3.2.1. Relative Density
Based on the above research, there were five process parameters of the stable molten track. With exposure time as the most important factor of energy input, four different exposure time process parameters were selected with 80 μs, 120 μs, 160 μs, and 200 μs. In addition to improving the poor efficiency, the point distance selected the larger one as much as possible in the range of the stable track area. Hence, the exposure time-point distance of 80 μs-20 μm, 120 μs-40 μm, and 160 μs-30 μm was selected to be applied to fabricate the multi-layer experiments. Also, there was a no stable molten track at the exposure time of 200 μs. Therefore, the process parameter of 200 μs-50 μm was chosen as a comparison group. The samples were fabricated by different exposure times, point distances, and hatch spaces, as shown in Table 3
. The influence of the hatch spaces, which is based on the molten width of the single track, was researched.
The multi-layer fabrications of different exposure times, point distances, and hatch spaces on the relative density were depicted as in Figure 5
, in which the high-relative-density blocks still reached 99.99% in SLM fabrication. There is a huge influence on relative densities by the exposure times, point distances, and hatch spaces. When the exposure times are from 80 μs to 200 μs, with a hatch space of 240 μm, the relative density ranges are from 99.53% to 99.99%. When the hatch spaces are from 120 μm to 240 μm, with an exposure time of 80 μs, the relative density ranges are from 99.53% to 99.99%. With the exposure times increasing, there is little damage to the hatch spaces on relative densities. At the exposure time of 120 μs, a series of the hatch spaces from 160 μm to 280 μm are nearly influences on relative densities from 99.98% to 99.99%. At the exposure times of 160 μs and 200 μs, the hatch spaces are affected on relative densities when the relative densities are from 99.82% to 99.99% and 99.52% to 99.95%, respectively.
The above results show that high-relative-density samples (99.99%) can be easily obtained within a range of SLM process parameters. However, in previous studies, high layer thickness fabricated was used to produce more pores and low relative densities. The reason mainly is the following two aspects.
Firstly, Ma et al. studied a maximum powder layer thickness of 80 μm that can be fabricated with high relative density block; the powder layer thickness increased to 150 μm, and the residual gas at the bottom of molten pool cannot come out in the rapid solidification in a timely manner [23
]. There is a significant influence on relative density with high layer thickness fabrication if the coarse powder is employed. The architecture of this layer particle for coarse powder is illustrated in Figure 6
with void area and hollow powder. There is a large void between each particle; meanwhile, the number of hollow powders increased along with the increasing powder particle size [27
]. Therefore, the layer thickness relative density of coarse powder is relatively low, and more gas components are contained in the powder layer. In Table 4
, the particle size is developed and significant recommendations are made. The relative densities can be reached 99% with low layer thickness fabrication on 316L through previous studies. Also, the layer thickness of 30–50 μm and the average powder (D50
) of 30–60 μm was massively used. Khairallah et al. studied the different powder grades of SLM, and it was confirmed that the fine powder exhibited low particle friction, high mobility, and reduced the pore between the large particles [28
]. This research aims to obtain a stable molten pool and an approximation full-relative-density block at a thicker powder layer of 150 μm; thus, the fine powder of average particle (D50
) of 18 μm was utilized. Fine powder could produce stable molten pools, which ensure that the powder of high layer thickness can be fully melted and the residual gas can come out in time.
Besides, with regard to the equipment factor, the common SLM machines were SLM Solution (Lubeck, Germany), Phenix (Riom, France), EOS (Planegg, Germany), Concept Laser (Lichtenfels, Germany), Renishaw (London, UK), Realizer (Borchen, Germany), and Trumpf (Munich, Germany), which are used in previous studies. Ahmadi et al. used Phenix to study the computational framework by SLM of 316L [16
]; Suryawanshi et al. used Concept Laser to study the effects of scanning strategy on mechanical properties [36
]; and Mao et al. used EOS to study the manufacturing feasibility of forming Cu-4Sn new material [37
]. In this experiment, the use of Renishaw is different from the previous three devices, which split the scanning speed into exposure time and point distance, called ‘’spot-to-spot formation’’, as shown in Figure 2
b. Extensive experiments indicate that point distance is divided by exposure time and is approximately equal to the scanning speed. For the same scanning speed, there is a multitude combination of exposure times and point distances. This leads to the fact the same scanning speed of Phenix, Concept Laser, or EOS was behaved to the same molten track morphology, and Renishaw can change the combination of different exposure times and point distances to obtain a more stable molten track, which makes it easier to produce high-relative-density samples during SLM processing.
3.2.2. Building Rate
The forming time of SLM mainly includes auxiliary processing time and laser manufacturing time. The auxiliary processing time is determined by the operation of equipment, including powder delivery time and substrate falling time. Also, the laser manufacturing time is determined by process parameters. Therefore, the laser manufacturing time was shortened mainly by adjusting process parameters. The building rate is defined as:
where “η” is building rate (mm3
/s), “δ” is layer thickness (mm), “h
” is hatch space (mm), “s
” is point distance (mm), and “t
” is exposure time (s). The building rate range of 0.9 mm3
/s can be realized from previous studies on SLM fabrication of 316L. Table 5
exhibits the summary of previous studies on 316L building rate, for example, Cherry et al. [21
] and Kamath et al. The authors of [20
] separately used AM250 (Renishaw, London, UK) and M2 (Concept Laser, Lichtenfels, Germany), the building rates of which are 2.48 mm3
/s and 1.67 mm3
/s, respectively. The building rate of this research is a range of 4.5 mm3
/s. The samples with a relative density of 99.99% were selected, and the building rate can be up to 12 mm3
/s, which is 3–10 times the previous 316L building rate.
