University of Southern California Dissertations and Theses

Properties of a ZK60 magnesium alloy processed by severe plastic deformation

(USC Thesis Other)

Properties of a ZK60 magnesium alloy processed by severe plastic deformation

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 1 PROPERTIES OF A ZK60 MAGNESIUM ALLOY PROCESSED BY SEVERE PLASTIC DEFORMATION by Seyed Alireza Torbati Sarraf A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (MECHANICAL ENGINEERING) August 2016 Copyright 2016 Seyed Alireza Torbati Sarraf i Acknowledgements I would like to express my sincere gratitude to my advisor, Professor Terence G. Langdon, for his encouragement, support and guidance during the entire course of this program. His support towards my career and invaluable advice made my achievement possible and his enlightening guidance and earnest attitude will also benefit me in my future path. My special thanks also extend to his wife, Mady Langdon, for her concern and kindness. I am grateful to my dissertation committee, Professors Michael Kassner and Edward Goo for their agreement to serve on my dissertation committee, for their feedback and evaluating my Ph.D. study. I would like to thank Professor Roberto Figueredo in the Division of Materials Science and Engineering at Universidade Federal de Minas Gerais in Brazil for his advices throughout my studies. It is my pleasure to acknowledge the significant help and beneficial discussions provided by Dr. Shima Sabbaghianrad and Dr. Mahmood Shirooyeh. Most importantly, none of this would have been possible without the love of my family. I would like to present this achievement as a gift to my wonderful family. It is their everlasting love, unconditional support and understanding that strengthened my determination in pursuing this degree. This work was supported by the National Science Foundation of the United States under Grant No. DMR-1160966. ii Table of Contents Acknowledgements i Table of Contents ii List of Tables v List of Figures vi Abstract xvi 1. Introduction 1 2. Literature review 2 2.1. Severe Plastic Deformation (SPD) 2 2.2. Equal-Channel Angular Pressing (ECAP) 3 2.2.1. Microstructural evolution after processing by ECAP 7 2.2.2. Grain refinement in magnesium alloys by ECAP 8 2.3. High-Pressure Torsion (HPT) 8 2.3.1. Microstructural evolution after processing by HPT 15 2.3.2. Grain refinement in magnesium alloys by HPT 16 2.3.3. Hardness homogeneity after processing by HPT 18 2.4. Superplasticity after processing by SPD 20 2.4.1. The implication of experimental results for superplasticity 22 2.5. Combination of SPD processes 23 2.6. Paradox of strength and ductility in ultrafine-grained materials 26 2.6.1. Low temperature ductility after HPT 29 3. Experimental Materials and Procedure 31 3.1. Experimental materials 31 3.2. Methods of severe plastic deformation 31 3.2.1. Processing by ECAP 31 3.2.2. Processing by HPT 33 3.2.3. ECAP followed by HPT processing 34 3.3. Microstructural analysis 35 3.3.1. Optical microscopy 35 3.3.2. Transmission electron microscopy (TEM) 35 iii 3.3.3. Scanning electron microscopy (SEM) 36 3.3.4. Electron-backscatter diffraction (EBSD) 36 3.4. Mechanical experiments 37 3.4.1. Microhardness measurements 37 3.4.2. Tensile measurements 39 3.4.3. Shear punch tests 40 3.5. Annealing processes 41 4. Experimental results 42 4.1. Microstructural analysis 42 4.1.1. Microstructural characteristics before and after HPT 42 4.1.2. Microstructural characteristics of the samples after combination of SPD processes (ECAP + HPT) 56 4.1.3. Microstructural characteristics of the samples annealed after processing by HPT 61 4.2. Textural analysis 66 4.2.1. Textural evolution of the samples processing by HPT 66 4.2.2. Textural evolution of the samples after processing by combination of SPD processes (ECAP + HPT) 67 4.2.3. Textural evolution of the samples after post-HPT annealing 70 4.3. Microhardness measurements 71 4.3.1 Microhardness values after processing by HPT 71 4.3.2. Microhardness values after processing by combination of SPD processes (ECAP + HPT) 74 4.3.3. Microhardness values after annealing HPT processed materials 76 4.4. Mechanical behavior 78 4.4.1. Tensile properties of the samples processed by HPT 78 4.4.1.1. Evaluating the flow behavior of ZK60 alloy: strain rate sensitivity, m, and activation energy, Q, after processing by HPT 83 4.4.2. Tensile properties after annealing sample processed by HPT 85 4.4.3. Shear punch testing 87 iv 5. Discussion 5.1. The effect of HPT 95 5.2. The effect of ECAP + HPT 113 5.3. The effect of post-HPT annealing process 121 5.4. Shear punch test 127 6. Summary and Conclusions 130 References 133 v List of Tables Table 2.1. Grain structures produced in pure magnesium and magnesium alloys using HPT and microhardness values. 20 Table 2.2. Reports of superplasticity in ultrafine-grained materials produced by ECAP. 21 Table 4.1. Results for the average grain sizes and the related uncertainty for disks processed by various numbers of turns in HPT at ambient temperature: results are shown for the center, half-radius and edge positions. 45 Table 4.2. Average grain size and area fraction percentile of grains smaller than 1.0 and 0.5 µm measured for studied materials. 51 Table 4.3. The HAGBs, LAGBs, 30 o and 85 o -90 o misorientation angles area fraction percentiles measured for studied materials. 53 Table. 5.1. Summary of the obtained saturation level and hardenability of HPT processed materials 298/Tm is homologous temperatures of processing at room temperature 108 Table 5.2. Reports of superplasticity in ultrafine-grained materials produced by HPT 111 Table 5.3 Superplastic characteristics in some magnesium alloys 112 vi List of Figures Figure 2.1 The principle of ECAP, showing the two angles Φ and Ψ [28]. 4 Figure 2.2 (a) The four fundamental processing routes for ECAP [34] and (b) the slip systems viewed on X, Y and Z planes for consecutive passes using processing routes in part (a) [35]. 5 Figure 2.3 Illustration of the X, Y and Z planes denote the transverse plane, the flow plane and the longitudinal plane, respectively [35]. 6 Figure 2.4 Variation of the equivalent strain, , with the channel angle, , for a single pass where N = 1 [37]. 6 Figure 2.5 A model for grain refinement in the central region of the billet in ECAP [38]. 7 Figure 2.6 A model for grain refinement of magnesium alloys processed by ECAP [26]. 10 Figure 2.7 Schematic illustration of quasi-constrained HPT processing [45]. 11 Figure 2.8 Schematic illustration of HPT for (a) unconstrained, (b) constrained and [46] (c) quasi-constrained conditions [47]. 12 Figure 2.9 Parameters used in estimating the total strain in HPT [48]. 13 Figure 2.10 Variation of equivalent strain as a function of number of turns in a disk with a thickness of 0.1 mm processed by HPT [46]. 14 Figure 2.11 Schematic illustration of microstructural evolution with straining in pure Al. Thin double lines depicted in region II represent low angle boundaries with some extension of boundary width and thick lines in region III represent high angle boundaries [57]. 16 vii Figure 2.12 Optical micrographs showing the microstructures after HPT after processing through 1, 3, 5 and 7 turns at a temperature of 423 K [18]. 17 Figure 2.13 Equivalent TEM micrographs in bright-field and dark-field, plus the associated SAED patterns, after HPT at 423K through 1, 3, 5 and 7 turns [18]. 18 Figure 2.14 Values of the Vickers microhardness plotted as a function of the equivalent strain in disks processed by HPT at temperatures of (a) 296 K, (b) 373 K, and (c) 473 K [50]. 20 Figure 2.15 Plots of strain rate versus (a) elongation to failure and (b) spatial grain size for a series of Al alloys produced using different processing methods superimposed on the plot, and delineated by the dashed ovals, are data collected for aluminum alloys after ECAP and HPT [74]. 23 Figure 2.16 Crystallize size distributions for nickel samples obtained by combinations of different methods: (a) ECAP, (b) ECAP + CR, (c) ECAP + HPT, and (d) ECAP + CR + HPT [81]. 24 Figure 2.17 EBSD images of the Al-7075 alloy in the (a ) annealed condition and after processing by (b) HPT through a total of 5 turns, (c) ECAP for 4 passes, (d) ECAP for 8 passes, and of ECAP for 4 passes and (e) HPT for 20 turns [86]. 25 Figure 2.18 Cold rolling of Cu and Al increases their yield strength but decrease their ductility. The extraordinary combination of both high strength and high ductility in nanostructured Cu and Ti processed by SPD clearly sets them apart from coarse-grained metals [87]. 26 Figure 2.19 Variation of (a) and (d) the normalized YS, (b) and (e) the normalized UTS with the normalized elongation to failure and (c) and (f) the variation of the normalized UTS with the normalized uniform elongation corresponding to the strain at UTS for samples processed by HPT at 298 and 445 K, respectively [88]. 27 viii Figure 2.20 The diagram showing the strength-ductility paradox as developed by Valiev where the shaded area represents the conventional behavior, the region of high strength and high ductility lies to the right of the solid and broken line and the numbers marked with percentages denote the strains imposed during rolling: the present results for the Al-7%Si alloy are superimposed on the plot and the displacement to the right with increasing number of HPT turns, N, is indicated by the arrow [88]. 28 Figure 2.21. Stress–strain curves at room temperature for the annealed material, after HPT through10 turns and after HPT through 10 turns and annealing at 423,473 and 523K at (a) 1.0 × 10 -4 , (b) 1.0 × 10 -3 and (c) 1.0 × 10 -2 s -1 respectively [6]. 30 Figure 3.1 Drawings of the die and plunger of ECAP. 32 Figure 3.2 Schematic of the combination of ECAP and HPT. 34 Figure 3.3 Near surface position of the disks for EBSD sample preparation with ion beam cross-sectional polishing machine. ND, TD and RD correspond to Normal Direction (torsion axis), Tangential Direction and Radial Direction, respectively. The shear direction is parallel to TD. 37 Figure 3.4 Schematic illustration of microhardness indentations configuration for (a) linear diagonal (b) rectilinear grid pattern methods on HPT disk. 38 Figure 3.5 Dimensions of tensile specimens after HPT processing [103]. 39 Figure 3.6 Schematic representation of (a) shear punch-die assembly [104] and (b) shear zone of sample inside the dies [105]. 40 Figure 3.7 The assembly of the specimen and SPT fixture between the compression platens. 41 Figure 4.1 Optical microstructures of As-received extruded ZK60. 42 Figure. 4.2 Optical micrographs of the microstructures after HPT at the center, half- radius and periphery positions of disks after processing through 1/4, 1/2, 1 and 5 turns at room temperature. 44 ix Figure 4.3 Optical micrograph after HPT processing for N = 1/4 turn at the periphery of the disk. 46 Figure 4.4 SEM micrographs after HPT processing for N = 1/4 turn near the center of the disk. 46 Figure 4.5 TEM images of ZK60 alloy processed by HPT after 5 revolutions: (a) bright field image, (b) conventional dark field image at the same position and (c) weak beam dark field at higher magnification for the grain marked in (b). 47 Figure 4.6 Image quality (IQ) maps of ZK60 (a) as-received and after processing by (b) 0 turn, (c) 1/2 turn, (d) 1 turn and (e) 5 turns of HPT in the mid-radius of the disks. 49 Figure 4.7 Grain size distribution of ZK60 in the (a) as-received extruded condition and after (b) 0 turn, (c) 1/2 turn, (d) 1 turn and (e) 5 turns of HPT in the mid-radius of the disks. 50 Figure 4.8 Number fraction of the misorientation angles of ZK60 in the (a) as- received extruded condition and after (b) 0 turn, (c) 1/2 turn, (d) 1 turn and (e) 5 turns of HPT in the mid-radius of the disks. 54 Figure 4.9 EBSD orientation micrographs of ZK60 in the (a) as-received condition and after (b) 0 turn, (c) 1/2 turn, (d) 1 turn and (e) 5 turns of HPT in the mid-radius of the disks. 55 Figure 4.10 Image quality (IQ) maps of ZK60 after (a) processing by initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT for (b) 5 and (c) 10 turns of HPT under applied pressure of 2.0 GPa. 57 Figure 4.11 Grain size distribution of ZK60 after (a) processing by initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT for (b) 5 and (c) 10 turns of HPT under applied pressure of 2.0 GPa. 58 x Figure 4.12 Number fraction of the misorientation angles (a) processing by initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT for (b) 5 and (c) 10 turns of HPT under applied pressure of 2.0 GPa. 59 Figure 4.13 EBSD maps of ZK60 after (a) processing by initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT for (b) 5 and (c) 10 turns of HPT under applied pressure of 2.0 GPa. 60 Figure 4.14 IQ maps of ZK60 processed by 5 turns of HPT after (a) long term annealing after 2500 minutes at 450 K and short term annealing after 20 min annealing at temperature of (b) 473 K and (c) 523 K. 62 Figure 4.15 Grain size distribution of ZK60 processed by 5 turns of HPT after (a) long term annealing after 2500 minutes at 450 K and short term annealing after 20 min annealing at temperature of (b) 473 K and (c) 523 K. 63 Figure 4.16 Number fraction of the misorientation angles after (a) long term annealing after 2500 minutes at 450 K and short term annealing after 20 min annealing at temperature of (b) 473 K and (c) 548 K. 64 Figure 4.17. EBSD maps of ZK60 processed by 5 turns of HPT after (a) long term annealing after 2500 minutes at 450 K and short term annealing after 20 min annealing at temperature of (b) 473 K and (c) 548 K. 65 Figure 4.18 The pole figure of {0001 } , {101 ̅ 0} and {21 ̅ 1 ̅ 0} crystallographic orientations on the TD-RD plane and normal to the ND of HPT direction from the mid-radius of the extruded disk and processed disks with N = 0, 1/2, 1 and 5 turns of HPT at 298 K. 68 Figure 4.19 The pole figure of {0001 } , {101 ̅ 0} and {21 ̅ 1 ̅ 0} crystallographic orientations on the TD-RD plane and normal to the ND of HPT direction from the mid-radius of the extruded disk of ZK60 after processing by initial ECAP for 4 passes or a combination of ECAP for 4 passes and HPT for 5 and 10 turns of HPT under applied pressure of 2.0 GPa. 69 xi Figure 4.20 The pole figure of {0001 }, {101 ̅ 0} and {21 ̅ 1 ̅ 0} crystallographic orientations on the TD-RD plane and normal to the ND of HPT direction from the mid-radius of samples processed by 5 turns of HPT and after post-HPT annealing. 70 Figure 4.21 Vickers microhardness across diameters of disks of the ZK60 processed through various numbers of turns by HPT: the lower dashed line shows the as-received condition. 73 Figure 4.22 Variation of the microhardness values versus equivalent strain of ZK60 after processing by 1/8, 1/4, 1/2, 1, 2 and 5 turns of HPT under applied pressure of 2.0 GPa. The lower dashed line represents the microhardness value of the as-received extruded material. 73 Figure 4.23 Color-coded contour maps showing the distributions of the Vickers microhardness values on the surfaces of disks processed by HPT through (a) 1/4, (b) 1/2, (c) 1 and (d) 5 turns: the significance of the colors is shown by the key at the lower right. 74 Figure 4.24 Vickers microhardness across diameters of disks of the ZK60 processed through through 4 passes of ECAP and various numbers of turns by HPT: the lower dashed line shows the as-received condition. 75 Figure 4.25 Variation of the Vickers microhardness values versus equivalent strain of ZK60 after processing by initial ECAP for 4 passes or a combination of ECAP for 4 passes and HPT for 1/2, 1, 5 and 10 turns of HPT under applied pressure of 2.0 GPa. 76 Figure 4.26 Vickers microhardness across diameters of disks of the ZK60 processed through 5 turns of HPT and annealed for (a) long term and (b) short term at different tempratures: the lower dashed line shows the as-received condition. 77 Figure 4.27 (a) Plots of engineering stress versus engineering strain (b) appearance of tensile specimens for the ZK60 magnesium alloy processed by different 79 xii turns of HPT pulled in tension to failure at an initial strain rate of 1.0 × 10 -3 and 473 K. Figure 4.28 Plots of engineering stress versus engineering strain for the ZK60 magnesium alloy pulled in tension to failure at: (a) 473, (b) 523 and (c) 573 K at initial strain rate ranging from 1.0 × 10 -1 to 3.0 × 10 -5 s -1 . 80 Figure 4.29 The tensile samples pulled in tension to failure at: (a) 473, (b) 523 and (c) 573 K at initial strain rate ranging from 1.0 × 10 -1 to 3.0 × 10 -5 s -1 . 81 Figure. 4.30 Elongation to failure versus initial strain rate for tests conducted at different temperatures. 82 Figure 4.31 The normalized stress versus temperature compensated strain rate for studied material shows a sigmoidal behavior of ZK60. 84 Figure 4.32 The variation in 𝜀 ̇( 𝑇 /𝐺 ) as a function of 1/T for the ZK60 magnesium alloy processed HPT. 85 Figure 4.33 Engineering stress vs engineering strain curves of samples annealed isothermally after HPT for 10 and 20 min at temperatures of 473, 523 548 K and tested at different strain rate of 10 -3 and 10 -4 s -1 . 86 Figure 4.34 Engineering stress vs engineering stress curves of ZK60 after long term annealing at 450 K for 2500 min at different initial strain rate of 10 -3 and 10 -4 s -1 . 87 Figure 4.35 Plots of shear stress versus normalized displacement for the ZK60 magnesium alloy processed by different turns of HPT at punch rate of 5.0 × 10 -2 mm/min and 473 K 88 Figure 4.36 Plots of shear stress versus normalized displacement for the extruded ZK60 magnesium alloy at: (a) 473, (b) 523 and (c) 573 K at different punch rate ranging from 8.0 to 1.0 × 10 -2 mm/min. 89 Figure 4.37 Plots of shear stress versus normalized displacement for the ZK60 magnesium alloy at: (a) 473, (b) 523 and (c) 573 K at different punch rate ranging from 1.0 to 1.0 × 10 -2 mm/min. 90 xiii Figure 4.38 Normalized shear yield stress versus temperature compensated strain rate for (a) A-received and (b) processed by 5 turns of HPT. 93 Figure 4.39 The variation in  y/G as a function of 1/T for the ZK60 for (a) as received and (b) processed by 5 turns of HPT. 94 Figure. 5.1 Distribution maps of ultrafine grains ( < 1 µm) during straining sample through (a) 1/2, (b) 1 and (c) 5 turns of HPT. 97 Figure 5.2 Analysis of (a) IQ maps of ZK60 after 1 turn of HPT containing fine and coarse grain aggregates, (b) grain size distribution of corresponding fine and coarse grain aggregates. 98 Figure 5.3 Schematic diagram illustrating the formation of a shear zone from the bands of recrystallized grains. 99 Figure 5.4 Grain size distribution of ZK60 in the as-received condition and after 0, 1/2, 1 and 5 turns of HPT in the mid-radius of the disks. 100 Figure 5.5 Comparison of number fraction of misorientation angles in ZK60 alloy for the extruded condition and after processing by HPT for 0, 1/2, 1, and 5 turns of HPT in the mid-radius of the disks 101 Figure 5.6 Orientation of crystals of newly recrystallized grains formed within the twin bands rotated in parallel around the c-axis of their parent twin by an average of 30 o . 103 Figure 5.7 Orientation of crystals of random A and B grains in the sliced disks of as- received extruded materials before HPT processing. 104 Figure 5.8 Schematic illustration of the variation of the Vickers microhardness across the disk at low total strains in HPT for materials having either slow or fast rates of recovery [138]. 106 Figure 5.9 Plot of Hv values against ε eq in double logarithmic scale for the ZK60 alloy after up to 5 turns of HPT under 2.0 GPa at ambient temperature. The inclined solid line demonstrates the relationship between data set up 107 xiv to ε eq ≈ 20 prior to saturation and dashed plateau line demonstrates the saturation limit. Figure 5.10 Evolution of grain size distribution of ZK60 processed by an initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT through up to 10 turns. 114 Figure 5.11 Schematic of Vickers microhardness against equivalent strain for samples processed only by HPT or by combinations of ECAP + HPT. 116 Figure 5.12 Plot of Hv values against ε eq in double logarithmic scale for the ZK60 after up to 10 turns of HPT under 2.0 GPa at ambient temperature. The inclined solid line demonstrates the relationship between data set up to ε eq ≈ 100 prior to saturation and dashed plateau line demonstrates the saturation limit. Initial material process by 4 passes of ECAP at 473 K prior to HPT. 118 Figure 5.13 Plots of hardenability lines of ZK60 alloy during HPT processing thorough up to saturation level for materials with different initial conditions. 119 Figure 5.14 TEM micrograph and EDS analysis of precipitates in the ZK60 alloy in the extruded condition (a) and (c) and after processing by ECAP at 473 for 8 passes (b) and (d). 120 Figure 5.15 Hall-Petch relationship of ZK60 processed by HPT and annealed at 473- 548 K for 20 min. 122 Figure 5.16 The SEM fractography of tensile specimens tested at room temperature for processed by 5 turns of HPT (a) and (c) and (e) and after post –HPT annealing at 473 K for 20 minutes (b) and (d) and (e). 124 Figure 5.17 Variation of the normalized UTS with the normalized elongation to failure of samples undergone post-HPT annealing and pulled into failure at room temperature at strain rate of (a) 1.0 × 10 -3 s -1 and (b) 1.0 × 10 -4 s -1 . 125 xv Figure 5.18 Micrograph of (a) dark field STEM image in the HPT deformed sample for 3 revolutions and bright field TEM images of precipitates in the sample after heat treatment at 423 K for (b) 600 s and (c) 128 h. Arrows in highlights the interaction between precipitates and grain boundaries. 127 Figure 5.19 Schmid factor distribution of (a) and (b) extruded material and (c) and (d) HPT processed materials for tensile and shear punch test respectively. 128 xvi Abstract A commercial extruded ZK60 magnesium alloy was processed by severe plastic deformation (SPD) techniques of high-pressure torsion (HPT) and a combination of equal-channel angular pressing (ECAP) followed by HPT. The relationship between processing, structure and properties were analyzed for this material. Post-HPT short term annealing was conducted on the HPT processed materials to improve poor ductility after HPT at room temperature. The microstructural properties were examined, microhardness measurements were recorded across the disk diameters and through the whole surfaces, and miniature tensile specimens were pulled to failure at temperatures ranging from 473 to 573 K. Shear punch tests were also conducted at identical temperatures and results were compared to tensile tests. EBSD analysis demonstrated that processing by HPT technique has a potential for producing an ultrafine-grained (UFG) structure magnesium alloy at room temperature containing reasonably equiaxed grains and majority of high-angle grain boundaries. Microstructures were refined to different levels of grain size depending on the processing operations. It was observed that the most grain refinement was achieved after processing by a single step SPD through 5 turns of HPT with a minimum average grain size of 700 nm. The grain refinement mechanism was consistent with nucleation of new grains along pre-existing grain boundaries and formation of bimodal structures during HPT processing and continuous evolution to homogeneous distribution of ultrafine grains. A strain hardening behavior towards hardness homogeneity detected for ZK60 processed by HPT and ECAP + HPT with a hardenability exponent, η, of 0.07 and 0.03 were measured for ε eq up to ~20 and ~100, respectively, where the hardness values reasonably saturated at ~125 kgf/mm 2 for HPT and 117 kgf/mm 2 under ECAP + HPT at room temperature and at a constant xvii pressure of 2.0 GPa and rotational speed of 1 rpm. Textural analysis at mid-radius of the un-processed and processed materials illustrated a gradual evolution from prismatic {101 ̅ 0} fiber of extruded materials to an ultimate basal {0001 } fiber texture with c-axis parallel to ND for both HPT and ECAP + HPT processed samples. The majority of the grain boundaries have misorientation larger than 15 o throughout processing which differs from that observed in face centered cubic metals. The superplastic behavior of the processed alloy was investigated through measuring strain rate sensitivity by miniature tensile testing and miniature shear punch testing (SPT). The tensile experiments were conducted at initial strain rates of 3.0 × 10 −5 to 1.0 × 10 −1 s −1 and SPT tests were conducted at punch rates in the range of 8.0 to 1.0 × 10 −2 mm/min at temperatures of 473 and 523 and 573 K. Results showed that the strain rate sensitivity index, m, has a maximum of ~0.5 in the intermediate range of strain rates, region II, in both SPT and tensile test methods and hence, SPT and tensile experimental results were in a good agreement with each other. The strain rate sensitivity index of ~0.5 and an elongation to failure of >400% with a maximum of 940% when testing with an intermediate initial strain rate of 1.0 × 10 −4 s −1 at 523 K were indicative of a superplastic deformation behavior for the material processed by HPT. Post-HPT short term annealing results showed that annealing condition for 20 min at 548 K could improve the elongation to failure of HPT process materials up to 2.8 times with a decrease in strength of 0.93 times. An optimum annealing condition also was obtained at 473 K for 20 min which could provide high strength-high ductility ZK60 after HPT with higher elongation to failure of 1.7 times and UTS of 1.06 times higher than HPT processed material. 1 1. Introduction In 1988, a revolutionary report was published demonstrating the potential for achieving very small grain sizes within the range of submicrometer by application of severe plastic deformation (SPD) to bulk coarse-grained solids [1]. The processing of metals through the application of SPD is a procedure for producing bulk fully-dense solids with ultrafine-grained (UFG) structures and grain sizes within the submicrometer or nanometer ranges [2]. Development of UFG metals provides an opportunity to understand flow behavior of materials within range of submicron grain sizes. Two SPD methods currently receiving considerable attention are equal- channel angular pressing (ECAP) [3] and high-pressure torsion (HPT) [4]. These methods of SPD have permitted fabrication of relatively large samples containing UFG microstructures. It is well established that UFG metals experience highly deformed microstructures containing high dislocation densities within the grains [5,6]. Having high-energy non-equilibrium grain boundaries is one of the characteristics of SPD grain structure containing an excess of extrinsic dislocations [7,8,9]. First applications of HPT may be traced back to the 1940s [10]. However considerable attention to develop this process was established only within the last decade because it is now recognized that it provides the ability to produce materials having exceptionally small grain sizes which are generally smaller than those produced using ECAP [11,12]. Limited ductility of some materials at room temperature such as Mg alloys [13] inhibits processing by ECAP at ambient temperature [14,15]. However earlier experimental studies [16,17] show that it is possible to process Mg alloys by HPT at room temperature which leads to both smaller grain sizes and a higher fraction of grain boundaries having high angles of misorientation than when processing using ECAP [18]. 2 It has been well established that the grain size is the most important structural parameter affecting the mechanical properties of polycrystalline materials [19]. A small grain size leads to a significantly higher strength through the Hall-Petch relationship in the low-temperature regime [20,21]. Additionally, at higher temperatures, small grain sizes lead to faster strain rates and enhance the feasibility of superplastic forming (SPF) for industrial processes [2]. Superplasticity is an important subject due to forming materials in reasonably rapid production rates for large volume of components containing geometrical complexity. Production of UFG materials via SPD techniques can provide feasibility of SPF by significantly reducing the grain sizes especially for materials with difficulties in formability such as magnesium alloys. Due to the exceptional density of Mg alloys, the use of these materials is increasing in biomedical [22] and aerospace applications [23] as well as in the automotive industries [24]. However, insufficient formability of Mg alloys due to the limited availability of slip systems in the hexagonal close-packed crystal structure leads to poor ductility at room temperature. There are comprehensive studies [25,26] to improve the mechanical properties of these alloys so that they will meet the requirements for automotive, aerospace and biomedical applications. The present investigation was undertaken to address this deficiency by processing a ZK60 alloy by SPD techniques and examining the microstructural and textural evolutions as well as the microhardness studies. The potential for achieving maximum superplasticity is also evaluated. 2. Literature review 2.1. Severe Plastic Deformation (SPD) There are archeological evidences for the use of techniques using principles of SPD in China as early as about 500 B.C. [27]. Nevertheless, the introduction of scientific principles to SPD 3 processing was presented by Professor P.W. Bridgman in 1943 [10] by combining the compression and torsion for the processing of hollow cylinders to obtain uniform distribution of stress through the sample. Later, he received a Noble Prize for his works on the effect of high pressure on bulk metals. Based on definition, SPD is a processing technique in which a very high strain is applied to a sample through extensive hydrostatic pressure leading to an ultrafine grain structure [9]. The UFG materials have a homogeneous and equiaxed microstructure with an average grain size of less than 1 µm with a large fraction of high-angle grain boundaries [4]. The unique feature of SPD processing is that a very high strain is imposed at relatively low temperatures (usually less than 0.4 Tm) without any significant change in overall geometrical dimensions of the workpiece [9]. This can prevent the free flow of materials and thereby produce a significant hydrostatic pressure. The presence of hydrostatic pressure is essential for achieving high strains and subsequently high densities of lattice defects. These defects are necessary for grain refinement [9]. Samples processed by SPD generally have improved mechanical properties and a uniform ultrafine-grain structure within the whole volume of sample. Samples should not have any mechanical damage or cracks after processing by SPD [28]. Several SPD processes have been developed among which equal-channel angular pressing (ECAP) and high-pressure torsion (HPT) were used in the early studies and have attracted much attention. Other SPD techniques developed include accumulated roll bonding (ARB) [29], multiaxial forging [30] and twist extrusion (TE) [31]. Nevertheless, processing by HPT and ECAP remained as the most popular techniques and further development was recently made on this processing [9]. 4 2.2. Equal-Channel Angular Pressing (ECAP) Segal et al. first developed the ECAP process in the early 1980’s [32]. Processing by ECAP is at present the most developed SPD technique. This process introduces intense plastic strain into the material via pure shear without changing the dimensions of sample. In this process a shear strain is introduced when a billet passes through a point of intersection of two channels. The cross-section of the sample could be round or square. Since the dimensions of the samples remain identical, high strains are attainable by a repetitive pressing [9]. The principle of this process is illustrated schematically in Fig. 2.1. Figure 2.1. The principle of ECAP, showing the two angles Φ and Ψ [28]. The die contains two channels that intersect with an inner angle  and an outer angle Ψ. Strain value imposed on the billet during pressing for N passes can be calculated using the following equation [33]: 5                            2 2 cosec 2 2 cot 2 3 N N  (2.1) where εN is the strain imposed after multiple pressings and N is the number of ECAP passes. Several basic processing routes have been practically developed so far. Four principle slip systems associated with these processing routes are depicted schematically in Fig. 2.2 [34] where the X, Y and Z planes as shown in Fig. 2.3 [35] represent the planes perpendicular to the pressing direction, parallel to the side face of the sample and parallel to the top face of the sample, respectively. These routes introduce different slip system while processing and they lead to significant difference in the produced microstructure. In route A the sample is repetitively pressed without any rotation. However, in route BA the sample is rotated by 90º in alternative directions between consecutive passes. The sample is rotated by 90º in the same manner between each pass in route BC and it is rotated by 180º in route C [36]. Figure 2.2. (a) The four fundamental processing routes for ECAP [34] and (b) the slip systems viewed on X, Y and Z planes for consecutive passes using processing routes in part (a) [35]. 