Influence of alloying elements on hot tearing susceptibility of Mg-Zn alloys bas

Zhenzhen Yng,Ke Wng,Penghui Fu,b,∗,Liming Peng,Bin Hu,Ming Liu,Anil K.Shev

aNational Engineering Research Center of Light Alloy Net Forming and State Key Laboratory of Metal Matrix Composite,Shanghai Jiao Tong University,Shanghai 200240,PR China

bShanghai Light Alloy Net Forming National Engineering Research Center Co.,Ltd.,Shanghai 201615,PR China

cGeneral Motors China Science Laboratory,Shanghai 201206,PR China

dGeneral Motors Research&Development Center,Warren,MI 48090,USA

1.Introduction

As one of the lightest structural materials,with their high specific strength and low density,magnesium(Mg)alloys are one of the most promising materials for lightweight applications.As most Mg alloys are initially prepared by casting processes,casting abilities are quite important for their actual applications.Unfortunately,Mg alloys are susceptible to cracking during solidification,which is one of the most fatal defects,namely,hot tearing or hot cracking.Previous studies demonstrate that hot tearing arises above the solidus temperature as a result of obstructed contraction during solidification,regularly in the region of mushy zone,where the liquid is not sufficient to fill the cavities,especially at a sharp change of cross section[1].

Considering the fact that the traditional trial-and-error method is time-consuming and laborious,a suitable computational model to predict the hot tearing susceptibility(HTS)of Mg alloys can help exclude alloy compositions that have poor hot tearing resistance during alloy design.Over decades,several models and criteria were presented for hot tearing.These criteria includes those based on nonmechanical aspects,such as Feurer[2],Clyne-Davies[3],and Katgerman[4],those based on mechanical aspects,such as Prokhorov et al.[5],and those that combined these two features,such as the Rappaz-Drezet-Gremaud(RDG)model[6].However,none of these existing criteria is very qualitatively consistent with the casting practice because of the complex mechanisms acting during the solidification of metals.Recently,Kou[7]proposed that tearing initiates at preexisting nucleation sites and propagates if the difference of separation rate of two neighboring grains due to tensile deformation and their growth rate toward each other exceeds the liquid feeding rate along the grain boundary.An index for the hot tearing tendency of an alloy during solidification was also put forward,that is,near,whereTis temperature andfsis the fraction of solid.Based on this index and thermodynamic software Pandat,Kou calculated the hot tearing tendencies of some aluminum alloys and compared the results with the practical HTS experimental data in casting practice.As a result,the calculations were in fairly accordance with the experimental values.Therefore,seems to be a good criterion for the prediction of HTS of an alloy.

Hot tearing resistances of Mg-Al[8],Mg-Zn[9],Mg-Gd[10],and Mg-Zn-Y[11]alloys have been studied.Zn addition to Mg alloy could improve the strength and hardness as well as the corrosion resistance[12].Mg-Zn based alloys have excellent mechanical properties,including both strength and ductility,along with low cost.However,they have a wide freezing range and are prone to hot tearing during casting.Therefore,if the hot tearing resistance of Mg-Zn based alloys could be improved,they will have wide applications.

In the present study,the hot tearing indexrecently proposed by Kou[7]is used to select alloying elements in order to reduce the hot tearing susceptibility(HTS)of Mg-Zn based alloys.Al,Cu,and Mn elements are picked up and their influence on hot tearing susceptibility of MgxZn(x=6,8,10,wt%)alloys was determined by constrained rod casting.

2.Prediction method and selection of alloying elements

2.1.Prediction of HTS based on the index

Hot tearing indexbased on the criterion proposed by Kou[7]indicates that a higher value ofnearmeans a higher hot tearing susceptibility.The index can be seen from the steepness of theTand(fs)0.5curve near,which can be calculated using commercial thermodynamic software Pandat and alloy database.Kou[7]confirmed that the approximation of indexin the range of 0.87＜fs＜0.94 agreed with some experimental data in Al casting alloys.The validity of the index to predict the hot tearing susceptibility of Mg casting alloys will be assessed in the present study.

