An Experimental Study of the Solidification Thermal Parameters Influence upon Microstructure and Mechanical Properties of Al-Si-Cu Alloys

Metals solidification involves the transformation of the molten metal back into the solid state. Solidification structures impact heavily on the final product's characteristics. The microstructure effects on metallic alloys properties have been highlighted in various studies and particularly the dendrite arm spacing influence upon the mechanical properties such as tensile strength has been reported. In the present investigation, Al-10wt%Si-2wt%Cu and Al-10wt%Si-5wt%Cu alloys were directionally solidified upward under transient heat flow conditions. The experimental results include solidification thermal parameters such as tip growth rate and cooling rate, optical microscopy, volume fraction of the eutectic mixture, primary dendritic arm spacing and ultimate tensile strength. Experimental growth laws of primary dendrite arm spacing as a function of the solidification thermal parameters are proposed. The Hall-Petch mathematical expressions were used to correlate the ultimate tensile strength as a function of the primary dendritic arm spacing. It was found that the alloy with higher copper content had a more refined structure. More refined structures had higher ultimate tensile strength values.


Introduction
The liquid transformation into a solid is probably the most important phase transformation in applications of science and engineering materials 1,2 .Currently a large part of the concepts and methods developed in support of solidification research, which is the main phenomenon that occurs during casting, can be applied on an industrial scale, allowing a visible improvement in the manufactured parts quality, making solidification studies a powerful tool for its economic potential 3 .The microstructure effects on metallic alloys properties has been highlighted in various studies and particularly, the grain size influence and dendrite arm spacing upon the mechanical properties has been reported 1,4 .There are a lack of studies about multicomponent alloys solidification in transitional arrangements on specific literature 5 .The ternary system (Aluminum, Silicon and Copper) has excellent properties as mechanical strength, low density as compared to ferrous alloys and good flowability.These properties make the system a good choice on automotive and aerospace industry, calling researchers attention 3,6 .Costa et al 7 investigated the influence of silicon on the microstructure comparing directional upward and horizontal solidification for the alloys Al-6wt%Cu-4wt%Si and Al-6wt%Cu, demonstrating that silicon alloying contributes to significant refinement of primary/secondary dendrite arm spacings.Araujo et al 8 investigated the interrelation between the secondary dendrite arm spacing and microhardness for Al-3wt%Cu, Al-3wt%Si and al-3wt%Cu-5.5wt%Sialloys in horizontal directional solidification.Vasconcelos et al 9 investigated the interconnection between microstructure and microhardness of horizontal directional solidification for Al-6wt%Cu-8wt%Si alloys.In addition to the primary dendritic arm spacing, the morphology of the eutectic mixture influences the mechanical properties.Yildirim et al 10 found that the addition of niobium to an iron aluminides modifies the eutectic mixture, so that the alloys exhibited higher compressive strength.After the heattreated, alloys exhibited ultra-high compressive strength and highly enhanced compressive fracture strain.There were no significant changes in the volumetric fractions of the phases and in the primary dendritic arm spacing after the heat treatment, being the change of the mechanical properties related to the morphology of the eutectic mixture.
In the present investigation, Al-10wt%Si-2wt%Cu (cited as 2Cu alloy) and Al-10wt%Si-5wt%Cu (cited as 5Cu alloy) alloys were directionally solidified upward under transient heat flow conditions.The experimental results include tip growth rate (V L ) and cooling rate (T R ), primary dendritic arm spacing (PDAS) and ultimate tensile strength (σ UTS ).These alloys were chosen so that only the Cu content was different, in order to analyze their influence.Controlling the thermic solidification parameters such as V L and T R , it is possible to predict the conditions to obtain the microstructures with the desired properties 11,12,13 .

