Applying Selective Laser Melting to Join Al and Fe: an Investigation of Dissimilar Materials

applied sciences

Article Applying Selective Laser Melting to Join Al and Fe: An Investigation of Dissimilar Materials

Dinh-Son Nguyen, Hong-Seok Park * and Chang-Myung Lee School of Mechanical & Automotive Engineering, University of Ulsan, Ulsan 680-749, Korea * Correspondence: [email protected]; Tel.: +82-052-259-1458

Received: 28 June 2019; Accepted: 23 July 2019; Published: 27 July 2019

Abstract: Combining aluminum and steel is a major goal of automobile manufacturers and other industries because the hybrid material reduces the weight of components. However, differences in chemical properties, thermal expansion, and physical characteristics of aluminum and steel are barriers to achieving this goal. In this article, selective laser melting (SLM), which is widely used in industrial fields, was applied to join dissimilar materials by printing aluminum on a steel substrate. Defects of joining during the SLM process, characteristics of the intermetallic reaction layer, and the effects of the process parameters were investigated. The analysis indicates that flake behavior could affect the quality of joining. The phases of the intermetallic layer found in this study were in agreement with other research, but the morphology of the layer was much different. A formula to estimate the join quality in terms of density energy is proposed. The results indicate that the SLM process is a promising method to manufacture a hybrid material.

Keywords: selective laser melting; dissimilar material joining; aluminum; stainless steel

1. Introduction The joining of aluminum, a low-density material with good corrosion resistance, and steel, a high-strength material with good formability, has great potential to meet the increasing demands of industries such as automotive manufacturing. The combination of aluminum and steel has unique physical and mechanical properties, such as a reduction of product weight, which is a significant advantage of aluminum–steel bimetallic parts. However, large differences in the melting temperatures, thermal conductivities, and thermal expansions of steel and aluminum, as well as low solubility, makes it quite difficult to join them. A variety of methods have been applied to join dissimilar materials, such as mechanical assembling (e.g., riveting, clinching, screwing) and heat joining (e.g., friction stir welding [1,2], arc welding [3,4], laser welding [5–9], hot-dip aluminizing [10]). Selective laser melting (SLM) is an additive manufacturing process that is widely used in many industrial areas. The SLM printing process includes of a series of steps, ranging from computer-aided design data preparation to the removal of a fabricated component from the building platform. A typical configuration of an SLM printer is shown in Figure1[ 11]. In the beginning, the piston head is raised to lift the material powder. At the same time, the substrate in the build cylinder is dropped to a distance that is equal to the layer thickness. The scrapper travels from the feed container to the overflow container to create a layer of powder on the substrate and then comes back to the initial position. The laser scans the surface of fabricated bed area basing on each slide data. The processes are repeated until the part has been finished. The SLM allows fabrication of complex-shaped parts that cannot be manufactured by other traditional methods, with high density and accuracy of the printed product. Therefore, SLM is attractive for manufacture of automotive, aerospace, and medical components. However, most SLM printers that are currently commercially available have been designed to produce parts from a single material. Hybrid parts manufactured by selective laser melting

Appl. Sci. 2019, 9, 3031; doi:10.3390/app9153031 www.mdpi.com/journal/applsci Appl. Sci. 2019, 9, x FOR PEER REVIEW 2 of 12 Appl. Sci. 2019, 9, 3031 2 of 12 designed to produce parts from a single material. Hybrid parts manufactured by selective laser melting were investigated [12]. The concept of multi-material fabrication by SLM has been reported were investigated [12]. The concept of multi-material fabrication by SLM has been reported in the in the literature [8–9]; however, these products were printed by physically mixing powders before literature [8,9]; however, these products were printed by physically mixing powders before layering. layering.

Figure 1. The selective laser melting (SLM) printing process [[11].11].

Flake Behavior When investigating the printing process paramete parametersrs of powder bed fusion printing, the ability to spread a susufficientfficient new layer of powder is the major determinant of product quality. However, However, some unwanted behaviors may occur duringduring thethe selectiveselective laserlaser meltingmelting process.process. Delamination or cracking of the support section (shown by the dashed circles in Figure2 2a,b)a,b) occuroccur insideinside thethe printedprinted part itself and are duedue toto processprocess parameters.parameters. Ho However,wever, the flaking flaking shown in Figure 22cc occurredoccurred because of material didifferencesfferences additionally.additionally.

Figure 2. Delamination ( a), support cracking ( b), and flaking ﬂaking (c).

Because of the flake, flake, the printed section and the scrapper are conflicted. conflicted. Therefore, Therefore, the spreading process cancan createcreate a a problem; problem; in in the the worst worst case, case, the the SLM SLM process process must must be stopped. be stopped. Therefore, Therefore,the flake the phenomenonflake phenomenon can represent can represent the join the qualityjoin quality of hybrid of hybrid materials. materials. In thisIn this study, study, the the SLM SLM process process of pureof pure aluminum aluminum powder powder on on stainless stainless steel steel 316 316 L substrateL substrate was was analyzed analyzed by by investigating investigating the the eff effectsects of processof process parameters parameters and and flake flake behavior. behavior.

