Tailoring of Microstructures and Tensile Properties in the ... - MDPI

metals Article

Tailoring of Microstructures and Tensile Properties in the Solidification of Al-11Si(-xCu) Brazing Alloys Bruno Monti Carmelo Donadoni 1 , Leonardo Fernandes Gomes 1 , Amauri Garcia 2 and José Eduardo Spinelli 1, * 1 2

*

Department of Materials Engineering, Federal University of São Carlos, São Carlos, SP 13565-905, Brazil; [email protected] (B.M.C.D.); [email protected] (L.F.G.) Department of Manufacturing and Materials Engineering, University of Campinas—UNICAMP, Campinas, SP 13083-860, Brazil; [email protected] Correspondence: [email protected]; Tel.: +55-16-3351-8512

Received: 12 September 2018; Accepted: 28 September 2018; Published: 30 September 2018







Abstract: Ternary Al-11wt %Si-(xwt %)Cu alloys are highly recommended as commercial filler metals for aluminum brazing alloys. However, very little is known about the functional inter-relations controlling the solidified microstructures characterizing processes such as torch and furnace brazing. As such, we evaluated two commercial brazing alloys, which are the Al-11wt %Si-3.0wt %Cu and Al-11wt %Si-4.5wt %Cu alloys: Cu contents typically trend in between the suitable alloying spectrum. We analyzed the effects of solidification kinetics over features such as the dendrite arm spacing and the spacing between particles constituting the eutectic mixture. Also, tensile properties were determined as a function of the dendrite microstructure dimensions. The parameters concerned for translating the solidification kinetics were either the cooling rate, or growth velocity related to the displacement of the dendrite tip, or the eutectic front. The relevant scaling laws representing the growth of these brazing alloys are outlined. The experimental results demonstrated that a 50% increase in Cu alloying (from 3.0 to 4.5 wt %) could be operated in order to obtain significant variations in the dendritic length-scale of the microstructure across the produced parts. Overall, the microstructures were constituted by an α-Al dendritic matrix surrounded by a ternary eutectic consisting of α-Al + Al2 Cu + Si. The scale measurements committed to the Al2 Cu eutectic phase pointed out that the increase in Cu alloying has a critical role on refining the ternary eutectic. Keywords: Al alloys; mechanical properties; microstructure; brazing; solidification

1. Introduction Over a wide range of mechanical and thermal applications, multicomponent alloys pertaining to Al–Si systems are most commonly fabricated through processes such as foundry, brazing, and welding [1]. The binary Al–Si alloys present a microstructure consisting of a primary phase, aluminum, or silicon, and a eutectic mixture of these two elements [2]. Silicon is added to aluminum alloys to promote good wear resistance, high heat transfer coefficient, and low thermal expansion coefficient. The addition of Cu to Al–Si alloys is widely used in automotive engine components, such as engine blocks, cylinder heads, and pistons, because of the good castability and fluidity of Al–Si–Cu alloys [1]. Al–Si–Cu ternary alloys are becoming increasingly important in the aerospace and automotive industries, due to their low relative weight and good mechanical strength at relatively high temperatures, and good resistance to abrasion and weldability. In general, Al–Si–Cu ternary alloys have higher mechanical strength than Al–Si alloys and higher corrosion resistance than Al–Cu alloys [3–5]. The increase in demand for materials with properties such as those of the Al–Si–Cu alloys, established the need to control the microstructure with more rigorous specifications. Therefore, good control of Metals 2018, 8, 784; doi:10.3390/met8100784

www.mdpi.com/journal/metals

Metals 2018, 8, 784

2 of 23

the solidification process is essential, since its influence can be noted even in the finished product [3,6]. According to Zeren et al. [7] more research efforts are necessary for a better understanding of the mechanisms responsible for strength variations in Al–Si–xCu alloys. The solidification process and the intrinsic characteristics of the alloy to be solidified have a direct influence on the microstructure formation, which determines the final properties of a casting. The obtained casting parts exhibit mechanical characteristics that depend on inherent aspects occurring during solidification, such as: grain size, the scale of the phases forming the microstructure such as dendrite arm spacings and interphase spacings, the size and distribution of such phases, chemical composition heterogeneities, inclusions, and porosity. The understanding of solidification of aluminum alloys has fundamental importance for planning of manufacturing processes, since it allows a better understanding of the factors affecting microstructure, and consequently, the product quality [8–13]. For the manufacturing of castings with predefined local properties, a widely used process is foundry technology (metal/mold casting processes). The cooling rate related to the process of heat transfer from the casting into the mold is an aspect acting in parallel, which conditions the change of the size and distribution of individual components within the resulting microstructure. The kind of crystallizing phases and the process of nucleation and growth of grains are influenced by the cooling rate. Affecting the mechanical properties, particularly for the reduction of plastic deformation in Al–Si–Cu alloys, the presence of copper also induces the formation of intermetallic phases, such as Al2 Cu [14]. Additionally, intermetallic phases can be crystallized due to the presence of impurities such as Fe in these alloys, degrading the mechanical properties of castings [14]. Commercial alloys such as 384.0 and A384.0 series, Al9Si3Cu(Fe) and ENAC 46000 are recommended for the die casting processes, and these alloys are in the range of between 10.5 to 12.0 wt % Si, and 3.0 to 4.5 wt % Cu [15]. Ceschini et al. [16] performed a study on the production of cast specimens of Al-10wt %Si-2wt %Cu alloy under controlled solidification conditions. The aim was to have samples that were associated with two values of secondary dendrite arm spacing (SDAS), of about 10 µm and 50 µm. The effect of the cooling rate and different Fe and Mn contents on the microstructure was evaluated, and consequently, the tensile and the fatigue properties of the Al-10wt %Si-2wt %Cu casting alloy were determined. The results showed that the cooling rate affected not only the SDAS values, but also the shape of the eutectic Si particles, and the size and volume fraction of the intermetallic compounds. A reduction of SDAS induced higher ultimate tensile strength (UTS) and elongation (EL) to the failure to be obtained. The highest UTS and EL reached 374 MPa and 12.1% for the alloy containing 0.5 wt % Fe and SDAS of 10 µm. On the contrary, in the samples with smaller SDAS, the EL degraded with increasing Fe and Mn contents, due to the larger volume fraction of Fe-rich intermetallic compounds [16]. Wang et al. [17] proposed an alternative Al-13wt %Si-5wt %Cu-0.8wt %Fe alloy fabricated by metal/mold casting followed by a T6 solid solution heat treatment. The UTS and EL values reached 336 MPa and 0.72%, respectively. Even though obvious correlations between such properties and the microstructural features have been declared, no functional correlations have been proposed. By seeking available functional types of correlations between mechanical strength and dendritic spacing, one could refer to the research developed by Okayasu et al. [18]. These authors reported excellent tensile properties for twin roll and Ohno continuous casting of Al-10.6wt %Si-2.5wt %Cu alloy samples, that is, UTS and EL at around 375 MPa and 10%, respectively. These properties have been associated with the fine round α-Al phase and tiny eutectic structures. A clear Hall–Petch relation has been derived, relating the yield tensile strength to the SDAS: σy=0.2 = 6.1(SDAS)−1/2 + 48.5, where σy (MPa) and SDAS (µm). Another important process technique is the joining process of aluminum alloys. Al–Si–Cu ternary alloys are also included among the alloys used for brazing [15]. Brazing of aluminum alloys is considered to be difficult due to the low melting temperature of Al alloys and the high affinity of Al to oxygen. As a reliable and economical method for the bonding of aluminum alloys, brazing with Al–Si alloys has been adopted among a variety of joining techniques. Commercial aluminum brazes such as BAlSi-3 and BAlSi-4, with silicon contents between 7 and 13 wt % Si, have demonstrated some success

