Keywords

Introduction

Additive manufacturing (AM) as a breakthrough in advanced manufacturing is utilized to fabricate near-net shaped products directly from a CAD model using a feedstock material (powder or wire material) [1]. A high power beam is applied to selectively melt the feedstock to join the layers of material on top of each other [2] to create a part with a tailored microstructure [3]. Direct metal laser sintering (DMLS) , also known as selective laser melting (SLM ), is amongst the most commonly used AM processes due to its freedom of design, short production cycle, and potential cost savings [1, 4]. The cooling rate during DMLS process can reach 103–108 K/s, which leads to evolution of ultrafine, metastable and gradient microstructure [2] and production of high strength components [5]. DMLS of light metals , specifically aluminum alloys has been under extensive studies because of these advantages [6]. In addition, lightweight characteristics, high specific strength , and good corrosion properties of DMLS -AlSi10Mg have drawn the attention of engineers and designers in recent years [7]. Despite the evolution of very fine microstructure during DMLS of metallic materials, the components usually suffer from a columnar grain structure due to epitaxial grain growth [8], which causes anisotropy of microstructure , texture, and mechanical properties [9]. It was shown in a previous study that, by changing the building direction from vertical to horizontal, columnar-to-equiaxed transition (CET) occurred in DMLS -AlSi10Mg due to variation of thermal boundaries and \( G/V \) ratio (where \( G \) is thermal gradient in \( K/m \) and \( V \) is growth rate in \( m/s \)) [10]. CET also affected Si precipitates characteristics (size and coherency) and strengthening behavior of the alloy [11]. In addition to the build geometry, height of the sample can also change the thermal boundaries (specifically, cooling rates) and microstructure of the DMLS -AlSi10Mg [12]. This could be an opportunity to design components with a gradient microstructure .

Despite the observation of microstructural variation through the height of a sample, it is not clear how both micron (i.e. dendrites and grains) and submicron (i.e. Si precipitates) characteristics of DMLS -AlSi10Mg change. Therefore, the aim of this study was to analyze the microstructure evolution in a vertically built DMLS -AlSi10Mg sample. A rectangular cuboid sample with dimensions of 10 mm × 20 mm × 40 mm was vertically built using DMLS process. The microstructure of the sample was analyzed using electron backscatter diffraction (EBSD ) and atom probe tomography (APT) techniques at various heights to understand how micron and submicron characteristics of the sample evolved during the process.

Experimental Procedure

A rectangular cuboid sample with dimensions of 10 mm × 20 mm × 40 mm was vertically built through DMLS process using an EOS M290 machine, as shown schematically in Fig. 1. The machine was equipped with a building plate of 250 mm × 250 mm × 325 mm, a 400 W Yb-fiber laser (maximum capability), and a laser beam with spot size of 100 μm. The process was conducted in an argon atmosphere with maximum 0.1% oxygen content. The sample was fabricated using AlSi10Mg virgin powder with an average particle size of 9 ± 7 µm and a chemical composition of Al-10 wt% Si-0.33 wt% Mg-0.55 wt% Fe. The process parameters used for DMLS of AlSi10Mg were recommended by EOS GmbH to obtain the least porosity in the core of the sample and reported elsewhere [10].

Fig. 1
figure 1

Schematic of DMLS -AlSi10Mg sample along with the samples used for microstructural studies

Full size image

Microstructure of the as-built sample was analyzed using EBSD at the mid-plane position along the centre line of the sample, as shown in Fig. 1. EBSD scans were conducted over an area of 400 µm × 800 µm with a step size of 0.3 µm in a field emission gun scanning electron microscope (FEG-SEM FEI Nova NanoSEM-650) equipped with a Hikari EBSD system.

APT analyses were conducted along the same locations of EBSD analyses, using a local electrode atom probe (LEAP) 4000X HR (CAMECA Scientific Instruments, Madison, WI), equipped with a reflectron. Measurements were performed in ultra-high vacuum (<4 × 10−11 Torr) at a temperature of about 60 K. Field evaporation from the samples were promoted using ultraviolet (UV) laser pulses (λ = 355 nm) with a pulse energy of 60 pJ and a pulse rate of 125–250 kHz. The target detection rate (controlled by varying the DC voltage applied to the sample) was 5 ions per 1000 pulses. Reconstruction of the acquired data was conducted using the integrated visualization and analysis software package (IVAS) v3.8 (CAMECA Scientific Instruments, Madison, WI). Sample preparation for APT was done using an NVision 40 focused ion beam (FIB) instrument (Carl Zeiss, Germany) and established lift-out techniques, with details described elsewhere [13]. Final cleaning of the specimen tips was done at a low voltage (5 kV) to minimize Ga implantation and damage to the samples.

Results and Discussion

Figure 2 shows the typical EBSD map of bottom and top of the sample. The microstructure of the sample consisted of overlapped melt pools as a result of layer-on-layer deposition [11]. A columnar microstructure is developed inside the melt pools due to epitaxial growth, while the pool boundaries are featured by equiaxed grains. It was shown previously that each elongated grain consisted of several α-Al dendrites bounded by eutectic Si walls [10].

