Microstructure and Selective Corrosion of Alloy 625 Obtained by Means of Laser Powder Bed Fusion

materials

Article Microstructure and Selective Corrosion of Alloy 625 Obtained by Means of Laser Powder Bed Fusion

Marina Cabrini 1,2,* , Sergio Lorenzi 1,2 , Cristian Testa 1,2 , Fabio Brevi 3, Sara Biamino 4 , Paolo Fino 4, Diego Manfredi 5 , Giulio Marchese 4 , Flaviana Calignano 6 and Tommaso Pastore 1,2 1 Consorzio INSTM, via G. Giusti, 9, 50121 Firenze, Italy; [email protected] (S.L.); [email protected] (C.T.); [email protected] (T.P.) 2 University of Bergamo, Department of Engineering and Applied Sciences, Viale Marconi 5, 24044 Dalmine, Italy 3 OMB Valves SpA, Via Europa, 7, 24069 Cenate Sotto, Italy; [email protected] 4 Department of Applied Science and Technology, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy; [email protected] (S.B.); paolo.ﬁ[email protected] (P.F.); [email protected] (G.M.) 5 Center for Sustainable Futures Technologies-CSFT@POLITO, Istituto Italiano di Tecnologia, Via Livorno 60, 10144 Torino, Italy; [email protected] 6 Department of Management and Production Engineering, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy; ﬂ[email protected] * Correspondence: [email protected]; Tel.: +390352052318

Received: 8 April 2019; Accepted: 27 May 2019; Published: 29 May 2019

Abstract: The effect of microstructure on the susceptibility to selective corrosion of Alloy 625 produced by laser powder bed fusion (LPBF) process was investigated through intergranular corrosion tests according to ASTM G28 standard. The effect of heat treatment on selective corrosion susceptibility was also evaluated. The behavior was compared to commercial hot-worked, heat treated Grade 1 Alloy 625. The morphology of attack after boiling ferric sulfate-sulfuric acid test according to ASTM G28 standard is less penetrating for LPBF 625 alloy compared to hot-worked and heat-treated alloy both in as-built condition and after heat treatment. The different attack morphology can be ascribed to the oversaturation of the alloying elements in the nickel austenitic matrix obtained due to the very high cooling rate. On as-built specimens, a shallow selective attack of the border of the melt pools was observed, which disappeared after the heat treatment. The results confirmed similar intergranular corrosion susceptibility, but different corrosion morphologies were detected. The results are discussed in relation to the unique microstructures of LPBF manufactured alloys.

Keywords: additive manufacturing; corrosion; alloy 625; selective corrosion; oil and gas; materials qualiﬁcation

1. Introduction Alloy 625 (UNS N06625) is one of the most used material in chemical and oil and gas industries for its corrosion resistance in either oxidant or non-oxidant acids. It is also used in the aerospace industry thanks to the resistance to thermal oxidation [1,2]. In the oil and gas industry, high aggressiveness of the environment requires the use of materials with outstanding corrosion resistance to ensure the reliability of assets, reduce the risk of breakage, prevent consequences on the environment and guarantee the safety of people. The continuous growth of demand and the depletion of reservoir lead the oil and natural gas extraction industry to exploit deeper wells, with extreme conditions (High Pressure High Temperature Wells—HPHT). The use of traditional alloys with low both mechanical and corrosion resistance is not suitable under such hypothesis and international standards i.e., NACE MR 0175/ISO

Materials 2019, 12, 1742; doi:10.3390/ma12111742 www.mdpi.com/journal/materials Materials 2019, 12, 1742 2 of 11

15156 defines such environment as “sour” even in presence of few ppm of H2S in the formation gas due to the high operating pressures. Materials resistant to sulfide stress corrosion cracking (SSCC) are required in these environments and high temperatures, high fraction of water in the formation and presence of CO2 and H2S implies the use of general and localized corrosion resistant alloys (Corrosion Resistance Alloys—CRA). Inconel 625® (Alloy 625) is a nickel-based superalloy strengthened by solid-solution hardening of Nb and Mo in a Ni-Cr matrix [3]. Alloy 625 has a combination of high yield strength, fatigue strength, and excellent corrosion resistance in aggressive environments, it has found widespread applications in the aerospace, marine, and nuclear industries where complex shapes are often required. The alloy shows general corrosion rates less than 0.5 mm/year in concentrated non-oxidizing organic and inorganic acids [4,5]. The resistance to pitting is very high, much higher than the traditional austenitic stainless steels. Thanks to the high nickel content, it is immune to chloride stress corrosion (SCC, Stress Corrosion Cracking). Alloy 625 is precipitation-hardening alloy, due to the formation of very fine phases, which strengthen the austenitic matrix, because of aging treatments between 550–750 ◦C. The formation of these phases or their decomposition modifies both the yield strength and ductility of the alloy [6]. Furthermore, the formation of carbides and secondary phases can also affect the corrosion resistance. In particular, the alloy can become susceptible to intergranular corrosion if subjected to improper solubilization treatment [7]. In the last years, a very promising additive manufacturing (AM) technique has been developed in order to find alternative processes based on Laser Powder Bed Fusion (LPBF) of metal powders—layer by layer—by using a laser following the 3D CAD model. This production method has advantages in terms of cost reduction and lead time, because it eliminates swarf—typical of traditional subtractive machining—welding operations, and assembly phases. Lot of efforts has been devoted to AM of Alloy 625 alloy [8–14]. Alloy 625 is sensitive to precipitation of intermetallic phases such as Ni3M phases, Ni2(Cr, Mo) Laves phase as well as MC primary carbides and M6C and M23C6 secondary carbides. Large amount of literature exists on the heat-treatment-induced phase evolution of wrought Alloy 625 [3,15,16]. The effect of post-weld heat treatment has been also studied by several groups [17–19]. As AM techniques are processes characterized by melting and solidification, microstructures are expected to be different due to different thermal gradients. Precipitation of phases that occurs in AM Alloy 625 are not present at all in wrought material or they only appear after tens to hundreds of hours [2]. Despite huge amount of data and scientific papers regarding the mechanical, physical and microstructural characterization of AM of nickel-based alloys, on the authors knowledge there are no works on the corrosion behavior of the alloy obtained by means of LPBF. On the contrary several works underline the effect of the different microstructures obtained by means of LPBF on the corrosion behavior of Ti6Vl4V alloy [20–22], on AISI 316L [23–25], high strength steels [26] on Cr/Co alloys [27,28] and AlSi10Mg [29–34]. The relation between microstructure and corrosion resistance is of fundamental importance for the qualification of materials selected for hostile environments. The paper is devoted to the study of susceptibility to intergranular corrosion of Alloy 625 produced by LPBF in as-built condition and after typical heat treatment usually recommended for oil and gas service. The behavior of LPBF alloy has been compared to hot-worked, heat treated commercial Alloy 625.

