Hot Tearing Behavior of Mg-Gd-Y-X Alloys

LIAO Jinge，LI Chengxing，ZOU Wenbing，KANG Jing，SONG Jiangfeng，3，JIANG Bin，3，ZHAO Hua，YANG Zhiyuan，PAN Fusheng，3

（1.National Engineering Research Center for Magnesium Alloy，Chongqing University，Chongqing 400044，China；2.Shanghai Spaceflight Precision Machinery Institute，Shanghai 201600，China；3.State Key Laboratory of Mechanical Transmissions，College of Materials Science and Engineering，Chongqing University，Chongqing 400044，China）

Abstract:Mg-Gd-Y-Zr alloy castings are widely used in the aerospace field owing to their high strength and excellent creep resistance.The castability of these alloys is also an important consideration for engineering application.Thus，the hot tearing susceptibilities（HTSs）of Mg-10Gd-1Y-1Zn-0.5Zr（VW91）alloy and Mg-10Gd-2Y-1Zn-0.5Zr（VW92）alloy are investigated with a constrained rod casting（CRC）mold.The microstructures and fracture surface are characterized by optical microscope and scanning electron microscope.The results unveil that the HTS of VW92 alloy is lower than that of VW91 alloy.The microstructures indicate that obvious tears can be observed in VW91 alloy，while the tears in VW92 alloy are tiny.The tear feeding and healing by eutectic are also observed in VW91 and VW92 alloys.Therefore，the lower hot tearing susceptibility of VW92 alloy is mainly attributed to the high amount of eutectic which feeds and heals tears.Besides，the effects of the coefficient of thermal expansion（CTE）and the fluidity of VW91 and VW92 alloys on their HTSs are discussed.

Key words:VW91 alloy；VW92 alloy；hot tearing susceptibility（HTS）；fluidity；coefficient of thermal expansion（CTE）；eutectic

0 Introduction

Magnesium and its alloys，exhibiting low densi‑ty，superior specific strength and stiffness，and bio‑compatibility，are promising light-weight metals for transportation，3C，aerospace，and biomedical appli‑cations.Since most of Mg alloys are originally prepared by casting processes，defect-free castings are of great significance for their subsequent actual applications.Hot tearing is a common defect in cast magnesium alloys，and occurs in the later stage of solidification，owing to the deficiency of feeding when the stresses exceed the strength of the partially solidified metal.Hot tears regularly initiate in the mushy zone，where the liquid is insufficient to fill the cavities，especially when the cross section changes sharply.Therefore，enhancing the hot tearing resis‑tance of magnesium alloys is of great theoretical and practical value for scientifically selecting materials and mass production.

The hot tearing behavior of several binary Mg al‑loys has been investigated，and the relationship be‑tween the hot tearing susceptibility（HTS）and the composition follows the typical“Λ”curve，in which the HTS rises at first，reaches a climax，and then de‑clines.

SRINVIVASAN et al.studied the hot tearing characteristics of Mg-Gd（=1%，2%，5%，and 10%）alloys.They found that the HTS augmented with the increase in the Gd content to reach a maxi‑mum at 2% Gd and then reduced with further in‑crease in the Gd content to arrive at a minimum at 10% Gd.WANG et al.studied the effect of the Y content（0.2%，0.9%，1.5%，and 4.0%）on the HTS of binary Mg-Y alloys，and reported that the HTS rose with the increase in the Y content，reached the climax at about 0.9% Y，and then decreased with further increasing the Y content.

The hot tearing of rare earth（RE）Mg alloys was investigated.WEI et al.studied the effect of the Zn content on the HTS of Mg-7Gd-5Y-0.5Zr，and found that when the Zn content was 3%，5 %，and 7%，the HTS of the alloys would be reduced by 27%，83%，and 100%，respectively.