6 Figure 2.3. Illustration of the X, Y and Z planes denote the transverse plane, the flow plane and the longitudinal plane, respectively [35]. Figure 2.4 demonstrates the equivalent strain, , for a channel containing an outer angle, , over the range of 45 to 180  for a single pass where N = 1 [37]. The equivalent strain measured for a channel with an outer angle of  = 90 and an inner angle of Ψ = 20 is 1. Figure 2.4. Variation of the equivalent strain, , with the channel angle, , for a single pass where N = 1 [37]. 7 2.2.1. Microstructural evolution after processing by ECAP Processing by ECAP can introduce remarkable grain refinement in the coarse-grained materials. A microstructural model was developed by Xu et al. [38] based on the observation of the microstructural evolution of pure aluminum processed by ECAP as shown in Fig. 2.5. This model is based on the mechanical shearing of grains. It was assumed that the banded structure is formed parallel to the theoretical shear planes. As a consequence of multiple passes, with multiple bands superimposed, the structure evolves to the ultimate equiaxed grains and the fraction of boundaries having high-angle misorientations in each substructure based on theoretical models are denoted by d and , respectively [39]. The equiaxed grains are formed when the elongated arrays of grains are subsequently sheared in other directions. An important feature of this model is that the ultimate size of the equiaxed grains in equilibrium is dictated by the width of the subgrain bands produced in the initial pass. Figure 2.5. A model for grain refinement in the central region of the billet in ECAP [38]. 8 2.2.2. Grain refinement in magnesium alloys by ECAP Microstructural investigations show that dynamic recrystallization (DRX) takes place on magnesium alloys deformed in the temperature range of 420-600 K [ 40- 43]. In dynamic recrystallization, the grains nucleate preferentially along the original boundaries of the coarse grain structure and along twin boundaries in a necklace-like pattern so that the inner cores of the coarse grains may not be refined. Fundamental investigations [15,26,41] report that the formation of new grains along the pre-existing grain boundaries by DRX is due to anisotropy of the shear stresses needed to activate slip on different systems in the hcp metals. Only systems having easy slip will be activated in the cores of the grains and this will be insufficient to successfully produce grain refinement. In fact, higher stresses at the vicinity of grain boundaries are due to the incompatibilities in slip between the neighboring grains. This will activate non-basal slip systems which supports the formation of three-dimensional arrays of dislocation and consequently leads to dynamic recrystallization in these areas [41]. Therefore, the grain refinement of magnesium alloys is directly related to numbers of available nucleation sites for new grains. Figure 2.6 depicts a schematic illustration of a comprehensive model for structural evolution during the processing of magnesium alloys by ECAP [26]. There is a concept of a critical grain size, dc, for homogeneous nucleation which is sufficiently small and leads to homogeneous nucleation throughout the original grains when processing. In fact, dc, is associated with the grain size at which non-basal slip is activated throughout the grains. 2.3. High-Pressure Torsion (HPT) HPT method currently receives much attention as a SPD technique mainly because the efficiency of the method in UFG structures with predominantly high-angle grain boundaries, which 9 started the new age of nanoSPD materials with novel properties [1,44]. However, the fundamental concept of HPT was first introduced by Bridgman in the early 1940s. He pioneered application of high torsional shearing stress combined with high hydrostatic pressure to many different kinds of materials such as pure elements, metallic materials, glasses, geological materials (rocks and minerals), biological materials, polymers and many different kinds of organic and inorganic compounds [44]. Bridgman stated that “If a bar is twisted while longitudinal compressive is simultaneously applied, it is possible to the bar through much greater angles without fracture than is possible without the compressive load. At the same time the magnitude of the torque that the bar can support increases.” [10]. Modern definition of HPT refers to a processing technique in which the samples generally in the form of a thin disk, is subjected to straining under hydrostatic pressure [9]. This technique has attracted considerable attention only within the last decade because it is now recognized that it provides the ability to produce materials having exceptionally small grain sizes, generally smaller than those produced using ECAP [11,12]. 10 Figure 2.6. A model for grain refinement of magnesium alloys processed by ECAP [26]. Figure 2.7 depicted HPT facility made of two anvils when the sample in the form of a disk is located between them. The sample is subjected to a compressive applied pressure, P, of several GPa at room temperature or at an elevated temperature and it is simultaneously subjected to a torsional strain which is imposed by rotation of the lower anvil [9]. 11 Figure 2.7. Schematic illustration of quasi-constrained HPT processing [45]. In practice three distinct types of HPT processing have been used namely unconstrained and constrained [46] and quasi-constrained HPT [47]. In unconstrained HPT, the sample is placed on the lower anvil and it is subjected to an applied pressure and torsional straining as shown in Fig 2.8(a). Outward flow of material expected under the applied pressure and only minor back-pressure is applied into the system because of the frictional forces acting between the sample and the anvil. But in constrained HPT, the sample is machined so that it fits in to the cavity of the lower anvil to avoid this disadvantage. Therefore there is no outward flow of material while torsional straining Fig. 2.8(b). However, experiments are often conducted under quasi-constrained condition where there is limited outflow of materials between the upper and lower anvils. The simulations by finite element analysis show the distribution of effective strain inside the quasi-constrained volume of the anvils is comparable to the prediction by ideal torsion. The applied pressure and the friction coefficient outside the quasi-constrained volume play a minor role in the distribution of effective strain [47]. 12 Figure 2.8. Schematic illustration of HPT for (a) unconstrained, (b) constrained and [46] (c) quasi-constrained conditions [47]. Deformation of the disk occurs due to the shear strain caused by the surface frictional forces under a quasi-hydrostatic pressure. For an infinitely small rotation, dθ, and a displacement, dl, it follows from Fig 2.9 that dl = r dθ where r is the radius of the disk. The incremental shear strain, dγ, is calculated by the following equation [48]: h r h l   d d d   (2.3) where h is the thickness of the disk. By further assuming that the thickness of the disk is independent of the rotation angle, θ, formal integrations suggest that, since θ = 2 N, the shear strain, γ, is given by: h r N     2 (2.4) where N is the number of revolutions. Finally, in many investigations the equivalent von Mises strain is then calculated using the relationship [9]: h r N vM 3 2 3       (2.5) 13 Figure 2.9. Parameters used in estimating the total strain in HPT [48]. The use of Eq. (2.4) is correct for small imposed shear strains, but for large strains where 8 . 0   , the equivalent strain is given by [49]:            2 4 1 ln ) 3 2 ( 2    (2.6) which converts to equation 2.5 when γ→0. It is deducted from the equations that there is an inhomogeneity in the imposed strain across the sample which consequently leads to an inhomogeneous microstructure. This variation is investigated in many studies [16,18,50,51]. Although there is an inhomogeneity in microstructure after low number of turns, the microstructure evolves to a homogeneous microstructure by increasing numbers of turns in many materials. Figure 2.10 illustrates the variation of equivalent strain as a function of the number of revolutions for different distance from the center of the disk processed by HPT using equations 2.4, 2.6 [46]. The thickness of the disk for this figure is fixed at 0.1 mm. 14 It is apparent from Fig. 2.10 that a quasi-saturation is achieved after total of ~2 turns of processing by HPT and the variation of accumulated strain is not very large after about 2 revolutions. Another equation was developed for evaluating the true strain. It considers the changes in thickness of the disk as a result of the applied pressure. The true strain for this condition is given by [52] 𝜀 𝑡𝑟𝑢𝑒 = ln [ 1 + ( 𝜃𝑟 ℎ ) 2 ] 1 2 + 𝑙𝑛 ℎ 0 ℎ (2.7) where  = 2 N, h0 and h are the initial and final thicknesses, respectively. The true accumulated strain, , can be obtained from the following equation [52]: 𝜀 = ln( 𝜃 .𝑟 ℎ ) = ln ( 2𝜋𝑁 .𝑟 ℎ ) (2.8) Recent studies investigated the feasibility of HPT processing on bulk materials with thicknesses higher than 0.1 mm [53-55]. The potential for using HPT with small cylinders were studied on an Al-Mg-Sc sample with a thickness of ~8 mm. However, after processing by HPT variations in the Figure 2.10. Variation of equivalent strain as a function of number of turns in a disk with a thickness of 0.1 mm processed by HPT [46]. 15 Vickers microhardness were inevitable [54]. Therefore this variation in microhardness and microstructure results in a gradient in mechanical properties of processed materials. 2.3.1. Microstructural evolution after processing by HPT The microstructural studies during the HPT processing provide essential information regarding the grain refinement mechanism and assist in optimizing the superplastic behavior at elevated temperatures. An early model for grain refinement was developed in Armco iron using HPT [56]. Based on this model, the sample attains a gradual grain refinement in the microstructure with increasing strain. However it is not consistent with the observations where the widths of the elongated sub-grains are similar to the average equiaxed grain sizes after a single turn [38]. Another model is suggested for grain refinement in pure Al after processing by HPT. Edalati et al. developed this model based on the change in the hardness and microstructure evolution after straining the samples [57]. Dislocation behavior and the formation of high-angle grain boundaries as a function of equivalent strain are the fundamental principles of this model. Figure 2.11 shows the hardness variation with equivalent strain and corresponding microstructure evolution was divided into three regions. The hardness increases with strain in region (I) due to the accumulation of dislocations and the formation of subgrain boundaries. However, some dislocations are annihilated at subgrain boundaries leading to enhancement of the misorientation angles in the second region (II), in the meantime, the hardness is still increasing. In the third region (III), the hardness remains constant which is representative of a balance between an increase in dislocation and absorption of dislocation at grain boundaries. 16 2.3.2. Grain refinement in magnesium alloys by HPT In a recent investigation, Edalati et al. [58] examined the hardness behavior of pure Mg- (99.9%) with imposed strain similar to pure Al (99.99%), having a maximum hardness followed by a steady state. An important advantage in the processing of hcp metals is that HPT provides an opportunity to avoid some of the difficulties associated with processing by ECAP. In practice, in order to avoid the problems of cracking at lower temperatures the ECAP processing of magnesium alloys is usually carried out at temperatures of ~473 K or higher. However, the imposition of a high hydrostatic pressure in HPT provides an opportunity for processing these same alloys at much lower temperatures. An AZ31 magnesium alloy was successfully processed by HPT at room Figure 2.11. Schematic illustration of microstructural evolution with straining in pure Al. Thin double lines depicted in region II represent low angle boundaries with some extension of boundary width and thick lines in region III represent high angle boundaries [57]. 17 temperature to give a grain size of ~900 nm [14,59]. The variations in microstructures obtained in the AZ61 alloy after HPT at 423K are shown in Fig. 2.12 after processing through 1, 3, 5 and 7 turns [18]. The corresponding TEM micrographs for the same samples are shown in Fig. 2.13 together with the associated selected area electron diffraction (SAED) patterns. The microstructure evolves more rapidly at the edge of the disk so that fewer turns are required to achieve significant grain refinement. At higher numbers of turns a similar evolution occurs in the center of the disks towards an ultrafine-grained microstructure that is reasonably similar to the microstructure at the peripheral regions. It is also apparent that the average grain size was relatively large after a single turn but decreased with increasing the numbers of turns. TEM observations show that the grains are better defined after larger number of turns. The visible rings in the SAED patterns confirms existence of grains separated by boundaries having high angles of misorientation [18]. Figure 2.12. Optical micrographs showing the microstructures after HPT after processing through 1, 3, 5 and 7 turns at a temperature of 423 K [18]. 18 Figure 2.13. Equivalent TEM micrographs in bright-field and dark-field, plus the associated SAED patterns, after HPT at 423K through 1, 3, 5 and 7 turns [18]. Many experiments have revealed a gradual evolution with increasing strain towards a homogeneous distribution of the measured hardness values. The significance and variations in the evolution towards homogeneity in HPT are examined in the following sections. 2.3.3. Hardness homogeneity after processing by HPT The degree of homogeneity in the microstructure within the disk can be anticipated by the hardness measurements [2]. Microhardness measurements combined with microstructural observation are two methods that are used to evaluate homogeneity across the disks processed by HPT. The experimental results obtained from austenitic steel [60], Cu [61] and high-purity Ni [62] show significant variations in the microhardness values across the diameters of disks with lower hardness values in the centers of the disks and higher values in the peripheral regions However, the results on commercially purity fcc materials such as Al [46] and Cu [63], other fcc aluminum alloys such as Al-Mg-Sc [64], and hcp materials including Magnesium alloys of ZK60 [17,65], AZ31 [50], AZ61 [18] and Mg-9Al [51] show a reasonable homogeneity in microstructure 19 across the disks after several numbers of turns when torsional straining is sufficiently high under a high imposed pressure. Experimental results obtained for AZ31 show a gradual evolution toward hardness homogeneity by increasing numbers of revolutions. The microhardness values in this material become saturated at a constant value after 5 turns at low temperatures [50]. The variations in hardness across HPT disks can be readily correlated by plotting the hardness values against the equivalent strain as shown in Fig. 2.14. These data are obtained at three separate temperatures 296, 373 and 473 K. Observations shows the points at 296 and 373 K fall along a curve with a gradual transition to a horizontal plateau which is typical of many materials. They correspond to HPT processing in the absence of any significant recovery whereas at 473 K they exhibit different shapes. The occurrence of this early peak in plots of hardness against equivalent strain shown in Figure 2.14 is well documented in case where the material exhibits recovery during the HPT processing [57,58]. The peak is associated with the occurrence of grain growth happening at a higher processing temperature of 473 K. Table 2.1 shows information regarding grain structure and microhardness after processing by HPT. 20 Figure 2.14. Values of the Vickers microhardness plotted as a function of the equivalent strain in disks processed by HPT at temperatures of (a) 296 K, (b) 373 K, and (c) 473 K [50]. Table 2.1. Grain structures produced in pure magnesium and magnesium alloys using HPT and microhardness values Alloy Initial grain Size (µm) final grain Size (µm) Structure after number of turns HPT Temperature (K) Applied Pressure (GPa) Initial Hardness (Hv) Final Hardness (Hv) References AZ31 10 0.5 5 296 6 55 105 [50] AZ31 10 0.8 373 6 55 105 AZ31 10 2.3 473 6 55 70 AZ61 22 0.22 7 423 3 62 - [18] AZ61 22 0.11 7 273 3 62 118 Mg-9Al 30 <0.4 5 423 3 102 111 [51] Mg-9Al 30 0.5 5 298 3 102 119 ZK60A 20±10 .<1.0 5 298 6 72 110 [63] Mg(99.9) 1600 <1.0 10 298 6 29 40 [58] 2.4. Superplasticity after processing by SPD Superplasticity is defined as a mechanical property of polycrystalline materials that delineates the ability to pull out to a very high elongation in tension without any significant necking. There are two fundamental requirements for achieving superplasticity in conventional polycrystalline materials [66,67]. First, the materials should contain grains sizes typically less than 10 µm. 21 Second, it requires temperatures at or above  0.5Tm where Tm is the absolute melting temperature. Therefore the superplasticity is anticipated to be achieved to a very small grain sizes. SPD techniques may lead to considerable grain refinement to submicrometer or even nano sizes. However, in evaluating the characteristics of UFG materials processed by SPD, it is very important to recognize that these ultrafine grains are unstable at elevated temperatures. Thus the UFG structures are lost through grain growth when the samples are heated and no advantageous properties are achieved in high temperature deformation. Nevertheless it is also possible to remain arrays of ultrafine grains even at very high temperatures [9]. Table 2.2. Reports of superplasticity in ultrafine-grained materials produced by ECAP Alloy or Composition (wt%) ECAP Grain Size (µm) Superplasticity Reference Number of passes Additional info Temperature (K) Testing Temperature (K) Strain Rate (s -1 ) Maximum elongation (%) Mg-0.6Zr 4 Cast+Ext-ECAP 573 1.4 573 3.3 × 1.0 -4 480 [68] Mg-9Al 2 473 0.7 373 1.0 × 1.0 -4 150 [69] 398 260 423 800 473 3.3 × 1.0 -1 90 3.3 × 1.0 -2 220 1.0 × 1.0 -2 270 3.3 × 1.0 -3 460 AZ31 6 P B = 218 MPa 423 ~1-3 623 1.0 × 1.0 -4 470 [70] 453 740 494 970 523 1210 ZK60 6 B C 473 0.8 400 1.0 × 1.0 -4 280 [71] 450 1140 473 1.0 × 1.0 -3 930 ZK60 473 0.8 473 3.3 × 1.0 -4 1000 [72] 2.0 × 1.0 -4 1310 1.0 × 1.0 -4 930 3.3 × 1.0 -5 900 1 473 0.8 1.0 × 1.0 -4 1500 [73] 2 1.0 × 1.0 -4 3050 3 1.0 × 1.0 -4 2130 4 1.0 × 1.0 -4 2100 In the high-temperature regime, diffusion is an important rate controlling factor in the flow rate of material. Also, the flow rate increases with decreasing grain size. Based on definition, superplasticity refers specifically to the ability of a material to pull out to a tensile elongation of at least 400% with a measured strain rate sensitivity of ~0.5 [66] and small grain sizes of <10 µm 22 [67]. Thus, reasonably stable UFG produced by SPD processing, have an opportunity for attaining exceptional superplastic elongations at elevated temperatures. Furthermore, high elongations may occur at unusually rapid strain rates that may even be within the regime of high strain rate superplasticity [2]. 2.4.1 The implication of experimental results for superplasticity In order to compare the nature of superplasticity in materials processed by ECAP and HPT with other processing techniques, two constructive useful diagrams are recently developed by Kawasaki and Langdon [74] as a consequence of the many reports of superplasticity in Al-based alloys. They are helpful to delineate the ranges of strain rates and grain sizes associated with the occurrence of superplasticity for a series of Al alloys fabricated through different approaches [75]. In Figs. 2.15(a) and 2.15(b) solid circles and ovals for various Al alloys processed with different procedures are representative of typical elongations to failure and the spatial grain sizes plotted against the strain rates for testing. The dashed ovals filled in blue and pink superimposed on each diagram are the appropriate ranges for UFG Al alloys processed by ECAP [76] and HPT, respectively. From the Figure 2.15 it is implied that the ECAP and HPT processing techniques expand the plastic forming rate of ingot materials to faster strain rates and the developed ranges generally overlap with the ranges associated with powder metallurgy (PM) materials which is the important advantage in using SPD techniques, because in general metals having ultrafine grain sizes without any contamination or porosity can be formed at higher strain rates. Thus, processing of bulk materials through the application of SPD techniques explain an important alternative approach for 23 future applications without utilizing the bottom up techniques such as condensation and deposition for fabricating nanostructured materials [77]. (a) (b) Figure 2.15 Plots of strain rate versus (a) elongation to failure and (b) spatial grain size for a series of Al alloys produced using different processing methods superimposed on the plot, and delineated by the dashed ovals, are data collected for aluminum alloys after ECAP and HPT [74]. 2.5. Combination of SPD processes To achieve UFG materials various SPD processing techniques have been applied to metals and alloys so far. However, recently the combination of various SPD processes has been investigated on different metals and alloys in order to achieve finer grain sizes and better 24 properties. The combination of accumulative roll bonding (ARB) and friction stir processing (FSP) [78], ECAP followed by HPT [79-83], ECAP followed by cold rolling (CR), and a combination of ECAP, CR, and HPT [81] are examples of the combination of SPD processes that have been studied recently. There are reports of various combinations of ECAP and HPT in studies of Cu [83,84], Ni [85], Ti [79] and Al [86] alloys. A comparison between different combinations of SPD processes on pure Ni is shown in Fig. 2.16 [81]. This figure illustrates further grain refinement after each additional SPD processing and an increase in the dislocation density is observed after the combination of ECAP and HPT [81]. Recent study on commercial Al-7075 alloy [86] processed by SPD procedures of ECAP followed by HPT showed that the combination of ECAP and HPT have smaller grain sizes and higher saturation harnesses than samples processed separately by ECAP or HPT. Figure 2.17 shows the microstructural observations during the process stages. It can be seen a significant grain Figure 2.16. Crystallize size distributions for nickel samples obtained by combinations of different methods: (a) ECAP, (b) ECAP + CR, (c) ECAP + HPT, and (d) ECAP + CR + HPT [81]. 25 refinement is achieved after each SPD technique, and the minimum grain size was 200 nm after processing by a combination of ECAP for 8 passes and HPT for 20 turns. Therefore it can be concluded that the saturation hardness is dependent upon the microstructure of the sample introduced in any processing step prior to the HPT processing. Figure 2.17. EBSD images of the Al-7075 alloy in the (a ) annealed condition and after processing by (b) HPT through a total of 5 turns, (c) ECAP for 4 passes, (d) ECAP for 8 passes, and of ECAP for 4 passes and (e) HPT for 20 turns [86]. 26 2.6. Paradox of strength and ductility in ultrafine-grained materials It is generally accepted that plastic deformation processes through conventional forming methods such as rolling, drawing or extrusion can enhance the strength of metals through the Hall-Petch relationship [20,21]. However these results are accompanied by a loss of ductility. Fig. 2.18 demonstrates that with increasing plastic deformation, the yield strength of Cu and Al monotonically increase while the ductilities decrease. The same trend is also true for other metals and alloys. However, literature [87] reports an extraordinary combination of high strength and high ductility produced in metals subject to severe plastic deformation (SPD). It is believed that this unusual mechanical behavior is caused by the unique nanostructures generated by SPD processing. The combination of ultrafine grain size and high-density dislocations appears to enable deformation by new mechanisms. Figure 2.18. Cold rolling of Cu and Al increases their yield strength but decrease their ductility. The extraordinary combination of both high strength and high ductility in nanostructured Cu and Ti processed by SPD clearly sets them apart from coarse-grained metals [87]. 27 Recently modified diagrams were developed which give a simple display of normalized stresses against normalized strains [88] and can provide more quantitative evaluation of the potential for achieving high strength and high ductility (HSHD) alloys than conventional diagram introduced by Valiev et al. [87,89]. The use of these diagrams is examined for Al-7%Si samples processed by HPT at temperatures of 298 and 445 K and shown in Fig. 2.19. From the plots at different temperatures, it can be seen that most of the data falling above the broken horizontal line and to the right of the broken vertical line display a simultaneous increase in strength and ductility compared to the as-cast material and this is designated the HSHD region whereas conventional behavior denotes an increase in ductility at the expense of strength or an increase in strength at the expense of ductility. Figure 2.20 is a superposition of the each (a) (b) (c) (d) (e) (f) Figure 2.19 Variation of (a) and (d) the normalized YS, (b) and (e) the normalized UTS with the normalized elongation to failure and (c) and (f) the variation of the normalized UTS with the normalized uniform elongation corresponding to the strain at UTS for samples processed by HPT at 298 and 445 K, respectively [88]. 28 datum obtained after processing Al-7%Si samples through up to 10 turns of HPT at 298 and 445 K on the classic report on the strength-ductility paradox illustrated in Fig 2.19 where the shaded region below the curve delineates the conventional behavior of high strength and low ductility. Therefore an HSHD materials is achieved in the Al-7%Si alloy at the grain sizes less than 0.5 µm after up to 10 HPT turns. On the other hand the strain rate sensitivity obtained for these conditions were 0.14 [88]. This value is significantly lower than m-values obtained for conventional GBS at temperatures above 0.5Tm [74,90]. Prediction in three-dimensional molecular dynamic simulations that were designed to model the behavior of nanocrystalline solids [91,92] together with experimental results such as topological observations using AFM [90] and Figure 2.20. The diagram showing the strength-ductility paradox as developed by Valiev where the shaded area represents the conventional behavior, the region of high strength and high ductility lies to the right of the solid and broken line and the numbers marked with percentages denote the strains imposed during rolling: the present results for the Al-7%Si alloy are superimposed on the plot and the displacement to the right with increasing number of HPT turns, N, is indicated by the arrow [88]. 29 plastic behavior in micro-pillars [93] of UFG alloys suggested the possibility of contribution of GBS at low temperatures and its effect in flow behavior mechanism. In order to increase ductility at low temperatures after SPD, It was found that controlled annealing process such as low temperature annealing [94] and short term annealing could lead to enhancement of strength and ductility [6]. 2.6.1 Low temperature ductility after HPT Generally, UFG metals exhibit high strength but poor ductility due to having both a low rate of strain hardening and a low strain rate sensitivity [89,95]. This problem is highlighted when facing hcp materials such as magnesium alloys with limited ductility at room temperature [13]. To enhance the ductility after SPD processing, several techniques have been proposed. Subjecting the processed material to a short-term annealing is one of the methods recently investigated on Al- 1%Mg alloy [6]. Studies on Cu processed by ECAP revealed that short term annealing has the potential of ordering the defect structures within the grain boundaries and thereby producing more equilibrated grain boundaries without facing major grain growth [96]. Other studies reported the ductility enhancement after processing Ti alloys by SPD processing followed by short term annealing [97-99]. It is notable in Fig. 2.20 that specimens processed by 10 turns without post- HPT annealing exhibit both the highest stresses and very small elongations before failure at all strain rates. However, samples experiencing post-HPT annealing at 473 and 523 K exhibit higher elongation but at significantly lower stress levels. The advantage of the short post-HPT annealing is clearly demonstrated in Fig. 2.21(b) and 2.21(c) at the faster strain rates where the samples processed by HPT exhibited brittle behavior and fractured before yielding but a post-HPT annealing treatment for only 10 min decreased the strength by > 25% however it provides reasonable elongations to failure with engineering strains of more than 0.2. Inspection of all tensile 30 stress - strain curves in Fig. 2.21 shows that the post-HPT short annealing leads generally to a decreasing strength and increasing ductility with increasing annealing temperature. Figure 2.21. Stress–strain curves at room temperature for the annealed material, after HPT through10 turns and after HPT through 10 turns and annealing at 423,473 and 523K at (a) 1.0 × 10 -4 , (b) 1.0 × 10 -3 and (c) 1.0 × 10 -2 s -1 respectively [6]. 31 3. Experimental Materials and Procedure 3.1. Experimental materials In these sets of experiments, a ZK60 Mg alloy with chemical composition of Mg-5.5 wt.% Zn-0.5 wt.% Zr was used in a rod shape with diameter of 10 mm. The rods were received after extrusion. Samples were sliced perpendicularly from the rods into disks with thicknesses of ~1.2 mm. By using abrasive papers, the disks were carefully polished on both sides and a series of samples was produced with ultimate thicknesses of ~0.82 mm. 3.2. Methods of severe plastic deformation The material was processed using severe plastic deformation techniques: ECAP, HPT, and a combination of ECAP and HPT. 3.2.1. Processing by ECAP An ECAP facility was composed of a die having two intersecting channels, a plunger and a hydraulic press. The two channels have an internal angle of  = 90 o and an outer angle at the intersection of the two channels of  = 20 o . The die and the plunger were made of a tool steel and were heat treated to obtain a Rockwell hardness of ~55. The tolerance of the plunger was kept extremely low to prevent materials from flowing in between the walls of the channel and the plunger. The drawings of the ECAP die and plunger are shown in Fig 3.1. 