Friction Stir Processing of Cast Magnesium Alloys

Home , Friction stir processing, Magnesium alloy

Scholars' Mine

Masters Theses Student Theses and Dissertations

Spring 2007

Friction stir processing of cast magnesium alloys

Timothy A. Freeney

Follow this and additional works at: https://scholarsmine.mst.edu/masters_theses

Part of the Materials Science and Engineering Commons Department:

Recommended Citation Freeney, Timothy A., "Friction stir processing of cast magnesium alloys" (2007). Masters Theses. 4613. https://scholarsmine.mst.edu/masters_theses/4613

This thesis is brought to you by Scholars' Mine, a service of the Missouri S&T Library and Learning Resources. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more information, please contact [email protected].

FRICTION STIR PROCESSING OF CAST MAGNESIUM ALLOYS

TIMOTHY ALAN FREENEY

A THESIS

Presented to the Faculty of the Graduate School of the

UNIVERSITY OF MISSOURI-ROLLA

In Partial Fulfillment of the Requirements for the Degree

MASTER OF SCIENCE IN MATERIALS SCIENCE AND ENGINEERING

2007

Approved by

______Dr. Rajiv S. Mishra, Advisor Dr. F. Scott Miller

______Dr. Von L. Richards

iii

PUBLICATION THESIS OPTION

This thesis consists of the following two papers that have been submitted for publication as follows:

Pages 1–23 are expanded publication in TMS 2007 Friction Stir Welding and

Processing IV Symposium.

Pages 24–53 are intended for submission in Materials Science and Engineering A.

ABSTRACT

This thesis investigates the feasibility and benefits of using friction stir processing as a thermo-mechanical microstructural modification tool for local property enhancement or casting repair of magnesium alloys, specifically high strength EV31A and WE43.

Castings often have usage limitations due to the inherent microstructural inhomogeneity from solidification; in addition magnesium cast alloys generally suffer from poor ductility and relatively low strength, yet the low density of magnesium is very attractive for weight critical applications.

Friction stir processing (FSP) is a proven tool for enhancement of aluminum alloys, imparting a fine-grained, homogeneous microstructure. The alloys in this study are designed for elevated temperature service in the aerospace and high performance automotive industries and are suitable in applications were lightweight and exceptional strength are desired. WE43 has excellent temperature resistance up to 300°C, and EV31A is a more industry cost-effective alternative to WE43 and has excellent strength retention up to 200°C. Results from the present work showed that both alloys are easily adapted to the use of FSP, and significant strength and ductility gains were achieved in both alloys over their respective as-cast peak strengthened states.

ACKNOWLEDGMENTS

I thank first and foremost my advisor; Dr. Rajiv Mishra, for his patience, guidance and support throughout my undergraduate and graduate studies. Dr. Mishra’s dedication to the group and his students has demonstrated that kindness fosters diligence. His generosity and mentorship have made my stay at UMR not only possible in many ways, but also enjoyable. In addition, I am grateful to my committee members, Dr. F. Scott Miller and Dr. Von L. Richards, for their guidance. I also take this opportunity to thank Dr. David G. Robertson for providing kind support and an open door through many difficult times. Also, I thank Dr. Kent D. Peaslee for being a pivotal figure of my introduction to this field and university. I am truly indebted to the Metallurgical Engineering faculty at the University of Missouri-Rolla, as they have been tremendous as educators and mentors; and I truly believe that without their caring and patience I would not have completed my studies. I also express my gratitude to Joyce Erkiletian, Pricilla Winner and Paula Cochran for their kind assistance and wonderful personalities. I further extend my warm thanks to all of my dear colleagues who have come and gone throughout the years; especially, Siddharth Sharma whose patience, mentorship and friendship were crucial to my development as a researcher and person. I also thank Indrajit Charit, Lucie Johannes, Vikas and Manisha Dixit, Dr. Yanwen Wang, Manasij Yadava; and all of the past and present group members who have assisted my learning and provided friendship through my tenure at UMR. I also thank my dear friends Steve, Bret, Jeff, Brad, Greg, Mark; and all of my friends from Sigma Nu and otherwise. This thesis is dedicated to my parents who brought me into this world; to my grandparents who have fostered my desire for education; and to all my family for shaping the person I am today. I will never forget my humble beginnings, and to those to whom I am indebted, “ Pigmaei gigantum humeris impositi plusquam ipsi gigantes vident”.

TABLE OF CONTENTS

Page PUBLICATION THESIS OPTION...... iii ABSTRACT...... iv ACKNOWLEDGMENTS ...... v LIST OF ILLUSTRATIONS...... viii LIST OF TABLES...... x PAPER I Effect of friction stir processing parameters on the microstructural and mechanical properties of high strength magnesium cast alloy Abstract ...... 1 1. Introduction ...... 1 2. Alloying Elements...... 3 2.1. Mg-Zr...... 3 2.2. Mg-Y-Nd ...... 3 3. Experimental Approach...... 4 4. Results and Discussion...... 6 4.1. Microstructure...... 6 4.1.1. As-Cast...... 6 4.1.2. Cast + T6...... 7 4.1.3. FSP Microstructure ...... 8 4.2. Material Flow...... 13 4.3. Mechanical Properties...... 15 5. Conclusions ...... 20 Acknowledgments...... 20 References ...... 20

vii

Page PAPER II Effect of friction stir processing on the microstructure and mechanical properties of a cast magnesium-rare earth alloy Abstract ...... 24 1. Introduction ...... 24 2. Effect of Alloying Elements...... 26 2.1 Mg-Zr...... 27 2.2 Mg-Zn ...... 28 2.3 Zr-Zn...... 28 2.4 Zn-RE...... 29 2.5 Mg-Nd-Gd ...... 29 3. Experimental ...... 29 4. Results and Discussion...... 31 4.1 Microstructure...... 31 4.1.1 Microstructure – As-Cast...... 31 4.1.2. Microstructure – As-Cast + T6 ...... 34 4.1.3. Microstructure – FSP ...... 37 4.1.4. Microstructure – FSP + HT ...... 40 4.2. Mechanical Properties...... 42 4.3. Fractography...... 44 4.4. DSC Results...... 45 5. Conclusions ...... 47 Acknowledgments...... 47 References ...... 47 VITA ...... 54

viii

LIST OF ILLUSTRATIONS

Figure Page PAPER I Effect of friction stir processing parameters on the microstructural and mechanical properties of high strength magnesium cast alloy 1. SEM images of as-cast WE43...... 7 2. OM (left) and SE SEM image (right) of cast + T6 condition. The location of the α- Zr clusters is apparent and is generally centrally located in a grain or on the boundary. SEM image shows incomplete dissolution of eutectic phase and solidification defects ...... 8 3. SEM (a,b) and OM (c) image of nugget grains at respective increasing heat inputs of 400 rpm 102 mm/min, 700 rpm 153 mm/min and 900 rpm 127 mm/min in the T5 condition...... 10 4. (a) Variation of mean linear grain size in the central region of the nugget with rotation rate. (b) A plot of frequency variation of grain size in the central region of the nugget for different heat indices ...... 11 5. Variation of mean linear grain size in the central region of the nugget with process pitch, which effectively demonstrates the heat input per unit volume of material processed...... 12 6. OM of WE43 FSP nugget in T6 condition illustrating abnormal grain growth caused by the instability of the fine-grained microstructure...... 13 7. Macro image of nugget geometry and onion ring structure for 400 rpm 102 mm/min, 700 rpm 153 mm/min and 900 rpm 127 mm/min...... 14 8. Variation of material flow regions with pitch ...... 15 9. Variation of tensile strength with various measures of processing thermal input.. 17 10. Variation of yield strength with various measures of processing thermal input... 18 11. Variation of elongation with various measures of processing thermal input...... 19

PAPER II Effect of friction stir processing on the microstructure and mechanical properties of a cast magnesium-rare earth alloy 1. Schematic of friction stir processing [8] ...... 25 2. Image of FSP EV31A surface texture ...... 30 3. BSE of EV31A in F temper...... 32 4. BSE of as-cast microstructure ...... 33 5. Variation of Nd, Zn and Gd content as a function of Mg concentration in matrix and intergranular phase of EV31A in as-cast condition...... 34 6. BSE image of EV31A in T6 condition, showing the intergranular Mg-Nd-Gd-Zn phase and Zr-Zn clusters...... 35

7. BSE of Zr 2Zn clusters in EV31A formed after T6 treatment at prior α-Zr rich clusters. Not all clusters were identical as shown; image on left has no apparent central core, while image on right has a clear core structure of same composition

as needles. Note that clusters analyzed contained Zr 3Zn 2...... 36 8. Variation of Zn, Nd and Gd content as a function of Mg concentration in matrix and intergranular phase of EV31A in T6 condition...... 37 9. Macro image of nugget cross section...... 39 10. BSE of as friction stir processed region in EV31A unetched and etched...... 39 11. BSE of friction stir processed region in EV31A after T6 heat treatment ...... 41 12. BSE of friction stir processed region in EV31A aged for 16 hrs at 200°C ...... 42 13. Representative tensile test curves for EV31A in various processing conditions at a strain rate of 10 -3 s -1 ...... 43 14. Average room temperature tensile properties according to experimental schedule at a strain rate of 10 -3 s -1...... 44 15. SEM images of fracture surface of EV31A in T6 condition and FSP condition.. 45 16. DSC analysis of samples in the FSP condition compared to the T6 condition..... 46

LIST OF TABLES

Table Page PAPER I Effect of friction stir processing parameters on the microstructural and mechanical properties of high strength magnesium cast alloy 1. Summary of probable precipitation sequence for Mg-Nd alloy...... 4 2. Summary of processing parameters and heat treatment conditions...... 5 3. Summary of mechanical properties for as-cast and cast + T6 heat treatments...... 16

PAPER II Effect of friction stir processing on the microstructure and mechanical properties of a cast magnesium-rare earth alloy 1. Physical properties of EV31A...... 26 2. Summary of probable precipitation sequence...... 27 3. Experimental heat treatment schedule with aging time of 16 hrs ...... 30 4. EDS analysis of as-cast microstructure...... 33 5. Results of EDS analysis of T6 microstructure ...... 36

PAPER I

Effect of friction stir processing parameters on the microstructural and mechanical properties of high strength magnesium cast alloy

T.A. Freeney and R.S. Mishra

Center for Friction Stir Processing, Materials Science and Engineering, University of Missouri, Rolla, MO 65409, USA

Keywords : Friction Stir Processing, WE43, cast magnesium, Yttrium, grain size

Abstract A heat treatable sand cast WE43 magnesium alloy was friction stir processed (FSP) at 3 heat indices to determine the effect on microstructural and mechanical properties of the processed region. In addition, the sequence of solutionizing and aging heat treatment was investigated to maximize the processing efficiency, which is defined as the percent increase of FSP nugget yield strength over base yield strength. The maximum processing efficiency achieved for the casting was 125%, when a yield strength of 250 MPa was achieved in the FSP + T5 heat treatment condition. Compared to a yield strength of 202 MPa in the cast + T6 condition, this increase in mechanical properties is an attractive prospect for the local enhancement of Mg-Y-Nd-Zr sand castings.

1. Introduction Cast magnesium alloys are becoming increasingly popular for use as structural components in the automotive and aerospace industries. The growing need among these industries to lower fuel consumption and satisfy emission guidelines has increased the attractiveness of using magnesium to replace components currently made from steel, aluminum, and polymer composites [1,2]. The use of casting processes in these applications is beneficial due to the ease of creating complex shapes and internal cavities,

2 which ultimately can result in the reduction of total part cost for assembly [3]. The benefit of magnesium casting is highlighted by exceptional fluidity, fast cycle time due to low latent heat, and good die life. State-of-the-art magnesium alloys offer high specific strength at room and elevated temperatures, good creep and corrosion properties, good castability, exceptional machineability and dampening characteristics [4]. Despite all these benefits of the casting process, mechanical properties can vary significantly due to solidification issues such as coarse interdendritic phases, large grain size, and interdendritic microporosity [5,6]. FSP has been proven to be an effective technology for the production of a fine- grained, homogenous, porosity-free microstructure in cast aluminum alloys, A356 and A319 [6,7]. In the studies by Ma et al. [6,8,9] on A356, a significant increase in mechanical properties, strength, ductility and fatigue life in the nugget region was reported. In magnesium casting, FSP may be a viable process for local mechanical property improvement as well as for the repair of casting defects. FSP has been demonstrated on magnesium-aluminum-manganese (AM) and magnesium-aluminum- zinc (AZ) casting alloys AM60B, AM50, AZ91/6, AZ91D4, AZ95D/5 [4,10,11,12,13]. The aluminum-manganese and aluminum-zinc alloy series, such as in the previous studies, are relatively inexpensive due to the relatively low cost of the alloying additions and ease of liquid alloying when compared to rare earth additions [14]. A notable study of single and multiple pass FSP of Mg-Y-Zn cast alloy demonstrated that a significant reduction in grain size was achieved while multiple passes resulted in no significant coarsening of grains in prior passes [15]. WE43 is a high strength casting alloy developed by Magnesium Elektron with excellent high temperature strength retention, corrosion resistance and long-term temperature stability up to 250°C. However, the use of rare earth elements in this alloy (Y, Nd), as well as required attention to melt practices, make the alloy more costly [16]. And cost is important when the decision is whether to repair or to scrap a casting. In the case of expensive castings, repair and modification processes like FSP can become justifiable from a cost perspective.

2. Alloying Elements 2.1. Mg-Zr An excellent summary of the current state of knowledge of grain refinement in cast Mg alloys is given by St. John et al [17]. There are several methods for grain refinement of Mg alloys during casting, mainly: superheating, carbon inoculation, incorporation of Fe or Al to form Fe and Al-rich intermetallics [18], and Zr addition [17], such as, in EV31A. The α-Zr cored structure, which is a characteristic phase in Mg-Zr type alloys, is believed to be significant in grain refinement during solidification [19,20,21]. The structure is thought to be formed through a peritectic reaction upon solidification [19,21]. Current understanding is that zirconium contributes to very effective grain refinement by two primary mechanisms: constitutional undercooling due to a high growth restriction factor [20]; and grain nucleation due to magnesium and zirconium having the same crystal structure and similar lattice parameters [19]. The end result of this reaction is spherical clusters of α-Zr particles in Mg matrix. For later discussion, note that Zr does not change the decomposition behavior of the super saturated solid solution (SSSS) [22] and thus will be ignored.

2.2. Mg-Y-Nd Neodymium has a relatively high solubility in Mg (12.5%) and is generally added to promote age hardenability and high temperature creep resistance [23]. A compilation of data for Mg-RE alloys [22] discusses current knowledge of compounds present in this alloy system. The magnesium-rich portion of the Mg-Nd diagram [24,25,26,27] shows

two compounds, Mg 12 Nd and Mg 41 Nd 5, of which the latter melts incongruently and the former is reported to be a metastable phase [24] due to its ability to decompose into a Mg solid solution near its formation temperature [22]. Although Mg 12 Nd is reported to be metastable, its existence has been reported in numerous alloys, at different cooling rates [27,28,29,30] and is part of the accepted Mg-Nd precipitation sequence (Table 1).

Table 1. Summary of probable precipitation sequence for Mg-Nd alloy

Mg-Nd [23] SSSS GP Zones β" β' β

plates D019 Mg 3Nd Mg 12 Nd

hcp plates fcc plates bct

Mg-Nd-RE SSSS β" β' β1[32] β

(Gd,Dy,Y) D019 [31] bc ortho fcc fcc

[23] platelets globular plates

The solubility of Nd in the equilibrium phase Mg 24 Y5 and the solubility of Y in the

equilibrium phase Mg 41 Nd 5 were assumed by Sviderskaya et al [33] and confirmed by Rokhlin et al [34]. This study determined that when compared to the binary diagram of Mg-Y [35], although precipitates of the same crystal structure form, decomposition of the ternary system occurred much faster than in the respective binary systems [34]. With the

addition of Nd to Mg-Y alloys, the stable Mg 24 Y5 phase was reported to change to a fcc structure isomorphous with Mg 5Gd [36,37,38].

3. Experimental Approach In this study, the as-received material was a sand cast WE43 component having a nominal composition of 4.0Y-2.25Nd-1.0RE-0.6Zr-bal. Mg (in wt. %). The component was sectioned into six approximately 1 ″ thick pieces that were machined so the backing and processing surfaces were parallel. Single pass FSP was performed on all samples such that adequate time was given in between runs for the backing plate and tool to return to room temperature. The processing tool made from MP159 had a 15.5 mm shoulder

with a conical threaded pin with d root = 7.1 mm, d tip = 4.9 mm and a length of 5 mm. For mechanical testing, mini-tensile samples with 4.2 mm gage length and 1 mm gage width were machined from the nugget oriented with gage length being entirely in the nugget through transverse section. Tensile samples were ground to a final thickness of 0.5-0.55

5 mm and given a 1 µm polished finish before testing. Tensile testing was done at a strain rate of 1x10 -3 s-1. Since heat treatable alloys are rarely used in the as-cast condition, the effects of heat treatment on the nugget properties are important. To understand the effect of heat treatment on microstructural and mechanical properties, the sequence of heat treatment was varied before and after FSP according to the experimental schedule shown in Table 2. The T6 treatment for WE43 casting involved solutionizing (SS - 520°C for 8h), warm water quench (60-80°C), and aging (250°C for 16h). The T5 treatment involved aging at 250°C for 16h. Metallographic samples were mechanically polished to 0.05 µm and etched using an acetic glycol etchant [23]. Images for microstructural analysis were taken with the use of optical microscopy (OM) and scanning electron microscopy (JEOL 840A); qualitative analysis was done with ImageJ software package. Grain size was determined through the mean linear intercept method where spatial grain size = mean linear intercept length x 1.78.

Table 2. Summary of processing parameters and heat treatment conditions

Tool Tool Heat treatment schedule Rotation Traverse Rate Speed Cast + Cast + FSP Cast + FSP SS + FSP (rpm) (mm/min) FSP +T6 +T5 +T5

400 102 x x x x

700 153 x x x -

900 127 x - x x

4. Results and Discussion 4.1. Microstructure 4.1.1. As-Cast Figure 1 shows SEM images of the polished as-received WE43 casting. The microstructure consists mainly of α-magnesium dendrites and a eutectic yttrium- neodymium rich lamellar interdendritic region. A study on a nearly identical Mg-Y-Nd alloy WN42 [39] determined the eutectic phase to have a composition of

Mg 5(Nd 0.67 Y0.33 ). EDS analysis of the eutectic phase in this study was analyzed with the same technique, and composition was near Mg 5(Nd 0.5 Y0.5 ). The needle-like phase located at grain boundaries and sparsely through the matrix was analyzed via EDS. Due to the small size of the phase, the matrix analysis when subtracted from the precipitate revealed an RE stoichiometry identical to the grain boundaries. The white circular particles throughout the matrix are the α-Zr cored structure, which is a characteristic phase in many Mg-Zr alloys reported earlier [19,20,21] and are believed to be formed through a peritectic reaction upon solidification [19]. Optical microscopy showed that some of the numerous α-Zr-rich structures had α-Zr particles in the center (visible in Fig. 1). Current understanding is that zirconium contributes to very effective grain refinement by two primary mechanisms: constitutional undercooling due to a high growth restriction factor [20]; and grain nucleation due to magnesium and zirconium having the same crystal structure and similar lattice parameters [19].

α-Zr

Fig. 1. SEM images of as-cast WE43. Apart from α-Mg, three phases were identified

through EDS: interdendritic lamellar Mg 5(Nd 0.5 Y0.5 ) eutectic phase, needle precipitate of

Mg 5(Nd 0.5 Y0.5 ), and α-Zr rich cored structure.

4.1.2. Cast + T6 In the cast + T6 condition (Fig. 2) the coarse eutectic phase is significantly reduced when compared to the as-cast material. The cast + T6 microstructure had an average grain size of 80 ± 5 µm, which was determined using the mean linear intercept technique.

According to [39], peak hardness is associated with the formation of fcc-β1 plates at the expense of β” and β’ precipitates. Apps et al. [39] also noted that the observed β1 was identical to the β1 phase in [38] and was isomorphous with Mg 3RE.

Fig. 2. OM (left) and SE SEM image (right) of cast + T6 condition. The location of the α- Zr clusters is apparent and is generally centrally located in a grain or on the boundary. SEM image shows incomplete dissolution of eutectic phase and solidification defects.

4.1.3. FSP Microstructure That dynamic recrystallization (DRX) results in grain refinement during FSP is well- accepted [40-43]. Also well-established is that the thermal cycle and high strain associated with FSP in both Al and Mg alloys is sufficient to dissolve secondary particles [6,8,9,13,44,45,46]. In this study, significant dissolution also occurs, and the analysis of microstructure after FSP can give insight to dissolution and understanding of mechanical properties. Processed regions were investigated in the T5 condition (Fig. 3) to visualize precipitation. At the lowest heat input, dissolution was moderate, and the undissolved eutectic phase was readily observable on the grain boundaries. At the middle heat input,

dissolution was almost complete and intragranular β1 precipitates were observed. At the highest heat input, the eutectic phase completely dissolves, re-precipitates and coarsens at grain boundaries as a large intergranular connected phase. The present results indicate an optimum heat input from the standpoint of eutectic dissolution and cooling rate for formation of a supersaturated solid solution (SSSS) available for the precipitation of β1. Analysis of the grain size was done from the central region of the nugget, as there was variation of grain size across the nugget due to banding. Average grain size for the three increasing heat indices were 3.9 ± 0.32 µm, 5.6 ± 0.32 µm and 7.1 ± 0.48 µm. This is a significant refinement over the as-cast + T6 grain size of 80 ± 5 µm.

In agreement with investigations done on FSP 1050A1 and 7075Al-T651 [47], the peak temperature of the FSW/P thermal cycle is the primary contributor to grain size in WE43. Charit et al. [47] investigated the grain size of aluminum alloys 7075 and 2024 during FSW. Their study found no monotonic increase in grain size versus pseudo-heat index over a very broad heat input range. Yet for similar heat input, the correlation between grain size and heat input achieved in 2024Al seemed to correlate fairly well with the results from this study. Figure 4a shows the correlation between the average grain size and tool rotation rate. The near linear relationship between tool rotation rate and grain size indicates that thermal input through tool rotation rate is the primary factor in dictating final grain size. Although the linear fit shows the high importance of tool rotation rate, it is unrealistic in predicting grain size, as there will be a saturation point for high heat inputs, due to transient melting and its effect on regulating peak temperature [48]. Thermal input is also commonly observed to be a function of rotation rate and traverse rate [47]. The grain size distribution in the center of the nugget for the three heat indices was analyzed to determine the influence of the FSP thermal cycle (Fig. 4b). Grain size distribution followed a log-normal distribution typical of a recrystallized material. Distribution of grain size can provide further insights to the microstructural thermal stability during post-FSP heat treatment and mechanical properties after FSP.

a b

Fig. 3. SEM (a,b) and OM (c) image of nugget grains at respective increasing heat inputs of 400 rpm 102 mm/min, 700 rpm 153 mm/min and 900 rpm 127 mm/min in the T5 condition. Grain size increases with increasing heat input, respectively 3.9 ± 0.32 µm, 5.6 ± 0.32 µm and 7.1 ± 0.48 µm. Bright white particles in SEM Image are α-Zr, and light gray particles are eutectic phase. In optical micrograph α-Zr are black, and dark gray particles are eutectic phase.

Rhodes et al. [49] postulated through quench studies of 7050Al that the fine-grained FSP microstructure was the result of nucleation and growth. In the post-FSP microstructure the mean grain size at each heat index was skewed toward the smaller grains. As the peak temperature and thermal cycle in the processed region increase, grain growth occurs more rapidly; and smaller grains are consumed, shifting the mean of the distribution away from the median value.

a b

Fig. 4. (a) Variation of mean linear grain size in the central region of the nugget with rotation rate. (b) A plot of frequency variation of grain size in the central region of the nugget for different heat indices.

Although tool rotation rate shows a strong correlation to grain size, it is not the only factor that contributes to thermal input. Thus grain size was plotted as a function of process pitch (Fig. 5), which effectively shows the thermal input in a finite volume of material associated with the traverse displacement per revolution of the tool. The resultant plot demonstrates that as more material is processed per revolution or thermal input per unit volume decreases, the grain size decreases.

Fig. 5. Variation of mean linear grain size in the central region of the nugget with process pitch, which effectively demonstrates the heat input per unit volume of material processed.

Figure 6 shows the resultant microstructure after a T6 treatment of the FSP nugget; abnormal and normal grain growth significantly increases grain size, thus this heat treatment sequence exhibited strength similar to the cast + T6 material. Ma et al. [9] reported that a T6 treatment after FSP in a cast A356 alloy resulted in peak strength. The microstructural stability of the nugget region is a critical issue that needs further study.

Fig. 6. OM of WE43 FSP nugget in T6 condition illustrating abnormal grain growth caused by the instability of the fine-grained microstructure at solutionizing temperature of 520°C.

4.2. Material Flow Figure 7 shows macro images of the cross section for the three FSP conditions. Figure 3 shows the microstructures taken from the center of the nugget for the 3 different heat inputs in the T5 condition. The clear onion structure present in all processing conditions has been observed previously and was first discussed by Krishnan [50]. Krishnan [50] proposed that onion rings form due to the extrusion of semicylindrical plugs of material around the tool and correspond to the distance traveled per revolution of the tool. The onion rings have also been reported to contain oxide particles [51] due to exposure of processed metal to air as it proceeds around the tool to fill the cavity left by the advancing tool [50]. In the study on Al-2195 the onion ring pattern had a strong ring corresponding to pitch as well as two weaker equispaced rings that were concluded to form due to the flats of the tool [52]. By comparing the three images in Fig. 7, the nugget can be separated into two distinct flow regions: pin influenced and shoulder influenced. These dynamic regions demonstrate that upon yielding of the initial extrusion band, the volume of material is separated into the two distinct shoulder flow and pin flow extrusion regions labeled 1 and

2, respectively. Upon superposition of the tool image, the reason for the flow separation is clear. Machining limitations precluded carrying the threads to the shoulder, and led to an upper region of the pin with no threads, thus no vertical displacement is encouraged in this region. The conclusion therefore is that the distinct difference of the two regions is due to the near complete radial displacement of material in the shoulder flow region (1) versus the pin flow region (2) that undergoes radial as well as vertical displacement.

Fig. 7. Macro image of nugget geometry and onion ring structure for (top) 400 rpm 102 mm/min, (mid) 700 rpm 153 mm/min and (bottom) 900 rpm 127 mm/min. Regions 1 and 2 designate the shoulder and pin flow regions respectively. Advancing side is on the right for all images, while traverse direction is into the page.

Figure 8 shows the variation in material flow versus process pitch. Clearly, the size of the nugget region increases as the volume of material processed per revolution of the tool increases.

Fig. 8. Variation of material flow regions with pitch, demonstrating the increase in nugget size with increasing volume of material processed per revolution.

4.3. Mechanical Properties To maximize the strength of the parent material, a T6 heat treatment involving a solutionizing and aging step is performed. These results will serve as a baseline comparison for further discussion (Table 3). Although a T6 heat treatment is necessary to strengthen the base casting, the fine-grained FSP microstructure is unstable at solutionizing temperature (Fig. 6), and results in a significant coarsening of grains. Due to the instability of the nugget microstructure, a sequence of heat treatment must be developed to maximize properties across the casting. The sequence investigated in this study involved solutionizing (SS) the casting prior to FSP, then age after processing. These results have been compared to cast + T6, as FSP, and FSP + T5. The processing condition that provided best strength was 700 rpm-153 mm/min. The microstructure that corresponded to this parameter had the combination of highest volume precipitation of β1 in the aged sample and fine grain size.

Table 3. Summary of mechanical properties for the as-cast and cast + T6 heat treatments

Tensile Sample Condition Properties As-Cast Cast + T6

UTS, MPa 181 ± 9.6 261.3 ± 5.2 0.2 %YS, MPa 136.8 ± 14 202 ± 6 % Elongation 1.9 ± 0.6 5

This data indicates that in WE43, dissolution, precipitation and subsequent coarsening of the eutectic and β1 phase more significantly impact strength than grain size variation due to grain growth. Figure 9 gives the tensile properties of all heat treatment steps and the variation with the different factors that influence heat input (rotation rate and traverse speed) as well as with a measure of heat input per unit volume (process pitch). From Fig. 9, note that the UTS shows a strong correlation to traverse speed and weak correlation to rotation rate, which results in a weak correlation to process pitch. Also noted is that solutionizing prior to processing imparts a slight increase in UTS.

Fig. 9. Variation of tensile strength with various measures of processing thermal input. Tensile strength shows strong dependence on traverse speed.

Figure 10 shows a large variation in yield strength due to heat treatment. The very strong correlation between yield strength and traverse rate, but weak correlation to rotation rate, is interesting. The lower heat input, or higher processing pitch, exhibits the lowest strength due to the incomplete homogenization of second phase particles. Solutionizing the casting prior to processing at this heat index has increased the YS by 15 MPa. The middle heat input shows the same trend, yet achieves a higher strength due to more complete homogenization because of the increased thermal cycle during processing. For the highest heat index, a reasonable argument is that the processing temperature is so high that coarsening of precipitates at grain boundaries depletes intragranular precipitates

(Fig. 3). Figure 11 shows a substantial increase in percent elongation from the lowest heat input to the middle heat input, but seems to reach a plateau of 25%. For both tensile and yield strength the properties across heat index rise and fall, yet the elongation decreases with decreasing heat input due to the decrease in uniformity and distribution of equiaxed grains.

Fig. 10. Variation of yield strength with various measures of processing thermal input. Yield strength shows strong dependence on traverse speed and weaker dependence on rotation rate.

Fig. 11. Variation of elongation with various measures of processing thermal input. Elongation shows strong dependence on both rotation rate and traverse speed, and a resultant good correlation with processing pitch.

5. Conclusions • Friction stir processing of a cast WE43 magnesium alloy resulted in significant grain refinement and dissolution of the β-Mg-Y-Nd phase. • The impact of the dissolution of β phase had more influence on strength than the variation in grain size due to FSP thermal cycle. • A solutionizing step prior to FSP and subsequent aging resulted in the best mechanical properties due to increased homogenization. • Maximum ultimate tensile strength of 303 MPa, 0.2% yield strength of 253 MPa and an elongation of 17% were achieved with processing parameters of 700 rpm and 153 mm/min. These values resulted in a 125% processing efficiency when compared to the cast + T6 condition.

Acknowledgments The authors thank the donations of material from Fansteel/Wellman Dymanics.

References [1] Aragones, J. Goundan, K. et al. “Development of the 2006 Corvette Z06 Structural Cast Magnesium Crossmember”, SAE International 2005-01-0340. [2] S. Schumann and H. Friedrich, “Current and Future Use of Magnesium in the Automobile Industry”, Materials Science Forum , 51(2003), 419-422. [3] N. Li et al., “2006 Ford GT Magnesium Instrument Panel Cross Car Beam”, SAE International 2005-01-0341. [4] M. Santella et al. “The Use of Friction-Stir Technology to Modify the Surfaces of AM60B Magnesium Die Castings”, JOM May 2006. V. 58 No.5. [5] B.J. Coultes, J.T. Wood, G.Wang, R. Berkmortel, “Mechanical Properties and Microstructure of Magnesium High Pressure Die Castings”, TMS Magnesium Technology 2003, 45-50. [6] Z.Y. Ma, S.R. Sharma, R.S. Mishra. “Effect of multiple pass friction stir processing on microstructure and tensile properties of a cast aluminum-silicon alloy”, Scripta Materialia v54(2006) 1623-1626.

[7] M.L. Santella. T. Engstrom, D. Storjohann, T.Y. Pan, Scripta Materialia 2005; 53:201. [8] Z.Y. Ma, S.R. Sharma, R.S. Mishra, M.W. Mahoney, Material Science Forum 2003; 426-432:2891. [9] Z.Y. Ma, S.R. Sharma, R.S. Mishra, “Effect of friction stir processing on the microstructure of cast A356 aluminum”, Mat. Sci. Eng. A: 433 (2006);269-278. [10] R. Johnson, Mater. Sci. Forum 419–422 (2003) 365. [11] W.B. Lee, Y.M. Yeon, S.B. Jung, Mater. Sci. Technol. 19 (2003) 785. [12] G. Kohn, S. Antonsson, A. Munitz, in: S.K. Das (Ed.), Automotive Alloys 1999, TMS, 2000, pp. 285–292. [13] A.H. Feng, Z.Y. Ma, “Enhanced mechanical properties of Mg-Al-Zn cast alloy via friction stir processing”, Scripta Mat:56/5. March 2007. pp 397-400. [14] P. Lyon, “New Magnesium Alloy for Aerospace and Specialty Applications”, TMS Magnesium Technology 2007: 311-315. [15] M. Tsujikawa, S.W. Chung, M. Tanaka, Y. Takigawa, S. Oki, K Higashi, “High Strengthening of Mg-Y-Zn Cast Alloy by FSP”, Materials Transactions V46/12, 2005, pp. 3081-3084. [16] P.Lyon – Elektron WE43, Foundry Processing, Phoenix 1998. [17] Ma Qian, D. Graham, L. Zheng, D.H. St. John and M.T. Frost. Material Science and Technology, 2003, 19, 156-162. [18] P. Cao, M. Qian, and D.H. St. John: Scripta Mater., 2004, vol. 51, pp. 125-29. [19] E.F. Emley, “Principles of Magnesium Technology” (Pergamon Press, Oxford ; New York, ed. [1st ]. 1966). [20] Ma Qian, D. Graham, L. Zheng, D.H. StJohn and M.T. Frost. Material Science and Technology, 2003, 19, 156-162. [21] M. Qian, D.H St. John and M.T. Frost: Scripta Mater.,2002, vol.46,pp. 649-54. [22] L.L. Rohklin, in: J.N. Fridlyander, D.G. Eskin (Eds.) “Magnesium Alloys Containing RE Metals: Structure and Properties”, Taylor and Francis, 2003. [23] ASM Specialty Handbook, Magnesium and Magnesium Alloys, Ed. Michael M. Avedesian and Hugh Baker, ASM International 1999. [24] S. Delfino, A. Saccone and R. Ferro, Metal. Trans. 21A (8), 2109-2114 (1990).

[25] A. Saccone, S. Delfino, D. Maccio and R. Ferro, J. Phase Equilibria, 14(4), 479-484 (1993). [26] L.L. Rokhlin, Izv. Akad. Nauk SSSR, Otd. Tekh. Nauk, Met. Toplivo, (2), 126-130 (1962) (in Russian). [27] L.L. Rokhlin. Magnesium Alloys Containing Rare Earth Metals, Nauka, Moscow (1980) (in Russian). [28] V.I. Evdokimenko and P.I. Kripyakevich, Kristallografiya, 8(2), 186-193 (1963) (Russian). [29] N.F. Lashko and G.I. Morozova, Zavodskaya Laboratoriya, 330(10), 1187-1189 (1964) (in Russian). [30] N.F. Lashko and G.I. Morozova, Kristallografiya, 9(2), 269-270 (1964) (in Russian). [31] Negishsi, Nishimura et al. Journal of Japan Institute of Light Metals. Vol 45 pt 2. [32] J.F. Nie, B.C. Muddle, “Characterization of Strengthening Precipitate Phases in Mg- Y-Nd Alloy”, Acta Mat. 48, 2000, 1691-1703. [33] Z.A. Sviderskaya, E.M. Padeezhnova, Izv. Akad. Nauk SSSR, Metally, (6) (1971) 200-204 (in Russian). [34] L.L. Rokhlin, T.V. Dobatkina, I.E. Tarytina, V.N. Timofeev, E.E. Balakhchi, “Peculiarities of the phase relations in Mg-rich alloys of the Mg-Nd-Y system”, Journal of Alloys and Compounds 367 (2004) 17-19. [35] M.E. Drits, L.L. Rohklin, I.E. Tarytina, Izv. Akad. Nauk SSSR, Metally, (3) (1983) 111-116 (in Russian). [36] G.W. Lorimer, in: Proceedings Magnesium Technology, Institute of Metals, London, 1986, p.47. [37] B. Smola, B.L. Mordike, in: 10 th European Crystallographic Meeting, Collected Abstracts, IUCr, Wroclaw, Poland, 1986, p.411. [38] J.F. Nie, B.C. Muddle, “Characterization of Strengthening Precipitate Phases in Mg- Y-Nd Alloy”, Acta Mat. 48, 2000, 1691-1703. [39] P.J. Apps, H. Karimzadeh, J.F. King, G.W. Lorimer, “Precipitation reactions in Magnesium-rare earth alloys containing Yttrium, Gadolinium or Dysprosium”, Scripta Mat. 48 (2003) 1023-1028. [40] K.V. Jata, S.L. Semiatin, Scripta Mater. 43, 2000, 743.

[41] J.-Q. Su, T.W. Nelson, R. Mishra, M. Mahoney, Acta Mater. 51, 2003, 713. [42] K.V. Jata, K.K. Sankaran, J.J. Rushau, Metall. Mater. Trans. A 31, 2000, 2181. [43] Y.S. Sato, M. Urata, H. Kokawa, Metall. Mater. Trans. A 33, 2002, 625. [44] K. Nakata, S. Inoki, Y. Nagano, T. Hashimoto, S. Johgan, M. Ushio, in: Proceedings of the Third International Symposium on Friction Stir Welding, Kobe, Japan, September 27–28, 2001. [45] W.B. Lee, J.W. Kim, Y.M. Yeon, S.B. Jung, Mater. Trans. 44 (2003) 917. [46] W.B. Lee, Y.M. Yeon, S.B. Jung, Mater. Sci. Technol. 19 (2003) 785. [47] R.S. Mishra, Z.Y. Ma, (eds) A.G. Cullis, S.S. Lau, “Friction Stir Welding and Processing: A Review Journal”, Materials Sci. 2005: v50: 1-78. [48] A. Gerlich, G. Avramovic-Cingara, T.H. North, “Stir Zone Microstructure and Strain Rate during Al 7075-T6 Friction Stir Spot Welding”, Met. Mat. Trans A 37A, Sept (2006) pp 2773-2786. [49] C.G. Rhodes, M.W. Mahoney, et al. “Fine-grain Evolution in Friction-Stir Processed 7050 aluminum”, Scripta Materialia 48 (2003) 1451-1455. [50] K.N. Krishnan, “On the Formation of Onion Rings”, Materials Science and Engineering, A327; 2002:246-251. [51] P.B. Pragnell, C.P. Heason, “Grain structure formation during friction stir welding observed by the ‘stop action technique’ ”, Acta. Mat. 53, 2005, 3179-3192. [52] H. Larson, et al., “Joining of dissimilar Al-alloys by Friction Stir Welding”, Second Int. Conf. on Friction Stir Welds, Gothenburg, Sweden, 2000.

PAPER II

Effect of friction stir processing on microstructure and mechanical properties of a cast magnesium-rare earth alloy

T.A. Freeney and R.S. Mishra

Center for Friction Stir Processing and Department of Materials Science and Engineering, University of Missouri-Rolla, 218 McNutt Hall, 1870 Miner Circle, Rolla, MO 65409, USA

Abstract Single pass friction stir processing (FSP) was used to increase the mechanical properties of a cast Mg-Zn-Zr-RE alloy, Elektron 21. A fine grain size was achieved through intense plastic deformation and the control of heat input during processing. The effects of processing and heat treatment on the mechanical and microstructural properties were evaluated. An aging treatment of 16 hours at 200°C resulted in a 0.2% proof stress of 275 MPa in the FSP material, a 61% increase over the cast + T6 condition.

Keywords: Friction Stir Processing, Elektron 21, EV31A, rare earth, cast magnesium, heat treatment

1. Introduction Interest in the use of magnesium in weight critical applications such as aircraft and automotives to reduce fuel consumption and emission levels has intensified recently [1,2,3]. The majority of current magnesium applications lies in die casting for the automotive industry, and during the 1990’s the percentage of total vehicle weight comprised of magnesium in automobiles more than doubled [4]. Implementation of cast magnesium components through the vehicle potentially offers 100 kg of weight reduction per automobile [1,5,6,7]. In the aircraft industry, where the cost of weight savings is 100

25 to 1000 times more than in the automotive industry, low strength and poor ductility are primary factors that inhibit the more widespread application of magnesium components [2]. FSP (Fig. 1.) has been demonstrated as a viable process to increase the strength of castings by reduction of intermetallic particle size, grain size, and elimination of porosity through forging pressure [8]. FSP is a thermo mechanical microstructural modification tool [9,10] developed from the concept of friction stir welding patented by TWI [11]. FSP uses frictional heat and pressure generated from mechanical work, transmitted through a tool consisting of a pin and shoulder, to impart localized grain refinement through intense plastic deformation [8]. Notably, microstructural modification via FSP has been conducted on cast aluminum alloys A356 and A319 [12,13]. In the studies by Ma et al. [12,14,15] on A356, a significant increase in mechanical properties, strength, ductility and fatigue life of the FSP material was reported. FSP has been demonstrated on magnesium-aluminum-manganese (AM) and magnesium-aluminum-zinc (AZ) casting alloys AM60B, AM50, AZ91/6, AZ91D4, AZ95D/5 [16,17,18,19]. The AM and AZ alloys, such as in the previous studies, are relatively inexpensive due to the relatively low cost of the alloying additions and ease of melt handling [20]. For aerospace and emerging automotive applications, high strength and creep resistance are important factors.

Fig. 1. Schematic of friction stir processing [8]

Mg alloys that contain rare earth elements exhibit excellent creep resistance and high temperature strength. Recent work by Tsujikawa et al. [21] on an Mg-Y-Zn cast alloy

26 found that multiple pass friction stir processing had no significant effect on the grain size of the prior pass. A study of high-strength, creep-resistant alloy WE43 [22] showed that for simultaneous peak strengthening of the nugget and casting, variation of processing thermal input was critical to the control of dissolution and precipitation morphology but only weakly affected grain size. Tsujikawa et al. [21] also found that a solutionizing step prior to FSP followed by an aging treatment yielded highest nugget strength. The ability to tailor local material properties within a bulk component offers designers more freedom and allows for synergistic casting design. Elektron 21 (ASTM EV31A) is a recently developed sand cast magnesium alloy designed and patented by Magnesium Elektron for specialty applications with temperature resistance of up to 200°C and good corrosion resistance [20]. The physical properties of EV31A are listed in Table 1. EV31A provides better castability and offers a low-cost alternative to high strength, creep resistant magnesium — yttrium alloy Elektron WE43 [20]. The benefits of magnesium casting are highlighted by exceptional fluidity, fast cycle time due to low latent heat, and good die life. State-of-the-art magnesium alloys offer high specific strength at room and elevated temperatures, good creep and corrosion properties, excellent machineability and dampening characteristics [16]. Despite all the benefits of the casting process, mechanical properties can vary significantly due to solidification issues such as coarse interdendritic phases, large grain size, and interdendritic microporosity [23,12].

Table 1. Physical properties of EV31A [24] Specific Coefficient of Thermal Electrical Specific Heat Gravity Thermal Conductivity Resistivity Jkg -1 K -1 (20°C) Expansion Wm -1 K -1 nΩm (20-200°C) 10 -6 K -1 (20°C) (20°C) (20-200°C)

1.82 26.7 86 94.6 997

2. Effect of Alloying Elements The microstructural constituents of the Mg-Zn-Zr-Nd-Gd alloy EV31A are not well- documented, especially with respect to ternary phases. The majority of work thus far has

27 been done with respect to interactions observed in the respective binary phase diagrams. The probable precipitation sequences of the elements present in EV31A have been summarized in Table 2.

Table 2. Summary of probable precipitation sequence

Mg-Zn [25] SSSS GP Zones MgZn 2 MgZn 2 Mg 2Zn 3 discs hcp rods hcp discs trigonal Mg-Nd [25] SSSS GP Zones β" β' β

plates D019 Mg 3Nd Mg 12 Nd hcp plates fcc plates bct Mg-Gd [26] SSSS ? β" β' β

D019 bc ortho Mg 5Gd hcp plates fcc plates

Mg-Nd-RE SSSS β" β' β1 [28] β (Gd,Dy,Y) D019 [27] bc ortho fcc fcc [25] platelets globular plates

2.1 Mg-Zr An excellent summary of the current state of knowledge of grain refinement in cast Mg alloys is given by St. John et al [29]. Of several methods for grain refinement of Mg alloys during casting, main ones include superheating, carbon inoculation, incorporation of Fe or Al to form Fe and Al-rich intermetallics [30]; and Zr addition [29], such as in EV31A. The α-Zr cored structure, which is a characteristic phase in Mg-Zr type alloys, is believed to be important to grain refinement during solidification [31,32,33]. The structure is believed to be formed through a peritectic reaction upon solidification [31,33]. Current understanding is that zirconium contributes to very effective grain refinement by two primary mechanisms: constitutional undercooling due to a high growth restriction factor [32]; and grain nucleation due to magnesium and zirconium having the same crystal structure and similar lattice parameters [31]. The end result of this reaction is spherical clusters of α-Zr particles in Mg matrix. For later discussion, note that Zr does not change the decomposition behavior of the SSSS [34] and thus will be ignored.

2.2 Mg-Zn The addition of Zn to Mg alloys has two main benefits. The first is the substantial increase in the critical resolved shear stress required for slip of the basal plane at room temperature when compared to Al, Ti, Cd, and In additions [35,36]; and when solubility is exceeded, precipitation strengthening of the MgZn type occurs near peak hardness values upon aging [35]. Later work [37] on an Mg-Zn-Zr alloy MB15, 5.0-6.0% Zn and

0.3-0.6% Zr identified the presence of MgZn 2 and Mg 2Zn 3. The presence of MgZn has also been reported in Mg-5% Zn- 5% Sn alloy [38] by EDS and XRD analysis. In this study the formation of MgZn precipitates was attributed to the increase in strength upon age hardening. An interesting point to note is that the accepted precipitation sequence of Mg-Zn alloys does not contain MgZn (Table 2). More recently a TEM and XRD analysis of the precipitates in an Mg-Y-Zn as-cast alloy was performed by Rao et al [39]. The

presence of nano sized MgZn and MgZn 2 precipitates suggested these precipitates improved the mechanical behavior of this alloy. The existence of a long periodic stacking

phase (LPS) 18R-Mg 12 ZnY due to cooling of a SSSS was also observed in this study; the existence of these ordered phases has previously been reported [39-44].

2.3 Zr-Zn Zn as a grain refiner in this alloy system has been the focus of recent work by Hildebrand et al [45]. In confirmation of results by Löhberg & Schmidt [46] and Bhan & Lal [47], Hildebrand et al. indicate that solubility of Zr increases with the addition of Zn up to a peak value. Further investigation of the effect of Zn on Mg-Zr binary by Arroyave et al. [48] suggests that the decrease in solubility of Zr in Mg with excess Zn is due to the formation of zinc zirconides. The existence of zinc zirconides most commonly in the form of ZrZn, Zr 2Zn, ZrZn 2 and Zr 3Zn 2 [48] has been observed by [31,37,45,49,50]. Based on phase equilibrium work [37,49,51] for the respective Mg-Zn-Zr binaries and lack of evidence of a ternary compound, Arroyave et al. [48] concluded that existence of a ternary was unlikely in the calculation of the Zn-Zr phase diagram. Although no ternary phases are believed to form between Mg-Zn-Zr, recent studies have shown formation of stable quasicrystals in this system for both annealed and cast samples [52].

2.4 Zn-RE Besides the beneficial effects of grain size reduction in the as-cast condition, the addition of Zn enhances precipitation strengthening [53] in RE containing alloys. A significant benefit of Zn addition to Nd-containing alloys is the decrease of Nd solubility in Mg [34], which can result in more effective use of Nd for precipitation strengthening. Zn addition has been reported to increase yield strength in both Mg-Nd and Mg-Y alloys, decrease plasticity in hot extruded Mg-Y alloys, increase plasticity in extruded Mg-Nd alloy in the T6 condition, and lower the necessary solutionizing temperature in Mg-Nd alloys [34,54-56]. The addition of Zn to Mg-RE system has been observed to change the eutectic morphology from dendritic to equiaxed [20,57,58]. Also note that the binary phase diagram of the Gd-Zn system shows several binary phases [59].

2.5 Mg-Nd-Gd The combined additions of Gd and Nd in Mg alloys have been reported to increase castability and precipitation strengthening [53]. According to the binary phase diagrams of Mg-Nd and Mg-Gd, there are several stable binary phases. In the Mg-rich portion of these diagrams the phases Mg 5Gd [60] and Mg 41 Nd 5 [61,62] form through eutectic reactions. The binary system of Nd-Gd demonstrates no stable compounds [63], and these elements have nearly zero solubility in each other at room temperature, yet a spinoidal δ phase forms around Gd 1Nd 1. The investigation of the Mg-3Nd-Gd ternary [27] revealed that the solid solubility of Gd in Mg is significantly reduced when Nd is added. This reduction in solubility is important in the aging response of this alloy system.

3. Experimental In this study, the as-received material was a sand cast EV31A plate in the T6 condition provided by Magnesium Elektron, with a nominal composition of 0.5%Zn- 3.1%Nd-1.7%Gd-0.3%Zr-bal. Mg (in wt. %). The coupon had dimensions of 64 mm x 9 mm x 150 mm. Single pass FSP runs were performed on the coupon with the following processing conditions: rotational speed of 400 rpm and a traverse speed of 102 mm/min with a tool tilt angle of 2.5° in air (Fig. 2). The MP159 alloy tool had a 25.4 mm shoulder

30 with a cylindrical threaded pin of 7.2 mm length and 6.4 mm diameter. For mechanical testing, mini-tensile samples with 4.2 mm gage length and 1 mm gage width were machined from the nugget. Samples were CNC machined from the transverse section so the gage length was entirely in the nugget. Four tensile samples for each heat treatment were ground to a final thickness of 0.5-0.55 mm and given a 1 µm polish before testing. Tensile testing was performed at a strain rate of 1x10 -3 s-1. Metallographic samples were mechanical polished to 0.05µm and etched using an acetic glycol etchant [25].

Fig. 2. Image of FSP EV31A surface texture. The bright smooth texture is indicative of lower process temperature.

To maximize the strength of EV31A a T6 treatment was performed; solutionizing (520°C for 8h), warm water quench (60-80°C) and aging (200°C for 16h). The effect of heat treatment was investigated according to the experimental schedule in Table 3.

Table 3. Experimental heat treatment schedule with aging time of 16 hrs

Heat Treatment Sample ID Solutionizing, °C Aging, °C Cast + T6 520 – 8 hr 200 FSP X X FSP + T6 520 – 8 hr 200 FSP + SS + Age 490 – 1 hr 200 FSP + Age X 200

Optical microscopy (OM), scanning electron microscopy (SEM JEOL 840A), energy dispersive spectroscopy (EDS) and differential scanning calorimetry (DSC) techniques were used for microstructural study; quantitative analysis was done with ImageJ software package. Grain size was determined through the mean linear intercept method, where spatial grain size = mean linear intercept length x 1.78.

4. Results and Discussion 4.1 Microstructure 4.1.1 Microstructure – As-Cast The dendritic arm spacing of the cast EV31A (Fig. 3) was calculated to be 52.8 ± 25 µm. An analysis of the as-received microstructure showed 3 main regions of interest (Fig. 4): lamellar eutectic, innerdendritic plate-like precipitates and Zr clusters. In the as- received microstructure the Zr rich clusters have no resolvable structure and are generally located toward the dendrite interior, where the volume of precipitates is lower than near the interdendritic regions. The spherical zirconium-rich peritectic regions ranged in radius from 6-20 µm. A microstructural study of Zn-free Mg-RE alloy GN72, which has same RE elements as EV31A, has been reported [64]. In particular, an analysis of the as-

cast eutectic, β and β1 phases was presented. Their study did not report on lamellar eutectic. All analyzed phases had an FCC structure, and the compositions for eutectic and

β phases were Mg 5(Nd 0.5 Gd 0.5 ), Mg 5(Nd 0.33 Gd 0.67 ). In another microstructural study on Mg-Zn-Zr-RE alloys ZE41A and EZ33A [65], a eutectic Mg-Zn-RE interdendritic phase was observed. In the higher Zn-containing alloy ZE41A (~4 wt% Zn), coring of Zn around the eutectic intergranular phase was strong. The Zn coring towards interdendritic regions lessened in alloy EZ33A (~2.4 wt% Zn), but the majority of Zn was still associated with the interdendritic eutectic phase.

Fig. 3. BSE of EV31A in F temper. Primary features are intergranular lamellar structure, Zr-rich clusters, and dense precipitates located heavily near grain boundaries. Black circles are artifacts from corrosion products due to reaction of Zr-rich clusters with colloidal polishing suspension.

A summary of the EDS analysis performed on the regions of interest (Fig. 4) in EV31A in the as-cast state is shown in Table 4. Although Zn was detected throughout the matrix, a high concentration of Zn was located in the eutectic phases. The interdendritic region consists of an Mg-Zn-RE lamellar structure, which is an indication of a quaternary eutectic transformation between α-Mg and Mg-Zn-Gd-Nd, although the existence of this would need further evaluation to confirm. The bright white interdendritic phase is significantly higher in concentration of Zn and Gd, than the gray phase. The gray interdendritic regions likely are present due to continued diffusion from the Mg-Zn-RE eutectic after solidification. Undoubtedly a large error is present in EDS measurements due to the large interaction volume and high Z values of the RE components, but for simplicity a relatively simple method for determining phase stoichiometry in thin foils through extrapolation of composition was presented by Lorimer et al. [66] and used in this study as applied to scanning electron microscopy.

To determine the stoichiometry of the as-cast eutectic phase, composition of the RE alloys was plotted versus Mg content (Fig. 5). The eutectic phase ranged in composition

from 50-70% Mg; this leads to an Mg 3X2 stoichiometry where X is the combined composition of Nd, Gd and Zn. Considering the fractions of the individual constituents, a stoichiometry of Mg 3(Nd 0.525 Zn 0.325 Gd 0.125 )2 was calculated.

Fig. 4. BSE of as-cast microstructure. Regions labeled 1-5 have been evaluated through EDS (Table 4).

Table 4. EDS analysis of as-cast microstructure

Phase Composition (at %)

Mg Zr Zn Nd Gd

matrix 96.5 0.7 0.4 1.5 0.9

1 79.8 ± 8.3 0.9 ± 0.9 3.5 ± 0.6 14.1 ± 6.7 1.6 ± 0.1

2 59.8 ± 5.8 0.6 ± 0.6 13.2 ± 3.8 21.2 ± 2.2 5.3 ± 1.6

3 80.8 ± 6.3 0.2 ± 0.1 3.5 ± 1 13.2 ± 4.6 2.4 ± 0.6

4 94.7 3.8 0.6 0.7 0.3

5 95 ± 1.2 0.6 ± 0.3 0.5 ± 0.6 3 ± 1.3 1.0 ± 0.2

Fig. 5. Variation of Nd, Zn and Gd content as a function of Mg concentration in matrix and intergranular phase of EV31A in as-cast condition.

An analysis of the suspected Zr-rich region (Table 4 #4) showed that it indeed was the well-documented α-Zr cluster. The EDS analysis also showed that the cluster had an increased concentration of Zn when compared to the matrix, which may indicate the local formation of zinc zirconides. Also interesting is that the cluster shows a decreased concentration of RE elements when compared to the matrix; this indicates that solute is rejected from this region. The plate-like phase (Table 4 #5) was fairly small, which made accurate analysis of its composition difficult. The plate structure is Nd-rich, but distinguishing composition from the matrix is unreliable and has not been conducted in this work.

4.1.2. Microstructure – As-Cast + T6 Peak mechanical properties at both elevated and room temperature as well as exceptional creep properties have been attributed to precipitation of β” and β’ on prismatic planes as they act as barriers to gliding dislocations [67-71]. Mg-Y-Nd alloy

WE43 [28] observations show that peak hardness corresponds to a majority of β’, β1 and

35 equilibrium β precipitates. The grain size of EV31A in the T6 condition (Fig. 6) was 78.5 ± 22 µm. The average size of the eutectic phase on the grain boundaries after the T6 treatment decreased and transformed to a consolidated Nd-rich phase. The Mg 3

(Nd 0.525 Zn 0.325 Gd 0.125 )2 phase present in the as-cast condition completely transformed during the T6 treatment.

Fig. 6. BSE image of EV31A in T6 condition, showing the intergranular Mg-Nd-Gd-Zn phase and Zr-Zn clusters.

An EDS analysis was done on the T6 microstructure to investigate phase composition (Table 5). The most notable feature of this microstructure is the transformation of Zr clusters into Zr-Zn clusters. A majority of the Gd and Zn present prior to T6 treatment in the as-cast eutectic phase diffused into the matrix. Since no clear location of Gd was found due to the high solubility of Gd in Mg, likely a fine dispersion of Mg-Gd type precipitates form upon aging of the SSSS; yet these structures are not identified in this study. Also likely is that Mg-Zn type precipitates will form upon decomposition of the SSSS, yet they were undetectable through SEM analysis. As the high concentrations of Zn present in the as-cast eutectic diffuse into the matrix, the formation of zinc zirconides of the type Zr 2Zn and less frequently Zr 3Zn 2 seems to be preferred (Fig. 7), as almost no Zn was left in the matrix or grain boundaries. The Zn-Zr

36 type precipitates also were oriented on prismatic planes and had the morphology of either needles or plates, while the central particles had a faceted structure. A final observation was that the Zn-Zr clusters, similar to the α-Zr clusters in the as-cast state, were highly susceptible to corrosion during polishing.

Table 5. Results of EDS analysis of T6 microstructure

Phase Composition (at %)

Mg Zr Zn Nd Gd

matrix 99.1 ± 1.3 0.4 ± 0.5 0 0.4 ± 0.5 0.2 ± 0.3

eutectic 89.8 ± 8.4 0 1.1 ± 0.4 8.3 ± 7.9 0.4 ± 0.3

needles 92.6 ± 6.8 4.9 ± 4.8 2.2 ± 1.8 0.2 ± 0.1 0.1 ± 0.1

cluster core 67.8 ± 26.7 21.9 ± 18.4 10.1 ± 8.3 0.2 ± 0.1 0

Fig. 7. BSE of Zr 2Zn clusters in EV31A formed after T6 treatment at prior α-Zr rich clusters. Not all clusters were identical as shown; image on left has no apparent central core, while image on right has a clear core structure of same composition as needles.

Note that clusters analyzed contained Zr 3Zn 2.

Similar to the calculation of composition for the as-cast eutectic, an extrapolation technique was used to determine the equilibrium eutectic (Fig. 8). The composition calculated at 75% Mg was Mg 3(Nd 0.88 Zn 0.08 Gd 0.04 ).The clusters of data around 90-94%

Mg represent a stoichiometry near Mg 12 Nd, which is the equilibrium β phase of the Mg- Nd precipitation sequence (Table 2). Thus the possibility that the presence of the

elements Zn and Gd in the EDS analysis originate from solubility in the Mg 12 Nd phase, or are spurious peaks from interaction volume, is not ruled out in this study.

Fig. 8. Variation of Zn, Nd and Gd content as a function of Mg concentration in matrix and intergranular phase of EV31A in T6 condition.

4.1.3. Microstructure – FSP This study demonstrates the benefits of using FSP as a modification tool to enhance the mechanical properties of an EV31A sand cast plate. The microstructure of the FSP region is governed by the inherent temperature and strain generated during FSP by frictional contact of the material at the interface with the rotating tool. The frictional condition, thus resultant microstructure, can be controlled at the tool interface by selecting appropriate rotation rate and traverse speed. Strains in excess of 40 and strain rates of 10 s-1 reportedly [72-74] are achievable, while temperature in Mg alloys can

38 range from 370°C to 500°C [75,76]. An extensive review of the microstructural evolution during FSP and effects of processing parameters are available [8]. Dynamic recrystallization (DRX) to govern grain refinement in FSP [77-80] is largely accepted, although the exact mechanisms that govern the resultant microstructure are still being debated. Largely the controversy is between the proposed multiple mechanism model (discontinuous DRX, grain growth, dislocation introduction, dynamic recovery and continuous DRX) [81] and geometric DRX followed by static annealing [82,83]. Pragnell et al. [82] found no evidence to support continuous DRX as proposed by Su et al [81], and remarked that the high temperature microstructural evolution during FSP can be directly compared to high strain hot torsion processing, despite the much lower strain rate present during hot torsion processing. Su et al. [81] cite work by Blum [84] in remarking that recrystallized grain size of Al alloys by means of geometric DRX is limited to 1-3 times the subgrain diameter [81]. To date no definitive recrystallization mechanism governing the microstructural evolution has been agreed upon. The FSP EV31A nugget was sectioned in the transverse direction (Fig. 9) for macro analysis of material flow. The nugget region shows the signature onion rings which have been reported to appear with a periodicity of pitch (distance advanced in welding direction/revolution or traverse speed/rotation rate) [85]. The onion rings have also been reported to contain oxide particles [86] due to exposure of processed metal to air as it proceeds around the tool to fill the cavity left by the advancing tool [85]. In the study [82] on Al-2195 the onion ring pattern had a strong ring corresponding to pitch as well as two weaker equispaced rings that were concluded to form due to the flats of the tool. In the present study the onion rings in EV31A were observed not to be continuous semicircles, but seemed to alternately align in several locations. The onion rings also showed a gradual change in grain size across each ring. This is believed to be preferential coarsening possibly due either to texture or to the presence of oxide particles; more investigation would be necessary to define the exact reason. SEM analysis revealed that the central region of the FSP nugget (Fig. 9) possessed the maximum grain size = 3.4 ± 0.26 µm, which is a significant reduction from the 78.5 ± 22 µm in the T6 condition. Recent studies of the effect of FSP on WE43, an Mg-Y-Nd casting alloy [22] show similar grain size of 3.9 ± 0.32 µm at same processing conditions,

39 with different tool geometry. This grain size also matches well with the reported ~3-4 µm grain size achieved with same tool geometry in A356 [15]. SEM images of nugget (Fig. 10) show the distribution of particles present after FSP.

Fig. 9. Macro image of nugget cross section oriented that welding direction into page. Maximum grain size is in central region of nugget = 3.4 ± 0.26 µm.

Fig. 10. BSE of as friction stir processed region in EV31A unetched (left) and etched

(right). Presence of both Zr 2Zn 3 (bright white) particles and β-Mg 12 Nd (grey) particles were detected via EDS; black regions correspond to dissolved phase from etching or particle dropout. Particle size in nugget region was determined by calculating average area and was 0.05 ± 0.04 µm 2, while for similar processing condition particle size in A356 was 2.84 ± 2.37 µm. This reduction in particle size has been reported to be a main contributor to elongation and fatigue property improvement in Al alloy A356 [87]. In the study of

A356, particle distribution was homogeneous; whereas in the present study (Fig. 10), particle banding is evident. To explain the significant difference in particle size post FSP, since the grain sizes are similar, the diffusivity of the second phase is considered. In the aluminum system silicon particles have very low solubility at processing temperatures, while the RE elements in the magnesium system have high solubility at processing temperatures. Deformation mechanism map construction for pure Mg and pure Al shows that at temperatures present during FSP, diffusion of Mg atoms is 5–6 times that of Al atoms [88]. This may be why upon similar processing conditions and expected similar strain rates, Mg particles are so much more refined.

4.1.4. Microstructure – FSP + HT Although the FSP microstructure exhibits enhanced ultimate strength, yield strength and elongation, optimization of the heat input and subsequent heat treatment are necessary. In application of FSP for casting modification, a local region of the casting will be processed for enhancement or repair. Thus the strength of the casting and FSP region must be maximized simultaneously. Through the correlation of input processing parameters to resultant microstructural properties, heat treatment can be tailored to design specifications. To maximize the strength of the FSP region, a T6 heat treatment was performed (Table 3) on the FSP material. The resultant microstructure (Fig. 11) experienced abnormal grain and particle coarsening. Abnormal grain growth is defined when grains in local regions of the microstructure grow at a faster rate than surrounding regions creating a bimodal grain distribution. Even after a solutionizing step of 520°C for 8 hrs, clusters of 5-10 µm grains are present on the prior noted FSP fine grain bands indicated by white arrows in Fig. 11. Unclear at this time is the reason for increased stability of these regions.

Fig. 11. BSE of friction stir processed region in EV31A after T6 heat treatment. Arrows represent clear regions of grain growth discontinuity at locations of prior banded microstructure.

The next heat treatment that was performed was a short solutionizing step and age (Table 3). This microstructure was not analyzed, but inferences will be drawn based on the mechanical data presented in the next section. The final heat treatment performed was a direct aging step on the FSP material (Fig. 12). This microstructure, as can be seen in the BSE image, contained a very dense concentration of precipitates. Due to the high volume fraction, proximity, and relative size of these precipitates, EDS analysis of the phases was not done. Yet the majority of the visible precipitates can be assumed to be the equilibrium β-Mg 12 Nd, based on the breakdown and dissolution of β-Mg 12 Nd during FSP.

Fig. 12. BSE of friction stir processed region in EV31A aged for 16 hrs at 200°C. Note extremely high precipitate density, which inhibited ability to distinguish between the separate phases.

4.2. Mechanical Properties The mechanical properties after processing are significantly improved due to a combination of reduction in grain size, breakage of secondary particles and dissolution of strengthening phase. Representative stress-elongation curves for the mini tensile specimens are shown in Fig. 13. EV31A in the T6 condition shows behavior similar to the FSP + T6 and FSP + SS + Age specimens. The FSP and FSP + Age specimens show yield point phenomena due to solute drag.

Fig. 13. Representative tensile test curves for EV31A in various processing conditions at a strain rate of 10 -3 s -1. Samples tested in FSP and FSP + Age condition show yield point phenomena.

Figure 14 shows a summary of the average properties obtained from mini tensile specimens. A moderate increase in yield strength was obtained in the FSP + Age condition.

Fig. 14. Average room temperature tensile properties according to experimental schedule at a strain rate of 10 -3 s -1. Maximum yield strength is achieved in the FSP + Age condition.

4.3. Fractography To gain a better understanding of the change in failure mechanisms due to FSP, SEM analysis of the fracture surface was performed. Figure 15 shows the fracture surface of EV31A in the T6 condition and the FSP condition. Fractograph of the T6 condition specimen consists of large flat cleavage planes. The dramatic difference in the fracture surface after FSP is a direct result of the grain size reduction in this study. This dimple fracture surface is the result of microvoid coalescence and not cleavage of grains as seen in the T6 sample.

Fig. 15. SEM images of fracture surface of EV31A in T6 condition (left) and FSP condition (right). Appearance of the fracture of the T6 condition and FSP condition are typical of high and low ductility fracture.

4.4. DSC Results The reaction kinetics of similar alloys has been reported to be quite slow [64], and natural aging of the SSSS does not occur in the 3.5 month evaluation [34]. During friction stir processing, although temperatures can reach around 500°C, very short time at temperature leads to the question of whether dissolution of RE elements is possible. Since diffusivity is known to be governed by temperature, time, and strain, a reasonable conclusion is that the large strains dramatically affect the diffusion kinetics. Analysis of deformation mechanism maps [88] shows that when considering a strain rate of around 40 s-1 and temperature around 500°C, deformation mechanism is slip and controlling mechanism for diffusion is pipe. Pipe diffusion is very efficient; and since dislocations are being generated at a large number during processing, diffusion is expected to occur rapidly, and to be extenuated by the fine grain size. A DSC analysis of the FSP material compared to the T6 condition (Fig. 16) demonstrates the significant dissolution that is achieved during FSP. Unfortunately, the lack of knowledge of precipitation in this alloy precludes quantitative analysis. Although the exact composition of the precipitates is unknown, the suggested precipitation sequence is shown in Table 2.

Fig. 16. DSC analysis of samples in the FSP condition compared to the T6 condition. Significant dissolution is obtained through FSP as evident by the presence of exothermic peaks P1 & P2.

A preliminary DSC investigation was done on EV31A and WE43 by Riontino et al. [89]; yet due to lack of literature on the precipitation sequence of this alloy, only one reaction was identified for EV31A. In their study the exothermic reaction at 300°C was attributed to formation of β1 phase based on the same reaction observed in WE43. In the FSP condition an exothermic peak, P1, present around 150°C was observed. The presence of this peak is not clearly understood and has not been previously observed in the air-cooled DSC analysis of EV31A [89]. The DSC study of the water-quenched WE43 sample demonstrated exothermic peak at the same temperature observed in this study and is attributed to the simultaneous formation of β” and β’. Thus P1 is likely due to the formation of similar β” and β’. In both the FSP- and the air-cooled samples [89], P2 is present, though the size of the peak in the FSP sample is much greater, thus indicating that the FSP time at temperature is a more effective method of retention of SSSS than air cooling from solutionizing

47 temperatures. The curve for the T6 sample shows no P2, but a very strong D1 and P3, which implies that although the sample is aged at 200°C, the precipitation is converted to its equilibrium state. The presence and strength of P2 in the FSP sample and the subsequent P3 suggest that P2 is due to the formation of β1 and P3 corresponds to its conversion to equilibrium β. The remaining peaks seen in the FSP sample are unique to the FSP microstructure. P4 was observed in the DSC curve of the FSP + T6 sample as well. The large peak at P5 is believed to be due to grain coarsening of the fine-grained FSP microstructure. Further TEM investigation would be needed to validate the precipitation sequence of this alloy.

5. Conclusions FSP of EV31A resulted in a reduction of grain size, breakage and dissolution of second phase particles. The grain size reduction was from 78.5 ± 22 µm in the T6 condition to 3.4 ± 0.26 µm. The FSP particle size was measured to be 0.05 ± 0.04 µm 2. A significant dissolution of strengthening precipitates achieved during friction stir processing resulted in a peak 0.2% yield strength in the FSP + Age condition of 266.8 ± 16 MPa.

Acknowledgments The author thanks Paul Lyon of Magnesium Elektron for supply of EV31A plates.

References [1] J. Aragones, K. Goundan, “Development of the 2006 Corvette Z06 Structural Cast Magnesium Crossmember”, SAE International (2005). [2] A. Wendt, K. Weiss, A. Ben-Dov, M. Bamberger, B. Bronfin, “Magnesium Castings in Aeronautics Applications-Special Requirements”, TMS Magnesium Technology (2005) 269-273. [3] S. Schumann and H. Friedrich, “Current and Future use of Magnesium in the Automobile Industry”, Materials Science Forum , 51(2003), 419-422. [4] D.A. Kramer, Magnesium, Its Alloys and Compounds, U.S. Geological Survey Open-File Report 01-341 http://pubs.usgs.gov/of/2001/of01-341/ .

[5] G. Cole, Magnesium Applications in Ford Motor Company’s PNGV Project P2000 Contour Proceeding of the 54 th International Magnesium Association, Toronto, CAN (1997). [6] H. Friedrich and S. Schumann, The Second Age of Magnesium- Research Strategies to Bring the Automotive Industry’s Vision to Reality, Proceedings of the 2 nd Israeli International Conference on Magnesium Science and Technology, pp 9-26, Dead Sea Israel, (2000). [7] N. Li, R. Osbourne, B. Cox, D. Penrod, Magnesium Engine Cradle – The USCAR Structural Cast Magnesium Development Project SAE International (2005). [8] R.S. Mishra, Z.Y. Ma, “Friction Stir Welding and Processing: A Review Journal”, Materials Science R (2005) v50: 1-78. [9] R.S. Mishra, M.W. Mahoney, S.X. McFadden, N.A. Mara, A.K. Mukherjee, Scripta Mater. 42 (2000) 163. [10] R.S. Mishra, M.W. Mahoney, Mater. Sci. Forum 357–359 (2001) 507. [11] W.M. Thomas, E.D. Nicholas, J.C. Needham, M.G. Murch, P. Templesmith, C.J. Dawes, G.B. Patent Application No.9125978.8 (1991). [12] Z.Y. Ma, S.R. Sharma, R.S. Mishra. “Effect of multiple pass friction stir processing on microstructure and tensile properties of a cast aluminum-silicon alloy”, Scripta Materialia v54(2006) 1623-1626. [13] M.L. Santella. T. Engstrom, D. Storjohann, T.Y. Pan, Scripta Materialia 2005; 53:201. [14] Z.Y. Ma, S.R. Sharma, R.S. Mishra, M.W. Mahoney, Material Science Forum 2003; 426-432:2891. [15] Z.Y. Ma, S.R. Sharma, R.S. Mishra, “Effect of friction stir processing on the microstructure of cast A356 aluminum”, Mat. Sci. A: 433 (2006); 269-278. [16] M. Santella et al. “The Use of Friction-Stir Technology to Modify the Surfaces of AM60B Magnesium Die Castings”, JOM May 2006. V. 58 No.5. [17] R. Johnson, Mater. Sci. Forum 419–422 (2003) 365. [18] W.B. Lee, Y.M. Yeon, S.B. Jung, Mater. Sci. Technol. 19 (2003) 785. [19] G. Kohn, S. Antonsson, A. Munitz, Automotive Alloys 1999, TMS, 2000, 285–292.

[20] P. Lyon, “New Magnesium Alloy for Aerospace and Specialty Applications”, TMS Magnesium Technology 2004: 311-315. [21] M. Tsujikawa, S.W. Chung, M. Tanaka, Y. Takigawa, S. Oki, K Higashi, “High Strengthening of Mg-Y-Zn Cast Alloy by FSP”, Materials Transactions V46/12, 2005, pp. 3081-3084. [22] T.A. Freeney, R.S. Mishra, G.J. Grant and R. Verma “Friction Stir Processing of a Cast WE43 Mg Alloy”, Friction Stir Welding and Processing IV, TMS 2007. [23] B.J. Coultes, J.T. Wood, G.Wang, R. Berkmortel, “Mechanical Properties and Microstructure of Magnesium High Pressure Die Castings”, TMS Magnesium Technology 2003, 45-50. [24] Magnesium Elektron UK: Datasheet 440: www.melpure.com . [25] ASM Specialty Handbook, Magnesium and Magnesium Alloys, Ed. Michael M. Avedesian and Hugh Baker, ASM International 1999. [26] B. Smola, I. Stulikova, F. von Buch, B.L. Mordike, Structural Aspects of high performance Mg alloys design, Mat Sci A324, 2002, pp. 113-117. [27] Negishsi, Nishimura et al. Journal of Japan Institute of Light Metals. Vol 45 pt 2. [28] J.F. Nie, B.C. Muddle, “Characterization of Strengthening Precipitate Phases in Mg- Y-Nd Alloy”, Acta Mat. 48, 2000, 1691-1703. [29] D.H. St. John, M. Qian, M.A. Easton, P. Cao, Z. Hildebrand. “Grain Refinement of Magnesium Alloys”, Metall. and Mat. Trans. A: v36A:2005;1669. [30] P. Cao, M. Qian, and D.H. St. John: Scripta Mater., 2004, vol. 51, pp. 125-29 [31] E.F. Emley, “Principles of Magnesium Technology” (Pergamon Press, Oxford ; New York, ed. [1st ]. 1966). [32] Ma Qian, D. Graham, L. Zheng, D.H. StJohn and M.T. Frost. Material Science and Technology, 2003, 19, 156-162. [33] M. Qian, D.H St. John and M.T. Frost: Scripta Mater.,2002, vol.46,pp. 649-54. [34] L.L. Rohklin, in: J.N. Fridlyander, D.G. Eskin (Eds.) “Magnesium Alloys Containing RE Metals: Structure and Properties”, Taylor and Francis, 2003. [35] E.D. Levine, W.F. Sheely, R.R. Nash, Trans. AIME, 1959, Vol 215, p 253. [36] C.R. Brooks, “Heat Treatment Structure and Properties of Nonferrous Alloys”, ASM 1982 255-256.

[37] Zhang, “Phase Constitution and Morphologies of Magnesium-Zirconium-Zinc Alloy MB15”, Acta Met. Sinica (English Edition) 3A (1989), 110-115. [38] S. Cohen, G.R. Goren-Muginstein, S. Avraham, G. Dehm, M Bamberger, “Phase Formation, Precipitation and Strengthening Mechanisms in Mg-Zn-Sn and Mg-Zn- Sn-Ca Alloys”, TMS, Mag Tech 2004, pp 301-305. [39] J.C. Rao, M. Song, K. Furuya, S. Yoshimoto, M. Yamasaki, Y. Kawamura, “TEM study of Mg-Zn precipitates in Mg-Zn-Y alloys”, J. Mater Sci, V41, 2006, 2573- 2576. [40] Y. Kawamura, K. Hayashi, A. Inoue, T. Masumoto, Mater. Trans A26, 1995, 1705. [41] A. Inoue, Y. Kawamura, Y. Matsushita, K. Hayashi, J. Koike, J. Mater. Res. 16, 2001, 1894. [42] E. Abe, Y. Kawamura, K. Hayashi, A. Inoue, Acta Mater. 50, 2002, 3845. [43] A. Inoue, Y. Matsushita, Y. Kawamura, K. Amiya, K. Hayashi, J. Koike, Mater. Trans. 43, 2002, 580. [44] D.H. Ping, K. Hono, Y. Kawamura, A. Inoue, Phil. Mag. Lett. 82, 2002, 543. [45] Z. Hildebrand, M. Qian, D. StJohn, M. Frost, “Influence of Zn on the soluble Zr content in Mg and the subsequent grain refinement by zirconium”, Mag Tech, 2004, pp. 241-245. [46] Löhberg, K. & Schmidt, G. (1975) Grain refining effect of zirconium in magnesium alloys, Giessereiforschung. 27(3): p.75-82. [47] S. Bhan and A. Lal, “The Mg-Zn-Zr System(Magnesium-Zinc-Zirconium)”, J. Phase Equilibria 14 (1993), 634-637. [48] R. Arroyave, A. van de Walle, Z.-K. Liu, “First-principle calculations of the Zn-Zr system”, Acta Mat 54, 2006, 473-482. [49] N.F. Lashko and G.I. Morozova, “Transformations of Nonequilibrium Dendrite Shapes in the Growth of the ZnZr Phase in Cast Alloys of the Mg-Zn-Zr System”, Soviet Phys.-Cryst. 14 (1969), 143-145. [50] T. Chun-Hu, Z. Shao-Qing, L. Li-Qi, “Influence of Homogenization on the Microstructure and Tensile Properties of Extruded Mg-Zn-Zr-RE Alloy”, Journal of Rare Earths, 1991, Vol. 9/2.

[51] G.I. Morozova, V.V. Tikhonova and N.F. Lashko, “Phase Composition and Mechanical Properties of Cast Mg-Zn-Zr Alloys”, Met. Sci. Heat Treat. (Russia) 20 (1978), 657-660. [52] J. Hasegawa, S. Takeuchi, and A.P. Tsai, “Stable Quasicrystals and approximants in Zn-Mg-Zr and Zn-Mg-Hf alloys”, Phil. Mag. Letters V85, N6, 2005, pp 289-297. [53] P. Lyon, T. Wilks, I. Syed, “The Influence of Alloying Elements and Heat Treatment upon the Properties of EV31A Alloy”, TMS, Mag Tech 2005, 303:308. [54] M.E. Drits, L.L. Rohklin, E.M. Padeezhnova I.I. Gurez, N.V. Miklina, T.V. Dobatkina, A.A. Oreshkina, “Magnesium Alloys with Yttrium”, Nauka, Moscow, 1979 (Russian). [55] M.E. Drits, Z.A. Sviderskaya and L.L. Rokhlin, Izv. Vyss. Uchebn. Zaved. Tsvetn. Metall., 3, 117-121, 1962 (Russian). [56] M.E. Drits, Z.A. Sviderskaya and L.L. Rokhlin, Metallurgiya, Metallovedenie I Fiziko-Khimicheskie Metody Issledovaniya, Russian Acad. Sci. Publishers, Moscow, 14, 120-129, 1963 (Russian). [57] P. Frankel — Elektron 21 Evaluation — Materials Science Investigation. Project for Magnesium Elektron 2003. [58] L. Gill, G.W. Lorimer, P. Lyon, Microstructure & Property Relationships of Mg-RE- Zn-Zr Alloys, DGM Conference, Germany, Nov. 2003. [59] G. Bruzzone, M.L. Fornasini, F. Merlo, “Rare Earth Intermediate Phases with Zn”, JLCM; 22, 1970, pp 253-264. [60] P. Manfrinetti, K.A. Gschneidner Jr., “Phase Equilibrium in the La-Mg and G-Mg Systems”, JLCM; 123, 1986, pp 267-275. [61] H. Okamato, “Mg-Nd”, J. Phase Equilibrium, 12, 1991, 249-250. [62] S. Delfino, A. Saccone, R. Ferro, “Phase Relationships in the Nd-Mg Alloy System”, Met. Trans. 21A, 1990, 2109-2114. [63] K.A. Gschneidner Jr., F.W. Calderwood, “The Gd-Nd System”, Bull. Alloy Phase Diagrams; 3(3), 1982; pp 351-353. [64] P.J. Apps, H. Karimzadeh, J.F. King, G.W. Lorimer, “Phase compositions in magnesium-rare earth alloys containing yttrium, gadolinium or dysprosium”, Scripta Materialia 48, 2003, pp 475-481.

[65] Y.M. Riddle, L.P. Barber, M.M. Makhlouf, “Characterization of Mg Alloy Solidification and As-Cast Microstructures”, TMS, Mag. Tech 2004 p. 203-208. [66] G.W. Lorimer, G. Cliff, P.E. Champness, C. Dickinson, F. Hasan, P.B. Kenway, Analytical electron microscopy. San Francisco Press (1984) p. 153. [67] G.W. Lorimer, in: Proceedings Magnesium Technology, Institute of Metals, London, 1986, p. 47. [68] P. Vostry, I. Stulokova, B. Smola, M. Cieslar, B.L. Mordike, Z. Metallkde. 79 (1988)340. [69] B. Smola, B.L. Mordike, in: 10 th European Crystallographic Meeting, Collected Abstracts, IUCr, Wroclaw, Poland, 1986, p. 411. [70] S. Kamado, Y. Kojima, S. Tanike, I. Seki, S. Hama, in: B.L. Mordike, K.U. Kainer (Eds.), Mg Alloys and Their Applications, Werkstoff-Informationsgesellschft, Frankfurt, Germany, 1998, p.229. [71] J.F. Nie, B.C. Muddle, in: B.L. Mordike, K.U. Kainer (Eds.), Mg Alloys and Their Applications, Werkstoff - Informationsgesellschft, Frankfurt, Germany, 1998, p. 229. [72] K.V. Jata, S.L. Semiatin, Scripta Mater. 43, 2000, 743. [73] P. Heurtier, C. Desrayaud, F. Montheillet, Mater. Sci. Forum 396-402, 2002, 1537. [74] X. Xu, S. Deng, in: Proceedings of 4 th Int FSW Symp, Park City, UT, USA. Cambridgae: TWI; 2003 (CDROM). [75] W.B. Lee, Y.M. Yeon, S.B. Jung, Mater. Sci. Technol. 19 (2003) 785. [76] S.H.C. Park, Y.S. Sato, H. Kokawa, in: S.A. David, T. DebRoy, J.C. Lippold, H.B. Smartt, J.M. Vitek (Eds.), Proceedings of the Sixth International Conference on Trends in Welding Research, Pine Mountain, GA, ASM International, 2003, p. 267. [77] .V. Jata, S.L. Semiatin, Scripta Mater. 43, 2000, 743. [78] J.-Q. Su, T.W. Nelson, R. Mishra, M. Mahoney, Acta Mater. 51, 2003, 713. [79] K.V. Jata, K.K. Sankaran, J.J. Rushau, Metall. Mater. Trans. A 31, 2000, 2181. [80] Y.S. Sato, M. Urata, H. Kokawa, Metall. Mater. Trans. A 33, 2002, 625. [81] J.-Q. Su, T.W. Nelson, C. J. Sterling, “Microstructure evolution during FSW/FSP of high strength aluminum alloys”, Mat. Sci.A:405, 2005, 277-286. [82] P. B. Pragnell, C. P. Heason, “Grain structure formation during friction stir welding observed by the ‘stop action technique’ ”, Acta. Mat. 53, 2005, 3179-3192.

[83] K.A.A. Hassan, B.P. Wynne, P.B. Pragnell, Fourth International Symposium on FSW: TWI, Park City, UT, USA, 2003 (CD ROM). [84] W. Blum, Q. Zhu, R. Merkel, H.J. McQueen, Mater. Sci. Eng. A 205, 1996, 23. [85] K.N. Krishnan, “On the formation of onion rings in friction stir welds”, Mater. Sci. Eng. A:327 246-251. [86] H. Larson, et al., “Joining of Dissimilar Al-alloys by Friction Stir Welding”, Second Int. Conf. on Friction Stir Welds, Gothenburg, Sweden, 2000. [87] Z.Y. Ma, S.R. Sharma, R.S. Mishra, “Microstructural Modification of As-Cast Al-Si- Mg Alloy by Friction Stir Processing”, Met. Trans. A: 37A (2006);3323-3336. [88] W.-J. Kim, S.W. Chung, C.S. Chung, D. Kum, “Superplasticity in Thin Mg Alloy Sheets and Deformation Mechanism Maps for Mg Alloys at Elevated Temperatures”, Acta Mat. 49, 2001, 3337-3345. [89] G. Riontino, D. Lussana, M. Massazza, “DSC investigation on WE43 and Elektron 21 Mg Alloys”, J. Mater. Sci. 41, 2006, 3167-3169.

VITA

Timothy A. Freeney was born in New Orleans, LA, on the 15 th of October 1981 to Arthur V. Freeney Jr. and Cynthia L.Y. Freeney. He was raised in Tangipahoa Parish, LA, where he attended schools in both Hammond and Independence, LA. He enrolled at the University of Missouri-Rolla in the fall of 1999, received his Bachelor of Science degree in Metallurgical Engineering in spring of 2005 and completed his Master of Science degree in Materials Science and Engineering in the spring of 2007. Timothy was a brother in Sigma Nu fraternity; he also was a member of Alpha Sigma Mu, ASM, TMS and Tau Beta Pi.