Journal of Materials Processing Technology 212 (2012) 1430– 1436 Contents lists available at SciVerse ScienceDirect Journal of Materials Processing Technology journa l h omepa g e: www.elsev ier .com/ Micros –3 proces R.G. Gua Liu a College of Ma b Department o th Kor a r t i c l Article history: Received 18 N Received in re 30 December 2 Accepted 2 Fe Available onlin Keywords: Semisolid Rheo-rolling Mg–3Sn–1Mn Vibration Microstructur Property lopin well equen f the nce t g tem 0 mm produ Introduction As an ad has been stu technique b to fabricate 2005). Up to developed, neto hydro plate techn direct chill et al., 2009 nique for pr this process and the sem Haga (2002 ing mill to d also carried et al. (2008 technique ( adhesion o gation duri ∗ Correspon E-mail add mechanical properties of the strip depend on its microstructure, 0924-0136/$ – doi:10.1016/j. vanced forming technology, semisolid metal processing died for more than thirty years since the finding of this y Flemings (1991). Semisolid forming is an ideal way magnesium alloy products (Martinez and Flemings, now, many semisolid processing techniques have been such as mechanical stirring (Brabazon et al., 2002), mag- dynamic (MHD) stirring (Jaworski et al., 2010), sloping ique (Cardoso Legoretta et al., 2008), melt conditioner (MC-DC) casting (Haghayeghia et al., 2010), etc. (Zhang ). The sloping plate processing is a convenient tech- eparing slurry due to its low cost and high efficiency. In , the alloy melt is poured onto a cooling sloping plate, isolid slurry is obtained after cooling and melt flow. ) has installed a sloping plate device onto a roll cast- evelop the semisolid rolling process. Many researchers out research on this technique, such as Kapranos ) and Grimmig et al. (2006) and Salarfar improved this Babaghorbani et al., 2006). However, the problems of f the slurry on the plate surface and the liquid segre- ng rheo-rolling process need to be resolved. Since the ding author. ress:
[email protected] (R.G. Guan). the microstructure of the strip should be controlled during the rheo-rolling process. In this investigation, vibrating sloping plate technique was developed, the problem of adhesion of the slurry on the plate surface was solved (Guan et al., 2007). In addition, a roll gap with constrained broadsides was designed to avoid the liq- uid segregation during the rheo-rolling. A semisolid rheo-rolling process was developed by combining the two techniques to pro- duce Mg–3Sn–1Mn (wt%) alloy strip. Microstructure evolution and mechanical properties were also studied in this investigation. Experimental procedure The experimental material was Mg–3Sn–1Mn (wt%) alloy whose main chemical compositions (wt%) are 3% Sn, 1% Mn, 0.5% Ca, 0.05% Cu, 0.01% Ni, 0.01% Fe, and Mg as a balanced. The liquidus and solidus temperatures of the alloy are 645 ◦C and 620 ◦C, respec- tively. Fig. 1 is a schematic illustration of the experimental appara- tus. The nominal diameter of the roller is 400 mm, The cross section size of the strip cast from semisolid slurry is 4 mm×160 mm. The maximum rolling speed is 22 m/min. The melt was poured onto the surface of the vibrating sloping plate. The semisolid slurry consist- ing of fine non-dendrites and remnant liquids can be obtained due to the high cooling rate caused by the sloping plate as well as stir- ring action caused by the vibration and metal flow. The semisolid slurry flowed directly into the roll gap between the convex and the see front matter © 2012 Elsevier B.V. All rights reserved. jmatprotec.2012.02.001 tructure evolution and properties of Mg sed by semisolid rheo-rolling na,∗, Z.Y. Zhaoa, H. Zhanga, C. Liana, C.S. Leeb, C.M. terials and Metallurgy, Northeastern University, Shenyang 110004, China f Materials Science and Engineering, Pohang University of Science and Technology, Sou e i n f o ovember 2011 vised form 011 bruary 2012 e 11 February 2012 (wt%) alloy e a b s t r a c t The high cooling rate caused by the s flow lead to a high nucleation rate as as well as dendrite growth and subs rosette primary grains. The change o solid fraction increases from the entra remained eventually. When the castin strip with a cross section of 4 mm×16 Mechanical properties of the present with the addition of 0.87 wt% Ce. locate / jmatprotec Sn–1Mn (wt%) alloy strip a ea g plate and