Effect of Pulse Electric Current on Macro-Segregation During Twin-Roll Casting of Al-Mg-Si Alloys

Abstract

The mechanical property, macro and micro-structure of AA6181 aluminum alloys sheets during twin roll casting (TRC) process with pulse electric current (PEC) of different intensity were tested and observed. With increased PEC intensity, the effect of controlling macro-segregation of sheet was more apparent the flow and temperature field during the TRC process was substantially unchanged with the different intensity of PEC. The primary dendrite arm of Al-Mg-Si alloy was increased, but the growth of secondary dendrite arm spacing was restrained. The elongation was increased from 8.1% to 24.2% when PEC strength reaches from 100 A to 500 A. But the yield strength and tensile strength with increasing strength of PEC show a change slightly.

1. Introduction

The centerline macro-segregation of solute elements (non-uniform distribution) in the sheets is a serious aspect even in twin roll casting process (TRCP). TRCP is a process where crystallization, solidification, deformation are all combined into a single process such that it is a simple, economical and effective method for production of aluminum. Controlling macro-segregation is important in developing the TRCP technology due to inhomogeneous mechanical properties in the product remarkable deteriorating the quality of Al-Mg-Si alloys sheets. However, at present there is no effective method to eliminate macro-segregation [1]. To control macro-segregation in the alloy sheets, an understanding of the cause of macro-segregation is essential. According to the solidification principles, the macro-segregation mainly occurs during the solidification stage when the solute moves from the solid to liquid under driving force [2].

The application of pulse electric current (PEC) has attracted significant attention in materials processing because it can change the distribution of solute and crystal morphology during solidification. Pan et al. [3] reported effect of pulse current on the carbon segregation. The results showed the behavior of non-uniform carbon segregation was attributed to Joule heat of the PEC. Zhang et al. [4] reported the effect of electric current on macro-segregation of primary silicon in hypereutectic Al-Si during directional solidification. The study revealed that the electric current caused thermoelectric magnetic convection and promoted the redistribution of solute. Li et al. [5] found that the distribution of different elements showed both positive and negative behavior when pulse magnetic field was applied to solidification. In addition, Liao et al. [6] also found that the electric current pulse can improve the homogeneous distribution of solute. According to studies of Kaldre et al. [7] the application of external physical field during solidification introduces three extra forces in the liquid as well as in the solid. The electromagnetic force that is generated by the interaction between the electric current and magnetic field, can effect the movement of liquid in the vicinity of the liquid-solid interface. This indirectly affects the macro and microstructure of the alloys.

The scholars mentioned above mainly studied the role of pulse electric current during the process of metal solidification. However, the twin-roll continuous casting is different from the solidification of ingot. Its solidification rate is 10²–10³°C/S during normal casting. The thickness of TRC Al-Mg-Si alloys sheets is 3–6 mm, such that the solidification is fast. When PEC is applied to the TRC process, its role in controlling the solute redistribution is worthy of study and is investigated in study described here.

The present study involves the following activities:

1. (a)    production of TRC Al-Mg-Si sheets,
2. (b)    macro and micro-structural examination of sheets with and without different PEC,
3. (c)    solute distribution (Mg and Si) in 6.8 mm thick twin-roll casting sheets,
4. (d)    evaluation of central macro-segregation under different PEC conditions and the mechanism involved, and influence of central segregation on mechanical properties.

2. Experimental Procedure and Simulations Preparation

2.1 Material preparation and processing technic

Figure 1 shows the process and experimental equipment of TRC process with PEC. The TRC process was carried out using a reversible horizontal twin-roll caster with parameters listed in Table 1. Before the experiment, the roller surface was smeared with graphite powder to avoid sticking. A refractory nozzle was used to directly melt into the two rolls gap. The 6181 aluminum alloy was melted in a 20 KW electrical resistance furnace. First, commercially pure (99.9%) aluminum ingot and Al-20%Si ingot were melted. When the melt temperature reached 730°C, it was held for 2 h, and commercially pure (99.9%) magnesium was added to the melt. When the entire pure (99.9%) magnesium was melted, the temperature of the electrical resistance furnace was adjusted to 695°C (Casting temperature). Next degassing was carried out and slag was removed from the melt, followed by pouring into the sluice, through a nozzle between the two roll gap, where the melt solidifies quickly and is deformed. During roll casting, a single PEC was simultaneously applied. Frequency and duty cycle of PECF were constant at 20 Hz and 0.15, respectively. The chemical composition of 6181 alloy in wt% was Al-1.2Si-1.0Mg-0.5Fe-0.312Cr-0.14Cu.

