2022 Volume 62 Issue 1 Pages 200-208
With the rapid development of selective laser melting technology, the effect of different process parameters on the quality of the printed parts was studied by many researchers in recent years. In this work, a comparison on the effect of laser power and scan speed which was considered as two main factors to affect laser power density, was studied. An inner structure part with overhanging surface was designed and printed to better study the influence on the surface quality caused by these two factors. The testing results revealed that with the same energy density, different performance can be observed on the overhanging and side surface quality caused by laser power and scan speed. With the increasing of the laser power, side surface roughness value showed an increasing trend due to the increasing of the temperature gradient of the molten pool while the overhanging surface quality had a descending trend. It was mainly due to the fact that to keep the same laser power density, the scan speed decreased which resulted to the increasing time for solidification of the molten pool. This phenomenon lead to the increasing of the sinking distance and the overhanging surface quality showed a decline trend.
In recent years, more and more attention has been attracted on additive manufacturing (AM) technology due to its solid freeform fabrication, material saving, time saving and so on.1,2,3,4,5) Selective laser melting (SLM), which is one of additive manufacturing technology, has got notice from a series of industries like military field, automotive field and aerospace field as it can achieve the formation of the metal part.6,7,8,9,10) What’s more, as the layer by layer forming from the bottom of the manufactured part, this method makes it possible to manufacture any complex structure part in theory.11,12,13,14) The schematic diagram was shown in Fig. 1.
The schematic diagram of the selective laser melting.
Precisely because of so many advantages this technology has, many researchers studied the effect of different parameters on the performance of the printed parts. Among all the process parameters, laser power and scan speed were considered as two main parameters due to the reason that it can highly affect the laser energy density. Creco et al.15) kept the same energy density in his work to tell the different effect of laser power, layer thickness and hatch spacing on the quality of the printed part. Chen et al.16) considered that laser power had a strong effect on the track profile which was one of the most significant factor or affect the lack of fusion. In another word, they thought laser power can highly affect the track profile to impact the overall quality of the printed part.
Some researchers focused on the effect of the laser power. Moraes et al.17) used finite element analysis to predict the temperature field distribution of the molten pool caused by laser power. He found that when SS-304L powder printed by 100 W and 200 W were recommended as 400 W was thought to generate too high temperature in the powder bed. Nakamoto et al.18) reviewed metallic powder using different laser power. They found that different laser power should be applied on different metal and a suitable laser energy was introduced to print pure titanium alloy. Dadbakhsh et al.19) used different laser power and scan speed to fabricate Al part with different ratio of Fe2O3. The testing results revealed that with the change of the ratio, the forming intermetallic showed a quite significant change which impacted the quality of the printed parts. Stopyra et al.20) varied five main process parameters, including laser power, time of exposure distance between line, layer thickness and distance between walls. They found that laser power, exposure time distance between points were three main factors to affect the quality of the printed parts. After optimization, the relative density of the titanium alloy can reach 99.9%.
As for the role of scan speed, it was also reported by some researchers. Zhang et al.21) tried to print metallic glasses using selective laser melting method. They found that laser scan speed can significantly affect the cluster changes in the printing process. Lower scan speed resulted an increasing fraction of BCC phase while higher scan speed leaded to a suppression of the nucleation of crystal phase. The effect of scan speed on the density and mechanical properties of the Ti-6Al-4V (TC4) part was investigated by Wang et al.22) and found that with the changing of the scan speed, the microstructure of the printed was different which resulted to the different property of the printed parts. Li et al.23) used finite element analysis method to explain the effect of scan speed and laser power on the quality of the printed parts. In this work, they used simulation method to illustrate the different behavior of the molten pool and microstructures were captured using FEM to verify the simulation results. Khorasania et al.24) tried to study the effects of different process parameters. They found that lower scan speed, laser power and scanning pattern angle can help to print a part with relative better density.
