Permeability Study of Austenitic Stainless Steel Surfaces Produced by Selective Laser Melting

Selective laser melting (SLM) is emerging as a versatile process for fabricating different metal components with acceptable mechanical properties and geometrical accuracy. The process has been used in the manufacturing of several parts (e.g., aerospace or biomedical components), and offers the capability to tailor the performance of several surface and mechanical properties. In this work, permeability properties and surface roughness of stainless steel (SS316L) surfaces were evaluated through experimentation with three different laser scanning patterns (chessboard, meander, and stripe), and different sloping angles between the fabricated surface and the laser beam incident on the process. Results showed that for each scanning pattern, the roughness decreased as the sloping angle increased consistently in all experimental trials. Furthermore, in the case of the permeability evaluation, the manufactured surfaces showed changes in properties for each series of experiments performed with different scanning patterns. The chessboard pattern showed a change of 67◦ to 107◦ in contact angle, while the meander and stripe patterns showed a variation in contact angle in a range of 65◦ to 85◦. The different scanning strategies in the SLM process resulted in an alternative method for surface enhancement with different hydrophobicity properties, valuable for designing the most appropriate permeability characteristics for specific applications.


Introduction
Selective laser melting (SLM) technology is an additive manufacturing (AM) process that enables the fabrication of components with complex shapes for medical, aerospace, and automotive applications directly from metal powder [1][2][3], offering a series of advantages compared with traditional processing techniques due to its versatility in producing both metal and polymer parts [4][5][6][7].The SLM process enables the direct melting of powders of a number of alloys, such as titanium (Ti6Al4V), stainless steel (SS316L), cobalt-chromium (Co-Cr), aluminum, and the building of net-shape parts through a layer-by-layer process.For each layer, a scanning laser beam supplies the energy to melt a layer of deposited metal powder and fuse it onto a previously melted layer.
The use of SLM technologies is now an asset in the manufacturing of biomedical devices [8,9], due to its use for customizing medical implants and orthodontic devices to the patient's anatomy.Another [10,11] increasing application is the development of aeronautic components with high degrees of shape complexity.However, SLM still faces limitations in terms of surface quality when it is compared to some conventional metal manufacturing processes.Surface quality is significantly influenced by the stair step effect resulting from the building angle measured from the building plate to the building orientation, which is the stepped approximation by layers of curved and inclined surfaces [12].This effect is present to a greater or lesser degree in all additive layer-manufacturing processes as consequence of the manufacturing process and the layer discretization.Surface quality (surface roughness, corrosion, and porosity properties) is also influenced by process parameters such as temperature profiles and densification ratios, among others [13][14][15][16].However, the layer thickness can be reduced in order to improve the surface finish [17], but a smooth surface is also constrained by the balling phenomenon, which refers to the particles either not melting at all or joining into rather large droplets, which occurs during laser melting [18][19][20].The balling phenomenon is related to the wettability of the powder; a poor wettability of the working powder increases the probability of developing this undesirable effect.The wettability also affects the mechanical and corrosion behavior of the fabricated parts, so it is considered as an important element to be consider in selective laser melting effectiveness.Poor wettability in the surface produces cracks that will further cause catastrophic failures in mechanical functional components [21].Thus, it can be determined that laser-melted surfaces with good wettability are less prone to corrosion and have higher wear resistance in comparison with damaged surfaces.
The balling effect limits the laser melting process resolution because it causes the formation of discontinuous tracks, therefore limiting the formation of accurate geometries.The ability to control the wetting properties of metal alloys is extremely beneficial, in many areas for scientific and industrial applications.Wetting controllability includes the enhancement of heat transfer [22][23][24], improvement of surface coatings [25], and self-cleaning and anti-icing surface fabrications [26,27].An alternative way to control wettability is the use of conical microstructures formed by the irradiation of surfaces with femtosecond laser pulses, in order to modify their properties for conducting contact angle evaluations [28,29].
