Heat-flow parameters affecting microstructure and mechanical properties of Al–Cu and Al–Ni alloys in directional solidification: an experimental comparative study

1 Introduction

Mechanical properties of metals and alloys are significantly affected by their microstructural state. Therefore, the study of their structure–property relationships have great technological importance. The processing path of a material is responsible for controlling its solidification structure and the development of defects that will ultimately determine the properties and performance of the material [1]. Solidification processing plays a central role in the production of metals and alloys, providing the first opportunity to influence microstructural development and mitigate defects that may result in premature failure of castings. Different processing conditions, i.e., cooling rate, thermal gradient and solidification speed, yield a variety of microstructural patterns, ranging from planar to cellular or to dendrite. Processing conditions also strongly influence the outcome of the cellular, dendrite arm spacing and the formation of a grain boundary in polycrystalline materials [2], [3], [4], [5]. Arm spacing is an important microstructural parameter resulting from the solidification process, as it exerts a significant effect on the properties of the castings. This parameter affects the microsegregation that takes place between cellular or dendrite arms, having a significant influence on mechanical behavior [6]. On the other hand, during solidification of alloys, the thermal parameters (thermal gradient, solidification speed and cooling rate) can be controlled through metal/mold heat transfer, chill thickness or by its build material. Hence, the appropriate control of thermal parameters offers an opportunity window in microstructural development during the solidification process. From a technological perspective, eutectic and off-eutectic alloys with low melting points (e.g., aluminum–copper, Al–Cu, or aluminum–nickel, Al–Ni) offer interesting mechanical properties. Intrinsic characteristics of Al-based alloys make them widely adopted in a variety of industrial applications. Addition of other elements, such as copper and nickel, provide, the higher range of strength needed for many industrial applications. Mechanical properties of Al–Cu alloys depend on solidification processing conditions and also on solute content, which is added intentionally to increase strength, hardness, fatigue endurance, and machinability. Aluminum alloys with solute content from 4 to 10 wt.% Cu were the first widely used in industrial applications. It is worth mentioning that success of these alloys also lies in the precipitation hardening phenomenon, which is in great part responsible for the final mechanical properties. One of the most detrimental physical mechanisms to the homogeneity and quality of a cast microstructure is gravity-induced convection. The influence of convection on microstructural selection has been acknowledged by several researchers [7], [8], [9]. It is known to play an important role in solidification of Al–Cu and Al–Ni alloys. In these, the rejected solute, which is heavier than the Al solvent, behaves differently, depending upon the growth direction of the solid phase [10, 11]. When the direction of solidification is upward, the melt rich in segregated solute flows downward. Convection, then, is mostly confined between the dendrite arms, since the thermal gradient tends to stabilize density stratification in the melt [12, 13]. This may lead to dendritic fragmentation, to accommodate high solute concentrations in the interdendritic regions [14, 15]. On the other hand, downward solidification may yield strong and unsteady convection, due to system instability with respect to heat transfer and segregated solute from the solid region. Ultimately, the convection induced by gravity often leads to a critical solute segregation and microstructural heterogeneity [16]. In the past few decades, emergence of directional solidification furnaces paved the way for experimental investigations of solidification in technologically relevant metal alloys [17], [18], [19], [20]. The use of this system may arguably be regarded as an important methodology in the study of solidification and microstructural morphology since the 1960s, when transparent organic analogs were adopted [21]. Here, we adopted a water-cooled system with upward directional solidification to investigate the growth of microstructures in binary aluminum alloys. We also highlight the effects of copper and nickel additions on dendrite-arm spacing and on mechanical properties (microhardness and modulus of elasticity). The experimental cooling curves allowed thermal parameters such as cooling rate, solidification speed, thermal gradient and local solidification time to be experimentally determined and correlated with the mechanical properties and dendrite-arm spacing.

