropores. Thus, the following power–velocity range of SLM parameters is deemed optimal for W-Ni-Fe alloy: a laser power of 650–700 W and a laser scanning velocity of 500–550 mm/s. At the same time, it should be noted that modern SLM equipment is able to maintain all the technological parameters of the process. However, a uniform structure of bulk SLM W-Ni-Fe specimens cannot be guaranteed since there is no control over the actual effect on the alloyed material. To tackle the issue of areas with various types of structural defects that appear in the structure of specimens and, above all, to prevent accidental formation of microcracks, it is deemed possible to shift the optimal range of SLM process parameters to microporosity and to apply hot isostatic pressing of SLM specimens to reduce the porosity of the material. At the stage of plasma spheroidization of W-Ni-Fe nanopowder granules, there is another opportunity to influence the structure of W-Ni-Fe materials produced by SLM. Depending on how long the liquid phase of particles lasts in the plasma process and their cooling rate, the internal structure of the resulting spherical particles may change from predominantly microdendritic (Figure 6b) to predominantly classical equiaxed (Figure 6c). The proposed approaches require further investigation. Likewise, the stability of the microstructure and properties of SLM W-Ni-Fe-based materials should be studied separately.

Comparison of Physical and Mechanical Properties of SLM Specimens Processed Using Various Technologies

The comparative research focuses on the structure and physical and mechanical properties of W-Ni-Fe specimens that have been processed using liquid phase sintering, with subsequent deformation (LPSD) and spark plasma sintering (SPS).

Studies of the microstructure of W-Ni-Fe specimens, using scanning electron microscopy, have shown the following results: optimal SLM modes make it possible to form a microdendritic structure with microdendrites of about 5–10 µm in size and a step of microdendrite branches of about 2 µm (Figure 15a); during SPS, a homogeneous equiaxed structure with an average grain size of about 1 µm is formed (Figure 15b); after LPSD, a classical two-phase structure is observed, which is typical of W-Ni-Fe alloys consisting of a solid tungsten-based solution and a matrix of a solid nickel and iron-based solution. The particle size of tungsten is about 20 µm (Figure 15c).

A comparison of the W-Ni-Fe samples’ density values has shown that the density of SLM samples is hardly inferior to the density of SPS specimens and LPSD specimens (

Table 2. Density and Vickers hardness of W-5Ni-2Fe samples produced by various technologies.

Samples Properties	SLM	SPS	LPSD
Density, ρ (g/cm3)	16.8	16.95	17
Vickers Hardness, HV (GPa)	4.8	5.4	4.8

) and is 16.8 g/cm³.

A comparison of the Vickers hardness values (at a load of 10 kg) of the W-Ni-Fe samples has shown that the hardness of the SLM samples is 4.8 GPa, meaning that it is not inferior to that of the samples processed using the traditional technology of deformation processing, whereas the hardness of the SPS samples is 5.4 GPa, which is 10% higher than that of the SLM samples (Table 2). The higher values of the hardness of the SPS samples are caused by a finely dispersed structure in the tungsten phase and the lower porosity of the material.

4. Conclusions

The research has resulted in the production of W-Ni-Fe heavy-alloy powder with spherical particles ranging from 15 to 50 μm in size and with a grain size of 0.5–3 μm, that has been processed for the first time by combining W-Ni-Fe nanopowder synthesis, granulation, and subsequent plasma spheroidization.

The research demonstrates the possibilities of using W-Ni-Fe spheroidized powder in 3D printing and processing W-Ni-Fe test specimens by SLM.

The maximum density and Vickers hardness of the W-Ni-Fe test specimens processed using SLM in optimal modes are shown to be 16.8 g/cm³ and 4.8 GPa, respectively, and are not inferior to the corresponding values for specimens obtained using a traditional technology of liquid phase sintering.

The internal structure of the W-Ni-Fe microparticles after plasma spheroidization of the nanopowder granules can largely determine the structure and properties of specimens processed by SLM. Additional studies to assess the effect of this factor can expand the potential of using W-Ni-Fe microparticles in SLM.

SLM parameters are shown to significantly affect microstructure formation in W-Ni-Fe alloy specimens. Within a narrow range of SLM parameters, a homogeneous equiaxed microstructure with an average grain size of about 1 μm has been formed. However, the inhomogeneity of the microstructure of bulk specimens along with the microcracks and micropores formed in their structure is the core issue of selective laser melting of W-Ni-Fe powder, which requires further study.

The presented results were obtained for the first time and have no prototypes. It has been established that the proposed approach makes it possible to obtain materials with properties corresponding to those manufactured using traditional technologies.

The studies carried out are the very initial stage of the study of the development of this approach, which does not yet allow us to determine the specific advantages and disadvantages of the proposed approach.