Some works have also been done on the laser energy density to observe its effect on the quality of the printed parts. Kluczynski et al.25) studied the influence of energy density on the quality of the printed parts. The testing results revealed that the energy density not only had an impact on the microhardness of the printed parts, the porosity also showed an improvement with a modification of the parameter combination. Choi et al.26) varied the laser energy density to study its effect on the microstructure and densification of the printed SS316L prats. The results indicated that the laser power density showed a quite significant impact on the property of the printed parts, especially the porosity of the 316L part. Han et al.27) revealed that both high relative densities and microhardness can be obtained by optimizing the energy density and the phase formation also showed some difference on different directions. Larimian et al.28) studied the scan strategy and energy density on the mechanical properties of 316L stainless steel samples and found that by optimizing the parameter combination, better densification and fine microstructure can be obtained. P. A. Lykov29) used different volumetric energy density parameters to process different kinds of metal and the obtained dependencies showed bad correlations with energy density and the sample density. According to the data obtained, author modified the volumetric energy density and found a god correlation with each other. Wang et al.30) revealed that the laser energy density can significantly affect the microstructure of the printed alloy while porous may be formed under an ultrahigh laser energy density.
From the illustration given above, it can be concluded that energy density which was determined by the laser power and scan speed showed a significant impact on the densification and porosity of the printed parts. However, the previous work done before to compare the effect between laser power and scan speed showed that their main differences were the porosity of the printed parts. It raised a question: Whether this difference was caused by the laser power density or the scan speed and laser power itself. Based on this, energy density was kept the same in this work and the laser power and scan speed were changed accordingly. To give a more intuitive understanding on the different performance of the printed part caused by laser power and scan speed, an inner structure was designed. Quantificat evaluation was proposed by observing the variation trend on roughness value of different surfaces caused by laser power and scan speed respectively.
TC4 powder used in this work was provided by SHENZHEN MINATECH Co., LTD, China manufactured using Plasma Rotating Electrode Process (PREP) method. To increase the flowability and sphericity to further improve the property of the powder, ball milling machine and tube furnace were employed this work and the relative information about the powder after processing was shown in Table 1. The preset temperature in the tube furnace was shown in Fig. 2.
| Powder | Sphericity | Flowability/s | D10/μm | D50/μm | D90/μm |
|---|---|---|---|---|---|
| Ti6Al4V | 0.982 | 12.3 | 16.3 | 18.5 | 21.2 |
| O/wt.% | N/wt.% | H/wt.% | C/wt.% | Al/wt.% | V/wt.% |
| 0.176 | 0.024 | 0.0064 | 0.010 | 5.74 | 3.83 |
Temperature set used for powder drying.
Selective laser melting machine used in this work was provided by Nanjing University of Aeronautics and Astronautics (NUAA). The model of this instrument was RAP-IV. The process parameters used in this work was shown in Table 2. To keep the value of the energy density, laser power and scan speed were changed according to the Eq. (1) and the relative process parameters were shown in Table 2.
| (1) |
| Parameter | Scan strategy | Hatch spacing | Spot size | Defocusing amount | Layer thickness |
|---|---|---|---|---|---|
| Value | Zigzag shape | 0.1 mm | 0.1 mm | 0.0 mm | 0.12 mm |
In this equation, ρ1 was the energy density (J/mm3) used in this work, P1 was the laser power (W) S1 was the scan velocity of the laser system (mm/s), hd was the hatch spacing (mm) and lt was the layer thickness (mm).
The flatness value used to describe the surface quality was measured and calculated using Trilinear Coordinates Measuring Instrument provided by Hexagon Metrocogy, RA-7525SEI-4, Sweden and Polyworks (software). Triangular laser measuring technique was taken into use to scan the printed parts as the existence of the inner hole and the scanner head was HP-L-20.8. The working distance of this scanner was 180±40 mm and the shape error was approximately 9 μm under the frequency of 100 Hz. The scan mode used here was ultrafine quality to ensure the accuracy of the obtained cloud point data. The fitting image was obtained by the point cloud data and the imported designed three-dimensional model. The flatness value was gained by the extraction of the specific surface and the calculation was done by the software.