2 Primary dendritic spacing models

Arm spacing is an important characteristic of as-cast microstructures as it can have a marked effect on the relationship between structure and mechanical properties of the materials and their alloys. Many experimental and theoretical models have been proposed, which focused on selecting primary and secondary dendrite arm spacing [3, 22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38]. Among the models existing in the literature, it is worth highlighting those proposed by Hunt and Lu [23], concerned with primary dendritic arm spacings, and Bouchard and Kirkaldy [22], which contemplated primary and secondary spacings. Both models assume solidification under transient heat flow conditions, which is the focus of the present paper. Hunt and Lu [23] derived analytical expressions which fit results from a numerical model describing non-steady-state array growth of an axisymmetric cell or dendrite. It was found that the primary dendrite spacings can be adequately approximated by the following expression:

where

Bouchard and Kirkaldy [22], in turn, have derived heuristically a spacing formula for steady-state solidification which, modified by a dimensionless leading coefficient a₁, has proved useful for the unsteady regime. This semi-empirical expression is

Kurz and Fisher [25] have derived a primary spacing equation of their own which, contrary to Bouchard and Kirkaldy’s, applies to steady-state conditions. The theorical expression for determination of primary dendritic-arm spacing is given in Equation (7).

where λ₁ represents the primary dendritic arm spacing, Γ is the Gibbs–Thomson coefficient, m_l the liquidus line slope, k_e the equilibrium partition coefficient, C₀ the alloy composition, D _l the solute diffusivity in the liquid region, ΔT_s the difference between the liquidus and solidus temperatures, S _P the solidification speed, G the temperature gradient in front of the liquidus isotherm, G_oɛ a characteristic parameter (600 × 6 K cm⁻¹), and a₁ the primary dendrite-calibrating factor. As discussed by Canté et al. [29], since the spacings proposed in the Hunt–Lu model (Equation (1)) refer to the radius rather than to the more commonly measured diameter, and since these spacings are minimum spacings, the values need to be multiplied by 2–4, for comparison with measured spacings. The thermophysical properties of the alloys considered in present work are summarized in Table 1, according to References [29, 30] and determined from Thermo-Calc with aluminum database v4.0.

3 Experimental procedure

First, cooling curves under low cooling rate were determined for both alloys considered, i.e., no cooling system was used during these experiments. Previous works state that the liquidus and solidus temperatures can be determined by said methodology [31]. The aluminum alloys were placed inside a graphite crucible and were then melted by using a muffle furnace at 850 °C. The experimental rig for low cooling rate was developed by using a wooden chamber with a thick insulating blanket to achieve a low cooling rate. The molten aluminum alloys were then placed into this rig and allowed to solidify. The cooling rate condition was conducted by allowing the molten to cool naturally at room temperature. The temperature profile of the melt was determined by using one K-type thermocouple immersed about half was into the molten metal height in the graphite crucible. The thermocouple was connected to a data-logger which linked to a notebook, the data logger was set at 0.001 s intervals. Two binary Al-based alloys containing 5 wt.% Cu and 5 wt.% Ni were prepared in a muffle furnace at 850 °C, from commercially pure aluminum, i.e., 99.9 % Al. During the melting, a steel rod with a 1 mm thick layer of insulating alumina was used to ensure homogeneity of the alloys. Subsequent to melting, each alloy was cast into a directional solidification apparatus with water-cooling system, which favors high solidification velocity at regions close to the bottom of mold. Figure 1 shows the details of the directional solidification casting assembly adopted for the present work. The solidification system was designed in such a way that the heat flux is extracted through the water-cooled bottom, promoting upward directional solidification. Adoption of such an experimental apparatus for the solidification experiments, allows natural convection to be minimized, as well as solute (Cu and Ni) convection due to buoyancy forces if the rejected solute from the solid region has higher density than the alloy melt. The inner vertical surface was covered with a layer of insulating alumina to reduce the radial heat flux, and a cover made of a refractory material was used on the mold top, to minimize heat losses from the metal/air interface. A steel mold was used having an internal diameter of 50 mm and height of 140 mm. The mold bottom was closed with a steel chill with a thickness of 3 mm. The directional solidification apparatus has lateral electric heaters, so a desired superheat can be achieved before the start of the solidification experiments per se. To begin solidification, the electric heaters are shut off while, at the same time, the water-cooling system is turned on. The solidification experiments were carried out with two binary alloys, Al-5.0 wt.% Cu and Al-5.0 wt.% Ni. Both experiments were carried out with superheats of 10 % above the liquidus temperature. Temperature profiles were determined by monitoring through a bank of type K thermocouples positioned along the castings at 5, 10, 15, 20, 35, 45, 60, 85 mm from mold bottom. Temperature readings were collected by a data acquisition system (data-logger hardware) at 0.001 s intervals, in order to make for an accurate determination of the thermal variables. The readings were automatically stored on a personal computer. After the solidification experiments, each cylindrical ingot was sectioned along its vertical axis, mechanically polished with abrasive paper, and subsequently etched with an acid solution (25 ml H₂O, 2.5 ml HF, 25 ml HNO₃; 60 ml HCl) in order to reveal the macrostructure. After the macrostructural analysis, selected transverse sections of the directionally solidified specimens at 5, 10, 15, 20, 35, 45, 60, 85 mm from the mold bottom were polished and etched with a solution 0.5 % HF for micrograph examination. An Olympus Optical Microscope (Olympus Corporation, Japan) was used to produce digital images that were analyzed using the Goitaca (https://sourceforge.net/projects/goitacaeq) image processing software in order to measure dendrite spacings. Figure 2 is a representative micrograph of a transverse section from which tertiary dendrite-arm spacing (λ₃) measurements were made. It is interesting to note that the solidified alloys (Al-5.0 wt.% Cu and Al-5.0 wt.% Ni alloy) formed a dendritic microstructure along the entire length of the castings. About 250–300 tertiary dendrite-arm spacings were measured for each specimen and average dendrite-arm spacing values were measured in the transverse-section on at least 10 different regions for each specimen. Both microhardness and modulus of elasticity tests were carried out at room temperature, for which a dynamic ultra microhardness tester was used, Figure 3. Both variables were measured at least in 15–20 different regions on the transverse section of castings, using a Bercovich pyramidal indenter with a load of 50 g and a loading time of 10 s.