2. Materials and Methods Commercial gas atomized Alloy 625 powder was used for LPBF process and hot-worked bar with 16 mm diameter was considered for comparison purposes. The bar was furnished in annealed condition, after heat treatment at 980 ◦C for 32 min and water quenching (Grade 1, according to ASTM B446 standard). The chemical composition of LPBF and hot-worked alloys as well as microstructures are shown in Figure1. The microstructure was revealed by grinding with silicon carbides emery papers up to 4000 grit, polishing with diamond paste up to 1 µm, and then etching with Kalling’s N◦2 reagent or electrolytic oxalic acid according to ASTM A262—Practice A. As-built LPBF specimens Materials 2019, 12, 1742 3 of 11 Materials 2019, 12, x FOR PEER REVIEW 3 of 11 werewere etched etched by by Mixed Mixed Acid Acid Solution Solution (15 (15 mL mL HCl, HCl, 10 10 mL mL H H3COOH3COOH and and 10 10 mL mL HNO HNO33,, CarloCarlo ErbaErba RPARPA reagents,reagents, Cornaredo, Cornaredo, Milan, Milan, Italy) Italy) (Figure (Figure1 ).1). The The microstructures microstructures of of LPBF LPBF specimens specimens were were observed observed onon planes planes parallel parallel or or perpendicular perpendicular to to the the building building direction, direction, whereas whereas hot-worked hot-worked specimens specimens were were onlyonly examined examined on on transverse transverse plane. plane.

[% wt.] C Si Mn P S Cr Mo Ni ** Nb Ti Al Co Ta Fe Nb+Ta Hot worked * 0.036 0.25 0.19 0.007 0.001021.6 8.26 61.92 3.66 0.243 0.199 0.020 0.010 3.11 3.67 LPBF 0.010 0.08 0.03 <0.001 0.002 22.4 8.2 64.59 3.73 0.18 <0.01 0.17 0.13 0.45 3.86 * Heat composition ** Balance

Hot worked – rod center Hot worked – rod border

Figure 1. Chemical composition and 3D optical image composite of as-built IN625 sample showing theFigure melt 1. pool Chemical contours composition (MPCs) along and the 3D building optical image direction composite (z-axis) andof as-built perpendicular IN625 sample to the buildingshowing directionthe melt (x-y)pool contours plane; mixed (MPCs) acids along reagent the building was used. direction (z-axis) and perpendicular to the building direction (x-y) plane; mixed acids reagent was used. 2.1. Specimens 2.1. Specimens Cubic specimens of 15 mm side were printed by means of an EOSINT M270 (EOS Gmbh, Krailling, Germany)Cubic Dual specimens Mode version of 15 mm working side underwere argonprinted atmosphere. by means of The an device EOSINT operates M270 at (EOS 195 W Gmbh, laser power,Krailling, 1200 Germany) mm/s scan Dual speed, Mode 0.09 version mm hatching working distance under asargon well atmosphere. as 0.02 mm ofThe layer device thickness. operates The at scanning195 W laser strategy power, involves 1200 mm/s the rotation scan speed, of laser 0.09 beam mm of hatching 67◦ between distance consecutive as well layersas 0.02 based mm of on layer the EOSthickness. strategy. The The scanning combination strategy of involves the process the parameters rotation of andlaser scanning beam of strategy67° between allows consecutive the production layers ofbased parts on with the EOS a residual strategy. porosity The combination less than 0.1%. of the Moreprocess in-depth parameters information and scanning about strategy optimization allows procedurethe production and building of parts parameters with a residual were previously porosity less published than 0.1%. [14,35 More]. Specimens in-depth in information as-built condition about andoptimization after post-processing procedure and annealing building heat parameters treatment were at 980previously◦C for 32published min were [14,35]. considered. Specimens Five in millimetersas-built condition high cylindrical and after specimens post-processing were obtained annealing by cuttingheat treatment the hot-worked at 980 bar.°C for 32 min were considered. Five millimeters high cylindrical specimens were obtained by cutting the hot-worked 2.2.bar. Corrosion Tests Boiling ferric sulfate/sulfuric acid corrosion tests were carried out according to ASTM G28 standard, method2.2. Corrosion A. Before Tests testing, the surface of specimens was grinded up to 1200 grit by silicon carbides emeryBoiling paper. ferric Two faces sulfate/sulfuric of LPBF specimens acid corrosion were ground tests were and polished carried upout to according 1 µm diamond to ASTM paste G28 in orderstandard, to permit method observations A. Before testing, of corrosion the surface morphology of specimens on main was planes, grinded i.e., up XY to plane 1200and grit XZby planesilicon parallelcarbides and emery perpendicular paper. Two to faces the buildingof LPBF sp plane,ecimens respectively. were ground One and face polished was polished up to 1 for μm observing diamond morphologypaste in order on to plane permit transversal observations to rolling of corrosi directionon morphology on disk specimens. on main planes, Afterwards, i.e., XY the plane specimens and XZ wereplane degreased parallel and in acetone perpendicular in ultrasonic to the bath, building rinsed inplane, water respectively. and dried before One theirface immersionwas polished in thefor boilingobserving test morphology solution. Before on andplane after transversal tests, the to specimens rolling direction were weighed on disk by specimens. means of analyticAfterwards, scale the to measurespecimens the were weight degreased loss and corrosionin acetone rate. in ultrasonic bath, rinsed in water and dried before their immersion in the boiling test solution. Before and after tests, the specimens were weighed by means 3. Results and Discussion of analytic scale to measure the weight loss and corrosion rate. Table1 shows the results of 120 h weight loss tests in boiling ferric sulfate /sulfuric acid solutions (ASTM3. Results G28 and Method Discussion A). The highest corrosion rate was observed on hot-worked alloys. The LPBF specimensTable showed 1 shows about the results the half of and120 h the weight third loss the corrosiontests in boiling rate measured ferric sulfate/sulfuric on hot-worked acid specimens solutions (ASTM G28 Method A). The highest corrosion rate was observed on hot-worked alloys. The LPBF specimens showed about the half and the third the corrosion rate measured on hot-worked specimens

