Contents lists available at ScienceDirect

Journal of Materials Processing Tech.

journal homepage: www.elsevier.com/locate/jmatprotec

Influence of laser parameters and complex structural features on the bio-inspired complex thin-wall structures fabricated by selective laser melting

Kaijie Lina,b, Luhao Yuana,b, Dongdong Gua,b,⁎

a College of Materials Science and Technology, Nanjing University of Aeronautics and Astronautics, Yudao Street 29, Nanjing, 210016, Jiangsu Province, PR Chinab Jiangsu Provincial Engineering Laboratory for Laser Additive Manufacturing of High-Performance Metallic Components, Nanjing University of Aeronautics andAstronautics, Yudao Street 29, Nanjing, 210016, Jiangsu Province, PR China

A R T I C L E I N F O

Associate Editor: A Clare

Keywords:Additive manufacturingSelective laser meltingThin-wall structuresBio-inspiredLobster-eyeAlSi10Mg

A B S T R A C T

The lobster-eye structure exhibits unique optical performance and finds its application in aerospace. However,the lobster-eye structure is highly complex and hard to process using traditional subtractive manufacturingmethods. In this work, complex thin-wall structures, inspired by the lobster eye, were fabricated by selectivelaser melting (SLM) with AlSi10Mg powder. The influence of key laser parameter, namely laser power, on thedensification behavior, dimensional accuracy, surface roughness and forming defect of SLM-processed lobster-parts was systematically investigated. Particular emphasis was placed on the impact of laser power and struc-tural features on the inner surface roughness. Results revealed that the relative density of SLM-processed lobster-eye parts exhibited a similar trend with bulk parts against the laser power. The small dimension feature of thin-wall structures contributed to the higher relative density of lobster-eye parts comparing with that of bulk parts.The roughness of inner surfaces and outside surfaces showed completely different responds to the laser power,and the inner surface morphology behaved differently among positions. All these special surface phenomenawere contributed to the combined impact of the laser parameter and the particular geometry of the lobster-eyestructure. Finally, considering the optical function and post-processing, an optimal laser power was determinedfor the SLM-processing of lobster-eye components.

1. Introduction

During the millions of years of evolution, the nature life has de-veloped the optimized natural structures to survive in the cruel naturalcompetition. According to specific living environment, natural struc-tures exhibit unique functions. Qin et al. (2015) claim that the topo-logical design of spider-web provides the properties of light-weight,high strength and elasticity. Habibi et al. (2015) find that the func-tionally-graded hollow structure of bamboo delivers excellent flexibilityand fracture toughness. Finnemore et al. (2012) find that the orderedmultilayer structure of nacre performs remarkably high toughness andresilience. Chen et al. (2016) believe that the continuous directionalwater transport on the peristome surface of nepenthes alata is con-tributed to the unique nano/micro-structure on the surface.

In the field of optical application, natural structures also showoutstanding performance. One of the most representative instances isthe lobster eye (Fig. 1a). The eye of lobster is composed of numeroussmall square channels arranging over a spherical surface. Each channel

is long and narrow, with its central axis goes toward to the center of thespherical surface. Angel (1979) finds that, light entering the channelarray from different angles is focused through grazing-incident reflec-tion and form a single image on the curved retina of lobster (Fig. 1b),thanks to the unique structure. The lobster-eye structure optics possessthe advantages of light-weight, small size and wide field of view, whichis perfected for the application of aerospace. Currently, the fabricationmethods applied to process lobster-eye structures are basically sub-tractive manufacturing. For instance, Peele et al. (2007) apply LIGAmethod to produce lobster-eye optics with a graphite substrate. How-ever, due to limitation of subtractive methods, the high aspect ratiochannel is hard to fabricate.

The channel-array structures of lobster eye can also be treated ascomplex thin-wall structures consisting of two groups of thin wallsperpendicular to each other. For the fabrication of these complex thin-wall structures, additive manufacturing (AM) technology has sig-nificant advantages over traditional subtractive (such as milling andmachining) or formative manufacturing (such as casting and plastic

https://doi.org/10.1016/j.jmatprotec.2018.12.004Received 25 June 2018; Received in revised form 2 November 2018; Accepted 1 December 2018

⁎ Corresponding author at: College of Materials Science and Technology, Nanjing University of Aeronautics and Astronautics, Yudao Street 29, Nanjing, 210016,Jiangsu Province, PR China.

E-mail address: dongdonggu@nuaa.edu.cn (D. Gu).

