2017 Volume 58 Issue 2 Pages 294-297
In selective laser melting (SLM), powder properties like size, distribution, shape, flow, and packing have effects on the process and the final parts quality, and several standards and methods are available for representing these characteristics. However, these are not enough to explain the actual packing state of the particles across the powder bed substrate. This work reports a novel method for evaluation of the powder layer density in the SLM process. The results show that the powder characteristics measured by conventional methods are not always appropriate for determining whether a powder material is suitable for SLM.
Selective laser melting (SLM) is a powder based additive manufacturing (AM) technique that allows the layer-by-layer production of 3D metal structures directly from a computer aided design (CAD) model by selective melting of powder layers using a focused laser beam1,2). This process has been recognized as a promising manufacturing technology in recent years in industries such as aerospace, medical, automotive, and consumer products, due to its ability to produce metallic parts of high geometric complexity, high density, and high dimensional accuracy3–6).
In the SLM process, it is critical to have a comprehensive understanding of the various process parameters and material properties that determine the quality of the final products1,7,8). In particular, the appropriate choice of raw materials (i.e., powder particles) for the SLM process has become increasingly important for generating the desired microstructure and mechanical properties. Furthermore, the powder spreading and deposition (i.e., formation of the powder bed) are also very important steps in SLM, and the powder layer density plays a decisive role in the density of final parts and related part properties. Generally, the powder shape, size, and distribution as well as the surface condition affect the flow and packing properties of the powder materials; and these characteristics have significant influence on the powder layer density1,9). It is also well known that the powder should have good flow characteristics to obtain a homogeneous and dense powder layer. Engeli et al.10) reported the influence of particle flow and packing behavior on the resulting part quality after SLM. Zhu et al.11) determined that the green density of the powder bed directly influenced the final density of the sintered parts in the direct laser sintering process.
Several standards and methods are available for measuring powder properties like particle size, distribution, flowability, and density12,13). However, the powder layer density as used for powder-based AM processes is different from the typical powder density measurements of apparent and tapped density. To date, few attempts have been made to determine the real density of the powder layer used for SLM. Therefore, in the present work, the evaluation of the powder layer density was investigated using two different 316L stainless steel powders.
Two types of gas atomized 316L stainless steel (SS316L) powders were used as starting materials in this work. One was supplied by Concept Laser GmbH (CL20ES, Germany) and the other by LPW Technology, Ltd. (LPW316AR, UK), respectively. The particle size distribution of each powder sample was in the 10–45 μm range with similar chemical composition, as provided by the suppliers. The morphology of the starting powders was observed with a field emission scanning electron microscope (FE-SEM, Tescan MIRA-3, Czech Republic), and the particle size distribution measurement was made with a laser particle size analyzer (Beckman Coulter LS 13 320, USA). The values of apparent (A.D) and tapped density (T.D) of SS316L powders were determined according to ASTM standards (B213-03 and B212-99) using a Hall flowmeter funnel with an orifice of 2.5 mm. Each measurement was repeated five times and the average value was used for analysis. The evaluation of particle flowability was carried out by measuring the Hall flow rate (s/50 g) as per ASTM B213-13, which is expressed as the time required for a 50 g powder sample to be discharged by gravitational force through a Hall flowmeter funnel. Furthermore, the Hausner ratio was calculated to express the flowability of the powder, which is the ratio between the tapped density and the apparent density of a powder, indicating the frictional condition in a moving powder14). In all the experiments for the characterization of the powder flowabiltiy, the powders used in this study were first dried to remove absorbed moisture.
To evaluate the powder layer density, a container was produced by SLM with the internal dimensions 10 × 10 × 10 mm (X, Y and Z) having a volume of 1000 mm3, as shown in Fig. 1(a). The experiment was conducted using a commercially available SLM machine (Mlab-Cusing, Concept Laser GmbH, Germany) which consisted of a continuous wave Yb:YAG fiber laser (wavelength 1075 nm) with a maximum power of 100 W and a laser spot diameter of 110 μm. The samples were prepared using a Z-increment under a high purity Ar atmosphere (O2 < 0.3%) at constant laser power, scanning speed and hatch spacing of 90 W, 800 mm/s and 80 μm, respectively, but with different layer thicknesses (from 25 to 75 μm). Figure 1(b) shows a schematic diagram for the SLM process used in the present study. The thin powder layers were uniformly deposited over a build plate by a moving blade and a focused laser was irradiated the powders to melt them. This process was repeated layer by layer until the final object was complete, based on the CAD data. During the process, only particles intended to become part of the container were melted; therefore un-melted power particles remained in the container. After the SLM process, the mass of the powder inside four different containers was measured using an electronic balance with an accuracy of 0.1 mg, and the powder layer density was quantitatively determined by calculating the average value.
The designed container (a) and schematic experimental procedure (b) for the powder layer density measurement.
