Journal Information
Vol. 8. Issue 1.
Pages 1121-1131 (January - March 2019)
Download PDF
More article options
Vol. 8. Issue 1.
Pages 1121-1131 (January - March 2019)
Original Article
DOI: 10.1016/j.jmrt.2018.09.002
Open Access
Microstructure and mechanical properties of Mg–Zn–RE–Zr alloy after thixoforming
Lukasz. Rogala,
Corresponding author

Corresponding author.
, Adrianna. Kaniab, Katarzyna. Berentb, Karol. Janusa, Lidia. Lityńska-Dobrzyńskaa
a Institute of Metallurgy and Materials Science, Polish Academy of Sciences, 25 Reymonta Str., 30-059 Krakow, Poland
b AGH University of Science and Technology. Academic Centre for Materials and Nanotechnology, 30, Mickiewicza Ave., 30-059 Krakow, Poland
This item has received

Under a Creative Commons license
Article information
Full Text
Download PDF
Figures (13)
Show moreShow less
Tables (5)
Table 1. Chemical composition of EZ33A magnesium alloy ingot.
Table 2. Chemical composition of phases in the EZ33A ingot measured with SEM/EDS and TEM/EDS techniques (areas marked in 2a-c and 3a).
Table 3. Characterization of the EZ33A according to Kazakov's criteria.
Table 4. Chemical composition of different phases in the EZ33 thixo-cast measured by SEM/EDS and TEM/EDS (areas marked in Fig. 8a–c).
Table 5. Chemical composition of different phases in the EZ33A thixo-cast after T6 measured by SEM/EDS and TEM/EDS (areas marked in Fig. 10a, b).
Show moreShow less

Magnesium is particularly challenging material, when formed from liquid phase because of high flammability risk. An alternative process for casting, which eliminates above mentioned disadvantage, is thixoforming, which involves a lower temperature of process and operation in the partially solidified state. Influence of semi-solid metal processing on EZ33A magnesium alloy (Mg–Zn–RE–Zr) microstructure and mechanical properties was studied. Ingot microstructure revealed globular grains with coarse eutectic mixture consisting of Mg7Zn3RE, T-phase – RE(Mg,Zn)11 and α(Mg). Heterogeneous nucleation of magnesium solid solution allowed obtaining structure appropriate for thixoforming. Using differential scanning calorimetry, temperature of process was determined to be 622°C, which corresponded to about 30% of the liquid phase. Thixo-cast microstructure consisted of α(Mg) globular grains with a size of 76±1.1 surrounded by fine eutectic mixture in a volume of 35%. T6 heat treatment (solution at 500°C for 6h and ageing at 190°C for 33h) caused increase of grain size to 92μm and the precipitation of two kinds of phases within the α(Mg): β′1 and β′2 responsible for the increase of yield strength to 135MPa, compression strength to 383MPa and hardness to 73HV5.

Magnesium alloys
Rare earth elements
Full Text

Magnesium alloys are attractive as light weight structural materials for potential automotive and aerospace applications [1]. A number of commercial magnesium casting alloys have been designed based on the Mg–Zn binary system with additions of rare earth elements (RE). The most commonly used are Ce, La, Gd, Nd and Y [2], which eliminate coarse dendritic microstructure, tendencies for hot cracking and shrinkage typical for the pure Mg–Zn binary alloy [2,3]. Additionally, the RE elements promote age hardening response during the T6 heat treatment due to the enhanced formation of the complex β’ type precipitations [4,5]. The Ce and La included in the EZ33A alloy are responsible for its strong grain boundary strengthening as well as forming intermetallic compounds with high melting points. The addition of rare earth elements also increases the high temperature creep resistance through the release of the stable phases which make dislocation glide difficult [6]. Nevertheless, the application of Mg–Zn, or Mg–Zn–Al systems is still limited in comparison with aluminium alloys due to their hot cracking tendency, serious micro-porosity and low mechanical properties [3,7]. In addition, the magnesium alloy melt can be extremely flammable and explosive in the process of industrial production [8]. An alternative technology, which eliminates casting disadvantages, being at the same time novel and promising, is semi-solid metal processing (SSM), which applied to EZ33A magnesium alloys, can be used to obtain significantly different microstructure. SSM processing involves the thixotropic behaviour of the alloys, which is characterized by a decrease in viscosity over a period of time under a constant shear rate [9,10]. The alloys for thixoforming should have a wide semi-solid temperature range and globular microstructure surrounded by a homogeneously distributed liquid phase [9]. It was confirmed that such a process improved the mechanical properties of aluminium and magnesium alloys [8,11] as well as steels [12]. The addition of RE has a positive effect on the globular microstructure formation (required in SSM) by heterogeneous nucleation of (α)Mg [4–6]. Zinc extends the solidus–liquidus temperature range of the Mg–Zn system [13] as well as actively takes part in strengthening the alloy by solution hardening and precipitation [4,5]. The composition of EZ33A magnesium alloys combines the appropriate features of the material for SSM with very good mechanical properties thanks to good response to T6 heat treatment. Nevertheless, this kind of alloy has not been so far subjected to the application of thixoforming. There are numerous examples in the literature confirming the positive effect of thixoforming on Mg–Zn–RE alloys. Zhang et al. [14] studied the effect of RE element on the globular microstructure formation in the semi-solid state of Mg–6Zn–4Sm–0.4Zr alloy. The thixoforming and heat treatment significantly improved the mechanical properties of Mg–Zn–RE according to work [15,16]. The aim of the present work was to study the influence of semi-solid metal processing on microstructure and mechanical properties of EZ33A magnesium alloy. The detailed characterization of the thixo-cast directly after forming allowed developing the optimal parameters of T6 heat treatment. The dominant mechanisms of strengthening after achieving peak-aged hardening were explained using advanced methods of TEM.