stirring action caused by the vibration and metal as two primary grain growth patterns, direct globular growth t breakage, which causes the formation of fine spherical or primary grain shape is not obvious in the roll gap. However, o the exit of the roll gap. The spherical or rosette grains were perature is 670 ◦C, and the vibrating amplitude is 1.5 mm, the was produced. Homogeneous microstructure was obtained. ct were higher than that of Mg–3Sn–1Mn (wt%) alloy casting © 2012 Elsevier B.V. All rights reserved. R.G. Guan et al. / Journal of Materials Processing Technology 212 (2012) 1430– 1436 1431 Fig. 1. Schematic diagram of semisolid rheo-rolling. concave rollers, as shown in Fig. 1. The process has two main advan- tages. One is that the temperature of the semisolid slurry is much lower than that of the melt in conventional roll casting. Therefore, shorter solidification time is required for the semisolid alloy, and the rolling speed of the semisolid alloy can be very high. The process is expected to be developed as a high-speed semisolid roll-casting technique. The other advantage is that the microstructure and mechanical properties of the strip can be improved by this method. In order to study the microstructure of semisolid Mg–3Sn–1Mn (wt%) alloy, samples of the slurry at different positions were taken and quenched in water for microstructure observation. In addi- tion, the test machine was stopped during working, and the alloy remained in the roll gap was taken for microstructure observation. Microstructure observation was carried out under an OLYMPUS PMG51 metallographic microscope. Scanning electron microscope (SEM) was also used to analyze the distributions of alloy elements on the sloping plate. Mechanical properties of Mg–3Sn–1Mn (wt%) alloy strip produced by the process were investigated by tensile test at room temperature. Results and discussion Microstructure evolution during the process Fig. 2 shows microstructures of the quenched melt at differ- ent positions on the sloping plate surface. Near spherical or rosette primary grains can be observed. No obvious coarse dendrite can be found when the casting temperature is 670 ◦C, and the vibrating amplitude is 1.5 mm. It is also found that the primary grain size of the quenched microstructure decreases gradually from the upper position A to the bottom position D (the positions are shown in Fig. 1), and the average grain size decreases from 76 �m to 36 �m. Solid fraction increases from the upper position A to the bottom position D. The grain shape also becomes rounder and rounder, and the average roundness decreases from 1.94 to 1.79. As the solidifi- cation microstructure of the alloy is closely related to the nucleation rate of the melt, the above observation indicates that more nuclei were generated during melt flow on the sloping plate. It was believed that the formation of fine and spherical microstructure is favoured by a high number of heterogeneous nucleation sites on the cooling sloping plate (Wu et al., 2008.). However, there is no clear evidence for this deduction. Nucleation is closely related to two factors, the strong cooling ability of the sloping plate and the stirring caused by vibration. In this study, high undercooling was realized in the whole melt on the surface of the sloping plate cooled by the circulating water. This derives from two factors, convection heat transfer caused by the metal flow and vibration as well as strong cooling ability of the plate. As we know, there exists a metal flow velocity boundary layer near the plate Fig. 2. Micros (c) position C; tructures of the quenched melt at different positions of the sloping plate (corresponding (d) position D. to the positions A, B, C and D in Fig. 1): (a) position A; (b) position B; 1432 R.G. Guan et al. / Journal of Materials Processing Technology 212 (2012) 1430– 1436 Fig. 3. Adhesi there is no sol surface and According t layers depe flat plate su 0.04–0.029 boundary la layer. This i ature layer Fig. 4. The com plate surface w on behaviors of the melt on the plate surface under different vibration conditions: (a) wi idification shell; (c) microstructure of the solidification shell attached on the plate surfac a temperature boundary layer during the processing. o the theory of classic heat transfer, the thickness of two nds on the Prandtl