Many hot tearing studies have revealed that the HTS is mainly affected by the eutectic amount and susceptible freezing range.LUO et al.reported that the eutectic fraction of the alloy exerted an effect on the formation of the tear，and the eutectic with a high fluidity facilitated the tear healing.SONG et al.in‑vestigated the HTS of Mg-2Ca-Zn alloys，and found a good correlation between the susceptible freezing range and the HTS.The high susceptible freezing range led to a high HTS.Similarly，the in‑crease in the Y content（0%，1%，2%，and 3%）was found to decrease the HTS of Mg-6Zn-1Cu-Y-0.6Zr alloys，while the addition of 0.5Cu and 0.3Mn enhanced the HTS of Mg-8/10Zn-1Al al‑loys，individually or together.

Mg-Gd-Y-Zr alloys have received a lot of atten‑tion in rare earth magnesium alloys，owing to the high strength and excellent creep resistance.Pre‑vious investigations demonstrated that adding Zn into Mg-Gd-Y-Zr can significantly augment its mechani‑cal properties by precipitating long-period stacking or‑dered（LPSO）phases.

CUI et al.studied the effects of Zn addition on the mechanical properties of Mg-10.5Gd-5Y-0.5Zr，and found that Mg-9Gd-3Y-0.5Zn-0.5-Zr alloy had a desirable combination of strength and elongation with 244 MPa in the yield strength，371 MPa in the ulti‑mate tensile strength，and 3.8% in the elongation.Through alloy composition optimization and heat treatment regulation，Chongqing University success‑fully solved the contradiction between the strength and plasticity of Mg-Gd-Zn alloys，and developed the Mg-10Gd-2Y-1Zn-0.5Zr（VW92）cast magne‑sium alloy，which had the tensile strength of 351 MPa，the yield strength of 252 MPa，and the elonga‑tion of 10.2%.Considering the fact that Mg-Gd-YZn-Zr alloys are promising material candidates for cast parts in the aerospace field，their products must have superior qualities without casting defects.Therefore，the resistance to hot tearing must be a sig‑nificant casting characteristic to be investigated.Therefore，in this paper，the HTSs of Mg-10Gd-1Y-1Zn-0.5Zr（VW91）and Mg-10Gd-2Y-1Zn-0.5Zr（VW92）alloys studied through experimental investi‑gation and simulated assessment.

1 Experimental procedures

1.1 Hot tearing tests

Mg-10Gd-1Y-1Zn-0.5Zr（VW91）and Mg-10Gd-2Y-1Zn-0.5Zr（VW92）alloys are prepared from pure Mg，pure Zn，Mg-30%Gd，Mg-30%Y，and Mg-30%Y.The actual chemical composition of the experimental alloys is detected through inductive‑ly coupled plasma atomic emission spectrum（ICPAES）apparatus，as listed in Tab.1.Pure Zn，Mg-30Gd，and Mg-30Y master alloys are added into the Mg melt in an electrical-resistance furnace under a protective atmosphere of high pure COand 0.2%SFat 720 ℃.When the melt temperature increases to 760 ℃，Mg-30Zr master alloy is added and stirred for 3 minutes.Then，the melt is held at 740 ℃for 5 minutes，and is poured into a constraint rod casting（CRC）mold preheated to 250 ℃.For each alloy，the hot tearing tests are repeated for at least three times.

Tab.1 Actual composition of the experimental alloys（%）

The schematic illustration of the CRC mold used in this study is presented in Fig.1（a），and the detailed description of hot tearing setup can be found in the investigation of SONG et al..In short，the device consists of a CRC mold，a temperature moni‑tor，and a load collection setup connected to a data acquisition system.During solidification，the temper‑ature and shrinkage force changes generated at the hot spot can be simultaneously recorded by thermo‑couples and load sensors and be transformed into a force-temperature-time curve.By analyzing the curve，the hot tearing initiation and propagation are determined via the force release.

1.2 Fluidity test

The fluidity of alloys is assessed using a spiral mold.The three-dimensional view of the fluidity ex‑periment mold is shown in Fig.1（c）.The pouring temperature is 740 ℃，and the spiral mold is preheat‑ed to 350 ℃.The fluidity melt composition is consis‑tent with the hot tearing composition.