Experimental Procedure
The experimental apparatus consists of a solidification furnace in cylindrical form (figure 1), with external steel casing and refractory bricks.Inside the oven there are an electrical resistors connected to a control panel and two support tubes.The inner one being made of stainless steel AISI 304 and the outer steel SAE 1020.Between these tubes there is a layer of refractory cement.An inner tube is responsible for directing the water jet to the inner surface of the plate, made of SAE 1020 steel.The thickness of the plate in the heat exchange region is 5 mm thick.In the upper surface, where the ingot is positioned, it was sanded with 1200 mesh sandpaper.The ingot was manufactured in stainless steel AISI 304 with internal diameter of 60 mm, external diameter of 76 mm and height of 160 mm.Type K thermocouples were positioned at distances, relative to the plate, of 4, 8, 12, 16, 35, 53 and 73 millimeters.The alloys were prepared in a graphite-clay crucible in a muffle-type electric furnace using commercially pure metals, which were analyzed by the X-ray spectrometry technique using a Spectro Spectromaxx Spectrometer (Table 1).Four casting leaks were performed for each of the alloys, all of them with a casting temperature of 700 ºC.The first ingot was used to analyze the solidification kinetics and preparation for micrographic analysis and the other three ingots were used to make test specimens for tensile tests.For each leak a sample was taken to check the composition of the prepared alloy.The analysis was performed by the X-ray spectrometry technique using a Bruker S2 Ranger Spectrometer (Table 2).After the pouring, the cooling started when the water jet was turned on at a flow rate of 21.6 L/min.All thermocouples were connected to a NI 9212 data acquisition module set and NI cDAQ 9171 chassis from National Instruments, which sent the collected data via USB cable to a computer, thereby allowing the collection of temperature data as a function of time at a rate of one die per second (Figure 2).The function P=f(t) was experimentally defined by the intersection of the straight line equivalent to liquidus temperature (T L ) of the alloy with the cooling curve of each thermocouple.By means of this intersection of the line the time of passage of the liquidus isotherm in that position of each thermocouple is determined.The derivative of the function P=f(t) allowed to obtain the respective experimental values for the displacement velocities of the liquidus isotherm (V L ), which corresponds to the liquid front passage through each thermocouple.The cooling rate (T R ) values for each position of the thermocouples were obtained experimentally from the intersections of the liquidus temperature line with the cooling curves of each thermocouple, through the result of the direct reading of the quotient of the temperatures before and after the temperature liquidus and the corresponding times (ΔT/Δt).
Transverse samples were taken from one of the solidified ingots for metallographic analysis at positions 5, 10, 20, 30, 40, 60 and 70 millimeters with respect to the cooling plate.These were embedded in bakelite, grinded and polished with diamond paste from 6 µm to 3 µm.The etchant used to reveal the microstructure was a 0.5% HF solution with a reaction time of approximately 22 seconds for the 2Cu alloy and 15 seconds for the 5Cu alloy.Optical microscopy was used to measure the PDAS for each selected position, and the triangle method was used 14 .At least 40 measurements were performed for each position, being performed only in the columnar region as a way to correlate this macrostructure with the thermal variables and tensile strength (Figure 3).The volume fraction of the phases observed in the metallographic examinations was calculated based on the systematic manual point count procedure described in ASTM E562 standards.The test specimens were made from three ingots, providing the tensile test of three test specimens for each position defined in Figure 4.The lower face of the ingot is the one that was in contact with theheat exchange plate.
The specimens were prepared according to ASTM E 8M standards with a diameter of 5 mm and a gage length of 25 mm.The equipment used for the test was a Kratos universal testing machine which has its own software for storing and displaying data.Figure 5 shows the sense of request of the specimen in relation to the microstructure obtained.