2. Materials and Methods In this this study, study, 99.9% 99.9% pure pure aluminum aluminum powder powder was was printed printed on ona stainless-steel a stainless-steel 316 L 316 substrate. L substrate. The Thepowder powder particle particle size of size the of aluminum the aluminum powder powder was 30–45 was 30–45µm. Aµ scanningm. A scanning electron electron microscopy microscopy (SEM) (SEM)image of image the aluminum of the aluminum powder powder(HKK Solution (HKK SolutionCo., Ltd., Co., Seoul, Ltd., South Seoul, Korea) South is shown Korea) in is Figure shown 3a. in Figure 33ba. Figureindicated3b indicatedthe particle the size particle distribution size distribution by the laser by thescattering laser scattering analyzer analyzerLA-960. LA-960.An SLM Anprinter SLM (MetalSys printer (MetalSys 150, WinforSys 150, WinforSys co., Ltd., Gyeonggi co., Ltd.,-do, Gyeonggi-do, South Korea) South with Korea) an air-cooled with an air-cooledytterbium Appl. Sci. 2019, 9, x FOR PEER REVIEW 3 of 12

fiber laser (YLR-200-AC-Y11 IPG Ytterbium Fiber Laser, 200 W maximum output) was used for processing. Table 1 describes the technical parameters of the printer.

Table 1. Technical parameters of the selective laser melting (SLM) MetalSys 150.

Parameter Value Wavelength 1075 nm Appl. Sci. 2019, 9, 3031 Maximum output power 200 W 3 of 12 Beam quality <1.1 Beam spot 70 µm ytterbium ﬁber laser (YLR-200-AC-Y11Building IPG size Ytterbium Fiber 150 × Laser,150 × 250 200 mm W maximum output) was used for processing. Table1 describesMaximum the technical scanning parameters speed of the 7 m/s printer.

FigureFigure 3. 3.(a )(a SEM) SEM image image of of pure pure aluminum aluminum powderpowder particles, particles, (b (b) )powder powder particle particle size size distribution, distribution, andand (c) (c laser) laser scanning scanning strategy. strategy.

A seriesTable of aluminum 1. Technical cubes parameters were ofprinted the selective on the laser stainless-steel melting (SLM) subs MetalSystrate without 150. support sections. All steel substrates investigated in this study were 10 mm thick to prevent deformation Parameter Value during processing. Prior to printing, the substrates were polished with 2000-grit paper and cleaned with ethanol. The printing processWavelength used the meander-scanning 1075 pattern nm shown in Figure 3c, in which the laser scan direction wasMaximum parallel to output the x-axis power in every layer. 200 Layer W thickness was 20 µm for every Beam quality <1.1 experiment. The fabricating chamber was set at 28 °C initially. The argon gas was pumped into the Beam spot 70 µm chamber at 25 L/min of flow rateBuilding and 0.05 size MPa of pressure 150 to150 obtain250 an mm oxygen level below less than × × 0.1 percent. After achievingMaximum a setup scanningoxygen level, speed the gas supply 7 m /wass cut off and atmosphere in the manufacturing chamber was recycled. Because SLM printing is an accumulative process, the flake was evaluated by observing the quantityA series of ofpossible aluminum printed cubes layers were before printed they on thereached stainless-steel the surface substrate of the substrate. without support The printing sections. Allprocess steel substrates was stopped investigated immediately in at this the study occurrence were of 10 flake. mm thickAfter printing, to prevent the deformation samples were during cut processing.using wire-cut Prior electrical to printing, discharge the substrates machining. were Then, polished they were with hot-press 2000-grit embedded paper and in cleanedan electro- with ethanol.conductive The printingresin, ground process with used SiC theabrasive meander-scanning paper with a grit pattern range shownup to 4000, in Figure and polished3c, in which with thea laserdiamond scan direction paste having was an parallel average to grain the size x-axis of 2 in µm. every Final layer. polishing Layer used thickness a colloidal was silica 20 µsuspensionm for every experiment.of 0.04 µm. The Scanning fabricating electron chamber microscopy was set (SEM) at 28 and◦C initially.energy dispersive The argon spectroscopy gas was pumped (EDS) intowere the chamberused for at intermetallic 25 L/min of observation flow rate and and 0.05 microstructural MPa of pressure characterization. to obtain an oxygen level below less than 0.1 percent. After achieving a setup oxygen level, the gas supply was cut off and atmosphere in the manufacturing3. Results and chamber Discussion was recycled. Because SLM printing is an accumulative process, the flake was evaluated by observing the 3.1. Deformation in the Selective Laser Melting (SLM) quantity of possible printed layers before they reached the surface of the substrate. The printing processDue was to stopped the layer-by-layer immediately principle at the occurrence of the SLM of flake.printing After process, printing, the theprinted samples section were becomes cut using wire-cutincreasingly electrical thicker, discharge and products machining. are formed Then, theyprogre weressively. hot-press Therefore, embedded to obtain in ana cubic electro-conductive thickness of resin,10 mm, ground the process with SiC must abrasive complete paper the with riskiest a grit period range first, up before to 4000, a 10 and × polished10 mm plate with of ametal diamond is pasteformed. having an average grain size of 2 µm. Final polishing used a colloidal silica suspension of 0.04 µm. Scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS) were used for intermetallic observation and microstructural characterization.

3. Results and Discussion

3.1. Deformation in the Selective Laser Melting (SLM) Due to the layer-by-layer principle of the SLM printing process, the printed section becomes increasingly thicker, and products are formed progressively. Therefore, to obtain a cubic thickness of 10 mm, the process must complete the riskiest period ﬁrst, before a 10 10 mm plate of metal is formed. × Appl. Sci. 2019, 9, 3031 4 of 12

Appl. Sci. 2019, 9, x FOR PEER REVIEW 4 of 12 Residual stress is always present in products printed by SLM [9,10], as presented in Figure4.