1.3 Measurement of the coefficient of thermal expansion (CTE)

The CTEs of VW91 and VW92 alloys are mea‑sured with an NETZSCH DIL402PC thermal expan‑sion tester with a specimen size of5 mm×25 mm.The experiments are carried out with high-purity ar‑gon gas protection.The argon flow rate and heating rate are 100 mL/min and 10 ℃/min，respectively，and the measured temperature ranges from 25 ℃to 400 ℃.Three duplicate samples are prepared for each alloy to ensure the reliability of the experimental data.

1.4 Characterization of microstructures

The sampling locations for the microscopic characterization are shown in Fig.1（b）.The samples for the microstructure observations are ground with the SiC abrasive paper with grit sizes of 120，400，600，800，1 000，1 200，and 2 000，and then etched in a solution with acetic acid（5 mL），picric acid（5 g），ethanol（100 mL），and distilled water（10 mL）after polishing.The microstructures of the etched samples are characterized by the optical micro‑scope（OM，ZEISS，Axiovert 40 MAT）.An obser‑vation of tear morphology is carried out by the scan‑ning electron microscope（SEM，FEI Nova 400）equipped with backscattered electron spectrometer（BES）and energy dispersive spectroscope（EDS）.

Fig.1 Schematic illustration of the mold and sample characterization

1.5 Calculation and numerical simulation

The phase diagram and solidification curves are plotted based on the Pandat software and the Scheil solidification model.OKAMOTOhave indicated that Zr did not form any intermetallic phases with Mg，i.e.，the Zr element has no effects on the phase transformation of magnesium alloys.Therefore，Zr is not considered in the phase diagram calculation.

The ProCAST software，based on the finite ele‑ment method，has been used to analyze the physical phenomena of the casting process，e.g.，liquid flow and casting defects.The simulation of hot tearing and fluidity is conducted by the ProCAST software with the thermal module，fluid module，and stress mod‑ule.It is worth noting that a finite element mesh with coarser density is used for the molds of the fluidity and hot tear to save computer running time.The sim‑ulation parameters of the casting are consistent with the experimental parameters.The hot tearing indica‑tor（HTI），based on Gurson’s constitutive model which is used to describe the internal damage of sol‑ids by considering the initiation and growth of the voids in the mushy zone of the casting，can be uti‑lized for a criterion of the hot tearing assessment.

2 Results

2.1 Phase diagram and solidification curves

The phase diagram of Mg-10Gd-Y-1Zn alloys is displayed in Fig.2.The two red lines marked in the diagram represent VW91 and VW92 alloys，re‑spectively.Based on the phase diagram and com‑bined with the non-equilibrium solidification curves calculated by the Pandat software，the solidification process of VW91 alloy is as follows.First，the α-Mg is precipitated from the liquid phase.Then，as the temperature decreases，the 18R LPSO phase is pre‑cipitated from the liquid phase，and then is trans‑formed into the 14H LPSO phase.At the end of so‑lidification，the residual liquid phase in the final stage is converted into MgY，MgGd，and α-Mg.

Fig.2 Phase diagram of Mg-10Gd-xY-1Zn alloys

During the solidification process of VW92 al‑loy，α-Mg is the first precipitation from the liquid phase.When the temperature declines，the 18R LP‑SO phase is precipitated from the liquid phase，and then a mixed LPSO phase of 18R and 14H appears.At the end of solidification，the residual liquid phase in the final stage is transformed into MgY，MgGd，and α-Mg.

The temperature range with solid fraction（）at 0.90-0.99 during solidification is defined as the sus‑ceptible freezing range（SFR）.In this temperature range，the plasticity of the alloy is very low，i.e.，the alloy is more liable to hot tearing with the increase in the SFR.The results and calculated temperature are shown in Fig.3 and Tab.2.With the increase in the Y content from 1% to 2%，the SFR value decreases from 14 ℃to 9 ℃.The calculated results indicate that the HTS of VW92 alloy is lower than that of VW91 alloy.

