Dendritic Segregation of Zn-Al Eutectoid Alloys

: The structural investigations performed and described in this paper are focused on both the segregation phenomenon at the dendrite level, and on the external structure in the equilibrium state. Qualitative and quantitative investigations of the Zn-Al eutectoid alloy (22% Al) structures in equilibrium and non-equilibrium states were considered and studied, revealing that, in this composition, dendritic segregation determines the eutectic transformation. Solidiﬁcation in non-equilibrium conditions controls both the eutectic transformation and the signiﬁcant variation of the chemical composition within the dendritic unit, as well as in other constituents. The most important effect of dendritic segregation is on the ratio of the different phases and on their geometry and distribution within the metallic matrix, which determines the material topology. The phenomena involved in such conditions are governed by the structural transformations, which take place during solidiﬁcation and cooling in a solid state.

With the introduction of Zn-Al alloys with increased Al content, the dendritic segregation phenomenon occurs as a peculiarity of single phase Al-Zn alloys [14]. The dimensional stability of Zn-Al alloys is significantly influenced by dendritic segregation and decreases over time. In Al-Zn systems the diffusion of the alloying elements takes place at a low velocity. This determines the significant structural modifications, compared to those in the equilibrium structure, like the presence of the eutectic transformation in the alloy with 40% (wt%) of Al [14].
At the same time, investigations concerning Zn-Al alloys with an Al content that is higher than the concentration corresponding to the eutectic point are difficult because of the lack of common opinion on the thermal equilibrium phase diagram. It is important to note that in regard to the Zn-Al system, until the first part of the 20th century, various forms of the phase diagrams [15][16][17] have been proposed. The research performed in the last 20-30 years, however, considers only two of the phase diagram options [17][18][19].
The most important differences between the two diagrams relate to the absence/presence of the peritectic transformation at 443 • C and to the lack/existence of the β-phase at high temperatures.
For the structural analysis of this research paper the phase diagram proposed by Presnyakov has been chosen, according to [18,19]. Table 1 reports the chemical composition of the binary Zn-Al alloy investigated in this research paper. An electric furnace with SiC resistance and graphite crucibles was used to develop the alloy. Gravity casting was performed and the effects of the solidification conditions were evaluated using dies made up of: (i) steel (OL, dimension 14 × 80 × 160 mm 3 ), (ii) refractory ceramic (C, Φ40 × 50 mm 2 ) and (iii) refractory ceramic cooled in the furnace (CC, Φ40 × 50 mm 2 ). The registered cooling curves were used to determine the transformation temperatures and the cooling rate (Table 2). The samples for the structural and thermal analyses have been extracted from semi-fabricated castings and, simultaneously, samples cast with steel die, of the size indicated in Figure 1a, were extracted and submitted to thermal treatment (hardening) using a specific ad-hoc tool projected and realized for this research. The schematic illustration of this device is reported in Figure 1b Figure 2 reports the microstructures obtained in the different casting conditions. Figure 3 reports the EDS results obtained along a line. The yellow line indicates where the analysis has been performed (along a line), the red line shows the Al concentration and the blue line relates to the Cu concentration.

Results and Discussion
The bottom of Figure 3 shows the variation of the Al concentration of the ZnAl22 alloy cast with steel die as a function of distance (l, expressed in µm).  Figure 2 reports the microstructures obtained in the different casting conditions. Figure 3 reports the EDS results obtained along a line. The yellow line indicates where the analysis has been performed (along a line), the red line shows the Al concentration and the blue line relates to the Cu concentration. The bottom of Figure 3 shows the variation of the Al concentration of the ZnAl22 alloy cast with steel die as a function of distance (l, expressed in µ m).   For the ZnAl22 alloy cast with steel die, the structural particularities have been illustrated in Figure 4a-c, while Table 3 reports the values of the concentrations obtained for the points indicated in Figure 4.  For the ZnAl22 alloy cast with steel die, the structural particularities have been illustrated in Figure 4a-c, while Table 3 reports the values of the concentrations obtained for the points indicated in Figure 4. For the ZnAl22 alloy cast with steel die, the structural particularities have been illustrated in Figure 4a-c, while Table 3 reports the values of the concentrations obtained for the points indicated in Figure 4.   The EDS analysis results for the thermal treated ZnAl22 alloy in the biphasic zone (S + L) are reported in Figure 5, where the same annotation of Figure 3 has been used for the analysis of the composition along a line. The lamellar shaped eutectoid phases can be seen when the alloy is cast with steel die. The thickness of the eutectoid phase is lower than 0.5-0.2 µm. The geometry and the size of this phase vary as a function of the conditions related to the development of the β-phase and are developed as a primary phase or following the eutectic reaction. The microstructural observations and the compositional analysis on micro-areas and along a line have shown that the concentration of the alloying element (Al) varies in a widely in both the solid solution α (s.s.α) and in the eutectic one. At the dendrite level, different structural regions can be observed with significant morphological differences. In the centre of dendrites of the s.s.α the concentration of Al is at its highest, reaching levels of 80-90%. In the central area, the appearance of a nearly concentric region can be observed with significant extension, whereas a characteristic of the eutectoid transformation is the appearance of the fine lamellar phase. On the surface of the dendrites there is a layer with quasi-uniform thickness, made up of two phases with polygonal shape, where the Al concentration is at about 50%. In the eutectic phase (eutectoid) of the interdendritic space, the Al concentration is minimal, 5-8%, which is close to the concentration which corresponds to the eutectic point in the thermal equilibrium phase diagram.  It was discovered that the structure is homogeneous; on the section of the β crystals the Al concentration was not significantly varied (between 22.37% and 27.14%) and they were superior (3-4%) to the average concentration of Al in the alloy. In the interdendritic regions (in the eutectic) the Al concentration was lower than the average concentration of Al in the alloy. It was discovered that the structure is homogeneous; on the section of the β crystals the Al concentration was not significantly varied (between 22.37% and 27.14%) and they were superior (3-4%) to the average concentration of Al in the alloy. In the interdendritic regions (in the eutectic) the Al concentration was lower than the average concentration of Al in the alloy.