Materials 2019, 12, x FOR PEER REVIEW 4 of 11 in as-built conditions and after heat treatment at 980 °C for 32 min, respectively. The corrosion morphology of these specimens is flat and even. On the contrary, hot-worked alloy showed penetratingMaterials 2019, 12corrosion, 1742 attack, more severe compared to LPBF specimens. 4 of 11

Table 1. Results of 120 h weight loss test in boiling ferric sulphate/sulfuric acid solution (ASTM G28). in as-built conditions and after heat treatment at 980 ◦C for 32 min, respectively. The corrosion Exposed Initial Final Mass Corrosion morphologyMetallurgical of these specimens is ﬂat and even. On the contrary, hot-worked alloy showed penetrating Specimens Surface Mass * Mass * Variation Rate corrosionCondition attack, more severe compared to LPBF specimens. [cm2] [g] [g] [g] [mdd] Hot worked HW1 6.05 6.73923 6.63074 0.10849 359 Table 1. Results of 120 h weight loss test in boiling ferric sulphate/sulfuric acid solution (ASTM G28). bar/Grade 1 HW2 6.13 7.16643 7.06459 0.10184 332 Metallurgical Exposed Surface Initial Mass * Final Mass * Mass Variation Corrosion Rate annealing Specimens Condition [cm2] [g] [g] [g] [mdd] LPBF Hot worked bar/Grade HW1 6.0513.50 6.7392328.23744 28.12392 6.63074 0.11352 0.10849 168.2 359 NHT1 1LPBF annealing as built HW2 6.13 7.16643 7.06459 0.10184 332 LPBF NHT1LPBF 13.50 28.23744 28.12392 0.11352 168.2 LPBF as built 13.35 27.03907 26.92816 0.11091 166.2 LPBF NHT2NHT2 13.35 27.03907 26.92816 0.11091 166.2 LPBF HT1 13.06 28.24230 28.16770 0.0746 114.2 LPBF/GradeLPBF 1 annealing/ Grade 1 LPBF HT1 13.06 28.24230 28.16770 0.0746 114.2 annealing LPBFLPBF HT2 HT2 13.12 13.12 28.34171 28.34171 28.26959 28.26959 0.07212 0.07212 109.9 109.9 * Average* Average of 33 measurements.measurements.

Figures2 and3 show the corrosion morphology of the specimens after 120 h exposure tests in Figures 2 and 3 show the corrosion morphology of the specimens after 120 h exposure tests in boiling ferric sulphate/sulfuric acid solution according to ASTM G28 standard. The building direction boiling ferric sulphate/sulfuric acid solution according to ASTM G28 standard. The building direction of LPBF specimens and the rolling direction of the hot-worked bar is represented by the yellow of LPBF specimens and the rolling direction of the hot-worked bar is represented by the yellow arrows. Figure2 shows the images of the metallographic sections of the alloys taken at low and high arrows. Figure 2 shows the images of the metallographic sections of the alloys taken at low and high magniﬁcation. HW specimens showed penetrating intergranular attack mainly located at the center of magnification. HW specimens showed penetrating intergranular attack mainly located at the center the specimen (Figure2a,d). of the specimen (Figure 2a,d).

Building direction Building direction

Rolling direction (a) (b) (c)

Building direction Building direction

Melt pool borders

Rolling direction (d) (e) (f)

Figure 2. SEMSEM images of Alloy 625 afterafter intergranularintergranular corrosioncorrosion tests.tests. Metallographic section of commercial rolled alloy along the rolling direction (a) detail of the attack on the the surface surface perpendicular perpendicular to the rolling direction (d). Metallographic section section of of LPBF alloys along the building direction in as produced conditions (b,,c)) andand afterafter heatheat treatmenttreatment ((ee,,ff).).

The intergranular attack produced also grain dropping, as evidenced by the analysis of the surface after exposure (Figure 3a). The high value of the corrosion rate is practically due to grain dropping. The morphology of the corrosion attack on LPBF as-built is far more even (Figure 2b,e), but the attack follows different paths. The morphological analysis emphasizes the unique melt pool microstructure-typical of LPBF processing—formed by overlapped laser scan traces (Figure 3b). The

Materials 2019, 12, x FOR PEER REVIEW 5 of 11 melt pool structure is well-defined and penetrating attacks along the borders can be clearly detected. Finally, the attack takes place in form of shallow attacks on the heat treated LPBF specimens (Figure 2c,f)Materials and2019 the, 12 melt, 1742 pool borders are no more evident. Equiaxed grain microstructure is evidenced5 of by 11 the corrosion attack (Figure 3c).

Materials 2019, 12, x FOR PEER REVIEW 5 of 11

melt pool structure is well-defined and penetrating attacks along the borders can be clearly detected. Finally, the attack takes place in form of shallow attacks on the heat treated LPBF specimens (Figure 2c,f) and the meltRollin poolg bordersdirection are no more evidenBuildint. Equiaxedg direction grain microstructureBuildin isg evidenceddirection by the corrosion attack (Figure 3c). (a) (b) (c)

Figure 3.3. SEMSEM images images of of the the polished polished surface surface of the of specimens the specimens after the after tests the of intergranulartests of intergranular corrosion. corrosion.(a) hot-worked (a) hot-worked bar, (b) LPBF bar, as-built, (b) LPBF (c )as-built, LPBF heat (c) treated.LPBF heat treated.