Figure 1:

Diagrammatic representation of experimental apparatus: (1) personal computer and data acquisition software; (2) data-logger hardware; (3) temperature control system; (4) type K thermocouples; (5) crucible; (6) melt; (7) unidirectional solidification furnace; (8) electric heaters; (9) ceramic fiber insulation; (10) steel mold; (11) steel plate; (12) water-cooling system (13) water pump.

Figure 2:

Transverse micrograph of a hypoeutectic Al-5.0 wt.% Cu alloy which illustrates measuring the tertiary dendrite-arm spacing (λ₃) of active arms.

Figure 3:

Schematic illustration of the microhardness and modulus of elasticity measurements of an Al-5.0 wt.% Cu alloy sample.

4 Results and discussion

4.1 Experimental cooling curves and thermal parameters

Figure 4a and d shows the phase diagrams for the Al–Cu and Al–Ni systems calculated by the Thermo-Calc software package, using the aluminum database v.4.0, emphasizing the liquid–solid transformation. For the Al-5.0 wt.% Cu and Al-5.0 wt.% Ni alloys, during cooling solidification begins at liquidus temperatures of 647 and 646 °C, respectively, and ends when they reach the solidus temperatures of 547.7 and 642 °C, also respectively, as indicated in Figure 4a and d. The solidification interval is determined by the liquidus and solidus temperatures (ΔT_S = T_L − T_S). As-cast alloys with wide solidification intervals tend to be susceptible to segregation during the solidification. For the Al–Cu and Al–Ni systems, the limit of solute solubility in the solid phase can be found at 5.8 wt.% Cu, with T = 547.7 °C, and 0.4 wt.% Ni, with T = 642 °C, respectively. In the hypoeutectic Al-5.0 wt.% Cu alloy, from 647 to 547.7 °C, the FCC_A1 phase will grow as a primary phase, while the liquid phase shrinks. Figure 4b and e shows corresponding experimental validation of the phase diagrams (Figure 4a and d) for both alloys examined. We introduce the cooling curves for Al-5.0 wt.% Cu and Al-5.0 wt.% Ni alloys, under a very slow solidification condition, i.e., no water-cooled solidification system was used during the experiments of solidification plotted in Figure 4b and e. For both Al-5.0 wt.% Cu and Al-5.0 wt.% Ni alloys, solidification begins at liquidus temperatures of 647 and 646 °C, respectively, as indicated by the change in the cooling curves caused by the release of latent heat. In the case of the Al-5.0 wt.% Cu alloy, the cooling rate was equal to 0.02 K s⁻¹, while for the Al-5.0 wt.% Ni alloy, it was 0.07 K s⁻¹. We can see that in the Al-5.0 wt.% Cu alloy case, solidification occurs in a range of temperatures wider when compared to those experimentally determined for Al-5.0 wt.% Ni alloy. In other words, for the Al-5.0 wt.% Cu alloy the solidification interval is equal to 99.3 °C, while for the Al-5.0 wt.% Ni alloy that interval is 4 °C. It is important to note that reduced solidification intervals result in better foundry properties by diminishing the tendency to defect (porosity and microshrinkage) formation and by reducing the width of the mushy zone [31]. The images shown in Figure 4c and f represent the as-cast microstructures of Al-5.0 wt.% Cu and Al-5.0 wt.% Ni alloys, respectively. At room temperature, the microstructure consists of primary dendrites (Al-rich) surrounded by a eutectic mixture of two solid phases Al₂Cu + FCC_A1 (Cu-rich), Figure 4c. On the other hand, for the Al-5.0 wt.% Ni alloy, the microstructure consists of primary dendrites (Al-rich) surrounded by a eutectic region of two solid phases Al₃Ni + FCC_L12, which is enriched in nickel solute, Figure 4f. The temperature files containing the temperature experimentally monitored by our data-logger hardware during solidification under transient heat-flow conditions in a water-cooled solidification setup are shown in Figure 5. Figure 5 shows the cooling curves obtained for both alloys (Al-5.0 wt.% Cu and Al-5.0 wt.