Ye Zhou,Pingli Mao,?,Le Zhou,Zhi Wang,Feng Wang,Zheng Liu

a School of Materials Science and Engineering,Shenyang University of Technology,Shenyang 110870,China

b Key Laboratory of Magnesium Alloys and the Processing Technology of Liaoning Province,China

Received 19 July 2019;received in revised form 13 March 2020;accepted 19 March 2020 Available online 24 June 2020

Abstract The effect of long-period stacking ordered(LPSO)phase on hot tearing susceptibility of Mg–1Zn–xY serial alloys were investigated experimentally using a home-made T-type hot tearing mold.The characteristic parameters related to HTS during the solidification process were calculated by thermal analysis.The microstructure and morphology of the crack zone were characterized by optical microscope(OM),scanning electron microscopy(SEM)and electron dispersive spectrometer(EDS),and the phases of the alloys were analyzed by X-ray diffraction(XRD).The result showed that the long-period stacking ordered(LPSO)phase formed when m(Zn)/m(Y)<1,and the LPSO content increased with increasing of Y.The presence of the LPSO phase in Mg–1Zn–xY alloys could benefit the hot tearing refilling and decrease the HTS of the alloys.With increasing the content of LPSO phase,the HTS of the alloys decreased.LPSO phase increased the skeleton strength,and reduced the HTS of Mg–1Zn–xY alloys.

Keywords:Mg–Zn–Y alloys;Hot tearing susceptibility;Thermal analysis;LPSO phase;Numerical simulation.

1.Introduction

Owing to low density,high specific stiffness,and corrosion resistance,magnesium(Mg)alloys have been widely used in aerospace,automobile,and electronics industry[1–3].However,their industrial and commercial applications have been restricted by the wide solidification temperature range and a considerable tendency to hot tearing.Hot tearing,also known as hot crack,has been thought as a major defect during solidification.And it usually occurs in the mushy zone,where the solid fraction is within the range of 0.9 and 0.99[4].The solidification process can be divided into four stages:(1)mass feeding,where the liquid can move freely and then the hot tearing hardly can form;(2)interdendritic feeding,where the dendrites start to contact with each other forming a solid network,the liquid must flow through the network.(3)interdendritic separation,where the solid network is torn in some area due to the contraction force causing by solidification shrinkage;(4)interdendritic bridging,where the solid network structure develops moderate strength.The last two stages play a major role in the occurrence of hot tearing[5,6].

Up to now,extensive research efforts have been devoted to the investigation of the hot tearing phenomenon during solidification of Mg alloys by using theoretical and experimental methods[7–9].Several factors influenced the severity of hot tearing,such as cooling rate of the alloy,properties of phases,the evolution of local stress and strain rate,and mold geometry[10–12].

Magnesium–zinc(Mg–Zn)based alloys exhibit high potential application as a high-temperature structural material in the automobile industry due to their preferable hyperthermia stability.However,the high hot tearing susceptibility(HTS)of Mg–Zn binary alloys limited their practical applications.Zhou et al.[13]reported that the HTS of Mg–Zn binary alloy followed the so-called“l(fā)ambda”shape curve,which was that the HTS of Mg–Zn binary alloy increased with increasing Zn content and then decreased with further addition of Zn.Effect of Ca content on the HTS of Mg–Zn alloys was also studied[14,15].The results indicated that element Ca played a significant role in HTS behavior of Mg–Zn–Ca ternary alloys.It exhibited that the increase of Ca content could reduce HTS of the alloys.

Recently,Mg–Zn–Y ternary alloys attracted intensive attention owing to the variety of the second phases generated in the alloy according to the Zn/Y ratio[16].In the meanwhile,it was found that the content of Zn,Y,and Zn/Y ratio were the important factors to affect the HTS of Mg–Zn–Y ternary alloys[16,17].The type of the second phase was closely related to the Zn/Y ratio.When Zn/Y>1,Zn/Y≈1,and Zn/Y<1,the second phase of the alloy is I phase(or I phase and W phase),W phase(or W phase and small amount of I phase),and LOPS phase(or small amount of W phase),respectively[18].LPSO phase as a specific strengthen phase has been investigated by many researchers in high Y content alloys[19].However,HTS of Mg–Zn–Y system alloys with LPSO phase was not reported so far.In this study,the effect of the LPSO phase on hot tearing susceptibility of Mg–Zn–Y system alloys was investigated by the experimental and numerical methods.