The effect of Zn content on the susceptibility of binary Mg-Zn alloys is discussed first as follows:in order to see whether the proposed index ofcan produce aλ-shaped curve and a peak susceptibility around 1.5%Zn(hereafter,all compositions are in weight percent unless otherwise stated)reported in several experimental studies,e.g.Le Zhou et al.[9].Based on Mg database and Pandat software,the curve ofTvs.fsof Mg-Zn based alloys during solidification was determined.The curves ofTvs.for binary Mg-Zn alloys are shown in Fig.1a.An average steepnesswith 0.9＜fs＜0.99(corresponding to,about in the region of mushy zone)is used as an approximation.The plot ofyields aλ-shaped curve with the peak around 1.5wt%Zn(Fig.1b),which agrees well with the experimental data of Zhou et al.[9].

Fig.2 shows the calculation results of the predicted susceptibility of four series alloys,whose hot tearing susceptibilities were tested in experiments[13-16].Fig.2a indicates that with the increase of Zn addition to Mg-9Al alloy,the predicted values of the HTS based on the index increase,which is in agreement with the experimental results demonstrated by Wang et al.[13].The effect of Al addition to Mg-1.5Zn alloy on the susceptibility was calculated and shown in Fig.2b,predicting that Al addition reduces the HTS of Mg-1.5Zn alloy.The prediction is consistent with the recommendation that Al addition can increase the hot tearing resistance of Mg-1.5Zn alloy,as proposed by Zhou et al.[14].Fig.2c demonstrates aλ-shaped curve with the peak around 0.5wt%Cu,consistent with Wang’s results[15],which indicates that 0.5wt%Cu addition can reduce the HTS of Mg-6Zn alloy.Furthermore,Zheng et al.[16]investigated the susceptibility of Mg-2.5Zn-xY-0.5Zr alloys,and the experimental results of HTS are as follows: MgZn2.5Y2Zr0.5＞MgZn2.5Y0.5Zr0.5＞MgZn2.5Y4Zr0.5＞MgZn2.5Y6Zr0.5,consistent with the predicted tendency using the index in Fig.2d.

Therefore,the predictions based on the indexare consistent with the hot tearing results of Mg-9Al-xZn[13],Mg-1.5Zn-xAl[14],Mg-6Zn-xCu[15]and Mg-2.5ZnxY-0.5Zr[16]alloys.It suggested that the indexbased on thermodynamic software Pandat and Mg database is useful and can be used to study the influence of chemical compositions on the hot tearing resistance of Mg-Zn based alloy.

2.2.Selection of alloying elements

Fig.3 shows the prediction results of hot tearing tendencies of Mg-Zn-Al alloys based on the index and Pandat.The alloys containing 6-10%Zn tend to have relatively lower hot tearing tendencies.Therefore,Mg-xZn(x=6,8,10,wt%)were chosen as basal alloys.From previous studies,it is observed that Al content addition to Mg-Zn alloy can increase the hot tearing resistance[14,17].Hence,1%and 4%Al addition were chosen as alloying elements.

Haoyu and co-workers[18]studied the susceptibility of Mg-Zn-Mn alloys and proposed that suitable amounts of Mn addition to Mg-6.5Zn alloy can reduce the hot tearing susceptibilities.Wang et al.[15]investigated the effect of Cu addition on the hot tearing susceptibility of Mg-6Zn-0.6Zr alloy and the results showed that 0.5%Cu-containing alloy exhibited lower hot tearing tendency.Therefore,in order to improve the hot tearing resistance of Mg-Zn-Al based alloys,0.5%Cu and 0.3%Mn were selected as alloying elements.Table 1 lists the prediction results of the influence of alloying elements on the hot tearing susceptibility(HTS)of Mg-Zn based alloys.It is shown that the addition of 1%and 4%Al,0.5%Cu,0.3%Mn to Mg-xZn(x=6,8,10,wt%)alloys in sequence can reduce their HTS,and they were selected as alloying elements.

Fig.1.Prediction of hot tearing susceptibility of Mg-Zn alloys based on steepness ofcurves in the fsrange of 0.9-0.99:(a)curves of T vs.;(b)predicted crack sensitivity essentially consistent with previous experiments[9];there was a hot tearing peak at 1.5wt%Zn.