Journal of Materials Processing Tech. 267 (2019) 34–43

Available online 03 December 20180924-0136/ © 2018 Elsevier B.V. All rights reserved.

T

forming) technologies. Due to the layer-wise nature, AM technology ismore competitive than traditional manufacturing to fabricate parts witha high level of complexity. Many efforts have been devoted into theresearch of additive manufacturing of thin-wall structures. Till now,various AM technologies have been applied to fabricate thin-wallstructures. Bontha et al. (2006) use laser melting deposition (LMD) tofabricate thin-wall Ti6Al4V parts and predict solidification micro-structure using thermal process maps. Denlinger et al. (2015) applyelectron beam melting (EBM) to build Ti6Al4V thin-walls and in-vestigate the evolution of residual stress. Boegelein et al. (2016) employ

selective laser melting to fabricate complex thin-wall structures. Due tothe high energy density and coarse size of feeding raw material, thesurface quality and dimensional accuracy of products, fabricated by thedeposition-based technologies and EBM, are worse than that of partsfabricated by SLM. Therefore, SLM technology has been widely used tofabricate thin-wall parts with fine feature and complex structures.

The structure of thin-wall parts is significantly different from that ofthe bulk parts, leading to completely different temperature distributionand residual stress generation. The optimized SLM processing para-meters for bulk parts, occurring high density and excellent mechanical

Fig. 1. (a) SEM image of lobster eye (NASA, 2018), (b) the optical mechanism of lobster-eye.

Fig. 2. The manufacturing process of the lobster-eye structure: (a) CAD model, (b) SEM image of powder and size distribution, (c) the schematic of SLM machine andthe photo of SLM processing and (d) the SLM-processed components.

K. Lin et al. Journal of Materials Processing Tech. 267 (2019) 34–43

35

performance, might not be preferred for thin-wall parts. Therefore,most of the researches on the SLM thin-wall parts concentrate on theprocess parameter study. Yadroitsev et al. (2007) study the influence ofscanning strategies, namely lengthwise, transversal and angular, on theformability of thin walls and find that lengthwise one is the optimalstrategy for forming thin walls, which can build the thinnest wall withsmallest dimensional error. Song et al. (2015) investigate the influenceof laser energy and scanning speed on the thickness of SLM-processedthin walls and find that the thickness of thin walls increases with theincreasing of linear laser energy density. Shishkovskii et al. (2016) in-vestigate the relation between powder size and the key structure as-pects of thin-wall parts. Results reveal that smaller powder size deliversbetter reproduction, thinner wall and more precise angle. Apart fromthe SLM process parameters, the thin-wall structure itself also affectsthe formability. Li et al. (2018) apply indirect coupled thermal-struc-tural finite element model to analyze the temperature and stress

distribution of thin-walls with different length during the SLM process.Results show that the different temperature history and the consequentdifference of stress distribution, causing by the various thin-walllengths, contributes to the different distortion and shrinkage behavior.Calignano et al. (2018) reveal that the constructed direction of SLM-processed walls, namely parallel and perpendicular to the recoatingblade moving direction, significantly affect the formability of thin-wallstructures. Calignano (2018) also find that the building angle affects theheat flow in the build component and consequently influence the sur-face roughness. The surface quality of thin-wall structures is a criticalindicator in many applications, especially for the medical implants. Liuet al. (2018) find that, similar with the SLM-processed bulk parts, thelaser processing parameters directly influence the surface roughness ofthin-wall parts. In particular, different parts of thin-wall structures re-sponds differently in term of surface roughness. Wang et al. (2016) findthat for a thin-wall part, built with inclination angle to the substrate

Fig. 3. (a) The relative density and (b) the cross-sectional OM images of SLM-processed bulk and lobster-eye parts (pores are marked by black arrows).

Fig. 4. Schematic of different thermal transfer: (a) the bulk part, (b) the thin-wall part.

K. Lin et al. Journal of Materials Processing Tech. 267 (2019) 34–43

36

plane, the roughness of downward surface is always higher than that ofthe upward surface. Mumtaz and Hopkinson (2009) investigate therelationship between process parameters and top/side roughness ofthin-wall parts. It is interesting to find that higher laser power tends tosimultaneously reduce the top surface roughness and side roughness,however, lower scan speed reduces top surface roughness but increasesside surface roughness.