2.Experimental procedure

2.1.Hot tearing tests

HTS of ternary Mg–Zn–Y alloys was evaluated in a Constrained Rod Casting(CRC,hereafter)mold,which was designed to capture the hot tearing in the corner between the sprue and the constrained rod.The details of the apparatus could be found elsewhere[20,21].The chemical compositions of Mg–Zn–Y alloys were measured using an Inductively Coupled Plasma(ICP)and the results are given in Table 1.

Table 1Chemical composition of the investigated alloys in weight percentage.

Resistance furnace with Si temperature controller was used for melting the alloys.Melting and pouring process was covered with the mixed gas of SF6(0.2%)and nitrogen with the flow rate of 1.6L min?1.Before melting,the inner surface of the stainless-steel crucible was sprayed with boron nitride(BN)coatings.Pure Mg was firstly melted,and then pure Zn and Mg–25%Y master alloy were put into the molten metal.The molten metal was mechanically stirred for 1min in order to homogenize the alloying elements and then held at 700°C for 30min to ensure complete dissolution and diffusion of the alloying elements.The molten metal was poured at 700°C into a“T-shaped”permanent mold,which was preheated to 250°C.

2.2.Differential thermal analysis

Thermal analysis was carried out based on the twothermocouple method.One of the thermocouples was placed close to the wall and the other at the center of the graphite crucible to measure the temperature difference between the center(Tc)and edge(Te)of molten metal during solidification.The thermal analysis experimental setup is illustrated in Fig.1[8].The top and the bottom graphite crucible were insulated with an insulating material to ensure Newton heat transfer.The first maximum temperature difference between the two thermocouples is defined as dendrite coherent temperature(Tcoh),which is considered as a key parameter to measure the feeding mechanism of casting Mg alloys.

2.3.Microstructure observation

The microstructure of the casting alloys was observed on the longitudinal section of the rod samples near the location of the hot spot(an area where hot tears were easy to occur).The samples were ground with SiC paper(from 600 grit up to 2000 grit)followed by mechanical polishing and then were chemically etched in a solution of nitric acid alcohol(4%).The microstructural observations were carried out by using an optical microscope(OM,ZEISS)and a scanning electron microscope(SEM,HITACHI S-3400N)equipped electron dispersive spectrometer(EDS)operated at 20kV.The phases of the investigated alloys were verified by X-ray diffraction(XRD,SHIMADZU-7000)with a scanning angle from 20° to 80° with a scanning speed of 2° per min.

Fig.1.Schematic illustration of thermal analysis setup[8].

Fig.2.Meshing results of(a)the casting sample and(b)hot tearing mold.

2.4.Numerical simulation

The ProCAST software was used to analyze the hot tearing behavior of the investigated Mg–Zn–Y system alloys,which is widely used for casting simulation to explain the phenomena occurring during the solidification process.The simultaneous operation of the thermal fluid flow and stress modules,as well as the hydrodynamic calculation of the stress and strain formation in the crystallization and cooling castings of casting operations are carried out simultaneously in accordance with international standards,and the solid fraction field and hot tearing indicator(HTI,hereafter)of the temperature field are displayed by a post-processor.The results of the field can be used to analyze the formation of hot cracks.

Before the beginning of the simulation,the grid generator MeshCAST module in ProCAST was used to divide the computational grid of casting and die models.Fig.2 is the result of meshing of the casting samples and hot tearing mold.

3.Results and discussion

3.1.Solidification analysis

Fig.3 illustrates the cooling curves recorded with the center thermocouple during the solidification process of Mg–1Zn–xY(x=1,2,3)alloys.It can be seen that there are two distinct cooling rate changes appearing in the cooling curves,as indicated by arrow“a”and“b”,implying the heat change caused by phase transformation during the solidification.The phase transformation during the solidification of Mg–1Zn–xY(x=1,2,3)alloys was verified by the first derivative of the cooling curves(dT/dt),as shown in Fig.4(a–c).

