Feasibility of Production of Multimaterial Metal Objects by Laser-Directed Energy Deposition

Abstract

The article focuses on the possibility of manufacturing bimetallic products for specific industrial applications using laser-directed energy deposition (LDED) additive technology to replace the traditional brazing process. Preferential process regimes were determined by parametric analysis for the nickel-alloy–steel and molybdenum–steel pairs. Comparative studies of the microstructure and hardness of the deposited layers and the transition layer at the boundary of the alloyed materials have been carried out. It is shown that LDED provides better transition layer and operational properties of the final part since the low-melting copper layer is no longer needed. A combined technological process has been developed, which consists in combining the traditional method of manufacturing a workpiece through the casting and deposition of a molybdenum layer by LDED.

1. Introduction

With the development of technology and the complication of the designs of machine parts and assemblies, the need for the manufacture of parts of complex geometry with inhomogeneous physical and mechanical properties is growing. Since known materials cannot fully provide all the physical and mechanical needs for such parts, it became necessary to manufacture heterogeneous parts, where individual parts of this part will be made of different materials. Even with conventional techniques, manufacturing multimaterial products (MMP) is a complex technological task. The main obstacles include time-consuming technological processes, production costs, and the properties of the parts. Therefore, it provokes a demand for new ways to solve these problems. Additive manufacturing, particularly laser-directed energy deposition (LDED), is considered to be a key technology for obtaining MMP and tackling the issues [1,2,3,4,5,6,7]. In this technology, many authors note the importance of the powder-feed scheme [8,9]. It is concluded that the coaxial method is the most beneficial, since powder feeding and optical system are independent of moving parts or a nozzle, providing a stable LDED process [10].

Today, additive manufacturing is developing rapidly, especially selective laser melting and LDED. New materials including ceramics are being introduced [11,12,13,14], improved laser optical schemes are being proposed [15], and multimaterial additive manufacturing is developing. The authors of [16] show the manufacture of bimetallic steel–copper products by selective laser melting. In spite of achieving a strong metallurgical bond, the authors emphasized that further studies are necessary in order to determine optimal processing parameters. However, being more productive, LDED is preferred in certain applications. This group of technologies includes several techniques capable of obtaining 3D objects without a powder bed, such as laser-directed energy deposition, welding, and cold gas dynamic spraying [17,18,19]. Nevertheless, LDED is considered to be the most widespread and developing technology. In [20,21], the LDED process of 316L steel and a Co-based superalloy was investigated both theoretically and practically. It was recommended to coat the powder layer with different thicknesses for each material due to distinctions in thermal and physical properties. The authors of [22,23] studied the influence of LDED working parameters on the structure and properties of the materials. The authors of [24,25,26] investigated the properties and showed the possibility of the practical application of the following pairs of metals: steel–copper, steel–aluminum, steel–titanium, and steel–nickel. For example, a steel–copper (bronze) pair can be used as heat exchangers (due to the high copper/bronze conductivity) or as a cam bushing (high tribological properties of copper/bronze), and a steel–aluminum pair can be used as adapters in oxygen regenerators or as parts of electrolytic refining equipment (high corrosion resistance of aluminum). The authors of [27] deposited molybdenum powder (Mo 99%) of a fine fraction by the preliminary layer deposition method, under the following LDED parameters: V = 0.8 cm/s, do = 0.3 cm, PL = 2 kW. It has been calculated that 51% of the absorbed power is spent on evaporation. A homogeneous structure was obtained, which, according to the authors, should positively influence the provision of high-performance properties. In [28], coatings from a high-entropy alloy (HEA) (CoCrFeNi)95Nb5) 100-xMox (x = 1, 1.5, and 2) were deposited on a substrate of steel 45 by LDED. The effect of laser radiation power and the percentage of molybdenum in the alloy on the microstructure and microhardness of coatings has been studied. The microstructure of the coating consisted of columnar dendrites which disappear with an increase in the molybdenum content; the grains are crushed and become more compact. The volume fraction of the interdendritic phase at a laser radiation power of 800 W was small and uneven. As the laser-beam power increased to 1000 W, the volume fraction of the interdendritic phase increased. At a laser-beam power of 1200 W, the volume fraction of the interdendritic phase decreased again. The coatings obtained at a power of 1000 W had the highest volume fraction of the interdendritic phase and a higher microhardness. The authors of [29] carried out the deposition of T15M composite powder on a steel substrate at different laser-scanning velocities. It was found that the coating obtained at a deposition rate of 200 mm/min had better wear resistance. In [30], it was shown that the microhardness and wear resistance of composite coatings based on Co on the titanium alloy Ti-6Al-4V increased with a decrease in the specific energy of laser radiation. Sun et al. deposited a composite Ni/Mo coating on the surface of a copper alloy using LDED [31]. The laser power was 6000 W, the scanning speed was 5 mm/s, and the powder feed rate was 10 g/min. The surface layer consisted of three Ni layers and two Mo layers. Due to the flowability and nonequilibrium solidification of molybdenum in the molten state, pores and cracks along the grain boundaries were observed in the Mo layer. The surface hardness of the Mo layer ranged from 200 to 460 HV. In [32], coatings from the Ni–Cr–Mo alloy with different Cr content on carbon constructional steel, obtained by LDED, were studied. An increase in the Cr content led to the formation of a dense protective passive film. Corrosion resistance first deteriorated with increasing Cr content from 18 wt.% to 22 wt.% and then improved with increasing Cr content from 22 wt.% to 26 wt.%. The authors of [33] studied the LDED of powders of high-entropy FeCoCrNi and FeCoNiCrMo0.2 alloys on the surface of stainless steel 304. It was found that the addition of Mo leads to a significant increase in the size of the dendrites in the middle region of the FeCoCrNiMo0.2 coating. The FeCoCrNiMo0.2 coating has a high corrosion resistance: a low current density in a NaCl solution with a concentration of 3.5 wt.%. A study of LDED using Ni–Mo–Si powder mixtures on an austenitic stainless steel substrate shows that the coating microstructure consists of primary dendrites and interdendritic eutectic [34]. Due to the presence of a large amount of hard and wear-resistant phase, the laser-deposited composite coating containing metal silicides has excellent wear resistance under conditions of sliding friction at high temperatures.