The lobster-eye structure is widely used in optical application, thesurface quality, especially the inner surface roughness, is curial forbetter optical performance. However, to the best of our knowledge, therelation between SLM-processing parameters and inner surface rough-ness of complex thin-wall structures has not been reported. In thepresent work, the lobster-eye parts were fabricated with the method ofselective laser melting using AlSi10Mg powder. The influence of SLMprocess parameter and the complex structural features on the form-ability of lobster-eye parts was investigated. Particular emphasis wasplaced on the impact of SLM process parameter and structural featuresto the inner surface roughness of lobster-eye parts.

2. Materials and methods

2.1. SLM processing of lobster-eye structures

2.1.1. Structure designThe thin-wall structure in this work was inspired by the lobster eye

(shown in Fig. 1) and the CAD model is shown in Fig. 2(a). The de-signed lobster-eye structure consisted of a 9 by 9 array of channels. Theopening of each channel is a square with a length of 1.5 mm, and theheight and cone angle (α) of each channel is 10mm and 2°, respectively.The thickness (t) of all the thin walls is set as 0.2mm. The cone-angle ofthe whole lobster-eye structure (θ) is 18°.

2.1.2. PowderThe powder size greatly influences the forming quality of SLM-

processed components (Sutton et al., 2017). Based on our previous re-search, gas-atomized AlSi10Mg powder (with the mean particle size of23 μm) possessed good formability and was used in this work. The SEMimage of the powder and the particle size distribution are presented inFig. 2(b).

2.1.3. SLM processingA self-developed SLM machine (Fig. 2(c)) was applied in this work.

The detail of this machine can be found elsewhere (Chen et al., 2017a).All the SLM processes were conducted under the Ar atmosphere withthe oxygen content lower than 10 ppm. SLM process parameters likelaser power, scanning speed and layer thickness affect the part quality(Fatemi et al., 2017). However, Sing et al. (2018) claim that the in-fluence of laser power on the forming quality of complex components ismore significant than other parameters. Therefore, the influence oflaser power was investigated in this work. The laser power was variedbetween 325W–425W in 25W increments. For all the SLM-processedlobster-eye components, the scanning speed, layer thickness and hatchspacing were constantly set as 2200mm/s, 30 μm and 50 μm, respec-tively. The selection of those parameters was based on our previousresearch (Gu and Dai, 2016). In order to minimize the area of over-hanging surface, the building direction of all components was parallelto the height direction of the center channel. During the fabrication,cubic parts with the dimension of 5mm×5mm×5mm were builtalongside with each lobster-eye part, using the identical parametersetting.

2.2. Characterization methods

2.2.1. Relative densityThe density of SLM-processed components (including lobster-eye

parts and cubic parts) were measured using the Archimedes’ principle.To avoid the trapping of bubble inside the small channels, ethanol wasemployed as the solution, because of the lower surface tension(0.0223 N/m at 20℃) than water (0.0728 N/m at 20℃). The relativedensity (ξ) were calculated by the following equation:

= ×ζρρ

100%M

T (1)

where ρM was the measured density of the SLM-processed part and ρT isthe theoretical density of AlSi10Mg (2.68 g/cm3).

2.2.2. Dimensional accuracyFor this particular lobster-eye structure, special emphasis was put

on the forming accuracy of the cone angle (θ) and the thickness of thinwalls (t). SLM-processed components were sliced along the height di-rection to facilitate the optical microscopic (OM) observation (OlympusPMG3, Japan). The key structure parameters were measured from theOM images by the software of ToupView.

2.2.3. Surface morphologyThe surface roughness of SLM-processed parts was measured by the

LEXT OLS4100 Laser Confocal Microscope (Olympus, Japan). In addi-tion, the surface morphology was characterized using a JSM-6360LVscanning electron microscope (Jeol, Japan).

3. Results and discussion

3.1. Densification behavior

The relative density of the bulk part and the lobster-eye part ex-hibited the same trend against the laser power (Fig. 3(a)). The relativedensity firstly increased with the laser power and reached the highestrelative density (99.7% for the lobster-eye part and 99.4% for the bulkpart) at the laser power of 400W, and then a decrease was observedwhen the laser power further increased to 425W. From the comparisonbetween the bulk part and the lobster-eye part, it is obvious that therelative density of lobster-eye parts was always higher than that of thebulk parts.

From the cross-sectional OM images shown in Fig. 3(b), it can beseem that when the laser power was low (325W), the porosity level ofthe cubic part and lobster-eye part was similar to each other. For the

Fig. 5. (a) Schematic diagram of cone angle and wall thickness of the SLM-processed lobster-eye part; (b) the cone angle and wall thickness of SLM-pro-cessed lobster-eye parts with different laser power.

