Upward Unsteady-State Solidification of Dilute Al–Nb Alloys: Microstructure Characterization, Microhardness, Dynamic Modulus of Elasticity, Damping, and XRD Analyses

Aluminium alloys form many important structural components, and the addition of alloying elements contributes to the improvement of properties and characteristics. The objective of this work is to study the influence of thermal variables on the microstructure, present phases, microhardness, dynamic modulus of elasticity, and damping frequency in unidirectional solidification experiments, which were performed in situ during the manufacturing of Al–0.8 Nb and Al–1.2 Nb (wt.%) alloys. Experimental laws for the primary (λ1) and secondary (λ2) dendritic spacings for each alloy were given as a function of thermal variables. For Al–0.8%wt Nb, λ1 = 600.1(Ṫ)−1.85 and λ2 = 186.1(VL)−3.62; and for Al–1.2%wt Nb, λ1 = 133.6(Ṫ)−1.85 and λ2 = 55.6(VL)−3.62. Moreover, experimental growth laws that correlate the dendritic spacings are proposed. An increase in dendritic spacing influences the solidification kinetics observed, indicating that metal/mold interface distance or an increase in Nb content lowers the liquidus isotherm velocity (VL) and the cooling rate (Ṫ). There is also a small increase in the microhardness, dynamic modulus of elasticity, and damping frequency in relation to the composition of the alloy and the microstructure.


Introduction
Materials engineering still faces major challenges in the modern world. There are a great variety of metal alloys, many of which are used due to their excellent properties. The importance of metallurgy field is evident, all metals, except for sintered parts, undergo a process of phase transformation and solidification during manufacturing, whether they are cast into shaped molds or produced into ingots. Alloys are studied with the aim to reduce weight, improve the combinations of mechanical properties, and predict their mechanical behaviour, such as damping capacity. Damping has practical engineering importance in limiting the amplitude of vibration at resonance conditions, and thereby reducing the probability of fatigue failure or it is capable to increase fatigue lives. Turbine blades, crankshafts, and overhead conductors are the typical applications where knowledge about damping capacity is crucial [1][2][3][4][5][6].
Nowadays, Nb applications in aluminum alloys continuously increase, with about 10% growth per year in the energy, automotive, and construction sectors [7]. The alloys contained in their Nb composition attract interest because of the potential to resist different types of corrosive media, from marine atmospheres to oxidizing atmospheres [8][9][10]. A schematic representation of the unidirectional solidifying device is shown in Figure 1. It is constituted by a cooling system (1, 2, and 3), a mold plate (4), and an ingot (5). The ingot is made of 314 stainless steel and has a square cross-section of 65 mm and a length of 100 mm and a 3 mm wall thickness. The mold plate was manufactured using a 1020 steel plate with 3 mm thickness. The cooling system is coupled to the mold plate to keep the temperature constant (at 23 • C) during the experiment. This device was kept inside a furnace capable of reaching a temperature of 1200 • C by means of two independently controlled zones of electrical resistance. The ingot was instrumented with seven type K thermocouples, positioned at 8, 10,14,18,22,27, and 35 mm from the refrigerated mold plate, which were connected to a data logger interface with a computer and the temperature data were acquired. Thermal data acquired in the casting were monitored during solidification with a frequency of 1 Hz. furnace for dissolution of the components and homogenization of the alloys.  Figure 1. It is constituted by a cooling system (1, 2, and 3), a mold plate (4), and an ingot (5). The ingot is made of 314 stainless steel and has a square cross-section of 65 mm and a length of 100 mm and a 3 mm wall thickness. The mold plate was manufactured using a 1020 steel plate with 3 mm thickness. The cooling system is coupled to the mold plate to keep the temperature constant (at 23 °C) during the experiment. This device was kept inside a furnace capable of reaching a temperature of 1200 °C by means of two independently controlled zones of electrical resistance. The ingot was instrumented with seven type K thermocouples, positioned at 8,10,14,18,22,27, and 35 mm from the refrigerated mold plate, which were connected to a data logger interface with a computer and the temperature data were acquired. Thermal data acquired in the casting were monitored during solidification with a frequency of 1 Hz.