32 Figure 3.1. Drawings of the die and plunger of ECAP. 33 The ECAP processing was conducted at a temperature of 473 K. This geometry and these angles lead to an imposed strain of ~1 on each separate passage through the die [37]. The billets were processed by ECAP for 4 and 8 passes which correspond to the total imposed strain up to ~6.4. All samples were processed using route BC. 3.2.2. Processing by HPT The HPT processing was conducted under quasi-constrained conditions [47,100] using an HPT facility with a rotating lower anvil made from high strength YXR3 tool steel. Both the upper and lower anvils contained central spherical depressions having nitrided surfaces with depths of 0.25 mm and diameters of 10 mm. By placing the disk in the lower depression, the anvils were then brought together and an applied pressure was imposed on the disk. All processing was conducted at room temperature. A pressure, P, of 2.0 GPa was applied during the process. By rotating the lower anvil at a constant speed of 1 rpm, the torsional strain was applied through total numbers, N, of 0, 1/8, 1/4, 1/2, 1, 2 and 5 revolutions. Processing for 0 turns means there is no torsional straining and the sample only experiences the applied pressure for a period of one minute. No damage or cracking was observed after processing by HPT at different revolutions at 2.0 GPa. To ensure that there was no slippage throughout the straining processing, a line was carefully marked along the diameter of the disks with an identical duplicate line on the other surface. The disks were placed in the depression in the HPT anvil afterwards and torsionally strained through the selected numbers of rotations. This method for monitoring the slippage in HPT samples is well established earlier [101] and careful measurements after processing showed no significant slippage of the samples during processing. The applied pressure was achieved by an imposed force of 157 kN and was determined using the following calculation: 34 𝐹 = (𝜋 × ( 𝑑 ℎ 2 ) 2 ) × 𝑃 ℎ (3.1) where F is the applied force, Ph is the hydraulic pressure, dh is the diameter of the hydraulic cylinder (250 mm in this facility). Therefore, this force can be transferred to the pressure applied on the 10 mm-diameter disk, Pd using: 𝑃 𝑑 = 𝐹 𝜋 ×5 2 (3.2) 3.2.3. ECAP followed by HPT processing The billets were processed by ECAP through 4 passes at 473 K and sliced in the direction perpendicular to the longitudinal axis using a diamond blade sawing with a thickness of 1.5 mm and grinded to final thickness of ~0.84 mm. Disks are then subjected to HPT for 1 to 10 turns under an applied pressure of 2.0 GPa at a speed of 1 rpm. The schematic of the combination of ECAP and HPT is illustrated in Fig. 3.2. Figure 3.2. Schematic of the combination of ECAP and HPT. Pressed and sliced sample 35 3.3 Microstructural analysis Microstructural observations were investigated using optical microscopy, transmission electron microscopy (TEM), scanning electron microscopy (SEM) and electron back-scatter diffraction (EBSD) techniques. 3.3.1. Optical microscopy In order to prepare the samples for optical microscopy, the samples were mounted in cold resin and grinded with abrasive papers. A thin layer of ~0.1 mm thickness was removed from the disk surfaces by grinding. Then, the samples were polished mechanically on a cloth with a final polishing using an OP-S solution (a Struers colloidal silica solution with an average particle size of 0.04 µm and pH of 9) to avoid pitting on the surface. An acetic-picral solution with a chemical composition of 4.2 g picric acid, 70 mL ethanol, 20 mL distilled water and 10 mL acetic acid was used to etch the surface and reveal the grain boundaries. A Keyence digital optical microscope was used for the microstructural observations. The central, half-radius and peripheral regions along random diameters of the disks were examined for microstructural observations and quantitative evaluations. The grain sizes were measured using the linear intercept method with a count of at least 500 grains. 3.3.2. Transmission electron microscopy (TEM) Transmission electron microscopy (TEM) was also utilized to investigate the microstructural characteristics after processing the samples by 5 turns of HPT. A TEM thin foil was sectioned from the half-radius area of the disk and specimens were processed by focused ion beam (FIB) for examination with an operating voltage of 20 kV using gallium ion to a final thickness of ~100 nm. TEM microscopy was performed using a Philips CM12 microscope operating at 200 kV working with a field emission gun. 36 3.3.3. Scanning electron microscopy (SEM) To observe the grains and grain boundary configurations of the HPT samples, an analytical field emission scanning electron microscope (SEM) JEOL JSM-7001F was used at an operating voltage of 15 kV. 3.3.4. Electron-backscatter diffraction (EBSD) Electron back-scatter diffraction (EBSD) and orientation imaging microscopy (OIM) were utilized to record microstructural characteristics of the processed samples in the half-radius of the cross-sections of the disks. To carry out the observation, an operating voltage of 7 kV was used in the same SEM (JEOL JSM-7001F). These observations were made on the near surface area of the disks. A JEOL IB09010CP ion beam cross-sectional polishing machine was used to polish the samples at an operating voltage of 6 kV for 5 hours. Low-angle grain boundaries (LAGBs) were set in the software as boundaries having misorientation differences between adjacent measuring points of 2 o – 15 o and high-angle grain boundaries (HAGBs) were defined as having misorientation differences of more than 15 o . Figure 3.3 shows the position at which the disks were ion polished by the cross-sectional polishing machine. The directions of ND, TD and RD correspond to Normal (torsion axis), Tangential and Radial Direction, respectively. The shear direction is parallel to TD. The pole figures and inverse pole figures in EBSD images were measured from the semi-oval area of sample in TD-RD plane shown in Fig. 3.3. 37 3.4 Mechanical experiments 3.4.1 Microhardness measurements Two different methods were utilized for microhardness measurements conducted on samples processed by HPT after various numbers of revolutions. In the first method the microhardness was measured along randomly selected diameters of the upper surfaces of the disks. Second, the microhardness tests were performed over the surface areas of the disks to obtain detailed information on the hardness evolution during the HPT processing. All samples were polished to a depth of ~0.1 mm prior to taking the measurements to achieve mirror-like surfaces. The Vickers microhardness values, Hv, were obtained at room temperature using an FM-1e microhardness instrument equipped with a Vickers indenter. The measurements were recorded under a load of 100 gf with a dwell time of 10 s for each separate measurement. The diagonal hardness values were recorded as an average of four separate positions configured in a cross-shape around the selected point at distances of 0.15 mm. Also, the individual selected points were separated from each other by incremental distances of 0.3 mm. These points are denoted by the ND RD TD ND RD TD 3 mm 1.5 mm Figure 3.3. Near surface position of the disks for EBSD sample preparation with ion beam cross-sectional polishing machine. ND, TD and RD correspond to Normal Direction (torsion axis), Tangential Direction and Radial Direction, respectively. The shear direction is parallel to TD. 38 small open circles along the bottom edge. For the second method, a rectilinear grid pattern was used with a spacing between each point of 0.3 mm in order to provide a visual representation of hardness values over the entire surface. Figure 3.4 depicted a schematic illustration of microhardness indentations configuration for (a) linear diagonal (b) rectilinear grid pattern methods on HPT disk. All of these individual values of Hv were then used to construct color-coded contour maps that provided clear visual presentations of the distributions in hardness across the surface of each disk [102]. (a) (b) | 0.3 | |0.15| Edge Center (0,0) Figure 3.4. Schematic illustration of microhardness indentations configuration for (a) linear diagonal (b) rectilinear grid pattern methods on HPT disk. 39 3.4.2. Tensile measurements The mechanical properties were evaluated after HPT processing by using tensile testing of miniature samples and shear punch test (SPT) on the processed disks. First, two miniature tensile samples were cut from each of the disks after processing by HPT. To avoid any potential problems associated with microstructural inhomogeneities near the disk centers, the samples were cut from off-center positions on the HPT disks shown in Fig. 3.5. An electrical discharge machining (EDM) was employed to cut the miniature tensile specimens from the disks to obtain dimensionally precise samples and to avoid mechanical damages. The gauge lengths of the tensile samples were 1 mm with cross-sectional areas of approximately 1 × 0.64 mm 2 . All of the tensile testings were carried out at temperatures of 473, 523 and 573 K using an Instron testing machine operating at a constant rate of crosshead displacement and initial strain rates ranging from 1.0  10 -1 to 3.0  10 -5 s -1 . Specimens were tested to failure and the recorded load-displacement curves were converted to engineering stress versus engineering strain. The flow stresses and the elongations to failure were determined directly from the stress-strain curves. Figure 3.5. Dimensions of tensile specimens after HPT processing [103]. 40 3.4.3. Shear Punch tests The shear punch test was carried out on as-received and HPT-processed samples. The as- received samples were sliced perpendicularly from the extruded rods into disks with thicknesses of ~1 mm. To avoid probable effect of thickness on obtained data, all disks were polished on both sides in a parallel manner to the thicknesses of ~0.65 mm which is identical to the thickness of processed samples. A shear punch fixture with a 3.175 mm diameter flat cylindrical punch and a 3.225 mm diameter receiving-hole was utilized for this experiment. The schematic view of the shear punch die and assembly are shown in Fig. 3.6. (a) (b) All shear punch tests were performed in the temperature of 473, 523 and 573 K using a screw driven MTS material testing system equipped with a three-zone split furnace. After locating the specimen in the fixture, the assembly of the specimen and SPT fixture were accommodated by the split furnace between the compression platens of the MTS testing system as shown in Fig. 3.7. Subsequently, the assembly was heated to the test temperature and held for 20 min to establish thermal equilibrium in the testing arrangement before the specimen was deformed by the punch. Tests were carried out with a load cell of 20 kN capacity and at cross-head speed ranging from 1.0 Figure 3.6. Schematic representation of (a) shear punch-die assembly [104] and (b) shear zone of sample inside the dies [105]. 41 to 1.0 × 10 -2 mm/min. The shear stress is calculated from τ = P/πdt, where P, is applied load, t is the specimen thickness and d is the average of the punch and die hole diameters. The τ was measured automatically as a function of punch displacement [106]. The data were acquired by a computer and were converted to the shear stress. The SPT curves were plotted as shear stress against normalized punch displacement. Three different samples were tested for each condition. The observations and measurements showed variation in the curves and measured ultimate shear strength values were negligible. 3.5. Annealing processes The annealing process was conducted on ZK60 alloy after processing through 5 turns of HPT. The study was investigated by the isothermal annealing. The EBSD analysis and tensile testing studies on the obtained samples. Isothermal annealing was performed for 10 and 20 min at 3 temperatures in the range 473-548 K, All annealing tests were conducted in an air circulated furnace with an accuracy of 1 K. Figure 3.7. The assembly of the specimen and SPT fixture between the compression platens. 42 4. Experimental results 4.1. Microstructural analysis 4.1.1. Microstructural characteristics before and after HPT Figure 4.1 exhibits a large grain structure distributed in the extruded material at a cross sectional area of the as-received rods. An initial inspection revealed a fairly equiaxed and uniform grain structure in the extruded material with an initial average grain size of ~9.4 µm. Figure 4.1. Optical microstructures of As-received extruded ZK60. Microstructural observation of the samples processed by HPT revealed the presence of many highly-refined grains. A comprehensive montage of optical micrographs after processing through 1/4, 1/2, 1 and 5 revolutions is presented in Fig. 4.2. The microstructures are shown with the three columns representing the center, half-radius and periphery of each disk. A gradual refinement of microstructure from the center to the periphery of the disks are noticeable in the early stages of processing. It is apparent that a minimum number of 1 turn is required to achieve a relatively uniform grain size. After 5 turns there is a highly-refined grain structure and the microstructure appears to be reasonably uniform throughout the disk. 100 µm 43 The measured grain sizes by optical microscopy are summarized in Table 4.1 for the disks processed up to 5 turns at the three different sampling matching to the positions shown in the micrographs of Fig. 4.2. The error bars correspond to the 95% confidence limits. In practice, there were some very small grains which were beyond the precision of the optical microscope and these were not included within the error bars. Further analysis of grain size measurements are provided on the next session by the EBSD analysis. The data in Table 4.1 show that the average grain sizes are initially relatively large in the early stages of processing but both the average grain sizes and the associated error bars decrease gradually as the microstructure evolves with additional straining. After 5 turns, the grain size is in the order of ~1 m at all three positions. 44 Figure. 4.2. Optical micrographs of the microstructures after HPT at the center, half-radius and periphery positions of disks after processing through 1/4, 1/2, 1 and 5 turns at room temperature. N = 1/4 N = 1/2 N = 1 N = 5 Center Half-radius Periphery 45 Table 4.1. Results for the average grain sizes and the related uncertainty for disks processed by various numbers of turns in HPT at ambient temperature: results are shown for the center, half-radius and edge positions. N Average grain size (µm) Center Half-radius Periphery 1/4 4.1 ± 2.5 2.4 ± 1.3 2.0 ± 1.2 1/2 2.3 ± 1.3 2.0 ± 1.1 1.8 ± 1.0 1 1.4 ± 1.0 1.2 ± 0.9 1.1± 0.8 5 1.2 ± 0.8 1.0 ± 0.6 0.9 ± 0.6 To provide more detailed information on the evolution of microstructure, the micrograph in Fig. 4.3 shows the periphery of the disk for the sample taken through 1/4 turn at a low magnification with the edge of the disk visible at the lower left and with a higher magnification shown in the inset at the upper right. Inspection suggests that dynamic recrystallization (DRX) occurred within the shear bands while imposing the torsional straining and this is indicated by arrows within the micrographs in Fig. 4.3 where small recrystallized grains can be detected. Several reports also observed the occurrence of dynamic recrystallization around the shear bands as well as within the dynamic recrystallized grains inside the shear bands while imposing large amounts of strains on magnesium alloys [107-109]. The grain boundary configurations are better defined by the SEM micrograph in Fig. 4.4 near the center of the disk after 1/4 turn where faceted grains are clearly separated by high-angle grain boundaries and twinning is visible within the larger grains as shown in the inset. Recent observations [110] on the texture and microstructure of the AZ31 alloy processed by HPT show twinning accompanied by basal <a> slip exerts a significant contribution to the deformation at low temperatures. 46 Figure 4.3. Optical micrograph after HPT processing for N = 1/4 turn at the periphery of the disk. Figure 4.4. SEM micrographs after HPT processing for N = 1/4 turn near the center of the disk. 47 (a) (b) (c) 500 nm 5 1/nm 500 nm 100 nm Figure 4.5. TEM images of ZK60 alloy processed by HPT after 5 revolutions: (a) bright field image, (b) conventional dark field image at the same position and (c) weak beam dark field at higher magnification for the grain marked in (b). 48 TEM images of a sample processed by HPT after 5 turns provide more details about microstructure and grain configuration of microstructure after final stage of HPT. The TEM image of the sample illustrated in Fig. 4.5(a) shows a bright field image and an electron diffraction pattern of the processed sample with a mean grain size of ~300 nm. It reveals an equiaxed ultrafine microstructure. The electron diffraction pattern shows a high-angle misorientation of neighboring grains. Figure 4.5(b) shows a dark field image at the same position depicted in Fig. 4.5(a). A considerable increase in the dislocation density inside the grain is observed comparing to the extruded Mg alloys [51]. For better observation of dislocations inside the grains, a weak beam dark field (WBDF) image was taken from a randomly selected and marked area in Fig. 4.5(b) and this is shown in Fig. 4.5(c). A high density of dislocations is observed. Figure 4.6 provides a montage of image quality (IQ) micrographs of the unprocessed extruded material prior to HPT followed by 0, 1/2, 1, and 5 turns of HPT in the mid-radius of the disks. Figure 4.7 demonstrates the corresponding grain size distribution of micrographs in Fig. 4.6. Although the grain size reported for the studied area of the extruded sample is about 5 μm, there are some relatively larger grains observed in the extruded materials as shown in previous paper [17]. Observations on microstructural changes show that there is a gradual grain size change from extruded condition to ultrafine grained structure obtained after 5 turns of HPT. After processing by 1/2 and 1 turns of HPT, Fig. 4.6(c) and (d) show areas with aggregates of small grain sizes but areas with coarse grains are still visible. After processing by 5 turns of HPT the majority of the grains are refined but some coarser grains remain. 49 Figure 4.6. Image quality (IQ) maps of ZK60 (a) as-received and after processing by (b) 0 turn, (c) 1/2 turn, (d) 1 turn and (e) 5 turns of HPT in the mid-radius of the disks. 10 µm (a) As-received 10 µm (b) 0 turn 5 µm (c) 1/2 turn 5 µm (d) 1 turn 5 µm (e) 5 turns 50 (a) (b) (c) (d) Figure 4.7. Grain size distribution of ZK60 in the (a) as-received extruded condition and after (b) 0 turn, (c) 1/2 turn, (d) 1 turn and (e) 5 turns of HPT in the mid-radius of the disks. (e) d = 5.0 2.5 µm d = 2.9 1.7 µm d = 1.9 1.2 µm d = 1.7 0.9 µm d = 0.7 0.5 µm 51 In order to facilitate the understanding of the structure refinement, the distribution of grain sizes are plotted for the different conditions in Fig. 4.7. It is observed that all the grains in the as received material are larger than 1 µm and a peak in the distribution is observed at ~5 µm. Processing by 1/2 and by 1 turn lead to the formation of fine grains and the peak of grain size distribution is reduced to 1~2 µm. Processing by 5 turns leads to an increase in the area fraction of fine grains and move the peak in the distribution to ~1 µm. It is worth noting that despite the differences in the location of the peak, the grain size distributions are spread over almost one order of magnitude of grain sizes in all conditions. This shows that the grain size distribution is highly heterogeneous in all conditions, including the material processed by 5 turns of HPT. Table 4.2 provides grain size values obtained from OIM analysis. The formation of bimodality of grain distribution was reported in magnesium alloys after HPT under a pressure of 2.0 to 6.0 GPa at room and elevated temperature [50,110-112]. At elevated temperatures the necklace structure is mostly responsible for formation of bimodal structure. However there are few evidences of necklace formation in the material processed at room temperature in this study and previous observations [16,50,59]. HPT turns Grain size (µm) Area fraction percentile (%) < 1.0 (µm) < 0.5 (µm) Extruded 5.0 ± 2.5 0 0 0 2.9 ± 1.7 10 0 1/2 1.9 ± 1.2 15 3 1 1.7 ± 0.9 25 3 5 0.7 ± 0.5 75 35 Table 4.2. Average grain size and Area fraction percentile of grains smaller than 1.0 and 0.5 µm measured for studied materials 52 The number fractions of misorientation angles are shown in Fig. 4.8 for as received extruded condition and after processing by HPT through 0, 1/2, 1 and 5 revolutions. All the plots in Fig. 4.8 correspond to the grain-to-grain misorientation distributions [113,114] rather than pixel to pixel misorientation. The solid curves are representative of the distributions of misorientation angles developed by Del Valle et al. [115-117] for an aggregate of randomly oriented hexagonal crystals using Monte-Carlo method. This method also was used by Mackenzie [118,119] to develop random distributions of misorientation for fcc materials. The distribution corresponding to a random aggregate of hexagonal crystals has a maximum of 90 o and a cut-off value of 93.84 o . The restrained distribution corresponds to a basal fiber where the neighboring grains have the c-axis at the angels lower than 30 o . This distribution has the maximum at 30 o and the cut-off value of 42.3 o [115]. It is noticed that the number fraction of misorientation between 85 o and 90 o raises at 0 turn and this peak remains noticeable throughout processing. Also, another peak is raised pronouncedly in about 30 o . To reveal the microstructure with many highly-refined grains and to precisely measure final grain size after processing of ZK60 by 5 turns of HPT, an EBSD analysis was carried out on a processed sample. Figures 4.9(a)-(e) demonstrate EBSD orientation and pattern quality micrographs respectively. The colors in the EBSD images are representative of misorientations by referring to the triangle colors. The depicted unit triangle with multi-color pattern shows each crystal orientation with a reference color. The information of the boundary misorientation distributions is given for HAGBs having misorientation angles, of  >15  and LAGBs having misorientations of ≈ 2-15 . It is observed in Fig. 4.9(d) that after 5 turns there is an ultrafine- grained structure containing many highly refined grains in the structure. Consequently, a bimodal microstructure and a coexistence of two populations of grains with distinctly different average size 53 and morphology is noticeable. The grain size distribution in mid-radius of the disks is demonstrated in Fig. 4.7(d). It is suggested that the microstructure is bimodal with some grains having sizes of ∼1.2 µm lying within arrays of relatively smaller grains with an average grain size of ~250 nm. The overall measured grain size was ~710 ± 30 nm. Table 4.3 provides a percentile analysis of misorientation angles measured from Fig. 4.8. It can be seen that the as-received extruded material has a misorientation angle distribution close to the random distribution in Fig 4.8(a) with ~95% of HAGBs. Observations show there is no decrease in the fraction of HAGBs during the HPT processing through different turns of HPT processing. However, there is drop in 0 turns when only compression applies to the material. HPT turns Fraction of misorientation angle percentile (%) HAGBs LAGBs 30 o 85-90 o Extruded 95 5 7 8 0 85 15 12 23 ½ 95 5 8 14 1 95 5 7 12 5 96 4 14 11 Table 4.3. The HAGBs, LAGBs, 30 o and 85 o -90 o misorientation angles area fraction percentiles measured for studied materials. 54 (a) (b) (c) (d) Figure 4.8. Number fraction of the misorientation angles of ZK60 in the (a) as- received extruded condition and after (b) 0 turn, (c) 1/2 turn, (d) 1 turn and (e) 5 turns of HPT in the mid-radius of the disks. (e) 55 10 µm (a) As-received 5 µm (c) 1/2 turn 5 µm (d) 1 turn 5 µm (e) 5 turns 15 µm (b) 0 turn Figure 4.9. EBSD orientation micrographs of ZK60 in the (a) as-received condition and after (b) 0 turn, (c) 1/2 turn, (d) 1 turn and (e) 5 turns of HPT in the mid-radius of the disks. 56 4.1.2. Microstructural characteristics of the samples after combination of SPD processes (ECAP + HPT) Figure 4.10 shows image quality (IQ) maps of ZK60 after (a) processing by initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT for 5 and 10 turns of HPT under applied pressure of 2.0 GPa. Corresponding grain size distributions are shown in Fig. 4.11. It is apparent from inspection of Fig. 4.10 that grain refinement takes place after processing by the ECAP technique at 473 K. However, there is a bimodality in microstructure in order of magnitude containing large grains with sizes of about 20 µm and grains with sizes of about 2 µm. Clearly, smaller grains containing more uniform distribution of grain size and no bimodality are achieved in Fig. 4.10(b). By continuing straining after processing by a combination of ECAP for four passes and HPT through 10 turns a fairly uniform structure is achieved with an average grain size of 2.2 µm. Corresponding grain to grain number fraction of the misorientation angles for different processing conditions in Fig. 4.10 are reflected in Fig. 4.12(a)-(c). It is noteworthy to see that after processing by ECAP and ECAP + HPT, the fraction of LAGB’s are lower than 10% in all processing conditions. However a high number of fraction misorientation observed in a peak about 85-90 o after ECAP and as material processed by HPT afterwards, this peak diminished significantly and completely different misoriention distributed randomly with values skewed to smaller angels concentrating around 30 o and 60 o . Final missorientation distributions after combination of two process is mostly close to merely single step SPD process by HPT than single step ECAP processes. EBSD maps of ZK60 after processing by initial ECAP for 4 passes and a combination of ECAP and HPT in Fig. 4.13 confirm a change in grain orientations from ECAP grain orientations to HPT grain orientations. 57 ECAP: 4p ECAP + HPT: 4p + 5t ECAP + HPT : 4p + 10t (a) (b) (c) 20 µm 10 µm 10 µm Figure 4.10. Image quality (IQ) maps of ZK60 after (a) processing by initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT for (b) 5 and (c) 10 turns of HPT under applied pressure of 2.0 GPa. 58 (a) (b) (c) Figure 4.11. Grain size distribution of ZK60 after (a) processing by initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT for (b) 5 and (c) 10 turns of HPT under applied pressure of 2.0 GPa. Grain size = 4.6 µm Grain size = 2.9 µm Grain size = 2.2 µm 59 (a) (b) (c) Figure 4.12. Number fraction of the misorientation angles (a) processing by initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT for (b) 5 and (c) 10 turns of HPT under applied pressure of 2.0 GPa. 60 ECAP: 4p ECAP + HPT: 4p + 5t ECAP + HPT : 4p + 10t (a) (b) (c) 20 µm 10 µm 10 µm Figure 4.13. EBSD maps of ZK60 after (a) processing by initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT for (b) 5 and (c) 10 turns of HPT under applied pressure of 2.0 GPa. 61 4.1.3. Microstructural characteristics of the samples annealed after processing by HPT In order to achieve High strength-high ductility (HSHD) in HPT processed Mg alloys, the effect of short term post-HPT annealing process is investigated and compared to long term annealing. Initial investigations in previous sessions demonstrated that this material contains UFG microstructure with and average grain size of 700 nm in mid-radius of the disks. The annealing treatment is investigated for ZK60 Mg alloy samples processed through 5 turns of HPT. Samples are annealed isothermally for 10 and 20 min at selected constant temperatures ranging from 473 to 548 K in order to investigate the differences in microstructure and properties comparing to long term annealing processes. Initial HPT processed samples are examined at temperatures as low as 450 K for 2500 min. Fig. 4.14 demonstrates the IQ maps of samples after (a) long term annealing for 2500 minutes for at 450 K and short term annealing for 20 minutes annealing at temperature of (b) 473 K and (c) 548 K where 14(a) has a higher magnification in Fig. 4.14. Corresponding grain size distributions are shown in Fig. 4.15. It can be observed that after long term annealing fairly equiaxed structure with uniform grain structure is obtained with grain size of 2.1 µm while few of the grain are in the range of less than 1 µm. However, after short term annealing for 20 min, clearly it can be observed that a bimodal structure is revealed due to abnormal grain growth during annealing process. The microstructure contains grains with sizes of less than 1 µm and some large grains in one order of magnitude in size. Fig. 4.16 corresponds to misorientation angle analysis of microstructures shown in Fig. 4.14. It can be observed that after long term annealing the distribution of misorientation angels tends toward normal Mckenzie distribution of misoriention for hcp materials. However the trend of distribution of misorientaion angle after short term annealing is similar to initial HPT processed samples containing higher intensities at peaks at 30 o and 90 o . 62 HPT + Anneal (450 K , 2500 min) HPT + Anneal (548 K , 20 min) HPT + Anneal (473 K , 20 min) (a) (b) (c) Figure 4.14. IQ maps of ZK60 processed by 5 turns of HPT after (a) long term annealing after 2500 minutes at 450 K and short term annealing after 20 min annealing at temperature of (b) 473 K and (c) 523 K. 63 (a) (b) (c) Figure 4.15. Grain size distribution of ZK60 processed by 5 turns of HPT after (a) long term annealing after 2500 minutes at 450 K and short term annealing after 20 min annealing at temperature of (b) 473 K and (c) 523 K. G.S. = 2.1 0.2 µm G.S. = 2.7 0.8 µm G.S. = 3.7 2.1 µm 64 (a) (b) (c) Figure 4.16. Number fraction of the misorientation angles after (a) long term annealing after 2500 minutes at 450 K and short term annealing after 20 min annealing at temperature of (b) 473 K and (c) 548 K. 65 HPT + Anneal (450 K , 2500 min) HPT + Anneal (548 K , 20 min) HPT + Anneal (473 K , 20 min) (a) (b) (c) Figure 4.17. EBSD maps of ZK60 processed by 5 turns of HPT after (a) long term annealing after 2500 minutes at 450 K and short term annealing after 20 min annealing at temperature of (b) 473 K and (c) 548 K. 66 Fig. 4.17 demonstrates the EBSD maps of ZK60 processed by 5 turns of HPT for (a) long term annealing after 2500 minutes for 450 K and short term annealing after 20 minutes annealing at temperatures of (b) 473 K and (c) 523 K. The map in Fig. 4.17(a) shows after long term annealing there is more random grain orientations due to recrystallization compared to short term annealed grains in Fig. 4.17(b) and (d). It is observed that the new recrystallized grain orientations are mostly dominated by the initial HPT grain orientations contained basal plains perpendicular to the TD-RD plane. Therefore short term annealing results in static recrystallization containing bimodality both in misorienation angles and grain size distribution with some very fine grains and some coarser grains. By increasing the annealing temperature the grain morphology changes from reasonably equiaxed with an average grain size in the range of 3-4 µm to structures containing abnormal elongated grains with bimodality in grain size distributions. It shows that recrystallization occurs when conducting short-term annealing at temperature of 473K and 528 K. 4.2 Textural analysis In order to understand the microstructural evolution and corresponding mechanical properties of processed materials, texture analysis were carried out by EBSD techniques and pole figures are studied from corresponding EBSD maps shown in session 4.1. 4.2.1 Textural evolution of the samples processing by HPT Figures 4.18 shows the texture variations for as-received extruded ZK60 and after processing different HPT turns through up to 5 revolutions. It can be seen that from {0001} and {101 ̅ 0} and {21 ̅ 1 ̅ 0} pole figures for the as-received ZK60 that the initial texture is {hki0} fiber which means {1010} planes in the extruded condition oriented perpendicular to the extrusion axis and the {0001} planes generally lie parallel to the extrusion axis. Therefore the hcp crystal c-axis are in the RD-TD plane of the disks before processing by torsion by HPT. The development of this 67 prismatic fibre texture in extruded Mg alloys has been reported in literate [110,120,121]. When N = 0 by applying only compression without torsion for 1 min under pressure of 2.0 GPa, due to the limited slip system in hcp materials at room temperature, compression is accompanied by outflow of material in the radial direction in the disk during quasi-constrained HPT processing, the c-axes of many hcp crystals tend to rotate and align with the compressive direction with basal planed parallel to the TD-RD plane of HPT. Therefore a strong basal fiber texture of {0001}<uvtw> were produced. The EBSD results confirms the XRD analysis investigated of ZK60 after 0 turns of HPT [65]. However by applying torsional straining there are transformation state while N = 1/2 and N = 1 turn which contains two components including the compressed {0001}<uvtw> basal fiber and second fiber. After straining with 5 turns of HPT and obtaining UFG microstructure the final texture is mostly dominated by basal fiber again while the maximum intensity of basal planes are much lower than fiber produced only by compression. 