Other methods have been employed to control the wetting properties of surfaces for preparing (a) hydrophilic surfaces via hydrophilic coatings [21,22], plasma treatments [23,24], or chemical treatments [25] and (b) hydrophobic surfaces via hydrophobic coatings [27,[30][31][32][33].However, these methods are generally expensive, complicated, and time consuming.Some simple methods, such as chemical treatments and inexpensive coatings are generally either evanescent or fragile.Moreover, no single method can cover the wide range of contact angles from the super-hydrophilic to super-hydrophobic condition.
SLM is capable of processing a variety of materials, and the development of powders is growing in variety.Accordingly, the research with metal alloys suitable for broad applications may be a better example for industrial scenarios.Stainless steel alloys accomplish this requirement since they may be used for biomedical device manufacturing, automotive component fabrication, and oil and gas instrumentation, among other applications [34,35].Furthermore, there are few studies related to the permeability of stainless steels.Chemical processes have been useful in controlling the hydrophobic characteristics of these alloys [36].On the other hand, Antony et al. [37] focus their studies of permeability by modifying scanning speeds and laser power during the melting process.Accordingly, it is well known after reviewing the literature that process parameters can be modified in order to change the final part properties, and thus eliminating conventional post-processing steps for the preparation of surfaces with good wettability.In order to contribute towards the complete evaluation of SLM in the manufacturing of functional surfaces, this study is intended to propose an alternative methodology for modifying the wettability of surfaces obtained by the process using different laser scanning strategies.
The distribution of particles size was calculated with the aid of scanning electron microscope (SEM) micrographs and with the use of image analysis software (imageJ, National Institutes of Health, 1.51r, Bethesda, MD, USA).After laser beam scanning of each cross-section, the manufacturing plate was lowered by the thickness of the layer to be deposited, thus maintaining the surface in the focal plane.Parts of 10 mm × 10 mm × 5 mm were fabricated at different sloping angles (θ) with respect to the normal laser beam (θ = 0 • , 45 • , 60 • , 70 • , 90 • ) as it is represented in Figure 1.Additionally, Table 1 shows the parameters used in the SLM process.After the manufacturing process, the samples were removed from the building plate and were brought to a cleaning process in an ultrasonic bath for 10 min in 95% ethanol, 10 min in acetone, and dried.The morphology and particle size distribution (30 images obtained) of the samples were carefully analyzed by using an EVO MA25 ZEISS scanning electron microscope (ZEISS, German, Jena, Germany).The surface roughness and contact angle measurements were made after each experimental run, as it is explained in the following.
The distribution of particles size was calculated with the aid of scanning electron microscope (SEM) micrographs and with the use of image analysis software (imageJ, National Institutes of Health, 1.51r, Bethesda, MD, USA).After laser beam scanning of each cross-section, the manufacturing plate was lowered by the thickness of the layer to be deposited, thus maintaining the surface in the focal plane.Parts of 10 mm × 10 mm × 5 mm were fabricated at different sloping angles (θ) with respect to the normal laser beam (θ = 0°, 45°, 60°, 70°, 90°) as it is represented in Figure 1.Additionally, Table 1 shows the parameters used in the SLM process.After the manufacturing process, the samples were removed from the building plate and were brought to a cleaning process in an ultrasonic bath for 10 minutes in 95% ethanol, 10 min in acetone, and dried.The morphology and particle size distribution (30 images obtained) of the samples were carefully analyzed by using an EVO MA25 ZEISS scanning electron microscope (ZEISS, German, Jena, Germany).The surface roughness and contact angle measurements were made after each experimental run, as it is explained in the following.Contact angle (CA) measured from the surface of a building part and a droplet of water was performed to evaluate the permeability characteristics of the fabricated parts [38].The contact angle has been drawn by the Dataphysics OCA 15EC system (DataPhysics Instruments GmbH, Filderstadt, Germany) in the measurement range of 1°-180°.The evaluation of the drop shape and the calculation  Contact angle (CA) measured from the surface of a building part and a droplet of water was performed to evaluate the permeability characteristics of the fabricated parts [38].The contact angle has been drawn by the Dataphysics OCA 15EC system (DataPhysics Instruments GmbH, Filderstadt, Germany) in the measurement range of 1-180 • .The evaluation of the drop shape and the calculation of the contact angle (Figure 2B) were performed with the aid of the SCA20_U software (DataPhysics Instruments GmbH, SCA 20, Filderstadt, Germany).10 µL of drops were placed onto the substrate using an automatic dispenser, and were injected slowly at 2 µL/s onto the solid surface by a syringe (Figure 2A).To measure contact angle values, droplet images were taken directly after 8 s of deposition in order to eliminate the impact of droplet evaporation (Figure 2B).The value of CA is the average of CA right and CA left.The contact angle tests were carried out at 23 • C.