Results and Discussion
The investigations focused on analysis of the stability of the β-phase, a metastable phase, in different thermal treatment conditions. Therefore, by reheating the thermal treated sample up to 400 • C with a holding time of 10 min and a cooling rate of 5 • C/min, the decomposition of the metastable structure was inevitable with the development of the phases (Figure 6a), according to the phase diagram.
The main particularities of the developed structure consist of a specific arrangement (like a rosette) of the eutectoid phase. When the heating temperature is above liquidus temperature (500 • C) and the cooling rate is 10 • C/min and again the development of the dendritic structure arises (Figure 6b).
(a) (b) Figure 5. EDS analysis for the ZnAl22 alloy (thermal treated condition) in the biphasic zone (S + L) for two (a-b) different areas.
It was discovered that the structure is homogeneous; on the section of the β crystals the Al concentration was not significantly varied (between 22.37% and 27.14%) and they were superior (3-4%) to the average concentration of Al in the alloy. In the interdendritic regions (in the eutectic) the Al concentration was lower than the average concentration of Al in the alloy.
The investigations focused on analysis of the stability of the β-phase, a metastable phase, in different thermal treatment conditions. Therefore, by reheating the thermal treated sample up to 400 °C with a holding time of 10 min and a cooling rate of 5 °C/min, the decomposition of the metastable structure was inevitable with the development of the phases (Figure 6a), according to the phase diagram.
The main particularities of the developed structure consist of a specific arrangement (like a rosette) of the eutectoid phase. When the heating temperature is above liquidus temperature (500 °C) and the cooling rate is 10 °C/min and again the development of the dendritic structure arises ( Figure  6b). It is very important to mention that, contrary to some scientific literature [25][26][27], the monophasic structure developed only from the β-phase and shows higher stability at room temperature, a stability which continues for some years (about four years).
Since the eutectic and eutectoid transformations play an important role in the development of the final structure, DSC and DIL analyses have been carried out to obtain some information concerning the transformation. This was done using small samples extracted from the castings It is very important to mention that, contrary to some scientific literature [25][26][27], the monophasic structure developed only from the β-phase and shows higher stability at room temperature, a stability which continues for some years (about four years).
Since the eutectic and eutectoid transformations play an important role in the development of the final structure, DSC and DIL analyses have been carried out to obtain some information concerning the transformation. This was done using small samples extracted from the castings obtained in different conditions. The heating/cooling of the samples was performed two times at different rates (5, 10 and 20 • C/min); first on the casted samples and second during the repetition of the DSC analysis on the same sample).
Analyzing the DSC curves (Figure 7), the parameters related to the phase transformations were evaluated, and, in particular, the temperature and transformation energy were considered. The DSC results are reported in Table 4 (there is a simplification of the data acquired during the analysis, where a total of 18 pairs of DSC heating/cooling curves were registered). different rates (5, 10 and 20 °C/min); first on the casted samples and second during the repetition of the DSC analysis on the same sample).
Analyzing the DSC curves (Figure 7), the parameters related to the phase transformations were evaluated, and, in particular, the temperature and transformation energy were considered. The DSC results are reported in Table 4 (there is a simplification of the data acquired during the analysis, where a total of 18 pairs of DSC heating/cooling curves were registered). The data obtained were analyzed in order to demonstrate how the casting and processing conditions and the heating/cooling during DSC analysis (the first heating/cooling step can be considered as thermal treatment) influence the phase transformations. This is relevant in Zn-Al alloys with eutectoid composition.
The first find was on the samples analyzed with DSC; the liquidus temperature registered during cooling (solidification) was 10-20 °C lower than that registered during heating (melting).
During cooling, the overcooling phenomenon was revealed, for both the eutectic and for the eutectoid transformations ( Figure 8).    The data obtained were analyzed in order to demonstrate how the casting and processing conditions and the heating/cooling during DSC analysis (the first heating/cooling step can be considered as thermal treatment) influence the phase transformations. This is relevant in Zn-Al alloys with eutectoid composition.