Fig.3.Cooling curves for Mg–1Zn–xY(x=1,2,3).

The baseline in Fig.4 is the first derivative of the cooling curve which assumed having no phase transformation during solidification.The peak of the first derivative of the cooling curves is the sign of the phase transformation.For Mg–1Zn–1Y alloy,the cooling and first derivative curves in Fig.4(a)show two well-defined peaks at 642 and 532°C,respectively.Combining the analysis of XRD results in Fig.4(d)and Mg–Zn–Y ternary phase diagram in Fig.4(e),it can be deduced that the corresponding phases of those two peaks are belonging toα-Mg and W phase.It is further confirmed by the microstructure observation shown in Fig.5(a).The phase transformation reactions of the two peaks are theα-Mg formation and the non-equilibrium eutectic reaction:L→α-Mg+W[22],respectively.The cooling and first derivative curves of Mg–1Zn–2Y alloy are shown in Fig.4(b).It can be seen that three obvious peaks appear at 635,538 and 530°C,respectively.Similarly,it can be concluded that the three peaks areα-Mg,W phase and LPSO phase according to the XRD results in Fig.4(d)and Mg–Zn–Y ternary phase diagram in Fig.4(e),and corresponding microstructure observation shows in Fig.5(b).The corresponding reactions areα-Mg formation,L→α-Mg+W,and L→α-Mg+LPSO eutectic transformations[23].A new phase of LPSO precipitates in the alloy due to the increment of Y(m(Zn)/m(Y)<1).The thermal analysis result for Mg–1Zn–3Y alloy is shown in Fig.4(c).It can be observed that three peaks showing in the first derivative curve,indicating three-phase transformation reactions.The corresponding temperature of the three peaks is 631,540,and 531°C,respectively.According to the XRD result in Fig.4(d)and the microstructure observation in Fig.5(c)that the reactions are the same as Mg–1Zn–2Y alloy.However,the amount of LPSO phase increases with increasing of Y content,as shown in the microstructure of the LPSO phase in Fig.5(c).

Fig.4.Thermal analysis results and XRD results of Mg–1Zn–xY alloys:(a)x=1,(b)x=2,(c)x=3[wt%],(d)XRD results and(e)phases in as-cast alloys.

For Mg–1Zn–3Y alloy,EDS mapping of the LPSO phase is measured and shown in Fig.5(d).It shows obviously that Mg element enriches in the matrix,and Zn,Y elements enrich in LPSO phase.Two adjacent grains are crossed by LPSO phase(Fig.5d)[24],which act as the bridging effect between the adjacent grains,resulting in the enhanced intergranular cohesion between two grains.At the later stage of solidification,the high strength of dendrites between two grains could resist the solidification contraction force.LPSO phase increases the strength of dendrites,and hinders dendrite separation,as a result,it reduces HTS of the alloy.

The phase reaction temperatures of Mg–1Zn–xY alloys(x=1,2,3)are summarized in Fig.6 according to the solidification curves.It shows that with the increasing of Y content,the formation temperature ofα-Mg decreases,the precipitation temperature of the LPSO phase slightly increases,and that of W phase seems stable.The temperature range from the reaction ofα-Mg to the precipitation of W phase are 110,105 and 100°C,respectively,which becomes smaller with increasing content of Y according to Fig.6.

The temperature of dendrite coherent(Tcoh)could be calculated through the data measured by thermal analysis during theα-Mg formation(Fig.7a)[25].Fig.7b shows the coherent temperature of Mg–1Zn–xY(x=1,2,3)alloys.The results show thatTcohdecreases with the increasing of Y content.

Fig.6.The phase reaction temperature of Mg–1Zn–xY alloys(x=1,2,3).

Fig.7.Cooling curve of the center and the edge of the temperature difference(T=Te?Tc)for(a)Mg–1Zn–2Y alloy,and(b)the dendrite coherent temperature for Mg–1Zn–xY alloys(x=1,2,3).

