In order to establish the superheat for the beginning of the horizontal solidification of the Al-3Cu-xNb alloys (x = 3 and 5wt.%), the solidification path of these alloys was simulated by the Thermo-Calc computational tool, as shown in Fig. 3, and experimental thermal analysis by DSC for one of the investigated alloys, as shown in Fig. 4a, and experimental cooling curves obtained by the methodology of Fig. 1c, as seen in Fig. 4b, were performed to contribute to the understanding of the microstructural evolution. As can be noted, both theoretical and experimental conditions converge to indicate the beginning of the phase transformation for precipitation of the Al-rich dendritic phase (Al-α), indicating the temperature approximately equal to 650 oC. It has indicated initial precipitation temperatures of the Al-α phase approximately equal to 650 oC and 645 oC for the Al-3Cu-3Nb and Al-Al-3Cu-5Nb alloys (wt.%), respectively. It was observed that after the formation of the Al-rich dendritic phase, the precipitation points agree with the temperatures indicated in all experimental curves of Fig. 4 for the formation of interdendritic-intermetallic phases second, such as the Fe and Al2Cu intermetallic compounds.
SEM/EDS microanalyses were performed for the Al-3Cu-3Nb alloy (wt.%), for two horizontally solidified samples at positions 10 and 90 mm from the heat transfer interface, as shown in Figs. 5 and 6, respectively. It is observed that the typical solidification microstructure is constituted by the following phases: Al3Nb + Al-α + (eutectic mixture of Al-α + Fe and Al2Cu intermetallic compounds), and this can also be seen in the theoretical solidification path predicted in Fig. 3b. In concomitant analysis of the thermal simulations from Fig. 3, it has led to the deduction of the occurrence of the following peritectic reaction: Liquid + Al3Nb → Al-α for temperatures above 650 oC, and it is in agreement the works in [15, 18–20, 23, 27]. XRD analysis results presented by [14] have shown Al3Nb intermetallic intensities in Al-xNb alloys (x = 0.8 and 1.2wt.%Nb).
On the other hand, in our recent study with the Al-3Cu-0.5Nb alloy (wt.%) (Cu/Nb = 6) [16], solidified under the same conditions and assumed in this work as one of the investigated alloys for comparison, the DSC analysis, experimental curves as well as the SEM/EDS micrographs showed the Nb dissolved in the Al-rich matrix, not verifying the formation of the aforementioned peritectic reaction. According to the results analyzed in Figs. 3 to 6, the TL temperatures were assumed to be the beginning of precipitation of the Al-α phase, as shown in Figs. 7a for Al-3Cu-xNb alloys (x = 3 and 5wt.%), and in [17] for Al-3Cu-0.5Nb alloy (wt.%).
Figure 7 shows the thermal data obtained during the solidification process, which were used to determine the VL and TR values, represented in Figs. 7b and 8, respectively. As expected, due to the imposition of a thermal resistance by conduction by the gradual formation of the solidification layer, higher VL and TR values at positions close to the heat transfer surface were observed.
The VL and TR effects on the secondary dendritic spacings (λ2) are shown in Figs. 9. It can be evidenced that lower λ2 values were obtained for as-cast samples distant from the cooled base of the ingot, in which higher VL and TR values are expected, and general power expressions given by λ2 = Constant.(P)0.36, λ2 = Constant.(VL)−2/3, and λ2 = Constant.(TR)−1/3 were proposed. From the analyses given in Figs. 9a, 9c and 9e, it was possible to propose a single mathematical expression for all investigated alloys, as shown in Figs. 9b, 9d and 9f, indicating that the Cu/Nb ratio (6, 1 and 0.6) did not affect the experimental secondary dendritic growth laws. As expected, lower λ2 values were obtained for higher VL and TR values. The values of exponents equal to -2/3 and − 1/3 of the experimental mathematical expressions that correlate λ2 with VL and TR, respectively, were induced in the graphics program used, and are in agreement with works cited in [11–13]. Furthermore, the index equal to -2/3 of the experimental relation VL x λ2 is absolutely equal to the value proposed by the Bouchard-Kyrkaldy mathematical model [31]. Recent work [15], the λ2 of the 6201 + NbB alloy was practically not affected by the cooling rate.