% Ni), by means of the thermocouples placed at different positions along the casting height, in this case, a water-cooled solidification system was used during the solidification experiments. In order to achieve a standardization for both experiments, the pouring temperature was set up at about 750 °C. One can see that the temperature profiles decrease faster at regions closer to the water-cooled bottom. The cooling rate then gradually dwindles toward completion of local solidification. The experimental results of temperature versus solidification time determined by the present work present a behavior similar to those found in the literature, Sales et al. [31], Baptista et al. [32] and Batista et al. [33]. The cooling curves were used to provide a function of corresponding time of the liquidus temperature passing by each thermocouple located at specific distances from the cooled bottom. The derivative of this function with respect to time gave the solidification speed as a function of time and then, from experimental functions ( P = f t a n d S t = f ( t ) , it was possible to determine an equation for solidification speed as a function of position (S _P = f(P)), as shown in Table 2. Figure 6 shows a comparison between the experimental results concerning solidification speed as a function of position of the Al-5.0 wt.% Cu and Al-5.0 wt.% Ni alloys. In any case, the solidification speed decreases faster in the region close to the mold bottom, followed by a gradual decrease over the casting length. This behavior is well known and discussed in previous papers [31], [32], [33]. The presence of a water-cooling system favors higher solidification speeds close to the chill. Farther away, the speed tends to decrease along the casting, because of the increased thermal resistance of the solidified region. An appreciable deviation between the speed profiles can be observed in Figure 6, which indicates that Cu addition in the commercially pure aluminum caused a greater increase in solidification speed when compared with those experimentally determined for our Al-5.0 wt.% Ni alloy. In the present work, the liquidus isothermal speed obtained in the solidification experiments was strongly favored by high conductivity metals such as aluminum and copper, Figure 6. In the figures that follow, the relationships between thermal variables (cooling rate, thermal gradient and local solidification time) of both alloys and positions along the ingot are plotted for comparison purposes. As a consequence of better thermal conductivity due to Cu addition in commercially pure aluminum, the values of a couple of its thermal variables (cooling rate and thermal gradient) are significantly higher than those experimentally determined for our Al-5.0 wt.% Ni alloy, Figure 7. According to Choi et al. [39], for Al–Cu alloy, the thermal conductivity is around 148 W m⁻¹ K⁻¹, while for binary Al–Ni alloy, its thermal conductivity is equal to 110 W m⁻¹ K⁻¹, Canté et al. [29]. This difference in thermal conductivity between aluminum alloys can be favoring the deviations shown in Figures 6 and 7. Previous works have shown that, with increasing cooling rates, the secondary dendrite-arm spacing tends to decrease, Sales et al. [31]. The time required for a particular location in the casting to solidify, once nucleation has begun and until solidification is completed, is usually known as local solidification time. Its variation for the Al-5.0 wt.% Cu alloy, as well as for the Al-5.0 wt.% Ni alloy, is displayed in Figure 8. From the comparison in Figure 8, one can see an appreciable deviation between results for each of those alloys. This suggests that a wide temperature range between the liquidus and solidus temperatures of the Al-5.0 wt.% Cu alloy favors a higher local solidification time, when compared with that for the Al-5.0 wt.% Ni alloy. Local solidification time is a key factor to control microsegregation, which defines the dendritic arm spacing. An understanding of the way segregation occurs during solidification should be of great importance for material properties and in designing post-casting processes such as heat-treatments.