Fig.3 Solidification curves of VW91 and VW92 alloys calculated by the Pandat software

Tab.2 Calculated temperatures at different solid fractions and content of the final residual liquid for VW91 and VW92 alloys

The calculated temperatures（℃）at different solid fractions for VW91 and VW92 alloys are listed in Tab.2，in whichis the temperature at=0.9，indicates the temperature at=0.99，Δ=-，andis the content of the final residual liquid，which means its content will become zero at the next moment.

2.2 Simulation for the HTS

The ProCAST software is utilized to simulate the solidification process of the alloys and predict the hot tearing defects during the solidification process.The hot tearing indicator（HTI）is a criterion for hot tearing evaluation，which can reflect the HTS of the alloys and the locations where tears occur.The HTI simulated results for the HTS of the alloys are dis‑played in Fig.4.From Fig.4，it can be seen that the hot tearing tendency of VW91 and VW92 alloys is very small，mainly concentrated in the hot spot re‑gion.With the increase in the Y content，the HTS of the alloys decline gradually.The simulated results of the ProCAST unveil that the HTSs of VW91 alloy is higher than that of VW92 alloy.

Fig.4 HTI simulated results for the HTSs of VW91 and VW92 alloys

Both results of the Pandat calculation and the ProCAST simulation show that VW92 alloy has a lower HTS than VW91 alloy.

2.3 Tear morphology and force-temperature-time curves

The force-temperature-time curves and macro‑photographs of VW91 and VW92 alloys are shown in Fig.5.From Fig.5（a），it can be seen that VW91 alloy has a smooth shrinkage force curve，the force drop corresponding to the hot tearing initia‑tion is tiny，and the force measured at 200 s exceeds 600 N.From Fig.5（b），it can be seen that there is an obvious force decline on the hot tearing curve of VW92 alloy，the temperature corresponding to the hot tearing initiation is 594 ℃，and the force mea‑sured at 200 s exceeds 200 N.From Fig.5（c），it can be seen that there are almost no visible tears at the hot spot of VW91 alloy.From Fig.5（d），it can be seen that there are also no visual tears at the hot spot of VW92 alloy.These indicate that the slope of force rise of VW92 alloy is less than that of VW91 alloy.

Fig.5 Force-temperature-time curves and macrophotographs of VW91 and VW92 alloys

2.4 Microstructures

The optical tear morphology at the hot spots of VW91 and VW92 alloys is exhibited in Fig.6.From Fig.6（a），it can be seen that although there are no tears visible to the naked eye，VW91 alloy has a small number of fine narrow tears at the hot spot and the tears propagate along the grain boundaries.From Fig.6（b），it can be seen that tear-like traces at the hot spots of VW92 alloy are observed，and are continuous and located at the grain boundaries.Com‑paring the grain sizes of VW91 and VW92 alloys，it is found that the grain sizes of the two alloys exhibit no significant difference.

Fig.6 Optical microstructures showing tears in VW91 and VW92 alloys

In order to further investigate the distributions of tears in the alloys，the SEM morphology with the BSE mode is taken at the hot spots of VW91 and VW92 alloys，and the results are shown in Fig.7.It can be seen that there are tears on both the upper and lower sides of the VW91 alloy hot spot，and the tears on the upper side are larger than those on the lower side.There are microtears on the lower side of the VW92 alloy hot spot，and almost no tears are observed on the upper side of the alloy.In addition，river-like bright white areas are observed near the hot tears in both alloys.The bright white areas in VW91 alloy are slender and discontinuous，and nu‑merous river-like areas in VW92 alloy are continu‑ously distributed in the field of view.Based on the previous investigations，the white river-like ar‑eas are eutectic and can heal the pre-formed hot tears.According to the non-equilibrium solidification curves calculated by the Pandat software，the four phases，i.e.，MgY，MgGd，α-Mg，and 18R LP‑SO are present in the final stage of solidification of VW91 and VW92 alloys.Therefore，it is judged that the eutectics consist of MgY，MgGd，α-Mg，and 18R LPSO.The presence of eutectics well ex‑plains that the tears on the upper side are larger than those on the lower side.Owing to the gravity，the eutectics are more likely to gather on the lower side to have a better feeding.Since the same area of VW92 alloy is observed in Figs.6（b）and 7（b），it can reach a conclusion that the morphology in the OM is the trace left after the tears are healed by the eutectics.Since the amount of eutectics in VW91 al‑loy is far less than that in VW92 alloy，the tears in VW91 alloy can be observed，while more traces can be spotted in VW92 alloy after the tears are fed and healed by the eutectic.