Figure 10 presents typical solidification structures in macrostructural and microstructural scales. Growth of columnar and equiaxed dendrites has allowed to define two distinct zones in the macrostructures, one consisting of columnar grains, and the other by equiaxed grains, as observed in Figs. 10b and 10c. Thus, it was observed the occurrence of a transition from columnar to equiaxed grains (CET) for a Cu/Nb ratio equal to 1 and 0.6 for higher VL and TR values. For the Cu/Nb = 6, the as-cast ingot was constituted of completely equiaxial grains, as reported in [16]. The experimental λ2 dependence on VL and TR, as shown in Fig. 9, can be evidence for obtaining fine Al-rich dendritic microstructures (low λ2). It is known that the solute, according to the mechanisms that explain the emergence of grains that make up the as-cast macrostructure, favors the emergence of equiaxed grains, contrary to the results found in the present work, but still according to these mechanisms, other variables affect grain morphology in the same way as the solute, such as growth and cooling rates of solidification, and thermal gradient [32]. On the other hand, it is also well known that these mechanisms were experimentally proven in the mid-20th century for binary alloys. Therefore, a more in-depth study is needed to investigate the concomitant effects of Nb variation and thermal solidification parameters on CTE formation for the investigated ternary alloy system.
Figure 11 shows the results of Vickers hardness measurement for all conditions of Cu/Nb ratio, and it was not possible to establish a mathematical relationship of HV = f(VL, TR and λ2), as shown in Fig. 11a. Thus, the mean values of Vickers hardness measurement for all analyzed positions were determined, and the results are shown in Fig. 11b. The hardening of the investigated alloys was observed as the Cu/Nb ratio was increased. Figure 11b also shows results of a comparative analysis with results from the literature for upward solidified Al-Cu binary alloys [14], and it can be seen that higher HV values were obtained for the ternary alloys of this work. It can be attributed to the presence of intermetallic phases additional to the microstructures, such as the presence of hard Fe and Al2Cu intermetallic phases that compose the eutectic mixtures of the multicomponent Al-3Cu-xNb alloys (x = 0.5, 3 and 5wt.%).
The results of the investigated electrical properties, measured according to Eqs. (1) to (4), are shown in Fig. 12. In general, it was observed that the resistance increased with the decrease of the Cu/Nb ratio, and as an obvious consequence predicted by Eqs. (3) and (4) the resistivity also increased, as well as the conductivity decreased with decreasing Cu/Nb ratio, as shown in Fig. 12a. It is deduced, therefore, that the best conductivity conditions are reached when Nb forms a solid solution with Al-rich matrix, i.e., for investigated content equal to 0.5% (mass fraction) (Cu/Nb = 6). It also indicated that the amount of the intermetallic Al3Nb phases in the as-cast microstructure of the Al-3Cu-xNb alloys (x = 3 and 5wt.%) plays an important role in the electrical characteristics of these alloys, acting as an obstacle to the electric current mobility, since higher ρ and lower σ were obtained for higher Nb contents.
The influence of the as-cast macrostructural characteristics, shown in Fig. 10, was evaluated on the electrical properties of the investigated alloys, and the results are presented in Fig. 12b. For this purpose, the properties were determined for an equivalent length (Leq) in the as-cast ingot. Eq. (2) was used to calculate the resistance for Leq at a position equal to 20 mm from the heat transfer interface. It was inverted for the resistivity calculation, that is, ρo.Leq = ρ.L, since ρ is inversely proportional to length. and the electrical conductivity for Leq = 20 mm was determined by means of Eq. (4). It is important to emphasize that the choice of Leq = 20 mm was due to the CET occurrence in the as-cast ingots of Al-3Cu-xNb alloys (x = 3 and 5wt.%). As expected, it can be noted that for the assumed equivalent length, the electrical properties follow the same behavior observed in Fig. 12a, where the Cu/Nb ratio also plays an important role, since it was expected that regions consisting of equiaxed grains, such as observed in the Al-3wt.%Cu-0.5wt.%Nb alloy, would act as barriers to the flow of current and, thus, reduce the electrical conductivity. Therefore, there is a predominance of the role of the Cu/Nb ratio in the electrical characteristics of the investigated alloys under the solidification conditions assumed in the present work.
In addition, it is important to emphasize that aluminum inevitably presents residual traces of some elements in its composition, mainly Fe, Ni, Cu, Si, Mg, and Cr [16], arising from its production process, which can negatively interfere in the electrical characteristics of as-cast Al alloys, designated as conductors of electricity. In fact, the measurement of electrical conductivity is even one of the possible methods of assessing the degree of purity of aluminum [33]. However, good electrical properties are not only important for aluminum in its pure state, but also for its alloys, especially in applications where good performance depends on a combination of electrical and mechanical properties. It was observed in the present study that the best combined condition of Vickers microhardness and electrical conductivity has been verified for the Cu/Nb ratio equal to 6, that is, average values of 53.8HV with a maximum equal to 60.5 HV contained between the error bars of the lowest Cu/Nb ratios (1 and 0.6), 4.0616 and 27.4156 MS/m, respectively, as noted in Figs. 11b and 12.