Hou et al.[26]found that if the solidification structure was in a small equiaxed crystal with short dendritic arms then the alloy would have lowTcoh,which can be proved by Fig.8(g–i).In contrast,if the alloy had highTcohthe well-developed dendrites would form a dendrite network and segment the liquid into isolated molten bath at an early stage of solidification,resulting in the hard of refilling the hot tears.According to theTcohanalysis result,it can be predicted that the HTSs of the three alloys are in the following order:Mg–1Zn–1Y>Mg–1Zn–2Y>Mg–1Zn–3Y.

3.2.Experimental and numerical simulation results of hot tearing test

The variation of contraction force curves and corresponding macroscopic samples of Mg–1Zn–xY(x=1,2,3)alloys are shown in Fig.8.The red line represents solidification temperature,and the black line represents the contraction force.In these force curves,the contraction force increases steeply at the beginning of the solidification,then suddenly drops or forms a plateau at a certain level.The appearance of the force dropping or the plateau indicates the initiation and propagation of hot tearing.The force curves exclusively show the formation of hot tears under all the experimental conditions.Fig.8(a)shows the contraction force of Mg–1Zn–1Y alloy,it indicates that the hot tearing initiates at the later stage of the solidification,and the propagation lasted for a period of time,implying high HTS of the alloy.It is consistent with the macroscopic observation of hot spot section of the Mg–1Zn–1Y alloy(Fig.8(d)).The contraction curve of Mg–1Zn–2Y alloy exhibits short period propagation after hot tearing initiation,indicating lower HTS than Mg–1Zn–1Y alloy,as shown in Fig.8(b)and(e).Fig.8(c)shows the contraction force curve of Mg–1Zn–3Y alloy,which exhibits a sudden decrease and then straight increases until the solidification completed.There is virtually no hot tear propagation in Fig.8(c),and there are no visible hot tears with naked eyes on the casting surface in Fig.8(f).The hot crack volumes of Mg–1Zn–xY(x=1,2,3)alloys are measured by the paraffin permeation method[20],which is 102.5,51.7 and 14.8mm3,respectively.The microstructures of Mg–1Zn–xY(x=1,2,3)alloys at the hot spot show in Fig.8(g–i).It is clear that the LPSO phase precipitates at grain boundary at Y content of 2 and 3wt%and the grain size decreases with increase of Y,which is 110,74,and 56μm,respectively.

The hot tearing simulation results are presented in Fig.9.When the alloy composition and boundary conditions are input into the software,the thermal physical parameters of the alloy are automatically adjusted,and the HTS of the experimental alloy is simulated by the hot tearing indicator module.The HTI is a strain-driven model based on the total strain that develops during the solidification process.The model computes the elastic and plastic strain,and it is based on Gurson’s constitutive model[27].Fig.9(a–c)shows the HTI results and the color labels show on the left,which indicated the level of the HTS at the hot spot of the alloy.The simulated hot spot of Mg–1Zn–1Y alloy displays a large area of red and the HTI is 0.0298,which exhibits high HTS.For Mg–1Zn–2Y alloy(Fig.9b),the red area of hot spot decreases,and corresponding HTI decreases.Meanwhile,the hot tearing sensitive area decreases.When the amount of Y increases to 3wt%,the HTI is only about 0.0199.As far as the simulation results are concerned,the increase of Y content has a great influence on the HTS of Mg–1Zn–xY alloys,that is,increasing Y content will reduce the HTS of Mg–1Zn–xY alloys.

Fig.8.The contraction force curves,macroscopic samples and microstructure of Mg–1Zn–xY alloys:(a),(d)and(g)x=1;(b),(e)and(h)x=2;(c),(f)and(i)x=3.