BecauseAppl. new Sci.Residual 2019 layers, 9, x stress FOR are PEER printed is always REVIEW on present other solidified in products layers, printed flake by behavior SLM [9,10], is theas presented result of an in accumulativeFigure4 of 124. process.Because After new printing layers aare layer printed shown on inother Figure solidified4a, the layers, solidified flake layers behavior increase, is the creatingresult of aan sheet indicatedaccumulativeResidual in Figure stressprocess.4b. is In always theAfter first printipresent severalng ina productslayer layers, shown theprinted thicknessin Figureby SLM of4a, [9,1 thethe0], printed solidifiedas presented layers layers in was Figureincrease, too 4. thin; therefore,Becausecreating there anew sheet was layers noindicated steep are temperature printedin Figure on 4b. other gradient In the solidified firs witht several thickness. layers, layers, flake Duringthe behaviorthickness heating, ofis thethe thermal printedresult compressive oflayers an was too thin; therefore, there was no steep temperature gradient with thickness. During heating, stressesaccumulative develop in process. the printed After layers, printing which a layer results shown in a in large Figure amount 4a, the of thermo-elasticsolidified layers strain, increase, possibly creatingthermal acompressive sheet indicated stresses in Figure develop 4b. Inin the the firs printet severald layers, layers, which the thickness results in of athe large printed amount layers of producing local thermo-elasto-plastic buckling of the material. When the laser beam of the printer wasthermo-elastic too thin; therefore, strain, possibly there wasproducing no steep local temper thermo-elasto-plasticature gradient with buckling thickness. of the During material. heating, When passes,thermalthe the laser buckle compressive beam isof generatedthe printerstresses passes, along developthe the directioninbuckle the isprinte generated of laserd layers,scanning along which the direction[results13]. in of alaser large scanning amount [13]. of thermo-elastic strain, possibly producing local thermo-elasto-plastic buckling of the material. When the laser beam of the printer passes, the buckle is generated along the direction of laser scanning [13].

FigureFigure 4. Temperature 4. Temperature gradient gradient mechanism mechanism [ 11[11]:]: schematicschematic of of the the SLM SLM processing processing (a) created (a) created a new a new layer onlayer other on solidifiedother solidified layers layers (b), thermal (b), thermal expansion expansion generated generated by heated by heated new layernew (layerc) and (c) shrinkage and processFigureshrinkage during 4. Temperature process cooling during (d ).gradient cooling mechanism (d). [11]: schematic of the SLM processing (a) created a new layer on other solidified layers (b), thermal expansion generated by heated new layer (c) and AftershrinkageAfter achieving achieving process a certain during a certain thickness, cooling thickness, (d). because because of heating, of heating, a steep thermala steep gradientthermal ingradient the manufactured in the layersmanufactured leads to diff erenceslayers leads in thermal to differences expansion in thermal throughout expansion the throughout thickness. the At thickness. the initial At stage,the initial the top stage, the top surface is heated because it faces the laser shown in Figure 4c. Rapid thermal expansion surface isAfter heated achieving because ita facescertain the thickness, laser shown because in Figure of heating,4c. Rapid a steep thermal thermal expansion gradient of the in topthe solid manufacturedof the top solid layers layers leads in contrast to differences to the bottom in thermal solid expansion layers, as throughout well as restriction the thickness. by the Atsurrounding the initial layers in contrast to the bottom solid layers, as well as restriction by the surrounding material, creates stage,material, the topcreates surface a concave is heated downward because it facesshape. th eAt laser a certain shown temperature, in Figure 4c. Rapidthe shape thermal and expansion degree of a concaveofbending the top downward produce solid layers maximum shape. in contrast Atelastic a to certain strain.the bottom Any temperature, solidaddition layers,al thethermal as shapewell expansion as andrestriction degree is converted by ofthe bending surrounding into plastic produce maximummaterial,compression. elastic creates After strain. a concave the Anylaser downward additionalmoves away, shape. thermal cooling At en expansiona sues,certain and temperature, shrinkage is converted occurs the intoshape shown plastic and in Figuredegree compression. 4d.of Afterbending the laser produce moves maximum away, cooling elastic ensues,strain. Any and addition shrinkageal thermal occurs expansion shown in is Figureconverted4d. into plastic compression.3.2. Intermetallic After Layer the laser moves away, cooling ensues, and shrinkage occurs shown in Figure 4d. 3.2. Intermetallic Layer The printing process creates molten aluminum on the solid substrate. That aluminum melts on 3.2. Intermetallic Layer Thethe solid printing surface process of the createssubstrate molten to create aluminum a diffusion on area the called solid an substrate. intermetallic That layer, aluminum which exists melts on the solidas theThe surface interface printing of between the process substrate the creates printed to molten create section aluminum a di andffusion the onsubstrate. area the solid called Formation substrate. an intermetallic of That the aluminumintermetallic layer, melts which phases on exists as thethebetween interface solid surfacethe between solid of the iron/steel the substrate printed antosectiond create liquid a and diffusioaluminum then substrate. area depends called Formation an onintermetallic the ofchemical the layer, intermetallic reactionwhich exists and phases interdiffusion of the elemental constituents. Figure 5 demonstrates the SEM morphology and EDS betweenas the the interface solid iron between/steel the and printed liquid section aluminum and the depends substrate. on theFormation chemical of the reaction intermetallic and interdi phasesffusion betweenmapping theresults solid of a iron/steelsample, which and indicateliquid aaluminum distribution depends of Fe in Figureon the 5b chemical and Al shown reaction in Figure and of the elemental constituents. Figure5 demonstrates the SEM morphology and EDS mapping results interdiffusion5c elements. The of theresults elemental show thatconstituents. the Fe element Figure di 5ffused demonstrates into molten the aluminumSEM morphology to a larger and extent EDS of a sample,mappingthan the whichAlresults element of indicate a diffusedsample, a distribution whichinto the indicate stainless-steel of a Fedistribution in substrate. Figure of5 bFe andin Figure Al shown 5b and inAl Figureshown 5inc Figure elements. The results5c elements.show The that results the Fe show element that the di ffFeused element into di moltenffused into aluminum molten aluminum to a larger to extenta larger than extent the Al elementthan di theffused Al element into the diffused stainless-steel into the stainless-steel substrate. substrate.