K. Lin et al. Journal of Materials Processing Tech. 267 (2019) 34–43

37

cubic part, several big pores with the size of about 80 μm can be ob-served. For the lobster-eye part, instead of big-size pores, many smallpores were found. With the increase of laser power, the number of poresignificantly reduced, resulting to the increase of relative density.Comparing the cubic part and the lobster-eye part, the porosity level ofthe former was higher than that of the latter, which was in agreementwith the result of relative density shown in Fig. 3(a).

The relation between the laser processing parameters and the den-sification behavior of SLM-processed AlSi10Mg has been well-in-vestigated. In the work of Chen et al. (2017b), the relative density ofSLM-processed AlSi10Mg firstly increases and then decreases with theincrease of laser power, which is in agreement with the results of thiswork (Fig. 3(a)). The underlying reason for this phenomenon could beexplained as follows: when increasing the laser power, more energy wasintroduced to the powder bed, resulting to better fusion of powder, andthus higher relative density; however, when the laser power was toohigh, the intensive energy input cause the formation of thermal cracks(Gu et al., 2012) or layer delamination (Chen et al., 2017b), leading tothe reduction of relative density.

Because of the small-dimension feature of lobster-eye parts, theprocessing condition was significantly different with that of the bulkparts. During SLM processing, the bulk-to-powder ratio near the moltenpool of the bulk part is remarkably larger than that of the thin-wall part(Fig. 4). Alkahari et al. (2012) and Dai and Gu (2015) report that thethermal conductivity of powder is grammatically lower than that of thebulk form of the same material. Therefore, under the same laser power,more heat stayed in the melt pool of the thin-wall part during SLMfabrication, leading to better fusion and consequent higher relativedensity.

3.2. Dimensional accuracy

For the specific lobster-eye parts, particular emphasis was placed onthe forming accuracy of wall thickness (t) and cone angle (θ), whichwere schematically shown in Fig. 5(a). The trends of wall thickness andcone angle against laser power were presented in Fig. 5(b). For the wallthickness, all values were higher than that of the designed value(0.2 mm), which might be related with the attached balls on the wall

Fig. 6. (a) The inner surface and outside surface roughness and (b) the SEM surface morphology of SLM-processed lobster-eye parts with different laser power.

K. Lin et al. Journal of Materials Processing Tech. 267 (2019) 34–43

38

surface. The thickness values increased when the laser power increasedform 325W to 350W, then slightly reduced to a relatively stable valueof 0.214mm when further increased laser power from 375W to 425W.In terms of the cone angle, with the increasing of laser power, the coneangle firstly increased and then reduced when the lase power exceeded375W.

Because the orientation of thin walls is not parallel to the SLMbuilding direction, all the thin walls were downfacing structures withvarious declined angle to the building direction. With the increase oflaser power, the size of melt pool and the volume of liquid metal in-creased. Song et al. (2015) claim that the fully melted liquid tended toflow downward under the influence of gravity, leading to the increase

of cone angle and the increase of thin wall thickness. Yu et al. (2016a)claim that when the laser energy is too high, excessive liquid formedand developed into “self-balling”. Therefore, when further increasedthe laser power, the liquid metal trended to form balls rather than flowdownward, leading to the reduction of cone angle.

3.3. Surface roughness

The surface roughness of the inner surface and outside surface of theSLM-processed lobster-eye parts was presented in Fig. 6(a). It is obviousthat the roughness of inner surface was always higher than that of theoutside surface. For the outside surface, the roughness increased

Fig. 7. The inner surface roughness of different positions in the SLM-process channel with the laser power of 425W: (a) roughness value; surface morphology of (b)bottom position, (c) middle position and (d) top position.

Fig. 8. The influence of lobster-eye structure on the thermal behavior during SLM processing: (a) the conductive heat flow and (b) the distribution of temperaturefield at different building heights.

K. Lin et al. Journal of Materials Processing Tech. 267 (2019) 34–43

39

slightly from 325W to 350W, and then decreased with the increasingof laser power from 350W to 425W. The relation between outsidesurface roughness and laser parameters was in agreement with thestudy of Liu et al. (2018) and Mumtaz and Hopkinson (2009). However,the surface roughness of the inner surface exhibited completely dif-ferent trend comparing with the outside surface. With the increase oflaser power, the roughness of inner surface sharply increased from325W to 350W and then slightly increased from 350W to 425W.