Metals 2018, 8, 784

3 of 23

in joining some Al alloys if the corrosion–chemical and mechanical strength aspects are considered. The working temperatures of these Al brazing alloys must be above 590 ◦ C, due to the fact that these Al brazes having melting temperatures in the range of 575–610 ◦ C. Hence, an important goal of the aluminum industry is the development of low-melting-point filler metals [19]. The interaction between the joint components and growth of the intermetallic compounds at the joint interface can be prevented by shorter brazing cycles/lower brazing temperature [20]. The solidification path of any typical aluminum alloy is modified by the addition of major and minor alloying elements, and these have a significant impact on the final microstructure [21]. With copper addition to the Al–Si alloy, the solute is rejected from both of the eutectic constituents. The equilibrium melting point changes locally from a low value at the interface in the direction of the higher liquidus temperature for the alloy, caused by solute segregation ahead of the eutectic interface. The melt in this boundary layer is constitutionally undercooled, when the actual temperature of the melt is less than the equilibrium liquidus temperature. Other changes might be expected by modifying the equilibrium liquidus temperature. This includes variations in the surface tension affecting the wetting angle between the nuclei present in the boundary solid layer and the melt, and changes in the chemical driving force for nucleation. Fundamentally, significant changes can be expected in the nucleation behavior of the Al–Si eutectic due to ternary solute segregation [22]. Kaya and Aker [23], for instance, demonstrated, through experimental data, that the Si flakes forming an Al–Si eutectic alloy could change with alloying elements. Additions of Cu, Co, Ni, Sb, and Bi to an Al-12.6% Si eutectic alloy resulted in finer Si flakes. A study was carried out by Chang et al. [19], where Al-10.8wt %Si-10wt %Cu, and Al-9.6wt %Si-20wt %Cu filler metals were used for the brazing of 6061 aluminum alloy at 560 ◦C. The results demonstrated that the addition of 10 wt % of copper into the Al-12wt %Si filler metal lowered the solidus temperature from 586 ◦ C to 522 ◦ C, and the liquidus temperature from 592 ◦ C to 570 ◦ C. With the increase in copper content to 20 wt % into the Al-12wt %Si filler metal, the liquidus temperature decreased from 592 ◦ C to 535 ◦ C. The highest value obtained for the shear strength referred to the 6061 Al alloy brazed with the Al-10.8wt %Si-10wt %Cu filler metal, which reached 67 MPa for a 60 min brazing time. According to these authors, the higher hardness of the 6061 aluminum alloy subtract near the butt joint interface after brazing with Al–Si–Cu filler metal could be associated with the formation of Al2 Cu intermetallic compounds, due to the fact that copper diffuses towards the 6061 Al alloy [19]. To further optimize the brazing processes associated with the aforementioned Al alloy, the development of a better understanding of the solidification characteristics of Al–Si–Cu alloys could assist with optimization of the brazing process parameters. The aim of the present work was to perform a detailed characterization of the microstructure of an Al-11wt %Si alloy with additions of 3.0 and 4.5 wt % of copper (Cu), directionally solidified (DS) under transient heat flow conditions. This means describing the morphologies, dimensions, and representative features of the eutectic phases for a wide range of solidification cooling rates. Both dendritic and eutectic length-scales of the microstructure were assessed. Correlations between the tensile properties, hardness, the dendritic arm spacing, and the spacing between particles constituting the eutectic were also investigated. Finally, advanced X-ray Diffraction (XRD) and Scanning Electron Microscopy with Energy Dispersive Spectroscopy (SEM-EDS) techniques were performed to determine the main characteristics of the eutectic phases formed along the length of the DS castings. 2. Experimental Procedure 2.1. Experiment to Follow Solidification Kinetics The Al-11wt %Si-3.0wt %Cu and Al-11wt %Si-4.5wt %Cu alloy castings were generated using a transient directional solidification system. A quantity of 1200 g of commercial purity Al, Si, and Cu was firstly melted in a dense high purity graphite crucible by induction heating up to 750 ◦ C in order to melt the Al and homogenize the other elements by diffusion. Then, the temperature was reduced

Metals 2018, 8, x FOR PEER REVIEW Metals 2018, 8, 784

4 of 23 4 of 23

was firstly melted in a dense high purity graphite crucible by induction heating up to 750 °C in order to melt the Al and homogenize the other elements by diffusion. Then, the temperature was reduced ◦ Cand to to700 700°C andheld heldfor for30 30min min before before the the directional directional solidification solidification procedure. procedure. A A detailed detaileddescription description of ofprocessing processingby bytransient transientdirectional directionalsolidification solidificationisisgiven givenelsewhere elsewhere[24,25]. [24,25].Iron Iron(Fe) (Fe)was wasfound foundto to be the main impurity in the tested samples, which remained inside the suitable commercial Fe wt. % be the main impurity in the tested samples, which remained inside the suitable commercial Fe wt % spectrum spectrumof of 0.205 0.205 ±±0.065. 0.065.Detailed Detaileddescriptions descriptionsofofextraction extractionofofsamples samplesfrom fromthe theDS DScastings castingsfor for metallography metallographyand andmechanical mechanicaltests testsare aregiven givenelsewhere elsewhere [12]. [12]. Liquidus Liquidusand andeutectic eutectictemperatures temperatureswere weredetermined determinedfor forthe thetwo twoAl–Si–Cu Al–Si–Cubased basedalloys alloysthrough through experiments experimentsin inwhich whichthe thealloy alloywas wasslowly slowly cooled cooledin in aa well-insulated well-insulatedcrucible, crucible,thus thuspermitting permittingthe the transformation temperatures to be determined, as illustrated in Figure 1. transformation temperatures to be determined, as illustrated in Figure 1.

Figure1.1. Overall Overallsketches sketchesshowing showingtwo twoadopted adoptedtechniques techniquesdealing dealingwith withthe themolten moltenAl-11Si(-xCu) Al-11Si(-xCu) Figure alloys in in the the present present research, research, which which are: are: the the determination determination of of cooling cooling curves curves and and consequently consequently alloys liquidus/eutectictemperatures, temperatures,and andan anassembly assemblyofofthe thedirectional directionalsolidification solidificationsystem systemwith withinsertion insertion liquidus/eutectic ofthermocouples thermocoupleswithin withinaasplit splitstainless stainlesssteel steelmold. mold. of

TheAl Alalloys alloyswere werefirst firstmelted meltedin inan aninduction inductionfurnace. furnace.After Afterthat, that,the themolten moltenalloy alloywas waspoured poured The into two cavities; that is, either a crucible dedicated to the determination of cooling curves or split into two cavities; that is, either a crucible dedicated to the determination of cooling curves or aasplit moldinserted insertedinto intothe thesolidification solidificationsystem. system.AAremelting remeltingoperation operationofofthe thealloy alloywas wasrun runinside inside the the mold mold since radial electrical wiring heated up the cylindrical stainless steel split mold (see Figure 1). mold since radial electrical wiring heated up the cylindrical stainless steel split mold (see Figure 1). When the melt temperature achieved 3% above the liquidus temperature, the furnace windings were When the melt temperature achieved 3% above the liquidus temperature, the furnace windings were disconnected,and andat atthe thesame sametime, time,the theexternal externalwater waterflow flowat atthe thebottom bottomof ofthe thecontainer containerbegan beganthe the disconnected, cooling down procedure, thus permitting the onset of solidification. cooling down procedure, thus permitting the onset of solidification. The solidification solidification system system permits K-type thermocouples along the The permits the theplacing placingofofa anumber numberofoffine fine K-type thermocouples along length of the casting. Eight thermocouples were strategically spaced between eacheach otherother untiluntil they the length of the casting. Eight thermocouples were strategically spaced between were 96 mm from the cooled bottom of the casting. The frequency of temperature data acquisition was they were 96 mm from the cooled bottom of the casting. The frequency of temperature data 1 Hz on each thermocouple. examination regarding the determination of determination exact positions acquisition was 1 Hz on eachPostmortem thermocouple. Postmortem examination regarding the of the thermocouple tips was carried out. of exact positions of the thermocouple tips was carried out. 2.2.Microstructural MicrostructuralCharacterization Characterizationand andTensile Tensile Tests Tests 2.2. Themacrostructure macrostructureof ofeach eachdirectionally directionallysolidified solidified(DS) (DS)casting castingwas wasrevealed revealedafter afterassessing assessingand and The grinding the whole longitudinal middle section surface with #600 grid paper. The etching solution grinding the whole longitudinal middle section surface with #600 grid paper. The etching solution used was composed of 95 mL of distilled water, 2.5 mL of HNO3 , 1.5 mL of HCl, and 1 mL of HF, which was applied for a couple of seconds.

Metals 2018, 8, 784

5 of 23

Longitudinal and transverse samples at various positions from the cooled bottom of the DS castings were mounted (i.e., 3 mm, 8 mm, 13 mm, 18 mm, 23 mm, 33 mm, 48 mm, 68 mm, and 88 mm), polished, and etched with a solution of 0.5% HF in water over 20 s, and then examined using an optical microscope (Olympus Co., Tokyo, Japan). The length-scale of the dendritic matrix was characterized by the primary (λ1 ), secondary (λ2 ), and tertiary (λ3 ) dendritic arm spacings. Forty measurements were performed for each microstructural spacing of each selected position when a certain alloy composition is considered. One of the targets of the present study was to correlate the dendrite spacing with the tensile properties measured in uniaxial tensile tests of the examined ternary Al–Si–Cu alloys. In order to obtain these measurements, several specimens were extracted along the length of the DS alloy castings. Each specific position chosen for tensile tests allowed three specimens to be extracted so that the average tensile properties regarding strength and ductility and their standard deviations could be determined. These specimens were subjected to tensile tests according to specifications of the ASTM Standard E 8M/04 at a strain rate of about 3 × 10−3 s−1 . Microhardness tests were performed on the transversal sections of the DS samples, using a test load of 1000 g and a dwell time of 10 s. The adopted Vickers microhardness (Hardness tester, Shimadzu, Kyoto, Japan) was the average of at least 10 indentation tests on each sample. The sizes of the eutectic phases were examined through a scanning electron microscope (SEM). Back-scattered electron (BSE) examinations were carried on deep etched samples (HCl over three minutes). The instrument used was a Philips SEM (XL-30 FEG, FEI, Hillsboro, OR, USA) equipped with an energy dispersive X-ray spectrometer (EDS). An analysis of the SEM images allowed the measurements of the spacing of the Al2 Cu particles, λAl2Cu as well those between the eutectic Si particles, λSi . Measurements of the eutectic-related spacings, λAl2Cu and λSi , were performed using the line intercept method [13,23]. Considering that the Al–Si eutectic has an anomalous structure, approximately 20–30 minimum spacing, λm , and maximum spacing, λM , values were measured in the various positions along the length of the DS castings in order to obtain the average λSi . The average spacing λSi is the arithmetic average between λm and λM . The X-ray diffraction (XRD) patterns of phases formed along the length of the Al–Si–Cu alloy castings were acquired by a Siemens D5000 diffractometer (Siemens, Munich, Germany) with a 2-theta range from 20◦ to 90◦ , CuKα radiation and a wavelength, λ, of 0.15406 nm. 3. Results and Discussion The as-cast macrostructures depicted in Figure 2 revealed the prevalence of very fine columnar grains after chemical etching of the DS Al-11wt %Si-3.0 and -4.5wt %Cu alloy castings. Such macro-morphologies prevailed along the entire length of the casting with a few equiaxed grains produced at the very top of the castings. These structures enabled a wide-ranging examination of the fabricated castings, with emphasis on the formed dendritic arrangements. The longitudinal sections could be assessed for the growth of secondary dendrite branches, whereas the cross sections along the length of the DS bodies were useful for the determination of either primary or tertiary dendritic spacing. The growth of well-aligned columnar structures also signifies that the heat flow during solidification remained unidimensionally driven. From the start of the water flow in the directional solidification experiment and considering the recording of experimental data, Figure 3 shows the variations of temperature that occurred for each thermocouple within the castings. It is worth noting that for thermocouples near the water-cooled surface, temperature changes quickly, whereas variation was much slower for positions monitored farther from the bottom cooled surface. The proper evaluation of these cooling curves provided the experimental variations of solidification cooling rates and growth velocities, as is presented next.

