1 Introduction

Selective laser melting is one of the powder-based additive manufacturing (AM) techniques that use high-intensity laser to selectively melt the metal powder bed layer by layer. Compared with traditional manufacturing technology, SLM has the following advantages: one-step forming; short production cycle, high material utilization rate, superior comprehensive mechanical properties, and production of complex parts [1,2,3]. Due to its excellent corrosion resistance, mechanical properties and biocompatibility, stainless steel 316L (SS316L) is widely used in marine, biomedical facilities and fuel cells [4, 5].

The SLM process parameters determine the surface quality of the processed parts, where laser power, scanning speed, layer thickness and hatch spacing are the most influential. Many researchers have evaluated the effects of these parameters on the performance of processed parts through the trial and error method [6,7,8]. Niendorf et al. [9] found that laser power had a great influence on the surface morphology and particle agglomeration of SLM SS316L parts. A much smoother surface was formed at a relatively higher laser power. However, at a lower laser power, many agglomerates appeared on the surface. Liverani et al. [10] studied the SLM SS316L parts processed with a laser power of 100 W to 150 W and a hatch spacing of 0.05 mm to 0.07 mm. The laser power was found to have the strongest influence on the mechanical properties, while the hatch spacing did not produce significant effects. Li et al. [11] reported SLM fabrication of SS316L using scanning speeds of 150 mm/s, 60 mm/s and 20 mm/s. With the decrease in the scanning speed, the width and depth of the molten pool increased and less powder particles cohered in the melt tracks. However, there are still some problems in the SLM technology, such as spheroidization, warpage, and microcracks, etc. [11,12,13,14,15,16]. Due to the diversity of materials and the complexity of the process, a large amount of experimental research is required, which needs a lot of cost. Moreover, the thermal properties of the metal powder change with the temperature during the experiment. It is found from the above literature that many scholars tend to study the influence of a certain process parameter or several process parameters on microstructure and mechanical properties of SLM-processed parts. In a fact, the quality of SLM-processed products is affected by multiple parameters. In order to provide a suitable processing window for the additive manufacturing process, laser energy density (LED) must be used as a comprehensive parameter to evaluate its impact on the SLM process. In this study, a finite element model (FEM) has been established to judge the temperature field of the molten region by using the thermal property parameters. After optimization of the parameters, the optimum factor levels were verified using experiments. The range of process parameters suitable for the SLM-processed SS316L was found. Then, the effects of laser energy density on the surface quality and mechanical properties of SLM SS316L parts were studied experimentally, which was beneficial to the optimization of the process parameters for processing complex parts.

2 Model for the SLM process

2.1 Finite element model

In this study, SOLID70 was employed to build the finite element analysis model. SOLID70 has a 3-D thermal conduction capability. The element has eight nodes with a single temperature degree of freedom at each node. The element is applicable to a 3-D, steady-state or transient thermal analysis. The element also can compensate for mass transport heat flow from a constant velocity field. The size of the top SS316L powder layer and the substrate layer was 2.6 × 0.6 × 0.12 mm3 and 2.4 × 0.4 × 0.03 mm3, respectively. Considering the computational efficiency, the powder bed was fine meshed with hexahedral elements (0.05 × 0.05 × 0.03 mm3), while the much coarser tetrahedron elements were adopted for the substrate (Fig. 1a). The time step used by the model calculation is 4.55 × 10−5 s. From Fig. 1b, the laser beam scans the metal powder in a zig–zag scanning path with four continuous scanning tracks along the x-axis. Six typical points were chosen to study the SLM process. P1, P2, and P3 were the starting point, the middle point, and the ending point of track I, respectively, which can reflect the temperature field distribution of the same track. P4, P5 and P6 were the middle points of tracks II, III and IV, respectively, which can reflect the combination between two adjacent tracks.

Fig. 1
figure 1

a Three-dimensional finite element model (side view); b laser scan strategy of SLM process

Full size image

2.2 Heat source model

In this study, a fiber laser was used during the SLM process (Nd:YAG, λ = 1064 nm), so the heat source model had a Gaussian distribution. The maximum heat flux density q(r) can be expressed as follows [17]:

$$ q\left( r \right) = \frac{2\alpha P}{{\pi \omega^{2} }} \cdot \exp \left( { - \frac{{2r^{2} }}{{\omega^{2} }}} \right) $$
(1)

where α represents the heat absorption rate of the laser beam into the powder, P is the laser power; ω represents the radius of the laser beam, r represents the radial distance between any node and the laser beam.