The surface roughness was also studied in this work to further illustrate the surface quality. It was mainly measured using Roughometer supplied by Mitutoyo, Japan. Due to the limitation of the total length of the printed samples, the sampling length taken here was 1.5 mm and the accelerating and decelerating length was both 0.75 mm. The interval number used in this work was 5 while the interval length was 1 μm to decrease the impact of the error happened in the measuring process.
To better explain the variation trend of the surface quality, Scanning Electron Microscope (SEM) provided by Carl Zeiss, Sigma 300, German was used here to observe the morphology of the inner structure surfaces. The accelerating voltage of this instrument was 20 KV with a SE2 type detector. The working distance of this detector was 8.7 mm and the magnification was 42X used in this work. White Light Interferometer (WLI) supplied by RTEC, MFD-D, USA was also taken into use to better observe the morphology of the inner surfaces. The magnification of this instrument was 20X while thew aperture was 0.4 mm with a working distance of 4.7 mm. The field of view, spatial sampling and resolution was 860*650 μm2, 0.34 μm and 0.35 μm respectively. The instruments used in this work was shown in Fig. 3.
The SLM machine (a), ball milling machine (b), trilinear coordinates measuring instrument (c), roughometer (d), white light interferometer (e) and scanning electron microscope (f) used in this work.
To have a better understanding on the different performance of the molten pool caused by laser power and scan speed, inner structure part was designed and the design sketch was shown in Fig. 4. The designed 3D model was converted to STL format, which is the most commonly used format in 3D printing, and imported into the software used for control. The process parameters were shown in Tables 2 and 3.
Design sketch of the inner structure part.
| Laser power | 160 W | 180 W | 200 W | 220 W | 240 W |
|---|---|---|---|---|---|
| Scan speed | 800 mm/s | 900 mm/s | 1000 mm/s | 1100 mm/s | 1200 mm/s |
The printed samples were cutted off from the baseplate to using low-speed wire cutting machine provided by Wuxi Institute of Technology to guarantee its precision. Trilinear coordinates measuring instrument was then employed to study the flatness of the positive surface (PS), side surface-right (SS-R), side surface-left (SS-L) and overhanging surface (OM). To better depict the variation trend of the surface quality, surface roughness was also discussed in this work. To measure the inner surface roughness, low-speed wire cutting was also used to segment into two pieces as shown in Fig. 5.
The lower half (a) and upper half (b) of the printed sample after cutting.
To better explain the observed variation trend and different performance of surface quality caused by laser power and scan speed, SEM and WLI were used to observe the morphology of the inner surface in full view and part separately and finite element analysis carried by ANSYS APDL was also used in this work to further verify the explanation given in this work.
Flatness of the four different inner surfaces were measured by Polyworks and the measured data was shown in Table 4. To make a more intuitive comparison, line chart was plotted shown in Fig. 6.
| 160 W/800 mm/s | 180 W/900 mm/s | 200 W/1000 mm/s | 220 W/1100 mm/s | 240 W/1200 mm/s | |
|---|---|---|---|---|---|
| Positive surface (PS)/mm | 0.060 | 0.058 | 0.063 | 0.062 | 0.062 |
| Side surface-Right (SS-R)/mm | 0.070 | 0.073 | 0.073 | 0.079 | 0.083 |
| Overhanging surface (OS)/mm | 0.102 | 0.096 | 0.087 | 0.085 | 0.084 |
| Side surface-Left (SS-L)/mm | 0.069 | 0.070 | 0.074 | 0.076 | 0.081 |
Variation of the flatness printed under different parameters combination.