3.2.3. Analysis of Forming Defects
In this research, the range of all the sample relative densities is from 99.53% to 99.99% (Figure 5
). To reveal the mechanism of forming defects, the specimen cross section was researched. There are three typical defects: the un-melted defect between the molten pools, the micro-pore defect within the molten pool, and the irregular distribution of splashing phenomenon (Figure 7
and Figure 9). When the hatch space is 240 μm at a small exposure time of 80 μs, the overlap rate is less than 20%, the defect shapes are irregular pores, and between the molten pools there is an unmelted area, as shown in Figure 7
a. The area of unmelted defect is more than 10,000 μm2
, having a great influence on the relative density. Consequently, the influence of the hatch space on overlap can result in uneven overlap and surface roughness because of surface deterioration. With the cumulative influence of layer-to-layer fabrication on many rugged depressions, the layers of molten pool cannot be spread fully and produce a larger pore between the molten pools. When the exposure times were increased from 120 to 180 μs, the unmelted defect was gone; at this time, the overlap rates were 40–50%, and the molten track was evenly arranged and overlapped with the adjacent fine overlap. This concept is illustrated in Figure 8
a, which is a schematic diagram of un-melted defects. Moreover, there are two reasons: the first is hatch space is determined by molten pool width, the second, molten pool width, is determined by exposure time. Hence, the hatch space is determined by the exposure time. It can be improved for defect-free interfaces at an overlap rate of 40–50%, and the hollow area in the previous solidified layer can be filled. As can be easily understood, the larger unmelted defects are mainly located in the interlayer binding site, resulting in poor metallurgical bonding between layers.
The residual micropore can be depicted within the molten pool as in Figure 7
b, in which the spherical pores are very small, with a general area around 100 μm2
. The location of the micropores is random within the samples, and the formation of spherical pores is not significantly related to process parameters. The experimental results indicated that micropore defects can be found both in the low exposure time of 80 μs and the small hatch space of 120 μm, or the high exposure time of 200 μs and the large hatch space of 360 μm. There are two reasons that some of the micropores are due to the SLM process chamber circulating gas and metal powder fabricated of the hollow powder with residual gas, which was dissolved within the molten pool as the molten pool in liquid phase state, and some of the residual gas at the bottom of the large molten pool failed to come out in time during the rapid solidification. As the exposure times increased from 80 μs to 200 μs, the micropore defects were slowly reduced; on the one hand, it is because a part of the dissolved gases at the bottom of molten pool came out. One the other hand, the other parts of the micropores were scanned secondarily, which also caused the micropores to be released. Therefore, it is considered that some of the micropores are caused by the circulation gas and hollow powder, which was caused not only by the process parameters, but also by the external factors. Optimization of process parameters can partly reduce the micropore phenomenon, but it cannot be completely eliminated.
At an even larger exposure time of 200 μs, there are some adhesions of the small particles on the formation surface with an average area of 100 μm2
to 1000 μm2
. The series of splashings includes spherical splashing, coarse spherical splashing, and irregular splashing [40
]. These splashing droplets are cooled and solidified in the process chamber environment to form small metal particles; some of these particles fall onto the surface of the solidified metal, and some of them fall into the molten pool or wrap around the molten pool [41
]; Figure 8
b gives the outline of a splash phenomenon. Figure 9
indicates the splash phenomenon for different exposure times from 80 μs to 200 μs. It can be observed that the smallest metal particles existed when the exposure time was 80 μs. The laser beam irradiation area temperature is presented as a Gaussian distribution. As the exposure time increased, the center temperature significantly increased, and a significant temperature gradient was produced within the molten pool has. Also, the splash of small particles continually influences the formation of the next deposited layer, reducing the ability to combine with the surrounding metallurgy.
3.2.5. Tensile Properties
The tensile properties of different exposure times of 80, 120, 160, and 200 μs are shown in Figure 12
, which possessed stable values of ultimate tensile strength (UTS) ranges from 550 MPa to 700 MPa, yield strength (YS) ranges from 450 MPa to 600 MPa, and elongation (EL) ranges from 39.7% to 41.8%. The average values of UTS, YS, and EL are 625 MPa, 525 MPa, and 39.9%, respectively. The exposure time of 80 μs possesses the highest UTS, YS, and a maximum EL of 160 μs. The SEM-fracture morphologies of tensile properties can be depicted as in Figure 13
. Obviously, all of the fracture presents a dimple rupture, and the dimple becomes coarse as the exposure times increase. The exposure time at 80 μs contains a finer dimple rupture, comparing the 200 μs with the coarser dimple rupture. Figure 13
e,f is the morphology of dimple rupture with 200 μs and 160 μs, respectively. Coarse dimple rupture leads to poor tensile properties easily, and a fine dimple rupture possesses little damage to the tensile properties. In addition, tensile properties are consistent with the microstructure. A finer equiaxed crystal can be obtained at exposure time 80 μs, resulting in a maximum of UTS and YS, and the tensile properties decrease as the grain size increases. The mechanical properties of metal are directly affected by average grain size from the Hall-Petch equation [43
]. Also, the fine grain can improve the tensile properties, and the coarse grain results in poor tensile properties easily. Grain sizes depend on the fabrication process parameters; reducing the exposure time can dominate the molten pool with a high nucleation rate and supercooling.
shows the tensile properties of different layer thicknesses; it indicates that the high layer thickness of 150 μm also can obtain excellent mechanical properties. There is no significant difference in the mechanical properties compared with the low layer thickness fabrication. It is due to the same metallurgical bonding and microstructure.