Qeud-r (J/g)
The first find was on the samples analyzed with DSC; the liquidus temperature registered during cooling (solidification) was 10-20 • C lower than that registered during heating (melting).
During cooling, the overcooling phenomenon was revealed, for both the eutectic and for the eutectoid transformations (Figure 8).   Observation: Meaning of the symbols used in the table: P 1 , P 2 and P 3 signify the eutectoid (eud), eutectic (eut) and liquidus (s.s.α) transformation temperatures, T l -lichidus signifies temperature and Q-phase signifies heat transformation. The second index "î" is related to the transformations which are produced during heating, while "r" is related to those which are formed during cooling. Q-phase signifies heat transformation. The second index "î" is related to the transformations which are produced during heating, while "r" is related to those which are formed during cooling. Please note that in Figures 8 and 9 the filled markers (square, triangle, circle) indicate the values related to the first round of heating/cooling, while the un-filled markers refer to the second cycle of heating/cooling in the DSC device. The samples cooled inside the DSC furnace, in all cases, show the same behavior concerning the relationship between the cooling rate and the overcooling grade of the phase transformations (eutectic and eutectoid, respectively). Overcooling for the transformation in solid state (eutectoid) is higher than that observed for the eutectic transformation and the eutectoid transformation temperature is affected only slightly by the cooling rate applied during solidification compared to the eutectic transformation temperature. For analysis of the thermal energy related to the phase transformations one has to consider the various relationships between different quantities, like latent heat of fusion (Qt), which represents the heat related to the inverse eutectic transformation, and the heat associated with the melting of the solid solution (Qα-î). Similarly, the latent heat of solidification (Qs) is created from the heat necessary for the solidification of the solid solution (Qα-r) and related to the eutectic transformation (Qeut-î). In this context, the following relations hold true: Please note that in Figures 8 and 9 the filled markers (square, triangle, circle) indicate the values related to the first round of heating/cooling, while the un-filled markers refer to the second cycle of heating/cooling in the DSC device. The samples cooled inside the DSC furnace, in all cases, show the same behavior concerning the relationship between the cooling rate and the overcooling grade of the phase transformations (eutectic and eutectoid, respectively). Overcooling for the transformation in solid state (eutectoid) is higher than that observed for the eutectic transformation and the eutectoid transformation temperature is affected only slightly by the cooling rate applied during solidification compared to the eutectic transformation temperature. Q-phase signifies heat transformation. The second index "î" is related to the transformations which are produced during heating, while "r" is related to those which are formed during cooling. Please note that in Figures 8 and 9 the filled markers (square, triangle, circle) indicate the values related to the first round of heating/cooling, while the un-filled markers refer to the second cycle of heating/cooling in the DSC device. The samples cooled inside the DSC furnace, in all cases, show the same behavior concerning the relationship between the cooling rate and the overcooling grade of the phase transformations (eutectic and eutectoid, respectively). Overcooling for the transformation in solid state (eutectoid) is higher than that observed for the eutectic transformation and the eutectoid transformation temperature is affected only slightly by the cooling rate applied during solidification compared to the eutectic transformation temperature. For analysis of the thermal energy related to the phase transformations one has to consider the various relationships between different quantities, like latent heat of fusion (Qt), which represents the heat related to the inverse eutectic transformation, and the heat associated with the melting of the solid solution (Qα-î). Similarly, the latent heat of solidification (Qs) is created from the heat necessary for the solidification of the solid solution (Qα-r) and related to the eutectic transformation (Qeut-î). In this context, the following relations hold true: For analysis of the thermal energy related to the phase transformations one has to consider the various relationships between different quantities, like latent heat of fusion (Q t ), which represents the heat related to the inverse eutectic transformation, and the heat associated with the melting of the solid solution (Q α-î ). Similarly, the latent heat of solidification (Q s ) is created from the heat necessary for the solidification of the solid solution (Q α-r ) and related to the eutectic transformation (Q eut-î ). In this context, the following relations hold true: The heat required for the eutectic transformation of the aggregation state, for both heating and cooling, is reported in Figure 10. An increase in the cooling rate during casting determines an improvement in the ratio of the eutectic transformation energy of the total melting energy and of the solidification energy, respectively. The ratio of this value is higher in the case of solidification.
The first observation relates to the energy ratio (heat) which corresponds to the inverse eutectic transformation of the total heat necessary for melting. This is about 8-12%, a ratio determined on the basis of the structural analysis [28]. This value of the eutectic ratio was considered as a starting point for the configuration of the phase diagram modified within the range of the concentration that corresponds to the eutectic and eutectoid points [28]. It is important to note that during cooling, the ratio of the eutectic transformation (11-14%) shows higher values than those registered during heating.
The energy related to the eutectoid transformation is about twofold higher than the eutectic transformation energy and it represents about 18% of the total melting heat required. Moreover, it can also be considered a function of the fraction of the constituents that make up the alloy. Therefore, for the cooling rate encountered in the case of gravity casting of the alloy in different dies (steel and two types of refractory materials), as well as in heating/cooling conditions of the samples submitted to DSC measurements (for a wide range of cooling rate), the development of about the same quantity of eutectic is observed.
The eutectoid transformation has been investigated by DIL analysis as well. For this analysis, samples have been heated and maintained in the monophasic domain, at 420 • C for 6 min, in order to obtain a homogenized structure. At the holding temperature the structure consists only of the β-phase.
Cooling has been performed in two ways: The first is controlled while the second occurs in an un-controlled way. For the controlled cooling and for the three cooling velocities considered, the contraction curves have been reported in Figure 10.

