Technological Basis of High-Speed Direct Laser Growth of products by Heterophase Powder Metallurgy Method
Intensive development of additive technologies in the recent years has significantly improved the methods of manufacturing and processing of products [1–6]. Growth technologies are replacing the conventional methods of casting and subsequent machining, during which up to 90% of the material can be removed. Replacement of casting and machining technologies with the growth can significantly reduce the cost of parts, which is particularly important in sectors such as gas turbine engine building, aviation, astronautics [7–9].
The technologies of growth based on the selective laser sintering and selective laser melting (SLS/SLM-technology) have been brought to the practical application [10–12]. Despite the large number of studies conducted in this area and the positive experience of industrial application of SLS/SLM-technologies, the potential of the growth technologies is still implemented only partially. A major problem in the development of additive technology is the need for a substantial increase in productivity while maintaining the required quality of the grown products.
This paper contains the results of theoretical and experimental studies of the process of growing direct laser metal products from the powder materials.
Materials and methods
of the research
The experimental studies of the processes of direct laser growth have been carried out at the Institute of Laser and Welding Technologies SPbPU (ILWT) using laboratory benches based on fiber lasers LS-15 and LS-5 with a capacity of 5 kW and 15 kW, respectively (see. Fig. 1).
In conducting the experiments on laser growth, the laser power, the growth rate of layers, determined by the rotational speed of the substrate and the rate of gradual ascent of the focusing head for applying the next layer, the powder consumption, the spot diameter of the laser radiation, the powder input angle have been varied. For experimental studies of gas-dynamic processes of powder transfer, high-speed camera Citius Centurino C100 and high resolution camera Basler acA-2000gm have been used. A series of experimental studies of the effect of growth mode on the width of gas-powder jet has been performed. Software in the programming environment LabVIEW 2012 has been developed for image processing automation. Noncoaxial nozzles with the outlet diameter of 1–2 mm and noncoaxial slotted nozzle with the slit width in the range of 0.2–1 mm have been used for the formation of gas-powder jet.
Superalloy powder EuTroLoy16625G.04 by Castolin Eutectic (Inconel 625) has been used as a material for the growth; the appearance of the powder is shown in Fig. 2, the chemical composition is given in the table. Fractional composition is 53–150 µm; the shape of the particles is spherical.
Metallographic studies of the grown products have been conducted using microscope DMI 5000 (Leica) with software Tixomet. The studies of the chemical composition and distribution of chemical elements have been carried out using a scanning electron microscope Phenom ProX and microscope Mira Tescan using consoles Oxford INCA Wave 500. In order to determine the microhardness of the deposited coatings microhardness tester Micromet 5103, Buehler has been used. To determine the mechanical properties of grown products the uniaxial tension tests have ben performed. The tests have been performed on a universal testing machine Zwick / Roell Z250, Allround series. The flat samples cut from the items grown from alloy Inconel 625 have been tested in the initial state and after heat treatment (stress relief annealing at T = 1000°C, 3 hours, air atmosphere).
Results and discussion
The development of technology of direct laser growth requires comprehensive theoretical and experimental research, careful selection of the process parameters, a detailed study of the phase-structural state of the obtained layers and items, as well as a comparative analysis of mechanical and operational characteristics. First of all, it requires an understanding of the flow of gas dynamic and thermal processes in the gas-powder jet at the drop on a substrate, the value of which considerably exceeds the amount of the melting bead. The next step is to establish the optimal conditions of melting single beads that meet the requirements and comprehensive analysis, including the structural studies . In the transition from melting single layers to direct laser growth, the task of flying metal particles in a jet of carrier gas is complicated by the decrease of specific surface area at which the item grows (previous layer). The theory of the processes occurring in the gas-powder jet when it falls on a substrate taking into account heating and partial melting of the powder particles by laser radiation has been developed by the authors of this study  on the basis of the joint solution of the problems of jet decomposition when hitting the barrier and transfer of the powder particles in the jet. The mathematical model built on the basis of the developed theory of transfer of the powder  made it possible to link the structure of gas-powder jet to carrier gas consumption, the size of the nozzle and the parameters of the powder particles. The experimental studies have also focused on the selection of process parameters to ensure maximum utilization of the powder and heterophase nature of the process with an incomplete melting of metal particles, which increases the mechanical properties of the product as compared to casting and SLS/SLM-technologies.