Fig.7 SEM micrographs of VW91 and VW92 alloys at the hot spots

The further magnification of the tear tip area is shown in Fig.8，where the bright white eutectic near the tears exhibits a broad lamellar structure.Since the amount of eutectics in VW91 alloy is much less than that of VW92 alloy，the area where eutec‑tics heal tears is limited.In the case of VW92 alloy，more manifest traces of eutectics healing tears can be observed.The tears in VW92 alloy are enriched with numerous lamellar eutectics，and the region where the tears initially propagate have been fed and healed by the eutectics.However，there still are ar‑eas not healed by the eutectics in the middle of the tears.Moreover，it can be seen that the eutectics on the left and right sides flow to the central area of the tears and almost connect as a whole.

Fig.8 SEM micrographs of the tear tips of VW91 and VW92 alloys

3 Discussion

In the hot tearing curve of VW91 alloy（see Fig.5（a）），the force drop corresponding to the hot tearing initiation is tiny，while an obvious force de‑cline is observed in the hot tearing curve of VW92 alloy（see Fig.5（b））.In addition，the slope of the force rise of VW92 alloy is less than that of VW91 alloy，indicating that the HTS of VW91 alloy might be smaller.However，from Fig.7，it can be seen that the tears in VW91 alloy are conspicuous，while the tears in VW92 alloy are tiny，which is contradic‑tory to the analysis based on the hot tearing curve.According to our previous studies，the susceptible freezing range（SFR）and eutectic amount have the greatest effects on the HTS of the alloy.However，the SFRs of VW91 and VW92 alloys are 14 ℃and 9 ℃，respectively，and the difference between them is not significant.We believe that the SFR has a slight effect on the HTS of the two alloys.However，comparing the BSE photos of the two alloys，it can be found that the eutectics in VW92 alloy are more enriched than in VW91 alloy.In addition，there is a difference in the residual liquid phase content of the two kinds of alloys.Therefore，we infer that the eu‑tectic amount is the main reason affecting the HTSs of the two kinds of alloys.

In fact，the occurrence of hot tearing in the alloy is closely correlated to its feeding capacity and shrink‑age force during solidification.On the one hand，the feeding capacity of the alloy is affected by the fluidity and eutectic amount.According to the liquid film the‑ory，in the later stage of solidification，the liquid phase is confined in the dendrites，and is difficult to flow.The intergranular tears can easily occur when the shrinkage is hindered.If the fluidity of the alloy is poor，the tears cannot be well fed and healed by the liquid phase，thereby leading to the initiation of hot tearing.On the other hand，the shrinkage force gener‑ated during solidification is mainly related to the CTE of the alloy.The CTE of a metal represents the changes in the macroscopic shape and length of the metal per 1 °C.It is able to reflect the thermal expan‑sion effect of a metal under the effect of a change in temperature.Since the thermal expansion behavior of a metallic material is the result of the non-simple harmonic vibrations of its atoms when heated.There‑fore，an alloy with a high CTE is bound to produce a higher shrinkage force during solidification，which augments the chance of tearing.Based on the above analysis，experiments are carried out to measure the fluidity and CTEs of VW91 and VW92 alloys.The relationships between the fluidity，CTE and the hot tearing of VW91 and VW92 alloys are discussed.