Fig. 2
figure 2

EBSD unique color grain maps of a bottom and b top of the sample

Full size image

To better interpret the microstructural features of the alloy and the differences between bottom and top of the sample, grains were analyzed considering the grain shape aspect ratio (\( \phi \)), which represents the ratio between the minor axis (\( L_{2} \)) and major axis (\( L_{1} \)) of each individual grain (\( \phi = L_{2} /L_{1} \)). The minor and major axes of each grain were evaluated by fitting an ellipse to each grain [10]. Figure 3a shows the area fraction of grains with different grain shape aspect ratios, at the bottom and top of the sample. Columnar and equiaxed grains were defined with \( \phi \) ≤ 0.33 and \( \phi \) > 0.33, respectively [14]. Referring to Fig. 3a, 76–78% of grains are columnar in both bottom and top specimens. Moreover, the grain size area (in µm2) of the DMLS -AlSi10Mg was very consistent at the bottom and top of the samples, as shown in Fig. 3b. Such a consistency in the grains characteristics seems to be a result of similar solidification conditions along the height of the sample.

Fig. 3
figure 3

a Grain shape aspect ratio and b grain size area at bottom and top of the sample

Full size image

It was shown in a previous study [10] that during DMLS of AlSi10Mg the constitutional undercooling which determines the solidification morphology (equiaxed or columnar) is dependent on \( G/V \) ratio. In addition, as the \( \psi \)-angle (the angle between the nominal growth rate and \( \left\langle {hkl} \right\rangle \) direction of the growing dendrites) decreased, \( G/V \) ratio increased, and a lower constitutional undercooling is obtained which led to the formation of a columnar structure. Since <001> direction is the major dendritic growth (easy-growing) direction in FCC metals and alloys [15], <001> texture represents the solidification mode in DMLS -AlSi10Mg [10]. Figure 4 shows the (001) pole figures at the bottom and top of DMLS -AlSi10Mg sample, obtained from EBSD analyses. As seen, in both locations, {001} fibre texture evolved as a result of < 001 > // \( z \) (building direction) growth. The consistency of \( \psi \)-angle at the bottom and top of the sample resulted in similar constitutional undercoolings and evolution of similar grain morphologies.

Fig. 4
figure 4

(001) EBSD pole figures of a bottom and b top of DMLS -AlSi10Mg . Note: BD and RD represent building direction and radial direction, respectively

Full size image

Despite the similarities in the microstructural characteristics of bottom and top of DMLS -AlSi10Mg , submicron characteristics of these two locations were different, as shown in Fig. 5. While both Si and MgSi precipitates developed on the top of the sample (Fig. 5b), only MgSi precipitates evolved in the bottom of the sample (Fig. 5a). Moreover, by moving from the top to the bottom of DMLS -AlSi10Mg , fine and dispersed MgSi precipitates are changed to large, needle-shaped ones. The absence of Si precipitates in the bottom of the sample is a result of various heating cycles experienced by the material, which led to dissolution of Si precipitates [12]. On the other hand, MgSi precipitates grew due to diffusion of Mg in the aluminum matrix. It was shown before that Si precipitates play an important role on the strength of DMLS -AlSi10Mg [5, 11]; therefore, their absence can compromise the strength of the alloy. On the other hand, variation of the characteristics of MgSi precipitates can affect the strength of DMLS -AlSi10Mg . Hence, it is essential to investigate the effect of submicron characteristics on the strength of the alloy in detail.

Fig. 5
figure 5

3D APT atom maps showing the distributions of Al, Mg and Si elements inside the melt pools of a bottom and b top specimens. Precipitates are highlighted by iso-concentration surfaces for 3 at.% Mg or 3 at.% Si

Full size image

Conclusions

In the current study, the microstructure of a vertically built DMLS -AlSi0 Mg sample was analyzed at two extreme locations; bottom and top of the sample using EBSD and APT techniques to investigate the evolution of micron and submicron characteristics of the alloy during the DMLS process. The microstructure of the sample at the bottom and top was consistent and comprised of columnar grains developed along the building direction. The majority of the grains (76–78%) were columnar in both locations and the grain area was similar. Evolution of {001} fibre texture as a result of <001> // \( z \) (building direction) growth was the main reason for columnar structure development, considering the fundamentals of solidification and constitutional undercooling. Despite the similarities in the microstructural characteristics of bottom and top of DMLS -AlSi10Mg , submicron characteristics of these two locations were different. While both Si and MgSi precipitates developed on the top of the sample, only MgSi precipitates evolved in the bottom of DMLS -AlSi10Mg . Moreover, by moving from the top to the bottom of DMLS -AlSi10Mg , fine and dispersed MgSi precipitates are transformed to large, needle-shaped ones. Differences in submicron characteristics can potentially affect the strength of DMLS -AlSi10Mg .