The variation trend of the flatness shown in Fig. 6 indicated that the flatness value of the positive inner surface was significant lower compared to other three inner surfaces while the overhanging surface flatness value was the highest in all these four inner surfaces. It was mainly caused by the different conditions of forming of these four inner surfaces. The positive surface was formed on the solid layer which was able to provide a strong support to the molten pool. The upper layer was also not supplied in the forming process which made it impossible to cause the powder bonding on the positive surface. As for the side surfaces (including right side surface and left side surface), although they were formed on the solid layer, powder beside the side surfaces had been supplied in the molten pool solidification process. This resulted to the extra powder bonding on the side surfaces which lead to the decreasing of the surface quality. Overhanging surface was formed on the unmelted powder which was unlikely provide enough support force for the molten pool. It lead to the sinking of the molten pool which caused the surface quality reduction. Moreover, the low thermal conductivity of the powder and capillary force between the molten pool and unmelted powder in the inner structure further decreased the overhanging surface quality. The schematic diagrams of the forming mechanism of different inner surfaces were shown in Fig. 7.
Overall image of the inner structure (a), schematic diagram of the formation of positive surface (b), side surface (c), overhanging surface (d) and the support force provide by solid metal (e) and powder (f) respectively.
Compared to the surface quality printed using different laser power and scan speed combination under the same energy density, it can be seen that the flatness value of the positive inner surfaces kept at about 0.06 mm and little changes can be seen with different parameters combination. As for the side surface flatness, it can be seen that with the increasing of the laser power, the flatness value showed a small reduction form about 0.070 mm to 0.083 mm. It was mainly caused by the extra powder bonding on the side surface. As is known to all, different laser power leaded to the change of the temperature gradient of the molten pool. Higher power increased the area of the melting zone which resulted to more powder bonding. With the reduction of the laser power, the measured flatness value of the overhanging surface showed a quite significant decreasing trend from 0.102 mm to 0.084 mm. It was mainly due to the fact that to keep the same energy density, the scan speed dropped with lower laser power value. It resulted to the addition of the solidification time which caused the further increasing of the sinking distance. This phenomenon was obvious harmful to keep the overhanging surface quality. The morphology of the inner surface was also listed here to verify the explanation as shown in Fig. 8.
The morphology of the positive inner surface (a), side surface (b) and overhanging surfaces printed under 160 W/800 mm/s (c) and 220 W/1100 mm/s (d) respectively.
The difference of the morphologies of the three different inner surfaces further verified the statement given above. The molten pool sinking can be clearly found on the overhanging surfaces while powder bonding was also quite significant on the side surfaces. The different morphology on the overhanging surfaces printed under different scan speed was a good corresponding to the measured data given before. Due to the lack of support structure in the printing process, the malformation morphology of molten pool can also be seen on the side surface caused by the surface tension in the forming process. Finite element analysis was also used in this work and the variation of the temperature of the molten pool on the overhanging surface (the bottom node of the overhanging layer) shown in Fig. 9. Higher maximum temperature and faster cooling speed can be seen on the sample printed using higher laser power while a much more stable temperature reduction can be found on the sample printed using higher scan speed. Different performance of the molten pool futher resulted to the different performance of the surfaces.
Nephogram of temperature in the printing process (a) and variation of the molten pool temperature (b) printed under different parameter combination. (Online version in color.)
The longer solidification time of the molten pool printed using lower scan speed further verified the explanation given above. To have a more comprehensive study, roughness of these four inner surfaces was also measured by Roughometer. The measured data and the line chart was shown in Table 5 and Fig. 10 respectively.
| 160 W/800 mm/s | 180 W/900 mm/s | 200 W/1000 mm/s | 220 W/1100 mm/s | 240 W/1200 mm/s | |
|---|---|---|---|---|---|
| PS/μm | 7.69 | 7.45 | 7.72 | 8.09 | 8.77 |
| SS-R/μm | 9.23 | 9.60 | 9.81 | 10.22 | 10.90 |
| OS/μm | 15.17 | 14.25 | 13.38 | 12.93 | 12.66 |
| SS-L/μm | 9.35 | 9.81 | 9.88 | 10.39 | 11.01 |
Variation of the roughness printed under different parameters combination.