Fig. 1

Schematic sketch of experimental setup. (online color)

Table 1 The parameters of twin roll casting.

2.2 Microstructure observation and tensile tests

Small samples of 20 mm × 20 mm were cut from the as-cast sheets. They were etched with 1:1 hydrofluoric acid-water for 20 s to reveal the structure of alloy sheets. The macro- and microstructure were observed by optical microscope and scanning electron microscopy (SEM). Small samples were cut into six layers along the thickness direction. The surface of the sheets was the first layer and thickness of each layer was 1 mm. The distribution of alloying components was studied by Optical Spectrum Analyzer.

To ensure influence of central segregation on mechanical properties, after cast-rolling, the Al-Mg-Si sheets with different PEC were processed by T6 heat treatment (solution treatment under 530°C for 30 min, water quenching and then artificial aging under 175°C for 0.5 h, cooling in room temperature). The tensile specimens were machined parallel to the rolling direction. The fracture morphology was observed by SEM.

2.3 Numerical simulation parameter

The finite element model of PEC TRC is shown in Fig. 2. Maxwell, which is computational software for calculation of electromagnetic field, was used to calculate the distribution of electromagnetic force. ANSYS was mainly used to simulate flow field and temperature field. The flow and temperature fields of the 6181 aluminum alloy were simulated during the casting and rolling process using the Fluent and Ansoft Maxwell 3D modules in ANSYS Workbench 13.0. Due to the complexity of solidification, heat transfer, and flow processes in the casting and rolling zone, it is difficult to create an accurate model that includes all the influencing factors. Therefore, in order to save calculation time and resources, the following assumptions and simplifications are made for the cast-rolling model:

Fig. 2

TRC model used in the simulation. (online color)

1. (1)    The short initial transition state of the casting and rolling process is ignored, and the temperature of each point in the melt pool area and the process parameters used are constant.
2. (2)    The entire dissolution pool region is regarded as an incompressible Newtonian fluid, and the influence of the solid phase region on the governing equation is ignored.
3. (3)    The flow of molten metal in the casting and rolling zone is regarded as a turbulent motion.

Based on the Fluent module, ProE5.0 was used to build the casting and rolling area model, which was imported into Fluent in STP format and the CFX meshing mechanism was used for the model. After setting the properties of the molten aluminum material in the casting and rolling zone, the boundary conditions of velocity inlet, pressure outlet, rolling wall and fixed wall are added to each surface of the casting and rolling zone, and the SIMPLE algorithm is used to solve the temperature and velocity fields of the casting and rolling zone after the set initial conditions are applied. Finally, the corresponding result is output in the processing module. Based on the Ansoft module, ProE5.0 was used to build the casting and rolling area model, and then the solution domain and traverse model were added to the Ansoft Maxwell 3D module in STP format. Then, the material properties of each part are set, and the winding resistance and electric magnetization are added, and the external circuit action is used as the boundary and initial conditions. By default, the dynamic meshing mechanism is used, and the calculation is performed after the solver is added. Finally, the magnetic field strength and electromagnetic force distribution in the casting and rolling zone are output. In this calculation, the Mesh meshing tool is used to set the minimum element size to 1 mm and then perform hexahedral meshing of the model, and the final number of mesh elements is 35194, the number of mesh nodes is 39197, the average quality of mesh elements is 0.98, and the average quality of mesh is 0.99. The solidification phenomenon of melt is mainly related to the fluid temperature, when the internal temperature of the plate is in a very stable state during the casting and rolling process, the calculation results can reflect the general characteristics of the internal temperature field and flow field in the casting and rolling area. Three different locations were selected in the casting and rolling area for mesh independence test. The results show that when the number of grids changes from 523 to 53693, the temperature change value of the selected three positions is very small. The inlet is set as the speed inlet, the direction is along the casting and rolling direction, the speed is 0.6 m/min, and the setting temperature is 695°C; The outlet boundary is outflow, and the outlet speed is the casting and rolling speed. Considering the complexity of heat transfer between the melt pool and the casting roller, the contact between the molten pool and the casting roller is divided into two areas to calculate separately. Among them, the first area is the aluminum alloy melt and the surface of the cast roll is set to 6 kw/m²k. The second area is the heat transfer of the aluminum alloy plate at the gas, with the increase of the air gap, the heat transfer coefficient gradually decreases, and the constant is set to 100 W/m²k in the paper. In order to improve the solution speed and ensure the solution accuracy, the coupled solution method was selected. Simulating procedure was divided into three steps: First, the electric circuit (generates PEC) was installed at both ends of the outlet and inlet of melt in TRC model. The distribution of electromagnetic force in the TRC model was calculated. Second, the fluid flow model (consistent with TRC) with the electromagnetic force was solved. Finally, the distribution of temperature field was solved by different PEC.