Fig.2.Calculated results based on the index:(a)the effect of Zn content on the susceptibility of Mg-9Al-xZn alloy,(b)the effect of Al content on the susceptibility of Mg-1.5Zn alloy,(c)the effect of Cu content on the susceptibility of Mg-6Zn alloy,and(d)the effect of Y content on the susceptibility of Mg-2.5Zn-xY-0.5Zr alloy,wt%.

Table 1 Prediction values of hot tearing susceptibility of Mg-Zn based alloys.

Fig.3.Prediction of hot tearing susceptibility tendencies of Mg-Zn-Al ternary system.

Fig.4.The values of HTS(hot tearing susceptibility)are determined by the factors of the crack location(flocation),the crack bars length(flength),and the crack width(fwidth)in constrained rod casting:Eq.(1)was proposed by Argo et al.[19].

3.Experimental

3.1.Casting

Pure Mg,Zn,Al,Cu and Mg-10wt%Mn master alloys were preheated to 473K before melting.Pure Mg was first melted in an electrical resistance furnace under a protective mixture gas of high pure CO2and SF6.Pure Zn and Al were added at 973K,and Cu and Mg-10wt%Mn master alloys were added at 993K.When the melt was heated to 1003K,and kept for about 30min,it was cast into a constrained rod casting mold[19]to test the hot tearing susceptibility.The mold was coated with a thin layer of graphite and placed in a heat insulation box and preheated to 523K.Two hot tearing test pieces were cast for each alloy.The mold temperature and pouring temperature were kept at 523K and 983K,respectively.

3.2.Hot tearing measurement

In the present study,Eq.(1)in Fig.4 proposed by Argo et al.[19]was used to calculate the hot tearing susceptibility in the constrained rod casting apparatus.The values of HTS are determined by the factors of the crack location(flocation),the crack bars length(flength),and the crack width(fwidth).The crack length factor is defined to be 4 for the longest rod as it is most likely to form cracks,and is 8,16,and 32 for the second,third,and fourth longest rod,respectively.The location factor is 1 for cracking at the sprue end,where hot tears occur easily,2 at the ball end,and 3 in the middle of the rod.The width factor is 1 for short hair line,2 for full hair line,3 for crack and 4 for broken.The values of HTS calculated by Eq.(1)can reflect the hot tearing tendency of an alloy.

3.3.Microstructure analysis

In order to clearly reveal grain boundaries,all of as-cast samples were aged for 16h at 473K and then polished and chemically etched in a solution of nitroxanthic acid and alcohol before observed using an optical microscope(OM)under polarized mode.The average grain size was counted according to the ASTM E112-Heyn lineal intercept procedure(2004)[21].

4.Results and discussion

4.1.Evaluations of hot tearing susceptibility

Fig.5 shows the influence of chemical composition on the hot tearing susceptibility of Mg-Zn alloys.Fig.5a illustrates that 1%Al addition to Mg-6Zn alloy can reduce the HTS from 256 to 112,and either 0.5%Cu or 0.3%Mn addition to Mg-6Zn-1Al(ZA61)alloy can also reduce the HTS,from 112 to 92 or 82.However,adding 0.5%Cu and 0.3%Mn together enhances the HTS of ZA61 alloy,from 112 to 128.Fig.5b reflects that the HTS of ZA64(80)is lower than that of ZA61(112),indicating that 4%Al addition to Mg-6Zn alloy leads to a decrease of HTS,and with the increase of Al content,the HTS of Mg-6Zn alloy decreases distinctly.The influence of Cu and Mn addition on the HTS of ZA64 alloy is similar to that of ZA61 alloy.Adding Cu and Mn individually reduces the HTS of ZA64 alloy,but adding together deteriorates the hot tearing resistance.Overall,the HTS of Mg-6Zn-4Al based alloys(Fig.5b)is lower than that of Mg-6Zn-1Al based alloys(Fig.5a).