During the melting process, when the temperature is higher than the melting point of the material, the powder undergoes a “solid–liquid–solid” state. Therefore, the application of temperature conditions is a problem that should consider the latent heat of phase change. The solution is achieved using ANSYS to define the enthalpy of the material at different temperatures [18]. The enthalpy H is related to the density ρ and specific heat capacity C of the material, which can be described as:

$$ H = \int\limits_{0}^{t} {\rho C\left( t \right){\text{d}}t} . $$
(2)

2.3 Initial and boundary conditions

In SLM, when the laser scans the metal powder bed, only a portion of the laser energy is absorbed by the powder. Apart from heat conduction, convection and radiation problems should be incorporated.

The initial condition of the temperature field is defined as [19]:

$$ T(x,y,z,t)|_{t = 0} = T_{0} $$
(3)

where T0 is the ambient temperature (293 K).

The boundary condition can be defined as:

$$ k\frac{\partial T}{\partial n} - q + q_{\text{c}} + q_{\text{r}} = 0\quad (x,y,z) \in S $$
(4)

where S is the surface of the laser irradiation, n is the normal vector of the surface S, q is the heat conduction, qc and qr are the heat convection and heat radiation, respectively, which can be represented by [20]:

$$ q_{\text{c}} = h(T - T_{0} ) $$
(5)
$$ q_{\text{r}} = \sigma \varepsilon (T^{4} - T_{0}^{4} ) $$
(6)

where, h is the heat transfer coefficient (10 W/mK), σ is the Stefan–Boltzmann constant (5.67 × 10−8 W/m2 K4), and ε is the thermal emissivity.

2.4 Material properties

The thermal properties of stainless steel are shown in Table 1. However, due to the gaps between the powder particles and the shielding gas, the thermal conductivity of the powder bed is significantly reduced. The thermal conductivity of the powder bed kp can be expressed as Eq. (7) by [21]:

$$ k_{\text{p}} = \frac{{k_{\text{s}} (1 - \varepsilon_{\text{p}} )}}{{1 +\Psi {\raise0.7ex\hbox{${k_{\text{s}} }$} \!\mathord{\left/ {\vphantom {{k_{\text{s}} } {k_{\text{g}} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${k_{\text{g}} }$}}}} $$
(7)

where, ks and kg are the thermal conductivity of the solid material and the ambient gas, respectively; εp is the porosity between the powders and the Ψ is the index factor.

Table 1 Thermal property parameters of 316L stainless steel [22]
Full size table

3 Experiments

In this study, gas-atomized SS316L powder was used, and Table 2 lists the composition of the powder. From Fig. 2, the powders presented with a mostly spherical shape with no agglomeration; some particles were attached to small satellites. The size of the powder particle ranged from 3 to 50 μm with a mean particle size of 27.82 μm.

Table 2 The compositions of SS316L powder (wt%)
Full size table
Fig. 2
figure 2

SEM morphology: a size distribution; B particle size of the powder

Full size image

An SLM-150 device with an Yb-Fiber laser (1064 nm) was used to fabricate the SS316L sample. In the SLM process, a “zig–zag” scanning strategy was applied when fabricating the sample. To avoid oxidation of the molten pool, the print chamber was filled with argon. The main direction of the SLM-processed cubic parts is shown in Fig. 3a. Detailed dimensions of the tensile samples are shown in Fig. 3b. The effect of laser power, scanning speed, hatch spacing and layer thickness on the SLM process was analyzed by optimization experiments (Table 4). Considering the combined effects of these four factors on the SLM process, the laser energy density (LED) was used as a comprehensive parameter to evaluate its impact on the SLM process. The laser energy density (LED), laser energy volume, is defined as (8):

$$ {\text{LED}} = \frac{P}{vsh} $$
(8)

where P is the laser power, v is the laser scanning speed, s is the hatch spacing, and h is the layer thickness of the powder bed.