The review shows that the development of multimaterial LDED is promising. However, several highly demanded material systems for LDED, such as molybdenum and low-carbon steel, corrosion-resistant steel, and nickel heat-resistant alloy, still require further research to meet industry needs. So far, no successful industrial-scale experience of LDED of these materials has been reported in the literature.

This work aims to study the possibility of manufacturing bimetallic products for specific industrial applications by LDED.

2. Materials and Methods

2.1. Raw Materials

The materials were taken according to aviation industry demands in the field of additive manufacturing. Particularly, in the first case it is proposed to replace the traditionally produced ‘body’ part with a LDED-processed one. Originally, the chosen ‘body’ part consists of a base made of 25L steel and a molybdenum plate brazed on, as shown in Figure 1, and the objective is to obtain the molybdenum plate (Figure 1) by LDED. This design is determined by the operational conditions of the part: the brazed protective layer must have wear-resistant properties at high temperatures to prevent coating destruction during exploitation. However, copper interlayer cannot withstand working temperature of the ‘body’ part (1083 °C) causing destruction of the coating. Therefore, to avoid this issue an alternative technology, LDED, was proposed for molybdenum on a cast 25L steel base without a copper interlayer. Pure molybdenum powder PMS-M99,9 (particle size 40–100 μm, JSC “POLEMA”, Tula, Russia,

Table 1. The chemical composition of the Mo powder PMS-M99,9 [35].

Material	Elements Composition, % Mass.
	Mo	Residual Elements 1, Total	O
Mo powder	>99.9	0.1	0.025

¹ Al, Fe, K, Ca, Si, W, Mg, Ni, Na, Mn, Zn.