Figure 5. (a) Energy dispersive spectroscopy (EDS) image, (b) Fe element distribution, and (c) Al element distribution. FigureFigure 5. ( a5.) Energy(a) Energy dispersive dispersive spectroscopy spectroscopy (EDS) image, image, (b ()b Fe) Fe element element distribution, distribution, and and(c) Al ( c) Al elementelement distribution. distribution. Appl. Sci. 2019, 9, 3031 5 of 12

Appl. Sci. 2019, 9, x FOR PEER REVIEW 5 of 12 Figure6 shows an EDS point analysis of a flaked sample with 31.11 at% iron and 68.89 at% aluminum.Figure The 6 analysisshows an confirms EDS point that analysis the Fe of2Al a5 flakedphase sample exists inwith the 31.11 intermetallic at% iron layer.and 68.89 The at% results agreealuminum. with previous The analysis research confirms using otherthat the methods Fe2Al5 phase [14–17 exists]. However, in the intermetallic the morphology layer. The of the results reaction layeragree during with theprevious SLM research process using is diff othererent methods from the [14–17]. previous However, methods. the morphology Although of previous the reaction studies reportedlayer tongue-like,during the SLM needle-like, process is ordifferent wave-like from [ 18th–e21 previous] morphology, methods. the Although SLM process previous shows studies a more chaoticreported morphology tongue-like, of theneedle-like, reaction or layer, wave-like as indicated [18–21] morphology, in Figure7. the The SLM chaotic process morphology shows a more can be chaotic morphology of the reaction layer, as indicated in Figure 7. The chaotic morphology can be explainedchaotic bymorphology the molten of material the reaction flow layer, during as indicated melting andin Figure evaporation 7. The chaotic during morphology the SLM process can be [22]. explained by the molten material flow during melting and evaporation during the SLM process [22]. Additionally, denudation [23], where a laser interacts with metal powder, results in non-repetitive Additionally, denudation [23], where a laser interacts with metal powder, results in non-repetitive formation of an intermetallic layer. formation of an intermetallic layer.

FigureFigure 6. Results6. Results of of the the EDS EDS point point analysisanalysis of the the intermetallic intermetallic layer layer of ofa flaked a ﬂaked sample. sample.

Figure 7. Morphology of the intermetallic layers formed during the selective laser melting process. Appl.Appl. Sci. Sci.2019 2019, 9,, 30319, x FOR PEER REVIEW 6 of6 12 of 12

Figure 7. Morphology of the intermetallic layers formed during the selective laser melting process. The intermetallic layer thickness was determined by detecting the Fe-Al change, as shown by the two dashedThe intermetallic lines in Figure layer8. thickness The thickness was determined is associated by with detecting the SLM the Fe-Al process change, parameters as shown of timeby the two dashed lines in Figure 8. The thickness is associated with the SLM process parameters of time and temperature. An increase in laser power increases the thickness of the reaction layer, whereas an and temperature. An increase in laser power increases the thickness of the reaction layer, whereas an increase in laser speed decreases the thickness of the metallic layer. An increase in laser power causes increase in laser speed decreases the thickness of the metallic layer. An increase in laser power causes an increase in energy and thus an increase in peak temperature. Higher temperatures and longer an increase in energy and thus an increase in peak temperature. Higher temperatures and longer reaction times increase the thickness of the reaction layer. Moreover, a higher laser scanning velocity reaction times increase the thickness of the reaction layer. Moreover, a higher laser scanning velocity resultsresults in in a lowera lower peak peak temperature temperature and and leads leads toto aa reduced thickness thickness of of the the bimetallic bimetallic layer. layer. The The EDS EDS resultsresults show show that that the thicknessthe thickness of the of metallic the metallic layer rangeslayer ranges from 3 µfromm to 3 70 µmµm. to When 70 µm. the intermetallicWhen the layerintermetallic is thin, the layer mechanical is thin, the performance mechanical performanc of the jointe isof notthe joint yet detrimental, is not yet detrimental, as shown as in shown Figure in8 a. However,Figure 8a. the However, joint will the fail joint when will the fail reaction when layerthe re thickens,action layer as thickens, shown in as Figure shown8b. in These Figure results 8b. These agree withresults previous agree studies with previous [24]. studies [24].