The differences intergranular in the attack morphology produced of also attack grain is stri dropping,ctly related as evidencedto the microstructure by the analysis of the ofalloy, the whichsurface depends after exposure upon the (Figure manufacturing3a). The high process value an ofd the corrosionheat treatments. rate is practicallyHW alloy shows due to a graindeep intergranulardropping. The attack morphology very close of to the the corrosion axis of the attack bar, where on LPBF strong as-built segregatio is far moren of several even (Figure precipitates2b,e), atbut the the grain attack boundary followsRollin digoccurred. ffdirectionerent paths. This Theis well morphological eviBuildindencedg direction by analysis the presence emphasizes of certain theBuildin unique amountg direction melt of both pool macromicrostructure-typical and micro-precipitates of LPBF (Figure processing—formed 4). The EDS anal byysis overlapped carried out laser on these scan precipitates traces (Figure confirmed3b). The themelt presence pool structure of Nb-rich(a) is well-defined phases, mainly and MC penetrating carbides.(b) attacks along the borders can be(c) clearly detected. Finally,Figure the attack3. SEM takes images place of in the form polished of shallow surface attacks of the on specimens the heat treated after the LPBF tests specimens of intergranular (Figure 2c,f) and thecorrosion. melt pool (a) hot-worked borders are bar, no (b more) LPBF evident. as-built, Equiaxed(c) LPBF heat grain treated. microstructure is evidenced by the corrosionc attack (Figure3c). a TheThe didifferencesfferences inin thethe morphologymorphology ofof attackattack is strictlystrictly related to the microstructure of the alloy, whichwhich dependsdepends uponupon thethe manufacturingmanufacturing processprocess andand thethe heatheat treatments.treatments. HWHW alloyalloy showsshows aa deepdeep intergranularintergranular attack attack very very close close to to the the axis axis of of the the bar, bar, where where strong strong segregation segregatio ofn of several several precipitates precipitates at theat the grain grain boundary boundary occurred. occurred. This This is well is well evidenced evidenced by the by presence the presence of certain of certain amountSpectrum amount of both of macro both andmacro micro-precipitates and micro-precipitates (Figure (Figure4). The 4). EDS The analysis EDS anal carriedysis carried out on out these60 on μm these precipitates precipitates confirmed confirmed the presencethe presence of Nb-rich of Nb-rich phases, phases, mainly mainly MC carbides.MC carbides. b c a

60 μm Spectrum Figure 4. Microstructure of HW specimens (a,b) (Kalling’s attack) and60 EDSμm spectrum of macro- precipitates (c) taken in the point evidenced in (a). b Owing Kim and Lee, the formation of carbides can be a factor contributing to the degradation of corrosion resistance of Alloy 625, because intergranular areas of carbides are thermodynamically more unstable and react more readily than other intergranular ones [36]. Cortial et al. studied the effect of heat treatment on the corrosion resistance of forged and welded nickel superalloys [37,38]. They concluded that hardening between 650 and 800 °C results in dislocation60 μm gliding of matrix, γ” precipitation and maximum of intergranular precipitation of fine M23C6 carbides. Tensile strength and intergranularFigureFigure 4. MicrostructureMicrostructure corrosion becomeof ofHW HW specimensmaxima, specimens impact (a,b)( a(Kalling’s, bstrength) (Kalling’s attack) and attack)reduction and EDS and ofspectrum EDS area spectrumbecome of macro- minima of in themacro-precipitatesprecipitates field of M 23(cC) taken6 and (c) in takenM the6C intergranularpoint in the evidenced point evidenced precipitation in (a). in (a). between 700 and 950 °C. Furthermore, the carbides have a higher Cr concentration than that in the base metal and they deplete the surrounding matrixOwingOwing in chrome, KimKim and thus Lee, decreasing the formation formation its corrosion of of carbides carbides resistance. can can be be However,a afactor factor contributing contributingthe precipitates to to the thefounded degradation degradation in this of worksofcorrosion corrosion were resistance mainly resistance enriched of of Alloy Alloy in 625,Nb. 625, Thebecause because role ofinterg intergranular Nb richranular precipitates areas areas of in carbides the corrosion areare thermodynamicallythermodynamically of Alloy 625 alloy more unstable and react more readily than other intergranular ones [36]. Cortial et al. studied the more unstable and react more readily than other intergranular ones [36]. Cortial et al. studied the eeffectﬀect ofof heatheat treatmenttreatment onon thethe corrosioncorrosion resistanceresistance ofof forgedforged andand weldedwelded nickelnickel superalloys [[37,38].37,38]. TheyThey concludedconcluded thatthat hardeninghardening betweenbetween 650650 andand 800800 ◦°CC resultsresults in dislocationdislocation gliding of matrix, γ” precipitation and maximum of intergranular precipitation of fine M23C6 carbides. Tensile strength and intergranular corrosion become maxima, impact strength and reduction of area become minima in the field of M23C6 and M6C intergranular precipitation between 700 and 950 °C. Furthermore, the carbides have a higher Cr concentration than that in the base metal and they deplete the surrounding matrix in chrome, thus decreasing its corrosion resistance. However, the precipitates founded in this works were mainly enriched in Nb. The role of Nb rich precipitates in the corrosion of Alloy 625 alloy

Materials 2019, 12, x FOR PEER REVIEW 6 of 11 Materials 2019, 12, 1742 6 of 11 was underlined by Tawancy et al. [39]. Authors made tests in boiling 10% nitric acid on Alloy 626 annealed and aged at different time and temperature. They concluded that Ni3Nb phase formed duringprecipitation aging and is maximumpreferentially of intergranular corroded owing precipitation the lower of fine chromium M23C6 carbides. content Tensile compared strength to andthe surroundingintergranular matrix. corrosion The become intense maxima, intergranular impact corros strengthion attack and reduction observed of on area the become hot-worked minima and in heat the field of M C and M C intergranular precipitation between 700 and 950 C. Furthermore, the carbides treated reference23 6 alloy6 can be then ascribed both to the precipitation of MC◦ carbides and Ni3Nb phase athave the agrain higher boundary Cr concentration which is favored than that at inthe the cent baseer of metal the bar and due they to deplete the production the surrounding process. matrix in chrome,The intergranular thus decreasing attack its corrosionis not so penetrating resistance. However,in as-built the LPBF precipitates Alloy 625 founded alloy, inand this it worksis not weredetectable mainly at all enriched on the inLPBF Nb. alloy The after role ofthe Nb heat rich treatment. precipitates The in microstructure the corrosion of the Alloy as-built 625 alloy samples was obtainedunderlined in bythis Tawancy work shows et al. columnar [39]. Authors grains made (CGs) tests developing in boiling epitaxially, 10% nitric acid thus on crossing Alloy 626several annealed melt poolsand aged along at dithefferent building time direction and temperature. (z-axis) They(Figur concludede 5). The marks that Ni underline3Nb phase the formed different during melt aging pool is borderspreferentially (MPCs) corroded along both owing the thebuil lowerding direction chromium (z-axis) content and compared perpendicular to the surroundingto the building matrix. platform The (x-yintense plane). intergranular The melt corrosionpools are not attack all observedplaced stra onight the along hot-worked the building and heat direction treated due reference to the alloyscan strategycan be then which ascribed involves both laser to the beam precipitation rotation of of 67° MC between carbides two and consecutive Ni3Nb phase layers at the [14,40,41]. grain boundary whichDuring is favored the LPBF at the process, center of the the solidification bar due to the rate production is different process. between the upper and lower zones of theThe melt intergranular pool. In the attacklower part is not of sothe penetrating layer the cooling in as-built speed LPBF is very Alloy high, 625 while alloy, the and upper it is part not coolsdetectable moreat slowly. all on the LPBF alloy after the heat treatment. The microstructure of the as-built samples obtainedFigure in this5a reveals work shows the FESEM columnar (Field grains Emission (CGs) Scanning developing Electron epitaxially, Microscope) thus crossing image several of as-built melt Alloypools along625 samples the building along directionthe building (z-axis) direction. (Figure The5). melt The markspools are underline made up the of di veryfferent fine melt primary pool dendriticborders (MPCs) structures along which both thehave building both cellular direction and (z-axis) columnar and perpendicular shapes, as shown to the in building Figure platform5b. The cellular(x-y plane). dendrites The melt structure pools is are due not to all the placed alteration straight of columnar along the dendrite building structures direction duecaused to theby very scan highstrategy cooling which rates, involves as previously laser beam reported rotation [42]. of 67◦ between two consecutive layers [14,40,41].