Roughness
In order to characterize surface roughness (Ra), samples were oriented in different angles with respect to the building plate, and different building patterns were used.The average surface roughness (Ra) measurements were performed using the ISO 4287 standard, which is one of the most adopted texture parameters in the literature [39,40].For this study, a non-contact 3D profilometry method was used.The surface topography information of the top surfaces of the samples was obtained using a focus variation Alicona infinite focus microscope G4 (Alicona Imaging GmbH, Graz, Austria).It was used with an objective with 20× magnification, with a vertical resolution of 50 μm and lateral resolution of 3.5 μm.Ra values were calculated with a cut-off length of 2500 μm.This cutoff captured the overall surface roughness, allowing the high-frequency roughness oscillations to be detected from the low-frequency waveforms [19].Three sections were selected in the samples (top, center, and bottom), so that subsequently three measurements were made (top, center, and bottom).

Scanning Electron Microscope
The particle size distribution obtained from the SEM images is shown in Figure 3, where it can be seen that the majority of the particles presented a spherical shape.The size distribution was found to be between 7 and 55 μm, with an average diameter of ~25 μm and D50 = 27.79 μm.For the surface morphology characterization, SEM analysis was carried out at θ = 90° and 60°, and for the three laser scanning strategies.Figure 4a-c shows a few partially sintered particles attached on the surface, where it can be seen that smoothed uniform layers were obtained in the

Roughness
In order to characterize surface roughness (Ra), samples were oriented in different angles with respect to the building plate, and different building patterns were used.The average surface roughness (Ra) measurements were performed using the ISO 4287 standard, which is one of the most adopted texture parameters in the literature [39,40].For this study, a non-contact 3D profilometry method was used.The surface topography information of the top surfaces of the samples was obtained using a focus variation Alicona infinite focus microscope G4 (Alicona Imaging GmbH, Graz, Austria).It was used with an objective with 20× magnification, with a vertical resolution of 50 µm and lateral resolution of 3.5 µm.Ra values were calculated with a cut-off length of 2500 µm.This cut-off captured the overall surface roughness, allowing the high-frequency roughness oscillations to be detected from the low-frequency waveforms [19].Three sections were selected in the samples (top, center, and bottom), so that subsequently three measurements were made (top, center, and bottom).

Scanning Electron Microscope
The particle size distribution obtained from the SEM images is shown in Figure 3, where it can be seen that the majority of the particles presented a spherical shape.The size distribution was found to be between 7 and 55 µm, with an average diameter of ~25 µm and D50 = 27.79 µm.

Roughness
In order to characterize surface roughness (Ra), samples were oriented in different angles with respect to the building plate, and different building patterns were used.The average surface roughness (Ra) measurements were performed using the ISO 4287 standard, which is one of the most adopted texture parameters in the literature [39,40].For this study, a non-contact 3D profilometry method was used.The surface topography information of the top surfaces of the samples was obtained using a focus variation Alicona infinite focus microscope G4 (Alicona Imaging GmbH, Graz, Austria).It was used with an objective with 20× magnification, with a vertical resolution of 50 μm and lateral resolution of 3.5 μm.Ra values were calculated with a cut-off length of 2500 μm.This cutoff captured the overall surface roughness, allowing the high-frequency roughness oscillations to be detected from the low-frequency waveforms [19].Three sections were selected in the samples (top, center, and bottom), so that subsequently three measurements were made (top, center, and bottom).