The representative low magnification SEM surface morphologies ofSLM-processed lobster-eye parts were illustrated in Fig. 6(b). Many ball-like features can be observed on the outside surface of the 325Wsample, which might be related to the balling-effect or the attached un-melted powder particles. With the increase of laser power to 425W, lessamount of ball-like features was observed on the outside surface thanthat of the 325W sample, and the proportion of smooth area (dark area)increased, contributing to the reducing surface roughness value(Fig. 6(a)). For the inner surface, with the increase of laser power, morebig size balls can be found, and balls tended to form clusters rather thansingle balls. All these contributed to the remarkable increase of innersurface roughness (Fig. 6(a)).

When the laser power was low, due to the lack of fusion, the laserscan track tended to break into discontinuous dots and consequentlyleaded to the formation of roughed surface. With the increase of laserpower, smooth laser scan tracks formed and the surface roughness

reduced. However, a completely opposite response to laser power waswitnessed for the inner surface roughness. This anomaly can be ex-plained by the particular structure of the lobster-eye. The lobster-eyestructure consists of many small channels, because of the layer-wisefeature of the SLM process, a square box was scanned for each layer.Therefore, the powder, locating in this square box, suffered morethermal cycles than the powder outside the lobster-eye structure. As aconsequence, the powder near the inner surface tend to gather togetherand form clusters (Fig. 6d) on the surface, resulting in the higherroughness of inner surface.

The inner surface roughness of different positions of the channel(laser power 425W) was measured and listed in Fig. 7. The roughnessof the middle part of the channel was the highest, and the roughness ofthe top part and bottom part was close to each other. From SEM imagesof surface morphologies on different positions (Fig. 7b–d), it is obviousthat more ball-like features and more clusters of balls can be observedfrom the middle position surface.

The surface roughness of thin-wall parts is greatly influenced by thelocal temperature of part during SLM processing. Xiong et al. (2018)find that decreasing the local temperature is associated with the de-crease of surface roughness of thin-wall parts. Gu et al. (2018) reportthat the local temperature varied significantly during the SLM proces-sing, especially along the building direction. At the bottom position, theportion being fabricated was close to the bulky substrate plate, which

Fig. 9. Low magnification SEM images of different positions inside the SLM-processed channel under different laser power: (a) top position, (c) middle position and(e) bottom position of 325W processed sample; (b) top position, (d) middle position and (f) bottom position of 400W processed sample.

K. Lin et al. Journal of Materials Processing Tech. 267 (2019) 34–43

40

absorbed a large amount of heat during processing, contributing to thelower local temperature of the portion and thus lower surface rough-ness (Fig. 8(a)). With the increase of building height, the molten poolmoved away from the bulky substrate plate and the heat conduction tothe substrate plate was weakened (Fig. 8(a)). As a consequence, thelocal temperature was relatively higher than that of the bottom posi-tion, resulting in the increase of surface roughness. However, the sur-face roughness at the top position was lower than that of the middleposition. This anomaly can also be explained by the particular structureof the lobster-eye. Because all the channels are arranged over a sphe-rical surface, the cross-section dimension of the channel increases frombottom to top, in other words, the distance between two parallel thinwalls increase with the building height. When a thin wall was underfabrication, the thermal effect to the previous fabricated thin wall was

weaker at the top position than that of the bottom position ((Fig. 8(b)),leading to the reduction of roughness at the top position. Under thecombined influence of those two factors, the inner surface roughnessexhibited the unique behavior at different positions.

3.4. Forming defects

The low-magnification SEM images of different positions inside thechannel with different SLM laser power were compared in Fig. 9. Whenthe laser power was 325W, many interlayer pores can be detected fromall positions. Most of those pores were located near the junction of twoperpendicular thin walls. Comparing the three different position, it canbe found that the amount of interlayer pores in the middle position wasthe highest among the three positions. For the 400W processed part,

Fig. 10. The top-view SEM images of SLM-processed parts with different laser power: (a) and (b) 325W, (c) and (d) 400W.

Fig. 11. (a) SLM-processed whole lobster-eye component, (b) back view of the whole lobster-eye component with a light source placing in front.

K. Lin et al. Journal of Materials Processing Tech. 267 (2019) 34–43

41

though more ball-like features can be seen from the inner surface thanthat of the 325W part, almost no interlayer pore can be found from thethree positions.