Direct laser growth
When studying the process of direct laser growth, a nozzle is an important part of the technological tool [13, 17]. It forms a gas-powder jet and thus has a decisive influence on the process of growth. Gas-powder nozzles can be divided into two classes by the presence of axial symmetry with respect to the axis of the laser beam: axisymmetric (or coaxial) and axially asymmetric (lateral or noncoaxial). The process heads, equipped with lateral nozzles, are asymmetric relative to the direction of the tool, when using, the process of a bead growth depends on the direction of motion. Therefore, when using the noncoaxial heads, the growth should be carried out without changing the direction of movement of the tool relative to the product. For technological experiments, a noncoaxial nozzle with the circular outlet orifice with a diameter of 2 mm (Fig. 3a) has been used. Fig. 3 b, c shows the examples of products grown using said nozzle.
A gas-powder jet formed by this nozzle has a simple structure that is symmetric relative to the axis of the channel and divert with the divergence angle of about 8–10°. The width of the jet is determined by the diameter of the channel, which limits the minimum width, since the reduction of the channel diameter increases the velocity of gas and powder particles, which, after a certain threshold, negatively affects the process stability due to deformation of the surface in the growth zone under the action of gas jet.
The experiments have shown that the use of noncoaxial nozzle allows producing a body of revolution with a minimum diameter of 6 mm, and growing the products of complex geometry, but with axial symmetry. The wall thickness of grown products varies from 0.6 to 3 mm, the surface roughness of the product does not exceed 50 µm. To manufacture the products with a complex shape it is necessary to use more complex coaxial nozzles.
The process heads, equipped with a coaxial nozzle, are characterized by the parameters of growth independent on the direction of the tool. This allows the use of complex toolpaths and produce items with a more complex geometry. The most common coaxial nozzle design is shown in Fig. 4 [18, 19].
The jet nozzle is a further development of the idea of the lateral nozzle. To ensure the isotropic melting process with respect to the direction of motion, 3–4 lateral nozzles are used which are arranged symmetrically around the axis of the laser beam. When processing, the intersection area of jets is located near the pool of molten metal, which is formed by the laser beam passing through the central orifice of the nozzle. If you use a jet nozzle, a wide neck (3–6 mm) is formed. The annular nozzles are more technological. A gas-powder jet is fed through the gap between the two conical surfaces that direct and focus it. The convergence is 40–70°, the slit width is 100 µm, the diameter of the annular slit in the bottom nozzle cut is 10–20 mm. The neck of the gas-powder jet has an average size of 1.5 to 6 mm, and depends not only on the design of the nozzle, but also on its geometry.
Due to the uniform distribution of powder around the circumference of the annular nozzle, a high degree of symmetry and isotropy relative to the direction of movement is achieved. To create a uniform distribution the nozzle comprises a manifold that distributes the gas-powder jet in a circle and extinguishes the tangential velocity of the powder particles. The main parameters that determines the width of the neck of the gas-powder jet is the width of the annular channel and the distance from the lower edge of the nozzle up to the neck. A series of model experiments within the channel has been carried out for the experimental investigation of the behavior of gas-powder jet. A nozzle with an annular channel has been replaced by a nozzle with a flat slit channel (Fig. 5) with a variable slit width, allowing us to facilitate the experimental setup and the registration of the gas-powder jet shape. The area of the nozzle was adjusted so as to obtain a similar rate during the flow of the carrier gas therethrough.
One of the important characteristics of the nozzle for direct laser growth is a gas-powder jet geometry formed by it. The main effect on the performance and stability of the growth process, using a coaxial nozzle is achieved by the following: convergence angle, the width of the neck and uniform distribution of the powder relative to the axis of the laser beam.