Scanning Electron Microscope
The particle size distribution obtained from the SEM images is shown in Figure 3, where it can be seen that the majority of the particles presented a spherical shape.The size distribution was found to be between 7 and 55 μm, with an average diameter of ~25 μm and D50 = 27.79 μm.For the surface morphology characterization, SEM analysis was carried out at θ = 90° and 60°, and for the three laser scanning strategies.Figure 4a-c shows a few partially sintered particles attached on the surface, where it can be seen that smoothed uniform layers were obtained in the For the surface morphology characterization, SEM analysis was carried out at θ = 90 • and 60 • , and for the three laser scanning strategies.Figure 4a-c shows a few partially sintered particles attached on the surface, where it can be seen that smoothed uniform layers were obtained in the process under these conditions.As θ increases, the sharpness of the edges diminishes due to the increase of unsintered particles that are trapped, and remain between the layers (Figure 4d-f).The formation of discontinuous edges is partly determined by the friction effect, which occurs during laser melting of the metal powder [41].During laser melting, heat at the edges of the layers is not enough to melt the particles; consequently, they do not fuse with the processed layer, tending to be adhered to the surface at this region as shown.Balling is a severe impediment in the layers' bonding, decreasing the density of the part and increasing the surface roughness [42].Increasing the temperature of the powder bed can increase permeability, thereby reducing the balling phenomenon [43].However, the scope of this work is to find the influence of the laser scanning strategies on the surface properties; thus, it is intended to contribute to the enhancement of the process in addition to other aspects in parameters modification that have been studied before.
Metals 2017, 7, 521 5 of 10 process under these conditions.As θ increases, the sharpness of the edges diminishes due to the increase of unsintered particles that are trapped, and remain between the layers (Figure 4d-f).The formation of discontinuous edges is partly determined by the friction effect, which occurs during laser melting of the metal powder [41].During laser melting, heat at the edges of the layers is not enough to melt the particles; consequently, they do not fuse with the processed layer, tending to be adhered to the surface at this region as shown.Balling is a severe impediment in the layers' bonding, decreasing the density of the part and increasing the surface roughness [42].Increasing the temperature of the powder bed can increase permeability, thereby reducing the balling phenomenon [43].However, the scope of this work is to find the influence of the laser scanning strategies on the surface properties; thus, it is intended to contribute to the enhancement of the process in addition to other aspects in parameters modification that have been studied before.

Roughness
Roughness (Ra) was measured on the samples without post-processing with the non-contact 3D instrument described above.Figure 5 shows the average roughness values of the different laser scanning strategies employed.It can be seen that as θ increases, the surface roughness values decrease; this tendency was observed in all samples during the experimental trials.It is observed that the three scanning patterns show a similar behavior.The lowest roughness value Ra is present at θ = 90°; in this case, the laser beam is perpendicular to the scanning surface and the metal particles are almost completely melted, forming a smooth surface and decreasing the balling effect, consistent with SEM results.Chessboard laser scanning presents a greater roughness for the different θ angles used.These Ra values vary from 12 to 19 μm.Meander and stripes laser scanning have smoother surfaces, the roughness changes from 12 to 6 μm with the variation of the building angle.Some authors [13,44] have reported that roughness values Ra vary depending on the parameters used during the melting process.For this work, the authors show that chessboard laser scanning creates rougher surfaces, the scanning shape changes continuously in orientation, and the areas to be consolidated are smaller than when using a one-dimensional scan.These changes of orientation in the patterns influence the construction time and properties of SLM components and supports.

Roughness
Roughness (Ra) was measured on the samples without post-processing with the non-contact 3D instrument described above.Figure 5 shows the average roughness values of the different laser scanning strategies employed.It can be seen that as θ increases, the surface roughness values decrease; this tendency was observed in all samples during the experimental trials.It is observed that the three scanning patterns show a similar behavior.The lowest roughness value Ra is present at θ = 90 • ; in this case, the laser beam is perpendicular to the scanning surface and the metal particles are almost completely melted, forming a smooth surface and decreasing the balling effect, consistent with SEM results.Chessboard laser scanning presents a greater roughness for the different θ angles used.These Ra values vary from 12 to 19 µm.Meander and stripes laser scanning have smoother surfaces, the roughness changes from 12 to 6 µm with the variation of the building angle.Some authors [13,44] have reported that roughness values Ra vary depending on the parameters used during the melting process.For this work, the authors show that chessboard laser scanning creates rougher surfaces, the scanning shape changes continuously in orientation, and the areas to be consolidated are smaller than when using a one-dimensional scan.These changes of orientation in the patterns influence the construction time and properties of SLM components and supports.The top surfaces of the samples with the meander scanning laser path at different sloping angles θ = 90° (Figure 6A), 70° (Figure 6B), 60° (Figure 6C), 45° (Figure 6D), and 0° (Figure 6E) are presented as topographies.The measured area was defined by 1.620 mm × 1.620 mm, and height was dependent on the surface profile having values from 133.507 to 180.583 μm.