The top-view SEM images of the 325W and 400W SLM-processedlobster-eye parts were presented in Fig. 10. All the thin walls, whetherparallel or perpendicular to the recoating direction, exhibited regularshape. The 325W parts showed many discontinuous scan tracks on thetop of thin walls, especially at the junction positions (marked by arrowsin Fig. 10(a)). Apart from that, several pores can also be detected at thejunction positions (marked by arrows in Fig. 10(b)), which might berelated with lack of fusion during SLM processing. For the 400Wsample (Fig. 10(c)), the scan track was more continuous and smooth.Instead of pores, some balls can be found on the top of thin walls(marked by arrow in Fig. 10(d)), which might be related with the “self-balling” effect (Yu et al., 2016b) caused by high laser power.

3.5. Optimal laser processing parameters

From the results presented above, none of the laser power coulddeliver the best outcome in all aspects, namely densification behavior,dimensional accuracy, surface roughness and forming defects. To de-termine the optimal laser power for the SLM-processed lobster-eyecomponent, those aspects should be ranked in order of importance.

The surface roughness of the lobster-eye part, especially the innersurface roughness is important for the optical application. However,due to the nature of SLM process, the side surface roughness of thin-wall structures is difficult to be controlled at a low level. Post-proces-sing, such as electro-chemical polishing, abrasive flow machining orsurface coating, must be conducted to obtain the required surface fin-ishing. The key structural parameters, namely the cone-angle and thethickness of thin-wall, directly affect the optical performance of lobster-eye structure. Between these two structural parameters, the wallthickness can be further modified by post-processing. While, the cone-angle is difficult to change after SLM-processing. Therefore, the di-mensional accuracy of cone-angle (θ) is more important than that ofwall thickness (t). The level of densification behavior directly governsthe forming defect. The pore formed within parts significantly influ-ences the mechanical behavior, especially fatigue property, which isharmful for long-term service of SLM-processed components. The porelocated on the surface, which is difficult to be eliminated by post-pro-cessing, might trap light and greatly diminish the optical performance.Based on the above discussion, the ranking of aspects in order of im-portance was drawn as follows: densification behavior > dimensionalaccuracy (cone-angle)> dimensional accuracy (wall thickness)> sur-face roughness.

According to the order of importance, the laser power of 400W,which delivered the highest relative density and relatively high di-mensional accuracy of cone-angle, was determined as the optimal laserpower. Using the optimal laser parameters, a demo component of full-size lobster-eye was fabricated and presented in Fig. 11.

4. Conclusion

Lobster-eye thin-wall components were fabricated by SLM withAlSi10Mg powder. The influence of key laser processing parameter,namely laser powder, on the densification behavior, dimensional ac-curacy, surface roughness and forming defect was systematically in-vestigated. Based on the results and discussion, several key conclusionsare highlighted as follows:

• The relative density of SLM-processed lobster-eye parts firstly in-creased with the laser power and reached the highest relative den-sity at the laser power of 400W, and then decreased when the laserpower further increased to 425W. Under the identical laser pro-cessing parameters, the relative density of the lobster-eye part washigher than that of the bulk parts, thanks to the small dimension

feature of thin-wall structures.

• The roughness of inner surface and outside surface showed com-pletely different responds to the laser power. With the increase oflaser power, the outside surface roughness gradually reduced,however, an increase of roughness was witnessed for the innersurface. In addition, the inner surface morphology behaved differ-ently along the building direction. All these phenomena were re-lated with the particular geometry of the lobster-eye structure.

• Considering the function and post-processing of lobster-eye parts, aranking of aspects in order of importance was proposed, which wasdensification behavior > dimensional accuracy > surface rough-ness. According to this ranking, 400W was the optimal laser powerfor the SLM-processed lobster-eye parts.

An important finding of this work was that not only the laserparameters, but also the structural features influenced the formabilityof SLM-processed components. Therefore, when employing SLM tech-nology to fabricate parts with complex and fine geometries, the influ-ence of structural features must be considered at the stage of structuredesign and SLM parameter design.

Acknowledgments

The authors gratefully acknowledge the financial support from theNational Natural Science Foundation of China (No. 51735005).

References

Alkahari, M.R., Furumoto, T., Ueda, T., Hosokawa, A., Tanaka, R., Abdul Aziz, M.S., 2012.Thermal conductivity of metal powder and consolidated material fabricated via se-lective laser melting. Key Eng. Mater. 523–524, 244–249. https://doi.org/10.4028/www.scientific.net/KEM.523-524.244.

Angel, J.R.P., 1979. Lobster eyes as X-ray telescopes. Astrophys. J. 233, 364–373.Boegelein, T., Louvis, E., Dawson, K., Tatlock, G.J., Jones, A.R., 2016. Characterisation of

a complex thin walled structure fabricated by selective laser melting using a ferriticoxide dispersion strengthened steel. Mater. Charact. 112, 30–40. https://doi.org/10.1016/j.matchar.2015.11.021.