Qs = Qα-r + Qeut-r (2)
The heat required for the eutectic transformation of the aggregation state, for both heating and cooling, is reported in Figure 10. An increase in the cooling rate during casting determines an improvement in the ratio of the eutectic transformation energy of the total melting energy and of the solidification energy, respectively. The ratio of this value is higher in the case of solidification.
The first observation relates to the energy ratio (heat) which corresponds to the inverse eutectic transformation of the total heat necessary for melting. This is about 8-12%, a ratio determined on the basis of the structural analysis [28]. This value of the eutectic ratio was considered as a starting point for the configuration of the phase diagram modified within the range of the concentration that corresponds to the eutectic and eutectoid points [28]. It is important to note that during cooling, the ratio of the eutectic transformation (11-14%) shows higher values than those registered during heating.
The energy related to the eutectoid transformation is about twofold higher than the eutectic transformation energy and it represents about 18% of the total melting heat required. Moreover, it can also be considered a function of the fraction of the constituents that make up the alloy. Therefore, for the cooling rate encountered in the case of gravity casting of the alloy in different dies (steel and two types of refractory materials), as well as in heating/cooling conditions of the samples submitted to DSC measurements (for a wide range of cooling rate), the development of about the same quantity of eutectic is observed.
The eutectoid transformation has been investigated by DIL analysis as well. For this analysis, samples have been heated and maintained in the monophasic domain, at 420 °C for 6 min, in order to obtain a homogenized structure. At the holding temperature the structure consists only of the βphase.
Cooling has been performed in two ways: The first is controlled while the second occurs in an un-controlled way. For the controlled cooling and for the three cooling velocities considered, the contraction curves have been reported in Figure 10. On these curves, the beginning and the end of the transformation, as well as the temperature when the transformation occurs with the maximum velocity (the peak on the derivative of the contraction curve), have been indicated. On these curves, the beginning and the end of the transformation, as well as the temperature when the transformation occurs with the maximum velocity (the peak on the derivative of the contraction curve), have been indicated. Figure 11 reports the cooling curves in the case of free cooling in the DIL apparatus.  Figure 11 reports the cooling curves in the case of free cooling in the DIL apparatus. Figure 11. Cooling curves registered during DIL analysis, when the cooling was un-controlled for the gravity casted alloy and using steel die.
For the curves reported in Figure 11, the cooling rates have been analytically calculated for the range before the transformation temperature. These curves give evidence of both the over-cooling phenomenon for the eutectoid transformation and the phenomenon related to the recoalescence of the alloy (Table 5 and Figure 12).   For the curves reported in Figure 11, the cooling rates have been analytically calculated for the range before the transformation temperature. These curves give evidence of both the over-cooling phenomenon for the eutectoid transformation and the phenomenon related to the recoalescence of the alloy (Table 5 and Figure 12).  Figure 11 reports the cooling curves in the case of free cooling in the DIL apparatus. Figure 11. Cooling curves registered during DIL analysis, when the cooling was un-controlled for the gravity casted alloy and using steel die.
For the curves reported in Figure 11, the cooling rates have been analytically calculated for the range before the transformation temperature. These curves give evidence of both the over-cooling phenomenon for the eutectoid transformation and the phenomenon related to the recoalescence of the alloy (Table 5 and Figure 12).   As reported in Figure 12, as in the case of the eutectoid transformation, an increase in the cooling rate determines an increase in the overcooling grade and in the re-heating temperature during the recoalescence phenomenon. When the sub-cooling rate increases, the finest microstructure is obtained. The variation of overcooling for the eutectoid transformation in the controlled and un-controlled mode is reported in Figure 13 and it is possible to observe that the eutectoid transformation during controlled cooling takes place at lower temperatures than in the case of free cooling. As reported in Figure 12, as in the case of the eutectoid transformation, an increase in the cooling rate determines an increase in the overcooling grade and in the re-heating temperature during the recoalescence phenomenon. When the sub-cooling rate increases, the finest microstructure is obtained. The variation of overcooling for the eutectoid transformation in the controlled and uncontrolled mode is reported in Figure 13 and it is possible to observe that the eutectoid transformation during controlled cooling takes place at lower temperatures than in the case of free cooling.