number in the case of the flow on a rface. The Prandtl number of liquid metal is in the range which is very small. The thickness of the temperature yer is much larger than that of the velocity boundary mplies that the distribution of the velocity and temper- in most part of the melt are homogenous, except that a velocity gra above analy of the temp addition, th to the conv the cooling which is mu ing process parison of the microstructures of the directly quenched melt and the solidified alloy on ithout flow. thout vibration, a solidification shell is remained; (b) with vibration, e. dient exists in a tiny layer close to the plate surface. The sis ignores the vibration action. Actually, the thickness erature boundary layer is increased by the vibration. In e sloping plate has a strong cooling ability. According ection heat transfer formula, it can be calculated that rate of the melt on the plate surface can reach 1000 K/s ch higher than that of the melt in the conventional cast- . In this way homogenous undercooling can be realized the plate surface: (a) directly quenched in water; (b) solidified on the R.G. Guan et al. / Journal of Materials Processing Technology 212 (2012) 1430– 1436 1433 Fig. 5. Microstructure and element Sn distribution along a line in the quenched melt taken from the position B in Fig. 1: (a) SEM micrograph; (b) element Sn distribution. in most part of the melt during the vibrating sloping plate process- ing, which is favorable for eruptive nucleation in the whole melt on the plate surface. The exp on the plate whereas th shown in Fi easily adhe each other. surface. Wh face will esc greatly enh ally promot Microstruct face is show along the n nucleus form before the v the plate wa mal directio this directio same time, eral directio condition, w Under the v face were d did not form of the quenched melt becomes smaller at the bottom part com- pared with the upper part. It is important to increase the effective nucleus density as the grain size is determined by the nucleation ener tem evel. ratur ions, lloy en th fusio ccele the h eneo the c and Whe plat form micr ed al ere a wate gra e coo tion nder Fig. 6. Micros erimental results show that the adhesion of the slurry surface is obvious before the application of vibration, e surface is smooth when the vibration is applied, as g. 3. When the sloping plate is stationary, the nuclei can re to the plate surface and grow until they contact with Finally, a solidification shell will remain on the plate en the vibration is applied, the nuclei on the plate sur- ape from the plate surface and enter into the melt. This ances the quantity of the effective nucleus and eventu- es the formation of fine and spherical solid structure. ure of the solidification shell attached on the plate sur- n in Fig. 3(C). Some columnar grains can be seen to grow ormal direction of the plate surface. This is because the ed on the plate surface could not escape from the plate ibration was applied. In addition, the cooling ability of s very strong, and the heat dissipated mainly along nor- n of the plate, which caused the nuclei to grow along n, leading to the formation of columnar grain. At the the columnar grains were seen to coarsen along the lat- n until they contacted with each other under stationary hich leaded to the formation of the solidification shell. ibration condition, the nuclei formed on the plate sur- ispersed in the melt. In this case, the solidification shell on the plate surface. It can be seen that the grain size rate. G casting a low l tempe condit (wt%) a Wh the dif ity is a causes homog when solute ation. cooling in the of the solidifi that th in the rosette that th nuclea large u tructures of the quenched slurry prepared by the vibrating sloping plate processing cast a ally, effective nucleus decreases with the increase of perature, so the casting temperature should be kept at By considering the cast ability of the alloy, the casting e should not be too low. Under the current experimental proper casting temperature of 670 ◦C for Mg–3Sn–1Mn was selected. e melt gradually flows down along the sloping plate, n coefficient increases, and the element diffusion veloc- rated because of the metal flow and vibration, which omogeneous solute distribution. At the same time, the us temperature field in the melt is