Fig.5.Effect of chemical composition on the hot tearing susceptibility of tested alloys under pouring temperature of 1003K and mold temperature of 523K:(a)ZA61(Mg-6Zn-1Al),ZA61-Cu(Mg-6Zn-1Al-0.5Cu),ZA61-Mn(Mg-6Zn-1Al-0.3Mn),ZA61-CuMn(Mg-6Zn-1Al-0.5Cu-0.3Mn);(b)ZA81(Mg-8Zn-1Al);(c)ZA64(Mg-6Zn-4Al);and(d)ZA101(Mg-10Zn-1Al).

Fig.5c shows that 1%Al addition to Mg-8Zn alloy can reduce the HTS from 108 to 72,while either 0.5%Cu or 0.3%Mn addition increases the HTS of ZA81 alloy,from 72 to 88 or 104.In addition,the addition of 0.5%Cu and 0.3%Mn together to ZA81 also improves the HTS from 72 to 136.The same tendency was observed in Mg-10Zn based alloys in Fig.5d,which have lower HTS than Mg-8Zn based alloys generally.At a higher content of Zn,the amount of low melt point eutectic phases increases,which improves the fluidity of the liquid and the tendency of hot tearing can also be alleviated[9].

As a whole,Al addition can significantly reduce the hot tearing susceptibility of Mg-xZn(x=6,8,10,wt%)alloys.Either 0.5%Cu or 0.3%Mn addition can decrease the HTS of ZA61 and ZA64 alloys,but the effect is contrary in ZA81 and ZA101 alloys.In addition,the addition of 0.5%Cu and 0.3%Mn together leads to higher susceptibility of all of the tested Mg-Zn-Al alloys.

4.2.Microstructure

Fig.6 presents the polarized optical microstructure images of tested alloys.For Mg-6Zn based alloys,from Fig.6a to c,the grain size decreases gradually with the addition of 1%Al and 0.5%Cu in succession to Mg-6Zn alloy.The addition of either 0.5%Cu or 0.3%Mn to ZA61 alloy can refine the grains,while the addition 0.5%Cu and 0.3%Mn together coarsen the grains,as shown in Fig.6b-e.The same change tendency can also be found in ZA64 based alloys,as shown in Fig.6f-i.In addition,the grain size of ZA64 alloy(177μm,Fig.6f)is less than that of ZA61 alloy(238μm,Fig.6b),which further reflects that the increase of Al content leads to a decrease of grain size.

For Mg-8Zn based alloys,from Fig.6j and k,the grain size decreases apparently with 1%Al addition.With the addition of 0.5%Cu and 0.3%Mn to ZA81 alloy either individually or together,the grain size increases,as shown in Fig.6k-n.The same change tendency exists in Mg-10Zn based alloys,as shown in Fig.6o-s.

Comparing the values of HTS with the average grain size of tested alloys,in Fig.5,it could be found that they have similar variation tendency.With the increase of grain size,the values of HTS increase and vice versa.The reason may be that grain refinement leads to an increase in the ductility and strength in mushy zone[5,20],and alleviates the initiation of hot tears.Therefore,the grain size is an important factor to influence the hot tearing susceptibility of Mg-Zn based alloys.

4.3.Comparison of thermodynamic calculation with experimental results

Fig.6.Optical microstructure images of tested alloys:(a)Mg-6Zn,(b)Mg-6Zn-1Al,(c)Mg-6Zn-1Al-0.5Cu,(d)Mg-6Zn-1Al-0.3Mn,(e)Mg-8Zn-1Al-0.5Cu-0.3Mn,(f)Mg-6Zn-4Al,(g)Mg-6Zn-4Al-0.5Cu,(h)Mg-6Zn-4Al-0.3Mn,(i)Mg-6Zn-4Al-0.5Cu-0.3Mn,(j)Mg-8Zn,(k)Mg-8Zn-1Al,(l)Mg-8Zn-1Al-0.5Cu,(m)Mg-8Zn-1Al-0.3Mn,(n)Mg-8Zn-1Al-0.5Cu-0.3Mn,(o)Mg-10Zn,(p)Mg-10Zn-1Al,(q)Mg-10Zn-1Al-0.5Cu,(r)Mg-10Zn-1Al-0.3Mn,and(s)Mg-10Zn-1Al-0.5Cu-0.3Mn,wt%.