Figure13.
The registered overcooling in the eutectoid transformation via the controlled and uncontrolled method during DIL analysis.

Conclusions
The research carried out has revealed that the investigated alloy solidification (the usual cooling rate during the casting of these alloys) determines, in non-equilibrium state, a pronounced segregation of Al at the level of the dendrites and the presence of the eutectic transformation. The most important effect of dendritic segregation is on the variation of the ratio of different phases and constituents with respect to the values indicated by the thermal phase diagram, as well as their form, geometry and fineness within the metallic matrix. The conditions leading to the decomposition of the phases during the eutectic and eutectoid transformation, when a hard segregation of Al occurs at the level of the dendrites, determine the structural modifications of the alloy.
The structural and thermal analysis results determine the fraction of the eutectic at about 10% in the structure of the ZnAl22 alloy with eutectoid composition. The dendritic segregation for this composition is accentuated, finding areas in the alloy's structure where the Al concentration reaches 90%. The β-phase obtained during thermal treatment (hardening) in the biphasic zone (L + S) does not show any segregation phenomenon. The monophasic structure, developed only from the βphase, shows stability for more than four years, contrary to some literature data.
During cooling, the registered ratio of heat, related to the eutectic transformation (11-14%) from the whole quantity of heat necessary for the phase transformation (liquid → solid), represents a higher level than the one obtained during heating. Eutectoid overcooling is about three times higher than that of eutectic. Eutectoid overcooling, in the case of un-controlled cooling, is lower than the one obtained during controlled cooling. In the case of non-controlled cooling of the samples, a higher cooling rate resulted in a more accentuated difference between the eutectoid transformation temperature and the recoalescence temperature.

Conclusions
The research carried out has revealed that the investigated alloy solidification (the usual cooling rate during the casting of these alloys) determines, in non-equilibrium state, a pronounced segregation of Al at the level of the dendrites and the presence of the eutectic transformation. The most important effect of dendritic segregation is on the variation of the ratio of different phases and constituents with respect to the values indicated by the thermal phase diagram, as well as their form, geometry and fineness within the metallic matrix. The conditions leading to the decomposition of the phases during the eutectic and eutectoid transformation, when a hard segregation of Al occurs at the level of the dendrites, determine the structural modifications of the alloy.
The structural and thermal analysis results determine the fraction of the eutectic at about 10% in the structure of the ZnAl22 alloy with eutectoid composition. The dendritic segregation for this composition is accentuated, finding areas in the alloy's structure where the Al concentration reaches 90%. The β-phase obtained during thermal treatment (hardening) in the biphasic zone (L + S) does not show any segregation phenomenon. The monophasic structure, developed only from the β-phase, shows stability for more than four years, contrary to some literature data.
During cooling, the registered ratio of heat, related to the eutectic transformation (11-14%) from the whole quantity of heat necessary for the phase transformation (liquid → solid), represents a higher level than the one obtained during heating. Eutectoid overcooling is about three times higher than that of eutectic. Eutectoid overcooling, in the case of un-controlled cooling, is lower than the one obtained during controlled cooling. In the case of non-controlled cooling of the samples, a higher cooling rate resulted in a more accentuated difference between the eutectoid transformation temperature and the recoalescence temperature.