4.2.2 Textural evolution of the samples after processing by combination of SPD processes (ECAP + HPT) The pole figures of {0001 }, {101 ̅ 0} and {21 ̅ 1 ̅ 0} crystallographic orientations in the TD-RD palne and normal to the ND of HPT directions are shown in Fig. 4.19. It is evident that there is a significant difference of basal planes orientation in the most grains where the prismatic fiber of extruded are distributed parallel to the extrusion direction changes in to basal fiber texture after HPT however after ECAP there is a large increase in (0001) pole density after four passes that could be attributed to creation of new refined grains. It is noted that most of basal planes depart from the extrusion direction and are located in a direction between the extrusion and transverse directions. The rotation of basal planes during ECAP process is believed to be led by shearing parallel to the basal planes. 68 Extruded N = 0 N = 1/2 N = 1 N = 5 - Max = 31.5 - Max = 16.5 - Max = 11.4 - Max = 14.1 - Min = 0.7 - Min = 0.7 - Min = 0.7 - Min = 0.7 RD TD - Max = 8.1 - Min = 0.7 Figure 4.18. The pole figure of {0001 }, {101 ̅ 0} and {21 ̅ 1 ̅ 0} crystallographic orientations on the TD-RD plane and normal to the ND of HPT direction from the mid-radius of the extruded disk and processed disks with N = 0, 1/2, 1 and 5 turns of HPT at 298 K. 69 After processing by combination of ECAP and HPT, it is evident that texture of basal plane evolve to majority of planes oriented near to the normal direction. However comparing to single step process of HPT, the basal planes distribution are more scattered. Extruded 5t 4p 4p + 5t 4p + 10t - Max = 14.1 - Max = 86.4 - Max = 15.2 - Max = 15.6 - Min = 0.7 - Min = 0.4 - Min = 0.6 - Min = 0.6 RD TD - Max = 8.1 - Min = 0.7 70 4.2.3. Textural evolution of the samples after post-HPT annealing To understand the textural evolution during annealing process after HPT, Figure 4.20 of samples processed by 5 turns of HPT and after post-HPT annealing. Figure 4.20. The pole figure of {0001 }, {101 ̅ 0} and {21 ̅ 1 ̅ 0} crystallographic orientations on the TD-RD plane and normal to the ND of HPT direction from the mid-radius of samples processed by 5 turns of HPT and after post-HPT annealing. From the pole figures in Fig. 4.20 it is apparent that generally there is no change on the position of basal planes neither after long term annealing for 4500 min at 450 K nor at short annealing for 20 min in higher temperature of 548 K. However it is observable that long term annealing at low RD HPT 5 turns HPT + Annealed (450 K, 2500 min) HPT + Annealed (473 K, 20 min) HPT + Annealed (548 K, 20 min) TD - Max = 14.1 - Max = 8.9 - Max = 10.5 - Max = 51.0 - Min = 0.7 - Min = 0.7 - Min = 0.7 - Min = 0.6 71 temperature decrease the maximum intensity of the basal planes while short annealing at higher temrprature increase the maximum intensity. Also annealing at 200 K for 20 min can hardly affect the texture of HPT processed material. 4.3. Microhardness measurements 4.3.1 Microhardness values after processing by HPT Figure. 4.21 demonstrates the Vickers microhardness, Hv, plotted against the distance from the centers of the disks processed by HPT through total numbers of 0, 1/8, 1/4, 1/2, 1, 2 and 5 turns: the lower dashed line is representative of the average hardness value obtained from the cross- sectional area of the as-received extruded material. Thus, the hardness values increase significantly after pressing the sample at 2.0 GPa for 0 turn where there is no torsional straining. Furthermore, it is readily apparent that the hardness values increase as the numbers of HPT revolutions increase. The trend of increasing hardness continues up to 2 turns in the disk peripheral areas but in the central regions the increase is initially less pronounced. Eventually, the hardness value saturates at Hv ≈ 124 after processing by HPT through 5 turns. Figure 4.22 shows microhardness recorded values on the cross-sectional planes longitudinally along randomly selected diameters against the calculated equivalent strains by means of Eq. 2.4 for the ZK60 alloy after HPT through 1/8, 1/4, 1/2, 1, 2 and 5 turns under a pressure of 2.0 GPa. The overall trend is indicated by the solid line in Fig. 4.22. The lower dashed line indicates the hardness value in the as-received condition. It can be observed that in the early stage of HPT, the hardness values rise quickly with accumulating strain but this slope decreases at equivalent strains above ~20 and ultimately the microhardness values become essentially saturated at a high Hv ≈125 at high strains. This evolution is consistent with a theoretical evaluation based on strain gradient plasticity modeling [122]. 72 The increase in hardness homogeneity is also confirmed by the color-coded contour maps. The color-coded contour maps were constructed to show the variations of hardness across the disks and the results are given in Fig. 4.23 after processing through totals of 1/4, 1/2, 1 and 5 turns: the Cartesian coordinates X and Y were randomly selected on the disk such that the center of each disk is located at (0,0). These illustrations provide a direct representation of the hardness evolutions with the various hardness values defined as shown in the key at the lower right. It can be observed that, by imposing torsional straining through numbers of HPT revolutions, the hardness values increase from an average of Hv ≈ 74 before processing to Hv ≈ 113 after a quarter turn followed by Hv ≈ 125 after 5 turns. A gradual transition towards higher values of hardness from the central to the peripheral areas is observed by increasing torsional straining. This gradient is more noticeable in the initial revolutions but it essentially vanishes at 5 turns and leaves a fairly uniform distribution of hardness throughout the surface area. Comparing these data with microhardness results of magnesium alloys including aluminum, such as AZ31 [123] and AZ91 [124] processed by HPT at room temperature with a pressure of 6.0 GPa, the microhardness of ZK60 saturates when using a pressure of 2.0 GPa and it also exhibits both a lower gradient towards the saturation level and a lower saturation level after imposing a large amount of deformation. 73 Figure 4.22. Variation of the microhardness values versus equivalent strain of ZK60 after processing by 1/8, 1/4, 1/2, 1, 2 and 5 turns of HPT under applied pressure of 2.0 GPa. The lower dashed line represents the microhardness value of the as-received extruded material. Figure 4.21. Vickers microhardness across diameters of disks of the ZK60 processed through various numbers of turns by HPT: the lower dashed line shows the as-received condition. 74 4.3.2. Microhardness values after processing by combination of ECAP and HPT The results shown in Fig. 2.24 exhibit the hardness values after ECAP for 4 passes and combination of ECAP and HPT through various numbers of revolutions from 1/2 up to 10 turns; the lower dashed line denotes the initial alloy in the as-received and extruded condition. The microhardness N = 5 turns N = 1/4 turn N = 1/2 turn N = 1 turn (a) (b) (c) Hv (d) ZK60 HPT: P = 2.0 GPa, RT Figure 4.23. Color-coded contour maps showing the distributions of the Vickers microhardness values on the surfaces of disks processed by HPT through (a) 1/4, (b) 1/2, (c) 1 and (d) 5 turns: the significance of the colors is shown by the key at the lower right. 75 values are reasonable uniform across the diameter horizontally oriented at the cross-section of a billet processed by ECAP for 4 passes. However by straining the material through the HPT process after ECAP, there is an increase in hardness values. This increase starts preferentially from the edge of the disk but eventually stabilize after about 10 turns. The experimental datum points from Fig. 4. 24 were replotted as the hardness values against the equivalent strain in HPT. It is readily apparent from Fig. 4. 25 that all points now cluster around a single line, the values of Hv increase rapidly at the very low strains from an initial value of Hv ≈ 93. A leveling off equivalent strains above takes place at strains higher than 200 with a well- defined saturation value of Hv ≈ 115. Figure 4.24. Vickers microhardness across diameters of disks of the ZK60 processed through through 4 passes of ECAP and various numbers of turns by HPT: the lower dashed line shows the as-received condition. 76 4.3.3. Microhardness values after post-HPT annealing process By annealing the samples for 2500 min at 450 K, it can be seen in Fig. 4.26(a) that the microhardness drops considerably from Hv ≈ 125 to Hv ≈ 85. It is noticeable in Fig. 4.16 that the fraction of LAGBs were remained unchanged after long term annealing. Fig. 4.26(b) demonstrate the effect of short term annealing on microharness values. It is noticeable that hardness values drops by annealing in the range of Hv = 80-90. Generally, by annealing samples at a temperature close to or higher than 0.5Tm, restoration processes such as recovery, recrystallization, and grain growth are enhanced [40]. As mentioned earlier the grain size are related to hardness values via Hall-Petch relationship. Formation of larger recrystallized grains after annealing of microstructure can make drop in micro hardness values. Figure 4.25. Variation of the Vickers microhardness values versus equivalent strain of ZK60 after processing by initial ECAP for 4 passes or a combination of ECAP for 4 passes and HPT for 1/2, 1, 5 and 10 turns of HPT under applied pressure of 2.0 GPa. 77 (a) (b) Figure 4.26 Vickers microhardness across diameters of disks of the ZK60 processed through 5 turns of HPT and annealed for (a) long term and (b) short term at different tempratures: the lower dashed line shows the as-received condition. 78 4.4 Mechanical behavior To investigate mechanical behavior of ZK60 alloy after processing by HPT, two sets of experiments including tensile testing and shear punch testing were conducted on specimens. 4.4.1 Tensile properties of the samples processed by HPT Variations of the engineering stress versus engineering strain of tensile samples pulled to failure are depicted in Fig. 4.27(a) using an initial strain rate of 1.0  10 −3 s −1 . Curves are plotted for the as-received material and samples processed by 0, 1/4, 1/2, 1 and 5 turns of HPT. It is observed that at 473 K the processed samples exhibit lower flow stresses and significantly higher elongations to failure by increasing the number of HPT revolution. The homogenized ultrafine grained structure achieved after 5 turns of HPT exhibit superplastic elongation of about 415%. The corresponding appearance of the specimens pulled to failure shown in Fig. 4.27(a) are depicted in Fig. 4.27(b) where the upper specimens is in an untested condition, the elongations to failure of all specimens are recorded in the right columns of the samples and the left column shows the testing strain rate, 𝜀 ̇. To investigate the mechanical properties of the obtained homogenized ultrafine-grained material after 5 turns of HPT experiments were conducted on the processed materials at initial strain rates ranging from 1.0 × 10 -1 to 3.0 × 10 -5 s -1 and temperature of 473, 523 and 573. The engineering stress versus engineering strain curves are depicted in Fig. 4.28. The curves are representative of the generally observed ultrafine-grained material behavior at elevated temperatures and it is apparent that the curves include a low strain hardening and high elongations to failure. The ultimate tensile stress drops significantly at the lower strain rates and higher temperatures. Nevertheless higher ductilities are then achieved. 79 (a) (b) Figure 4.27. (a) Plots of engineering stress versus engineering strain (b) appearance of tensile specimens for the ZK60 magnesium alloy processed by different turns of HPT pulled in tension to failure at an initial strain rate of 1.0 × 10 -3 and 473 K. 80 (a) (b) (c) Figure 4.28. Plots of engineering stress versus engineering strain for the ZK60 magnesium alloy pulled in tension to failure at: (a) 473, (b) 523 and (c) 573 K at initial strain rate ranging from 1.0 × 10 -1 to 3.0 × 10 -5 s -1 . 81 It can be seen in Fig. 4.28 that the specimen exhibiting the highest measured elongation of ~940% corresponding to 1.0 × 10 -4 s -1 at 523 K exhibits a general thinning of the deforming material in a very uniform fashion rather than initiating a neck which is typically observed in regular metals. Pulling out in a uniform manner without the development of local necking is a characteristic feature of superplastic flow [125]. The ductility obtained for this material is higher than the maximum ductility achieved for an AZ61 alloy in tensile testing after processing by 5 turns of HPT where a maximum superplastic elongation of 620% was reported in a specimen at the same temperature and initial grain size using an initial strain rate of 3.3 × 10 -4 s -1 [18]. A comprehensive montage of the corresponding appearance of the specimens related to the curves pulled to failure at wide range of strain rate and temperature is shown in Fig. 4.29 are depicted in Fig. 4.30 where the upper specimens is in an untested condition. Figure 4.29. The tensile samples pulled in tension to failure at: (a) 473, (b) 523 and (c) 573 K at initial strain rate ranging from 1.0 × 10 -1 to 3.0 × 10 -5 s -1 . 82 It is well documented that the superplastic properties of metallic materials are associated with the length scale and with the thermal stability of their grain structure [67]. In the present investigation, the measured elongations of >400% in Figs 4.28 confirms the occurrence of superplastic flow. Furthermore, the stress-strain curves obtained from the material processed by HPT are representative of typical superplastic behavior with a relatively high yield stress at higher initial strain rates and lower yield stresses at lower strain rates due to the advent of less strain hardening. Within the strain range that the experiments were conducted, a strain rate sensitivity of ~0.3 - 0.5 was estimated using the flow stress values at a nominal strain of 0.1 at the testing temperature of 473 K. This result is consistent with earlier studies of ultrafine-grained ZK60 processed by ECAP with a similar grain size of ~0.8 µm reported after 6 passes of ECAP using processing route Bc [72] and where a similar strain rate sensitivity of about ~0.4 was obtained using the same strain rate range and temperature. Figure. 4.30. Elongation to failure versus initial strain rate for tests conducted at different temperatures. 83 The variations of the elongations to failure versus strain rate are plotted in Fig. 4.29. This result confirms that the HPT processed ZK60 magnesium alloy exhibits superplasticity at moderate strain rates in this temperature range. 4.4.1.1 Evaluating the flow behavior of ZK60 alloy: strain rate sensitivity, m, and activation energy, Q, after processing by HPT It was shown in Fig. 4.3 that processing by severe plastic deformation retains an ultrafine- grained microstructure in ZK60 magnesium alloy. The obtained microstructure leads to superplastic behavior at elevated temperatures as shown in Fig. 4.29. This behavior is representative of changes in strain rate sensitivity and deformation mechanisms. Flow curves for ZK60 magnesium alloy at different deformation temperatures are shown in Fig. 4.28. Typically, the flow stress increases to a maximum and then decreases to a finally attained steady state. This flow behavior is the characteristic for hot working accompanied by DRX [126]. It is now well established that polycrystalline materials exhibit a steady-state creep rate, 𝜀 ̇, which varies with the experimental conditions through a relationship of the form [127]: 𝜀 ̇ = 𝐴𝐷𝐺 𝐛 𝑘𝑇 ( 𝐛 𝑑 ) 𝑝 ( 𝜎 𝐺 ) 𝑛 (4.1) where D is the diffusion coefficient (=Do exp(-Q/RT)), where Do is a frequency factor, Q is the activation energy of an appropriate diffusion process, R is the gas constant and T is the absolute temperature), G is the shear modulus, b is the Burgers vector, k is Boltzmann’s constant, d is the grain size,  is the applied stress, p and n are the exponents of the inverse grain size and the stress, respectively, and A is a dimensionless constant [127]. Due to the constancy of Q at a given temperature, it is possible to determine the, m, from: 𝑚 = ( 𝜕 ln ( 𝜎 𝐺 ) 𝜕 ln ( 𝜀 ̇ 𝑇 𝐺 ) ) 𝑇 (4.2) 84 There are several ways to determine the value of m. In concept, it is possible to conduct series of tensile experiments and varying strain rates at a series of temperatures. To obtain the m-value of the HPT processed material, the normalized stress versus temperature compensated strain rate for studied material are depicted in Fig. 4.31. The relationship between the logarithmic flow stress and the logarithmic strain rate showed a sigmoidal shape, as has been often observed for many superplastic metallic alloys [128]. The results are in agreement with superplastic elongations obtained in Fig. 4.29 within an intermediate strain rate region where the strain rate sensitivity, m, is higher. The m-values tended to decrease with increasing and decreasing strain rate. 0.5 0.30 0.20 . Figure 4.31. The normalized stress versus temperature compensated strain rate for studied material shows a sigmoidal behavior of ZK60. 85 Considering the fact that tests conducted at temperatures higher than 0.5Tm, the grain boundary sliding in superplastic flow leads to a constitutive relationship of the form shown in Eq. 4.1 with A = 10, n = 2, p = 2, and D = Dgb [51]. Therefore, for a constant σ/G, the activation energy, Q, is evaluated as: 𝑄 = −𝑅 ( 𝛥 [ln 𝜀 ̇ ( 𝑇 /𝐺 ) ( 𝑑 /𝑏 ) 2 ] 𝛥 ( 1/𝑇 ) ) 𝜎 /𝐺 (4.4) 4.4.2. Tensile properties after annealing sample processed by HPT The objective of the post-HPT annealing treatment was to evaluate the potential for achieving HSHD material by influencing the strain hardening rate and the flow stress in tensile testing at room temperature. The results of the tensile tests at room temperature after annealing are shown in Fig. 4.33 and 4.34. Figure 4.32. The variation in 𝜀 ̇( 𝑇 /𝐺 ) as a function of 1/T for the ZK60 magnesium alloy processed HPT. 86 (a) (b) (c) (d) Figure 4.33. Engineering stress vs engineering strain curves of samples annealed isothermally after HPT for 10 and 20 min at temperatures of 473, 523 548 K and tested at different strain rate of 10 -3 and 10 -4 s -1 . 87 It is apparent that after short term annealing, by increasing the annealing time and temperature at constant strain rate, the elongation to failure increases while strength didn’t decrease significantly. 4.4.3. Shear Punch Testing Although superplasticity is conventionally examined under tensile tests [66], alternative localized tests such as indentation [129], nano indentation [112], impression [105] and shear punch testing (SPT) [130] methods have been studied in cases where the material is only available in small amounts, such as the samples produced by severe plastic deformation (SPD) processes. HPT method with extreme grain refining capability can only provide small samples in the form of thin disks and these are suitable candidate materials for miniature tests such as SPT to examine the superplastic behavior. By plotting shear stress against normalized displacement, SPT curves are obtained which are comparable to those obtained in uniaxial tensile tests. Figure 3.35 provides comparative results of SPT techniques at 473 K and punch rate of 5.0 × 10 -2 mm/min for the as- received samples and samples processed different turns of HPT through up to 10 turns. It is observed that the SPT curves consist of an elastic linear part, yielding followed by a work Figure 4.34. Engineering stress vs engineering stress curves of ZK60 after long term annealing at 450 K for 2500 min at different strain rate of 10 -3 and 10 -4 s -1 . 88 hardening region and then ultimate shear strength which leads to fracture, similar to tensile curves shown in Fig. 4.30. Also, both SPT and tensile results demonstrate that the strength of the material decreases with increasing numbers of turns at elevated temperature due to the development of the UFG microstructure. These curves are typical of hot working that is accompanied by dynamic recrystallization (DRX) or recovery. It is clear that the flow stress of the alloy significantly decreased with increasing temperature and decreasing the punch rates. The as-received extruded materials and obtained ultrafine grain ZK60 after 5 turns of HPT was tested in shear punch tests at 473, 523 and 573 K at different punch rate ranging from 8.0 to 1.0 × 10 -2 mm/min and the results are shown in Fig. 4.31 and 32 respectively. The SPT curves in Fig. 4.32 are representative of the generally observed behavior of UFG materials at elevated temperatures, including low strain hardening due to microstructural restoration accompanied by decline in ultimate tensile and shear stress at lower strain rates. Figure 4.35. Plots of shear stress versus normalized displacement for the ZK60 magnesium alloy processed by different turns of HPT at punch rate of 5.0 × 10 -2 mm/min and 473 K. 89 (a) (b) (c) Figure 4.36. Plots of shear stress versus normalized displacement for the extruded ZK60 magnesium alloy at: (a) 473, (b) 523 and (c) 573 K at different punch rate ranging from 8.0 to 1.0 × 10 -2 mm/min. 90 (a) (b) (c) Figure 4.37. Plots of shear stress versus normalized displacement for the ZK60 magnesium alloy at: (a) 473, (b) 523 and (c) 573 K at different punch rate ranging from 1.0 to 1.0 × 10 -2 mm/min. 91 For all curves, the ultimate shear stress drops significantly at the lower punch rates (strain rates) and indicative of a positive strain rate sensitivity index of the material. The high-temperature tensile flow stress, σ, of materials can be related to tensile strain rate, 𝜀 ̇, with a power-law relationship [23]. Considering the von-Mises yield criterion for pure shear of kinematically hardening materials, this equation can be adapted for evaluating superplastic behavior in the SPT technique by substituting 𝜀 ̇ with 𝛾 ̇ and σ with τ and rewriting the modified power-law relationship [130]: 𝛾 ̇ 𝑇 𝐺 = 𝐴 ́ ( 𝜏 𝐺 ) 1 𝑚 𝑒𝑥𝑝 ( −𝑄 𝑅𝑇 ) (4.5) where 𝐴 ́ is a material parameter, G is the shear modulus, m denotes the SRS index, Q represents the deformation activation energy, R is the universal gas constant and T is the absolute temperature. Due to the constancy of Q at a given temperature, it is possible to determine the SRS index, m, from: 𝑚 = ( 𝜕 ln ( 𝜏 𝐺 ) 𝜕 ln ( 𝛾 ̇ 𝑇 𝐺 ) ) 𝑇 (4.6) Figure 4.33 shows y versus temperature-compensated, 𝛾 ̇, where values were normalized to the shear modulus to eliminate the changes in shear modulus with temperature. There is a sigmoidal relationship that divides the curve into three distinct regions of flow, with variable values of m. This is in consistent with the with the occurrence of three distinct regions of flow with superplastic elongations up to ~940% occurring in region II over a limited range of intermediate strain rates spanning about two orders of magnitude. The sigmoidal relationship with a maximum m-value of ~0.50 in region II of shear strain rate can be observed. The obtained results confirm the repeatability of the data to produce reliable curves for evaluating flow properties of metallic materials. 92 The deformation activation energy may be calculated at constant shear strain rate as: 𝑄 = 𝑅 𝑚 ( 𝜕 ln ( 𝜏 𝐺 ) 𝜕 ln ( 1 𝑇 ) ) 𝛾 ̇ 𝑇 𝐺 (4.7) Consequently, the normalized 𝛾 ̇ values are plotted against the reciprocal of absolute temperature at constant temperature-compensated shear strain on a semi-logarithmic scale in Fig. 4.34. Calculations were made in the temperature range of 473- 573 K in which the material showed a maximum value for the strain rate sensitivity. The calculate the activation energy of the material after 5 turns of HP shows a significant decrease in average activation energy value from ~119 ± 7 kJ/mol in the as-received extruded material to ~98 ± 14 kJ/mol after 5 turns of HPT. 93 (a) (b) Figure 4.38. Normalized shear yield stress versus temperature compensated strain rate for (a) A-received and (b) processed by 5 turns of HPT. 0.2 0.2 0.4 . 0.5 0.3 0.3 . 94 (a) (b) Figure 4.39. The variation in  y/G as a function of 1/T for the ZK60 for (a) as received and (b) processed by 5 turns of HPT. Q ave = 119 7 kJ/mol . Q ave = 98 14 kJ/mol . 95 5. Discussion 5.1. The effect of HPT The microstructural evolution studies during the HPT process provide essential information regarding the grain refinement mechanism and assist in optimizing the superplastic behavior at elevated temperatures. The imposition of an applied pressure of 2.0 GPa during torsional straining permits the process to be conducted successfully at ambient temperature. A similar grain refinement was also reported for the same material processed by ECAP after 6 passes at 473 K [72]. Equation 2.5 predicts the development of heterogeneity in HPT processing since the imposed strain varies across the disk from zero at the center to a maximum in the peripheral areas. Nevertheless, by continuing the straining up to 5 turns there is ultimately a fairly homogeneous distribution of fine grains with an average size within the range of UFG materials. It is noteworthy to mention there are some discrepancies between the measured grain sizes by optical microscopy and data obtained with EBSD in Table 4.1 and 4.2 respectively. Several reports also obtained smaller grain sizes for magnesium alloys after HPT processing [18,112] and this apparent difference is attributed, at least in part, to the use of different measuring procedures. Specifically, grain sizes less than ~0.5 µm are beyond the resolution capability of optical microscopy so that the initial experimental procedure may miss some observations at smaller scales. However, it is believed that this effect has only a very minor influence on the present measurements. Several grain refinement mechanisms have been proposed for fcc metals such as high- purity polycrystalline aluminum [132], Ni [133] and Cu [134]. It has been suggested that, in the early stages of straining, the relatively large initial grains are divided into sub-grains or bands of elongated cells. After experiencing more strains, these bands evolve into an essentially uniform 96 distribution of ultrafine grains [135,136]. As a result of the limited slip systems, a significantly different grain refinement mechanism is observed in hcp metals such as magnesium alloys [135]. Formation of bimodal grain structure at elevated temperatures in deformation temperature in the range of ~400-600 K is consistent with formation of the necklace-like microstructure as a result of dynamic recrystallization of magnesium alloys [26] where new finer grains form along the original grain boundaries of the initial coarser structure and these finer grains gradually consume the larger grains and thereby produce an ultrafine structure [110]. The basal slip and deformation twinning operate to comply with strain compatibility at the grain boundaries in the early stages of straining [137]. As a consequence of continued straining, the internal stresses lead to large elastic distortions in the vicinities of the grain boundaries where the stress concentrations exceed the critical resolved shear stress for non-basal slip [40]. Eventually, a rearrangement of the dislocations leads to the formation of high-angle boundaries containing a high density of grain boundary dislocations. This process leads to a homogeneous microstructure with uniform ultrafine grains. This mechanism of grain refinement predicts a multi-modal grain size distribution. To study the distribution formation of ultrafine grains ( < 1 µm) within the microstructure, Fig. 5.1 provided corresponding analysis of IQ maps shown in Fig. 4.6 during HPT process. By using OIM analyzer, Fig. 5.1(a)-(c) were divided by two fractions of ultrafine grains (grey color) and grains having sizes more than 1 µm (white color). By comparing the maps between these two figures at different stages of straining, it is apparent that new recrystallized ultrafine grains were created in the form of aggregates along the grain boundaries of large grains. They consume ~15 and ~25% of microstructure after 1/2 and 1 turn of HPT in a form of island-like structure. Continual straining up to 5 turns expands this UFG aggregates covering ~75% of the area whereas still coarser grains remained in the microstructure result in spreading the grain size distributions. These observations 97 are consistent with the model for the formation of a shear zone from the bands of recrystallized grains [40]. (a) 1/2 turn 5 µm (b) 1 turn 5 µm (c) 5 turns 5 µm Figure. 5.1 Distribution maps of ultrafine grains ( < 1 µm) during straining sample through (a) 1/2, (b) 1 and (c) 5 turns of HPT. 98 To have a better understanding of bimodality of microstructure within different aggregates in intermediate steps of HPT, another analysis conducted on IQ maps of ZK60 after 1 turn of HPT shown in Fig. 5.2 based on the configuration of the aggregates in an intermediate step of HPT and the grains size distributions were analyzed. (a) (b) Figure 5.2. Analysis of (a) IQ maps of ZK60 after 1 turn of HPT containing fine and coarse grain aggregates, (b) grain size distribution of corresponding fine and coarse grain aggregates. 5 µm 5 µm 5 µm Aggregate Area fraction (%) dave (µm) Fine grains 39 1.1 ± 0.1 Coarse grains 61 2.0 ± 0.5 99 The micrograph divided by two categories containing fine grains and coarse grains aggregates. It is apparent that the fine grains tend to aggregate as island-like structure while large grains are mostly aggregated in the rest of the area. The corresponding grain size distributions of fine and coarse aggregates in Fig 5.2(b) showed that the fine grains aggregates skewed to smaller grain sizes occupying 40% of the area fraction of micrograph with an average grain size of 1.1 µm and the rest of the area fraction was occupied by coarse grain aggregates distributed with an average grain size of 2.0 µm. It was proposed in the model of grain refinement of magnesium alloys [40] that very small recrystallized grains formed along the original grain boundaries where satisfactorily oriented for basal slip as shown in Fig. 5.3. Shear cannot still easily occur due to constraints imposed by adjacent large grains. However by applying large strains such as torsional straining in HPT, the recrystallized bands broaden and can provide a pathway for easy slip throughout the specimen. High local stresses in the advances of these bands will promote local strain and successive recrystallization. When a clear path for shear deformation is formed through the specimen, deformation will be more concentrated in this region and will produced a shear zone detected in Fig. 4.3. Stress concentration Figure 5.3. Schematic diagram illustrating the formation of a shear zone from the bands of recrystallized grains [40]. 100 The evolution of grain size distribution during HPT process shown in Fig. 5.4 together with the quantitative analysis of grains sizes and area fractions gathered in Table 4.2 confirms that there is a gradual refinement through HPT processing at room temperature such that the distribution of grain sizes move sequentially from a peak at ~2 µm in 1/2 turn to a peak at ~1 µm after 5 evolution of HPT and formation of UFG structure with grains having sizes < 1 µm by torsional straining through HPT processing. Nevertheless the microstructure obtained after 5 turns demonstrated another peak at grains size of < 0.5 µm. Few evidences of formation of this grain within this size range are available at early stages of HPT obtained in Table 4.2. Figure 5.5 illustrates an evolution of misorientation angles in ZK60 alloy from the extruded condition through up to 5 turns of HPT in the mid-radius of the disks obtained from Fig. 4.8(a)- (e). It can be noticed that there are some distinct changes in misorientation angle distribution from extruded condition through up to 5 turns of HPT. These changes are mostly concentrated on Figure 5.4. Grain size distribution of ZK60 in the as-received condition and after 0, 1/2, 1 and 5 turns of HPT in the mid-radius of the disks. 