Metals2018, 2018,8, 8,784 x FOR PEER REVIEW Metals Metals 2018, 8, x FOR PEER REVIEW

of 23 23 66 of 6 of 23

Figure 2. Directionally solidified (DS) macrostructures characterizing: (a) Al-11wt. %Si-3wt. %Cu and Figure 2.2.Directionally solidified (DS) macrostructures Al-11wt %Si-3wt Figure Directionally macrostructures characterizing: characterizing: (a) (a) Al-11wt. %Si-3wt.%Cu %Cuand and (b) Al-11wt. %Si-4.5wt. solidified %Cu alloy(DS) castings. (b) Al-11wt %Si-4.5wt %Cu alloy castings. (b) Al-11wt. %Si-4.5wt. %Cu alloy castings. 650 650 Al-11wt.%Si-3.0wt.%Cu Al-11wt.%Si-3.0wt.%Cu

Positions from metal/mold interface: Positions3from mm metal/mold 8interface: mm 3 mm 37 mm 37 mm

600 600

(°C) Temperature (°C) Temperature

8 mm 51 mm 51 mm

13 mm 13 66 mm mm 66 mm

18 mm 18 96 mm mm 96 mm

TL= 850.6K (577.6 °C) TL= 850.6K (577.6 °C)

550 550

TTE= 788K (515 °C) TTE= 788K (515 °C)

500 500

450 450

400 400 0 0

200 200

400 400

Time (s) Time (s) (a) (a)

Figure 3. Cont.

600 600

Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

7 of 23 7 of 23

650 Al-11wt.%Si-4.5wt.%Cu

Positions from metal/mold interface: 5 mm

10 mm

15 mm

20 mm

38 mm

53 mm

68 mm

98 mm

600

Temperature (°C)

TL= 849.6K (576.6 °C) 550

TTE= 788K (515 °C) 500

450

400 0

200

400

600

Time (s)

(b) Figure 3.3. Experimental Experimental cooling cooling curves curves considering considering different different positions positions along along the the length length of of the the Figure (a) Al-11wt. %Si-3.0wt. %Cu and (b) Al-11wt. %Si-4.5wt. %Cu alloy castings. Each mentioned position (a) Al-11wt %Si-3.0wt %Cu and (b) Al-11wt %Si-4.5wt %Cu alloy castings. Each mentioned position referstotoaathermocouple thermocoupleinserted insertedwithin withinthe theDS DScasting. casting. refers

Figure 44 shows shows the the time time evolutions evolutions of of the the liquidus liquidus front front along along the the lengths lengths of of both both the the Figure Al-11wt.%Si-3.0wt %Si-3.0wt. %Cu Al-11wt. %Si-4.5wt. castings during cooling. plots Al-11wt %Cu andand Al-11wt %Si-4.5wt %Cu%Cu alloyalloy castings during cooling. TheseThese plots were were generated by the previous information on the temperatures liquidus temperatures of and eachby alloy, and by generated by the previous information on the liquidus of each alloy, monitoring monitoring their transit onengaged each of thermocouples the engaged thermocouples at various positions (P) along the their transit on each of the at various positions (P) along the length of the length of the casting. As a consequence of these plots, growth velocities (v L ) experimental tendencies casting. As a consequence of these plots, growth velocities (vL ) experimental tendencies in Figure 5a,b in Figure 5a,b were established foralloys. both evaluated alloys.are These valuesofare results of the were established for both evaluated These vL values directvLresults thedirect time-derivatives of time-derivatives of the experimental in Figure 4. vofL displacement represents theofrate displacement of the experimental functions in Figure 4.functions vL represents the rate theof liquidus isotherm the liquidus isotherm (v L ). There is no significant change between the v L plots of both examined alloys. (vL ). There is no significant change between the vL plots of both examined alloys. Analogously, Analogously, rates of displacements of the Al–Si binary eutectic BE) and ofCu theternary Al–Si–Al 2Cu ternary rates of displacements of the Al–Si binary eutectic (vBE ) and of the(vAl–Si–Al eutectic (vTE ) 2 eutectic TE) couldfrom be obtained fromcurve the cooling curve analyses of both alloys. The calculation of such could be (v obtained the cooling analyses of both alloys. The calculation of such velocities is very important, since their is magnitude is directly toscales the eutectic and the isvelocities very important, since their magnitude directly related to therelated eutectic and thescales morphology morphology developed from reaction. each eutectic reaction. developed from each eutectic The determination of the tipcooling coolingrate, rate,T˙ṪLL,, as as aa function function of of position position (P) (P) in in the thecasting casting was was The determination of the tip carried out by computing the time-derivative of each cooling curve (dT/dt) right after the passage of carried out by computing the time-derivative of each cooling curve (dT/dt) right after the passage thetheliquidus be seen seen in in of liquidusisotherm isothermbybyeach eachthermocouple. thermocouple.AAlarge largespectrum spectrumof of cooling cooling rates rates can can be Figure 5c,d. Further, when comparing the cooling-down regimes of both alloys it can be observed Figure 5c,d. Further, when comparing the cooling-down regimes of both alloys it can be observed that thatalloy the alloy containing higher Cu content is associated higher of cooling Although the containing higher Cu content is associated with with higher levelslevels of cooling rates.rates. Although the thecontents Cu contents lie inside the commercial spectrum of braze Al fillers, the levels the cooling Cu here here lie inside the commercial spectrum of braze Al fillers, the levels of the of cooling rates rates eachdiffer alloy from differeach fromother. each other. of eachofalloy

Metals 2018, 8, x FOR PEER REVIEW Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

8 of 23 8 of 23 8 of 23

100

Al-11wt.%Si-3.0wt.%Cu

100

Al-11wt.%Si-4.5wt.%Cu

100

Al-11wt.%Si-3.0wt.%Cu

100

Al-11wt.%Si-4.5wt.%Cu

80

80

80

(mm) Position (mm) Position

(mm) Position (mm) Position

80 60 60 40 40 20

0.7

P = 2.4 (t) 0.7 P 2==2.4 (t) R 0.99 2 R = 0.99

20 0 0 0

50

0

50

100

150

60 60 40 40 20

0.7

P = 2.07 (t) 0.7 P 2==2.07 R 0.99(t) 2 R = 0.99

20 0 0

200

250

0

50

Liquidus front passing 100 150 time (s) 200

250

0

50

Liquidus front passing time (s)

200

250

Liquidus front passing 100 150 time (s) 200

100

150

250

Liquidus front passing time (s)

Figure 4. Experimental time-dependent (t) displacements of the liquidus isotherm along the casting Figure 4. time-dependent (t) displacements displacements of of the the liquidus liquidus isotherm isotherm along along the the casting Figure 4. Experimental Experimental time-dependent casting length (P) of the Al-11Si-xCu alloys. R22 is(t) the coefficient of determination. length (P) of the Al-11Si-xCu alloys. R is the coefficient of determination. length (P) of the Al-11Si-xCu alloys. R2 is the coefficient of determination. 1.4

Al-11wt.%Si-3.0wt.%Cu

1.2

Al-11wt.%Si-3.0wt.%Cu

1.4

(a) (a)

1.4

Al-11wt.%Si-4.5wt.%Cu

1.2

Al-11wt.%Si-4.5wt.%Cu

1.2

1.2

1.0

1.0

(mm/s) Growth (mm/s) raterate Growth

(mm/s) Growth (mm/s) raterate Growth

1.4

1.0 0.8 -0.4

vL = 2.2 (P) -0.4 vL2==2.2 (P) R 0.95

0.8 0.6 0.6

2

R = 0.95

0.4 0.4 0.2 0.2

0.0

1.0

0.8 0.6 0.6 0.4 -0.45

vL = 2.9 P -0.45 v 2= 2.9 P RL = 0.8 2 R = 0.8

0.4

0.2 0

20

40

60

0

20

40Position (mm) 60

80

100

80

100

0.0 0.0

0

20

40

0

20

40Position (mm) 60

Position (mm) 21

21

(c)

Al-11wt.%Si-4.5wt.%Cu

21 18

Al-11wt.%Si-4.5wt.%Cu

18 15

(°C/s) Cooling (°C/s) raterate Cooling

(°C/s) Cooling (°C/s) raterate Cooling

(c)

Al-11wt.%Si-3.0wt.%Cu

15 12 12 9

-1.4

= 67(P) -1.4 T L 2= 67(P) R = 0.99 TL

9 6

2

R = 0.99

6 3

80

100

80

100

(d) (d)

18 15 15 12 12 9

-1.3

= 88 (P) -1.3 (P) L R ==88 0.99 TL

9 6

T 2

2

R = 0.99

6 3 3 0

3 0 0

60

Position (mm)

Al-11wt.%Si-3.0wt.%Cu

21 18

(b)

0.8

0.2

0.0

(b)

0 0

20 20

40 40 Position

60 60 (mm)

Position (mm)

80 80

100 100

0

0

20

40

0

20

Position 40

60

80

100

(mm) 60

80

100

Position (mm)

Figure 5. 5. Experimental Experimental variations variations of of (a,b) (a,b) growth growth rate rate and and (c,d) (c,d) tip tip cooling cooling rate rate during during unsteady unsteady state state Figure 2 2 directional solidification of the ternary Al-11Si-xCu alloys. R is the coefficient of determination. Figure 5. Experimental of (a,b) growth rate and (c,d) rateof during unsteady state directional solidificationvariations of the ternary Al-11Si-xCu alloys. R is tip thecooling coefficient determination. directional solidification of the ternary Al-11Si-xCu alloys. R2 is the coefficient of determination.