2Experimental procedure2.1Material preparation

The commercially available EZ33A magnesium alloy ingot, supplied by Magnesium Elektron Ltd., was used as the raw material for semi-solid metal processing. Its chemical composition is presented in Table 1. The thixoforming was conducted using a specially built prototype of press. The feedstock of diameter 30mm and height 30mm was placed inside the coil of an inductive furnace and heated up to 622̊C (with the heating rate of about 100̊C/min), which corresponded to about 30% of the liquid phase (according to the DSC curve). The temperature of the feedstock was measured with an S type thermocouple. The billet was then moved to the vertical shot sleeve and forced by a piston at velocity 1.5m/s into the die cavity in the axial direction at the pressing force of 250kN. The die of HS6-5-2 steel, sprayed with BN had temperature 20̊C. A few thixo-formed cuboid shape samples (50mmx40mmx10mm) were produced in that way (Fig. 1).

Table 1.

Chemical composition of EZ33A magnesium alloy ingot.

Element [wt.%]
Zn  Ce  Nd  La  Zr  Mg 
2.3±0.1  0.9±0.3  0.8±0.1  1.2±0.1  0.4±0.05  Bal. 
Fig. 1.

EZ33A thixo-casts.

2.2Analysis of microstructure and mechanical properties

The samples taken from the longitudinal direction of the ingot and thixo-casts (directly after the process as well as after T6 treatment) were subjected to the structural characterization. The materials used for optical microscopy were polished and etched in Nital. The observations were carried out using a Leica DM IRM metallographic microscope. The quantitative and qualitative analyses were performed using the Snyder-Graff, Jeffries Planimetric method [17,18]. In order to obtain reasonable statistical results about 1000 particles were evaluated in each specimen. ImageJ program was used for calculating the size distribution and characteristic values. Microstructural examinations were also carried out using a Scanning Electron Microscope (FEI VERSA 3D FEG). A series of BSE microphotographs were taken to identify the structural components. The chemical composition of selected micro areas was established using the Energy Dispersive X-ray spectroscopy (EDS) technique with an EDAX Apollo XP spectrometer. Five measurements were made for each phase and the mean value and the standard deviation were calculated. The experiments were performed at HV = 30 kV, WD = 10 mm, mag = 500× using a CBS detector. The phase analysis was carried out with a Panalytical Empyrean X-ray diffractometer with Co Kα radiation in the scan mode 20-100 2-Theta range at the anode voltage of 40kV and current 40mA. The PDF-4+ cryptographic database (The International Center for Diffraction Data), the Panalytical Data Viewer, High Score Plus and Match programs were used to identify the phases. The Vickers hardness was measured using a Zwick/ZHU 250 tester under the load of 5kg in accordance with ASTM E92. The compression test was performed following the PN-57/H-04320 standard using an INSTRON 3382 machine and samples 3mm in diameter and 4.5mm in height.

3Results and discussion3.1Characterization of EZ33A ingot

The metallographic analysis based on scanning electron microscopy image presented in Fig. 2a showed that the EZ33A ingot consisted of α(Mg) globular grains with the circularity index of 0.9±0.01 and the average size of 65.6±1.1μm. The alloy contained additions of rare-earth elements, e.g. Ce, Nd, La (EDS point 1, Fig. 2a, Table 2), which had a significant impact on microstructure, leading to the grain refinement by the heterogeneous nucleation of α(Mg). Zr enriched precipitations were identified inside the grains (Fig. 2b, point 2, Table 2) to form the ZnZr2 phase [19]. They were also present at globule boundaries limiting microstructure coarsening. Around them, the eutectic mixture in the volume of 10.4%, and high concentration of alloying elements (light area in Fig. 2a, EDS point 3 in Fig. 2c, Table 2) suggesting the presence of complex intermetallic phases were observed.

Fig. 2.

(a) SEM-BSE microstructure of EZ33A ingot, (b) Zr enriched precipitations identified in the grains, and (c) eutectic in the interglobular area.

Table 2.

Chemical composition of phases in the EZ33A ingot measured with SEM/EDS and TEM/EDS techniques (areas marked in 2a-c and 3a).

Place of analysis  Element [wt.%/at.%]
  Zn  Ce  Nd  La  Zr  Mg 
Fig. 2a, point 11.3±0.1/  0.4±0.1/  0.2±0.1/  –  1.1±0.1/  Bal. 
0.5±0.1  0.1±0.0  >0.1±0.0    0.3±0.1   
Fig. 2b, point 22.5±1.1/  0.5±0.5/  –  –  29.0±4.8/  Bal. 
1.2±0.5  0.1±0.1      10.1±2.1   
Fig. 2c, point 312.9±1.4/  14.4±2.0/  2.8±0.7/  8.6±0.7/  –  Bal. 
6.9±1.1  3.6±0.7  0.7±0.2  2.1±0.3     
Fig. 3a, point 116.4±1.2/  17.4±0.7/  11.0±1.1/  9.3±0.5/  –  Bal. 
10.4±0.8  10.±0.9  3.2±0.3  2.8±0.6     

The TEM observations of grain boundaries revealed the presence of precipitates with the size of about 3.5μm (Fig. 3a). The electron diffraction pattern (Fig. 3b) from the area marked with ring in Fig. 3a showed (200) and (02-1) reflections which corresponded to the T phase with the [012] zone axis of the orthorhombic c-centered lattice and the parameters of a0.91nm, b1.22nm, c0.89nm. The T-phase described for the first time by Wei [4,5] was of orthorhombic c-centered structure with lattice parameters a0.96nm, b1.12nm, and c0.94nm for the Mg–8Zn–1.5RE alloy and a1.01nm, b1.16nm, c0.99nm for the Mg–5Zn–10RE one. The change of lattice parameters yielded from different amounts of Zn and Mg in the chemical composition of the T phase. It was established that the Zn content could change from 19.3 to 43.6at.% and depending on the chemical composition of the starting alloy [4,5,20–22]. Additionally, the EDS analysis of the identified phase (point 1 in Fig. 3a, Table 2), confirmed the presence of Mg, Zn and RE in amounts 45.9, 16.4 and 37.7wt.%, respectively. The XRD analysis of ingot (Fig. 4a) also revealed the predominant amount of α(Mg) at a small volume of the T-phase.

Fig. 3.

(a) TEM-BF micrograph of the area of eutectic with marked points of EDS analysis, (b) selected electron diffraction pattern of areas marked in the image (a).

Fig. 4.

X-ray diffractograms of EZ33 magnesium alloy (a) ingot, (b) thixo-cast, (c) thixo-cast after T6 heat treatment.


Fine precipitates, observed within the grains of α(Mg) solid solution in the EZ33A alloy, often occurred in the form of agglomerates. Fig. 5 presents the STEM-HAADF image together with the distribution of elements in the area with precipitates, in which two kinds of phases can be established; one of oval shape and another in the form of laths. Based on maps of the distribution of elements, it could be stated that both types of precipitates were enriched mainly in zinc and rare earth elements (Ce, La and Nd). The BF image and the analysis of electron diffractions confirmed, that the lath-like elongated precipitates of length below 1μm revealed the hexagonal structure of MgZn2 (P63/mmc) Laves phase with lattice constants a0.5221nm, c0.8567nm (Fig. 6a, b). The exemplary microstructure and the diffraction pattern of the individual particle are shown in Fig. 6a,b. They contained dissolved elements of rare earths and little amount of zirconium. Their chemical composition determined with the EDS point analysis was in wt.%: 58.7 Mg, 18.3 Zn, 13.6 Nd, 7.5 Ce, 0.7 La and 1.2 Zr. The established content indicates a higher concentration of neodymium and depletion in lanthanum. The oval particles were identified as the T phase (Fig. 6c, d), which was composed of 69.5wt.% Mg, 17.2 Zn, 5.9 Ce, 5.3 Nd, 1.2 La and 0.9 Zr. In both analyses the Mg content was too high compared with its real contribution, due to an additional signal from the matrix.

Fig. 5.

(a) STEM-HAADF image of the EZ33A ingot with large frame of mapping and surface distributions of alloy constituents: (b) Mg; (c) Zn; (d) Zr; (e) Ce; (f) La; (g) Nd.

Fig. 6.

Microstructure of EZ33A ingot (a, b) TEM-BF from area of elongated precipitations with SAEDP, (c, d) TEM-BF from area of the irregular shape of precipitation with SAEDP.

3.2Characterization of solidus–liquidus temperature range

The results of calorimetric studies carried out during heating and cooling at 15°C/min of the EZ33A alloy (ingot) in the solidus–liquidus range were graphically composed as heat flow curves with changes of liquid fraction contribution, both versus temperature in Fig. 7 a and b, respectively.

Fig. 7.

DSC results of the EZ33A ingot examination: (a) heating/cooling flow curves, (b) liquid fraction curve as a function of temperature.


Melting of α(Mg) of the EZ33A alloy was preceded with the dissolution or partial melting of two probable phases of the eutectic, which were Mg7Zn3RE and RE(Mg,Zn)11 (T-phase) (Fig. 7a bottom curve) [4–6,20,21]. The effect is visible as endothermic reactions in the range of temperatures 560–585°C. It should be noted, that the phases contained in their lattices the rare earth elements, whose contribution in the examined alloy amounted to about 2.9wt.%. That could be responsible for the local changes in their melting temperatures. The effective melting process of the alloy, which has a decisive influence on the increase of liquid phase in function of temperature, occurs in the range from 592 to 656°C. It is visible in the DSC curve as a big endothermic effect. According to the equilibrium system [13] the liquidus temperature for 2.3wt.% of Zn in Mg is 578°C, while that of solidus is 645°C. The large differences observed are due to the high rate of cooling (DSC, 15°C/min) and the contribution of rare earths elements, which have higher melting temperatures compared with magnesium (Ce – 795°C, Nd – 1010°C, La – 920°C). During the crystallization of EZ33A alloy (Fig. 7a, upper curve), the liquidus temperature was estimated to be 636°C and it was shifted by 20°C in regard to that estimated for heating, which was due to undercooling of the alloy required at crystallization. The process ended at 590°C followed with a small exothermic reaction in the temperature range 548–557°C brought about by the crystallization of low-melting phases, probably enriched with zinc and rare earths. This broadened peak was observed at relatively high temperature compared with small endothermic peaks visible during heating, which could be caused by a local segregation. However, it might also be due to the crystallization of Mg7Zn3RE and RE(Mg,Zn)11 T-phase at lower temperatures [4–6,20,21]. A strong supersaturation of α(Mg) solid solution with zinc or RE resulting in the decrease of the eutectic amount can be one more explanation. The dependence of liquid fraction in function of temperature was drawn in order to obtain the detailed description of solid–liquid range and to select a proper temperature of thixoforming which was shown in Fig. 7b. Based on the Kazakov criteria, the thermo-physical properties of the EZ33A alloy were determined for the solid–liquid range [23,24]. The obtained results are collected in Table 3. The values for ΔT40%-20% and ΔT60%-40% corresponded to the temperature interval between 40–20% and 60–40% of liquid fraction (fL), respectively (values have been chosen due to they are a typical ranges of the process window used in various kinds of SSM technology [10]). This temperature range should be wide enough to allow easier control of solid fraction during the process. Its means that small deviations in temperature (because of inaccuracy of temperature measuring instruments) should not significantly affect the changes of liquid fraction. Point A, marked in Fig. 7b, corresponding to 622°C (liquid content 30%) was taken as the temperature of thixoforming casting process.

Table 3.

Characterization of the EZ33A according to Kazakov's criteria.

Alloy  TS  T30%  T50%  TL  TLTS  ΔT40%-20%  ΔT60%-40% 
EZ33A  592  622  628.5  656  64 
3.3Analysis of EZ33A thixo-cast

The EZ33A magnesium alloy ingot was used as the feedstock for semi-solid processing to assess the influence of technology on microstructure and mechanical properties. As was confirmed in several works [25,26], SSM allows obtaining a unique structure, thanks to specific conditions of the process, e.g. high shearing during the flow of the alloy suspension, rapid cooling, exerted pressure during the solidification and, in consequence, the formation of the strong non-equilibrium state of material. The selection of parameters proved the thixotropic cast free of defects. The cavity of the mould was completely filled by the metal and the micro-porosity was not observed. The product did not require more post-processing than the removal of fine fragments of the filler residue. The SEM-BSE microstructure of EZ33A thixo-cast revealed the presence of α(Mg) globular grains with the average size of 76±1.1μm, surrounded by the eutectic mixture in the volume of 35% (Fig. 8a). The XRD analysis of the thixo-cast (Fig. 4b) confirmed the presence of α(Mg) as well as T-phase which were probably located in interglobular areas (Fig. 8b). The differences in the content of the liquid phase observed, based on the DSC (Fig. 7a, b) and microstructure analyses, presumably resulted from different cooling rates as well as inaccuracy during the temperature control. The results of EDS point analysis in the grain (area 1 in Fig. 8a) as well as in the eutectic (point 2 in Fig. 8b) are presented in Table 4. The significantly increased segregation in comparison with the EZ33A ingot is visible. It is connected with remaining of the sample in the semi-solid state which leads to the diffusion control phenomena such as the increase of average grain sizes and the decrease of concentration of RE elements in the grains, leading in consequence to the promotion of segregation. Nevertheless, it can be seen that the eutectic was refined and so was the grain size in comparison with the ingot structure. Additionally, the precipitations enriched in RE elements were identified in the center areas of grains (Fig. 8c, EDS point 3, Table 4). In the Mg–Zn–RE–Zr alloys with high contents of rare-earth the appearance of the (Mg, Zn)12RE precipitates at the grain boundaries of a secondary solid solution based on the Mg12Ce intermetallic phase [4,5] was observed in [27,29]. In the case of Mg–Zn–RE–Zr alloys with high content of zinc (ZE and EZ type of alloys) the structure was similar, however with differences in the concentration of eutectic components (Mg7Zn3RE and T-phase with formula RE(Mg, Zn)11) [27–29]. The studies of phase diagrams, carried recently on Mg–Zn–Ce system have shown, that in Mg-rich alloys, apart from the Mg solid solution, the Ce(Mg, Zn)12 phase was established, which appeared together with the Ce(Mg, Zn)11 one of tetragonal structure (space group I4/mmm) and lattice constants a1.01–1.03nm, c0.57–0.6nm [28,29]. Similarly to the case of the Ce(Mg, Zn)11 phase, the compound revealed the variable content of Zn in the range of 10.4–43.9 at.%. In the present work, the RE(Mg, Zn)11 was identified, probably due to the temperature of the process. High temperatures increase the rate of diffusion, leading to grain growth [6]. The share of eutectic phases decreased due to the thixotropic process as a result of their partial dissolution in the matrix and, as a consequence, microstructure strengthening was achieved. The separation of fine eutectics on the grain boundary was also obtained, which was a positive result of the use of thixotropic technology to increase the strength of alloy.

Fig. 8.

SEM-BSE microstructure of EZ33A thixo-cast with marked areas of EDS point analysis; (a) in Mg solid solution, (b) eutectic, and (c) precipitate.

Table 4.

Chemical composition of different phases in the EZ33 thixo-cast measured by SEM/EDS and TEM/EDS (areas marked in Fig. 8a–c).

Place of analysis  Element [wt.%/at.%]
  Zn  Ce  Nd  La  Zr  Mg 
Fig. 8a, point 11.3±0.2/  –  –  –  0.8±0.1/  Bal. 
0.5±0.1        0.6±0.1   
Fig. 8b, point 215.5±0.1/  13.4±0.1/  3.9±0.1/  6.9±0.2/  –  Bal. 
8.2±0.1  3.3±0.1  0.9±0.1  1.8±0.1     
Fig. 8c, point 34.3±1.3/  3.0±1.0/  1.3±0.5/  1.4±0.4/  –  Bal. 
1.8±0.6  0.6±0.2  0.2±0.1  0.3±0.1     
3.4EZ33 thixo-cast after thixoforming and T6

In the next stage of the study, a heat treatment of thixo-cast samples was performed. Due to the fact that the SSM microstructure was significantly different from that typically obtained after casting or plastic deformation, new conditions of T6 heat treatment (time, temperatures of solution treatment and aging) were developed. The saturation temperature was determined based on the DSC results obtained during heating (Fig. 7a). It can be noticed that the beginning stage of the melting process of intermetallic phases started at ≈560̊C. To eliminate the influence of differences in heating rates of the samples (15̊C/min in the calorimeter and about 60̊C/min in the furnace for heat treatment), which could have led to local melting, the temperature of saturation was decreased to 500̊C. The time of saturation was 6h and it was proposed based on literature [29,30], where alloys of similar chemical composition were considered. The optimal ageing temperature to produce the β’ type precipitation in the Mg–Zn system was reported to be in the range of 100–170̊C [5,27,31]. However, in the present work, 190̊C was applied, while ageing time was shortened. Next, the kinetics of the supersaturated magnesium solid solution decomposition was determined, using hardness measurements in function of time. This method is very often applied due to its sensitivity to microstructure changes [6]. Fig. 9 shows the curves of hardness changes versus ageing time (in the range of 0.15–50h) at the constant temperature of 190̊C for the EZ33A thixo-cast after solution at 500̊C for 6h. Two characteristic hardness peaks after 16h and 33h of aging at 190̊C can be seen. The precipitations formed in this range are probably coherent with the α(Mg) matrix. Additionally, another maximum can be seen (third one) after 45h at 190̊C, which was preceded by a sharp decline of hardness to 54HV. It could be because that early formed precipitations (obtained at first peak) lost the coherence with the matrix [1]. In accordance with Rokhlin [6] the aging behavior of Mg based alloys depends on types of rare earth elements used. Two main groups of rare earth metals were proposed: cerium one which included Nd, Sm, La and the yttrium group with Gd, Tb, Ho, Er and Tm. The metals have different effects on changes of hardness with time of aging. The addition of cerium group to Mg alloy leads to obtaining a significantly lower hardness than that of the yttrium group metals. Referring to the current research, the increases of hardness at beginning of aging time (after 16h) are connected with the formation of Zn enriched precipitations (which lost coherence after 36h), while second and third peak corresponding to 33h and 45h of aging at 190̊C respectively, are results of formation the RE containing phases, Two different groups of rare earths in the EZ33A thixo-cast have direct influence on its aging curve. Hence, a synergic effect of their action is observed. The positive effect of Ce on precipitation hardening in Mg–Zn based alloys has been reported in [4]. In the MgZn4Ce0.1 (wt.%) alloy the hardening rate was initially low, but its value increased after 3 hour-ageing to achieve the peak hardness of 75HV after 168h at 180°C. The differences, in comparison with the present work, are results of the higher concentration of Ce and higher temperature of ageing. What is interesting, a very weak effect of hardness changes was observed after T6 treatment in the binary system Mg2Zn, while in the Mg7Zn (all in wt.%) alloy the hardness significantly increased [4,5]. It suggested that the volume of Zn played a crucial role in the hardening effect. Moreover, microalloying the Mg–Zn alloys with e.g. Ce can lead to advantageous hardening effects [4].

Fig. 9.

Hardness of the EZ33A thixo-cast after solution at 500°C/6h versus aging time at 190°C for 0.15–50h.


Fig. 10a and b present the SEM-BSE microstructure of a cross-section of thixo-formed sample after the T6 treatment (solution at 500°C for 6h and ageing at 190°C for 33h). Large changes in the microstructure are visible in the form of a smaller amount of the secondary phase. The eutectic partially dissolved, resulting in coarsening of primary grains. The average grain size increased up to 92μm (Fig. 10a). The chemical analysis was performed in the areas marked with squares in Fig. 10a, b and the results are presented in Table 5. The content of alloying elements in the solid solution (area 1 in Fig. 10a, Table 5) is higher than in the sample directly after processing, which suggests their partial dissolution in globules, although the fine β’ precipitates resulting from ageing could not be seen at that magnification. As a result of T6 heat treatment, the T-phase and such partially dissolved in the magnesium solid solution, which was confirmed by the EDS analyses, in which smaller amounts of Zn and RE elements were observed in comparison with the thixoformed material (area 2 in Fig. 10a, Table 5). It was also confirmed by the XRD examination, in which the intensity of T-phase decreased (Fig. 4c). Spherical and rectangular shape precipitations enriched in RE elements (area 3–5, Fig. 10a, b, Table 5) as well as micro-eutectic (area 6, Fig. 10b, Table 5) were observed in the microstructure of grains. Solutioning and aging of these alloys reported in the study [27] did not bring the desired effect, because it led to the breakup of RE(Mg, Zn)11 phase and its partial dissolution. It could be connected with the coarse eutectic present in the ingot microstructure and also with a too low temperature of annealing [15,32,33]. The microstructural results are consistent with reference [26], confirming, that the T6 treatment after the SSM could influence the improvement of the mechanical properties like hardness and compression strength. Fig. 11 shows the TEM bright field image, in which small, coherent with matrix precipitations with the average thickness of 3nm and length up to 12nm are visible as dark points homogenously distributed in the α(Mg) matrix. They are responsible for strengthening of material by the precipitation hardening mechanism, which introduces stress to the α(Mg) lattice [34], thanks to the early-stage decomposition of the supersaturated magnesium solid solution. In accordance with Wei et al. [35] the precipitations observed in the present work are similar to β′1, which have a rod-like morphology. Rare earth elements have a very low solubility in α(Mg) solid solution [13]. Nevertheless, the presence of Ce addition in the alloy improves the precipitation hardening effect of the Mg–Zn based alloy [4]. The TEM bright field image was taken along [2–1–10] zone axis and the corresponding diffraction pattern as an insert (Fig. 12a) of the thixo-formed part after T6 peak-aged condition shows the presence of β′1 and β′2 transition phases. The morphology of these phases is different: the β′1 phase precipitates as rods perpendicular to (0001) α(Mg), visible in Fig. 12a as dot points, whereas the β′2 phase precipitates as plates parallel to (0001) Mg(α) with the size of 100–120nm. At the matrix orientation [0001], the β′2 particles are visible as discs [35]. The β′1 in the shape of rods and blocks are similar to Mg4Zn7 structure, while the β′2 phase in the form of plates or laths is like MgZn2 phase [5]. These transition phases transform into equilibrium β ones with the MgZn or Mg2Zn3 structures. Nevertheless, in the literature there is some controversy regarding the transition phases, e.g. according to Sturkey and Clark [36]β′1 and β′2 have structures similar to the Laves MgZn2 phase. Significant differences in the density between these two types of precipitations are visible in Figs. 11 and 12. Further investigation is required to assess the influence of treatment conditions or the role of the alloying elements on the type of formed precipitates. The electron diffraction pattern from the image area shows the reflections from α(Mg) (01–10) and (0002) at zone axis [2–1–10]. Unfortunately, due to the small size of precipitations, the reflections from them were not observed. Fig. 12b shows a high-resolution transmission electron microscopy (HRTEM) image taken of the area presented in Fig. 12a. The atomic planes of α(Mg) matrix with clearly visible thin precipitates as well as a high number of dislocations located around them indicate the presence of nano precipitates partially coherent with the matrix. Based on the lattice distances and angles between them, measured from a fast Fourier transform (FFT) obtained from the area marked with rectangular frame (as insert in Fig. 12b), the α(Mg) matrix with zone axis [2–1–10] was confirmed. Additionally, less visible reflections (marked with arrows) which corresponded to the β′1 were also present [37].

Fig. 10.

(a, b) SEM-BSE microstructure of EZ33A thixo-cast after T6 with marked EDS points of analysis.

Table 5.

Chemical composition of different phases in the EZ33A thixo-cast after T6 measured by SEM/EDS and TEM/EDS (areas marked in Fig. 10a, b).

Place of analysis  Element [wt.%/at.%]
  Zn  Ce  Nd  La  Zr  Mg 
Fig. 10, point 11.7±0.2/  0.4±0.1/  0.6±0.2/  07±0.2/  0.4±0.1/  Bal. 
0.6±0.1  0.1±0.1  0.1±0.1  0.1±0.1  0.1±0.1   
Fig. 10, point 212.2±1.1/  11.1±0.6/  2.9±0.2/  5.1±0.2/  –  Bal. 
5.9±0.1  2.5±0.1  0.6±0.1  1.2±0.1     
Fig. 10, point 33.3±1.3/  2.0±1.0/  1.1±0.4/  0.9±0.4/  –  Bal. 
1.3±0.6  0.4±0.2  0.2±0.1  0.2±0.1     
Fig. 10, point 47.1±1.0/  2.4±0.9/  0.9±0.4/  0.4±0.4/  –  Bal. 
2.8±0.5  0.4±0.2  0.2±0.1  0.1±0.1     
Fig. 10, point 56.1±0.6/  2.9±0.9/  0.4±0.4/  0.6±0.4/  –  Bal. 
2.4±0.3  0.5±0.2  0.1±0.1  0.1±0.1     
Fig. 10, point 617.3±1.1/  13.1±1.0/  –  6.3±0.4/  –  Bal. 
8.8±0.6  3.1±0.2    1.5±0.1     
Fig. 11.

TEM-BF micrograph of EZ33A thixo-cast after T6 heat treatment (solutioned at 500°C for 6h and aged at 190°C for 33h).

Fig. 12.

(a) TEM-BF micrograph of EZ33A thixo-cast after T6 heat treatment (solutioned at 500°C for 6h and aged at 190°C for 33h) and SAEDP as insert with indexed reflections, (b) HR-TEM image from area visible in (a) and FFT obtained from the marked area.

3.5Mechanical properties

The room temperature compression test was carried out to determine the deformation behavior of the EZ33A magnesium alloy. The stress–strain plots of the ingot (continuous curve; feedstock for SSM), directly after thixoforming (double dotted dashed curve), after two kinds of T6; the first: solution at 500°C for 6h and ageing at 190°C for 13h (dotted curve), the second one: solutioned at 500°C for 6h and aged at 190°C for 33h (dashed curve) are presented in Fig. 13a. The hardness results are collected in the form of the block diagram in Fig. 13b. The ingot had the lowest compression strength 210MPa at hardness 52±3HV and plasticity 4.3%. It was connected with coarse eutectic and intermetallic components, which limited the plastic deformation at room temperature and with lack of nano-precipitates in the α(Mg) matrix. The EZ33A thixo-cast revealed the high plastic strain of 22.5% at the compression strength of 335MPa and, similarly to the examined ingot, the yield strength 90MPa and hardness 54±2HV. Relatively better mechanical properties are the results of refining the eutectic by rapid quenching to a cooled steel die. Additionally, the analysis showed that more eutectic phases appeared in the structure, resulting in strengthening of the interglobular space compared with the input material (ingot). This was due to the fragmentation of the eutectic phase. The significantly better properties were achieved during the measurements of the samples after the heat treatments (solutioning at 500°C for 6h and aging at 190°C for 13h) and (solutioning at 500°C for 6h and aging at 190°C for 33h). Hardness was 67 and 73HV for the sample aged for 13h and 33h, respectively. The yield strengths of both samples after T6 were similar and reached 135MPa. The differences in compression strength: 341MPa, 383MPa and plasticity 17% and 22% were observed for the sample aged for 13h and 33h, respectively. They were due to different lengths of ageing time of α(Mg), which could have affected the number of precipitations responsible for strengthening. The higher hardness of thixo-cast after T6 was in agreement with its increased compressive yield strength. According to Rokhlin [6] who studied EZ33 ingot after T5 during tensile strength test, the yield strength (YS) was 96MPa. Higher YS value of the thixo-cast after T6 (at 30%) could be connected with an increased number of β’ precipitates (due to the solution treatment), while the high plasticity was connected with the fine eutectic in the SSM-ed samples. When comparing the plasticity and compressive strength (CS) of the EZ33 alloy, which contained about 3wt.% of RE and ZE41 one with about 1wt.% RE, it could be seen that the strain and CS were higher for EZ33, in spite of YS, which was lower. The increase in ductility and CS were possibly linked to weakening of the basal texture due to the formation of hard eutectic phases and secondary precipitations brought about by the presence of RE [37].

Fig. 13.

Compression curves of the EZ33A: (a) ingot (continuous curve), directly after thixoforming (double dotted dashed curve), after two kinds of T6: solutioned at 500°C for 6h and aged at 190°C for 13h (dotted curve), and also solutioned at 500°C for 6h and aged at 190°C for 33h (dashed curve); (b) block diagram representing hardness analysis results.


Detailed characterization of EZ33A magnesium alloy after mould casting, thixoforming as well as after T6 heat treatment established differences in microstructure and mechanical properties.

  • (1)

    The EZ33A ingot microstructure showed a globular structure with a coarse eutectic mixture consisting of Mg7Zn3RE, T-phase – RE(Mg, Zn)11 and α(Mg). Heterogeneous nucleation of magnesium solid solution allowed obtaining structure appropriate for SSM in the as-cast state. The temperature of thixoforming was determined using DSC analysis to be 622̊C, which corresponded to about 30% of the liquid phase.

  • (2)

    The thixo-cast microstructure consisted of α(Mg) globular grains with size of 76±1.1 surrounded by fine eutectic mixture in volume of 35%. The T6 heat treatment of thixo-cast (solution at 500°C for 6h and aging at 190°C for 33h) caused light increase of grain size to 92μm and precipitation within the α(Mg) grains of two kinds of phases: β′1 and β′2 responsible for the increase of yield strength to 135MPa, compression strength to 383MPa and hardness to 73HV.

Conflicts of interest

The authors declare no conflicts of interest.


The authors gratefully acknowledge the financial support by the Polish National Centre for Research and Development, Grant No.: LIDER/007/151/L-5/13/NCBR/2014.

X.H. Shao, Z.Q. Yang, X.L. Ma.
Strengthening and toughening mechanisms in Mg–Zn–Y alloy with a long period stacking ordered structure.
Acta Mater., 58 (2010), pp. 4760-4771
L.L. Rokhlin.
Advanced magnesium alloys with rare-earth metal additions.
Adv Light Alloys Compos., 59 (1998), pp. 443-448
I.J. Polmear.
Magnesium alloys and applications.
Mater Sci Technol, 10 (1994), pp. 1-16
L.Y. Wei, G.L. Dunlop, H. Westegen.
The intergranular microstructure of cast Mg–Zn and Mg–Zn-rare earth alloys.
Metall Mater Trans A, 26 (1995), pp. 1947-1955
L.Y. Wei, G.L. Dunlop.
Crystal symmetry of the pseudo-ternary T-phases in Mg–Zn-rare earth alloys.
J Mater Sci Lett, 15 (1996), pp. 4-7
L.L. Rokhlin.
Advances in metallic alloys. Magnesium alloys containing rare earth metals: structure and properties,
Z. Yang, J.P. Li, J.X. Zhang, G.W. Lorimer, J. Robson.
Review on research and development of magnesium alloys.
Acta Metall Sin., 21 (2008), pp. 313-328
F. Czerwinski.
Magnesium injection molding.
Springer, (2008),
Ł. Rogal, J. Dutkiewicz, T. Czeppe, J. Bonarski, B. Olszowska-Sobieraj.
Characteristics of 100Cr6 bearing steel after thixoforming process performed with prototype device.
Trans Nonferrous Met Soc China, 20 (2010), pp. 1033-1036
G. Hirt, R. Kopp.
Thixoforming: semi-solid metal processing.
WILEY-VCH Verlag GmbH & Co. KGaA, (2009),
Ł. Rogal, J. Dutkiewicz, H.V. Atkinson, L. Lityńska-Dobrzyńska, T. Czeppe, M. Modigell.
Characterization of semi-solid processing of aluminum alloy 7075 with Sc and Zr additions.
Mater Sci Eng A, 580 (2013), pp. 362-373
J. Dutkiewicz, L. Rogal, K. Solek, A. Mitura.
Thixoforming technology of high carbon X210CrW12 steel.
Int. J. Mater. Form., 2/1 (2009), pp. 753-756
T.B. Massalski, H. Okamoto, P.R. Subramanian, L. Kacprzak.
Binary alloy phase diagrams.
2nd ed., ASM International, (1990),
Y. Zhang, X.F. Huang, Y. Ma, T.J. Chen, Y. Hao.
Effects of Sm addition on microstructural evolution of Mg–6Zn–0.4Zr alloy during semi-solid isothermal heat treatment.
China Foundry, 14 (2017), pp. 85-92
Z. Zhao, Q. Chen, Y. Wang, D. Shu.
Microstructural evolution of an ECAE-formed ZK60-RE magnesium alloy in the semi-solid state.
Mater Sci Eng A, 506 (2009), pp. 8-15
Ł. Rogal, F. Czerwiński, L. Litynska-Dobrzyńska, P. Bobrowski, A. Wierzbica-Miernik, J. Dutkiewicz.
Effect of hot rolling and equal-channel angular pressing on generation of globular microstructure in semi-solid Mg–3% Zn alloy.
Solid State Phenom, (2015), pp. 381-388
Z. Jeffries, A.H. Kline, E.B. Zimmer.
The determination of grain size in metals.
Trans AIME., 54 (1916), pp. 594-607
ASTM E 91-51T.
Method for estimating the average grain size of non-ferrous metals other than copper and their alloys.
D. Qiu, M.X. Zhang.
Effect of active heterogeneous nucleation particles on the grain refining efficiency in an Mg–10wt.% Y cast alloy.
J Alloys Compd, 488 (2009), pp. 260-264
Z. Szklarz, M. Bisztyga, H. Krawiec, L. Lityńska-Dobrzyńska, Ł. Rogal.
Global and local investigations of the electrochemical behavior the T6 heat treated Mg–Zn–RE magnesium alloy thixo-cast.
Appl Surf Sci., 405 (2017), pp. 529-539
K. Gusieva, C.H.J. Davies, J.R. Scully, N. Birbilis.
Corrosion of magnesium alloys: the role of alloying.
Int Mater Rev, 60 (2015), pp. 169-194
M. Huang, H. Li, H. Ding, Y. Ren, G. Qin, S. Hao.
Partial phase relationships of Mg–Zn–Ce system at 350°C.
Trans Nonferrous Met Soc China, 19 (2009), pp. 681-685
W. Solano-Alvarez, Z. Szklarz, H. Bhadeshia.
Feasibility study for thixoforming nanostructured bainitic steels.
Mater Sci Eng A, 651 (2016), pp. 708-719T
A.A. Kazakov.
Alloy compositions for semisolid forming.
Adv Mater Manuf Process, 157 (2000), pp. 31-34
D.H. Kirkwood.
Semisolid metal processing.
Int Mater Rev, 39 (1994), pp. 173-189
Ł. Rogal, J. Dutkiewicz.
Globular microstructure formation in X210CrW12 steel for semi solid processing using plastic deformation or boron modification.
Arch Metall Mater, 58 (2012), pp. 751-755
V. Pavlyuk, B. Marciniak, E. Różycka-Sokołowska.
The isothermal section of the phase diagram of Ce–Mg–Zn ternary system at 470K.
Intermetallics., 20 (2012), pp. 8-15
R. Kocurek, J. Adamiec.
Evaluation of structure and properties of welded joint of magnesium alloy EZ33A-T5.
Arch Foundry Eng, 13 (2013), pp. 65-71
S.N. Mathaudhu, W.H. Sillekens, N.R. Neelameggham, N. Hort.
Magnesium technology.
A John Wiley & Sons, Inc., Publication, (2012),
F. Czerwinski.
Magnesium alloys-design, processing and properties.
InTech – Open Access Publisher, (2011),
I.J. Polmear.
Light alloys: metallurgy of the light metals.
3rd ed., Arnold, (1995),
J.F. Nie.
Precipitation and hardening in magnesium alloys.
Metall Mater Trans A, 43 (2012), pp. 3891-3939
Y.C. Tsai, C.Y. Chou, R.R. Jeng, S.L. Lee, C.K. Lin.
Effect of rare earth elements addition on microstructures and mechanical properties of A356 alloy.
Int J Cast Mater Res, 24 (2011), pp. 83-87
M. Ashby, K. Johnson.
The bicycle: materials and form.
Materials and design: the art and science of material selection in product design, 3rd ed., Butterworth & Heinemann is an Imprint of Elsevier, (2002),
L.Y. Wei, G.L. Dunlop, H. Westengen.
Precipitation hardening of Mg–Zn and Mg–Zn–RE alloys.
Metall Mater Trans A, 26 (1995), pp. 1705-1716
L. Sturkey, J.B. Clark.
Mechanism of age-hardening in magnesium zinc alloys.
J Inst Met, 88 (1959–1960), pp. 177-181
S. Tekumalla, S. Seetharaman, A. Almajid, M. Gupta.
Mechanical properties of magnesium-rare earth alloy systems: a review.
Met Open Access Metall J, 5 (2015), pp. 1-39
Copyright © 2018. Brazilian Metallurgical, Materials and Mining Association
Journal of Materials Research and Technology

Subscribe to our newsletter

Article options
Cookies policy
To improve our services and products, we use cookies (own or third parties authorized) to show advertising related to client preferences through the analyses of navigation customer behavior. Continuing navigation will be considered as acceptance of this use. You can change the settings or obtain more information by clicking here.