Figure 4:

Solidification path: (a) Partial Al–Cu phase diagram calculated by the Thermo-Calc software considering aluminum database v.4.0; (b) experimental cooling curve obtained for Al-5.0 wt.% Cu alloy without cooling system and under slow cooling rate conditions; (c) solidification microstructure of hypoeutectic Al-5.0 wt.% Cu alloy; (d) partial Al–Ni phase diagram calculated by the Thermo-Calc software considering aluminum database v.4.0; (e) experimental cooling curve obtained for Al-5.0 wt.% Ni alloy without cooling system and under slow cooling rate conditions; and (f) solidification microstructure of hypoeutectic Al-5.0 wt.% Ni alloy.

Figure 5:

Temperature versus solidification time: (a) Al-5.0 wt.% Cu alloy; and (b) Al-5.0 wt.% Ni alloy.

Table 2:

Experimental functions for position and solidification speed.

	Al-5.0 wt.% Cu	Al-5.0 wt.% Ni
Position as a function of time, P = f(t)	P = 1.22t0.89	P = 1.64t0.76
Solidification speed as a function of time, St = f(t)	St = 1.10t−0.11	St = 1.25t−0.24
Solidification speed as a function of position, Sp = f(P)	SP = 1.12p−0.12	SP = 1.46p−0.31

Figure 6:

Solidification speed (S_p) versus position (P).

Figure 7:

Thermal variables versus position (P): (a) cooling rate (R); and (b) thermal gradient (G).

Figure 8:

Local solidification time (L_st) versus position (P) along ingot.

4.2 Effect of thermal parameters on the as-cast microstructure

Admittedly, these thermal parameters (S _p , R, G and L_st) also have an effect on the solidification macrostructure. One of the aims in the present work is to study the possible effect on the solidification macrostructure of solute (Cu and Ni) addition to commercially pure aluminum. Figure 9 depicts the solidification macrostructure of both alloys examined. With respect to Figure 9, one can observe the occurrence of two types of morphology in it: columnar and equiaxed. Solidification occurred dendritically from the chill surface, forming a typically columnar structure; this indicates that the heat flux was unidirectionally oriented during the experiments. On the other hand, in regions closer to the mold top, omnidirectional growth can be observed, characterizing equiaxed grains. In the equiaxed growth case, heat flows from the crystal into the undercooled melt, while in the columnar growth heat flows from the superheated melt into the cooler solid, Kurz et al. [3]. In the present work, a dendritic structure can be observed in both cases analyzed. It is interesting to point out that equiaxed grains occurred in the liquid region ahead of the columnar zone. This way, a sharp transition from columnar to equiaxed grains (CET) can be observed. Although structural transition has occurred in both cases, it seems that the CET position was not changed during our experiments, Figure 9. The origin of columnar and equiaxed grains has been the subject of numerous experimental investigations in the field of metallurgy. The aforementioned transition has been reported to be dependent on thermal variables associated with the casting process, including heat transfer coefficients at the metal–mold interface, solidification speed, thermal gradients, cooling rates, alloy composition, and melt superheat, Willers et al. [36]. In the dendritic network shown here, the dendritic arms play an important role, since they contribute to a more extensive distribution of solutes (Cu and Ni) and, consequently, to the mechanical properties. It is important to emphasize that thermal variables (R, G and L_st) will also have an important role during solidification, since they vary continuously along the casting height. Figure 10 shows the transverse microstructures along the casting height of both alloys examined. To the right of each microstructure, we find information on sample position (P) in the casting, thermal variables (R, G and L_st) and tertiary (λ₃) and primary (λ₁) dendrite arm spacings. It can be seen that λ₃ and λ₁ increase in both cases, with the decrease in R and G. It can also be seen in Figure 10 that dendrite spacing values of the Al-5.0 wt.% Cu alloy lie above those experimentally obtained for the Al-5.0 wt.% Ni alloy. Addition of copper solute into commercially pure aluminum has a significant thickening effect on λ₃ and λ₁ when compared to addition of nickel into commercially pure aluminum. This work presents the results obtained for dendrite arm spacings versus thermal variables (R and G). The effect of R and G on λ₃ was assessed. It was found that the λ₃ is highly dependent on R and G, Figure 11. The relationship between dendrite-arm spacing (λ₃) and local solidification time (L_st) has been studied for this paper. Experimentally determined algebraic functions have been proposed for representing this relationship, which fitted well for aluminum alloys, Figure 12. Tertiary dendrite-arm spacing (λ₃) is seen to increase faster at the onset of solidification, when the lowest local solidification times are obtained. The results show that with the high solidification speeds found close to the water-cooled chill, decrease of the L_st means that the sidebranch perturbation has insufficient time to develop, which favors a more refined microstructure. The investigation also showed that the results for the Al-5.0 wt.% Cu alloy lie above those obtained for Al-5.0 wt.% Ni alloy. This is due to a wider temperature range between liquidus and solidus temperatures of the Al-5.0 wt.% Cu alloy, which favors a higher local solidification time when compared with the Al-5.0 wt.% Ni alloy, as previously discussed.