Fig. 3
figure 3

a The schematic of the main directions of SLM-processed parts; b geometry of SLM-processed sample for tensile testing

Full size image

The Archimedes method was used to obtain the relative density of the samples. An automatic microhardness tester (FALCON 500) was utilized to obtain the microhardness under a load of 0.981 N with a duration of 10 s. The surface roughness of the SS316L was tested and analyzed using a laser confocal microscope (VK-X250). Samples were cut, polished and etched for metallographic examination. The top-surface microstructures were observed by a scanning electron microscope (SEM) (JSM-7001F). X-ray diffraction (XRD) measurement (Bruker D8 Discover diffractometer) was performed to analyze the phase constitution of the samples with 2θ ranging from 20° to 80° with a scanning speed of 2°/min. Samples for tensile were prepared according to GB/T 228.1-2010 standard. The tensile tests were performed using a universal testing machine (UTM4104) for all samples with a cross head velocity of 1 mm/min at ambient temperature. The tensile fracture surfaces of all the samples were characterized using a scanning electron microscope (SEM).

4 Results and discussions

4.1 Simulation results

Figure 4a depicts the temperature distribution measured at P4 under different laser powers as a function of laser scanning time, where the slope of curve represents the cooling rate at P4. When the laser approaches P4, its temperature rises rapidly; when the laser is far from P4, its temperature drops rapidly, which in turn produces a high cooling rate. As the laser power is increased from 100 to 200 W, the maximum cooling rate of P4 is increased monotonically from 5.3 × 106 to 8.5 × 106 K/s and the maximum temperature is increased from 1636 to 2585 K. At a laser power of 100 W, the maximum temperature of 1636 K is slightly lower than the melting point of the stainless steel powder (1672 K). A low laser power causes the powder to melt insufficiently, resulting in the printed part having low density and many pores, thereby affecting the mechanical properties. In order to ensure that the stainless steel powder can be fully melted, the laser power of the SLM needs to be in the range of 100–200 W.

Fig. 4
figure 4

Temperature distributions of P4 as a function of time a at various laser power (v = 1100 mm/s; s = 0.08 mm; h = 0.03 mm); b at various scan speed (P = 200 W; s = 0.08 mm; h = 0.03 mm); c at various hatch spacing (P = 200 W; v = 1100 mm/s; h = 0.03 mm); d at various layer thickness (P = 200 W; v = 1100 mm/s; s = 0.08 mm)

Full size image

Figure 4b presents that the maximum cooling rate of P4 is increased from 3.8 × 106 to 5.3 × 106 K/s, and the maximum temperature decreases from 3637 to 2585 K as the scanning speed is increased from 500 to 1100 mm/s at P = 200 W. At a high scanning speed (1100 mm/s), the SLM process produces a large cooling rate, resulting in a large temperature gradient. When the scanning speed is low (500 mm/s), the SLM process produces a lower cooling rate and a higher temperature. Too low or too high scanning speed will cause residual stress, cracks, coarse grain, etc., which will cause a decrease in mechanical properties in the SLM parts.

The effect of the hatch spacing on the temperature distribution measured at P4 is shown in Fig. 4c. As the hatch spacing is increased from 0.04 to 0.12 mm, the maximum temperature decreases from 2815 to 2280 K. For a given laser power and scanning speed, as the hatch spacing is increased, less energy per unit area and relatively lower temperature are obtained. A relatively low temperature tends to result in insufficient melting of the powder particles and poor metallurgical bonding between adjacent tracks.

The layer thickness is also an important parameter in the SLM process. A larger layer thickness will affect the interlayer bonding, while a smaller layer thickness requires more layers to be melted, which will affect the forming efficiency. As the layer thickness of the powder is increased, the maximum temperature is reduced (Fig. 4d). An increase in the layer thickness results in less energy penetration into the powder at the bottom of each layer; this means that the powder at the bottom of each layer cannot be fully melted, which results in many defects in the finished part.

After finite element simulation, the optimum factor levels suitable for SLM processing SS316L is laser power of 100–200 W, scanning speed of 500–1500 mm/s, hatch spacing of 0.04–0.12 mm, and layer thickness of 0.03–0.06 mm.