FigureFigure 8. Intermetallic8. Intermetallic layer layer thickness thickness determined determined by EDS by line EDS mapping line mapping analysis: analysis: (a) a thin intermetallic(a) a thin layerintermetallic [11] and ( blayer) a thick [11] and reaction (b) a layerthick reaction that creates layer a that flake. creates a flake. 3.3. Influence of Process Parameters on Join Quality of Dissimilar Materials 3.3. Influence of Process Parameters on Join Quality of Dissimilar Materials InIn the the SLM SLM process, process, the the intermetallic intermetallic reactionreaction layerlayer plays a a crucial crucial role role against against the the residual residual stress stress generatedgenerated by by the the thermal thermal gradient. gradient. The The intermetallicintermetallic laye layerr is is brittle brittle in in the the area area close close to to the the side side of ofthe the substratesubstrate [25 [25],], and and the the strength strength depends depends onon itsits thickness.thickness. A A thinner thinner intermetallic intermetallic layer layer has has greater greater strength,strength, and and vice vice versa. versa. Therefore, Therefore, toto preventprevent a low joining strength, strength, the the thickness thickness of of the the reaction reaction layerlayer should should be be limited limited [26 [26].]. The The flake flake behaviorbehavior waswas investigated by by observing observing the the number number of oflayers layers thatthat could could be be printed printed before before flaking flaking occurred. occurred. FigureFigure9 9 showsshows the experimental experimental results results of of the the observed observed quantitiesquantities of of printed printed layers layers in in the the boundary boundary areaarea inin whichwhich the flaking flaking occurred. occurred. AtAt the the beginning beginning of of the the process, process, whenwhen thethe printedprinted section is is not not thick thick enough, enough, residual residual stress stress fromfrom the the thermal thermal gradient gradient in in the the X X and and Y Y directionsdirections is overwhelming; thus, thus, an an intermetallic intermetallic layer layer can can preventprevent cracking. cracking. Our Our experiment experiment showed showed thatthat flakingflaking appeared appeared around around the the 11th 11th layer layer but but no nosooner. sooner. WhenWhen the the printed printed section section became became increasingly increasingly thicker, thicker, the the temperature temperature di ffdifferenceerence between between the the top top and intermetallicand intermetallic layers waslayers greater. was greater. The intermetallic The intermetalli layerc alsolayer su alsoffers suffers from greaterfrom greater residual residual stress stress created bycreated the thermal by the gradient thermal in gradient the thickness in the direction.thickness direction. However, However, when the when printed the section printed achieved section a certainachieved thickness, a certain the thickness, effect of the effect top layer of the on to thep layer intermetallic on the intermetallic layer was layer reduced, was reduced, and the riskand of the risk of flaking decreased. It is reasonable to assume that further flaking did not occur after the flaking decreased. It is reasonable to assume that further flaking did not occur after the 25th layer. 25th layer. Figure9a demonstrates the laser power e ffect at 1.5 m/s, 2 m/s, and 4 m/s presented by the triangle, Figure 9a demonstrates the laser power effect at 1.5 m/s, 2 m/s, and 4 m/s presented by the circle, and rectangle makers, respectively. As shown, the printed layers decreased when the laser triangle, circle, and rectangle makers, respectively. As shown, the printed layers decreased when the power increased. Increasing laser power results in a high peak temperature, increasing the residual laser power increased. Increasing laser power results in a high peak temperature, increasing the stressresidual by increasing stress by increasing the thermal the transient thermal transient in each layer.in each Increasing layer. Increasing laser powerlaser power results results in a flakedin a happeningflaked happening sooner.Figure sooner.9a Figure also shows 9a also that, shows at a that, fixed at laser a fixed power, laser flaking power, occurred flaking occurred also sooner alsoat a lower laser scanning speed. At a certain printed thickness, a lower laser velocity creates a weaker thermal transient between the top layer and intermetallic layers, reducing residual stress. Additionally, Appl. Sci. 2019, 9, x FOR PEER REVIEW 7 of 12 Appl. Sci. 2019, 9, 3031 7 of 12 sooner at a lower laser scanning speed. At a certain printed thickness, a lower laser velocity creates a weaker thermal transient between the top layer and intermetallic layers, reducing residual stress. aAdditionally, reduction in a laser reduction velocity in raisedlaser velocity the peak raised temperature. the peak Higher temperature. temperatures Higher and temperatures longer reaction and timeslonger lead reaction to an times increase lead of to reaction an increase layer of thickness reaction andlayer an thickness easier flaked and an appearance. easier flaked appearance. The influenceinfluence ofof hatchhatch distancedistance on on flaking flaking is is shown shown in in Figure Figure9b. 9b. Laser Laser power power was was investigated investigated at 100at 100 W, 150W, W150 and W 180and W 180 displayed W displayed by the rectangle,by the rectangle, circle and circle triangle and shapes, triangle respectively. shapes, respectively. Increasing theIncreasing hatch distance the hatch increased distance the incr numbereased ofthe layers number that of could layers be printed.that could Increasing be printed. the Increasing hatch distance the alsohatch reduced distance the also heating reduced time the in heating certain time printed in ce areas,rtain printed reducing areas, the thermalreducing transient the thermal in a transient printing layerin a printing and residual layer stress.and residual stress.

Figure 9. EffectsEﬀects of laser power (a), and hatch distance (b) on ﬂaking.flaking.

The function of of the the intermetallic intermetallic layer layer is isto to act act as asprotection protection against against the the residual residual stress stress force. force. As Asmentioned mentioned in thein the previous previous section, section, high high temperature temperature tends tends to to cause cause concave concave bending bending in in the printed layers, whichwhich leadsleads to to the the flake flake phenomenon. phenomenon. The The initial initial residual residual stress stress force force is not is not enough enough to disrupt to disrupt the surfacethe surface of the of substrate the substrate because becaus of thee of thin the printed thin printed layers. layers. During During accumulative accumulative processes, processes, an increase an ofincrease printed of layers printed causes layers the causes residual the stress residual force stre toss increase. force to With increase. an unsuitable With an processunsuitable parameter, process theparameter, residual the stress residual force overwhelmsstress force overwhelms the opposing the force opposing existing force in the existing intermetallic in the intermetallic layers and leads layers to flaking.and leads However, to flaking. when However, the printed when layersthe printed achieve layers a certain achieve thickness, a certain concave thickness, bending concave transforms bending intotransforms convex into bending convex [27], bending which is [27], an advantage which is ofan flaking. advantage In our of case,flaking. experimental In our case, results experimental show that theresults 11th show and 25ththat layersthe 11th are and the 25th lower layers and upperare the boundaries lower and of upper flaking boundaries behavior, of respectively. flaking behavior, respectively.