Contact Angle
CA measurements were performed on the surfaces of the samples in order to observe the change in permeability for each scanning strategy used.For each measurement, the samples were analyzed under the same conditions, and five random areas of the surface of each sample were evaluated.Figure 7 shows the variation of the CA for different sloping angles and each scanning strategy; the error bars indicate the variability for each combination.It can be noticed that the equilibrium CAs at the surfaces are approximately between 70° and 120°.
It can be seen that the CAs decrease with the increase of θ.The chessboard scanning strategy shows a larger change in CA from 67° to 107° (Figure 7A).In the cases of meander and stripes scanning strategies, the behavior is similar and the variation of the CA oscillates from approximately 65° to 85° (Figure 7B,C).Some authors have associated the permeability with the roughness, but it has been seen that materials with the same roughness have different values of CA; this shows that the permeability cannot be controlled solely by the roughness modification [45].Other authors found The top surfaces of the samples with the meander scanning laser path at different sloping angles θ = 90 • (Figure 6A), 70 • (Figure 6B), 60 • (Figure 6C), 45 • (Figure 6D), and 0 • (Figure 6E) are presented as topographies.The measured area was defined by 1.620 mm × 1.620 mm, and height was dependent on the surface profile having values from 133.507 to 180.583 µm.The top surfaces of the samples with the meander scanning laser path at different sloping angles θ = 90° (Figure 6A), 70° (Figure 6B), 60° (Figure 6C), 45° (Figure 6D), and 0° (Figure 6E) are presented as topographies.The measured area was defined by 1.620 mm × 1.620 mm, and height was dependent on the surface profile having values from 133.507 to 180.583 μm.

Contact Angle
CA measurements were performed on the surfaces of the samples in order to observe the change in permeability for each scanning strategy used.For each measurement, the samples were analyzed under the same conditions, and five random areas of the surface of each sample were evaluated.Figure 7 shows the variation of the CA for different sloping angles and each scanning strategy; the error bars indicate the variability for each combination.It can be noticed that the equilibrium CAs at the surfaces are approximately between 70° and 120°.
It can be seen that the CAs decrease with the increase of θ.The chessboard scanning strategy shows a larger change in CA from 67° to 107° (Figure 7A).In the cases of meander and stripes scanning strategies, the behavior is similar and the variation of the CA oscillates from approximately 65° to 85° (Figure 7B,C).Some authors have associated the permeability with the roughness, but it has been seen that materials with the same roughness have different values of CA; this shows that the permeability cannot be controlled solely by the roughness modification [45].Other authors found

Contact Angle
CA measurements were performed on the surfaces of the samples in order to observe the change in permeability for each scanning strategy used.For each measurement, the samples were analyzed under the same conditions, and five random areas of the surface of each sample were evaluated.It can be seen that the CAs decrease with the increase of θ.The chessboard scanning strategy shows a larger change in CA from 67 • to 107 • (Figure 7A).In the cases of meander and stripes scanning strategies, the behavior is similar and the variation of the CA oscillates from approximately 65 • to 85 • (Figure 7B,C).Some authors have associated the permeability with the roughness, but it has been seen that materials with the same roughness have different values of CA; this shows that the permeability cannot be controlled solely by the roughness modification [45].Other authors found that the permeability of stainless steel surfaces can be changed by modifying SLM parameters like laser power, scanning speed, and beam diameter [39].Nevertheless, the sloping angle variation and the laser scanning strategy modification turn out to be alternative methods to control the permeability of the parts produced by SLM.The influence of the scanning strategy has been a useful tool for controlling different surface properties in the sintering process [15].Thus, the different conditions employed for SLM show a change in hydrophobicity that is possible to control depending on the required need.
Metals 2017, 7, 521 7 of 10 that the permeability of stainless steel surfaces can be changed by modifying SLM parameters like laser power, scanning speed, and beam diameter [39].Nevertheless, the sloping angle variation and the laser scanning strategy modification turn out to be alternative methods to control the permeability of the parts produced by SLM.The influence of the scanning strategy has been a useful tool for controlling different surface properties in the sintering process [15].Thus, the different conditions employed for SLM show a change in hydrophobicity that is possible to control depending on the required need.