In order to observe the change of temperature field in the same scanning orbit and different orbits, the continuous scanning process under a set of parameters was also studied. Figure 5 presents the temperature contours of the typical points (P1–P6). The initial process parameters are shown in Table 3. The maximum temperature of the molten pool rises rapidly from the preheat temperature (373 K) to above the powder’s melting temperature (1673 K). It can be observed from Fig. 5 that as the laser moves, the thermal field and the molten pool move along the track. The temperature gradient in front of the center point of the heat source is much larger than that behind the center of the heat source. In the other temperature simulations, there have also been reports of a similar phenomenon. Rapidly cooled molten material has better thermal conductivity than untreated powder [22].

Fig. 5
figure 5

Temperature contours of the typical points of laser action on the powder layer a P1; b P2; c P3; d P4; e P5 and f P6

Full size image
Table 3 Initial selected process parameters
Full size table

The temperature curves over time of the points (P1, P2 and P3) on the same track are depicted in Fig. 6a. The temperature trend of the points on the same track is similar. The highest peak appears at the time when the laser spot is closest to the point. When the laser spot is far away, the temperature drops. Compared to the heating process, the cooling process of the point has a lower temperature gradient. From Fig. 6a, the maximum temperatures of P1, P2 and P3 were 2284 K, 2289 K and 2294 K, respectively. Due to the pre-heating effect the laser energy has on the powder, the maximum temperature gradually increases.

Fig. 6
figure 6

Temperature curves over time of a the points (P1, P2 and P3) at the same track and b the points (P2, P4, P5 and P6) at the adjacent tracks

Full size image

The temperature curves over time for the points (P2, P4, P5 and P6) on adjacent tracks are also shown in Fig. 6b. The maximum temperatures of P2, P4, and P5 are 2289 K, 2585 K, and 2637 K, respectively, and the maximum temperature rises due to the accumulation of energy. However, it is found that the maximum temperature rises by 296 K from P2 to P4, while from P4 to P5, it only rises by 52 K. The increase in amplitude is small, because the heat accumulation effect reaches its limit. Due to the laser beam scan finishing, the maximum temperature drops as the laser moves to P6 on the last track.

4.2 Experimental results

4.2.1 Surface roughness, density and surface topography

The surface roughness Sa and the relative density of all samples processed at different LEDs are reported in Table 4 and Fig. 7. As shown in Table 4, the single factor analysis method was employed to analyze the influence of SLM parameters on the performance of the SLM part. Each factor is divided into 4 levels based on the range of the finite element analysis result. For instance, the influence of scanning speed can be analyzed from the test results of sample 1-1–1-4 and the influence of laser power can be got from that of sample 1-1, 2-1 and 3-1. The two curves in Fig. 7 are fitted curves for surface roughness and density. It is clear from the results that the surface roughness and relative density of SLM SS316L depends on the value of the LED. Within the entire range of the LED (37.88–250 J/mm3), it can be divided into three regions, A, B and C, corresponding to three different melting mechanisms.

Table 4 SLM process parameters, density and surface roughness for all SLMed SS316L samples
Full size table
Fig. 7
figure 7

Density, surface roughness Sa versus LED for SLM-processed SS316L samples

Full size image
  1. (I)

    Region A: when the laser energy density is 37.88–75.76 J/mm3, this is called the partial melting zone. In this region, the sample has a relative density of less than 99% and a surface roughness Ra of more than 10 μm. Insufficient laser energy density cannot induce significant melting of SS316L powders. This will inevitably induce unfavorable surface tension and wetting characteristics [13]. The balling effect on the surface of the samples is a serious problem.

  2. (II)

    Region B: when the LED is between 75.76 J/mm3 and 151.52 J/mm3, the relative density of this region is higher than 99% and the surface roughness is less than 10 μm. Enough liquid phase and favorable wetting characteristics are suitable for the formation of continuous melting tracks [23, 24]. The balling effect obviously eases, and the surface quality becomes better.

  3. (III)

    Region C: as the LED increases above 151.52 J/mm3, the relative density is greatly reduced, and the surface roughness sharply increases. A high LED causes the temperature of the molten pool to increase sharply. The molten pool is affected by the recoil pressure generated by the evaporating gas, causing the liquid in the molten pool to flow violently and severely splash [25].