Figure 2 displays the morphology and size distribution of gas atomized SS316L powder particles. Both the CL20ES (Fig. 2(a)) and the LPW316AR (Fig. 2(b)) powder samples consisted mainly of spherical particles but with some satellites on their surfaces. However, the LPW316AR powder contained more irregularly shaped particles with rough surfaces than that of the CL20ES powder. The particle size distribution of the both powders is revealed in Fig. 2(c). It was found that the LPW316LAR powder had a wider particle size distribution range and contained a greater amount of fine (< 15 μm) and coarse (> 45 μm) particles in comparison with CL20ES powder. The particle size values of D10, D50, and D90 (in μm) were 18.49, 31.90, and 48.10, respectively for the LPW316AR powder and 19.76, 29.39, and 42.35, respectively for the CL20ES powder. Given these conditions, the sphericity factor (S)15) was introduced to characterize the particle shapes quantitatively, as depicted in Fig. 2(d). This is one of the commonly used to express the deviation from spherical shape of a particle, and can be defined as follows:
| \[S = 4 \pi A/P^2 \] |
SEM micrographs of (a) CL20ES and (b) LPW316AR powder, (c) their particle size distribution, and (d) their sphericity factor. The inserts in (a) and (b) are the corresponding cross-sectional images.
Regarding the SLM process, the physical properties of the particles are considered the major material factors that directly affect part qualities including density, surface condition, microstructure, and mechanical performance1,7). Table 1 shows the flow rate, apparent density, tapped density, and Hausner ratio of the SS316L powder samples. The results show better flow and packing characteristics for the CL20ES powder than for the LPW316AR powder. In particular, the flow rate could not be measured for the LPW316AR powder sample, while the CL20ES powder exhibited high flowability. This can be attributed to the fact that the LPW316AR powder had a wide particle size distribution with a large amount of small particles and low sphericity (described in Fig. 2), which could increase interparticle friction more that in the CL20ES powder16,17). The high flowability of the CL20ES powder was also indicative of higher values of the other properties listed in Table 1.
| CL20ES | LPW316AR | |
|---|---|---|
| Hall flow, s/50 g | 17.13 | No flow |
| Apparent density, g/cm3 | 4.45 | 4.08 |
| Tapped density, g/cm3 | 5.04 | 4.88 |
| Hausner ratio, - | 1.13 | 1.20 |
As a general rule in the SLM process, use of powder materials with high flowability tends to guarantee reproducible deposition of single powder layers, higher powder packing density, and better the part properties1). From this point of view, the LPW316AR powder may not be regarded as the most suitable material for the SLM process. However, the powder properties discussed above are not adequate to explain the actual packing state of the particles across the powder bed substrate. In this context, Fig. 3 reveals the powder layer density of SS316L powder samples with different thicknesses of applied-layer. It was found that both powder samples showed specific layer densities that had values intermediate between the apparent and tapped density, as presented in Table 1. This could be due to compaction of the powder layers by the coater blade during deposition of the powder18).
Powder layer density of (a) CL20ES and (b) LPW316AR powder samples with different layer thicknesses.
As seen in Fig. 3(a) and (b) respectively, the powder layer density measured for all three layer thicknesses (25, 50, and 75 μm) yielded similar values, indicating that the powder layer density is independent on the layer thickness. The average layer density for the CL20ES powder was 4.61 g/cm3, corresponding to 58.4% of the theoretical density (TD), whereas the LPW316AR powder had a lower density value of 4.41 g/cm3 (55.6% TD). Interestingly, the LPW316AR powder, despite having no measurable flow rate, showed a density value comparable with that of the CL20ES. This implies, to some extent, that conventional measurement methods cannot provide enough information for determining a raw powder suitable for the SLM process.
From this, we propose a possible relationship between the powder layer density and the final part density; that the final part density increases with increase in the powder layer density. This can be attributed to the low powder layer density leading to a local lack of particles on the powder bed, and generation of irregular-shaped pores in these areas because non-uniform fusion/melting19–21). The influence of the powder layer density on the final density and resultant part properties (i.e., microstructure and mechanical performance) in the SLM process is currently under investigation, and will be presented later in a separate report.
In this study, we introduced a new methodology adequate for evaluating the true powder layer density for selective laser melting. The particle properties of CL20ES and LPW316AR powder were examined according to conventional standards and methods, including particle size, distribution, flowability, apparent and tapped density, and shape factor. It was found that the CL20ES powder exhibited better flow and performance than did the LPW316AR powder. However, the powder layer density of both powder samples had a similar value (> 55% TD), regardless of the layer thickness. As a result, we determined that these conventional methods are not always suitable for verifying the applicability of a powder for use in the SLM process.
This study was supported financially by Fundamental Research Program of the Korean Institute of Materials Science (KIMS). This work was also supported by the National Research Council of Science & Technology (NST) grant by the Korea government (MSIP), (No. CRC-15-03-KIMM).