3.4. Join Quality in Term of Energy Appl. Sci. 2019, 9, 3031 8 of 12

3.4. Join Quality in Term of Energy Relationship between supplied energy and join quality was investigated. The energy, E, is described by laser power, P, and volume of molten material, V, as E = P/V. Other studies [28,29] calculated energy as follows: P E = (1) 1 v.h.t Here, v, h, and t are laser scanning speed, hatch spacing, and layer thickness, respectively. However, Equation (1) can present a rough estimation for comparable process parameters [30]. Meanwhile, by assuming that the melt tracks are semi-circular, energy is calculated as follows [31]:

P E2 = (2) 1 2 h2 v. 2 π t + 4

Figure 10a,b illustrates a comparison between experimental observation and energy calculated by E1, E2 respectively. The orange color represents a flaked area, whereas the green color represents a non-flaked area. In this case, laser power was kept at 180 W and layer thickness was fixed at 20 µm. The results indicate that it is easier to obtain a good joining when laser velocity and hatch distance are increased, as mentioned in Figure9. This can be explained by a decrease in the reaction layer when laser scanning speed and hatch space are increased, or energy is reduced. The diagonal stripes in Figure 10 indicate energy calculated by the formulas. The diagonal striped areas cover not only the non-flaked area, but also broadly cover the flaked space. The more covering of the flaked area, the less the accuracy of the formula. Figure 10a,b indicate that the energy formulas E1 E2 cannot be used to describe the flake behavior and become less accurate as the flaked area increases. Moreover, in the SLM process, hatch overlap (d/h) was investigated as an important parameter [32] (here, d is laser spot size). Figure 11 illustrated the effect of the hatch overlap. The current line track nth of a laser generates a post-heating zone during printing, which reheats the solidified material of the last scan line n 1th, as well as a pre-heating zone, which affects the melting process of the − next scanning line n + 1th. Finally, hatch overlap affects melt geometry, heat-affected zone, formation of a solidification microstructure, and residual stress. In this analysis, the hatch overlap parameter was considered to affect energy. Combined with the assumption that the melt track is semi-circular, the following formula was generated:

P E3 = (3) 1 2 h2 d v. 2 π t + 4 h

Figure 10c illustrates the ﬁt of the E3 formula calculation and experimental area. Although there is an existing small gap between the calculated area and experimental result, the E3 formula could be applied to determine the relationships between joining quality and process parameters in terms of energy. Appl. Sci. 2019, 9, 3031 9 of 12 Appl. Sci. 2019, 9, x FOR PEER REVIEW 9 of 12

FigureFigure 10.10. ComparisonComparison betweenbetween experimentalexperimental resultsresults andand energyenergy calculatedcalculated byby ((aa)) E𝐸11,(, b(b))E 𝐸22,, andand ((cc)) E𝐸3 formulas. Appl. Sci. 2019, 9, 3031 10 of 12 Appl. Sci. 2019, 9, x FOR PEER REVIEW 10 of 12

Figure 11. SchematicSchematic representation representation of the hatch overlap effect. eﬀect.

4. Conclusions Conclusions Joining of pure aluminum and stainless steel by the SLM printing process was analyzed.analyzed. In the SLM method, the flake flake behavior could be represented by Al-Fe join strength. Process parameters such as laser power, laser scanning speed, and hatch dist distanceance significantly significantly affect affect the final final hybrid-material join quality. The results indicate a high-risk area between the 11th and 25th layers, where we should be most concerned about flaking flaking behavior. After After the the 25th 25th layer, layer, bimetallic joining joining can can be be guaranteed. guaranteed. The intermetallicintermetallic layer layer plays plays a significanta significant role role in the in combinationthe combination of hybrid of hybrid materials materials by the SLMby the process. SLM Theprocess. results The showed results theshowed Fe2Al the5 phase Fe2Al existing5 phase inexisting the intermetallic in the intermetallic layer of the layer flaked of the samples. flaked Thatsamples. was Thatagreed was with agreed other with joining other methods. joining The methods. research The figured research out thefigured dependence out the ofdependence the flake behavior of the flake and behaviorthe applied and energy: the applied the increasing energy: the energy increasing leads to energy a sooner leads disruption to a sooner the join.disruption The research the join. found The researchout that thefound hatch out overlap that the a ffhatchected overlap to Al-Fe affected join strength. to Al-Fe Based join onstrength. investigating Based on the investigating hatch overlap the as = hatchan effective overlap factor, as an the effective formula factor,E3 theP/V formula3 can be 𝐸 used =𝑃 to⁄ estimate 𝑉 can be theused approximate to estimate valuesthe approximate of process parametersvalues of process to achieve parameters a good to join achieve quality a ofgood aluminum join quality and of stainless aluminum steel. and stainless steel.