To study the effect of the flow of powder and carrier gas on a spatial structure of the gas-powder jet, a series of experiments using a slit nozzle has been carried out (Fig. 6a), during which the flow of fed powder has been changing in the range of from 5 to 20 g/min at a constant slit width (400 µm) and flow rate of carrier gas (8 l/min). The results have shown that the width of the gas-powder jet is virtually independent on the flow rate of the powder, which is in good agreement with the theory . The volume fraction of powder in the gas-powder jet is 0.3% at a flow rate of 20 g/min and 8 l/min for powder and gas, respectively. Thus, in a first approximation, the interaction between the particles can be ignored, and the powder flow has almost no effect on the geometry of the gas-powder jet.
To determine the effect of the volume flow of the transport gas on the gas-powder jet width, a series of experiments with the consumption of 3 to 10 l/min has been carried out. They showed that the geometry of the jet is weakly dependent on the gas flow. This can be explained by the fact that the speed of individual particles at the nozzle outlet increases with increasing carrier gas flow only to a certain value (5 m/s for the powder particles with a density of about 7–9 g/cm3). For a selected range of flow rates of the carrier gas of 3–10 l/min with the slit width of 400 µm, the transport gas velocity at the nozzle is from 6 to 20 m/s and the speed of the grains of powder is locked at a maximum level of 4.5–5.5 m/s. Thus, in a first approximation, the influence of the flow of the metal powder and the carrier gas on the width of the gas-powder jet can be neglected.
The main factor affecting the width of the gas-powder jet after exiting the nozzle is a slit width of the nozzle. While moving through the channel of the nozzle, the powder particles collide with the walls of the channel thus losing a part of the normal speed relative to the wall. Due to multiple reflections of particles, their directional movement is formed, and after leaving the nozzle the jet has a directional structure with a small divergence angle. Figure 6 shows examples of images of the obtained gas-powder jets.
The processing of data showed that the jet has a normal distribution in cross section, starting from a distance of about double width of the nozzle. Fig. 7 shows the experimentally measured values of width of a gas-powder jet at a distance of 5; 7.5 and 10 mm from the nozzle, depending on the width of the slit. It is seen that the width of the jet increases monotonically with increasing slit width.
Since the thickness of the jet increases linearly with increasing distance from the nozzle, a divergence angle of the jet has been measured (Fig.8). The angle increases monotonically with increasing slit width of the nozzle, having a minimum value with a slit width of 200 µm.
In terms of the technology of direct laser growth, the reduction of the width of the neck of the gas-powder jet gives certain advantages: increased utilization of the powder, as it increases its concentration and most of it is directed in the growth zone; it increases the productivity of the process by increasing the speed of the tool at a constant thickness of the melt bead; it increases the stability of the process, as it increases the powder density gradient along the axis of the jet in the neck; it simplifies the manufacturing process of thin walls and small-scale elements.
On the basis of the experimental studies conducted, the following process parameters can be recommended for large complex shapes: slit width – 250–300 µm, the angle of convergence of the nozzle cone is 60°, the distance from the nozzle to the area of the neck of gas-powder jet – 9–10 mm. With this set of parameters, the calculated neck of gas-powder jet is ~ 2 mm. Commonly available powder fractions are also used, the nozzle is suitable for high-power laser radiation (up to 5 kW), the high consumption of the powder, and a suitable concentration and metal powder gradient is also provided.
Using the established process parameters, the bodies of revolution made of superalloy Inconel 625 have been grown, the cross-section of the wall of a grown product is shown in Fig.9.
The porosity on the test samples did not exceed 0.05 vol.%, No cracks, inclusions are detected, thus it is found that the process of intensive oxidation do not occur during direct laser growth. The microstructure is preferably cast; the longitudinal size of the dendrites varies in the range of 50–250 µm, the size of individual dendrites up to 500 µm. Layered structures with distinct boundaries between the deposited layers, which can be seen in the products manufactured using the SLM-technology , have been found. Formation of the cast structure could also be due to the original structure of the powder used, which is produced by gas atomization method. Figure 10 shows the microstructure of the powder used (a) and the sample grown at a power of 0.75 kW (b).