Fig.9(d–i)shows the solid fraction and corresponding effective stress of a point in the hot spot,where exhibits high HTI.According to previous study that hot tears occurs at the later stage of solidification when the solid fraction range is 0.9–0.99.Fig.9(d)clearly shows that the mushy zone(0.9

3.3.Fracture surface morphology

Fig.10 shows the fracture surface morphology of Mg–1Zn–1Y and Mg–1Zn–2Y alloys.The bubble-like protrusions are the dendrites covered with liquid film before final solidification.The traces of the liquid flow can be seen among the dendrites in the fracture surface,which are caused by the dendrites separation and residual liquid shrinkage in the mushy zone.Fig.10(a)shows the fracture surface morphology of Mg–1Zn–1Y alloy,a lot of torn bridge can be seen on the surface indicating the torn of the dendrite due to the solidification contraction and the bridge enriches Mg element,indicating that the dendrite skeleton forms earlier in the solidification stage and the solidification stage changes from mass feeding to interdendritic feeding prematurely,which will cause high HTS.Fracture surface morphology of Mg–1Zn–2Y alloy is exhibited in Fig.10(b).It can be seen that a lot of liquid flow evidence can be found in the fracture surface.It means the liquid flow with certain strength can resist contraction force at the last stage of solidification,implying that the dendrites have certain strength before hot tearing,resulting in decreasing of HTS with high Y element content.

Fig.9.Numerical simulation results of Mg–1Zn–xY(x=1,2,3)alloys:HTI results(a)x=1,(b)x=2,(c)x=3;Solid fraction of hot spot(d)x=1,(f)x=2,(h)x=3;Effective stress of hot spot(e)x=1,(g)x=2,(i)x=3.

The typical microstructures of longitudinal cross-section near the hot crack region of Mg–1Zn–1Y and Mg–1Zn–2Y alloy are shown in Fig.11.It is clear that the hot tears are not rapidly refilled in Mg–1Zn–1Y alloy.While in Mg–1Zn–2Y alloy,the refilling races are clearly presented,which insure as LPSO phase(Fig.11b).Fig.12 shows the liquid fraction(fL)curves of Mg–1Zn–xY(x=1,2,3)alloys during the solidification.ThefLof the hot tearing initiation moment is marked on the curves,which was 3%,8% and 27%,respectively.ThefLincreases with the increase of Y content,i.e.,the remaining eutectic liquid increases,which is believed to be capable of refill the hot tear.Hence,a large amount of liquid effectively promotes the crack healing and the LPSO phase refilling in Mg–1Zn–2Y alloy guarantees the low HTS of the alloy.As the addition of Y increase to 3wt%,the HTS of the alloy is low enough so that there is no hot crack can be found with the naked eye near the hot spot(Fig.8f).LPSO phase plays an important role in refilling at the later stage of solidification,which can reduce the HTS.

Fig.10.Fracture surfaces of hot tearing in(a)Mg–1Zn–1Y and(b)Mg–1Zn–2Y.

Fig.11.The typical microstructures near the hot crack region of(a)Mg–1Zn–1Y and(b)Mg–1Zn–2Y alloys.

Fig.12.The liquid fraction curves of Mg–1Zn–xY(x=1,2,3)alloys:(a)x=1,(b)x=2,(c)x=3.

4.Conclusions

In the present study,the effects of long-period stacking ordered(LPSO)phase on hot tearing susceptibility(HTS)of Mg–1Zn–xY(x=1,2,3)alloys were investigated.The main conclusions can be summarized as follows:

(1)Whenm(Zn)/m(Y)<1,LPSO phase formed in the Mg–1Zn–xY alloys,and the content of the LPSO phase increased with increasing of the Y element.

(2)The experimental results show that the HTS decreases with the increasing of Y,and the temperature of dendrite coherent(Tcoh)also decrease.Corresponding hot crack volume is 102.5,51.7 and 14.8mm3,respectively.

(3)The simulation results have the same trend as the experimental results,that is,with the increasing of Y content,the hot tearing indication(HTI)decreased.The effective stress of the hot spot during the solidification also decrease with the increase of Y.

(4)LPSO phase benefited to the liquid flow with a certain strength and increased refilling traces at a later stage of solidification.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors would like to acknowledge the financial support from High Level Innovation Team of Liaoning Province(XLYC1908006).Innovation Talent Program in Science and Technology for Young and Middle-aged Scientists of Shenyang(No.RC.180111),Project of Liaoning Education Department(Nos.LQGD2019002 and LJGD2019004),Liaoning nature fund guidance plan(No.2019-ZD-0210),and Liaoning Revitalization Talents Program(Nos.XLYC1807021 and 1907007).