Author Contributions: D.-S.N. and H.-S.P. conceived the idea of this study. D.-S.N. developed, validated, Author Contributions: D.-S.N. and H.-S.P. conceived the idea of this study. D.-S.N. developed, validated, and and performed the experimental works. D.-S.N. wrote the manuscript and H.-S.P. and C.-M.L. revised it. Allperformed authors the discussed experimental the results works. and D.-S.N. commented wrote onthe the ma manuscript.nuscript and H.-S.P. and C.-M.L. revised it. All authors discussed the results and commented on the manuscript. Funding: This work was supported by the ICT R&D program of MSIP/IITP [B0101-14-1081, Development of Funding:ICT-based This software work platforms was supported and service by the technologies ICT R&D prog for medicalram of 3DMSIP/IITP printing [B0101-14-1081, application]. Development of ICT-basedConﬂicts of software Interest: platformsThe authors and declareservice notechno conﬂictslogies of for interest. medical 3D printing application]. Conflicts of Interest: The authors declare no conflicts of interest. References References 1. Liyanage, T.; Kilbourne, J.; Gerlich, A.P.; North, T.H. Joint Formation in Dissimilar Al Alloy/Steel and Mg 1. Liyanage,Alloy/Steel T.; Friction Kilbourne, Stir Spot J.; Gerlich, Welds. A.P.;Sci. Technol.North, T.H. Weld. Joint Join. Formation2009, 14, 500–508. in Dissimilar [CrossRef Al Alloy/Steel] and Mg 2. Alloy/SteelTaban, E.; Gould, Friction J.E.; Stir Lippold, Spot Welds. J.C. Dissimilar Sci. Technol. Friction Weld. WeldingJoin. 2009 of, 14 6061-T6, 500–508. Aluminum and AISI 1018 Steel: 2. Taban,Properties E.; Gould, and Microstructural J.E.; Lippold, J. Characterization.C. Dissimilar FrictionMater. Welding Des. 2010 of, 6061-T631, 2305–2311. Aluminum [CrossRef and AISI] 1018 Steel: 3. PropertiesLi, J.; Li, H.; and Huang, Microstruc C.; Xiang,tural T.; Characterization. Ni, Y.; Wei, H. Welding Mater. Des. Process 2010 Characteristics, 31, 2305–2311. of Pulse on Pulse MIG Arc 3. Li,Brazing J.; Li, ofH.; Aluminum Huang, C.; Alloy Xiang, to StainlessT.; Ni, Y.; Steel. Wei, Int.H. Welding J. Adv. Manuf. Process Technol. Characteristics2017, 91, 1057–1067.of Pulse on [ CrossRefPulse MIG] 4. ArcArivazhagan, Brazing of N.;Aluminum Singh, S.; Alloy Prakash, to Stainless S.; Reddy, Steel. G.M. Int. InvestigationJ. Adv. Manuf. on Technol. AISI304 2017 Austenitic, 91, 1057–1067. Stainless Steel 4. Arivazhagan,to AISI 4140 Low N.; Singh, Alloy SteelS.; Prakash, Dissimilar S.; Reddy, Joints byG.M. Gas Investigation Tungsten Arc, on ElectronAISI 304 BeamAustenitic and FrictionStainless Welding. Steel to AISIMater. 4140 Des. Low2011 Alloy, 32, 3036–3050. Steel Dissimilar [CrossRef Joints] by Gas Tungsten Arc, Electron Beam and Friction Welding. 5. Mater.Pardal, Des. G.; Meco,2011, 32 S.;, 3036–3050. Ganguly, S.; Williams, S.; Prangnell, P. Dissimilar Metal Laser Spot Joining of Steel to 5. Pardal,Aluminium G.; Meco, in Conduction S.; Ganguly, Mode. S.; Williams,Int. J. Adv. S.; Manuf. Prangne Technol.ll, P. Dissimilar2014, 73, 365–373.Metal Laser [CrossRef Spot Joining] of Steel to Aluminium in Conduction Mode. Int. J. Adv. Manuf. Technol. 2014, 73, 365–373. 6. Yang, J.; Chen, J.; Zhao, W.; Zhang, P.; Yu, Z.; Li, Y.; Zeng, Z.; Zhou, N. Diode Laser Welding/Brazing of Aluminum Alloy to Steel Using a Nickel Coating. Appl. Sci. 2018, 8, 922. Appl. Sci. 2019, 9, 3031 11 of 12