) and carbon steel 25L substrate (

Table 2. The chemical composition of the 25L steel substrate.

Material	Elements Composition, % Mass.
	Fe	C	Mn	Si	S	P
25L steel	balance	0.22–0.30	0.45–0.90	0.20–0.52	0.05	0.05

) were used within the experiment.

Replacing brazing in the production of multimaterial objects is relevant for many industries. Thus, in the second case we attempted to obtain an ‘electromagnet case’ (Figure 2) from heat-resistant nickel alloy powder HN71MTYuB (particle size 0–70 μm, JSC “VILS”, Moscow, Russia,

Table 3. The chemical composition of the HN71MTYuB alloy powder.

Material	Elements Composition, % Mass.
	Fe	C	Si	Ni	S	Cr	Ce	Mo	Nb	Ti	Al	B	Pb
HN71MTYuB alloy	≤2.0	0.03–0.07	≤0.5	71.034–78.46	≤0.007	13.0–16.0	≤0.01	2.8–3.2	1.9–2.2	2.35–2.75	1.45–1.8	0.011–0.013	≤0.001

) and corrosion-resistant austenitic steel powder PR-12H18N9T (particle size 20–63 μm, JSC “POLEMA”, Tula, Russia,

Table 4. The chemical composition of the 12H18N10T steel powder.

Material	Elements Composition, % Mass.
	C	Cr	Fe	Mn	Ni	P	S	Si	Ti	Cu
PR-12H18N9T steel	≤0.12	17.0–19.0	balance	≤2.0	8.0–9.5	≤0.035	≤0.02	≤0.8	1.9–2.2	≤0.3

) by LDED. Such objects are traditionally made by brazing dissimilar materials, which negatively affects the quality. The work proposes the sequential growth of parts from metal powders by LDED.

2.2. LDED Equipment

LDED process was performed on a 3D TruCell 3008 CNC machine (Trumpf, Ditzingen, Germany) equipped with a Yb:YAG disk laser with an output power range of 80–2000 W and a wavelength of 1030 nm. A special LDED head with coaxial powder feed with a nozzle gap of 7 mm was used. The nozzle has a two-channel powder-feeding system to work with two powders sequentially or simultaneously. The powder was fed into the working area through a nozzle using a shielding gas of argon and helium as the carrier gas at a rate of 10 L/min each. The optical system provides a focused laser spot with a diameter of 200 μm. Since the goal was to provide a precision LDED process to manufacture objects, the experiments were carried out with a laser diameter spot of 200 μm. The transverse energy distribution inside the laser beam was Gaussian (normal distribution).

2.3. Microstructure Analysis and Properties Evaluation

The main indicators for the quality of powder materials for LDED are their granulometric, morphological, and chemical composition [36,37]. Granulometric analysis of powders was carried out on an Occhio 500 nano-optical morphometer (Occhio S.A., Angleur, Belgium) with software for statistical image analysis. A morphological and elemental analysis was performed on a Tescan Vega 3 LMH scanning electron microscope (SEM) (Tescan, Brno, Czech Republic) equipped with an energy dispersive X-ray microanalyzer (Oxford Instruments, Abingdon, UK).

The microstructure and microrelief of the steel surface were studied using a Carl Zeiss Axio Observer D1m optical microscope (Zeiss AG, Oberkochen, Germany) and a PHENOM G2 PRO scanning electron microscope (Thermo Fisher Scientific, Waltham, MA, USA) with a built-in energy-dispersive EDX analyzer (Thermo Fisher Scientific, Waltham, MA, USA).

To determine the phase composition of the samples, X-ray diffraction phase analysis was used. The diffraction patterns were taken on a PANalytical Empyrean Series 2 X-ray diffractometer (Malvern Panalytical, Malvern, Worcestershire, UK) using monochromatized CuKα radiation. The phase composition was analyzed using the PANalytical High Score Plus v3.0 software (Malvern Panalytical, Malvern, Worcestershire, UK) and the ICCD PDF-2 database.