To highlight the characteristics of different regions, the top surface topography of the SS316L samples printed at three different LEDs, 37.88, 119.05 and 166.67 J/mm3, is shown in Fig. 8. At a lower LED (37.88 J/mm3) within zone A, it is found that ellipsoidal and spherical particles and some large cavities appear on the surface (Fig. 8a). A lower LED results in a decrease in the molten pool’s temperature and a reduction in the contact region between the molten pool and the metal powder particles. Significant transverse shrinkage, distortion and discontinuous tracks are found to occur. In the case of the optimum LED (119.05 J/mm3) within zone B, a dense surface with continuous tracks and fewer spherical particles are formed (Fig. 8b). At a higher LED of 166.67 J/mm3, remelting and scorching on the surface can be observed due to the too high energy level (Fig. 8c).

Fig. 8
figure 8

SEM images showing topography of top surface of SS316L printed at three different LEDs: a 37.88 J/mm3; b 119.05 J/mm3 and c 166.67 J/mm3

Full size image

4.2.2 Defects

Figure 9 shows the defects in the cross section of SLM SS316L at different LEDs. In Fig. 9a1, many irregular cavities associated with unmelted powders can be observed in the sample printed at an LED of 37.88 J/mm3. It is found that the powders do not completely melt due to insufficient laser energy, which results in poor fusion between adjacent melt tracks. In addition, at a lower LED, the energy absorbed by the powder per unit time is also small, causing the temperature in the molten pool to be too low, thereby increasing the melt viscosity and lowering the melt fluidity and wettability [26, 27]. Some spherical particles are formed on the surface. Since the shape and size of the particles are consistent with the original SS316L powder (Fig. 9a2), the defects are most likely generated by incomplete melting. From Fig. 9b, an LED of 119.05 J/mm3 results in fewer pores and more dense parts, which agrees with the result of the density test. The high cooling rate causes gas entrapment in the molten pool, resulting in some small pores in the sample. However, the samples processed by SLM at a higher LED create a melt pool with a long liquid lifetime [28], which provides sufficient time for gas to diffuse from the molten pool [29]. In Fig. 9c, some small and shallow pores and hot cracks appear in the sample printed at a much higher LED (166.67 J/mm3); no spherical particles are found in these porosities. Due to too much higher LED, some of the stainless steel powders evaporate and splash. Some gas is trapped in the samples to form small porosities.

Fig. 9
figure 9

Porosity distribution of SS316L printed at three different LEDs: a1, a2 37.88 J/mm3; b 119.05 J/mm3 and c 166.67 J/mm3

Full size image

4.2.3 Microstructure

Figure 10 shows OM images of the top surface microstructure of the SS316L samples processed at different LEDs. It is obvious that the LED plays an important role in the evolution of the microstructure. When the LED is 37.88 J/mm3 (as shown in Fig. 10a), the microstructure consists of relatively straight and regular polygonal austenitic structure. Since the weak Marangoni effect causes heat flow in various directions, the grains have no apparent preferential orientation [30]. Figure 10b, c are the microstructure of the sample fabricated at an LED of 119.05 and 166.67 J/mm3, respectively. Comparing the microstructures of these three LEDs, it is found that the higher the LED, the larger the austenite grains. At a higher LED, the scanning speed is slower, which increases the energy input and stability of the melt pool. A higher temperature in the melt pool provides enough energy as well as nuclei for the growth of grains.

Fig. 10
figure 10

OM images showing microstructure of top surface of SS316L printed at three different LEDs: a 37.88 J/mm3; b 119.05 J/mm3 and c 166.67 J/mm3

Full size image

The high-magnification SEM images of the top surface of the SS316L under different LEDs are shown in Fig. 11. At an LED of 37.88 J/mm3, the grains are fine, but many large pores are visible (Fig. 11a1). Figure 11a2 is an enlarged view of the grain boundary of Fig. 11a1; there are large pores in the grain boundary, which are caused by poor fusion of the stainless steel powder. Since the LED is too low, the adjacent melting tracks do not fuse together. There are also equiaxed grains near the grain boundary (Fig. 11a3). Figure 11b1 presents the microstructure of a sample printed at an LED of 119.05 J/mm3. At larger magnifications (Fig. 11b2), the cellular structure can be identified without pores in the fusion line. Fine columnar grains are also formed near the fusion line, which are similar to the “subgrain” referred to in welding (Fig. 11b3). The grain grows preferentially either perpendicular or at an angle to the fusion line. At a high LED of 166.67 J/mm3, the cellular grains are large (Fig. 11c3) and there are some pores in the grain boundary (Fig. 11c2). The shape of the pores is mainly spherical. This is due to the higher laser energy and the lower scanning speed, which makes it difficult for the gas in the melt pool to escape in a very short time.