Figures 6 and 7 present the optical micrographs of the Al-11wt %Si-3.0 and 4.5wt %Cu braze Figures 6 and 7 present the optical micrographs of the Al-11wt. %Si-3.0 and 4.5wt. %Cu braze alloyFigures samples. and In the micrographs, the aluminium-rich dendrites can be seen being enveloped by present the optical micrographs of the Al-11wt. %Si-3.0 and 4.5wt. %Cuby braze alloy samples.6 In the7 micrographs, the aluminium-rich dendrites can be seen being enveloped the the products of the reactions; is, the Al–Si eutectic and the ternary eutectic. 2 Cu alloy samples. theeutectic micrographs, thethat aluminium-rich dendrites beAl–Si–Al seen2Cu being enveloped byThe the products of theIn eutectic reactions; that is, the Al–Si eutectic and thecan Al–Si–Al ternary eutectic. The micrographs refer to different positions along the length of the DS castings. Very fine to large products of the eutectic reactions; that is,along the Al–Si eutectic thecastings. Al–Si–AlVery 2Cu ternary eutectic. The micrographs refer to different positions the length of and the DS fine to large lengthlength-scales of the formed dendritic structures can be observed. This is explained due to the wide micrographs refer to different positions along the length of the DS castings. Very fine to large lengthscales of the formed dendritic structures can be observed. This is explained due to the wide range of range of solidification cooling experienced across the fabricated DS wide castings. scales of experimental the formed dendritic structures canrates be observed. This explained due to the range of experimental solidification cooling rates experienced across theisfabricated DS castings. experimental solidification cooling rates experienced across the fabricated DS castings.

Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

9 of 23 9 of 23

Figure 6.6.Optical the casting casting length; length;that thatis, Figure Opticalmicrostructures microstructuresrepresenting representingdifferent differentpositions positions (P) (P) along along the is,different differentcooling coolingrates ratesduring during solidification of the Al-11wt %Si-3.0wt %Cu alloy: (a,b) P = 3 mm, solidification of the Al-11wt. %Si-3.0wt. %Cu alloy: (a,b) P = 3 mm, (c,d) (c,d) 8 mm 48 mm. P =P8 =mm andand (e,f)(e,f) P = P48=mm.

Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

10 of 23 10 of 23

Figure 7. Optical microstructures representing different positions along the casting length, that is, Figure 7. Optical microstructures representing different positions along the casting length, that is, different cooling rates during solidification of the Al-11wt %Si-4.5wt %Cu alloy: (a,b) P = 3 mm, different cooling rates during solidification of the Al-11wt. %Si-4.5wt. %Cu alloy: (a,b) P = 3 mm, (c,d) (c,d) P = 8 mm and (e,f) P = 48 mm. P = 8 mm and (e,f) P = 48 mm.

Figure 8 shows the magnitude modifications that occurred in the microstructural spacing values shows thealong magnitude modifications that occurred the microstructural spacing values for the Figure various8 positions the length of the DS castings. Largeindendritic variations in scale could the various positions lengthdendrite of the DSarm castings. Large variations in930 scale could befor noted. For example, thealong meanthe primary spacing, λ1 , dendritic varied from 45 µm to µm in be noted. For example, the mean primary dendrite arm spacing, λ 1, varied from 45 μm to 930 μm in the samples extracted along the length of the Al-11wt %Si-3.0wt %Cu alloy casting. The secondary the samples extracted alongfrom the length alloy casting. Theacross secondary dendritic spacing, λ2 , varied 7.5 µmoftothe 30Al-11wt. µm as a %Si-3.0wt. function of%Cu the relative position the dendritic spacing, λ 2, varied from 7.5 μm to 30 μm as a function of the relative position across the AlAl-11wt %Si-4.5wt %Cu alloy casting. 11wt. %Si-4.5wt. %Cu alloy casting.

Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

11 of 23 11 of 23

1200

1200

(a) Primary dendritic spacing, 1 (m)

Primary dendritic spacing, 1 (m)

1000

800

600

400

200

Experimental data 0.82 2 1 = 23 (P) R = 0.94

1000

800

600

400

Experimental data 0.71 2  = 42.5 (P) R = 0.98

200

0

0 20

40

60

80

20

100

60

80

100

40

40

(c)

Al-11wt.%Si-3.0wt.%Cu

35

Secondary dendritic spacing,  (m)

Secondary dendritic spacing,  (m)

40

Position (mm)

Position (mm)

30 25 20 15 10

Experimental data  = 3.8 P0.43 R2 = 0.94

5 0

(d)

Al-11wt.%Si-4.5wt.%Cu

35 30 25 20 15 10

Experimental data  = 5.8 P0.38 R2 = 0.98

5 0

0

20

40

60

80

100

0

20

Position (mm)

40

60

80

100

Position (mm) 35

35

(e)

Al-11wt.%Si-3.0wt.%Cu

Tertiary dendritic spacing,  (m)

Tertiary dendritic spacing,  (m)

(b)

Al-11wt.%Si-4.5wt.%Cu

Al-11wt.%Si-3.0wt.%Cu

30 25 20 15 10

Experimental data  = 4.8 P0.35 R2 = 0.95

5

(f)

Al-11wt.%Si-4.5wt.%Cu 30 25 20 15 10

Experimental data  = 4.5 P0.41 R2 = 0.89

5 0

0 0

20

40

60

Position (mm)

80

100

0

20

40

60

80

100

Position (mm)

Figure 8. 8. Primary Primary (a,b), (a,b), secondary secondary (c,d), (c,d), and and tertiary tertiary (e,f) (e,f) dendrite dendrite arm arm spacing spacing variations variations across across the the Figure length of the Al-11Si-xCu braze alloy castings. length of the Al-11Si-xCu braze alloy castings.

Figure 9 depicts the mean experimental values, along with the standard variation, of primary (λ1 ), Figure 9 depicts the mean experimental values, along with the standard variation, of primary secondary (λ2 ), and tertiary (λ3 ) dendritic spacings as a function of cooling rate (primary, tertiary), and (λ1), secondary (λ2), and tertiary (λ3) dendritic spacings as a function of cooling rate (primary, growth rate (secondary). Power function fittings were suitably derived to represent the experimental tertiary), and growth rate (secondary). Power function fittings were suitably derived to represent the scatter of each alloy. Analyzing the experimental tendencies in the graphs of Figure 9, it can be experimental scatter of each alloy. Analyzing the experimental tendencies in the graphs of Figure 9, inferred that the 50% increase in the Cu alloying (from 3.0 to 4.5 wt %) had significant impacts on it can be inferred that the 50% increase in the Cu alloying (from 3.0 to 4.5 wt. %) had significant λ1 , λ2 , and λ3 . This means increases in λ1 , λ2 , and λ3 by 88%, 45%, and 42% respectively, for the impacts on λ1, λ2, and λ3. This means increases in λ1, λ2, and λ3 by 88%, 45%, and 42% respectively, alloy containing 4.5 wt % of Cu, as compared to the other alloy composition. The multipliers of the for the alloy containing 4.5 wt. % of Cu, as compared to the other alloy composition. The multipliers of the experimental equations varied when compared with the experimental tendencies of the two evaluated alloys for a particular microstructural parameter. However, the representative exponents

Metals 2018, 8, 784

12 of 23

experimental equations varied Metals 2018, 8, x FOR PEER REVIEW

when compared with the experimental tendencies of the two evaluated 12 of 23 alloys for a particular microstructural parameter. However, the representative exponents referring to referring to each microstructural parameter wereregardless preservedofregardless of thealloy. considered alloy. A −2/3 each microstructural parameter were preserved the considered A −2/3 power law power law characterizes the experimental λ2 with L, while a −0.55 exponent is ablethe to characterizes the experimental variations ofvariations λ2 with vLof , while a −v0.55 exponent is able to represent represent the two for λ1. two tendencies fortendencies λ1 . (a)

Al-11wt.%Si-3.0wt.%Cu

Primary dendritic spacing, 1 (m)

1 = 265 (T L)

-0.55

2

R = 0.96

Al-11wt.%Si-4.5wt.%Cu 1 = 500 (T L)