Figure 5. FESEMFESEM images images at at different diﬀerent magnification magniﬁcation ( a,,b)) exhibiting melt pools with columnar and cellular primary dendrites for as-built IN625 sample; mixed acids reagent.

AtDuring higher the magnification LPBF process, the(Figure solidification 6), very rate fine is bright different precipitates between the along upper the and dendritic lower zones areas of indicatedthe melt pool. by arrows In the lower1 and partbright of elongated the layer the phases cooling along speed the isinter-dendritic very high, while regions the upper pointed part out cools by arrowsmore slowly. 2 can be observed. Previous TEM investigations performed on as-built Alloy 625 (produced with Figurethe same5a revealsprocess theparameters) FESEM (Field [43] reported Emission a re Scanninglatively high Electron concentration Microscope) of Nb image and ofC for as-built these Alloyprecipitates, 625 samples suggesting along the buildingearly stage direction. formation The of melt fine pools MC are carbides. made up EDS of veryresults fine revealed primary enrichmentdendritic structures in Nb and which Mo havewithin both the cellular inter-dendriti and columnarc areas, shapes,indicated as shownpossible in segregation Figure5b. The of cellularNb and Modendrites due to structure their elevated is due tendency to the alteration to segregat of columnare, as confirmed dendrite by structures the literature caused [44,45]. by very high cooling rates,These as previously precipitates reported did not [42 ].influence the corrosion resistance of the alloys owing to their very smallAt sizes. higher On magnification the other hand, (Figure the attack6), very seems fine bright to be precipitatespreferentially along located the dendritic along the areas border indicated of the meltby arrows pools. 1 and bright elongated phases along the inter-dendritic regions pointed out by arrows 2 can be observed.Dinda et Previousal. hypothesized TEM investigations the precipitation performed of intermetallic on as-built γ Alloy” [Ni3 625Nb] (producedand δ [Ni3Nb] with phases the same in theprocess heat parameters)affected zone [43 at] the reported border a of relatively melt pool high [9], concentrationtaking into account of Nb also and the C for reticular these precipitates, parameters modificationssuggesting the in early the stageNi matrix. formation Anam of et fine al. MC reported carbides. the EDSpresence results ofrevealed γ” [Ni3Nb] enrichment precipitates in Nb at andthe borderMo within of melt the inter-dendritic pool [46]. A areas,similar indicated attack morphology possible segregation was observ of Nbed and also Mo for due aluminum to their elevated alloys [29,32,34].tendency to However, segregate, the as preferential confirmed byattack the literatureat the border [44, 45of ].the melt pool does not penetrate in depth for Alloy 625.

Materials 2019, 12, 1742 7 of 11 Materials 2019, 12, x FOR PEER REVIEW 7 of 11

Figure 6. FESEMFESEM images images of of as-built as-built IN625 IN625 sample sample at at high magnification magniﬁcation ( a) columnar dendritic structures withwith nanoprecipitatesnanoprecipitates (1) (1) and and segregation segregation of of Nb Nb and and Mo Mo within within the the inter-dendritic inter-dendritic zones zones (2); (2);(b) cellular(b) cellular dendritic dendritic structures structures with with nanoprecipitates nanoprecipitates (1) and(1) and segregation segregation of Nb of Nb and and Mo Mo within within the theinter-dendritic inter-dendritic areas areas (2). (2).

TheThese heat precipitates treatment didat 980 not °C influence for 32 min the dissolved corrosion the resistance melt pool of microstructure the alloys owing obtained to their by very the LPBFsmall sizes.process On (Figure the other 7a). hand, The specimens the attack seemspresented to be columnar preferentially grains located along alongthe building the border direction, of the coupledmelt pools. with the formation of fine carbides as displayed in Figure 7b,c. The carbides had dimensions from Dindananometric et al. hypothesizedsize (around 40 the nm) precipitation as well as larger of intermetallic size roughlyγ” [Niaround3Nb] 500 and nm,δ [Ni as3 canNb] be phases seen in Figuresthe heat 7c aff andected 7b, zone respectively. at the border In order of melt to detect pool [ 9a], chemical taking into composition account also variation, the reticular the EDS parameters analysis wasmodifications performed in on the the Ni largest matrix. carbides. Anam et al. reported the presence of γ” [Ni3Nb] precipitates at the borderEDS of meltscan poolline [46(Figure]. A similar 7d) revealed attack morphology that the carb wasides observed are enriched also for in aluminum Nb and alloysdepleted [29, 32of, 34Cr]. suggestingHowever, the the preferential formation attackof MC at carbides, the border which of the is melt in accordance pool does not with penetrate the TTT in (Time depth Temperature for Alloy 625. Transformation)The heat treatment diagram at of 980 Alloy◦C for625 32 [47]. min dissolved the melt pool microstructure obtained by the LPBF process (Figure7a). The specimens presented columnar grains along the building direction, coupled with the formation of fine carbides as displayed in Figure7b,c. The carbides had dimensions from nanometric size (around 40 nm) as well as larger size roughly around 500 nm, as can be seen in Figure7c,b, respectively. In order to detect a chemical composition variation, the EDS analysis was performed on the largest carbides. EDS scan line (Figure7d) revealed that the carbides are enriched in Nb and depleted of Cr suggesting the formation of MC carbides, which is in accordance with the TTT (Time Temperature Transformation) diagram of Alloy 625 [47]. The very low susceptibility to intergranular corrosion of the heat-treated specimens could be attributed to the absence of macroscopic precipitation as the size and distribution of the precipitate are coherent with the rounded pits on the surface of the specimen at the end of the intergranular corrosion test (Figure7a). The precipitation of second phases is not continuous. This promotes a quite even not penetrating corrosion attack that does not penetrate significantly in depth along the grain boundaries, without any dropping. Thus, very low loss of weight was measured due to this fact.