101 misorientation angles of 30 o and 90 o . Table 4.3 showed that ~14% of the boundaries within larger grains contain misorientation angle of ~90 o at early stages of HPT by straining the sample for 1/2 revolution. However further analysis showed that this peak does not exist in as-received extruded material but after compression by 0 turn of HPT the intensity of this peak raise from 8% to 23% After starting the HPT the intensity of the peak decreases by 11% after 5 turns of HPT. It is well documented [138] by study of a line profile of point-to-point misorientation angle across the twin boundaries of a compressed magnesium alloy sample that the lattice misorientation angle of 86.3 o was caused by the {1012 ̅ } twinning. EBSD analysis of magnesium single crystal during unidirectional compressive deformation at room temperature also showed that [139] the early stage of deformation in single crystal can be characterized by profuse {1012 ̅ } extension twining. They can convert the whole sample into softer orientations for slip and completely eliminating the initial Figure 5.5. Comparison of number fraction of misorientation angles in ZK60 alloy for the extruded condition and after processing by HPT for 0, 1/2, 1, and 5 turns of HPT in the mid- radius of the disks 102 unfavorable orientation of the single crystal. Observations in microstructural studies by optical microscopy, SEM and EBSD in Figs. 4.2, 4.4 and 4.9 respectively showed that twins existed more abundantly in the partitions of coarse grains located at the center and half-radius areas at early stages of HPT. It indicates the contribution of the twining to accommodate deformation during of HPT processing. Successively secondary and tertiary {1012 ̅ } extension twins were found to form in the primary extension twinned areas. Therefore the observed peak in the range of 85-90 o at early stages of HPT are attributed to misorientation angles at the boundaries between a parent grain and twin bands. Another peak with value of ~14% of the misorientation angle was found at ~30 o at final stages of HPT with formation of UFG structure. Although this peak existed after 0 turn of HPT, it loses its intensity during the intermediate stages of HPT and finally it appears again after 5 turns of HPT by formation of ultrafine grain materials. Investigation on ductility in single crystal of magnesium from the specimen strained up to -1.0 showed that a very high frequency peak gradually appeared at about 30 o ± 5 o in the misorientation angle distribution [139]. Comprehensive textural observations on this phenomena confirmed that newly recrystallized grains formed within the twin bands were rotated in parallel around the c-axis of their parent twin by an average of 30 o . Figure 5.6 illustrate the orientation of this configuration with respect to HPT coordinate. It is suggested that their formation was attributed to the successive fragmentation of the twin bands due to prismatic slip and further continuous dynamic recrystallization [139,140]. It is worth noting that a peak in misorientation distribution is also expected in magnesium with a basal fiber texture [115-117] and the present results suggests the formation of such texture after 5 turns of HPT. 103 It is well documented that magnitude of critical resolved shear stress (CRSS) for basal plane slip in magnesium single crystals are 100 times lower than for non-basal plane slip at ambient temperature [141]. Therefore the basal <a> slip system set of {0001}<112 ̅ 0> are the easy glide systems in Mg. The prismatic <a> slip system {101 ̅ 0}<112 ̅ 0> and 1 st -order pyramidal <a> slip system {101 ̅ 1}< 112 ̅ 0> also can accommodate deformation in magnesium, however none of these systems is capable to incorporate the strain along the c-axis direction. The activation of the 2 nd - order pyramidal <c+a> slip system {112 ̅ 2}< 112 ̅ 3> with a c-slip component may provide the required additional degree of freedom. However it is significantly dependent on temperature [110]. On the other hand, deformation twinning in Mg is usually easier than pyramid <c+a> slip. Therefore, twinning plays an important role to accommodate the deformation, particularly at low temperatures. The evidence of abundant twining was noticed in EBSD in Fig. 4.9 at early stages of HPT and specifically after compression by N = 0. Misorientation distribution evolution also confirmed the existence of twin boundaries with high number of fraction of 30 o and 90 o . Based on the extension and contraction of the crystal along the c-axis direction, two twin modes have been Figure 5.6. Orientation of crystals of newly recrystallized grains formed within the twin bands rotated in parallel around the c-axis of their parent twin by an average of 30 o . 104 observed in magnesium alloys. They were classified as extension twins {101 ̅ 2}<1011 ̅ ̅ ̅ ̅ > and contraction twins {101 ̅ 1}<1012 ̅ ̅ ̅ ̅ > [142]. Corresponding pole figure of EBSD orientation maps in Fig. 4.18 demonstrated that extruded and sliced disk-like samples prepared for HPT contain grains with basal planes typically having c-axis parallel to the extrusion direction which is perpendicular to torsional axis in HPT coordinates. However, the angle between the c-axis in TD-RD plane are mostly depend on sampling location. This configuration was shown in Fig 5.7. It can be seen in Fig. 4.18 that texture changes significantly from prismatic fiber to basal fiber after applying very low straining during compression. The direction of compressive load in this situation is perpendicular to the c-axis of the material. Therefore the extension twinning {1012} becomes active and subsequently, the c-axis of crystals tends to reorient nearly parallel to the loading direction. A rotation of the c-axis by ~90º was observed at strains less than 20% in compression of HPT when N = 0. Therefore, imposition of compressive pressure of 2.0 GPa prior to torsional straining in HPT can lead to rotation and realignment of the c-axis parallel to the compression axial direction. This is in agreement with the observed variations in texture. ODF analysis during HPT of AZ31 at room temperature [110] α c-axis c-axis A B RD ND TD Figure 5.7. Orientation of crystals of random A and B grains in the sliced disks of as-received extruded materials before HPT processing. 105 showed that by imposition of torsional straining some grains reorient their c-axis about 10° from the ND under the compressive pressure and they oriented almost perpendicular to the shear direction and this make basal slip <a> become the easiest process for accommodating shear deformation. Some other grains that have their c-axis tilted at angles more than 10° from the ND, due to the limited slip systems available at lower temperatures, mainly basal slip <a> and deformation twinning may make contributions to accommodate the torsion deformation in HPT so that the orientation between the c-axis and the torsion axis will not change significantly. However, it is worth noting that the maximum intensities observed in the pole figures decreases after 1/2 turn and decreases even more after 1 turn which suggests a gradual dissolution of the initial strong compression texture. After processing by 5 turns of HPT the pole figure shows that the majority of basal {0001 } planes are oriented parallel to ND. It is well documented [138] that for materials with low stacking fault energy and slow recovery rate, Fig. 5.8, a higher hardness will be reached in the peripheral region in the early stages of HPT and therefore the hardness is initially higher at the edge and lower in the center of the disks processed by HPT. The observation of microhardness evolution during the HPT processing in Fig. 4.21 are in good agreement with this study. The frequent equation using to correlate true stress and true strain data with the strength coefficient and the strain-hardening exponent are described by the Hollomon-equation [143]: 𝜎 = 𝐾 𝜀 𝑝 𝑛 (5.1) where σ is the true stress, K is a material constant, εp is the true strain and n is the strain hardening index or the n-value. This equation can be simply modified for hardness values for evaluating strain hardening behavior occurring in the early stage of HPT. Brinell [144] found that for steels with a wide range of carbon contents, there is a linear relationship: 106 𝜎 = 𝑐𝐻 (5.2) where c is related to material. It was shown by O’Neill [145] that although it seems not to be a constant value of c for all metals, if they showed little work hardening then c = 0.36. A quantitative analysis was conducted to estimate the extent of strain hardening by Hv measurements. In this method by replacing σ with Hv from the Eq. 5.2 and εp with ε eq (or εvM) obtained from Eq. 2.5 the modified arrangement of Eq. 5.1 can be rewritten in the form of: 𝐻𝑣 = 𝐾 ′ 𝜀 𝑒𝑞 𝜂 (5.3) where K´ is a new materials constant and η is the hardenability exponent which corresponds to the slope in a double logarithmic plot of the Hv values versus equivalent strain of HPT. Figure 5.9 demonstrates the plot of Hv values against ε eq for the ZK60 alloy after up to 5 turns of HPT extracted from Fig. 4.22 in a double logarithmic scale for. It can be seen that data up to 𝜀 𝑒𝑞 ≤ 20 can be well fitted by a straight inclined line before saturation. Figure 5.8. Schematic illustration of the variation of the Vickers microhardness across the disk at low total strains in HPT for materials having either slow or fast rates of recovery [138]. 107 The slope of this data-set is representative of η which was measured ~0.07 for this experiment. The dashed line illustrated in the figure is the saturated value of microhardness at higher equivalent strains. The obtained hardenability exponent is in a good harmony with other hardenability exponents found for magnesium alloys processed by HPT [112,146,147]. Table 5.1 gathered strain hardening data of materials processed by HPT. These data are helpful to develop a phenomenological model to estimate minimum level of accumulated strain or number of HPT turns required to obtain a level of saturated hardness values for different materials. By replacing ε eq from Eq. 2.5 in to hardenability Eq. 5.3: 𝑁 𝑃 ,𝑇 = ℎ√3 2𝜋𝑟 ( 𝐻 𝑠𝑎𝑡 𝐾 ′ ) −𝜂 (5) 1 0.07 Fig. 5.9. Plot of Hv values against ε eq in double logarithmic scale for the ZK60 alloy after up to 5 turns of HPT under 2.0 GPa at ambient temperature. The inclined solid line demonstrates the relationship between data set up to ε eq ≈ 20 prior to saturation and dashed plateau line demonstrates the saturation limit. 108 where NP,T is the estimated a minimum number of HPT turns to achieve saturation hardness at constant applying pressure and temperature of HPT. Material Initial Condition Processing condition Ave. grain size (µm) Hv (kgf/mm 2 ) Hardenability Ref. 298/Tm εeq P(GPa) di df Initial Saturation K’(kgf/mm 2 ) η Ti-6Al-4V 1 h at 1023 K 0.15 30 6.0 10.1 0.13 295 365 31 0.031 [148] Ti-6Al-4V 4 h at 873 K 0.15 30 6.0 9.7 0.07 305 405 32 0.052 [148] CP Al 1 h at 773 K 0.32 0-3 6.0 1000 1.4 20 43 42 0.18 [146] CP AL 1 h at 773 K 0.32 3-10 6.0 - 1.4 43 37 63 -0.24 [146] Zn-22Al 1 h at 753 K 0.51* 40 3.0 1.4** 0.35 68 24 61 -0.22 [149-152] ZK60 Extruded 0.32 20 6.0 10 - 72 110 84 0.07 [65] AZ31 Extruded 0.32 30 6.0 10 0.9 60 115 83 0.08 [153] ZK60 Extruded 0.32 20 2.0 10 0.7 74 120 99 0.07 - * With respect to eutectic alloy ** The grain size at equiaxed region was 1.4 µm, and alternating a and b phases lamellar structure having average thicknesses of 100 nm Based on a comparison of the results from different materials studied through HPT processing gathered in Table 5.1, three strain hardening behavior of materials were distinguished during the HPT processing. In each hardness evolution trend, the degree of the strain hardening or softening is expressed by hardenability exponent, η. When η > 0, it represents the general hardness trend for the materials demonstrating that the hardness increases initially with increasing equivalent strain and then saturates to the maximum hardness at reasonably high strains. In Table 5.1 the high-purity aluminum exhibits a similarity with the strain hardening metals presenting overall high hardness after HPT with early stages of straining hardening with η > 0. However, there is a subsequent softening with η < 0 with microstructural recovery to achieve saturation at high hardness values after reasonably high numbers of HPT turns [146]. For typical materials such as Zn-22Al with η < 0 [148,150,154], it is discovered that strain softening with a decrease in hardness is followed by an ultimate saturation with increasing equivalent strain. This hardness behavior is essentially the opposite of the conventional strain hardening behavior Table. 5.1. Summary of the obtained saturation level and hardenability of HPT processed materials 298/Tm is homologous temperatures of processing at room temperature 109 obtained in this experiment even with significant grain refinement. It is reported that this weakening behavior in a Zn-Al alloys is due to the decomposition of hard Zn precipitates as a result of alteration in the precipitation kinetics in the Al-rich phase by applying the severe plastic deformation through HPT processing [151] and also transformation of lamellar structure into an ultrafine-grained equiaxed structure [152]. The strengthening mechanisms of metals can be classified by dislocation strengthening, precipitation strengthening, solid solution strengthening, and grain boundary strengthening. The first three mechanisms deal with hindering dislocation movement in grains and the last mechanism inhibits extension of plastic deformation in the entire structure. A recent investigation on pure Mg [155] shows not only grain refinement and dislocation density should be considered but also the texture after HPT plays an important role in hardness of the SPD processed materials. Since the basal plane is the primary slipping system active in low temperature deformation of Mg alloys, changes in the basal plane texture is perhaps influential in the mechanical behavior of the alloy. Experimental results in textural evolution in this study showed that even after the early stages of HPT by compression a remarkable increase of microhardness took place with respect to the extruded ZK60 alloy where no such extreme grain refinement and high dislocation density occurred [16]. It can be deduced that the texture strengthening is an important factor in hardening of materials especially in the early levels of processing in which dislocation hardening and grain size play only minor roles. It was also shown in this study and also reported in literature [120] that in magnesium alloys the majority of basal planes are oriented parallel to the extrusion direction after processing. Inevitably, by applying shear through torsional straining in HPT, the basal planes must rotate by 90 degree to positions parallel to the shear direction. The critical resolved shear stress, CRSS, of non-basal slip in single crystal Mg [156] is two times the magnitude of the basal 110 slip systems at ambient temperature. Considering the point that deformation along the forming direction limits the basal slip, it is beneficial to activate non-basal slip which leads to texture strengthening at room temperature. Therefore texture hardening contributes to the microhardness values obtained at room temperature after processing by HPT. The elongations to failure are frequently reported to increase with decreasing strain rate in magnesium-based alloys at testing strain rates at and below ~1.0 × 10 -4 s -1 [70,121,157]. On the other hand, it has been well established for many superplastic alloys that the relationship between stress and strain rate is often sigmoidal, with a low strain rate sensitivity at low strain rates in region I and at high strain rates in region III while there is a high strain rate sensitivity at intermediate strain rates in region II where the material exhibits maximum superplasticity [67,158] where the strain rate sensitivity, m, is high (>0.3). The present results in Fig. 4.30 and 4.31 are consistent with the occurrence of three separate flow regimes with a loss in the superplastic behavior at both faster and slower strain rates. This result is also consistent with earlier data for the ZK60 alloy processed by ECAP for 6 passes and then tested in tension at the same temperature of 473K [71-73]. Successful processing of different magnesium alloys through ECAP at high temperatures generally results in grain sizes no smaller than about 2 µm [121, 157,159] and this leads to superplastic elongations at moderate strain rates. However, the results for the ZK60 specimens processed by ECAP with similar grain size of ~0.8 µm led to elongations up to ~1300% at 473 K when testing with a strain rate of ~2  10 -4 s -1 [71] whereas in the present investigation the maximum elongation was about 940% at a similar strain rate of 1.0  10 -4 s -1 at 523 K. The reason for this difference lies in the size of the tensile samples since the ECAP tensile samples were cut from billets with regular dimensions whereas in the present investigation using HPT disks, it was necessary to cut miniature specimens with gauge lengths of 1 mm and gauge cross- 111 sectional areas of only about 1 × 0.64 mm 2 . Despite many investigations reporting superplasticity after ECAP, less number of reports are available after HPT. Table 5.2. has summarized some superplastic repots after obtaining homogeneous microstructures and compared to this study results. Table 5.2. Reports of superplasticity in ultrafine-grained materials produced by HPT Alloy or Composition (wt%) HPT Grain Size (nm) Superplasticity Reference Number of Turns Pressure (GPa) Temperature (K) Testing Temperature (K) Strain Rate (s -1 ) Maximum elongation (%) AZ61 5 3 423 230 473 3.3 × 1.0 -3 620 [18] Mg-9Al 5 3 298 210 473 3.3 × 1.0 -3 600 [51] 423 330 1.0 × 1.0 -3 550 5.0 × 1.0 -3 620 3.3 × 1.0 -4 810 Mg-9Al 423 370 1.0 × 1.0 -3 590 Mg-10Gd 5 6 RT 100 673 1.0 × 1.0 -2 470 [162] 1.0 × 1.0 -3 580 ZK60 5 2 RT 700 473 1.0 × 1.0 -4 535 523 1.0 × 1.0 -4 940 573 3.0 × 1.0 -5 620 Comparing the ductilities generally obtained for the materials presented in Tables 2.2 and 5.2 for same materials processed by ECAP and HPT respectively and pulled into failure at identical conditions, it is recognized that a limited occurrence of very high ductilities in alloys is available after processing by HPT. Nevertheless, it was been reported that the occurrence of high ductilities after HPT often does not satisfy the conditions for superplastic flow because elongations to failure are generally less than 400%. It is now well established that the use of miniature specimens will affect measurements of the post-necking elongations in tensile testing [160,161]. Thus, there is a general consistency between the results on the ZK60 specimens processed by HPT in the present investigation and the ZK60 specimens processed earlier by ECAP [71] but it is difficult to directly compare the measured superplastic elongations because of the significant differences in the sizes of the tensile specimens. A similar trend of superplastic behavior has been observed in magnesium alloys processed by HPT at room temperature with grain sizes smaller than 1 µm [51,59,162] . 112 Another circumstance of samples size effect is processing bulk samples of an Al-3.0%Mg-0.2%Sc alloy by HPT in the form of small cylinders with diameters of 10 mm and heights of 8.6 mm attained a grain size of ~130 nm at the periphery [55]. Depending on the position of the specimens within the cylinder, tensile elongations of ~1000% and ~1600% were documented in the vertical and horizontal specimens, respectively. This experiment proves that the elongations to failure are enhanced by increasing the gauge cross-sectional area of the tensile specimens. However, direct comparisons of the superplastic elongations between two different SPD processing routes at same strain rates are not feasible due to the effect of the different sample sizes. Therefore an alternative method of mechanical testing is favorable in which the sample size does not make significant changes in final results. It is well documented that the dominant mechanism for superplasticity is GBS [74]. Therefore grain refinement is essential to extending superplastic regime to wider stress ranges and higher strain rate region. UFG materials contain higher number of intergranular dislocations and high fraction of HAGBs. They therefore contribute further to sliding activities by lowering the activation energy of diffusion close to grain boundary diffusion. To compare the obtained superplastic characteristics of ZK60 in this study, Table 5.3 gathered superplastic characteristics of different magnesium alloys processed through different routes. Table 5.3 Superplastic characteristics in some magnesium alloys Material Superplastic characteristics Reference Alloy Processing d (µm) T (K) m Q (kJ/mol) ZK60 HPT 0.7 473-573 0.5 89 This study AZ91 ECAP + annealing 0.7-3.1 423-523 0.5 89-96 [128] AZ31 Extruded 5 573-623 0.5 - [163] ZK60 Extrusion + annealing 6.5 453-523 0.5 104.5 [164] AZ31 Rolled 25-30 623-773 0.38 145 [165] ZK60 Compressed 85 423-723 0.1-0.22 92-135 [41] AZ31 Rolled 130 573-623 0.33 127 [163] 113 It is apparent from Table 5.3 that, by increasing the grain size from UFG scale to very high grain sizes, the temperature required to achieve ductilites within the range of superplastic elongations increased while the strain rate sensitivities decreased when the grain size exceeds 10 µm. The grain boundary activation energy and self-diffusion activation energy of Mg are 92 and 135 kJ/mol respectively [167]. It is apparent that the activation energies obtained from fine grain Mg are close to the activation energy for grain boundary diffusion of Mg while by increasing the grain sizes, they are close to Mg self-diffusion activation energy in the superplastic region. The activation energy obtained for superplastic flow region in this experiment was 89 ± 6 kJ/mol for the HPT processed material after 5 turns. This value is in good agreement with value of activation energy for grain boundary diffusion. The results of the tensile tests in the present investigation also showed that for superplastic region at temperature ranging from 473-573 K ( 0.52-0.62Tm), the strain rate sensitivity of 0.5 and the activation energy was in agreement with that for grain boundary diffusion of magnesium. It is therefore likely that a dominant superplastic deformation mechanism for the magnesium alloys is grain boundary sliding accommodated by dislocation movement controlled by grain boundary diffusion. This is in agreement with other studies on ZK60 magnesium alloys [128,163] 5.2. The effect of ECAP + HPT The present investigation showed that processing of ZK60 with 4 passes via route BC at 473 K produced a grain size of 4.6 µm. The microstructure in Fig. 4.10(a) is consistent with the model proposed for formation of necklace structure during ECAP of Mg alloys [26] in Fig. 2.6. Figure 5.10 summarized the grain size distribution evolution during processing by HPT for the specimens processed by ECAP. By imposition of additional strength by HPT for 5 turns, grain size of 2.9 µm was achieved. This grain size is about 4 times larger than single step SPD process by applying 114 5 turns of HPT on the as-received extruded materials. Nevertheless, the grain size distribution is more uniform with homogeneous microstructure. After applying 10 turns, the microhardness value was saturated at Hv ≈ 117 while the obtained grain size after 10 turns of HPT was 2.2 µm. It was shown in Fig. 4.25 that the datum points for ECAP + HPT level off at 𝜀 𝑒𝑞 ≈ 100 and there is well-defined saturation at a microhardness value of Hv ≈ 117. The same type of plot were discussed earlier for the sample processed by HPT where the strain rate plateaued at 𝜀 𝑒𝑞 ≈ 20 with a well-defined saturation limit of Hv ≈ 125. The analysis in the datum point of the applied pressure in Fig. 5.11 summarized the saturation trend for HPT and ECAP + HPT processed materials. The horizontal dashed line correspond to hardness values of ~125 for HPT alone, ~117 for ECAP for 4 passes and HPT. It is also readily apparent that the 95 % error bars in datum point shown in Fig. 4.21 and 4.24 are sufficiently small to unequivocally establish these separate lines in Fig.5.11 to delineate different values for the saturation hardness under these various processing conditions. Thus, it is apparent that processing only by HPT leads to the higher saturation value, and the Figure 5.10. Evolution of grain size distribution of ZK60 processed by an initial ECAP for 4 passes and a combination of ECAP for 4 passes and HPT through up to 10 turns. 115 saturation value decreases when initial ECAP processing was imposed through 4 passes and larger number of HPT turns required to a level-off saturation value. However, the saturation hardness achieved at high strains in HPT processing is dependent upon the microstructural conditions within the material prior to the processing. Similar experiments conducted for A7075 processed by ECAP and HPT at room temperature [86] showed that there was additional grain refinement by processing materials prior to HPT and thus the saturation hardness was increased. Considering the grain size effect, the microhardness results obtained both for aluminum and magnesium alloys are consistent with microstructure achieved after ECAP + HPT. It is well documented that the saturation hardness anticipated in HPT processing is related directly to the minimum attainable grain size, dmin. There are some assumptions in the literature regarding an appropriate value of dmin in materials processed by HPT [168-170]. A model has been proposed based on the stabilized grain size concept when there is a balance between the dislocation structure Figure 5.11. Schematic of Vickers microhardness against equivalent strain for samples processed only by HPT or by combinations of ECAP + HPT. 116 formation because of the introduction of dislocations during processing and the coexisting grain size increase in due to recovery [12]. This model was developed initially to describe the attrition to a stable and minimum grain size during milling [171] of hcp materials subsequently a similar approach was used to predict the minimum grain sizes in HPT [172] and in ECAP [173] of bcc and fcc materials. Therefore several SPD processes may be modeled in a similar manner including ECAP, HPT, accumulative roll bonding and ball milling [174]. According to the model, the normalized minimum grain size in HPT is given by an expression of the form [172]: 𝑑 𝑚𝑖𝑛 𝐛 = 𝐴 3 exp ( −𝛽𝑄 4𝑅𝑇 )( 𝐷 𝑝𝑜 𝐺 𝑏 2 𝑣 𝑜 𝑘𝑇 ) 0.25 ( 𝛾 𝐺 𝐛 ) 0.5 ( 𝐺 𝜎 ) 1.25 (5.6) where b is the Burgers vector, A3 is dimensionless constant, β is a constant having a value of 0.04, Q is the activation energy for self-diffusion, R is the gas constant, T is the absolute temperature, Dpo is the frequency factor for pipe diffusion, G is the shear modulus, νo is the initial dislocation velocity, k is Boltzmann's constant, γ is the stacking fault energy and  is the yield stress. Inspection of Eq. 5.6 showed that, all terms are constant for any selected material at constant temperature except the yield stress, . In the present investigation, the HPT processing and hardness measurements were all undertaken at room temperature so that the only variable in Eq. 5.6 is the yield stress  which is equal to H/C based on Eq. 5.2. Thus, Eq. 5.6 predicts that a change in hardness will lead to a change in dmin. The initial hardness values prior to HPT processing were different for the disks processed only by HPT and those processed by a combination of ECAP + HPT. For the disks processed earlier by HPT without ECAP, the initial Vickers microhardness value was equal to the as-received extruded hardness as shown in Fig. 4.21 which is given by Hv 117  75. By contrast, the ECAP + HPT disks were processed by HPT after processing by ECAP and this means the initial condition, immediately prior to HPT, was equal to the hardness after 4 passes of ECAP with value of Hv  95 based on Fig. 4.25. The expected difference in grain size between these two processing conditions can be estimated by substituting hardness values of 75 and 95 into Eq. 5.6 so that it is anticipated the grain sizes will differ by a factor of ~1.3. Therefore the model anticipated the grain size of 500 nm. In practice, the obtained grain size failed to predict the larger grain size obtained after processing by HPT + ECAP. Therefore additional factors are requireed to take into account for the saturation hardness values obtained under different conditions. Accordingly the hardenability of the ECAP + HPT processed material measured during the HPT processing. Figure 5.12 demonstrates the plot of Hv values against ε eq for this material extracted from Fig. 4.25 in a double logarithmic scale. It can be seen that data up to 𝜀 𝑒𝑞 ≤ 100 can be well fitted by a straight inclined line before saturation. The slope of this data-set is representative of η which was measured ~0.03 for this experiment. The dashed line illustrated in the figure is the saturated value of microhardness at higher equivalent strains. By comparing Fig. 5.9 and 5.12, it is apparent that the material processed by ECAP + HPT at a higher level of applying strain while a lower hardenability obtained for this material. In order to have a better comparison of hardenability of ZK60 after processing by HPT and HPT + ECAP, Fig. 5.13 provided equations for the hardenability during the HPT processing of these two processed materials. 118 From the plots in Fig. 5.13, it is apparent that although both materials have positive hardenability exponent, the material processed only by HPT exhibited significantly higher magnitude of exponent. Therefore smaller grain sizes and larger hardness values after identical number of HPT revolution was anticipated. It is expected that that initial microstructure prior to HPT is important for achiving smaller grain sizes. 1 0.03 Figure 5.12. Plot of Hv values against ε eq in double logarithmic scale for the ZK60 after up to 10 turns of HPT under 2.0 GPa at ambient temperature. The inclined solid line demonstrates the relationship between data set up to ε eq ≈ 100 prior to saturation and dashed plateau line demonstrates the saturation limit. Initial material process by 4 passes of ECAP at 473 K prior to HPT. 119 The ZK60 alloy is a precipitation-hardening magnesium alloy containing zinc as its primary alloying element. Structural stability of different precipitates were investigated for the ZK60 alloy [175]. Recent TEM study [176] on commercial extruded ZK60 and after processing by ECAP at 473 K illustrated in Fig. 5.14. Many precipitates are present in rod-shaped in Fig. 5.14(a) although some spherical particles are also visible. Evidence of dislocation interactions with the rod-shaped precipitates in extruded condition are noticeable. However after processing by ECAP, Fig. 5.14 (b), the morphology of precipitates changes to generally spherical particles with sizes below ∼100 nm and with a reasonably uniform distribution both throughout the grains and along the grain boundaries. The EDS analysis confirmed that the majority of the rod-shaped precipitates are identified as Mg1(Zn,Zr)1, whereas after ECAP the spherical particles are Mg1(Zn,Zr)1 and Mg1Zn1. Therefore, the imposition of a high hydrostatic stress in ECAP at 473K Figure 5.13. Plots of hardenability lines of ZK60 alloy during HPT processing thorough up to saturation level for materials with different initial conditions. 120 has a major effect on the nature of the precipitate morphology the rod-shaped precipitates visible in the extruded condition. (a) (b) (c) (d) Figure 5.14. TEM micrograph and EDS analysis of precipitates in the ZK60 alloy in the extruded condition (a) and (c) and after processing by ECAP at 473 for 8 passes (b) and (d) [175]. It is well document that the main precipitate in Mg-Zn system is Mg1Zn1 which may occur as a transition phase (Mg1Zn´1) with a rod-shaped structure. They are oriented perpendicular to the basal plane or as an equilibrium phase (Mg1Zn1) with a disk shape lying parallel to the basal plane [177]. Considering the fact that basal slip is the primary slip system, for deformation of Mg, the transition phase (Mg1Zn´1) can contribute more efficiently as obstacles for dislocations. It is therefore the peak of strength expected when the majority of precipitates occur as Mg1Zn´1. It is 121 noteworthy to know that the transition phase precipitates in the early stages of aging and they transform into the equilibrium phase if overaged. Because of high temperature and high imposed strain, the ZK60 alloy could undergo overaging during ECAP. The smaller size of the Mg1Zn1 after ECAP suggests that the initial rod-shaped precipitates were also fragmented prior to overaging. Fragmentation of precipitates also reported for A7034 alloy after first pass of ECAP [178]. It can be concluded that the morphology of precipitates at the early stages of aging are more effective in pining the dislocations and by overaging to stable Mg 1Zn1 the capacity for storage of dislocations was reduced. 