Based on the calculation results in Table 1,the addition of Al,Cu,and Mn all should reduce the HTS of Mg-xZn(x=6,8,10,wt%)based alloys,which is the reason why those elements were picked up.However,it could be seen in Fig.7 that the variation tendency of HTS(experimental values)and the hot tearing index|ΔT/Δ(fs)0.5|(prediction)with chemical composition are not very consistent,especially for multicomponent alloys(C=5),such as Mg-6Zn-1Al-0.5Cu-0.3Mn.For Mg-6Zn-1Al based alloys,in Fig.7a,the experimental results show that the HTS of Mg-6Zn-1Al-0.5Cu-0.3Mn(ZA61-CuMn)is higher than that of Mg-6Zn-1Al(ZA61),against the calculation results of the index.The similar disagreement can be seen in Mg-6Zn-4Al based alloys in Fig.7b.The disagreements are much worse in Mg-8Zn and Mg-10Zn based alloys(Fig.7c and d).With 1%Al addition,the values of HTS decrease as the index predict,while with the addition of 0.5%Cu and 0.3%Mn,there are different variation tendencies between the experimental HTS and the calculations of the index.

As well known,the multi-component alloys have a complicated solidification path and complex hot tearing behavior.The phase diagrams of multi-component alloys in Pandat software are not well optimized;therefore,the calculation results of ΔT andfs,based on Pandat software and Mg database,are not completely true.In addition,the prediction index mainly concerns the tensile deformation and liquid re filling rate along the grain boundary,and may overlook some other factors,which can influence the nucleation of hot tearing,such as the grain size.It is worth mentioning that the values of the HTS and the grain size of tested alloys present similar variation tendencies,as shown in Fig.5.

Fig.7.Comparison between the experimental data(HTS)of tested alloys and the prediction values using indexwith the range of 0.9＜fs＜0.99:(a)ZA61(Mg-6Zn-1Al),ZA61-Cu(Mg-6Zn-1Al-0.5Cu),ZA61-Mn(Mg-6Zn-1Al-0.3Mn),ZA61-CuMn(Mg-6Zn-1Al-0.5Cu-0.3Mn-0.5Y);(b)ZA81(Mg-8Zn-1Al);(c)ZA64(Mg-6Zn-4Al);and(d)ZA101(Mg-10Zn-1Al),wt%.

Fig.8.Relativity of the original index,the grain size d,and modified indexes with different values of n(n=1,2),and the experimental results of hot tearing susceptibility.

Consequently,taking the grain size into consideration,the indexof HTS can be modified to,wheredis the average grain size,n is the constant value.Fig.8a and b shows that the relativity of the original indexand the experimental results(HTS)is only 0.43(R2,Fig.8a),even lower than the relativity of the grain sizedand the HTS(R2=0.66,Fig.8b).The relativity of the modified indexeswith different values ofn(n=1,2)and the experimental results(HTS)are shown in Fig.8c and d.The modified indexes taking grain size into consideration have better relativity than original index.Especially,the indexhas much better relativity(R2=0.94 in Fig.8d)with experimental results(HTS).Therefore,at least for Mg-Zn based alloys,the modified indexis a better prediction tool to design hot tearing resistant alloys.Unfortunately,the prediction of grain size based on the composition of an alloy is still an unsolved scientific issue.Therefore,it is still impossible to predict the hot tearing resistance of an alloy based on its chemical composition.Anyways,the modified indexfurther confirms the importance of grain refinement to Mg casting alloys.

5.Conclusion

The hot tearing index|ΔT/Δ(fs)0.5|recently proposed by Kou[7]and constrained rod casting were used to study the influence of alloying elements on hot tearing susceptibility(HTS)of Mg-Zn based alloys.Based on the index|ΔT/Δ(fs)0.5|and the thermodynamic calculations of Pandat software,Al,Cu,and Mn elements were chosen and their influence on hot tearing susceptibility of Mg-xZn(x=6,8,10,wt%)alloys was checked by experimental.Some conclusions can be drawn as follows:

(1)Al addition can significantly reduce the hot tearing susceptibility of Mg-Zn alloys.The addition of either 0.5%Cu or 0.3%Mn individually can reduce the HTS of Mg-6Zn-(1,4)Al alloys,while adding together increases the susceptibility.The addition of 0.5%Cu and 0.3%Mn individually or together increases the HTS of Mg-8/10Zn-1Al alloys.