3.1 Effects of the CTE

The curves of the average CTE and temperature of VW91 and VW92 alloys are shown in Fig.9.It can be seen that the CTEs of both alloys rise with the increase in the temperature.When the temperature is below 250 °C，the CTEs of VW91 and VW92 alloys are almost identical.When the temperature exceeds 250 °C，the CTE of VW92 alloy is greater than that of VW91 alloy.Therefore，when the VW92 melt is poured into the mold，owing to its bigger CTE，a larger shrinkage force is generated during solidifica‑tion，resulting in larger tears at the beginning of solid‑ification.This is reflected as a more pronounced force drop in the hot tearing curve.In the case of VW91 alloy，it has a less shrinkage force during solidifica‑tion，and thus the tears at the beginning of solidifica‑tion are smaller，and no obvious force decline is ob‑served in the hot tearing curve.On the other hand，comparing the force values in Figs.5（a）and 5（b），at 200 s，the measured force values in VW91 and VW92 alloys are around 650 N and 200 N，respec‑tively，which means that more stress is released in VW92 alloy during the tear formation.Hence，the high CTE of VW92 alloy results in big hot tearing initiation at the original stage.

Fig.9 Average CTE of VW91 and VW92 alloys

3.2 Effects of fluidity

The simulated and experimental results of the fluidity tests are shown in Fig.10.In the numerical simulation of the casting process，the heat exchange between the alloy liquid and the castings follows the law of conservation of energy during the casting fill‑ing process，and the fluidity of the present simula‑tion is demonstrated by the temperature field calcu‑lated during the material filling.The simulated re‑sults indicate that there is no big difference between the fluid lengths of VW91 and VW92 alloys.The experimental results demonstrate that the lengths of VW91 and VW92 fluidity samples are 286 mm and 240 mm，respectively，which indicates that the fluidi‑ty of VW91 alloy is better than that of VW92 alloy.However，according to Fig.7，conspicuous tears can be observed in VW91 alloy，while the tears in VW92 alloy are tiny.Combined with Figs.7 and 10，it can be found that the fluidity of VW91 alloy is better than that of VW92 alloy.

Fig.10 Simulated and experimental results of fluidity for alloys

However，the eutectic amount of VW91 alloy is 0.5%，which is less than that of VW92 alloy（1%）.Since eutectics can feed and heal tears，when large tears initiate in VW92 alloy at the original stage of solidification，the eutectics will flow to the tears for feeding and healing.Therefore，the trace of large tears being healed by the eutectics can be seen in the BSE（see Fig.7（b））.However，due to the poor flu‑idity of VW92 alloy，there are still some tears which have not been fed by the eutectics in time.

Although VW91 alloy has good fluidity，only a few tears can be fed and healed owing to the less eu‑tectic amount，and most of the tears are eventually retained.Therefore，the healed hot tearing is mainly correlated to the eutectic amount，and the fluidity of the melt has a slight effect on the healing of tears.

In summary，the high CTE of VW92 alloy leads to large tear initiation，which is reflected by the low final force in the hot tearing curves.Due to the high amount of eutectics，the formed tears in VW92 alloy are healed by the eutectics，which is reflected by the tear-like eutectic traces at the hot spot.The observed hot tearing size in VW91 alloy is larger than that of VW92 alloy.

4 Conclusions

In the present work，the HTSs of VW91 and VW92 alloys are studied through experimental inves‑tigation and simulation assessment.The main conclu‑sions can be summarized as follows.

1）The simulated results are in good agreement with the experimental results.The HTS of VW92 al‑loy is lower than that of VW91 alloy.

2）Since the CTE of VW92 alloy is larger than the CTE of VW91 alloy，large tears are produced in VW92 alloy at the beginning of solidification owing to the greater shrinkage force generated during solidifica‑tion，while small tears are produced in VW91 alloy owing to the less shrinkage force during solidification.

3）Although the fluidity of VW91 alloy is better than that of VW92 alloy，its eutectic amount is much less than that of VW92 alloy.The tears initiated in the early stage of solidification in VW92 alloy can be fed and healed by numerous eutectics，but some tears are still not healed by the eutectics owing to the poor fluidity.In VW91 alloy，only a few tears can be healed as a result of the less eutectic amount，and most of the tears are retained.