Figure 7. (a) Optical microscopy image of LPBF Alloy 625 annealed at 980 °C for 32 min and water quenched showing columnar grains along the building direction; (b,c) FESEM images of IN625 sample annealed at 980 °C for 32 min and water quenched revealing fine precipitates throughout the material, with the largest located along the grain boundaries; (d) EDS scan line (at higher magnification) on a carbides showing the enrichment in Nb and the depletion of Cr with respect to the austenitic matrix.

Materials 2019, 12, x FOR PEER REVIEW 7 of 11

Figure 6. FESEM images of as-built IN625 sample at high magnification (a) columnar dendritic structures with nanoprecipitates (1) and segregation of Nb and Mo within the inter-dendritic zones (2); (b) cellular dendritic structures with nanoprecipitates (1) and segregation of Nb and Mo within the inter-dendritic areas (2).

The heat treatment at 980 °C for 32 min dissolved the melt pool microstructure obtained by the LPBF process (Figure 7a). The specimens presented columnar grains along the building direction, coupled with the formation of fine carbides as displayed in Figure 7b,c. The carbides had dimensions from nanometric size (around 40 nm) as well as larger size roughly around 500 nm, as can be seen in Figures 7c and 7b, respectively. In order to detect a chemical composition variation, the EDS analysis was performed on the largest carbides. EDS scan line (Figure 7d) revealed that the carbides are enriched in Nb and depleted of Cr suggestingMaterials 2019 ,the12, 1742formation of MC carbides, which is in accordance with the TTT (Time Temperature8 of 11 Transformation) diagram of Alloy 625 [47].

Figure 7. (a) Optical microscopy image of LPBF Alloy 625 annealed at 980 ◦C for 32 min and water Figurequenched 7. ( showinga) Optical columnar microscopy grains image along of the LPBF building Alloy direction;625 annealed (b,c) at FESEM 980 °C images for 32 of min IN625 and sample water

quenchedannealed atshowing 980 ◦C for columnar 32 min and grains water along quenched the building revealing direction; fine precipitates (b,c) FESEM throughout images the of material, IN625 withsample the annealed largest located at 980 °C along for 32 the min grain and boundaries; water quen (chedd) EDS revealing scan line fine (at precipitates higher magnification) throughout on the a material,carbides showingwith the the largest enrichment located in Nbalong and the depletiongrain boundaries; of Cr with respect(d) EDS to thescan austenitic line (at matrix.higher magnification) on a carbides showing the enrichment in Nb and the depletion of Cr with respect to 4. Conclusionsthe austenitic matrix. The susceptibility to selective corrosion behavior of Alloy 625 obtained by means of LPBF was studied. The behavior was compared to commercial hot-worked, heat treated Grade 1 Alloy 625. The morphology of attack after boiling ferric sulfate-sulfuric acid test according to ASTM G28 standard is less penetrating for LPBF 625 alloy compared to hot-worked and heat treated alloy both in as-built condition and after heat treatment at 980 ◦C for 32 min. Furthermore, lower weight loss of LPBF alloy were measured. The different attack morphology can be ascribed to the oversaturation of the alloying elements in the nickel austenitic matrix obtained due to the very high cooling rate. The precipitates of second phases along and inside the micro-dendrites are too small to promote preferential dissolution paths. On as-built specimens, a shallow selective attack of the border of the melt pools was observed. The Grade 1 heat treatment at 980 ◦C for 32 min followed by water quenching caused the melt pools microstructure to disappear and prevent the in-depth penetration of the selective attack.

Author Contributions: Methodology, S.L., C.T., D.M., G.M., F.C.; investigation, S.L., C.T., D.M., G.M.; data curation, S.B., P.F., M.C.; all the authors contributed to writing the original draft of the paper. writing—review & editing T.P., P.F., M.C., S.B.; supervision, T.P., S.B.; project administration, T.P.; funding acquisition, T.P., F.B. Funding: The work was realized with the contribution and reliance on the Collaboration Agreement of INSTM and the Lombardy Region signed on 24th September 2015. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Eiselstein, H.; Tillack, D. The Invention and Deﬁnition of Alloy 625. Superalloys 1991.[CrossRef]

2. Radavich, J.F.; Fort, A. Eﬀects of Long Time Exposure in Alloy 625 at 1200◦F, 1400◦F and 1600◦F. Superalloys 1994, 718, 635–647. Materials 2019, 12, 1742 9 of 11