(2)The indexalone cannot predict the influence of alloying elements on the hot tearing susceptibility very well based on the current thermodynamic calculations of Pandat software and its Mg database.It indeed provides a powerful tool to pick up the suitable elements and their contents to enhance the hot tearing resistance of Mg alloys.

(3)Based on the experimental and calculation results,the index can be modified tofor more accurate prediction on the hot tearing resistance of Mg-Zn based alloys.Grain refinement improves the hot tearing resistance of Mg-Zn based alloys significantly.

Acknowledgment

This research work is collaborated by General Motors Global Research and Development(GM R&D),Warren,MI,USA,and Shanghai Jiao Tong University(SJTU),Shanghai,China.It is supported by the National Key Research and Development Program of China(2016YFB0701204),Shanghai Rising-Star Program(15QB1402700),Special Fund of Jiangsu Province for the Transformation of Scientific and Technological Achievements(BA2016039).

[1]G.Cao,C.Zhang,H.Cao,Y.A.Chang,S.Kou,Metall.Mater.Trans.A 41(2010)706-716.

[2]U.Feurer,Giessereiforschung 28(1976)75-80.

[3]T.W.Clyne,G.J.Davies,Br.Foundryman 74(4)(1981)65-73.

[4]L.Katgerman,JOM 34(1982)46-49.

[5]N.N.Prokhorov,Russian Castings Production 2(1962)172-175.

[6]M.Rappaz,J.-M.Drezet,M.Gremaud,Metall.Mater.Trans.A 30(1999)449-455.

[7]S.Kou,Acta Mater.88(2015)366-374.

[8]G.Cao,S.Kou,Mater.Sci.Eng.A 417(2006)230-238.

[9]L.Zhou,Y.D.Huang,P.L.Mao,K.U.Kainer,Z.Liu,N.Hort,Int.J.Cast Metals Res.24(2013)170-176.

[10]A.Srinivasan,Z.Wang,Y.Huang,F.Beckmann,K.U.Kainer,N.Hort,Metall.Mater.Trans.A 44(2013)2285-2298.

[11]P.Gunde,A.Schiffl,P.J.Uggowitzer,Mater.Sci.Eng.A 527(2010)7074-7079.

[12]M.M.Avedesian,H.Baker,ASM Specialty Handbook:Magnesium and Magnesium Alloys,ASM International,1999.

[13]Y.Wang,Q.Wang,G.Wu,Y.Zhu,W.Ding,Mater.Lett.57(2002)929-934.

[14]L.Zhou,Y.Huang,P.Mao,K.U.Kainer,Z.Liu,N.Hort,Investigations on Hot Tearing of Mg-Zn-(Al)Alloys,John Wiley&Sons,Inc.,2011.

[15]Z.Wang,Y.Li,F.Wang,J.Song,Z.Liu,P.Mao,J.Mater.Eng.Perform 25(2016)5530-5539.

[16]Z.Liu,S.-B.Zhang,P.-L.Mao,F.Wang,Trans.Nonferr.Metals Soc.China 24(2014)907-914.

[17]J.Song,F.Pan,B.Jiang,A.Atrens,M.-X.Zhang,Y.Lu,J.Magnes.Alloys 4(2016)151-172.

[18]B.Yuanyuan,L.Haoyu,Z.Haitao,W.Xin,Z.Zhiqiang,L.Qichi,Acta Metall.50(2014)1237-1243.

[19]D.Argo,M.Pekguleryuz,P.Vermette,M.Lefebvre,in:Proceedings of NADCA Transactions,Cincinnati,Ohio,USA,2001,pp.181-186.

[20]D.G.Eskin,Suyitno,L.Katgerman,Progress in Mater.Sci.49(2004)629-711.

[21]Standard Test Methods for Determining Average Grain Size.ASTM designation E112-96e3.West Conshohocken(PA):ASTM;2004.