The alloy was previously prepared in a muffle furnace, where the liquidus temperature (TL) was measured, and subsequently poured into the mold in the liquid state. For the experiment, the alloy was remelted inside the experimental furnace, avoiding the convection caused by the initial leak. The cooling system was triggered when the temperature of the first thermocouple reached a value 5% higher than the measured TL, ensuring that the whole alloy was in the liquid state, thus promoting solidification in a vertical and ascending manner. The mold plates were polished to ensure the same contact surface with the solidified metal for both alloys. In this way, the procedures and analyses performed resemble and corroborate with the recent solidification studies found in the literature [27][28][29][30][31][32][33][34][35][36]. In order to examined the macro and microstructures, metallographic techniques were used according to ASTM E3-11 [37]. The ingots were sectioned on a midplane, grinded, polished, and etched with a Flick solution (10 mL HF, 15 mL HCL, and 10 mL H2O) for macroscopy examination. The evaluated microstructures were taken from the center of the ingot to ensure directionality for the measurement of dendritic spacings by the method proposed by Gunduz and Çadirli [38]. Transverse sections, perpendicular to the growth direction, from the directionally solidified specimens at different positions along the ingot length were polished and etched with Keller solution (2 mL HF, 3 The alloy was previously prepared in a muffle furnace, where the liquidus temperature (T L ) was measured, and subsequently poured into the mold in the liquid state. For the experiment, the alloy was remelted inside the experimental furnace, avoiding the convection caused by the initial leak. The cooling system was triggered when the temperature of the first thermocouple reached a value 5% higher than the measured T L , ensuring that the whole alloy was in the liquid state, thus promoting solidification in a vertical and ascending manner. The mold plates were polished to ensure the same contact surface with the solidified metal for both alloys. In this way, the procedures and analyses performed resemble and corroborate with the recent solidification studies found in the literature [27][28][29][30][31][32][33][34][35][36].
In order to examined the macro and microstructures, metallographic techniques were used according to ASTM E3-11 [37]. The ingots were sectioned on a midplane, grinded, polished, and etched with a Flick solution (10 mL HF, 15 mL HCL, and 10 mL H 2 O) for macroscopy examination. The evaluated microstructures were taken from the center of the ingot to ensure directionality for the measurement of dendritic spacings by the method proposed by Gunduz and Çadirli [38].
Transverse sections, perpendicular to the growth direction, from the directionally solidified specimens at different positions along the ingot length were polished and etched with Keller solution (2 mL HF, 3 mL HCL, 5 mL HON 3 , and 190 mL H 2 O) for measurement of primary dendritic spacings and longitudinal section for measurement of secondary dendritic spacings. The image processing was carried out using the Olympus LEXT OLS 4000 confocal laser microscope (Shinjuku, Japan). At least 40 measurements were performed for each selected position for l1 and 50 measurements for l2, for both alloys. For the determination of the dynamic elastic modulus and damping frequency, the impulse excitation technique was used (Sonelastic, ATCP Physical Engineering) following ASTM E1876-09 [39]. For the microhardness test, 20 random indentations were performed on each sample with 50 gf load with an EmcoTest microdurometer (model DuraScan 20) according to the norm ASTM E384-08 [40]. X-ray diffraction (XRD, Bruker Smart Apex) was carried out over a wide range of diffraction angles (2θ = 20 • to 120 • ), with an angular pitch of 0.05 • and a counting time per point equal to 2.4 s. Figure 2a,b shows the typical cooling curves acquired during the upward directional solidification for alloys Al-0.8 Nb and Al-1.2 Nb (wt.%) and the liquidus temperature (T L ) measured for alloys were 645.8 • and 653.39 • (Celsius), respectively. It can be observed that the Al-0.8 wt.% Nb alloy solidifies more quickly than Al-1.2 wt.% Nb, probably as a function of thermal diffusivity and solute redistribution. The experimental cooling curves refer to thermocouples located at specific distances (P in mm) from the cooled metal/mold interface. mL HCL, 5 mL HON3, and 190 mL H2O) for measurement of primary dendritic spacings and longitudinal section for measurement of secondary dendritic spacings. The image processing was carried out using the Olympus LEXT OLS 4000 confocal laser microscope (Shinjuku, Japan). At least 40 measurements were performed for each selected position for l1 and 50 measurements for l2, for both alloys. For the determination of the dynamic elastic modulus and damping frequency, the impulse excitation technique was used (Sonelastic, ATCP Physical Engineering) following ASTM E1876-09 [39]. For the microhardness test, 20 random indentations were performed on each sample with 50 gf load with an EmcoTest microdurometer (model DuraScan 20) according to the norm ASTM E384-08 [40]. X-ray diffraction (XRD, Bruker Smart Apex) was carried out over a wide range of diffraction angles (2θ = 20° to 120°), with an angular pitch of 0.05° and a counting time per point equal to 2.4 s. Figure 2a,b shows the typical cooling curves acquired during the upward directional solidification for alloys Al-0.8 Nb and Al-1.2 Nb (wt.