Based on the results of microprobe spectral analysis and mapping of the selected area, the grey area is γ-solid solution based on nickel, white mesh along the grain boundaries represents niobium and molybdenum carbides, black dots represent a finely divided silica, alumina and manganese [21,22]. Grown samples have hereditary microstructure of the powder used, there is a certain mesh carbide dissolution and the subsequent selection of individual inclusions, dispersion of which is 0.5–2 mm, the size and form of the oxides during direct laser growth is unaffected.
To evaluate the effect of the microstructure of the grown products on the mechanical properties the heat treatment has been carried out (stress relief annealing at T = 1000°C, 3 hours). The results of the tensile tests are shown in Fig. 11.
The mechanical properties of the alloy before the heat treatment: average tensile strength is 866 MPa, yield strength is 488 MPa, relative elongation is 28%. The mechanical properties after heat treatment: average tensile strength is 855 MPa, yield strength is 479 MPa, relative elongation is 27%. These values correspond to the reference data in terms of the properties of INCONEL 625 alloy in the rolling conditions: tensile strength – 827–1103 MPa, yield strength – 414–758 MPa, relative elongation – 60–30%. In order to assess the type of destruction, fractographic study has been carried out the results of which are shown in Figure 12.
The photographs show a large number of facets with traces of plastic deformation, indicating the viscous nature of the fracture. Fine grained veneers should be noted, even in the samples without heat treatment. The samples after heat treatment also have a large number of facets with traces of plastic deformation, which also speaks of the viscous nature of the fracture.
Comparing the microstructure of polished and etched samples (Fig.13) before and after heat treatment, a fine grain structure in both cases is worth mentioning.
The structure has regions with small and large grains, which indicates certain unevenness. After the heat treatment the size of carbide and oxide inclusions does not change as well as the size of the grains. The mechanical properties remain at the level of rolling, so it can be argued that the process of direct laser growth allows making products with satisfactory performance. Further heat treatment is not required.
Direct laser growth technology is a complex and multifactorial process with a large number of parameters that affect the final result. Therefore, to understand the relationships between the process parameters and optimize the technology of obtaining products with specified characteristics, as well as reducing material and time costs and risks of negative results, the study should be carried out comprehensively, i. e. experimental studies should be preceded by the theoretical studies and mathematical modeling, and the results should be verified experimentally.
The built research stand for experimental studies of gas-dynamic processes of transfer of the powder allowed us to study the effect of the geometry of the nozzle, the flow of the transport gas, the flow and fractional composition of the powder on the spatial structure of the gas-powder jet using the methods of high-speed shooting, and various options of laser illumination. The results of experimental studies of gas-dynamic processes of transfer of the powder are in good agreement with the results of mathematical modeling, which confirms the physical adequacy of the developed mathematical model. We recommend the following process parameters: the slit width – 250–300 µm, the angle of convergence of the nozzle cone – 60°, the distance from the nozzle to the area of the gas-powder neck – 9–10 mm.
The results of investigating the structure and properties of grown products showed that the mechanical properties of the alloy before and after heat treatment are at the level of the metal rolled: tensile strength averages 850–900 MPa, yield strength – 470–490 MPa, relative elongation – 28%. The microstructure of the samples from EuTroLoy 16625G.04 alloy obtained by direct laser growth are characterized by equiaxic structure with the release of the second phase (mainly carbides) along the grain boundaries, and is hereditary from metal powder, used for growth. Fracture photomicrographs have expressed patching character peculiar to materials with a sufficiently high ductility. The tests on low-cycle fatigue showed that the fatigue strength of samples grown from the alloy EuTroLoy 16625G.04 of 271 MPa at 2 · 106 cycles, which exceeds the fatigue strength of the alloy castings.
The results of the research showed that the developed technology of direct laser growth, in spite of its technological complexity, can replace the currently used technologies, providing multiple increase in productivity and material savings. The products made by direct laser growth do not require further isostatic pressing or heat treatment, which significantly reduces the time of their manufacture in comparison with both SLS/SLM-technologies and the technologies based on the casting and subsequent heat treatment and machining.
The paper was supported by the Ministry of Education of the Russian Federation, unique project ID: ПНИЭР – RFMEFI5814X0010 (PNIER – RFMEFI5814X0010).