Figure 9:

As-cast grain macrostructures: (a) Al-5.0 wt.% Cu alloy; and (b) Al-5.0 wt.% Ni alloy.

Figure 10:

Photomicrographs of samples taken from transverse sections along the castings.

Figure 11:

Tertiary dendrite arm spacing (λ₃) versus thermal variables: (a) cooling rate (R); and (b) thermal gradient (G).

Figure 12:

Tertiary dendrite-arm spacing (λ₃) versus local solidification time (L_st).

4.3 Comparison between dendrite arm spacing calculated by predictive models and experimental values

Each of Figures 13 and 14 shows a comparison between the values of experimental results of the primary dendritic-arm spacing (λ₁), with predictions of theoretical values determined through predictive models, for the Al-5.0 wt.% Cu and Al-5.0 wt.% Ni alloys, respectively. These models are the Hunt–Lu model (non-steady-state solidification), represented by Equations (1) and (5), and the Bouchard–Kirkaldy model (non-steady-state solidification), given by Equation (6), with a calibration factor of a₁ = 250, and the Kurz–Fisher model (steady-state solidification). One can see, in Figures 13a and 14a, that the predictions of the Bouchard–Kirkaldy model with non-steady-state solidification overestimate the experimental results for any alloy examined. However, for any composition assessed, Figures 13a and 14a, clearly show that the experimental data lie between the maximum (x4) and minimum (x2) limits of Hunt–Lu model predictions. On the other hand, the predictions of the Kurz–Fisher model with steady-state solidification conditions have revealed, for both alloys considered (Figures 13 and 14b), a poor agreement with the experimental data. The predicted values of primary dendrite-arm spacings calculated using the Hunt–Lu, Bouchard–Kirkaldy and Kurz–Fisher models are liable to suffer deviations caused by thermophysical properties, such as diffusion coefficient, equilibrium partition coefficient and liquidus line slope. Assumptions that the partition coefficients and the liquidus line slope remain constant throughout the entire solidification process reveal themselves quite inaccurate for binary systems. The Bouchard–Kirkaldy model, in turn, is strongly dependent on the primary dendrite-calibrating factor (a₁). All these uncertainties must be considered when comparing experimental values with calculated ones.

Figure 13:

Comparison of experimental and theoretical values of primary dendritic-arm spacing as a function of position for Al-5.0 wt.% Cu alloy: (a) Hunt–Lu and Bouchard–Kirkaldy models for non-steady-state solidification; and (b) Kurz–Fisher model for steady-state solidification.

Figure 14:

Comparison of experimental and theoretical values of primary dendritic-arm spacing as a function of position for the Al-5.0 wt.% Ni alloy: (a) Hunt–Lu and Bouchard–Kirkaldy models for non-steady-state solidification; and (b) Kurz–Fisher model for steady-state solidification.