formed. In this case, ritical condition for nucleation is satisfied, the uniform temperature distributions can induce eruptive nucle- n the low temperature melt is rapidly cooled by the e, the eruptive nucleation takes place and finally results ation of fine spherical primary grains. The comparison ostructures of direct quenched melt in water and the loy on the plate surface are shown in Fig. 4. It is observed re obvious dendrites, and the grain ripened completely r-quenched melt, but there are some fine spherical or ins in the solidified alloy on the plate, which implies ling plate can provide a big cooling rate to increase the rate. Once the melt was poured onto the cooling plate, a cooling could occur, which resulted in the appearance t 690 ◦C under different vibrating amplitudes: (a) 0.5 mm; (b) 1.5 mm. 1434 R.G. Guan et al. / Journal of Materials Processing Technology 212 (2012) 1430– 1436 Fig. 7. M of eruptive grains (Lego In additi directly wit also be foun that the hap Direct globu the conditi twin-screw process (Gu found that t with the ab direct globu ture distrib that determ growth patt It is difficult case, the pr formation o the stirring cause an id tions showe �-Mg and M icrostructures of the solidified melt in the roll gap: (a) on the whole the section along th nucleation and the formation of fine spherical or rosette retta et al., 2008.). on, it has been found that the primary grain could grow h a globular pattern (Kim and Lim, 2000), which can d in some areas in Fig. 2. However, it is worth noticing pening conditions for this behavior are quite different. lar growth is generally considered to take place under on that the melt is subjected to shearing action, e.g., stirring process (Ji et al., 2001), shear cooling roll (SCR) an et al., 2009). However, in the current studies, it is he direct globular growth still exists in other processes sence of a strong shearing. So the reason that causes lar growth is not shearing force, solute and tempera- ution as well as nucleation rate are the essential reasons ine the grain growth pattern. Besides the direct globular ern, dendrite growth and breakage were also observed. to avoid constitutional supercooling completely. In this otuberance of the spherical interface will result in the f dendrites along the preferential direction. Perhaps, action caused by the vibration and metal flow cannot eal dendrite to break. However, previous investiga- d that Mg–3Sn–1Mn (wt%) alloy is mainly composed of g2Sn phases, no any Mn-containing phases are formed, so the micr Mn and are The aggreg sponding a heat flux an further and tion of the is shown in show the se aries are vi are weak a vibration an greatly affe amplitude, tude increa conditions, be 1.5 mm. The high good micro when the c tude is 1.5 results and used to pre e rolling direction; (b) position O; (c) position P; (d) position Q. ostructures of the alloys are less affected by the element mainly affected by Sn distribution (Yang et al., 2011). ation of Sn at dendritic arm root can cause the corre- reas to be weak, and these areas may neck down under d stirring easily. The dendritic arm fragments can grow evolve into fine spherical grains. Element Sn distribu- quenched melt taken from the position C on the plate Fig. 5. After etching, the positions m and n in Fig. 5 gregations of the elements. Many primary grain bound- sible after etching, which also proves that these areas nd can be easily broken under the actions of heat flux, d metal flow. The microstructures of the alloys are also cted by the vibrating amplitude. In a scope of vibrating microstructure becomes finer as the vibrating ampli- ses, as shown in Fig. 6. Under the current experimental the reasonable vibrating amplitude was suggested to quality Mg–3Sn–1Mn (wt%) alloy semisolid slurry with structures can be prepared by the vibrating sloping plate asting temperature is 670 ◦C, and the vibrating ampli- mm, as shown in Fig. 2(d). Based on the experimental the previous reports, the vibrating sloping plate can be pare other Mg alloy semisolid slurry (Guan et al., 2007). R.G. Guan et al. / Journal of Materials Processing Technology 212 (2012) 1430– 1436 1435 Fig. 8. The pr strip obtained The effects alloy are ap The sem device ente the concave the slurry w fied melt in microstruct the primary solid fractio of the roll g the