From the line chart, it can be found clearly that the positive surface roughness value had a quite significant increasing from 7.72 μm to 8.77 μm with the increasing of the laser power from 200 W to 240 W. It was mainly caused by the powder splashing in the printing process due to the large deviation gap of the temperature gradient of the molten pool with such a high laser power. This phenomenon can also be observed in the forming process shown in Fig. 11. Extra powder bonding on the side surfaces showed a significant impact on the side surface roughness. The roughness value showed an increasing trend with the increasing of the laser power. As for the overhanging surface roughness, its value was more related to the sinking distance of the molten pool instead of the powder bonding combined to the measured data and the morphology shown in Fig. 12 and the overall variation trend showed a good agreement with the variation trend of the overhanging surface flatness. To better validate the explanation given here, surface morphology was observed by WLI and defects of the printed parts were captured by OM shown in Figs. 12 and 13 respectively.
The powder splashing phenomenon observed under 240 W laser power.
The morphology of the positive inner surface (a), side surface (b) and overhanging surface (c) captured by WLI.
The defects of the samples printed under 200 W (a) and 240 W (b) respectively.
From Fig. 12(c), it can be seen that obvious altitude difference was seen on the overhanging surfaces caused by the sinking of the molten pool and the powder bonding was clearly found on the side surface. As for the positive inner surface, it was in a quite flat status. The defects shown in Fig. 13 verified the presence of the splashing phenomenon of the powder in the printing process under high laser power due to the existence of the pores in the sample printed under 240 W laser power while little obvious pores can be seen in the samples printed under 160 W–200 W laser power.
Combined to the experimental data and morphology of the samples printed under different parameters combination, it can be concluded that scan speed showed a significant impact on the solidification of the molten pool which can greatly affect the overhanging surface quality due to the lack of hard support from solid metal. Laser power, however, had an influence on the powder bonding on the side surface caused by the change of the temperature gradient. When excessive laser power was employed in the printing process, powder splashing caused by the large deviation of the temperature gradient between molten pool and adjacent powder leaded to the decreasing of the positive inner surface quality. Taken all these factors into consideration, the parameters combination of laser power 200 W and scan speed 1000 mm/s was considered rather suitable in this work to print a well-performed inner structure part.
In this work, an inner structure part was designed and printed to compare the different effects of the laser power and scan speed which were considered as two main factors to affect energy density. Energy density was set to 0.06 J/mm3 and these two parameters were changed accordingly to prevent the influence of energy density. Positive, side and overhanging inner surface quality were studied in this work to give a more intuitive understanding on the different performance of the printed parts caused by laser power and scan speed respectively. Detail conclusions were drawn as follows.
(1) Laser power had a greater impact on the shape of the temperature gradient of the molten pool. This may result to the extra powder bonding on the side surface which caused the decreasing of the side surface quality. Excessive laser power also leaded to the powder splashing in the forming process due to the too large deviation of the temperature between molten pool and adjacent powder. It resulted to the increasing of the pores in the printed parts.
(2) Overhanging surface quality was greatly affected by the scan speed due to the time cost in solidification process. Lower scan speed leaded to the increasing of the solidification time while the sinking distance of the molten pool also increased in this process. It likely resulted to the droplets phase morphology of the molten pool which significantly lower the quality of the overhanging surface.
(3) Combined to all the measured data and morphology observed in this work, 200 W laser power and 1000 mm/s scan speed were considered as a suitable parameter combination to print inner structure part under 0.06 J/mm3 laser energy density.
This study is supported by The Startup Foundation for Introducing Talent of NUIST Binjiang College (550220018#2020r018), the National Defense Science and Technology Innovation Zone Project and the Additive Manufacturing Products Supervision and Inspection Center of Jiangsu Province, Wuxi Institution of Supervision & Testing on Product Quality.