5.3. The effect of post-HPT annealing process It was discussed earlier that an excellent grain refinement was achieved for ZK60 by HPT processing at room temperature and excellent superplasticity was obtained at high temperature. Despite the success in achieving a highly-refined microstructure of HPT processed ZK60 with excellent strength up to 270 MPa, it is readily apparent from Fig. 4.33 that ductilities with elongations to failure of only less than 15% were obtained for the tests conducted at room temperature and strain rate of 10 -4 and 10 -3 s -1 . Alternatively, imposing a short term annealing after the HPT processing has significant effect on the microstructure as shown in Fig. 4.14-4.17. Annealing at 473 K and 548 K for 20 min increases the grain size rapidly to range of 2-6 µm. The smaller grain sizes after different annealing condition cause a drop in microhardness to the range of 80-90 kgf/mm 2 . Investigation of grain size and microhardness evolution of HPT processed AZ31 during annealing in the range of 373 to 673 K for 1800 s [179] showed that the material underwent recovery up to 423 K, recrystallization at 423 K and grain growth at higher temperatures while there is no evidence for a difference in the annealing behavior of samples processed to different numbers of turns of HPT. Also it is reported that the material followed a linear trend of 122 increasing hardness as a function of the inverse of the square-root of the grain size up to grain sizes in the range of 1 µm. Similar behavior observed after annealing the materials at different temperatures for 20 min in this study and data are in a good agreement with Hall-Petch relationship in Fig.5.15. The corresponding stress-strain curves in Fig. 4. 28 for the samples processed by HPT and then subjected to short post-HPT annealing provide a very clear demonstration of the advantage of imposing a short term annealing on ZK60. The short post-HPT annealing at temperature ranging from 473 to 528 K for 10 and 20 min can improve the ductility of material to a reasonable ductility with elongations to failure up to 40%. Although similar elongation were achieved after long post-HPT annealing at 450 K, Fig. 4.34. The strength of the material dropped significantly at faster strain rates. Hv = 47 + 65d -1/2 Figure 5.15. Hall-Petch relationship of ZK60 processed by HPT and annealed at 473 to 548 K for 20 min. 123 Fractography analysis of samples after failure in Fig. 5. 16 compared the fracture surface of HPT processed sample and post-HPT annealed samples after tensile testing. It is apparent that the broken sample in Fig. 5.16(a) was representative of brittle fracture. The plate-like features in Fig. 5.16(c) strongly suggested cleavage fracture either along twin boundaries or shear bands however the post-HPT annealed sample showed a 45 o inclined in the fracture surface in Fig. 5.16(b). This suggests the deformation may occurred at about maximum Schmid factor. At higher magnification in Fig. 5.16(f), the microstructures exhibited ductile fracture in the form of microvoids. The size and distribution of the microvoids are uniform and matches the grain size in this microstructure suggesting that the large microvoids are in fact grain pullouts in an otherwise ductile fracture. It is reasonable to anticipate generally an increase in ductility at the expense of a decrease in strength. Figure 5.17 summarized the variation of UTS and elongation to failure of post-HPT annealed ZK60 by comparing them to the initial HPT material. The measured value of UTS divided by UTS for the HPT, is plotted against the elongation to failure normalized to the value for the HPT material, fl/ fl-HPT. Datum points are shown for two testing strain rates. The horizontal and vertical broken lines correspond to the points where these normalized values are equal to 1.0. The definition of conventional behavior and high strength-high ductility (HSHD) region in Fig. 5.17 were deliberated earlier [88] and described in session 2.6. It is apparent that by post annealing the ductility of the HPT processed materials were improved. Although the a slight UTS drop were observed in most of the conditions, a datum point with an optimum annealing condition of 473 K for 20 min were detected in HSHD region at strain rate of 10 -4 s -1 . 124 (a) (b) (c) (d) (e) (f) Figure 5.16. The SEM fractography of tensile specimens tested at room temperature for processed by 5 turns of HPT (a) and (c) and (e) and after post –HPT annealing at 473 K for 20 minutes (b) and (d) and (e). 125 (a) (b) Figure 5.17. Variation of the normalized UTS with the normalized elongation to failure of samples undergone post-HPT annealing and pulled into failure at room temperature at strain rate of (a) 1.0 × 10 -3 s -1 and (b) 1.0 × 10 -4 s -1 . 473 K, 10 min 473 K, 20 min 523 K, 10 min 523 K, 20 min 548 K, 10 min 548 K, 20 min 450 K, 2500 min Strain rate: 1.0 x 10 -3 s -1 Annealing Condition Conventional behavior High strength-High Ductility 473 K, 10 min 473 K, 20 min 523 K, 10 min 523 K, 20 min 548 K, 10 min 548 K, 20 min 450 K, 2500 min Annealing Condition Strain rate: 1.0 x 10 -4 s -1 Conventional behavior High strength-High Ductility 126 Considering the fact that the analysis showed short post-HPT annealing has few effects on the texture and misorientation angle distribution of HPT process materials, the simultaneous increase in strength and ductility while the grain size increased may be attributed to microstructural modifications by changes in distribution and morphology of precipitates. A recent study on heat treatment of ZK60 [166] showed that when the as-cast ZK60 alloy was solution treated at 400 °C for 10 h and aged at 200 °C for 15-20 h, the volume fraction of precipitates reached a peak value. This high density of the second phase precipitates was beneficial to improve both the strength and ductility. Solution treatment at 400 °C for 10 h and artificial aging at 150 °C for 30 h is considered the optimum heat treatment condition to obtain a good combination of strength and ductility. Observations on microstructural of a UFG Mg-3.4%Zn alloy processed by HPT [180] showed that there are two kinds of precipitates: one from the dynamic precipitation during HPT and the other from the ageing treatment precipitates formed from annealing treatment are observed inside of grains after 4h at 423K. TEM study of precipitates in the sample before and after heat treatment in Fig. 5.18(a) revealed that the occurrence of dynamic precipitation during deformation, these precipitates, Mg4Zn7, are mainly concentrated on the grain or sub-grain boundaries. It is apparent that annealing for 600 s leads to the continuous growth of the precipitates in Fig. 5.18(b). The size of the precipitates shifts to higher values with an average diameter of 40 nm and they are still more concentrated within the grain boundaries. As it is apparent there is significant growth from the HPT processed condition. Further long term annealing for 128 h increased the precipitates size to 45 nm and they mostly located both inside of the grains and grain boundaries with formation of precipitates free zone (PFZ) around the boundaries accompanied by depletion of solutes. Therefore in the current study on ZK60, short term post-HPT annealing at 473 K for 20 min can 127 provide an optimum condition for size and dispersion of precipitates to enhance both strength and ductility of the HPT processed material. 5.4. Shear Punch Test It is well documented for many materials that the tensile strength can be related to the effective shear strength obtained through the shear punch testing (SPT) method of thin specimens [104,105,130]. Therefore SPT method is an attractive test to provide a means for evaluating the flow properties of materials when processed with small amounts such as HPT and ECAP samples with limited dimensions. This method was initially employed for testing thin rolled materials, its application has been recently extended to magnesium alloys processed by ECAP [130]. The CRSS of magnesium alloys are sensitive to deformation temperature [141,156]. It is more pronounced for non-basal slip systems. There is a quick drop in the CRSS in latter system by increasing the temperature. Therefore the activation of non-basal slip systems at elevated temperatures improves the formability of magnesium alloys. However, regardless of temperature the most dominant deformation mode is the basal slip system of {0002}<112 ̅ 0> for magnesium. The basal slip plane is generally aligned with the deformation direction. Tensile testing introduces tensile stress on the (a) (b) (c) Figure 5.18. Micrograph of (a) dark field STEM image in the HPT deformed sample for 3 revolutions and bright field TEM images of precipitates in the sample after heat treatment at 423 K for (b) 600 s and (c) 128 h. Arrows in highlights the interaction between precipitates and grain boundaries [180]. 128 material while SPT introduces pure shear stress. Therefore it is valuable to compare the Schmid factor for the extruded and HPT processed materials with respect to the testing method. Figure 5.19 measured the Schmid factor of extruded and HPT materials for tensile and SPT test. (a) (b) (c) (d) Figure 5.19. Schmid factor distribution of (a) and (b) extruded material and (c) and (d) HPT processed materials for tensile and shear punch test respectively. It is apparent in the Fig. 5.19 that for shear punch test that the average Schmid factor is higher than tensile test which indicated that the basal slip system can operate relatively easier during SPT test along the ND direction than tensile along TD direction. Also the average Schmid factor were 0.31 and 0.29 for tensile test and 0.35 and 0.36 for shear punch test of extruded and HPT materials respectively. It showed that since there is a rotation by 90 o from prismatic fiber texture in extruded material to basal fiber texture in HPT processed material. The average of Schmid factor remained unchanged although some slight changes in Schmid factor distributions were predicted. Average number = 0.31 Average number = 0.35 Average number = 0.29 Average number = 0.36 129 A comparative analysis on the results of tensile and shear punch test in Fig. 4.31 and 4.38, respectively, showed that there is a good agreement between the flow properties obtained from different tests. Figure 4.38 exhibited a sigmoidal relationship for the HPT processed material that divided the curves into three distinct regions of flow, with m-values of ~0.30, ~0.50 and ~0.2 in regions I, II and III, respectively. A similar trend of superplastic behavior was obtained for the earlier for tensile test results in Fig. 4.31 with a sigmoidal relationship and a maximum m-value of ~0.50 in region II. Therefore, according to the similarity of the obtained results by SPT and tensile methods, it seems that SPT can be used for evaluation of flow properties of materials. This would be worthwhile especially for superplastic studies of HPT processed samples where samples are available as small thin disks. The activation energy obtained from SPT flow curves in Fig. 4.39 for the extruded and HPT processed ZK60 were 119 ± 7 and 98 ±14 kJ/mol. It is therefore a decrease in the activation energy of extruded material containing average grain size of 10 µm with a value expanded between the self-diffusion, 135 kJ/mol, and grain boundary diffusion, 92 kJ/mol [167], to the activation energy close to grain boundary diffusion after processing by HPT with an average grain size of less than 1 µm. The obtained activation energy from SPT test is in good agreement with results obtained from tensile testing with value of 89 ± 7 kJ/mol. Experimental results in superplasticity showed that the grains move over each other in the superplastic region II, therefore grain boundary sliding (GBS) is an important deformation mechanism [66]. By considering the UFG microstructure of the material and the sigmoidal dependency of SRS on strain rate with m-value close to 0.5 in the intermediate region for both tensile and shear results, the dominant deformation mechanism of the material in the superplastic region was GBS. 130 6. Summary and Conclusions In the current study samples of an Mg-Zn-Zr alloy (ZK60) were processed successfully by high-pressure torsion (HPT) for several numbers of turns at room temperature. The effect of processing by HPT has been investigated on the mechanical properties and microstructural behavior of the samples using Vickers microhardness testing, tensile testing and shear puch testing at elevated temperature, transmission electron microscopy and scanning electron microscope (SEM) equipped with electron backscatter diffraction (EBSD). Several points can be concluded from the current study on ZK60 magnesium alloy: 1. The microstructural observations show a gradual formation of ultrafine grained microstructure with an intermediate stages of grain refinement containing aggregates of fine grains surrounded by aggregates of coarse grains. The final average grain size of ~700 nm and a reasonable homogeneity achieved after 5 turns of HPT are achieved although the range of grain distribution remains in an order of magnitude. 2. EBSD analysis illustrates the evolution of grains orientation from extruded texture with prismatic {101 ̅ 0} planes lying parallel to the surface of the examined disk to an ultimate basal {0001 } fiber texture with c-axis parallel to ND after 5 turns of HPT. 3. The misorientation angle evolution during HPT processing show peaks in the range of 85-90 o . These peaks are attributed to the contribution of tensile twinning in torsional deformation of magnesium under quasi-constraint pressure of HPT. The peak detected by genesis of ultrafine grained structure at ~30 o could be attributed to the basal fiber texture and successive fragmentation of the twin bands due to prismatic slip and further continuous dynamic recrystallization within the twin bands. 131 4. ZK60 demonstrates a strain hardening behavior towards a hardness homogeneity after 5 turns of HPT with hardenability exponent, η, of ~0.07 measured for ε eq up to ~20 where the hardness values reasonably saturates at Hv  125. 5. Mechanical testing at a temperature range of 473 to 573 K reveals superplastic elongations in tensile testing of samples processed by 5 turns of HPT with a maximum elongation of about 940% at 523 K when testing with an initial strain rate of 1.0 × 10 −4 s −1 . 6. A sigmoidal relationship between flow stress and strain rate is detected, such that there is a linear relationship with maximum slope of 0.5 at intermediate strain rates. This result together with elongations to failure up to 940% confirm the superplastic behavior of the HPT-processed ZK60 alloy. Activation energy of 89 ± 7 kJ/mol obtained within the range of experiment temperature suggest that the dominant superplastic deformation mechanism for the processed magnesium alloys is grain boundary sliding accommodated by dislocation movement controlled by grain boundary diffusion. 7. The ZK60 alloy was also processed by ECAP for 4 passes and then by HPT through various total turns up to 10 revolutions. OIM analysis shows that the grain size reduces after ECAP and reduces more after a combination of ECAP and HPT. However comparing to single step SPD process by HPT, larger grain size is obtained. The grains with size of ~4.6 µm after ECAP are refined to grain size of ~2.2 µm after processing by subsequent HPT for 10 turns. 8. Larger grain sizes after combination process may be attributed to the change in morphology, size and distribution of precipitates after ECAP at 473 K where the initial rod-shaped precipitates were fragmented and overaged prior to HPT processing. 132 9. The microhardness values increase after ECAP to the value of Hv ≈ 95 and then further straining by HPT up to 10 turns, the hardness value saturates at Hv ≈ 117 with hardenability exponent, η, of ~0.03 measured for ε eq up to ~100. 10. Post-HPT short term annealing within the range of 473 to 548 K for 10 and 20 min improves the elongation to failure from value of 15% to 40%. There is an optimum condition at 473 K for 20 min which can provide high strength and high ductility ZK60 after HPT with higher elongations to failure of 1.7 times and UTS of 1.06 times higher than HPT processed material. 11. The mechanical behavior of the HPT-processed ZK60 also evaluated by shear punch test (SPT) at different temperatures and shear strain rates. The SPT results are in a very good agreement with miniature tensile testing. Therefore, The SPT could be an alternative method to assess the properties of UFG materials. 12. A sigmoidal relationship between shear yield stress and shear strain rate was detected, such that there is a linear relationship with a maximum slopes 0.5 at intermediate strain rates for SPT curves. These results are in a good agreement with the superplastic behavior of the HPT- processed ZK60 alloy obtained by miniature tensile samples. The obtained activation energy of 98 ±14 kJ/mol from shear punch test is consistent with activation energy obtained from tensile test results. It is indicative of GBS contribution as dominant deformation mechanism of the material in the superplastic region. 133 References 1. R.Z. Valiev, O.A. Kaibyshev, R.I. Kuznetsov, R.Sh. Musalimov, N.K. Tsenev, Dokl. Akad. Nauk SSSR 301(1988) 864. 2. T.G. Langdon, Acta Mater. 61 (2013) 7035. 3. R.Z. Valiev, T.G. Langdon, Prog. Mater. Sci. 51 (2006) 881. 4. A. P. Zhilyaev, T.G. Langdon, Prog. Mater. Sci. 53 (2008) 893. 5. J. Wang, Z. Horita, M. Furukawa, M. Nemoto, N.K. Tsenev, R.Z. Valiev, Y. Ma, T.G. Langdon, J. Mater. Res. 8 (1993) 2810. 6. O. Andreau, J. Gubicza, N.X. Zhang, Y. Huang, P. Jenei, T.G. Langdon, Mater. Sci. Eng. A 615 (2014) 231. 7. Z. Horita, D.J. Smith, M. Furukawa, M. Nemoto, R.Z. Valiev, T.G. Langdon, J. Mater. Res. 11 (1996) 1880. 8. Z. Horita, D.J. Smith, M. Nemoto, R.Z. Valiev, T.G. Langdon, J. Mater. Res. 13 (1998) 446. 9. R.Z Valiev, A.P. Zhilyaev, T.G. Langdon, Bulk Nanostructured Materials: Fundamentals and Applications, John Wiley and Sons, Hoboken, New Jersey (2014) pp. 7-20. 10. P.W. Bridgman, J. Appl. Phys. 14 (1943) 273. 11. A.P. Zhilyaev, B.K. Kim, G.V. Nurislamova, M.D Baró, J.A. Szpunar, T.G. Langdon, Scripta Mater. 46 (2002) 575. 12. A.P. Zhilyaev, B.K. Kim, J.A. Szpunar, M.D. Baró, T.G. Langdon, Mater. Sci. Eng. A 391 (2005) 377. 13. B. Mordike, T. Ebert, Mater. Sci. Eng. A 302 (2001) 37. 14. Y. Huang, R.B. Figueiredo, T.G. Langdon, Rev. Adv. Mater. Sci. 31 (2012) 129. 134 15. R.B. Figueiredo, T.G. Langdon, J. Mater. Sci. 44 (2009) 4758. 16. S.A. Torbati-Sarraf, T.G. Langdon, J. Alloy Compd. 613 (2014) 357. 17. S.A. Torbati-Sarraf, T.G. Langdon, Adv. Mater. Res. 922 (2014) 767. 18. Y. Harai, M. Kai, K. Kaneko, Z. Horita, T.G. Langdon, Mater. Trans. 49 (2008) 76. 19. M. Kawasaki, R.B. Figueiredo, T.G. Langdon, J. Mater. Sci. 47 (2012) 7719. 20. E.O. Hall, J. Proc. Phys. Soc. B 64 (1951) 747. 21. N.J. Petch, J. Iron Steel Inst. 174 (1953) 25. 22. M.P. Staiger, A.M. Pietak, J. Huadmai, G. Dias, Biomaterials 27 (2006) 1728. 23. F.H. Froes, D. Eliezer, E. Aghion, JOM 50(9) (1998) 30. 24. K. Johnson, Adv. Mater. Proc. 160 (2002) 62. 25. A.A. Luo, Int. Mater. Rev. 49 (2004) 13. 26. R.B. Figueiredo, T.G. Langdon, J. Mater. Sci. 45 (2010) 4827. 27. R.B. Figueiredo, T.G. Langdon, Mater. Sci. Forum 247(9) (2011) 667. 28. R.Z. Valiev, R.K. Islamgaliev, I.V. Alexandrov, Prog. Mater. Sci. 45 (2000) 103. 29. Y Saito, H. Utsunomiya, N. Tsuji, T Sakai, Acta Mater. 47 (1999) 579. 30. O.R. Valiakhmetov, R.M. Galeyev, G.A. Salishchev, Fiz. Metall. Metalloved 10 (1990) 204. 31. D.V. Orlov, V.V. Stolyarov, H.S. Salimgareyev, E.P. Soshnikova, A.V. Reshetov, Y.Y. Beygelzimer, S.G. Synkov, V.N. Varyukhinet. In: Y.T. Zhu, T.G. Langdon, R.Z. Valiev, S.L. Semiatin, D.H. Shin. T.C. Lowe. Editors. Ultrafine Grained Materials III. Warrendale, PA: TMS (2004) 457. 32. V.M. Segal, V.I. Reznikov, A.E. Drobyshevskiy, V.I. Kopylov, Russian Metall. 1 (1981) 99. 33. Y. Iwahashi, J.T. Wang, Z. Horita, M. Nemoto, T.G. Langdon, Scripta Mater. 35 (1996) 143. 34. K. Nakashima, Z. Horita, M. Nemoto, T.G. Langdon, Mater. Sci. Eng. A 281 (2000) 82. 135 35. P.B. Berbon, M. Furukawa, Z. Horita, M. Nemoto, T.G. Langdon, Metall. Mater. Trans. A 30A (1999) 1989. 36. M. Furukawa, Z. Horita, T.G. Langdon, Metals. Mater. Int. 9 (2003) 141. 37. K. Furuno, H. Akamatsu, K. Oh-ishi, M. Furukawa, Z. Horita, T.G. Langdon, Acta Mater. 52 (2004) 2497. 38. C Xu, M. Furukawa, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 398 (2005) 66. 39. M. Furukawa, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 332 (2002) 97. 40. S.E. Ion, F.J. Humphreys, S.H. White, Acta Metall. Mater. 30 (1982) 1909. 41. A. Galiyev, R. Kaibyshev, G. Gottstein, Acta Mater. 49 (2001) 1199. 42. M.M. Myshlyaev, H.J. McQueen, A. Mwembela, E. Konopleva, Mater. Sci. Eng. A 337 (2002) 121. 43. A.G. Beer, M.R. Barnett, Mater. Sci. Eng. A 423 (2006) 292. 44. K. Edalati, Z. Horita, Mater. Sci. Eng. A 652 (2016) 325. 45. C. Xu, Z. Horita, T.G. Langdon, Acta Mater. 56 (2008) 5168. 46. A.P. Zhilyaev, T.R. McNelley, T.G. Langdon, J. Mater. Sci. 42 (2007) 1517. 47. R.B. Figueiredo, P.R. Cetlin, T.G. Langdon, Mater. Sci. Eng. A 528 (2011) 8198. 48. A.P. Zhilyaev, G.V. Nurislamova, B.K. Kim, M.D. Baró, J.A. Szpunar, T.G. Langdon, Acta Mater. 51 (2003) 753. 49. N.H. Polakowski, E.J. Ripling, Strength and structure of engineering materials, (1966) Englewood cliffs, New Jersey: Prentice-Hall. 50. Y. Huang, R.B. Figueiredo, T. Baudin, F. Brisset, T.G. Langdon, Adv. Eng. Mater. 14 (2012) 1018. 51. M. Kai, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 488 (2008) 117. 136 52. M.V. Degtyarev, T.I. Chashchukhina, L.M. Voronova, L.S. Davydova, V.P. Pilyugin, Phys. Metal. Metall. 90 (2000) 604. 53. G. Sakai, K. Nakamura, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 406 (2005) 268. 54. G. Sakai G, K. Nakamura K, Z. Horita, T.G. Langdon, Mater. Sci. Forum 503-504 (2006) 391. 55. Z. Horita Z, T.G. Langdon, Scripta Mater. 58 (2008) 1029. 56. R.Z. Valiev, Y.V. Ivanisenko, E.F. Rauch, B. Baudelet, Acta Mater. 44 (1996) 4705. 57. K. Edalati, Z. Horita, Mater. Trans. 50 (2009) 92. 58. K. Edalati, A. Yamamoto, Z. Horita, T. Ishihara, Scripta Mater. 64 (2011) 880. 59. Y. Huang, R.B. Figueiredo, T. Baudin, A. Helbert, F. Brisset, T.G. Langdon, J. Mater. Sci. 47 (2012) 7796. 60. A. Vorhauer, R. Pippan, Scripta Mater. 51 (2004) 921. 61. H. Jiang, Y.T. Zhu, D.P. Butt, I.V. Alexandrov, T.C. Lowe, Mater. Sci. Eng. A 290 (2000) 128. 62. Z. Yang, U. Welzel, Mater. Lett. 59 (2005) 3406. 63. Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 410-411 (2005) 422. 64. G. Sakai, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 393 (2005) 344. 65. H.J. Lee, L. S.K. Lee, K.H. Jung, G.A. Lee, B. Ahn, M. Kawasaki, T.G. Langdon, Mater. Sci. Eng. A 630 (2015) 90. 66. T.G. Langdon, J. Mater. Sci. 44 (2009) 5998. 67. T.G. Langdon, Metall. Trans. A 13A (1982) 689. 68. K. Matsubara, Y. Miyahara, Z. Horita, T.G. Langdon, Metall. Mater. Trans. A 35A (6) (2004) 1735. 137 69. K. Matsubara, Y. Miyahara, Z. Horita, T.G. Langdon, Acta Mater. 51 (11) (2003) 3073. 70. R. Lapovok, Y. Estrin, M.V. Popov, T.G. Langdon, Adv. Eng. Mater. 10 (2008) 429. 71. R.B. Figueiredo, T.G. Langdon, Mater. Sci. Eng. A 503 (2009) 141. 72. R.B. Figueiredo, T.G. Langdon, Mater. Sci. Eng. A 430 (2006) 151. 73. R.B. Figueiredo, T.G. Langdon, Adv. Eng. Mater. 10 (2008) 37. 74. M. Kawasaki, T.G. Langdon, J. Mater. Sci. 51 (2016) 19. 75. K. Higashi, Mater. Sci. Forum 170-172 (1994) 131. 76. M. Kawasaki M, T.G. Langdon, J. Mater. Sci. 42 (2007) 1782. 77. H. Gleiter, Prog. Mater. Sci. 33 (1989) 223. 78. Y.S. Sato, Y. Kurihara, S.H.C. Park, H. Kokawa, N. Tsuji, Scripta Mater. 50 (2004) 57. 79. V.V. Stolyarov, Y.T. Zhu, T.C. Lowe, R.K. Islamgaliev, R.Z. Valiev Nanostruct. Mater. 11 (1999) 947. 80. A.P. Zhilyaev, G.V. Nurislamova, S. Surinach, M.D. Baro, T.G. Langdon, Mater. Phys. Mech. 5 (2002) 23. 81. A.P. Zhilyaev, J. Gubicza, G. Nurislamova, A. Revesz, S. Surinach, M.D. Baro, T. Ungar Phys. Status. Solidi. A 198 (2003) 263. 82. V.V. Popov, E.N. Popova, A.V. Stolbovskiy, Mater. Sci. Eng. A 539 (2012) 22. 83. J. Wongsa-Ngam, H. Wen, T.G. Langdon, Mater. Sci. Eng. A 579 (2013) 126. 84. N. Lugo, N. Llorca, J.M. Cabrera, Z. Horita, Mater. Sci. Eng. A 477 (2008) 366. 85. A.P. Zhilyaev, M.D. Baro, T.G. Langdon, T.R. McNelley Rev Adv Mater Sci 7 (2004) 41. 86. S. Sabbaghianrad, T.G. Langdon, J. Mater. Sci. 50 (2015) 4357. 87. R.Z. Valiev, I.V. Alexandrov, Y.T. Zhu, T.C. Lowe, J. Mater. Res. 17 (2002) 5. 88. T. Mungole, P. Kumar, M. Kawasaki, T.G. Langdon, J. Mater. Res. 29 (2014) 2534. 138 89. R. Valiev, Nature 419 (2002) 887. 90. T.G. Langdon (1994) Acta Metall. Mater. 42 (1994) 2437. 91. H. Van Swygenhoven, P.M. Derlet, Phys. Rev. B64 (2001) 224105. 92. K.S. Kumar, H. Van Swygenhoven, S. Suresh, Acta Mater. 51 (2003) 5743. 93. N.Q. Chinh, T. Gyori, R.Z. Valiev, P. Szommer, G. Varga, K. Havancsák, T.G. Langdon, MRS Comm. 2 (2012) 75. 94. R.Z. Valiev, A.V.Sergueeva, A.K. Mukherjee, Scripta Mater. 49 (2003) 669. 95. R. Valiev, Nat. Mater. 3 (2004) 511. 96. T. Suo, Y.L. Li, F. Zhao, Q. Deng, K. Xie, Mater. Res. Innov. 15(s1) (2011) S69. 97. I. Semenova, G. Salimgareeva, G. DaCosta, W. Lefebvre, R. Valiev, Adv. Eng. Mater. 12 (2010) 803. 98. A.V. Polyakov, I.P.Semenova, R.Z. Valiev, Y. Huang, T.G. Langdon, MRS Commun. 3 (2013) 249. 99. A.V. Polyakov, I.P.Semenova, Y. Huang, R.Z. Valiev, T.G. Langdon, Adv. Eng. Mater. 16 (2014) 1038. 100. R.B. Figueiredo, P.H.R. Pereira, M.T.P. Aguilar, P.R. Cetlin, T.G. Langdon, Acta Mater. 60 (2012) 3190. 101. K. Edalati, Z. Horita, T.G. Langdon, Scripta Mater. 60 (2009) 9. 102. M. Kawasaki, T.G. Langdon, Mater. Sci. Eng. A 498 (2008) 341. 103. Y. Zhang, S. Sabbaghianrad, H. Yang, T.D. Topping, T.G. Langdon, E.J. Lavernia, J.M. Schoenung, S.R. Nutt, Metall. Mater. Trans. 46A (2015) 5877. 104. R. Alizadeh, R. Mahmudi, Mater. Sci. Eng. A 527 (2010) 3975. 105. R. Alizadeh, R. Mahmudi, M.J. Esfandyarpour, Scripta Mater. 64 (2011) 521. 139 106. R.K. Guduru, K.A. Darling, R. Kishore, R.O. Scattergood, C.C. Koch, K.L. Murty, Mater. Sci. Eng. A 395 (2005) 307. 107. H. Yan, S.W. Xu, R.S. Chen, S. Kamado, T. Honma, E.H. Han, Scripta Mater. 64 (2011) 141. 108. M. Hakamada, A. Watazu, N. Saito, H. Iwasaki, J. Mater. Sci. 43 (2008) 2066. 109. T. Al-Samman, G. Gottstein, Mater. Sci. Eng. A 490 (2008) 411. 110. Y. Huang, R.B. Figueiredo, T. Baudin, A.L. Helbert, F. Brisset, T.G. Langdon, Mat. Res. 16 (2013) 577. 111. S.A. Torbati-Sarraf, S. Sabbaghianrad, T.G. Langdon, Lett. Mater. 5(3) (2014) 287. 112. I.C. Choi, D.H. Lee, B. Ahn, K. Durst, M. Kawasaki, T.G. Langdon, J. Jang, Scripta Mater. 94 (2015) 44. 113. L.S. Tóth, B. Beausir, C.F. Gu, Y. Estrin, N. Scheerbaum, C.H.J Davies, Acta Mater. 58 (2010) 6706. 114. L.S. Tóth, C. Gu, Mater. Charact. 92 (2014) 1. 115. J.A. Del Valle, M.T. Pérez-Prado, O.A. Ruano, Rev. Metal. Madrid 38 (2002) 353. 116. J.A. Del Valle, M.T. Pérez-Prado, O.A. Ruano, Metall. Mater. Trans. A 36A (2005) 1428. 117. J.A. Del Valle, O.A. Ruano, Acta Mater. 55 (2007) 455. 118. J.K. Mackenzie, Biometrika 45 (1958) 229. 119. J.K. Mackenzie, M.J. Thomson, Biometrika 44 (1957) 205. 120. S.A. Torbati-Sarraf, R. Mahmudi, Mater. Sci. Eng. A 627(2010) 3515. 121. H.K. Lin, J.C. Huang, T.G. Langdon, Mater. Sci. Eng. A 402 (2005) 250. 122. Y. Estrin, A. Molotnikov, C.H.J. Davies, R. Lapovok, J. Mech. Phys. Solids 56 (2008) 1186. 140 123. R.B. Figueiredo, T.G. Langdon, Mater. Sci. Eng. A 528 (2011) 4500. 124. A. Al-Zubaydi, R.B. Figueiredo, Y. Huang, T.G. Langdon, J. Mater. Sci. 48 (2013) 4661. 125. T.G. Langdon, Metal. Sci. 16 (1982) 176. 126. F.J. Humphreys, M. Hatherley, Recrystallization and Related Annealing Phenomena. Pergamon Press, Oxford, 1996, p. 363. 127. Y. Li, T.G. Langdon, Mater. Sci. Eng. A 265 (1999) 276. 128. M. Mabuchi, K. Ameyama, H. Iwasaki, K. Higashi, Acta Mater. 47(7) (1999) 2047. 129. R. Mahmudi, H. Mahjoubi, P. Mehraram, Int. J. Modern Phys. B 22 (2008) 28232. 130. R. Mahmudi, R. Alizadeh, Sh. Azhari, Mater. Lett. 97 (2013) 44. 131. M. Kawasaki, T.G. Langdon, J. Mater. Sci. 47 (2012) 7726. 132. M. Kawasaki, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 524 (2009) 143. 133. K. Neishi, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 325 (2002) 54. 134. S. Komura, Z. Horita, M. Nemoto, T.G. Langdon, J. Mater. Res. 14 (1999) 4044. 135. R.B. Figueiredo, T.G. Langdon, Mater. Sci. Eng. A 501 (2009) 105. 136. T.G. Langdon, Mater. Sci. Eng. A 462 (2007) 3. 137. M.R. Barnett, M.D. Nave, C.J. Bettles, Mater. Sci. Eng. A 386 (2004) 205. 138. C. Xu, Z. Horita, T.G. Langdon, Acta Mater. 55 (2007) 203. 139. K.D. Molodov, T. Al-Samman, D.A. Molodov, G. Gottstein, Acta Mater. 76 (2014) 314. 140. K.D. Molodov, T. Al-Samman, D.A. Molodov, G. Gottstein, Metall. Mater. Trans. A, 45 (2014) 3275. 141. R.E. Reed-Hill RE, W.D. Robertson, JOM 9 (1957) 496. 142. Y.N. Wang, J.C. Huang, Mater. Chem. Phys. 81 (2003) 11. 143. J.H. Hollomon, Trans. Am. Inst. Min. Metall. Pet. Eng. 162 (1945) 268. 141 144. J.A. Brinell, J. Iron. Steel Inst. 59 (1901) 243. 145. H. O’Neill, The hardness of metals and its measurement, Chapmanand Hall, Ltd., (1934) London, England. 146. M. Kawasaki, J. Mater. Sci. 49 (2014) 18. 147. M. Kawasaki, R.B. Figueiredo, T.G. Langdon, Acta Mater. 59 (2011) 308. 148. Y.C. Wang, T.G. Langdon, J. Mater. Sci. 48 (2013) 4646. 149. M. Kawasaki, B. Ahn, T.G. Langdon, Acta Mater. 58 (2010) 919. 150. N.X. Zhang, M. Kawasaki, Y. Huang, T.G. Langdon, J. Mater. Sci. 48 (2013) 4582. 151. M. Furukawa, Z. Horita, M. Nemoto, R.Z. Valiev, T.G. Langdon, J. Mater. Res. 11 (1996) 2128. 152. T.S. Cho, H.J. Lee, B. Ahn, M. Kawasaki, T.G. Langdon, Acta Mater. 72 (2014) 67. 153. M. Kawasaki, R.B. Figueiredo, Y. Huang, T.G. Langdon, J. Mater. Sci. 49 (2014) 6586. 154. M. Kawasaki, B. Ahn, T.G. Langdon, Mater. Sci. Eng. A 527 (2010) 7008. 155. X.G. Qiao, Y.W. Zhao, W.M. Gan, Y. Cheng, M.Y. Zheng, K. Wu, N. Gao, M.J. Starink, Mater. Sci. Eng. A 619 (2014) 95. 156. E.W. Kelly, W.F. Hosford, Trans. Metall. Soc. AIME 242(5) (1968) 654. 157. R.B. Figueiredo, T.G. Langdon, J. Mater. Sci. 41 (2008) 7366. 158. H. Ishikawa, F.A. Mohamed, T.G. Langdon, Phil. Mag. 32 (1975) 1269. 159. R. Lapovok, Y. Estrin, M.V. Popov, S. Rundell, T. Williams, J. Mater. Sci. 43 (2008) 7372. 160. Y.H. Zhao, Y.Z. Guo, Q. Wei, A.M. Dangelewicz, C. Xu, Y.T. Zhu, T.G. Langdon, Y.Z. Zhou, E.J. Lavernia, Scripta Mater. 59 (2008) 627. 161. Y.H. Zhao, Y.Z. Guo, Q. Wei, T.D. Topping, A.M. Dangelewicz, Y.T. Zhu, T.G. Langdon, E.J. Lavernia, Mater. Sci. Eng. A 525 (2009) 68. 142 162. O.B. Kulyasova, R.K. Islamgaliev, A.R. Kil’mametov, R.Z. Valiev, Phys. Met. Metallogr. 101 (2006) 585. 163. E. Aghion, B. Bronfin, Mater. Sci. Forum 350-351 (2000) 159. 164. H. Watanab, T. Mukai, M. Kohzu, S. Tanabe, K. Higashi, Mater. Trans. JIM, 40 (1999) 809. 165. X. Wu, Y. Liu, Scripta Mater. 46 (2002) 269. 166. X. Chen, X. Huang, F. Pan, A. Tang, J. Wang, D. Zhang, Trans. Nonferrous. Met. Soc. China 21 (2011) 754. 167. H. J. Frost, M. F. Ashby, Deformation-Mechanism Maps. Pergamon Press, Oxford, 1982, p.44. 168. R. Pippan, F.Wetscher, M. Hafok, A.Vorhauer, I.Sabirov, Adv. Eng. Mater. 8 (2006) 1046. 169. R. Pippan, J. Int. Mater. Res.100 (2009) 1653. 170. R. Pippan, S. Scheriau, A. Taylor, M. Hafok, A. Hohenwarter, A. Bachmaier, Ann. Rev.Mater.Res. 40 (2010) 319. 171. F.A. Mohamed, Mater. Sci. Eng. A 527 (2010) 2157. 172. F.A. Mohamed, S. S. Dheda, Mater. Sci. Eng. A 558 (2012) 59. 173. F.A.Mohamed, S.S. Dheda, Mater. Sci. Eng. A 580 (2013) 227. 174. Y. Xun, F. A. Mohamed, Mater. Sci. Eng. A 528 (2011) 5446. 175. H. Watanabe, K. Moriwaki, T. Mukai, T. Ohsuna, K. Hiraga, K. Higashi, Mater. Trans. 44 (2003) 775. 176. B. Li, S. Joshi, K. Azevedo, E. Ma, K.T. Ramesh, R.B. Figueiredo, T.G. Langdon, Mater. Sci. Eng. A 517 (2009) 24. 177. G. Mima, Y. Tanaka, Trans. Jpn. Inst. Met. 12 (1971) 71. 143 178. C. Xu, M. Furukawa, Z. Horita, T.G. Langdon, Acta Mater. 53 (2005) 749. 179. L.R.C. Malheiros, R.B. Figueiredo, T.G. Langdon, J. Mater. Res. Technol. 4 (2015) 14. 180. F. Meng, J.M. Rosalie, A. Singh, K. Tsuchiya, Mater. Sci. Eng. A 644 (2015) 386.