melt is shape is not so the solid ond solidifi the interfac existed. Alt could hinde grains were Microstructu strip produc When t amplitude a cross sec reported th alloy can re of the alloy refinement additions o ous dendrit Mg–3Sn–1M was 165 MP In the prese (wt%) alloy grains, and ultimate ten and the elo fore, the ul Fig tively lloy ably ent and f 0.87 esent etter rheo uld b sion his p alloy Mg– s we e stro ion c cleat oduct and three-dimensional microstructures of Mg–3Sn–Mn alloy at the casting temperature of 670 ◦C. of vibrating sloping plate on the Mg–3Sn–1Mn (wt%) plicable to other Mg alloys (Guan et al., 2007). isolid slurry prepared by the vibrating sloping plate red into the rectangular gap between the convex and roller. Under the action of the effective friction force, as rolled directly. The microstructures of the solidi- the roll gap are shown in Fig. 7. It can be seen that the ure consists of spherical or rosette grains. The change of grain shape is not obvious in the roll gap. However, the n can be found to increase from the entrance to the exit ap. The reason is that the shearing time of the rollers on short in the roll gap, so the change of the primary grain obvious. But the alloy solidified further in the roll gap, fraction increases gradually. However, during the sec- cation period in the roll gap, the melt was subjected to ial friction of the rollers, and the shearing action still hough the shearing time was short, the shearing action r the formation of dendrite, and the spherical or rosette respec (wt%) a is prob refinem tensile tions o The pr with b by the ties co Conclu In t (wt%) ties of proces (1) Th act nu maintained. res and properties of the Mg–3Sn–1Mn (wt%) alloy ed by the present process he casting temperature is 670 ◦C, and the vibrating is 1.5 mm, the Mg–3Sn–1Mn (wt%) alloy strip with tion size of 4 mm×160 mm can be obtained. It was at the addition of 0.87 wt% Ce to the Mg–3Sn–1Mn fine the grains and improve the tensile and elongation (Pan and Yang, 2011). The average grain size before was 275 �m for the alloy and it was 102 �m with the f 0.87 wt% Ce (Pan and Yang, 2011). However, obvi- es have been found. The ultimate tensile strength of n (wt%) alloy casting with the addition of 0.87 wt% Ce a, and the elongation was 3.9% (Pan and Yang, 2011). nt work, the internal microstructures of Mg–3Sn–1Mn strip were mainly composed of fine spherical or rosette the average grain size was 47 �m, as shown in Fig. 8. The sile strength of Mg–3Sn–1Mn (wt%) alloy was 175 MPa, ngation to failure was 5.6%. as shown Fig. 9. There- timate tensile strength and elongation are improved direct g quent b rosette (2) The cha gap, but of the r eventua (3) When t amplitu a cross present Mechan that of 0.87 wt Acknowled The auth Foundation Chinese Na ment under . 9. Stress-strain curve of Mg–3Sn–1Mn (wt%) alloy strip. by 24.8% and 84.6% in comparison with Mg–3Sn–1Mn casting with the addition of 0.87 wt% Ce. This situation related to the grain refinement. Therefore, the grain effects are probably responsible for the difference in elongation properties between the alloys with the addi- wt% Ce and the alloy produced by the present process. technique can produce Mg–3Sn–1Mn (wt%) alloy strip mechanical properties. Additionally, the strip produced -rolling usually needs deep rolling, so the final proper- e improved significantly. s aper, a semisolid rheo-rolling process of Mg–3Sn–1Mn was proposed. Microstructure evolution and proper- 3Sn–1Mn (wt%) alloy strip processed by the proposed re studied. The following conclusions were drawn: ng cooling rate by the sloping plate and the stirring aused by the vibration and metal flow lead to high ion rate and two primary grain growth patterns, the lobular growth as well as dendrite growth and subse- reakage, leading to the formation of fine spherical or primary grains. nge of the primary grain shape is not obvious in the roll the solid fraction increases from the entrance to the exit oll gap. The spherical or rosette grains were remained lly. he casting temperature is 670 ◦C, and the vibrating de is 1.5 mm. Mg–3Sn–1Mn (wt%) alloy strip with section size of 4 mm×160 mm was prepared by the process. Homogenous microstructure can be obtained. ical properties of the present product were higher than Mg–3Sn–1Mn (wt%) alloy casting with the addition of % Ce. gements ors thanks for the supports of National Natural Science of China under grant Nos. 51034002 and 50974038, and tional Program for Fundamental Research and Develop- grant No. 2011CB610405. 1436 R.G. Guan et al. / Journal of Materials Processing Technology 212 (2012) 1430– 1436 References Brabazon, D., Browne, D.J., Carr, A.J., 2002. Mechanical stir casting of aluminium alloys from the mushy state: process, microstructure and mechanical properties. Materials Science and Engineering A 326, 370–381. Babaghorbani, P., Salarfar, S., Nili-Ahmadabadi, M., 2006. Kinetics of coarsening and solid sphericity during reheating of ductile iron and Al alloys. Diffusion and Defect Data B, Solid State Phenomena 116–117, 205–208. Cardoso Legoretta, E., Atkinson, H.V., Jones, H., 2008. Cooling slope casting to obtain thixotropic feedstock I: observations with a transparent analogue. Journal of Materials Science 43, 5448–5455. Flemings, M.C., 1991. Behavior of metal alloys in the semisolid state. Materials Trans- actions A 22, 957–981. Grimmig, T., Ovcharov, A., Afrath, C., Bunck, M., Buhrig-Polaczek, A., 2006. Poten- tial of the rheocasting process demonstrated on different aluminum based alloy systems. Diffusion and Defect Data B, Solid State Phenomena 116–117, 484–488. Guan, R.G., Xing, Z.H., Shi, L., Wang, C., Wang, Y., 2007. Preparing semisolid light alloys by vibrating wavelike sloping plate process and thixoforming. Materials Science Forum 561–565, 865–868. Guan, R.G., Chen, L.Q., Li, J.P., Wang, F.X., 2009. Dynamical solidification behaviors and metal flow during continuous semisolid extrusion process of AZ31 alloy. Journal of Materials Science and Technology 25, 395–400. Haga, T., 2002. Semisolid strip casting using a twin roll caster equipped with a cooling slope. Journal of Materials Processing Technology 130–131, 558–561. Haghayeghia, R., Zoquib, E.J., Greenc, N.R., Bahaia, H., 2010. An investigation on DC casting of a wrought aluminium alloy at below liquidus temper- ature by using melt conditioner. Journal of Alloys and Compounds 502, 382–386. Jaworski, M.A., Gray, T.K., Antonelli, M., Kim, J.J., Lau, C.Y., Lee, M.B., Neumann, M.J., Xu, W., Ruzic, D.N., 2010. Thermoelectric magnetohydrodynamic stirring of liq- uid metals. Physical Review Letters 104, 1–4. Ji, S., Fan, Z., Bevis, M.J., 2001. Semi-solid processing of engineering alloys by a twin-screw rheomoulding process. Materials Science and Engineering A 299, 210–217. Kim, K.S., Lim, S.G., 2000. Microstructure of A356 semi-solid metal fabricated by inclination cooling plate. Journal of the Korean Institute of Metals and Materials 38, 529–533. Kapranos, P., Haga, T., Bertoli, E., Pola, A., Azpilgain, Z., Hurtado, I., 2008. Thixo- extrusion of 5182 aluminium alloy. Diffusion and Defect Data B, Solid State Phenomena 141–143, 115–120. Legoretta, E.C., Atkinson, H.V., Jones, H., 2008. Cooling slope casting to obtain thixotropic feedstock II: observations with A356 alloy. Journal of Materials Sci- ence 43, 5456–5469. Martinez, R.A., Flemings, M.C., 2005. Evolution of particle morphology in semisolid processing. Metallurgical and Materials Transactions A 36, 2205–2210. Pan, F.S., Yang, M.B., 2011. Preliminary investigations about effects of Zr, Sc and Ce additions on as-cast microstructure and mechanical properties of Mg–3Sn–1Mn (wt%) magnesium alloy. Materials Science and Engineering A 528, 4973–4981. Wu, S.S., Xie, L.Z., Zhao, J.W., Nakae, H., 2008. Formation of non-dendritic microstruc- ture of semi-solid aluminum alloy under vibration. Scripta Materialia 58, 556–569. Yang, M.B., Qin, C.Y., Pan, F.S., 2011. Effects of heat treatment on microstructure and mechanical properties of Mg–3Sn–1Mn magnesium alloy. Transactions of Nonferrous Metals Society of China 21, 2168–2174. Zhang, X.L., Li, T.J., Xie, S.S., Wang, T.M., Cao, Z.Q., Jin, J.Z., 2009. Research and devel- opment of semisolid slurry processing technology. Rare Metal Materials and Engineering 38, 1495–5000. Microstructure evolution and properties of Mg–3Sn–1Mn (wt%) alloy strip processed by semisolid rheo-rolling Introduction Experimental procedure Results and discussion Microstructure evolution during the process Microstructures and properties of the Mg–3Sn–1Mn (wt%) alloy strip produced by the present process Conclusions Acknowledgements References