%) and the liquidus temperature (TL) measured for alloys were 645.8° and 653.39° (Celsius), respectively. It can be observed that the Al-0.8 wt.% Nb alloy solidifies more quickly than Al-1.2 wt.% Nb, probably as a function of thermal diffusivity and solute redistribution. The experimental cooling curves refer to thermocouples located at specific distances (P in mm) from the cooled metal/mold interface. By plotting the time at which the liquidus isotherm passes through a certain position against the height of the corresponding thermocouple, velocity curves of the liquidus isotherm were obtained as a function of position, as shown in Figure 3. The cooling rate (Ṫ) was obtained by dividing the temperature difference by the time difference (Ṫ = ΔT/Δt) of the two points as the liquidus isotherm passed their positions [41] according to Figure 4.  The curves of the thermal variables VL and Ṫ (Figures 4 and 5) have a decreasing profile from the moment they move away from the metal/mold interface. This phenomenon occurs as the solidified layer creates a solid/liquid interface, thereby increasing the thermal resistance of the solid, making it difficult for heat to travel toward the mold. The thermal gradient at the front of the solid/liquid interface indicates how the temperature is distributed per unit length, as shown in Figure 5.   The curves of the thermal variables VL and Ṫ (Figures 4 and 5) have a decreasing profile from the moment they move away from the metal/mold interface. This phenomenon occurs as the solidified layer creates a solid/liquid interface, thereby increasing the thermal resistance of the solid, making it difficult for heat to travel toward the mold. The thermal gradient at the front of the solid/liquid interface indicates how the temperature is distributed per unit length, as shown in Figure 5.  The curves of the thermal variables V L andṪ (Figures 4 and 5) have a decreasing profile from the moment they move away from the metal/mold interface. This phenomenon occurs as the solidified layer creates a solid/liquid interface, thereby increasing the thermal resistance of the solid, making it difficult for heat to travel toward the mold. The thermal gradient at the front of the solid/liquid interface indicates how the temperature is distributed per unit length, as shown in Figure 5.  The curves of the thermal variables VL and Ṫ (Figures 4 and 5) have a decreasing profile from the moment they move away from the metal/mold interface. This phenomenon occurs as the solidified layer creates a solid/liquid interface, thereby increasing the thermal resistance of the solid, making it difficult for heat to travel toward the mold. The thermal gradient at the front of the solid/liquid interface indicates how the temperature is distributed per unit length, as shown in Figure 5.  To ensure sequential solidification along the ingot and to prevent the growth of equiaxial grains in areas of constitutional sub-cooling within the molten liquid, a high thermal gradient is needed during the solidification process [42]. This gradient also reduces segregation and allows the materials to be used at higher operating temperatures [43]. Figure 6 shows the solidification macrostructure of alloys Al-0.8%wt Nb and Al-1.2%wt Nb. The macrostructure is composed of columnar grains. This directionality is due to the flow of heat extraction, which confirms the unidirectional solidification. The amount of Nb studied does not seem to be enough to cause grain refinement under the solidification conditions of the experiments. To ensure sequential solidification along the ingot and to prevent the growth of equiaxial grains in areas of constitutional sub-cooling within the molten liquid, a high thermal gradient is needed during the solidification process [42]. This gradient also reduces segregation and allows the materials to be used at higher operating temperatures [43]. Figure 6 shows the solidification macrostructure of alloys Al-0.8%wt Nb and Al-1.2%wt Nb. The macrostructure is composed of columnar grains. This directionality is due to the flow of heat extraction, which confirms the unidirectional solidification. The amount of Nb studied does not seem to be enough to cause grain refinement under the solidification conditions of the experiments.  Figure 7 shows the optical images of the longitudinal and transverse microstructures of the alloys Al-0.8%wt Nb and Al-1.2%wt Nb. The images are associated with thermal parameters such as, VL and Ṫ, and the dendritic spacings λ1 and λ2. The higher the velocities and cooling rates, the nearer the plate mold gives a refinement of the microstructure. Once the thermal resistance increases due to the formation of the solidified layer, the dendritic formation profile increases. As such, the primary and secondary dendritic spacings increase along the ingot; however, it is possible to identify that although the alloy with 1.2%wt Nb presents lower rates and velocities this also has smaller spacings, which leads to believe that Nb may have influenced this parameter.