Results and Discussion
The liquidus temperature for the alloys under study were obtained in the Al-10%Si-XCu pseudo-binary diagram generated in the Thermocalc software version 2017a and is shown in Figure 6.It is observed that there is a difference of approximately 10 ºC between the two compositions.Figure 7 shows the macrograph of an as-cast ingot evidencing the columnar structure obtained in the solidification.This structure is located from the heat exchange surface to a distance of 70 mm.In this position, for both alloys, there is a columnar/equiaxed transition (CET).Above this distance there is an equiaxed structure.This same procedure was performed for the other ingots, which were later cut for the preparation of the specimens.
The solidification thermal parameters obtained are shown in Figure 8.It can be seen that both V L and T R decreased as far from the surface, where heat is exchanged.The 5Cu alloy has lower value variation than 2Cu alloy from V L .
For T R , the experimental points of the two alloys were very close (Figure 8 b).Thus, a single experimental equation was proposed for both alloys, evidencing that the percentage of Cu did not significantly change the T R .
It is shown in Figure 9 the correlation between PDAS (λ 1 ) and distance from heat extraction surface (P).It is possible to see that 5Cu alloy has lower λ 1 values than 2Cu alloy, showing an influence of the copper on the microstructure.For both alloys, higher position values (P) tends to increase the λ 1 values.The correlation among PDAS (λ 1 ), tip growth rate (V L ) and cooling rate (T R ) are shown in Figure 10.The value of the exponent is set to adjust the equation that relates λ 1 and T R was -0.55 proposed by Bouchard and Kirkaldy 15 for binary alloys.Several authors 16,17,18,19 have used this value for ternary alloys obtaining satisfactory results.In Figure 10 a, λ 1 values tend to decrease with increasing T R values.It is also observed that, for the 5Cu alloy, the values of T R are lower than for the 2Cu alloy.In Figure 10 b, λ 1 values tend to decrease with increasing V L values.The value of the exponent of the equation relating λ1 and V L was different from -1,1 proposed by Bouchard and Kirkaldy 15 for binary alloys.Some authors 20,21,22 have also found different values of exponent for V L in ternary alloys.It is also observed that, for the 5Cu alloy, the values of V L are lower than for the 2Cu alloy, demonstrating the influence of copper on the microstructure.
It is shown in Figure 11a, the correlation among ultimate tensile strength (σ UTS ) and the distance from heat-exchange surface.For higher position values the σ UTS value tends to decrease.To correlate σ UTS and λ 1 , the Hall-Petch equation was used.The coefficients of the equations were defined by  the least squares method.Several authors 19,23,24,25,26 used this methodology.These authors also primarily use PDAS, not the secondary dendritic arm spacing (SDAS), in columnar structures.As in this work only the columnar structure was analyzed it was decided to use λ 1 .It is also possible to verify that the higher the copper rate, the higher the values of σ UTS .
Figure 12 shows the transverse micrographs of the two alloys studied at positions 5, 20, 40 and 70 millimeters with respect to the heat extraction surface.It is observed that, for both alloys, the positions closest to the heat extraction surface have more refined structures.The 5Cu alloy presents lower λ 1 values than the 2Cu alloy in all positions, evidencing the influence of copper in the refinement of the microstructure.Because it was more refined, it was impossible to measure EDP at the 5 mm position for the 5Cu alloy.In this position we obtained an equiaxed structure, typical of the chill zone.The volumetric fraction of the eutectic mixture is presented in table 3. It is observed that the values of the volumetric fraction of the eutectic mixture decrease for larger distances of the surface of heat extraction, therefore for higher cooling rates, the greater the eutectic mixture fraction.This is in accordance with the literature 27 because higher cooling rates imply solidification out of balance, increasing the fraction of eutectic mixture.

Conclusions
In this experimental study on the influence of the solidification thermal parameters on the mechanical properties and microstructures of the 2Cu and 5Cu alloys, we can conclude: 1.The values of the solidification thermal parameters V L and T R decrease to positions furthest from the heat exchange surface.The V L values are higher for the 2Cu alloy, evidencing the influence of the lower copper content.The values of T R were very close to the two alloys, evidencing that the copper content does not influence this parameter.2. PDAS values tend to decrease as the distance from the heat exchange surface increases.The 5Cu alloy presents lower values of PDAS than the 2Cu alloy for all positions, evidencing that a higher copper content contributes to the refinement of the microstructure.3. When PDAS is correlated with the solidification parameters VL and TR, it is observed that, for higher values of VL and TR, the values of PDAS tend to decrease.The 5Cu alloy presents lower values in this correlation than the 2Cu alloy evidencing the influence of the higher copper content.4. When the σ UTS is correlated with the PDAS, it is observed that smaller values of PDAS present higher values of σ UTS .This fact shows that more refined structures have higher σ UTS values.The 5Cu alloy presents higher σ UTS values than the 2Cu alloy in this correlation, evidencing the influence of a higher copper content. 5.The volumetric fraction values of the eutectic mixture increase for higher cooling rates for both alloys.

Figure 1 .
Figure 1.Schematic illustration of the upward unidirectional solidification furnace

Figure 2 .
Figure 2. Upward unidirectional solidification furnace with control panel, data acquisition unit connected to thermocouples and computer to present and store experimental data.

Figure 3 .
Figure 3. Representative schematic of the technique used to quantify the PDAS in a cross section.

Figure 4 .
Figure 4. Location of test specimens for tensile testing (dimensions in millimeters).

Figure 5 .
Figure 5. Illustrative drawing of the specimen as a function of the microstructure, direction of growth and direction of application of the tensile test 15 .

Table 1 .
Chemical composition of the raw materials in weight %

Table 2 .
Chemical composition of ingots in weight %

Table 3 .
Volume fraction of the eutectic mixture