Fig. 11
figure 11

SEM images showing microstructure of top surface of SS316L printed at three different LEDs: a1, a2 37.88 J/mm3; b1, b2 119.05 J/mm3; c1, c2 166.67 J/mm3; a3, b3 and c3 are magnified views of squares D, E and F in a2, b2 and c2, respectively

Full size image

Figure 12 depicts the XRD patterns of the powder particles and SLM SS316L samples. The crystalline phase of the SS316L powder consists of austenite γ (fcc phase) and ferrite δ (bcc phase). After the SLM consolidation, the austenite γ is the only crystalline phase and no trace of ferrite δ is present. Lin [30] and Liu [31] found that since the growth temperature of austenite γ was higher than ferrite δ, austenite γ took the lead during the SLM SS316L process. The phase transition of SLM SS316L is (L) − (γ + L) − (γ + δ + L) − (γ + δ) − (γ), where L represents the liquid phase. Since the (γ + δ + L) phase and the (γ + δ) phase are only formed in a narrow temperature range and the cooling rate is extremely quick, the generation of ferrite δ is reduced [32]. Therefore, the SLM SS316L is mainly composed of austenite γ. The XRD peaks of the SLM SS316L are considerably broader compared to the sharp peaks of the powder. This phenomenon agrees with the previous research findings, which is attributed to the residual stress and the lattice distortion caused by the SLM process. In addition, an obvious difference in the preferred orientation is observed in these SLM samples. The sample processed at an LED of 119.05 J/mm3 exhibits a strong (220) preferred orientation, while the other two samples exhibit a maximum peak at (111) and a weak peak at (220). This behavior can be attributed to the growth of preferential crystallite induced by SLM at the different LEDs, as E. Liverani [10] and K. Saeidi [33] studied the crystallite growth under different scanning modes and discovered the preferred orientation of the phase.

Fig. 12
figure 12

XRD patterns of SS316L powder and the SLMed SS316L samples

Full size image

4.2.4 Microhardness

The microhardness distribution of the top surface of SLM SS316L under different LEDs is presented in Fig. 13. As the LED increases, the average microhardness first increases and then decreases. The microhardness is above 300 HV when the LED ranges from 75.76 to 151.52 J/mm3. The maximum microhardness is 330.05 HV when the LED is 119.05 J/mm3. When the LED is moderate, the powders melt sufficiently to form enough liquid phase with a small wetting angle θ. At high temperature, the melted powder particles have good fluidity, and a continuous melt pool is formed. When the LED is lower than 75.76 J/mm3, the microhardness is lower than 300 HV. At a low LED, a sufficient liquid phase cannot be formed during the SLM process, thus, the wetting angle is obtuse. The powder melts insufficiently, and porosity and spheroidization occur and the density of the SLM parts is relatively low, resulting in reduced hardness. At a high LED (higher than 151.52 J/mm3), the hardness decreases almost exponentially with the increase of the LED. The higher the LED, the more liquid phase is generated during SLM process; as a result, the metal powder partially evaporates, and the gas remains in the SLM parts, causing spheroidization and some porosity.

Fig. 13
figure 13

Effect of LED on the microhardness of the top surface of SLMed SS316L

Full size image

4.2.5 Tensile property

The influence of the LED on the mechanical properties of SLM 316L samples is presented in Fig. 14. When the LED ranges from 75.76 to 151.52 J/mm3, the ultimate tensile strength (UTS) is more than 600 MPa, the yield strength (YS) is above 400 MPa, and the elongation (EL) is about 30%. If the LED is too low or too high, both the tensile strength and yield strength significantly decrease. As shown in Fig. 9a, at a low LED, many irregular cavities associated with unmelted powders, large pores will act as the weakest initiator position and accelerate the fracture during tensile testing. While at a high LED (Fig. 9d), some pores caused by evaporation and splashing of stainless steel powder also cause a decrease in strength. Figure 15a shows the stress versus strain curves for three different LEDs. From Fig. 15b, the UTS, the YS and the EL of the sample manufactured at an LED of 119.05 J/mm3 are superior to those of the samples manufactured at an LED of 37.88 J/mm3 and an LED of 166.67 J/mm3. That is to say, the LED has a significant influence on the tensile properties.