Microstructure
Al 0.8%wt. Nb Al 1.2%wt. Nb  Figure 7 shows the optical images of the longitudinal and transverse microstructures of the alloys Al-0.8%wt Nb and Al-1.2%wt Nb. The images are associated with thermal parameters such as, V L andṪ, and the dendritic spacings λ 1 and λ 2 . The higher the velocities and cooling rates, the nearer the plate mold gives a refinement of the microstructure. Once the thermal resistance increases due to the formation of the solidified layer, the dendritic formation profile increases. As such, the primary and secondary dendritic spacings increase along the ingot; however, it is possible to identify that although the alloy with 1.2%wt Nb presents lower rates and velocities this also has smaller spacings, which leads to believe that Nb may have influenced this parameter. To ensure sequential solidification along the ingot and to prevent the growth of equiaxial grains in areas of constitutional sub-cooling within the molten liquid, a high thermal gradient is needed during the solidification process [42]. This gradient also reduces segregation and allows the materials to be used at higher operating temperatures [43]. Figure 6 shows the solidification macrostructure of alloys Al-0.8%wt Nb and Al-1.2%wt Nb. The macrostructure is composed of columnar grains. This directionality is due to the flow of heat extraction, which confirms the unidirectional solidification. The amount of Nb studied does not seem to be enough to cause grain refinement under the solidification conditions of the experiments.  Figure 7 shows the optical images of the longitudinal and transverse microstructures of the alloys Al-0.8%wt Nb and Al-1.2%wt Nb. The images are associated with thermal parameters such as, VL and Ṫ, and the dendritic spacings λ1 and λ2. The higher the velocities and cooling rates, the nearer the plate mold gives a refinement of the microstructure. Once the thermal resistance increases due to the formation of the solidified layer, the dendritic formation profile increases. As such, the primary and secondary dendritic spacings increase along the ingot; however, it is possible to identify that although the alloy with 1.2%wt Nb presents lower rates and velocities this also has smaller spacings, which leads to believe that Nb may have influenced this parameter.

Microstructure
Al 0.8%wt. Nb Al 1.2%wt. Nb                      Figure 14 shows the obtained microhardness values for alloys Al-0.8 Nb and Al-1.2 Nb (wt.%) as a function of the primary dendritic spacing. The microhardness measurements along the ingot did not show significant changes with the change in microstructure, remaining at an average value of 42.7 HV for Al-0.8%wt Nb alloy and 44 HV for Al-1.2%wt Nb alloy. There is no difference in the microhardness values as the volumetric amount of the intermetallic is insufficient to strengthen the microstructure.      Figure 15 shows the experimental values obtained by the impulse excitation test for the (a) modulus of elasticity, (b) the damping frequency, and (c) damping, in relation to the position of the metal/mold interface. Both the internal and external variables are directly dependent on the rate of heating and cooling, frequency, grain size, and the presence of precipitates [44][45][46]. The addition of Nb provided an approximately 70% increase in Al-1.2 wt.% Nb alloy's damping (Figure 15c) forming a larger number of Al 3 Nb as shown by XRD. It indicates that the precipitate can be the reason of this increase. The XRD analyses presented in Figure 16 were performed with the objective of characterising the crystalline phases present in the microstructure of the alloys. According to the Elliott [47] diagram, the intermetallic phase identified using XRD was Al3Nb, as shown in Figure 16. The intensity of the Al3Nb peaks in the studied alloys confirms that the Al-0.8 wt.% Nb alloy has a smaller volumetric fraction of the Al3Nb solid solution. Thus, the damping seems to be influenced by the quantity of this volumetric fraction. Furthermore, Figure 14a-c suggests that it may also influence the microhardness. However, there was no significant change in the modulus of elasticity, as shown in Figure 15a. The XRD analyses presented in Figure 16 were performed with the objective of characterising the crystalline phases present in the microstructure of the alloys. According to the Elliott [47] diagram, the intermetallic phase identified using XRD was Al 3 Nb, as shown in Figure 16. The intensity of the Al 3 Nb peaks in the studied alloys confirms that the Al-0.8 wt.% Nb alloy has a smaller volumetric fraction of the Al 3 Nb solid solution. Thus, the damping seems to be influenced by the quantity of this volumetric fraction. Furthermore, Figure 14a-c suggests that it may also influence the microhardness. However, there was no significant change in the modulus of elasticity, as shown in Figure 15a. Metals 2019, 9,

Conclusions
The Nb content in the alloy influenced the characteristics related to the solidification kinetics, indicating that with the Nb addition, the spacings were more refined with the decrease in the values of the thermal parameters ( and VL). The primary dendritic growth of the alloys Al-0.8 Nb and Al-1.2 Nb (% wt), solidified in the upward unidirectional solidification device, can be represented by the experimental laws: λ1 = 600.1( ) −1.85 and λ1 = 133.6( ) −1.85 , respectively. The experimental laws obtained for secondary dendritic growth: λ2 = 186.1(VL) −3.62 and λ2 = 55.6(VL) −3.62 , respectively for 0.8%wt Nb and 1.2%wt Nb.
The Nb content and the thermal variables had tiny influence on the microhardness values, with average values of 42.7 HV and 44 HV, respectively for alloys with 0.8%wt Nb and 1.2%wt Nb.
The increase in the volumetric fraction of Al3Nb seems to influence the damping frequency and the damping, with increases of 18% and 70%, respectively. However, no significant differences were observed in the elastic modulus (71.07 GPa and 72.16 GPa, respectively for 0.8%wt Nb and 1.2%wt Nb) indicating that further studies are needed to clarify the influence of Al3Nb on the Al-Nb alloy properties.