3. Rai, S.K.; Kumar, A.; Shankar, V.; Jayakumar, T.; Rao, K.B.S.; Raj, B. Characterization of microstructures in Inconel 625 using X-ray diffraction peak broadening and lattice parameter measurements. Scr. Mater. 2004, 51, 59–63. [CrossRef] 4. Kolts, J.; Wu, J.B.C.; Manning, P.E.; Asphahani, A.I. Highly Alloyed Austenitic Material for Corrosion Resistance. Corros. Rev. 1986, 6, 279–326. 5. Metals Handbook, 9th ed.; ASM International: Metals Park, Geauga, OH, USA, 1987. 6. Shankar, V.; Bhanu Sankara Rao, K.; Mannan, S. Microstructure and mechanical properties of Inconel 625 superalloy. J. Nucl. Mater. 2001, 288, 222–232. [CrossRef] 7. Harris, J.A.; Scarberry, R.C. Effect of metallurgical reactions in INCONEL nickel-chromium-molybdenum alloy 625 on corrosion resistance in nitric acid. JOM 1971, 23, 45–49. [CrossRef] 8. Murr, L.E.; Martinez, E.; Amato, K.N.; Gaytan, S.M.; Hernandez, J.; Ramirez, D.A.; Shindo, P.W.; Medina, F.; Wicker, R.B. Fabrication of metal and alloy components by additive manufacturing: Examples of 3D materials science. J. Mater. Res. Technol. 2012, 1, 42–54. [CrossRef] 9. Dinda, G.P.; Dasgupta, A.K.; Mazumder, J. Laser aided direct metal deposition of Inconel 625 superalloy: Microstructural evolution and thermal stability. Mater. Sci. Eng. A 2009, 509, 98–104. [CrossRef] 10. Paul, C.P.; Ganesh, P.; Mishra, S.K.; Bhargava, P.; Negi, J.; Nath, A.K. Investigating laser rapid manufacturing for Inconel-625 components. Opt. Laser Technol. 2007, 39, 800–805. [CrossRef] 11. Witkin, D.B.; Adams, P.; Albright, T. Microstructural evolution and mechanical behavior of nickel-based superalloy 625 made by selective laser melting. In Proceedings of the SPIE, San Francisco, CA, USA, 16 March 2015. 12. Yadroitsev, I.; Gusarov, A.; Yadroitsava, I.; Smurov, I. Single track formation in selective laser melting of metal powders. J. Mater. Process. Technol. 2010, 210, 1624–1631. [CrossRef] 13. Keller, T.; Lindwall, G.; Ghosh, S.; Ma, L.; Lane, B.M.; Zhang, F.; Kattner, U.R.; Lass, E.A.; Heigel, J.C.; Idell, Y.; Williams, M.E.; Allen, A.J.; Guyer, J.E.; Levine, L.E. Application of finite element, phase-field, and CALPHAD-based methods to additive manufacturing of Ni-based superalloys. Acta Mater. 2017, 139, 244–253. [CrossRef] 14. Marchese, G.; Garmendia Colera, X.; Calignano, F.; Lorusso, M.; Biamino, S.; Minetola, P.; Manfredi, D. Characterization and Comparison of Inconel 625 Processed by Selective Laser Melting and Laser Metal Deposition. Adv. Eng. Mater. 2017, 19, 1600635. [CrossRef] 15. Suave, L.M.; Cormier, J.; Villechaise, P.; Soula, A.; Hervier, Z.; Bertheau, D.; Laigo, J. Microstructural evolutions during thermal aging of alloy 625: Impact of temperature and forming process. Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 2014, 45, 2963–2982. [CrossRef] 16. Petrzak, P.; Kowalski, K.; Blicharski, M. Analysis of phase transformations in Inconel 625 alloy during annealing. Acta Phys. Pol. A 2016, 130, 1043–1044. [CrossRef] 17. Dupont, J.N.; Robino, C.V.; Marder, A.R.; Notis, M.R. Solidification of Nb-bearing superalloys: Part II. Pseudoternary solidification surfaces. Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 1998, 29, 2797–2806. [CrossRef] 18. Xu, F.; Lv, Y.; Liu, Y.; Xu, B.; He, P. Effect of heat treatment on microstructure and mechanical properties of Inconel 625 alloy fabricated by pulsed plasma arc deposition. Phys. Procedia 2013, 50, 48–54. [CrossRef] 19. Li, S.; Wei, Q.; Shi, Y.; Zhu, Z.; Zhang, D. Microstructure Characteristics of Inconel 625 Superalloy Manufactured by Selective Laser Melting. J. Mater. Sci. Technol. 2015, 31, 946–952. [CrossRef] 20. Dai, N.; Zhang, L.-C.; Zhang, J.; Chen, Q.; Wu, M. Corrosion behavior of selective laser melted Ti-6Al-4V alloy in NaCl solution. Corros. Sci. 2016, 102, 484–489. [CrossRef] 21. Yang, J.; Yang, H.; Yu, H.; Wang, Z.; Zeng, X. Corrosion Behavior of Additive Manufactured Ti-6Al-4V Alloy in NaCl Solution. Metall. Mater. Trans. A 2017, 48, 3583–3593. [CrossRef] 22. Liu, L.; He, M.; Xu, X.; Zhao, C.; Gan, Y.; Lin, J.; Luo, J.; Lin, J. Preliminary study on the corrosion resistance, antibacterial activity and cytotoxicity of selective-laser-melted Ti6Al4V-xCu alloys. Mater. Sci. Eng. C 2017, 72, 631–640. [CrossRef] 23. Zi˛etala,M.; Durejko, T.; Polański,M.; Kunce, I.; Płociński,T.; Zieliński,W.; Łazińska,M.; St˛epniowski,W.; Czujko, T.; Kurzydłowski, K.J.; et al. The microstructure, mechanical properties and corrosion resistance of 316 L stainless steel fabricated using laser engineered net shaping. Mater. Sci. Eng. A 2016, 677, 1–10. [CrossRef] Materials 2019, 12, 1742 10 of 11

24. Trelewicz, J.R.; Halada, G.P.; Donaldson, O.K.; Manogharan, G. Microstructure and Corrosion Resistance of Laser Additively Manufactured 316L Stainless Steel. JOM 2016, 68, 850–859. [CrossRef] 25. Sander, G.; Thomas, S.; Cruz, V.; Jurg, M.; Birbilis, N.; Gao, X.; Brameld, M.; Hutchinson, C.R. On The Corrosion and Metastable Pitting Characteristics of 316L Stainless Steel Produced by Selective Laser Melting. J. Electrochem. Soc. 2017, 164, C250–C257. [CrossRef] 26. Schmidt, D.P.; Jelis, E.; Clemente, M.P.; Ravindra, N.M. Corrosion of 3D Printed Steel. Available online: https://www.researchgate.net/publication/283346384_CORROSION_OF_3D_PRINTED_STEEL (accessed on 28 May 2019). 27. Tuna, S.H.; Özçiçek Pekmez, N.; Kürkçüo˘glu,I. Corrosion resistance assessment of Co-Cr alloy frameworks fabricated by CAD/CAM milling, laser sintering, and casting methods. J. Prosthet. Dent. 2015, 114, 725–734. [CrossRef] 28. Puskar, T.; Jevremovic, D.; Williams, R.J.; Eggbeer, D.; Vukelic, D.; Budak, I. A comparative analysis of the corrosive effect of artificial saliva of variable pH on DMLS and cast Co-Cr-Mo dental alloy. Materials 2014, 7, 6486–6501. [CrossRef] 29. Cabrini, M.; Lorenzi, S.; Pastore, T.; Pellegrini, S.; Ambrosio, E.P.; Calignano, F.; Manfredi, D.; Pavese, M.; Fino, P. Effect of heat treatment on corrosion resistance of DMLS AlSi10Mg alloy. Electrochim. Acta 2016, 206, 346–355. [CrossRef] 30. Cabrini, M.; Lorenzi, S.; Pastore, T.; Pellegrini, S.; Testa, C.; Manfredi, D.; Calignano, F.; Fino, P.; Cattano, G. Evaluation of corrosion resistance of aluminum silicon alloys manufactured by LPBF | Valutazione della resistenza a corrosione di leghe di alluminio-silicio ottenute per LPBF. Metall. Ital. 2017, 109, 123–126. 31. Cabrini, M.; Lorenzi, S.; Pastore, T.; Testa, C.; Manfredi, D.; Lorusso, M.; Calignano, F.; Pavese, M.; Andreatta, F. Corrosion behavior of AlSi10Mg alloy produced by laser powder bed fusion under chloride exposure. Corros. Sci. 2019, 152, 101–108. [CrossRef] 32. Cabrini, M.; Calignano, F.; Fino, P.; Lorenzi, S.; Lorusso, M.; Manfredi, D.; Testa, C.; Pastore, T. Corrosion Behavior of Heat-Treated AlSi10Mg Manufactured by Laser Powder Bed Fusion. Materials 2018, 11, 1051. [CrossRef] 33. Cabrini, M.; Lorenzi, S.; Pastore, T.; Testa, C.; Manfredi, D.; Cattano, G.; Calignano, F. Corrosion resistance in chloride solution of the AlSi10Mg alloy obtained by means of LPBF. Surf. Interface Anal. 2016, 48, 818–826. [CrossRef] 34. Cabrini, M.; Lorenzi, S.; Testa, C.; Pastore, T.; Manfredi, D.; Lorusso, M.; Calignano, F.; Fino, P. Statistical approach for electrochemical evaluation of the effect of heat treatments on the corrosion resistance of AlSi10Mg alloy by laser powder bed fusion. Electrochim. Acta 2019, 305, 459–466. [CrossRef] 35. Marchese, G.; Lorusso, M.; Calignano, F.; Ambrosio, E.P.; Manfredi, D.; Pavese, M.; Biamino, S.; Ugues, D.; Fino, P. Inconel 625 by direct metal laser sintering: Effects of the process parameters and heat treatments on microstructure and hardness. In Proceedings of the 13th International Symposium of Superalloys, Seven Springs, PA, USA, 11–15 September 2016. 36. Kim, J.S.; Lee, H.W. A study on effect of intergranular corrosion by heat input on Inconel 625 overlay weld metal. Int. J. Electrochem. Sci. 2015, 10, 6454–6464. 37. Corrieu, J.; Vernot-Loier, C.; Cortial, F. Influence of Heat Treatments on Corrosion Behavior of Alloy 625 Forged Rod. Superalloys 1994.[CrossRef] 38. Cortial, F.; Corrieu, J.M; Vernot-Loier, C. Influence of heat treatments on microstructure, mechanical properties, and corrosion resistance of weld alloy 625. Metall. Mater. Trans. A 1995, 26, 1273–1286. [CrossRef]

39. Tawancy, H.M.; Allam, I.M.; Abbas, N.M. Effect of Ni3Nb precipitation on the corrosion resistance of Inconel alloy 625. J. Mater. Sci. Lett. 1990, 9, 343–347. [CrossRef] 40. Manfredi, D.; Calignano, F.; Krishnan, M.; Canali, R.; Ambrosio, E.P.; Atzeni, E. From powders to dense metal parts: Characterization of a commercial AlSi10Mg alloy processed through direct metal laser sintering. Materials 2013, 6, 856–869. [CrossRef] 41. Kreitcberg, A.; Brailovski, V.; Turenne, S. Elevated temperature mechanical behavior of IN625 alloy processed by laser powder-bed fusion. Mater. Sci. Eng. A 2017, 700, 540–553. [CrossRef] 42. Antonsson, T.; Fredriksson, H. The effect of cooling rate on the solidification of INCONEL 718. Metall. Mater. Trans. B-Process Metall. Mater. Process. Sci. 2005, 36, 85–96. [CrossRef] Materials 2019, 12, 1742 11 of 11

43. Marchese, G.; Lorusso, M.; Parizia, S.; Bassini, E.; Lee, J.W.; Calignano, F.; Manfredi, D.; Terner, M.; Hong, H.U.; Ugues, D.; Lombardi, M.; Biamino, S. Inﬂuence of heat treatments on microstructure evolution and mechanical properties of Inconel 625 processed by laser powder bed fusion. Mater. Sci. Eng. A 2018, 729, 64–95. [CrossRef] 44. Harrison, N.J.; Todd, I.; Mumtaz, K. Reduction of micro-cracking in nickel superalloys processed by Selective Laser Melting: A fundamental alloy design approach. Acta Mater. 2015, 94, 59–68. [CrossRef] 45. Tobergte, D.R.; Curtis, S. Modeling and Experimental validation of Nickel-based super alloy (Inconel 625) made using Selective Laser Melting. Solid Free. Fabr. 2013.[CrossRef] 46. Anam, M.A.; Dilip, J.J.S.; Pal, D.; Stucker, B. Eﬀect of scan pattern on the microstructural evolution of Inconel 625 during selective laser melting. In Proceedings of the 25th Annual International Solid Freeform Fabrication Symposium, Austin, TX, USA, 4–6 August 2014. 47. Floreen, S.; Fuchs, G.E.; Yang, W.J. The Metallurgy of Alloy 625. In Superalloys 718, 625, 706 and Various Derivatives; Minerals, Metals and Materials Society: Pittsburgh, PA, USA, 1994; pp. 13–39.

© 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/).