Bontha, S., Klingbeil, N.W., Kobryn, P.A., Fraser, H.L., 2006. Thermal process maps forpredicting solidification microstructure in laser fabrication of thin-wall structures. J.Mater. Process. Technol. 178, 135–142. https://doi.org/10.1016/j.jmatprotec.2006.03.155.

Calignano, F., 2018. Investigation of the accuracy and roughness in the laser powder bedfusion process. Virtual Phys. Prototyp. 13, 97–104. https://doi.org/10.1080/17452759.2018.1426368.

Calignano, F., Cattano, G., Manfredi, D., 2018. Manufacturing of thin wall structures inAlSi10Mg alloy by laser powder bed fusion through process parameters. J. Mater.Process. Technol. 255, 773–783. https://doi.org/10.1016/j.jmatprotec.2018.01.029.

Chen, H., Zhang, P., Zhang, L., Liu, H., Jiang, Y., Zhang, D., Han, Z., Jiang, L., 2016.Continuous directional water transport on the peristome surface of Nepenthes alata.Nature 532, 85–89. https://doi.org/10.1038/nature17189.

Chen, H., Gu, D., Xiong, J., Xia, M., 2017a. Improving additive manufacturing proces-sability of hard-to-process overhanging structure by selective laser melting. J. Mater.Process. Technol. 250, 99–108. https://doi.org/10.1016/j.jmatprotec.2017.06.044.

Chen, Z., Wei, Z., Wei, P., Chen, S., Lu, B., Du, J., Li, J., Zhang, S., 2017b. Experimentalresearch on selective laser melting AlSi10Mg alloys: process, densification and per-formance. J. Mater. Eng. Perform. 26, 5897–5905. https://doi.org/10.1007/s11665-017-3044-5.

Dai, D., Gu, D., 2015. Tailoring surface quality through mass and momentum transfermodeling using a volume of fluid method in selective laser melting of TiC/AlSi10Mgpowder. Int. J. Mach. Tools Manuf. 88, 95–107. https://doi.org/10.1016/j.ijmachtools.2014.09.010.

Denlinger, E.R., Heigel, J.C., Michaleris, P., 2015. Residual stress and distortion modelingof electron beam direct manufacturing Ti-6Al-4V. Proc. Inst. Mech. Eng. Part B: J.Eng. Manuf. 229, 1803–1813. https://doi.org/10.1177/0954405414539494.

Fatemi, S.A., Ashany, J.Z., Aghchai, A.J., Abolghasemi, A., 2017. Experimental in-vestigation of process parameters on layer thickness and density in direct metal lasersintering: a response surface methodology approach. Virtual Phys. Prototyp. 12,133–140. https://doi.org/10.1080/17452759.2017.1293274.

Finnemore, A., Cunha, P., Shean, T., Vignolini, S., Guldin, S., Oyen, M., Steiner, U., 2012.Biomimetic layer-by-layer assembly of artificial nacre. Nat. Commun. 3, 966. https://doi.org/10.1038/ncomms1970.

Gu, D., Dai, D., 2016. Role of melt behavior in modifying oxidation distribution using aninterface incorporated model in selective laser melting of aluminum-based material.J. Appl. Phys. 120. https://doi.org/10.1063/1.4961410.

Gu, D., Hagedorn, Y.C., Meiners, W., Meng, G., Batista, R.J.S., Wissenbach, K., Poprawe,R., 2012. Densification behavior, microstructure evolution, and wear performance ofselective laser melting processed commercially pure titanium. Acta Mater. 60,

K. Lin et al. Journal of Materials Processing Tech. 267 (2019) 34–43

42

3849–3860. https://doi.org/10.1016/j.actamat.2012.04.006.Gu, D., Shi, Q., Lin, K., Xi, L., 2018. Microstructure and performance evolution and un-

derlying thermal mechanisms of Ni-based parts fabricated by selective laser melting.Addit. Manuf. 22, 265–278. https://doi.org/10.1016/J.ADDMA.2018.05.019.

Habibi, M.K., Samaei, A.T., Gheshlaghi, B., Lu, J., Lu, Y., 2015. Asymmetric flexuralbehavior from bamboo’s functionally graded hierarchical structure: underlying me-chanisms. Acta Biomater. 16, 178–186. https://doi.org/10.1016/j.actbio.2015.01.038.