The parameters of the melt and mold with change in temperature are listed in Table 2. To improve the authenticity of the simulation results, it should be noted that the specific heat, density, viscosity and thermal conductivity was varied with temperature. The behavior with change of temperature is described in Fig. 3 by the software of JMatPro. The main parameters of PEC used in the simulation include peak value, duty cycle and frequency. The value of duty cycle was 15%, frequency was 20 Hz and the variation of PEC was from 100–500 A.

Table 2 Therm-physical properties of Al-Mg-Si alloy.

Fig. 3

Thermal parameters and resistivity of Al-Mg-Si alloy. (online color)

2.3.1 Governing equations

When a square PEC field is applied to the TRC process, it induces alternating magnetic field in the conductor, whose interaction generates an electromagnetic force, known as Lorentz force. The Maxwell equations of electromagnetic field can be described by Tillack et al. [8]:

  

\begin{equation} \nabla \times E = - \frac{\partial B}{\partial t} \end{equation} (1)

  

\begin{equation} \nabla \times B = 0 \end{equation} (2)

  

\begin{equation} \nabla \times D = \rho_{0} \end{equation} (3)

  

\begin{equation} \nabla \times H = j + \frac{\partial D}{\partial t} \end{equation} (4)

Where B and D are flux density for magnetic and electric field respectively. H and E are intensity of magnetic and electric field, respectively. ρ₀ is electric charge density while j is electric current density vector. Equations (1) and (4) represent how the field varies in space and eqs. (1) and (4) represent how it circulates in the space.

The pulsed electromagnetic field was induced by PEC in the coil, the current profile in the coil is similar to a square wave, and the discharge time is much shorter than the charging time. The relationship I(t) between the current and time is described by Zhao et al. [9] and Nakada et al. [10]:

  

\begin{equation} I(t) = A_{0} + \sum_{n = 1}^{\infty }A_{n} \sin (n\omega_{1} + \phi) \end{equation} (5)

Where w₁ is fundamental frequency. A₀ is the amplitude. φ is phase. In eqs. (5), the A_n can be calculated from:

  

\begin{equation} A_{n} = \frac{2V}{n\pi }\left| \sin \left(\frac{n\pi \Delta }{T}\right) \right| \end{equation} (6)

Where T and Δ are constants and represent cycle and pulse width, respectively.

The electromagnetic force, which induces force convection in the melt, can be considered a momentum source imposed on the flow. To study this, the electromagnetic force f that is generated by PEC can be calculated using following equation:

  

\begin{equation} f = \nabla [B(t) \cdot I(t)]dl \end{equation} (7)

Where B(t) is the magnetic flux density perpendicular to the current direction. l is distance of current transmitting.

2.3.2 Volume-of-fluid (VOF)

According to Li et al. [11], when PEC is applied to the aluminum melt, a cyclic magnetic field is induced. The electromagnetic force f is generated by their interaction as shown by eq. (7). The VOF during TRC process with PEC was fixed and reconstructed by introduction of the value of f. It can be described as follows [12]:

  

\begin{equation} \frac{\partial F}{\partial t} + \frac{1}{\nu_{F}}\nabla \cdot (\nu F) = F_{s} \end{equation} (8)

Where ν_f is the fractional volume open to flow and F represents the volume fraction occupied by the fluid.