To analyze the microhardness, a Qness Q10A microhardness tester (Qness GmbH, Golling, Austria) was used with a maximum indenter load of 10 kg, which makes it possible to determine the hardness by the Vickers method with a measurement error HV = 0.01.

3. Results and Discussion

3.1. Granulometric and Morphological Analysis of the Raw Materials

To confirm the compliance of the raw powder with the required parameters, an input control of the powder material was carried out. Chemical analysis showed that the composition of the powder according to the main elements corresponds to the one declared by the manufacturer.

Based on the results of the granulometric analysis of powders, integral and differential curves of the distribution of powder particles by size were constructed.

Each point on the curve corresponds to the sum of fractions less than a certain diameter. The distribution histogram of the molybdenum PMS-M99.9 powder particles is shown in (Figure 3a).

The average particle size d₅₀ is defined as the mathematical expectation of the differential particle size distribution curve. It has been established that the average particle size of the PMS-M99.9 powder is d₅₀ = 76.79 µm. The total content of particles that do not correspond to the size of the main fraction declared by the manufacturer (from 40 to 100 microns) is 9.75%. Powder particles have a spherical shape (Figure 3b) and have a high sphericity index of more than 90%, which has a positive effect on the LDED process.

Figure 4 shows the integral curves and histograms of powders HN71MTYuB (a) and PR-H18N9T (b). For these powders, the distribution of particle sizes corresponds to the normal law, which is clearly seen in the graphs. The graphs also show the main parameters of the fractional composition. The maximum particle size for the HN71MTYuB powder is 103.2 μm, the minimum is 12.21 μm. The remaining particle sizes are not considered due to their insufficient content in the powder. The maximum particle size for PR-12H18N9T is 62.15 μm, the minimum is 15.65 μm. It should be noted that the powder material HN71MTYuB has a significantly larger fractional particle size compared to PR-12H18N9T, which adversely affects the geometric characteristics of the transition layers in multimaterial products.

The HN71MTYuB powder particles have a high sphericity index (average value 71.7%) and low roughness value (average value 2.8%), which ensures the application of a uniform deposited layer (Figure 5a).

Figure 5b clearly shows that the majority of PR-12H18N9T powder particles also have a spherical shape (average value of the sphericity index 66.3%) and a low roughness value (average value 4.5%), with a small number of satellites and irregularly shaped particles.

Thus, the analysis confirmed that the powders meet the requirements for powders for additive technologies [1,2,4].

3.2. Study of the LDED Processing of the ‘Body’ Part from Molybdenum Powder on a Steel Substrate

In order to determine the optimal parameters for LDED of molybdenum on a steel substrate, a parametric analysis of single beads of the materials was performed. It was carried out by varying the laser radiation power in the range from 200 to 400 W, and the deposition rate from 300 to 700 mm/min. A stable coaxial powder flow is ensured at a powder-flow rate F_pow = 1.5 g/min and a carrier-gas-flow rate of at least F_cgas = 4 L/min. Effective protection of the melt zone during deposition, without the threat of deformation, is ensured at a flow rate of shielding gas of F_shgas = 4 L/min.

The criteria for choosing the preferred parameters for the single beads included the following indicators: track appearance (uniformity of penetration, absence of cracks), microstructure, their geometric characteristics (width, height, and depth of penetration), and hardness values of the deposited beads (

Table 5. The hardness of the deposited beads (HV) depending on the scanning speed of the laser beam, P = 400 W.

Track Number	1	2	3	4	5	6	7	8	9
Laser-scanning velocity, mm/min	300	350	400	450	500	550	600	650	700
Hardness, HV	430	436	428	421	425	418	423	421	426

, Figure 6). As a result, a beam-scanning speed of V = 350 mm/min and a laser power of 400 W were found to be optimal.

Hatch distance varied in a range from 0.7 to 2.0 mm while the scanning speed, laser-radiation power, powder-feed rate, and carrier- and shielding-gas consumption were constant (V = 350 mm/min, P = 400 W, F_pow = 1.5 g/min, F_cg = 4 L/min, and F_shg = 4 L/min). The distance between the axes of two adjacent beads 2.0 mm and 1.5 mm (Figure 7a) is insufficient to obtain a monolithic and uniform layer.