Abstract (if available)

Abstract A commercial extruded ZK60 magnesium alloy was processed by severe plastic deformation (SPD) techniques of high-pressure torsion (HPT) and a combination of equal-channel angular pressing (ECAP) followed by HPT. The relationship between processing, structure and properties were analyzed for this material. Post-HPT short term annealing was conducted on the HPT processed materials to improve poor ductility after HPT at room temperature. ❧ The microstructural properties were examined, microhardness measurements were recorded across the disk diameters and through the whole surfaces, and miniature tensile specimens were pulled to failure at temperatures ranging from 473 to 573 K. Shear punch tests were also conducted at identical temperatures and results were compared to tensile tests. ❧ EBSD analysis demonstrated that processing by HPT technique has a potential for producing an ultrafine-grained (UFG) structure magnesium alloy at room temperature containing reasonably equiaxed grains and majority of high-angle grain boundaries. Microstructures were refined to different levels of grain size depending on the processing operations. It was observed that the most grain refinement was achieved after processing by a single step SPD through 5 turns of HPT with a minimum average grain size of ∼700 nm. The grain refinement mechanism was consistent with nucleation of new grains along pre-existing grain boundaries and formation of bimodal structures during HPT processing and continuous evolution to homogeneous distribution of ultrafine grains. ❧ A strain hardening behavior towards hardness homogeneity detected for ZK60 processed by HPT and ECAP + HPT with a hardenability exponent, η, of 0.07 and 0.03 were measured for εeq up to ~20 and ~100, respectively, where the hardness values reasonably saturated at ~125 kgf/mm² for HPT and 117 kgf/mm² under ECAP + HPT at room temperature and at a constant pressure of 2.0 GPa and rotational speed of 1 rpm. ❧ Textural analysis at mid-radius of the un-processed and processed materials illustrated a gradual evolution from prismatic {101 ̅0} fiber of extruded materials to an ultimate basal {0001} fiber texture with c-axis parallel to ND for both HPT and ECAP + HPT processed samples. The majority of the grain boundaries have misorientation larger than 15° throughout processing which differs from that observed in face centered cubic metals. ❧ The superplastic behavior of the processed alloy was investigated through measuring strain rate sensitivity by miniature tensile testing and miniature shear punch testing (SPT). The tensile experiments were conducted at initial strain rates of 3.0 × 10⁻⁵ to 1.0 × 10⁻¹ s⁻¹ and SPT tests were conducted at punch rates in the range of 8.0 to 1.0 × 10⁻² mm/min at temperatures of 473 and 523 and 573 K. Results showed that the strain rate sensitivity index, m, has a maximum of ~0.5 in the intermediate range of strain rates, region II, in both SPT and tensile test methods and hence, SPT and tensile experimental results were in a good agreement with each other. The strain rate sensitivity index of ~0.5 and an elongation to failure of >400% with a maximum of ∼940% when testing with an intermediate initial strain rate of 1.0 × 10⁻⁴ s⁻¹ at 523 K were indicative of a superplastic deformation behavior for the material processed by HPT. ❧ Post-HPT short term annealing results showed that annealing condition for 20 min at 548 K could improve the elongation to failure of HPT process materials up to ∼2.8 times with a decrease in strength of ∼0.93 times. An optimum annealing condition also was obtained at 473 K for 20 min which could provide high strength-high ductility ZK60 after HPT with higher elongation to failure of 1.7 times and UTS of 1.06 times higher than HPT processed material.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Microstructural evolution and mechanical properties of an aluminum alloy processed by equal-channel angular pressing and high-pressure torsion