3. Results

3.1 Microstructure morphology

Sun et al. [13] had studied alloy sheets produced by TRC process, exhibited severe macro-segregation defect near the centerline. However, when the 300 A PEC was applied to TRC, the degree of macro-segregation was mitigated. In order to further study the influence law of different PEC on macro-segregation, the PEC of different intensities (100–500 A) were applied to TRC. Figure 4 shows the macro-segregation in sheets with 100–200 A and 400–500 A PEC, and mark the location of the quantification analysis. It can be observed that the application of 100–200 A PEC was not effective in controlling the macro-segregation. When the intensity of PEC exceeded 400–500 A, the segregation can be divided into small regions. With increased PEC intensity, segregation was gradually decreased. In other words, macro-segregation was gradually controlled with increased PEC intensity. Figure 5 presents the quantitative analysis of Mg and Si. In the thickness direction of the sheet, with increase in PEC intensity, the plots of Mg and Si showed less variation with distance (across thickness).

Fig. 4

Macro-segregation features of sheets on the TRC process with different PEC. (a) 100 A, (b) 200 A, (c) 400 A, (d) 500 A. (online color)

Fig. 5

The distribution of alloys elements in the sheet under the different PEC. (a) Si element, (b) Mg element. (online color)

In order to analyze precipitated phases formation in segregation position, the EDS spot-scan was also carried out and the results are shown in Fig. 6. The morphology of precipitated phases are found to be evolved as dagger like, worm shaped as well as irregular morphology for samples under PEC fields. The existence of these uncertain brittle phase, such as Mg:Al:Si:Fe = 8:79:11:2 and Mg:Al:Si = 1:34:15 with small amount of iron in segregation can also be determined by means of quantitative analysis from EDS. According to the analysis chart, with the increase of PEC fields, the degree of segregation in the cast rolled plate decreases.

Fig. 6

The spot-scan EDS of compound phases of segregation in center of sheet. (a) 100 A, (b) 200 A (c) 500 A. (online color)

Additionally, it is necessary to mention the effect of PEC on the microstructure. Under the action of PEC, the crystal morphology transforms from rosiness to dendritic structure. The primary dendrite arm of Al-Mg-Si alloy was increased and the growth of secondary dendrite arm spacing was restrained in comparison to the conventional structure as Fig. 7. The primary dendritic arms are refined with improvement of PEC intensity.

Fig. 7

Schematic diagram of the micro-structure in the TRC process. (online color)

3.2 Mechanical properties

The tensile tests of different samples at room temperature were carried out to research effects of the pulse electric current on the mechanical properties. Figure 8 shows the tensile mechanical property of 6181 alloy in roll casting state and after T6 heat treatment. Those results indicate that the PEC fields does not have a prominent influence on the tensile strength and yield strength. The average tensile strength becomes from 194 MPa (100 A) to 205 MPa (500 A). However, under the condition of 500 A pulse electric current, the average elongation increases by 199% (from 8.1% to 24.2%) when compared to the samples solidified with 100 A PEC (serious macro segregation). Therefore, an evident effect of pulse electric current on the mechanical property of samples was presented.

Fig. 8

The tensile strength and the elongation properties of alloy 6168 samples with different PEC treatment. (online color)

Figure 9 shows the fracture appearance of samples in TRC with different PEC and after T6 heat treatment. The fracture morphology of samples can be seen from Fig. 9(b), (d), (f), (h) and (j). It is very obvious that crack growth area with 100A and 200A PEC are composed of tear ridge patterns and flat surface, which indicate features of brittleness fracture. More dimples are observed in the samples under 300A PEC field and the dimples are more and deeper with improvement of pulse current Intensity. At 500A PEC field, honeycomb-shaped dimples were observed in the sample. This result indicates that the plasticity of sample begins to increases when the PEC reaches 300A. This phenomenon shows that the fracture mechanisms transform from brittle rupture to ductile fracture.

Fig. 9

Tensile fracture of the tensile specimen during TRC with different PEC treatment, (b), (e), (f), (h), and (j) form the backscattered electron images and (a), (c), (e), (g) and (i) form the secondary electron images. (a), (b): 100 A, (c), (d): 200 A, (e), (f): 300 A, (g), (h): 400 A, (i), (j): 500 A. (online color)

Figure 9(a), (c), (e), (g) and (i) are observed by backscattered electron images. It can be seen that the macro segregation near t he center do not be eliminated by T6 treatment. The fracture path mainly occurs in the area of brittle phase concentration, which are detected as Mg:Al:Si:Fe = 8.5:77:11.5:3 by EDS. The results are similar with Fig. 6. With improvement of Pulse Current Intensity, the white precipitates gradually evenly distribute in the area of fracture morphology. As mentioned above, the macro segregation near the center of sheets processed by TRC directly affects material plasticity.

3.3 Forced melt and temperature distribution

Flow stabilizes after a certain time and fluctuates as a regular pattern within the pulsed period of the magnetic field during solidification. However, during TRC, with increase in pulse current, the flow was almost unchanged in Fig. 10. The velocity of flow was maximum near the top and bottom of the roll surface, reached 0.8 m·min⁻¹ and gradually decreased towards the sheet center. Figure 11 shows the effect of different PEC intensity on temperature distribution during TRC. It is apparent that the temperature distribution n is in agreement with flow distribution with no impact on electromagnetic force. The kiss points were also almost coincident for different PEC intensity. Therefore, the pulse current does not affect the redistribution of alloying elements by altering the flow of the aluminum melt. When the pulsed current is introduced into the casting and rolling process, the distribution of solute elements on both sides of the solid-liquid interface is changed, mainly because the pulsed current enhances the diffusion of the solute. When a pulsed current passes through a metal melt, the electromigration phenomenon occurs in charged particles under the action of an electric field. In addition, due to the difference in the charge of Mg, Si and Al ions, the traction of the pulsed electromagnetic force is also different, so local convection will be formed between the dendritic crystal, which will affect the redistribution of alloying elements.

Fig. 10

Distribution of velocity vectors with different PEC intensity, (a) 100 A; (b) 200 A; (c) 300 A; (d) 400 A; (e) 500 A. (online color)

Fig. 11

(a) Distribution of temperature with different PEC intensity, (b) Liquid cave morphology. (online color)

4. Discussion

According to study of Lv et al. [14], the macro-segregation during TRC was associated with the location of kiss point. The kiss point if located close to the molten melt ensures reduced segregation. Concentration of alloys components in liquid phase increases gradually with the flow of molten metal, and concentration reaches maximum at kiss point zone on enrichment as Fig. 12. The vortex will bring some solute elements to return back to the liquid region in molten pool. Li et al. also observed that PEC can produce forced flow to reduce macro-segregation during solidification. However, it was noted that from the above mentioned observations, PEC had almost no impact on the temperature field and flow field and the effect of PEC is weak, and the location of kiss point remains unchanged. This was the main reason that the time was not enough for PEC to affect the flow of melt in 6 mm thick sheets with rapid solidification during TRC process.

Fig. 12

Schematic diagram of macro segregation formation during TRC. (online color)

Mitsuo [15] mentioned that when the PEC passed through the melt, a large number of microscopic particles and clusters would lead to electromigration phenomenon. This is mainly manifested in the driving force of diffusion of solute atoms F, which can be described by following equation,

  

\begin{equation} F = EZ^{*}e \end{equation} (9)

where E is electric field intensity, Z is effective charge number of ion, e is charge of electricity. Equation 10 implies that the greater the electric field intensity, the stronger is the atomic diffusion in the electric field direction. This leads to solute transport flux given by Liao [16]:

  

\begin{equation} J = \mu C_{L} = \frac{D_{L}jF_{a}Z^{*}}{RT\sigma_{L}} \end{equation} (10)

Where µ is electromigration rate, D_L is solute diffusion coefficient, C_L is solute concentration in liquid, R is gas constant, T is melt concentration at the solid-liquid interface. j is current density, σ_L is melt conductivity and F_a is Faraday constant.

In addition, there is a driving force in the melt gradient of temperature and concentration. The concentration of solute that is excluded from solid-liquid per second is given by:

  

\begin{equation} J_{s} = - AD_{L}\frac{dz}{dt} \end{equation} (11)

where dz/dt represents the volume of liquid converted into solid per unit time. Thus, the total diffusion flux of the solute at the front of the solidifying interface with PEC can be expressed by:

  

\begin{equation} J = J_{l} + J_{s} = - AD_{L}\frac{dC_{l}}{dz} + \frac{D_{L}jF_{a}Z^{*}}{RT\sigma_{L}} \end{equation} (12)

Here, it is assumed that the diffusion coefficient does not vary with temperature, and solute diffusion in solid is ignored. According to Fick’s second law of diffusion and Navier-Stoke equation:

  

\begin{equation} C_{L} = C_{L}^{*} - (C_{L}^{*} - C_{0})\frac{1 - e^{ - \left(\frac{v}{D_{L}} + \frac{F_{a}Zj}{RT\sigma_{L}}\right)z'}}{1 - e^{ - \left(\frac{va}{\root 3 \of{D_{L}{}^{2}}j} + \frac{\root 3 \of{D_{L}}F_{a}Za}{RT\sigma_{L}}\right)}} \end{equation} (13)

where C_L is solute concentration at solid-liquid interface, C₀ is original composition of Al-Mg-Si alloy, ν is growth rate of dendrite. Equation (14) can be calculated by using eq. (13) for z,

  

\begin{equation} \frac{\partial C_{L}}{\partial z'} = - \left(\frac{v}{D_{L}} + \frac{FZj}{RT\sigma_{L}}\right)\frac{(C_{L}^{*} - C_{0})e^{ - \left(\frac{v}{D_{L}} + \frac{FZj}{RT\sigma_{L}}\right)z'}}{1 - e^{ - \left(\frac{va}{\root 3 \of{D_{L}{}^{2}}j} + \frac{\root 3 \of{D_{L}}F_{a}Za}{RT\sigma_{L}}\right)}} \end{equation} (14)

Equation (14) shows the distribution of solute at the front of solid-liquid interface is related to the growth rate of dendrite, solute diffusion coefficient, effective charge number of ion, and conductivity, viscosity, density, temperature distribution of melt and electric field intensity. However, when the PEC was applied to TRC process, the $\frac{\partial C_{L}}{\partial z'} < 0$, the solute distribution was monotonically decreased. In other words, with increased PEC intensity, solute concentration at the front of solid-liquid interface was gradually decreased.

Under the action of electromigration as Fig. 13, the positive and negative ions moved toward both ends of the magnetic field and formed a new electric field of strength E. The electric field promotes Al³⁺, Si⁴⁻ and Mg²⁺ clusters to become electric dipoles. The positive and negative ions moved along the direction of the electric field. Thus, the primary dendrite arm grew along the direction of electric field and secondary dendrite arm was restrained as Fig. 7. The phenomenon is consistent with the experimental results.

Fig. 13

Electromigration mechanism PEC during TRC. (online color)

As can be seen from Fig. 8 and Fig. 9, there is an inconsistent relationship between the elongation and macro-segregation in TRC sheets. Some uncertain brittle phase are formed when the alloys components gather in center of sheets at the final stage of solidification. These brittle phase don’s be eliminated after T6 treatment, which directly cause brittle rupture of 6181 sheets during the stretching process. The mechanical properties of all samples with PECF were effectively improved, which can be mainly attributed to these two factors: the refinement of primary dendritic arms and uniform distribution of alloys components. With the application of the pulse electric current, the microstructures of α-Al were transformed from coarse rosiness structure to nearly fine dendritic grains which would significantly improve elongation. The uniform distribution of solute and weaken of brittle phases also can enhance the elongation.

5. Conclusions

1. (1)    With increased intensity of PEC, the effect of control of centerline macro-segregation was more apparent.
2. (2)    The effect of PEC on flow field and temperature during TRC Al-Mg-Si thin sheets was not effective in controlling centerline macro-segregation.
3. (3)    Electromigration was the main mechanism that controlled centerline macro-segregation and altered dendrite morphology.
4. (4)    Under the action of PEC, the primary dendrite arm of Al-Mg-Si alloy became longer, but the growth of secondary dendrite arm spacing was restrained.
5. (5)    The elongation was increased by 199% with 500 A PEC strength. But the yield strength and tensile strength with increasing strength of PEC show a change slightly.

Acknowledgments

The authors would like to thank the Fundamental Research Funds for the National Natural Science Foundation of China (No. 52104377).

REFERENCES

• 1)   K.  Inoue,  K.  Takata,  K.  Ichitani and  Y.  Shirai: Vacancy behavior during aging at 50 and 100°C in Al–Mg–Si alloys with excess Si studied by positron annihilation spectroscopy, Mater. Trans. 60 (2019) 2255–2259. doi:10.2320/matertrans.L-M2019846
• 2)   M.  Morishita,  M.  Abe,  K.  Tokuda and  M.  Yoshida: Prediction method of crack sensitivity during DC casting of Al-Mn and Al-Mg alloys, Mater. Trans. 52 (2011) 166–172. doi:10.2320/matertrans.L-M2010823
• 3)   L.  Pan,  W.  He and  B.  Gu: Non-uniform carbon segregation induced by electric current pulse under residual stresses, J. Mater. Process. Technol. 226 (2015) 247–254. doi:10.1016/j.jmatprotec.2015.07.017
• 4)   Y.H.  Zhang,  X.C.  Miao,  Z.Y.  Shen,  Q.Y.  Han,  C.J.  Song and  Q.J.  Zhai: Macro segregation formation mechanism of the primary silicon phase in directionally solidified Al–Si hypereutectic alloys under the impact of electric currents, Acta Mater. 97 (2015) 357–366. doi:10.1016/j.actamat.2015.07.002
• 5)   Y.J.  Li,  Y.F.  Teng,  X.H.  Feng and  Y.S.  Yang: Effects of Pulsed Magnetic Field on Microsegregation of Solute Elements in a Ni-Based Single Crystal Superalloy, J. Mater. Sci. Technol. 33 (2017) 105–110. doi:10.1016/j.jmst.2015.12.021
• 6)   X.  Liao,  Q.  Zhai,  C.  Song,  W.J.  Chen and  Y.G.  Gong: Effects of electric current pulse on stability of solid/liquid interface of Al–4.5 wt.% Cu alloy during directional solidification, Mater. Sci. Eng. A 466 (2007) 56–60. doi:10.1016/j.msea.2007.02.022
• 7)   I.  Kaldre,  Y.  Fautrelle,  J.  Etay,  A.  Bojarevics and  L.  Buligins: Influence on the macrosegregation of binary metallic alloys by thermoelectromagnetic convection and electromagnetic stirring combination, J. Cryst. Growth 402 (2014) 230–233. doi:10.1016/j.jcrysgro.2014.06.029
• 8)   M.S.  Tillack and  N.B.  Morley: Flow balancing in liquid metal blankets, Fusion Eng. Des. 27 (1995) 735–741. doi:10.1016/0920-3796(94)00148-Z
• 9)   J.  Zhao,  Y.  Cheng,  K.  Han,  X.  Zhang,  Z.  Xu and  Q.  Zhai: Numerical and experimental studies of surface-pulsed magneto-oscillation on solidification, J. Mater. Process. Technol. 229 (2016) 286–293. doi:10.1016/j.jmatprotec.2015.09.027
• 10)   L.  Jianming,  L.  Shengli,  L.  Jin and  L.  Hantong: Modification of Solidification Structures by Pulse Electric Discharging, Scr. Metall. Mater. 31 (1985) 1691–1694. doi:10.1016/0956-716X(94)90465-0
• 11)   X.B.  Li,  F.G.  Lu,  H.C.  Cui and  X.H.  Tang: Migration behavior of solidification nuclei in pure Al melt under effect of electric current pulse, Trans. Nonferrous Met. Soc. China 24 (2014) 192–198. doi:10.1016/S1003-6326(14)63047-8
• 12)   F.  Liu and  L.Y.  Zhang: Numerical simulation of magnetic field and flow field distributions during pure aluminum solidification under pulse magneto-oscillation, Foundry 61 (2012) 285–290.
• 13)   K.M.  Sun,  L.  Li,  S.D.  Chen,  G.M.  Xu,  G.  Chen,  R.D.K.  Misra and  G.  Zhang: A new approach to control centerline macrosegregation in Al-Mg-Si alloys during twin roll continuous casting, Mater. Lett. 190 (2017) 205–208. doi:10.1016/j.matlet.2016.12.109
• 14)   Z.  Lv,  F.  Du,  Z.  An,  H.G.  Huang,  Q.Z.  Xu and  J.N.  Sun: Centerline segregation mechanism of twin-roll cast A3003 strip, J. Alloy. Compd. 643 (2015) 270–274. doi:10.1016/j.jallcom.2015.04.132
• 15)  S. Mitsuo: Liquid metal, Japan, (1987).
• 16)  X.L. Liao: Effect of electric current pulse on solidification structure of metal, Shanghai, Shanghai University, (2007).