At a hatch distance of 0.7 mm, the first layer was monolithic, but not uniform. With such a distance between the beads, defects in the form of cracks were found which is the most common defect for refractory materials. With a hatch distance of 1.0 mm, it was possible to obtain a uniform surface.

To ensure the isotropy of properties of the parts, the scanning rotated 90° from layer to layer while the vertical hatch distance Δz ≈ 0.25 mm was 0.25 (Figure 7b). As a result of parametric analysis, the preferential modes for the LDED of molybdenum PMS-M99.9 powder cast steel 25L were determined: laser-radiation power P = 400 W, scanning speed V = 350 mm/min, hatch distance of 1.0 mm, powder consumption F_pow =1.5 g/min, and flow rate of carrier gas F_cg = 4 L/min and shielding gas F_shg = 4 L/min. The process parameters were: scanning speed V = 350 mm/min, laser power P = 400 W, beam scanning step 1.0 mm, powder consumption F_pow = 1.5 g/min, carrier-gas flow F_cg = 4 L/min and shielding gas F_shg = 4 L/min, and hatch distance along the vertical axis Δz ≈ 0.5 mm.

As mentioned earlier, the traditional way of making the ‘body’ part involved vacuum brazing to connect the molybdenum plate and the steel body using copper brazing alloy. Figure 8 shows the microstructures of the layers during brazing and LDED. The microstructure of a brazed joint consists of a mixture of solid solutions based on copper (ε-phase) and iron (α-phase) (Figure 8a). Figure 8b shows the microstructure of the deposited layer in the cross-section at the boundary of the melting zone and the steel substrate. It can be seen that there are no structural defects (cracks and pores) in the deposited layer and at the boundary of the melting zone.

In the deposited Mo layer, a dendritic structure is observed; columnar dendrites grow from the boundary of the melting zone to the center of the molten pool, i.e., in the direction opposite to the heat sink (Figure 9).

The distribution of elements on the steel–molybdenum boundary during traditional brazing is presented in Figure 10.

The analysis of the microstructure (Figure 8a) of the brazed joint (Figure 10) shows that it has a large width (up to 65 microns), in which there are large areas of copper commensurate in size with grains of ferrite and perlite in the steel substrate. This leads to the destruction of the coating during operation, since the operating temperature during the operation of the ‘body’ part exceeds the melting point of the copper brazing alloy (1083 °C).

This disadvantage of brazing is eliminated by LDED. First of all, within the LDED process, there is no need to use additional materials to ensure a strong bond between Mo and steel. Figure 11 shows a distribution map of the main elements in the region of the boundary between the deposited material and the steel substrate.

An analysis of the concentration distribution of Mo and Fe in the region of the boundary of the deposited material and the steel substrate (Figure 11) shows that mutual diffusion of Mo and Fe is observed along the width of the transition layer of the melting zone (MZ) and the heat-affected zone (HAZ), which further increases the adhesion of the deposited layer.

A comparative analysis of the hardness of the molybdenum plate after brazing and the deposited molybdenum layer showed the advantage of LDED. The hardness of the molybdenum layer obtained by LDED is much higher (

Table 6. Hardness and density depending on the Mo deposition technology.

Deposition Technology	Hardness, HV	Density, g/cm3
Vacuum brazing (t = 1120 °C, τ = 10 min)	182–192	10.0–10.2
LDED	412–443	10.0–10.1

). The low hardness of the molybdenum plate on the samples made according to the traditional technology is observed as a result of the annealing of the molybdenum plate. During LDED, hardening from the liquid state is realized, which leads to an increase in hardness. Table 6 shows the hardness and density values of the molybdenum plate after brazing and the molybdenum layer obtained by LDED.

It should be noted that the operating time of the ‘body’ part is limited by the melting temperature of the constituent parts. Thus, LDED technology provides better operational properties.