3

10

-0.55

2

R = 0.96

2

10

-1

10

0

1

10

10

Secondary dendritic spacing, 2 (m)

Cooling Rate, T L (‫؛‬C/s)

(b)

Al-11wt.%Si-3.0wt.%Cu -2/3 2 2 = 11 (vL) R = 0.90

2

10

Al-11wt.%Si-4.5wt.%Cu -2/3 2 2 = 16 (vL) R = 0.90

1

10

0

0

10

2x10

Growth rate, vL (mm/s)

Tertiary dendritic spacing, 3 (m)

(c)

Al-11wt.%Si-3.0wt.%Cu

2

10

3 = 12 (T L)

-1/4

2

R = 0.98

Al-11wt.%Si-4.5wt.%Cu 3 = 17 (T L)

-1/4

2

R = 0.98

1

10

-1

10

0

10

1

10

Cooling rate, T L (‫؛‬C/s)

Figure 9. Experimental Experimental variations variations characterizing: characterizing: (a) (a) primary primary dendrite dendrite arm spacing, λ cooling Figure λ1, vs. cooling rate, (b) secondary (c)(c) tertiary dendrite arm spacing, λ3 , rate, secondary dendrite dendritearm armspacing, spacing,λλ vs.growth growthrate, rate,and and tertiary dendrite arm spacing, 2 ,2,vs. 2 is the 2 vs. cooling rate of the Al-11wt %Si-xwt %Cu alloys. R coefficient of determination. λ3, vs. cooling rate of the Al-11wt. %Si-xwt. %Cu alloys. R is the coefficient of determination.

Circulating between a liquid richrich in solute. The The dendritic arrayarray is formed by lateral Circulating betweenthe thedendrites dendritesis is a liquid in solute. dendritic is formed by instabilities of higher order, which are the secondary and tertiary branches developed from the primary lateral instabilities of higher order, which are the secondary and tertiary branches developed from the primary stems. The higher order tertiary formations are those that are located closer to the interdendritic portions, which are prone to commence the eutectic reaction. It appears that such proximity with the eutectic mixture may affect the growth of tertiary branches, which becomes less sensitive to the cooling rate variations, resulting in a relationship λ3 = constant ṪL−1/4.

Metals 2018, 8, 784

13 of 23

stems. The higher order tertiary formations are those that are located closer to the interdendritic portions, which are prone to commence the eutectic reaction. It appears that such proximity with the eutectic mixture may affect the growth of tertiary branches, which becomes less sensitive to the cooling rate variations, resulting in a relationship λ3 = constant T˙ L −1/4 . Metals 2018, 8, x FOR PEER REVIEW 13 of 23 The solidification paths of the Al–Si–Cu braze alloys were computed using the Thermo-Calc software The (Thermo-Calc Software Sweden). Thiswere was possible using assumption of solidification paths ofAB, the Solna, Al–Si–Cu braze alloys computedby using thethe Thermo-Calc Scheilsoftware conditions and the TCAL5 Al-based Alloys Database. Figure 10 shows the isopleth (Thermo-Calc Software AB, Solna, Sweden). This was possible by using the assumption of simulation Scheil relative to the Al-11Si-xCu system, considering then a Figure parameterization 11wt simulation %Si. The solidification conditions and the TCAL5 Al-based Alloys Database. 10 shows the of isopleth relative to the of Al-11Si-xCu system, then as a parameterization of bottom 11wt. %Si. The evolutions these alloys were considering also calculated, can be seen in the plots of solidification Figure 10. In this of these calculated, as canin bethe seen in thecomposition bottom plots to of permit Figure 10. In this case, evolutions an impurity level alloys of 0.2 were wt %also Fe was considered alloys more realistic case, an impurity level of 0.2 wt. % Fe was considered in the alloys composition to permit more sequences of precipitation to be estimated. According to Thermo-Calc results, the eutectic reaction realistic sequences of precipitation to be estimated. According to Thermo-Calc results, the eutectic may occur at 522 ◦ C for both Al-11wt %Si-3.0 and -4.5wt %Cu alloys. The products of this reaction are: reaction may occur at 522 °C for both Al-11wt. %Si-3.0 and -4.5wt. %Cu alloys. The products of this α-Al + AlFeSi + Si + Al2 Cu. A comparison of the graphs at the bottom of Figure 10 shows that a mass reaction are: α-Al + AlFeSi + Si + Al2Cu. A comparison of the graphs at the bottom of Figure 10 shows fraction 9% isfraction associated with the eutectic for structure the alloyfor containing 3.0 wt % Cu, whereas it thatof a mass of 9% is associated withstructure the eutectic the alloy containing 3.0 wt. % is 14% for the Al-11wt %Si-4.5wt %Cu alloy. Cu, whereas it is 14% for the Al-11wt. %Si-4.5wt. %Cu alloy.

Figure 10. Partial pseudo-binary diagram,and and graphs of temperature versus Figure 10. Partial pseudo-binaryAl-11wt Al-11wt.%Si-xCu %Si-xCu phase phase diagram, graphs of temperature versus the solid fraction (solidification paths), computed by the Thermo-Calc software for alloys containing the solid fraction (solidification paths), computed by the Thermo-Calc software for alloys containing 0.2 wt0.2%wt. Fe.% Fe. The following precipitations (fractions occurred the Al-11wt. %Si-3.0and and-4.5wt -4.5wt.%Cu %Cualloys The following precipitations (fractions %)%) occurred forforthe Al-11wt %Si-3.0 alloys respectively: 9% and 6% of α-Al, followed by nearly 50% and 41% of the solid fraction of Si; that, respectively: 9% and 6% of α-Al, followed by nearly 50% and 41% of the solid fraction of Si; after after that, the growth of 32% and 39% of solid fractions related to the AlFeSi phase, before the eutectic reaction occurring in the remaining liquid. Accumulation of the proportions of α-Al with that for the binary eutectic Si resulted in ~60% for the Al-11wt. %Si-3.0wt. %Cu alloy, while less than 50% of the

Metals 2018, 8, 784

14 of 23

the growth of 32% and 39% of solid fractions related to the AlFeSi phase, before the eutectic reaction occurring in the remaining liquid. Accumulation of the proportions of α-Al with that for the binary eutectic resulted ~60% for the Al-11wt %Si-3.0wt %Cu alloy, while less than 50% of the14fraction of MetalsSi2018, 8, x FORin PEER REVIEW of 23 these phases is associated with the Al-11wt %Si-4.5wt %Cu alloy. Such a combination of mass fractions fraction of these since—according phases is associatedtowith the Al-11wt. %Si-4.5wt. %Cu alloy. a combination of may remains important Okayasu [18]—fine α-Al phases and Such tiny silicon structures mass fractions remains important since—according to Okayasu [18]—fine α-Al phases and tiny provide superior tensile and fatigue properties. silicon structures may provide superior tensile and fatigue properties. The phases within the directionally solidified alloy samples have been identified by XRD patterns, The phases within the directionally solidified alloy samples have been identified by XRD as can be seen Figure 11.inThe XRD to different coolingcooling rates along the length patterns, asin can be seen Figure 11. spectra The XRDcorresponding spectra corresponding to different rates along of the castings reveal the occurrence of four different phases, which are α-Al, Al Cu, and Si and 2 the length of the castings reveal the occurrence of four different phases, which are α-Al, Al2Cu, andAlFeSi. The Si presence of the Fe-bearing was nextphase confirmed through SEMthrough analysis of the microstructures, and AlFeSi. The presence ofphase the Fe-bearing was next confirmed SEM analysis of the as seen in Figures 12–15. microstructures, as seen in Figures 12 to 15. Al-11wt.%Si-3.0wt.%Cu -Al Si Al2Cu AlFeSi

Intensity (a.u.)

(a)

P=48mm

T L= 0.4°C/s

P=8mm

T L= 4.6°C/s P=3mm

T L= 14.4°C/s

10

20

30

40

50

60

70

80

90

Angle 2 (Deg.) (b)

Intensity (a.u.)

Al-11wt.%Si-4.5wt.%Cu Al Si Al2Cu AlFeSi

P=48mm

T L= 0.3°C/s

P=8mm

T L= 2.2°C/s

P=3mm

T L= 20.8°C/s 10

20

30

40

50

60

70

80

90

Angle 2 (Deg.)

Figure 11. X-ray diffraction (XRD) patterns of of thethe (a)(a) Al-11wt %Si-3.0 andand (b)(b) -4.5wt %Cu alloy samples Figure 11. X-ray diffraction (XRD) patterns Al-11wt. %Si-3.0 -4.5wt. %Cu alloy samples different cooling rates. solidified at solidified differentat cooling rates. Backscattered electron (BSE) images using SEM, as well as EDS mapping point Backscattered electron (BSE) images using SEM, as well as EDS mapping andand EDSEDS point analyses, analyses, were undertaken to characterize the secondary phases (morphology and chemistry) in the were undertaken to characterize the secondary phases (morphology and chemistry) in the Al–Si–Cu The images12 in and Figures 12 and 13 indicate that the microstructure alloyAl–Si–Cu castings.alloy Thecastings. images in Figures 13 indicate that the microstructure containscontains binary Al–Si binary Al–Si eutectic (gray color areas) and ternary eutectic structures consisting of Al + Si + Al2Cu eutectic (gray color areas) and ternary eutectic structures consisting of Al + Si + Al2 Cu (brighter areas (brighter areas in the images). It can be clearly seen that both eutectics change appreciably in scale as in the images). It can be clearly seen that both eutectics change appreciably in scale as a function of a function of the concerned cooling rate sample as characterized by the images at the left (fast cooling) the concerned cooling rate sample as characterized by13. the images at the left (fast cooling) and right and right (intermediate cooling) sides of Figures 12 and

(intermediate cooling) sides of Figures 12 and 13.

Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

15 of 23 15 of 23

Figure Figure12.12. Transverse Transverse scanning scanning electron electron microscopy microscopy (SEM) (SEM)microstructures microstructuresrelated relatedtotodistinct distinct ◦ C/s and (e–h): 4.6 ◦ C/s) during directional magnifications and cooling rates (i.e., (a–d): 14.4 magnifications and cooling rates (i.e., (a–d): 14.4 °C/s and (e–h): 4.6 °C/s) during directional solidification %Si-3.0wt %Cu alloy. solidificationofofthe theAl-11wt Al-11wt. %Si-3.0wt. %Cu alloy.

Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

16 of 23 16 of 23

Figure Figure13. 13.Transverse TransverseSEM SEMmicrostructures microstructuresrelated relatedto todistinct distinctmagnifications magnificationsand andcooling coolingrates rates(i.e., (i.e., ◦ C/s and (e–h): 5.7 ◦ C/s) during directional solidification of the Al-11wt %Si-4.5wt %Cu alloy. (a–d): 20.8 (a–d): 20.8 °C/s and (e–h): 5.7 °C/s) during directional solidification of the Al-11wt. %Si-4.5wt. %Cu alloy.

Metals 2018, 8, x FOR PEER REVIEW Metals 2018, 8, 784

17 of 23 17 of 23

Figure 14. SEM images in the SE signal of the (a,f) Al–Si–Cu alloy samples along the casting length at Figure 14. SEM images in the SE signal of the (a,f) Al–Si–Cu alloy samples along the casting length at different positions from the cooled bottom surface, and respective X-ray elemental mappings through different positions from the cooled bottom surface, and respective X-ray(d,i) elemental mappings energy dispersive X-ray spectrometry (EDS): Al–K (b,g), Si–K (c,h), Cu–K and Fe–K (e,j). through energy dispersive X-ray spectrometry (EDS): Al–K (b,g), Si–K (c,h), Cu–K (d,i) and Fe–K (e,j).

Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

18 of 23 18 of 23

Figure 15. Representative SEM microstructures microstructures and and EDS microprobe microprobe measurements measurements (at. (at. %) %) related related to Figure the positions positions(P) (P)(a) (a)8 8mm, mm,and and mm from cooled bottom of Al-11wt the Al-11wt. %Si-4.5wt. %Cu the (b)(b) 4848 mm from thethe cooled bottom of the %Si-4.5wt %Cu alloy. alloy.

Higher magnification backscattered electron images were used to identify the growth of the Fe-bearing needles, as indicated by arrows inwere Figure 13. to Compositional analysisofofthe these Higherintermetallic magnification backscattered electron images used identify the growth Feparticles using EDS showed the phasesbywere: 1. α-Al; 2. Al13. 3. Si; 4. AlFeSi. The detailed bearing intermetallic needles,that as indicated arrows in Figure Compositional analysis of these 2 Cu; quantitative chemical composition analyses solid phases different rate samples of particles using EDS showed that the phasesofwere: 1. α-Al;for 2. two Al2Cu; 3. Si; cooling 4. AlFeSi. The detailed the Al-11wt %Si-4.5wt %Cu alloy areanalyses given inof Figure These can be cooling considered be coarse quantitative chemical composition solid 15. phases forparticles two different ratetosamples of enough to permit an accurate their compositions. the Al-11wt. %Si-4.5wt. %Cu determination alloy are givenofin Figure 15. These particles can be considered to be Samples to the slow-cooled regions of of the DS compositions. castings were chosen to be subjected to coarse enoughrelated to permit an accurate determination their examination SEM-EDS regions mapping, seen in Figure 14. Thetofinal distribution Samplesthrough related elemental to the slow-cooled of as thecan DSbecastings were chosen be subjected to of the elements can beelemental seen within the phases and constituents. Thein red contrast for Al showed a high examination through SEM-EDS mapping, as can be seen Figure 14. The final distribution intensity withincan thebe α-Al dendrite branches, expected, whileThe thered green for Si for wasAlconcentrated in of the elements seen within the phases as and constituents. contrast showed a high both the binary the ternary eutectic Si phase. Spots concentrated in Fefor were noted in the fourth intensity withinand the α-Al dendrite branches, as expected, while the green Si was concentrated in EDS mapping in the bottom images of Figure 14. both the binary and the ternary eutectic Si phase. Spots concentrated in Fe were noted in the fourth the fast-cooled samples of Figures 12 and EDS In mapping in the bottom images of Figure 14.13, the morphology of the binary eutectic Si phase is polyhedral-like. However, for allofthe other 12 samples related to lower cooling theeutectic morphology of In the fast-cooled samples Figures and 13, the morphology of therates, binary Si phase the eutectic Si is flake-like as can in Figures 12–15. The formation of polyhedral silicon particles is polyhedral-like. However, forbe allseen the other samples related to lower cooling rates, the morphology in Al–Si based alloys hasbe not been reported the openof literature. of hypoeutectic the eutectic Si is flake-like as can seen incommonly Figures 12–15. The in formation polyhedral silicon In the investigation, the Si precipitates inhibited from growing in the samples, particles inpresent hypoeutectic Al–Si based alloys has not although been commonly reported in the open literature. solidified faster conditions.the Under a fast regime of solidification, the liquid becomes In theunder present investigation, Si precipitates although inhibited from growing in the enriched samples, in Si, which remains in between the primary stems. Asthe the temperature declines, solidified under faster constrained conditions. Under a fast regime of solidification, liquid becomes enriched such particles may grow preferentially in the areas surrounding the Al Cu particles, that is, the previous in Si, which remains constrained in between the primary stems. As2 the temperature declines, such Cu-enriched a consequence, a flake-like morphology notparticles, be attained theprevious growth particles mayzones. grow As preferentially in the areas surrounding the may Al 2Cu thatsince is, the is interruptedzones. forming Si particles. The preferentialmay grown polyhedral silicon Cu-enriched As apolyhedral consequence, a flake-like morphology not be attained since theparticles growth in fast-cooled samples are located Si neighbouring Al2 Cu particles, can be seen in Figures 12d is interrupted forming polyhedral particles. Thethe preferential grown as polyhedral silicon particles in and 13d. These morphologies in neighbouring the Al-11wt %Si-3.0wt %Cu and Al-11wt alloys fast-cooled samples are located the Al2Cu particles, as can %Si-4.5wt be seen in %Cu Figures 12dwere and 13d. These morphologies in the Al-11wt. %Si-3.0wt. %Cu and Al-11wt. %Si-4.5wt. %Cu alloys were

Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

19 of 23 19 of 23

shown to to be be associated associated with with cooling cooling rates rates of of 14.4 14.4K/s K/s and and 20.8 20.8 K/s K/s respectively, relative shown respectively, i.e., i.e., for for aa relative position in the casting of P = 3 mm. position in the casting of P = 3 mm. The experimental experimental evolutions evolutionsof ofthe theSiSispacing, spacing,λλSi, ,and andthe theAl Al2Cu Cuspacing, spacing,λλAl2Cu,, as a function of The 2 Si Al2Cu as a function of binary eutectic growth velocity, v BE, and as a function of the ternary eutectic growth velocity, vTE, binary eutectic growth velocity, vBE , and as a function of the ternary eutectic growth velocity, vTE , were were experimentally determined, andplotted are plotted in Figure 16. Experimental growth laws in the form experimentally determined, and are in Figure 16. Experimental growth laws in the form of of power functions were established. power functions were established.

Si interflake spacing,  Si (m)

Al-12.6wt.%Si-2wt.%Cu -0.41

Al-11wt.%Si-3.0wt.%Cu Al-11wt.%Si-4.5wt.%Cu -1.0 2 Si = 1.4 (vBE) - R =0.82

Si=1.57 (v)

(from ref [23]) 1

10

0

10

(a) -1

0

10

10

Al2Cu interparticle spacing,  Al2Cu (m)

Binary eutectic growth rate, vBE (mm/s) Al-11wt.%Si-3.0wt.%Cu -3/4 2 Al2Cu = 0.60 (vTE) R = 0.92 Al-11wt.%Si-4.5wt.%Cu -3/4 2 Al2Cu = 0.42 (vTE) R = 0.91

0

10

(b) -1

0

10

10

Ternary eutectitc growth rate, vTE (mm/s)

Figure Figure 16. 16. Experimental Experimental variations variations characterizing: characterizing: (a) (a) Si Si interflake interflake spacing, spacing, λλSiSi,, vs. vs. eutectic eutectic growth growth velocity (b) Al inter-particle spacing vs.spacing eutectic growth velocity of the Al-11wt %Si-xwt 2 Cu Al velocity and and (b) 2Cu inter-particle vs. eutectic growth velocity of %Cu the 2 is the coefficient of determination. alloys. R 2 Al-11wt. %Si-xwt. %Cu alloys. R is the coefficient of determination.

The binary Al–Si eutectic was characterized through the experimental evolution of λSi as a The binary Al–Si eutectic was characterized through the experimental evolution of λ Si as a function of the eutectic growth velocity, as can be seen in Figure 16a. Figure 16a also synthesizes function of the eutectic growth velocity, as can be seen in Figure 16a. Figure 16a also synthesizes the the experimental tendency (dot line) from the study by Kaya [23] (steady-state solidification of the experimental tendency (dot line) from the study by Kaya [23] (steady-state solidification of the Al-12.6wt %Si-2.0wt %Cu alloy) as a function of growth rate. It can be seen that the experimental Al-12.6wt. %Si-2.0wt. %Cu alloy) as a function of growth rate. It can be seen that the experimental eutectic growth law, derived in the present study, shows a higher slope if compared to the experimental eutectic growth law, derived in the present study, shows a higher slope if compared to the tendency of the stationary regime. This demonstrates that the unsteady-state growth of eutectic Si experimental tendency of the stationary regime. This demonstrates that the unsteady-state growth of in ternary Al–Si–Cu alloys seems to remain more sensitive to the variations in the solidification eutectic Si in ternary Al–Si–Cu alloys seems to remain more sensitive to the variations in the thermal parameters. solidification thermal parameters.

Metals 2018, 8, 784 Metals 2018, 8, x FOR PEER REVIEW

20 of 23 20 of 23

a lower Al2 Cu characterizes the alloy content compared ItIt was was seen seenthat that a lower Al2spacing Cu spacing characterizes the with alloyhigher with Cu higher Cuascontent as to those oftothe other alloy (i.e., alloying of 3.0 wt % Cu). because the multiplier of the compared those of the other alloy (i.e., alloying of 3.0 wt. %This Cu). is This is because the multiplier of experimental equations in Figure 16b decreases from from 0.60 to0.60 0.42towith in the alloy Cualloy content, the experimental equations in Figure 16b decreases 0.42increase with increase in the Cu while preserving the samethe exponent for both for experimental equations. content, while preserving same exponent both experimental equations. The exponents exponentsofofthe the reported eutectic growth a function the growth ratefound were The reported eutectic growth lawslaws as a as function of theofgrowth rate were found to be close[26,27] to −1/2 [26,27] as caninbe in the of present of to be close to −1/2 rather thanrather −3/4, asthan can − be3/4, observed theobserved present results Figureresults 16b. The Figure 16b. The −1/2established exponent is well for the regular growth in binary Al–Cu −1/2 exponent is well for theestablished regular eutectic growth in eutectic binary Al–Cu alloys. The power 2v = constant alloys. The power growthislaw v = constant is that which was by originally and growth law λ thatλ2which was originally proposed Jacksonproposed and Huntby forJackson the growth Hunt for the growth[26]. of regular eutectics [26]. of regular eutectics However, itit is is important important to to remember remember that that the the microstructure microstructure of of the the referred referred ternary ternary eutectic eutectic is is However, formed by by aa three-phase three-phase mixture mixture of of silicon and and Al22Cu Cu in in an an α-Al α-Al matrix. matrix. As As such, such, two two major major factors factors formed appear to to contribute contribute to to the the higher higher sensitivity sensitivity of of the the variations variations in in the thegrowth growth rate raterendered rendered by by aahigher higher appear exponent: (i.) the thermal instability induced by the unsteady-state regime of heat flow extraction exponent: (i.) the thermal instability induced by the unsteady-state regime of heat flow extraction during the the growth growth of of the Al22Cu Cu eutectic eutectic phase, phase, and and (ii.) the the solute-driven solute-driven unsteadiness unsteadiness due due to to the the during buildup of of Si Si rejected rejected during during the the growth growth of the AlSi eutectic. buildup The evolutions evolutions of of the the experimental experimental tensile tensile mechanical mechanical properties properties versus versus the the primary primary dendritic dendritic The spacingare areshown shownininFigure Figure σu —ultimate tensile strength, (b)—yield σy=0.2 —yield strength, spacing 17 17 for:for: (a) σ(a) u—ultimate tensile strength, (b) σy=0.2 strength, (c) δ— (c) δ—elongation-to-fracture, and (d) HV—Vickers Hall–Petch-type were elongation-to-fracture, and (d) HV—Vickers hardness.hardness. Hall–Petch-type correlationscorrelations were adopted to adopted to represent of the experimental scatters. represent some of the some experimental scatters. 250

160

Yield tensile strength (MPa)

225

200

175

150

y=0.2

125 2

R =0.99 100 0.02

0.03

0.04

0.05 1/2

1/() ( µm )

0.06

130 120 110 100 90 0.02

0.07

0.04

1/()

0.05 1/2

( µm )

0.06

0.07

- 1/2

140

(c)

Al-11 wt.% Si-4.5 wt.% Cu Al-11 wt.% Si-3.0 wt.% Cu

130

(d)

120

2

R =0.97

Vickers hardness (HV)

Strain to failure (%)

,

0.03

- 1/2

Al 11 wt. Si - 3.0 wt. Cu Al 11 wt. Si - 4.5 wt. Cu  = 50.4 (11/2) +0.2

7

5 4 3 2 1 0 0.02

(b)

140

8

6

Al-11 wt.% Si-4.5 wt.% Cu Al-11 wt.% Si-3.0 wt.% Cu

150

,

Al-11wt.Si-4.5wt.Cu Al-11wt.Si-3.0wt.Cu 1/2 u = 1200 (1 ) + 140



u,

Ultimate tensile strength (MPa)

(a)

110 100 90 80 70 60

0.03

0.04

1/(1)

0.05 1/2

( µm )

- 1/2

0.06

0.07

50 0.02

0.03

0.04

0.05

0.06

(1)

0.07 1/2

0.08

(m)

0.09

0.10

0.11

0.12

-1/2

Figure17. 17.(a)(a) Ultimate, yield tensile strength, (c) elongation, (d) Vickers as a Figure Ultimate, (b) (b) yield tensile strength, (c) elongation, and (d)and Vickers hardnesshardness as a function 2 is the coefficient function of the primary the Al-11wt %Si-xwt alloys. of the primary dendriticdendritic spacing spacing for the for Al-11wt. %Si-xwt. %Cu %Cu alloys. R2 isRthe coefficient of of determination. determination.

Some aspects aspects related relatedto tothe theAl-11wt. Al-11wt %Si-3.0wt. %Si-3.0wt %Cu The combined combined Some %Cu alloy alloy may may be be synthetized. synthetized. The mass fraction of α-Al + Si is ~60%, the ternary eutectic is 9% in fraction, and the eutectic spacing is is mass fraction of α-Al + Si is ~60%, the ternary eutectic is 9% in fraction, and the eutectic spacing higher in of about 43%, as compared to the other alloy. In contrast, the Al-11wt. %Si-4.5wt. %Cu alloy is characterized by ~47% of mass fraction of α-Al + Si, and 14% of the mass fraction of the ternary

Metals 2018, 8, 784

21 of 23

higher in of about 43%, as compared to the other alloy. In contrast, the Al-11wt %Si-4.5wt %Cu alloy is characterized by ~47% of mass fraction of α-Al + Si, and 14% of the mass fraction of the ternary eutectic, as typified by a smaller eutectic spacing. Based on the mentioned data, it is possible that these changes in the different microstructure characteristics balance each other out, resulting in a similar evolution of the ultimate tensile strength and the strain-to-failure, as observed in Figure 17. Single Hall–Petch-type formulations are proposed in these cases to represent both alloys. σu and δ increase with decreasing λ1 along the length of the DS Al–Si–Cu alloy castings. This is because lower spacings contribute to a more extensive distribution of the second phases. If these hard particles are better distributed throughout the microstructure, higher strength values can be expected. The variations in λ1 were shown to be not significant to change for both the yield tensile strength and the hardness. These properties are associated with lower stresses, as compared to σu and δ. In the case of σy , only the start of plastic deformation is achieved. Hardness indentations, in turn, exhibit relatively low plastic deformation. Under such lower stresses, the Al-11wt %Si-4.5wt %Cu alloy properties (σy=0.2 and HV) are higher, as can be seen in Figure 17b,d. This is explained by the higher proportions of Fe-bearing intermetallic particles and interdendritic fine scale ternary eutectic, as compared to those related to the Al-11wt %Si-3.0wt %Cu alloy. 4. Conclusions The following conclusions can be drawn from the present experimental investigation:

•

•

•

The solidification microstructures of the Al-11wt %Si-3.0 and 4.5wt %Cu braze alloy samples were shown to be characterized by aluminum-rich dendrites enveloped by the products of the eutectic reactions; that is, the Al–Si eutectic and the Al–Si–Al2 Cu ternary eutectic. The fraction of phases forming the Al-11wt %Si-3.0 and -4.5wt %Cu alloys changed, respectively: from 9% to 6% of α-Al, followed by nearly 50% and 41% of the solid fraction of Si; after that, the growth of 32% and 39% of the solid fraction related to the AlFeSi phase before the eutectic reaction occurred in the remaining liquid. Accumulating the proportions of α-Al with that for the binary eutectic Si resulted in ~60% for the Al-11wt %Si-3.0wt %Cu alloy, while less than 50% of fraction of these phases was associated with the Al-11wt %Si-4.5wt %Cu alloy. Large dendritic variations in scale were shown to occur from the bottom to the top of the directionally solidified castings of both examined alloys, associated with a wide range of experimental solidification growth rates and cooling rates. This permitted the establishment of power function growth laws relating the primary (λ1 ), secondary (λ2 ), and tertiary (λ3 ) dendritic spacings as a function of the cooling rate (primary, tertiary) and the growth rate (secondary): Al-11wt %Si-3.0wt %Cu λ1 = 265 T˙ L −1/2 λ2 = 11 vL −2/3 λ3 = 12 T˙ L −1/4

•

Al-11wt %Si-4.5wt %Cu λ1 = 500 T˙ L −1/2 λ2 = 16 vL −2/3 λ3 = 17 T˙ L −1/4 ,

where λ1;2;3 (µm), vL (mm/s), and T˙ L (K/s). This means that an increase in Cu alloying of 50% (from 3.0 to 4.5 wt %) was shown to be associated with an increase in λ1 , λ2 , and λ3 by 88%, 45%, and 42%, respectively, for a given value of T˙ L or vL . The experimental evolutions of the Si spacing, λSi , and the Al2 Cu spacing, λAl2Cu , as a function of binary eutectic growth velocity, vBE , and as function of the ternary eutectic growth velocity, vTE , were experimentally determined, and experimental growth laws in the form of power functions established: Al-11wt %Si-3.0wt %Cu λSi = 1.4 vBE −1.0 λAl2Cu = 0.60 vTE −3/4

Al-11wt %Si-4.5wt %Cu λSi = 1.4 vBE −1.0 λAl2Cu = 0.42 vTE −3/4

Metals 2018, 8, 784

•

22 of 23

where λSi;Al2Cu (µm); vBE;TE (mm/s). The experimental evolutions of ultimate tensile strength (σu ) and the elongation-to-fracture (δ) were shown to be affected by the length scale of the dendritic phase, and they were related to λ1 by a single Hall–Petch-type correlation for both examined alloys; Al-11wt %Si-3.0wt %Cu, and Al-11wt %Si-4.5wt %Cu: 1

σu = 1220 (1/λ1 2 ) + 140 1

δ = 50.4 (1/λ1 2 ) + 0.2 where σu (MPa) and δ (%). Author Contributions: B.M.C.D. completed his Master’s degree on solidification in ternary Al-Si-Cu alloys. L.F.G. contributed to the SEM analysis. J.E.S. contributed to the supervision and funding acquisition; and A.G. contributed to the final analysis and revision of the writing. Acknowledgments: The authors are grateful to FAPESP (São Paulo Research Foundation, Brazil: grant 2017/12741-6) and CNPq—National Council for Scientific and Technological Development, for their financial support. Conflicts of Interest: The authors declare no conflict of interest.

References 1.

2.

3.

4. 5.

6. 7. 8. 9. 10. 11. 12.

13.

14.

Choi, S.W.; Kim, Y.M.; Lee, K.M.; Cho, H.S.; Hong, S.K.; Kim, Y.C.; Kang, C.S.; Kumai, S. The effects of cooling rate and heat treatment on mechanical and thermal characteristics of Al–Si–Cu–Mg foundry alloys. J. Alloys Compd. 2014, 617, 654–659. [CrossRef] Cerri, E.; Evangelista, E.; Spigarelli, S.; Cavaliere, P.; Dericcardis, F. Effects of thermal treatments on microstructure and mechanical properties in a thixocast 319 aluminum alloy. Mater. Sci. Eng. A 2000, 284, 254–260. [CrossRef] Costa, T.A.; Dias, M.; Gomes, L.G.; Rocha, O.L.; Garcia, A. Effect of solution time in T6 heat treatment on microstructure and hardness of a directionally solidified Al-Si-Cu alloy. J. Alloys Compd. 2016, 683, 485–494. [CrossRef] Ponweiser, N.; Richter, K.W. New investigation of phase equilibria in the system Al-Cu Si. J. Alloys Compd. 2012, 512, 252–263. [CrossRef] [PubMed] Costa, T.A.; Moreira, A.L.; Moutinho, D.J.; Dias, M.; Ferreira, I.L.; Spinelli, J.E.; Rocha, O.L.; Garcia, A. Growth direction and Si alloying affecting directionally solidified structures of Al–Cu–Si alloys. Mater. Sci. Technol. 2015, 31, 1103–1112. [CrossRef] Kurz, W.; Fisher, D.J. Fundamentals of Solidification, 4th ed.; Trans Tech Publications Ltd.: Aedermannsdorf, Switzerland, 1992; pp. 1–16, ISBN 0-87849-522-3. Zeren, M.; Karakulak, E.; Gümüs, S. Influence of Cu addition on microstructure and hardness of near-eutectic Al-Si-xCu-alloys. Trans. Nonferrous Met. Soc. China 2011, 21, 1609–1702. [CrossRef] Garcia, A. Solidificação: Fundamentos e Aplicações, 2nd ed.; Editora da Unicamp: Campinas, Brazil, 2007; pp. 1–399, ISBN 978-85-268-0782-2. Silva, B.L.; Garcia, A.; Spinelli, J.E. Complex eutectic growth and Bi precipitation in ternary Sn-Bi-Cu and Sn-Bi-Ag alloys. J. Alloys Compd. 2017, 691, 600–605. [CrossRef] Elliot, R. Eutectic solidification. Int. Metals. Rev. 2013, 22, 161–186. [CrossRef] Tiller, W.A. Liquid Metals and Solidification, 1st ed.; ASM: Cleveland, OH, USA, 1958; pp. 276–318. Reyes, R.V.; Bello, T.S.; Kakitani, R.; Costa, T.A.; Garcia, A.; Cheung, N.; Spinelli, J.E. Tensile properties and related microstructures aspects of hypereutectic Al-Si alloys directionally solidified under different melt superheats and transient heat flow conditions. Mater. Sci. Eng. A 2017, 685, 235–243. [CrossRef] Kaya, H.; Çadirli, E.; Gündüz, M.; Ulgen, A. Effect of the temperature gradient, growth rate, and the interflake spacing on the microhardness in the directionally solidified Al-Si eutectic alloy. J. Mater. Eng. Perform. 2003, 12, 544–551. [CrossRef] Hajkowski, J.; Ignaszak, Z. Experimental and simulation tests on the impact of the conditions of casting solidification from AlSi9Cu3 alloy on their structure and mechanical properties. Arch. Foundry Eng. 2018, 18, 167–175. [CrossRef]

Metals 2018, 8, 784

15. 16.

17. 18.

19. 20. 21. 22. 23. 24.

25. 26. 27.

23 of 23

Bray, J.W. ASM Handbook Volume 2: Properties and Selection: Nounferrous Alloys and Special-Purpose Materials, 10th ed.; ASM Intl.: Cleveland, OH, USA, 1990; p. 148, ISBN 0-87170-378-5. Ceschini, L.; Boromei, I.; Morri, A.; Seifeddine, S.; Svensson, I.L. Microstructure, tensile and fatigue properties of the Al-10%Si-2%Cu alloy with different Fe and Mn content cast under controlled conditions. J. Mater. Process. Technol. 2009, 209, 5669–5679. [CrossRef] Wang, E.R.; Hui, X.D.; Wang, S.S.; Zhao, Y.F.; Chen, G.L. Improved mechanical properties in cast Al–Si alloys by combined alloying of Fe and Cu. Mater. Sci. Eng. A 2010, 527, 7878–7884. [CrossRef] Okayasu, M.; Ohkura, Y.; Takeuchi, S.; Takasu, S.; Ohfuji, H.; Shiraishi, T. A study of the mechanical properties of an Al–Si–Cu alloy (ADC12) produced by various casting processes. Mater. Sci. Eng. A 2012, 543, 185–192. [CrossRef] Chang, S.Y.; Tsao, L.C.; Li, T.Y.; Chaung, T.H. Joining 6061 aluminum alloy with Al–Si–Cu filler metals. J. Alloys Compd. 2009, 488, 174–180. [CrossRef] Muhrat, A.; Puga, H.; Barbosa, J. Low-temperature brazing of titanium using Al-based filler alloys. Adv. Mater. Sci. Eng. 2018, 2018, 4570120. [CrossRef] Djurdjevic, M.B.; Vicario, I. Description of hypoeutectic Al-Si-Cu alloys based on their known chemical compositions. Rev. Metal. 2013, 49, 340–350. [CrossRef] Darlapudi, A.; Mcdonald, S.D.; Stjohn, D.H. The influence of ternary Cu additions on the nucleation of eutectic grains in a hypoeutectic Al-10 wt % Si alloy. J. Alloys Compd. 2015, 646, 699–705. [CrossRef] Kaya, H.; Aker, A. Effect of alloying elements and growth rates on microstructure and mechanical properties in the directionally solidified Al-Si-X alloys. J. Alloys Compd. 2017, 694, 145–154. [CrossRef] Canté, M.V.; Spinelli, J.E.; Cheung, N.; Garcia, A. The correlation between dendritic microstructure and mechanical properties of directionally solidified hypoeutectic Al-Ni alloys. Met. Mater. Int. 2010, 16, 39–49. [CrossRef] Rosa, D.M.; Spinelli, J.E.; Ferreira, I.L.; Garcia, A. Cellular growth during transient directional solidification of Pb–Sb alloys. J. Alloys Compd. 2006, 422, 227–238. [CrossRef] Jackson, K.A.; Hunt, J.D. Lamellar and rod eutectic growth. Trans. Metall. Soc. AIME 1966, 236, 1129–1142. [CrossRef] Morando, C.; Fornaro, O. Morphology and phase formation during the solidification of Al-Cu-Si and Al-Ag-Cu ternary eutectic systems. Mater. Res. 2018, 21, E20170930. [CrossRef] © 2018 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/).