Fig. 14
figure 14

Effect of LED on tensile property: a ultimate tensile strength (UTS); b yield strength (YS); and c elongation (EL)

Full size image
Fig. 15
figure 15

a Stress–strain curve at three LEDs and b the ultimate tensile strength (UTS), yield strength (YS), and elongation (EL) of the three samples

Full size image

The representative fracture morphologies of SLM SS316L printed at different LEDs is shown in Fig. 16. Clearly, a number of unmelted particles and pores can be observed on the fracture surfaces of the sample printed at 37.88 J/mm3 (Fig. 16a1). The crack originated from these defects can also be seen in Fig. 16a2. Some shallow dimples, typical of ductile failure, can also be observed on the fracture surfaces. Typically, the pores and cracks act as the weakest initiator position that will propagate and accelerate the fracture during tensile testing. Therefore, it can be inferred that increased defects are responsible for the low mechanical properties for the sample printed at 37.88 J/mm3. There are tiny pores in the fracture surface of the sample printed at an LED of 119.05 J/mm3 (Fig. 16b1). An obvious ductile transgranular fracture mode can be seen in Fig. 16b1. The SLM building direction (BD) has been marked by a red arrow in Fig. 16b1. From Fig. 16b2, it can be seen that more deeper dimples are formed, indicating much better mechanical properties for the sample printed at 119.05 J/mm3. As the LED increases to 166.67 J/mm3, large pores are obvious in the fracture surface. A sliding stripe with cleavage steps, chopping waves and serpentine patterns can be seen on the fracture surface (Fig. 16c1). Therefore, the fracture mechanism of the sample printed at an LED of 166.67 J/mm3 belongs to the slip separation of ductile fracture with low strength and plasticity.

Fig. 16
figure 16

SEM images showing fracture morphologies of SS316L printed at three different LEDs: a1, a2 37.88 J/mm3; b1, b2 119.05 J/mm3; and c1, c2 166.67 J/mm3

Full size image

5 Conclusions

  1. (1)

    A finite element model was established to study the effect of laser power, scanning speed, layer thickness and hatch spacing on the temperature fields of the SLM SS316L powder system. As the laser power increased, the temperature at the nodes of the powder layer increased. However, as the scanning speed, layer thickness and hatch spacing were increased, the temperature decreased. Following the optimization that was carried out, the optimal parameters for SLM processing of SS316L were found to be as follows: laser power of 100–200 W, scanning speed of 500–1500 mm/s, hatch spacing between 0.04 and 0.12 mm, and layer thickness between 0.03 and 0.06 mm.

  2. (2)

    A series of experiments were carried out, where different LEDs were divided into three regions corresponding to three melting mechanisms: the partial melting zone (37.88–75.76 J/mm3), the full melting zone (75.76–151.52 J/mm3), and the over-melting zone (larger than 151.52 J/mm3). The results showed that SLM SS316L with a relative density higher than 99.66% and lower porosity was obtained at a LED of 119.05 J/mm3.

  3. (3)

    It was found that LED was a suitable parameter to determine the microstructure of the SLM SS316L. At the optimum LED (119.05 J/mm3), the interior of the SLM SS316L part was substantially non-porous. The columnar and cellular grains were the main microstructural features.

  4. (4)

    LED was also found to be the most influential parameter on the mechanical properties of SS 316L parts. The UTS (693.13 MPa), the YS (616.24 MPa) and the EL (34.60%) of the sample manufactured at the optimal LED of 119.05 J/mm3 were superior to those of all the other SLM samples. Many dimples appeared on the fracture surface, indicating ductile fracture.

The results demonstrate that a suitable window suitable for SLM processing of 316L has been found by finite element model and experiments. Future work will be conducted on the application of the processes to the real components with complex geometries.