6. Yang, J.; Chen, J.; Zhao, W.; Zhang, P.; Yu, Z.; Li, Y.; Zeng, Z.; Zhou, N. Diode Laser Welding/Brazing of Aluminum Alloy to Steel Using a Nickel Coating. Appl. Sci. 2018, 8, 922. [CrossRef] 7. Mehrpouya, M.; Gisario, A.; Barletta, M.; Natali, S.; Veniali, F. Dissimilar Laser Welding of NiTi Wires. Lasers Manuf. Mater. Process. 2019, 6, 99–112. [CrossRef] 8. Tomashchuk, I.; Grevey, D.; Sallamand, P. Dissimilar Laser Welding of AISI 316L Stainless Steel to Ti6-Al4-6V Alloy via Pure Vanadium Interlayer. Mater. Sci. Eng. A 2015, 622, 37–45. [CrossRef] 9. Mehrpouya, M.; Gisario, A.; Broggiato, G.B.; Puopolo, M.; Vesco, S.; Barletta, M. Effect of Welding Parameters on Functionality of Dissimilar Laser-Welded NiTi Superelastic (SE) to Shape Memory Effect (SME) Wires. Int. J. Adv. Manuf. Technol. 2019, 103, 1593–1601. [CrossRef] 10. Tanaka, Y.; Kajihara, M. Kinetics of Isothermal Reactive Diffusion between Solid Fe and Liquid Al. J. Mater. Sci. 2010, 45, 5676–5684. [CrossRef] 11. Park, H.S.; Nguyen, D.S. Study on Flaking Behavior in Selective Laser Melting Process. Procedia CIRP 2017, 63, 569–572. [CrossRef] 12. Santos, L.; de Jesus, J.; Ferreira, J.; Costa, J.; Capela, C. Fracture Toughness of Hybrid Components with Selective Laser Melting 18Ni300 Steel Parts. Appl. Sci. 2018, 8, 1879. [CrossRef] 13. Kruth, J.P.; Leu, M.C.; Nakagawa, T. Progress in Additive Manufacturing and Rapid Prototyping. CIRP Ann. Manuf. Technol. 1998, 47, 525–540. [CrossRef] 14. Demir, A.G.; Previtali, B. Multi-Material Selective Laser Melting of Fe/Al-12Si Components. Manuf. Lett. 2017, 11, 8–11. [CrossRef] 15. Shishkovsky, I.; Missemer, F.; Kakovkina, N.; Smurov, I. Intermetallics Synthesis in the Fe–Al System via Layer by Layer 3D Laser Cladding. Crystals 2013, 3, 517–529. [CrossRef] 16. Kempen, K.; Vrancken, B.; Buls, S.; Thijs, L.; Van Humbeeck, J.; Kruth, J.P. Selective Laser Melting of Crack-Free High Density M2 High Speed Steel Parts by Baseplate Preheating. J. Manuf. Sci. Eng. 2014, 136, 061026. [CrossRef] 17. Zhang, M.; Liu, C.; Shi, X.; Chen, X.; Chen, C.; Zuo, J.; Lu, J.; Ma, S. Residual Stress, Defects and Grain Morphology of Ti-6Al-4V Alloy Produced by Ultrasonic Impact Treatment Assisted Selective Laser Melting. Appl. Sci. 2016, 6, 304. [CrossRef] 18. Tanaka, Y.; Kajihara, M. Morphology of Compounds Formed by Isothermal Reactive Diffusion between Solid Fe and Liquid Al. Mater. Trans. 2009, 50, 2212–2220. [CrossRef] 19. Peyre, P.; Sierra, G.; Deschaux-Beaume, F.; Stuart, D.; Fras, G. Generation of Aluminium-Steel Joints with Laser-Induced Reactive Wetting. Mater. Sci. Eng. A 2007, 444, 327–338. [CrossRef] 20. Eggeler, G.; Auer, W.; Kaesche, H. On the Influence of Silicon on the Growth of the Alloy Layer during Hot Dip Aluminizing. J. Mater. Sci. 1986, 21, 3348–3350. [CrossRef] 21. Pasche, G.; Scheel, M.; Schäublin, R.; Hébert, C.; Rappaz, M.; Hessler-Wyser, A. Time-Resolved X-ray Microtomography Observation of Intermetallic Formation between Solid Fe and Liquid Al. Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 2013, 44, 4119–4123. [CrossRef] 22. Verhaeghe, F.; Craeghs, T.; Heulens, J.; Pandelaers, L. A Pragmatic Model for Selective Laser Melting with Evaporation. Acta Mater. 2009, 57, 6006–6012. [CrossRef] 23. Matthews, M.J.; Guss, G.; Khairallah, S.A.; Rubenchik, A.M.; Depond, P.J.; King, W.E. Denudation of Metal Powder Layers in Laser Powder Bed Fusion Processes. Acta Mater. 2016, 114, 33–42. [CrossRef] 24. Achar, D.R.G.; Ruge, J.; Sundaresan, S. Metallugical and Mechanical Investigations of Aluminium-Steel Fusion Welds (I). Alum. Dusseld. 1980, 56, 391–397. 25. Rathod, M.J.; Kutsuna, M. Joining of Aluminum Alloy 5052 and Low-Carbon Steel by Laser Roll Welding. Weld. Res. 2004, 83, 16S–26S. 26. Borrisutthekul, R.; Yachi, T.; Miyashita, Y.; Mutoh, Y. Suppression of Intermetallic Reaction Layer Formation by Controlling Heat Flow in Dissimilar Joining of Steel and Aluminum Alloy. Mater. Sci. Eng. A 2007, 467, 108–113l. [CrossRef] 27. Hu, Y.; Xu, X.; Yao, Z.; Hu, J. Laser Peen Forming Induced Two Way Bending of Thin Sheet Metals and Its Mechanisms. J. Appl. Phys. 2010, 108, 073117. [CrossRef] 28. Shah, L.H.; Ishak, M. Review of Research Progress on Aluminum-Steel Dissimilar Welding. Mater. Manuf. Process. 2014, 29, 928–933. [CrossRef] 29. Yusuf, S.M.; Gao, N. Influence of Energy Density on Metallurgy and Properties in Metal Additive Manufacturing. Mater. Sci. Technol. 2017, 33, 1269–1289. [CrossRef] Appl. Sci. 2019, 9, 3031 12 of 12

30. Hagedorn, Y. Additive Manufacturing of High Performance Oxide Ceramics via Selective Laser Melting; RWTH Aachen University: Berlin, Germany, 2013. 31. Yap, C.Y.; Chua, C.K.; Dong, Z.L. An Eﬀective Analytical Model of Selective Laser Melting. Virtual Phys. Prototyp. 2016, 11, 21–26. [CrossRef] 32. Criales, L.E.; Arısoy, Y.M.; Özel, T. A Sensitivity Analysis Study on the Material Properties and Process Parameters for Selective Laser Melting of Inconel 625. Am. Soc. Mech. Eng. 2015, 86, V001T02A062.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).