Li, Z., Xu, R., Zhang, Z., Kucukkoc, I., 2018. The influence of scan length on fabricatingthin-walled components in selective laser melting. Int. J. Mach. Tools Manuf. 126,1–12. https://doi.org/10.1016/j.ijmachtools.2017.11.012.

Liu, Y., Zhang, J., Pang, Z., Wu, W., 2018. Investigation into the influence of laser energyinput on selective laser melted thin-walled parts by response surface method. Opt.Lasers Eng. 103, 34–45. https://doi.org/10.1016/j.optlaseng.2017.11.011.

Mumtaz, K., Hopkinson, N., 2009. Top surface and side roughness of Inconel 625 partsprocessed using selective laser melting. Rapid Prototyp. J. 15, 96–103. https://doi.org/10.1108/13552540910943397.

Peele, A.G., Vora, K.D., Shew, B.Y., Loechl, B., Harvey, E.C., Hayes, J.P., 2007. Productionissues for high aspect ratio Lobster-eye optics using LIGA. Microsyst. Technol. 13,511–515. https://doi.org/10.1007/s00542-006-0188-x.

NASA, 2018. Proposed NASA Mission Employs “Lobster-Eye” Optics to Locate Source ofCosmic Ripples. https://www.nasa.gov/feature/goddard/2017/proposed-nasa-mission-employs-lobster-eye-optics-to-locate-source-of-cosmic-ripples.

Qin, Z., Compton, B.G., Lewis, J.A., Buehler, M.J., 2015. Structural optimization of 3D-printed synthetic spider webs for high strength. Nat. Commun. 6, 1–7. https://doi.org/10.1038/ncomms8038.

Shishkovskii, I.V., Safronov, V.A., Protasov, K.E., 2016. Producing thin-walled BrA9bronze components by selective laser melting. Russ. Eng. Res. 36, 335–338. https://doi.org/10.3103/S1068798X16040171.

Sing, S.L., Wiria, F.E., Yeong, W.Y., 2018. Selective laser melting of lattice structures: astatistical approach to manufacturability and mechanical behavior. Robot. Comput.Manuf. 49, 170–180. https://doi.org/10.1016/j.rcim.2017.06.006.

Song, C., Yang, Y., Liu, Y., Luo, Z., Yu, J.K., 2015. Study on manufacturing of W-Cu alloythin wall parts by selective laser melting. Int. J. Adv. Manuf. Technol. 78, 885–893.https://doi.org/10.1007/s00170-014-6689-3.

Sutton, A.T., Kriewall, C.S., Leu, M.C., Newkirk, J.W., 2017. Powder characterisationtechniques and effects of powder characteristics on part properties in powder-bedfusion processes. Virtual Phys. Prototyp. 12, 3–29. https://doi.org/10.1080/17452759.2016.1250605.

Wang, D., Wang, Y., Wang, J., Song, C., Yang, Y., Zhang, Z., Lin, H., Zhen, Y., Liao, S.,2016. Design and fabrication of a precision template for spine surgery using selectivelaser melting (SLM). Materials (Basel) 9. https://doi.org/10.3390/ma9070608.

Xiong, J., Li, Y., Li, R., Yin, Z., 2018. Influences of process parameters on surfaceroughness of multi-layer single-pass thin-walled parts in GMAW-based additivemanufacturing. J. Mater. Process. Technol. 252, 128–136. https://doi.org/10.1016/j.jmatprotec.2017.09.020.

Yadroitsev, I., Bertrand, P., Smurov, I., 2007. Parametric analysis of the selective lasermelting process. Appl. Surf. Sci. 253, 8064–8069. https://doi.org/10.1016/j.apsusc.2007.02.088.

Yu, G., Gu, D., Dai, D., Xia, M., Ma, C., Chang, K., 2016a. Influence of processing para-meters on laser penetration depth and melting/re-melting densification during se-lective laser melting of aluminum alloy. Appl. Phys. A Mater. Sci. Process. 122, 1–12.https://doi.org/10.1007/s00339-016-0428-6.

Yu, G., Gu, D., Dai, D., Xia, M., Ma, C., Shi, Q., 2016b. On the role of processing para-meters in thermal behavior, surface morphology and accuracy during laser 3Dprinting of aluminum alloy. J. Phys. D Appl. Phys. 49. https://doi.org/10.1088/0022-3727/49/13/135501.

K. Lin et al. Journal of Materials Processing Tech. 267 (2019) 34–43

43