3.3. Study of the LDED Process of Manufacturing the ‘Electromagnet Case’ from Steel 12H18N10T and Alloy HN71MTYuB Powders

To improve performance of the ‘electromagnet case’ part, brazing was replaced by LDED of alternating layers of corrosion-resistant 12H18N10T steel and HN71MTYuB heat-resistant nickel alloy and optimized through a parametrization process. The laser-radiation power was varied from 80 W to 300 W, the beam scanning speed was varied from 100 m/s to 1700 m/s, and the powder feed was from 1 rpm to 8 rpm. (1 rpm = 1.25 g/min for steel and 1 rpm = 1.34 g/min for nickel alloy). Figure 12 shows the microstructures of single beads of steel 12H18N9T (Figure 12a) and alloy HN71MTYuB (Figure 12b) obtained in different regimes.

Based on the results of the analysis of the geometric characteristics of the deposited beads, the assessment of the porosity of the material, the assessment of the stability of the deposited beads and the presence or absence of cracks, the following preferential regimes for single bead deposition are proposed: for steel PR-12H18N9T—power P = 100 W, V = 1150 mm/min, F = 8.75 g/min (Figure 12a-7), for alloy HN71MTYuB—power P = 100 W, V = 1400 mm/min, F = 9.38 g/min (Figure 12b-6).

The criterion for the optimality of the deposited layer and the multilayer structure is the absence of porosity and the uniformity of the surface of the deposited layer.

For two- and three-dimensional objects for both materials, it was found that the optimal scanning step along the X-Y axis is 0.2 mm and along the Z-axis 0.22 mm.

Figure 13 shows a uniform surface of deposited multilayer structures for steel PR-12H18N9T. Figure 14 demonstrates the structure of a multilayered bimetallic specimen (PR-12H18N9T and HN71MTYuB).

Figure 14 shows the microstructure of alternating layers of the bimetallic specimen. Often, due to the difference in thermophysical properties of materials, microstructural imperfections occur. The problem is addressed by correcting the working parameters. Additionally, the flatness of the surface of the underlayer plays an important role in the deposition of subsequent layers. The deposited layers repeat irregularities of the surface and previous layers.

Hardness analysis was carried out to evaluate the mechanical properties of the deposited layers. The maximum hardness of the steel material was 250 HV_0.1, and that of the nickel alloy was 350 HV_0.1.

4. Conclusions

As a result of parametric analysis, the preferential parameters for the LDED process of grade PMS-M99.9 molybdenum powder on 25L cast steel were determined: laser radiation power P = 400 W, scanning speed V = 350 mm/min, hatch distance 1.0 mm, powder consumption F_p = 1.5 g/min, flow rate of carrier gas F_cg = 4 L/min, and shielding gas F_shg = 4 L/min.

Comparative study of the microstructure and hardness of molybdenum layers obtained by LDED and vacuum brazing have shown that LDED will provide better performance properties due to the absence of additional fusible material in the joint (copper).

It is shown that the preferential LDED parameters of molybdenum powder on 25L carbon steel provide high hardness of the deposited layer and a high-quality transition layer at the Mo–steel boundary (no cracks and pores, mutual diffusion of Mo and Fe is observed in the boundary area).

A combined technological process for manufacturing the ‘body’ part is proposed, which consists in combining the traditional method of manufacturing a workpiece by casting from 25L steel and depositing a molybdenum layer from PMS-M99.9 powder by LDED.

The following preferential LDED parameters are proposed for PR-12H18N9T powder: power P = 100 W, V = 1150 mm/min, powder consumption F_p = 8.75 g/min) and for alloy HN71MTYuB.: power P = 100 W, V = 1400 mm/min, powder consumption F_p = 9.38 g/min, hatch distance along the X-Y axis 0.2 mm, and hatch distance along the Z-axis 0.22 mm.

The possibility of replacing brazing with LDED in the manufacture of multimaterial products of steel–nickel alloy is shown with the example of the ‘electromagnet case’ part.