4.4 Effect of dendrite arm spacing on the mechanical properties of as-cast material

Mechanical properties such as modulus of elasticity (E) and microhardness (H) can be influenced by dendrite spacings and the addition of solute (Cu or Ni). Because of this, our solidification experiments were specifically conducted so as to investigate the influence of each of these factors. Although no relationship was observed between modulus of elasticity and tertiary dendrite-arm spacings in Figure 15a and b, it is interesting to highlight that 5.0 wt.% Cu addition (Figure 15a) favored a reduction in the average value of the modulus of elasticity by about 50 %, when compared with those reported in connection with the 5.0 wt.% Ni addition (Figure 15b). The results of microhardness versus tertiary dendrite-arm spacings (λ₃) are shown in Figure 16. Tests were carried out to investigate the influence of microstructural parameter on microhardness. The present results are in agreement with previous ones (Sales et al. [31]). Microhardness after vertical solidification of both alloys analyzed in present work, are proved to decrease with increasing dendrite arm spacing. On the other hand, solute addition of Cu into commercially pure aluminum favored a microhardness higher than that verified during upward unidirectional solidification of the Al-5.0 wt.% Ni alloy. This is due to the fact that Cu is the main hardening element in Al-based alloys, i.e., Cu solute as the main alloying element can improve this mechanical property by the precipitation of the metastable Al₂Cu phase, whereas Ni addition improves corrosion resistance and reduces the expansion coefficient of the material.

Figure 15:

Modulus of elasticity (E) versus tertiary dendrite arm spacing (λ₃): (a) Al-5.0 wt.% Cu alloy; and (b) Al-5.0 wt.% Ni alloy.

Figure 16:

Microhardness versus tertiary dendrite arm spacings (λ₃).

5 Conclusions

The results herein represent a response to information required concerning the influence of thermal parameters on dendrite-arm spacing and mechanical properties in two binary Al-based alloys containing, respectively, 5 wt.% Cu and 5 wt.% Ni. After experiments under a condition of very slow solidification, one can see that, in the case of the Al-5.0 wt.% Cu alloy, solidification occurs in a temperature range wider when compared to those experimentally obtained with the Al-5.0 wt.% Ni alloy. At the same time, a deviation between results of solidification speed can be observed. This goes to indicate that Cu addition into commercially pure aluminum caused a greater increase in solidification speed when compared to those determined for the Al-5.0 wt.% Ni alloy. The relationship between thermal variables of the Al-5.0 wt.% Cu and Al-5.0 wt.% Ni alloys and position in ingot is plotted for comparison purposes. Due to better thermal conductivity of Al-5.0 wt.% Cu alloy, the values of its thermal variables (R and G) are above those determined for the Al-5.0 wt.% Ni alloy. For both alloys, the local solidification time was also experimentally determined. This experimental result suggests that, for the Al-5.0 wt.% Cu alloy, a wider temperature range between liquidus and solidus favors a higher local solidification time as compared with results for the Al-5.0 wt.% Ni alloy. Although structural transition has occurred, from CET, in both cases analyzed here, it seems that CET position is not changed during solidification. The effect of R and G on λ₃ was evaluated. It was found that the λ₃ is highly dependent on R and G. Experimental algebraic functions have been proposed for representing the relationship between λ₃ and L_st. The corresponding plots fitted well for both aluminum alloys. The lowest tertiary dendrite spacings were observed in the Al-5.0 wt.% Ni alloy with the lowest local solidification times. One can see that primary dendrite-arm spacing values of the Al-5.0 wt.% Cu alloy lie above those experimentally obtained for the Al-5.0 wt.% Ni alloy. It means that the addition of copper solute into commercially pure aluminum has a significant thickening effect on dendrite-arm spacing as compared to addition of nickel. During directional solidification under unsteady-state conditions, primary-arm spacing (λ₁) was observed to increase the farther out the position considered, for all of the alloys examined. Analytical models from Hunt–Lu, Bouchard–Kirkaldy and Kurz–Fisher were used to determined primary dendritic-arm spacing. Good agreement was found between experimental data and the values predicted by the Hunt–Lu and Bouchard–Kirkaldy models. This because these two models assume solidification takes place under a transient heat-flow condition. By contrast, predictions made with the Kurz–Fisher model, which considers a steady-state condition, yielded poor agreement for all the alloys examined. Although no clear relationship was observed between modulus of elasticity and tertiary dendrite spacing, 5.0 wt.% Cu addition into commercially pure aluminum favored a reduction in the average value of the modulus of elasticity of about 50 % when compared with that noted in the Al-5.0 wt.% Ni alloy. The tests carried out to investigate the influence of λ₃ and solute on microhardness indicate that the microhardness resulting from vertical solidification, decrease with increasing λ₃. On the other hand, Cu solute addition into commercially pure aluminum favored a microhardness higher than that verified after upward unidirectional solidification of Al-5.0 wt.% Ni alloy.