Conclusions
This work discusses the application of different laser scanning strategies in order to modify the permeability characteristics of selective laser melted surfaces made of austenitic stainless steel.The main conclusions that can be drawn are: 1.The lowest Ra values were found on the surfaces with greater sloping angle.Additionally, these surfaces present fewer desegregated particles.2. As the sloping angle diminishes, the accumulation of semi-melted particles increases, increasing the roughness of the surface.3. The meander and stripes scanning strategies showed one-dimensional scenes in which the balling effect is smaller with respect to the chessboard scanning laser.The chessboard scanning strategy showed a continuous change in the scanning direction in the SLM process, which generates a surface with greater roughness.4. The layer-by-layer overlay in the SLM process showed that the one-dimensional scanning direction creates a smooth surface with a slight change in the contact angle. 5.The three conditions studied showed that the surface roughness might have the same value but different contact angles, indicating that the permeability can be controlled without a significant detriment to surface quality.
In addition to the findings in previous works, in which the surface permeability is calibrated by the variation of laser parameters in SLM, the contribution of the work presented here turns out to be an alternative methodology for the preparation of surfaces for different applications.In order to make a comprehensive assessment of selective laser melting operations for the fabrication of functional surfaces, the combination of powder materials, process parameters, surface orientation, and scanning strategies will be studied in future works, including the densification of the parts changing these parameters and its influence on the mechanical and surface development.

Conclusions
This work discusses the application of different laser scanning strategies in order to modify the permeability characteristics of selective laser melted surfaces made of austenitic stainless steel.The main conclusions that can be drawn are: 1.
The lowest Ra values were found on the surfaces with greater sloping angle.Additionally, these surfaces present fewer desegregated particles.2.
As the sloping angle diminishes, the accumulation of semi-melted particles increases, increasing the roughness of the surface.

3.
The meander and stripes scanning strategies showed one-dimensional scenes in which the balling effect is smaller with respect to the chessboard scanning laser.The chessboard scanning strategy showed a continuous change in the scanning direction in the SLM process, which generates a surface with greater roughness.4.
The layer-by-layer overlay in the SLM process showed that the one-dimensional scanning direction creates a smooth surface with a slight change in the contact angle. 5.
The three conditions studied showed that the surface roughness might have the same value but different contact angles, indicating that the permeability can be controlled without a significant detriment to surface quality.
In addition to the findings in previous works, in which the surface permeability is calibrated by the variation of laser parameters in SLM, the contribution of the work presented here turns out to be an alternative methodology for the preparation of surfaces for different applications.In order to make a comprehensive assessment of selective laser melting operations for the fabrication of functional surfaces, the combination of powder materials, process parameters, surface orientation, and scanning strategies will be studied in future works, including the densification of the parts changing these parameters and its influence on the mechanical and surface development.
Metals 2017, 7, 521 4 of 10 deposition in order to eliminate the impact of droplet evaporation (Figure2B).The value of CA is the average of CA right and CA left.The contact angle tests were carried out at 23 °C.

Figure 2 .
Figure 2. Contact angle configuration: (A) Stainless 316L surface and water drop; (B) Contact angle analysis.

Figure 2 .
Figure 2. Contact angle configuration: (A) Stainless 316L surface and water drop; (B) Contact angle analysis.
to eliminate the impact of droplet evaporation (Figure2B).The value of CA is the average of CA right and CA left.The contact angle tests were carried out at 23 °C.

Figure 2 .
Figure 2. Contact angle configuration: (A) Stainless 316L surface and water drop; (B) Contact angle analysis.
Figure 7 shows the variation of the CA for different sloping angles and each scanning strategy; the error bars indicate the variability for each combination.It can be noticed that the equilibrium CAs at the surfaces are approximately between 70 • and 120 • .

Figure 7 .
Figure 7. Contact angle measurements results with different scanning strategies and sloping angles: (A) chessboard, (B) meander and (C) stripes.

Figure 7 .
Figure 7. Contact angle measurements results with different scanning strategies and sloping angles: (A) chessboard, (B) meander and (C) stripes.

Table 1 .
Process parameters for Stainless steel 316L.

Table 1 .
Process parameters for Stainless steel 316L.