PDF

Processing of magnesium alloys with ultrafine grain structure

PDF

Microstructural evolution and mechanical properties of a copper-zirconium alloy processed by severe plastic deformation

PDF

Investigation of the influence of severe plastic deformation on the microstructure and mechanical properties of Al-7136 alloy

PDF

Mechanical behavior and deformation of nanotwinned metallic alloys

PDF

Processing, mechanical behavior and biocompatibility of ultrafine grained zirconium fabricated by accumulative roll bonding

PDF

Mechanical behavior and microstructure optimization of ultrafine-grained aluminum alloys and nanocomposites

$Using X-ray microbeam diffraction to study the long range internal stresses in plastically deformed materials$

PDF

Using X-ray microbeam diffraction to study the long range internal stresses in plastically deformed materials

PDF

Mechanical behavior of materials in extreme conditions: a focus on creep plasticity

PDF

Processing and properties of phenylethynyl-terminated PMDA-type asymmetric polyimide and composites

PDF

Chemical and mechanical deformation of porous media and materials during adsorption and fluid flow

PDF

Atomistic modeling of the mechanical properties of metallic glasses, nanoglasses, and their composites

PDF

Sustainable manufacturing of out-of-autoclave (OoA) prepreg: processing challenges

PDF

Long range internal stresses in cyclically deformed copper single crystals; and, Understanding large-strain softening of aluminum in shear at elevated temperatures

Asset Metadata

Creator Torbati Sarraf, Seyed Alireza (author), Torbatisarraf, Seyedalireza (author)

Core Title Properties of a ZK60 magnesium alloy processed by severe plastic deformation

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Mechanical Engineering

Publication Date 08/04/2016

Defense Date 04/14/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag EBSD,equal-channel angular pressing,high-pressure torsion,magnesium alloy,OAI-PMH Harvest,severe plastic deformation,superplasticity,texture,ultrafine grain materials,ZK60

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Langdon, Terence G. (committee chair)

Creator Email sarraf.alireza@gmail.com,torbatis@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-299222

Unique identifier UC11281389

Identifier etd-Torbatisar-4749.pdf (filename),usctheses-c40-299222 (legacy record id)

Legacy Identifier etd-Torbatisar-4749.pdf

Dmrecord 299222

Document Type Dissertation

Format application/pdf (imt)

Rights Torbatisarraf, Seyedalireza; Torbati Sarraf, Seyed Alireza

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA