PHYSICAL SIMULATION of FRICTION STIR PROCESSED TI-5Al-1Sn-1Zr-1V-0.8Mo

PHYSICAL SIMULATION OF FRICTION STIR PROCESSED TI-5Al-1Sn-1Zr-1V-0.8Mo

A Thesis

Presented in Partial Fulfillment of the Requirements for The Degree of Master of Science in the Graduate School of The Ohio State University

By Melissa Joanne Rubal, B.S. ********

The Ohio State University 2009

Dissertation Committee:

Approved by

Professor John Lippold, Advisor ______Doctor Mary Juhas, Advisor Advisor Professor Jim Williams Welding Engineering Graduate Program i

ABSTRACT

Friction stir processing (FSP) can be employed to modify the grain size and microstructure of a material. In titanium alloys, the refined microstructure achieved during processing can improve the mechanical properties, such as yield stress and fatigue crack initiation resistance. Documenting the microstructural evolution of Ti-5111 (5Al- 1Sn-1Zr-1V-0.8Mo) during FSP, as well as simulating the observed microstructure in a Gleeble® 3800 thermo-mechanical simulator can determine the link between strain, strain rate and temperature during processing. In this study, FSP of Ti-5111 was performed above and below the beta transus temperature allowing for investigation of the microstructural evolution in both conditions. Each processed panel was instrumented with thermocouples to record the thermal histories in the stir zone and adjacent heat-affected zone. Single sensor differential thermal analysis (SS-DTA) was used to determine the - transformation during processing. Transverse sections of the processed panels were analyzed using optical and scanning electron microscopy, electron backscatter diffraction (EBSD) and hardness mapping. FSP produced extreme grain refinement in both processing conditions – reducing the 200-500 m prior- base material grains to 1-20 m. The stir zone in the panel processed above the transus exhibited a strong transformation microtexture, governed by the Burgers orientation relationship, while the sub-transus panel displayed a shear deformation texture. Vicker’s hardness mapping revealed two distinct hardness regions: the base material and a more uniform and slightly harder stir zone. The microstructures observed in the FSP panels were simulated using hot torsion testing on a Gleeble® 3800. Ideally, the strain and strain rate data may be used to verify ii

FSP modeling programs of titanium to reduce the parameter selection phases of future friction stir projects. However, strain localization observed during hot torsion testing necessitates a different sample design for titanium alloys to take into consideration the ease of adiabatic shear band formation and the low thermal conductivity. A continuous cooling transformation (CCT) diagram was constructed for Ti-5111 to separate the effects of deformation and non-equilibrium temperature profiles on the depression of the - transformation observed during processing. Similar transformation temperatures were observed during testing, indicating the significant reduction in transformation temperature is independent of the extreme deformation.

iii

DEDICATION

Dedicated to those who love me no matter what: my parents, my sisters, Ken and Otis

ACKNOWLEDGEMENTS

I would like to thank my co-advisors, Professor John Lippold and Dr. Mary Juhas for their guidance and support throughout this project. I could not have asked for more encouraging and caring mentors. Thanks are also extended to Dr. Boian Alexandrov for assistance with thermocouples, thermal data acquisition and phase transformation analysis. His expertise was greatly appreciated. Seth Shira and Brian Thompson of the Edison Welding Institute are acknowledged for performing the friction stir processing. The funding for this project was provided by the Office of Naval Research through Julie Christodoulou and the cognizant program officers Johnnie Deloach and Richard Fonda. I am sincerely grateful for the support and direction offered by ONR. I would also like to thank Professor Jim Williams and Adam Pilchak of the Material Science and Engineering Department at Ohio State University for assistance with titanium metallurgy and electron backscatter diffraction analysis. Finally, I would like to extend my appreciation to the members of the Welding and Joining Metallurgy Group, both past and present. I value the friendships that were formed while you helped to shape me into a true researcher.

VITA

December 19, 1985 ……………………..Born - Mayfield Heights, OH. United States.

June, 2008 ………………………………B.S.W.E. The Ohio State University Columbus Ohio

2007 – Present …………………………..Graduate Research Associate The Ohio State University

FIELD OF STUDY

Major Field: Welding Engineering

TABLE OF CONTENTS Abstract ...... ii Dedication ...... iv Acknowledgements ...... v Vita ...... vi List of Figures ...... x List of Tables ...... xiv List of Abbreviations ...... xv

Chapters:

1. Introduction ...... 1

2. Background ...... 3 2.1. Titanium Metallurgy ...... 3 2.1.1. Equilibrium Phases ...... 3 2.1.2. Alloying Additions...... 4 2.1.3. Titanium Alloy Classifications ...... 5 2.1.3.1. Alloys ...... 6 2.1.3.2. + Alloys ...... 6 2.1.3.3. Alloys ...... 7 2.2. Ti-5111 ...... 7 2.3. High Temperature Deformation of Titanium Alloys ...... 8 2.4. Friction Stir Welding ...... 10 2.4.1. Background ...... 10 2.4.2. Friction Stir Processing ...... 12 2.4.3. Friction Stir Welding of Titanium Alloys ...... 13 2.4.3.1. Tooling ...... 13 2.4.3.2. Weld Regions ...... 14 2.4.3.3. FSW/FSP Studies ...... 15 2.4.4. FSW and FSP Thermal History Acquisition ...... 17 2.4.4.1. Single Sensor Differential Thermal Analysis (SS-DTA) ...... 17 2.5. Gleeble ® Simulations ...... 18 2.5.1. Background ...... 18 2.5.2. Hot Torsion Tests ...... 18 2.5.2.1. Hot Torsion Test Samples ...... 19 2.5.3. Hot Compression Tests ...... 19

3. Objectives ...... 20

vii

4. Material, Equipment and Experimental Procedures ...... 21 4.1 Material – Ti-5111 ...... 21 4.2 Equipment ...... 22 4.2.1. Optical Microscope ...... 22 4.2.2. Scanning Electron Microscopes ...... 22 4.2.3. Friction Stir Machine ...... 23 4.2.4. Gleeble 3800 System ...... 24 4.3 Experimental Procedures ...... 24 4.3.1. Friction Stir Processing ...... 24 4.3.1.1. Thermal history acquisition system ...... 24 4.3.1.2. Thermocouple Placement...... 25 4.3.1.3. Thermocouple Types and Preparation ...... 27 4.3.1.4. Processing Parameters ...... 28 4.3.2. Single Sensor Differential Thermal Analysis ...... 29 4.3.3. Characterization ...... 30 4.3.3.1. Metallography ...... 30 4.3.3.2. EBSD ...... 30 4.3.3.3. Hardness Mapping ...... 30 4.3.4. Gleeble Simulations ...... 31 4.3.4.1. Hot Torsion ...... 31 4.3.4.2. CCT Diagram ...... 34

5. Friction Stir Processing - Results and Discussion ...... 36 5.1 Macro Examination ...... 36 5.2 Thermal Histories...... 37 5.3 SS-DTA...... 44 5.4 Metallographic Analysis ...... 45 5.5 Microtexture Analysis ...... 52 5.6 Hardness Mapping ...... 55

6. Gleeble® Simulations - Results and Discussion ...... 58 6.1 Hot Torsion Simulations ...... 58 6.2 Continuous Cooling Transformation Diagram ...... 65

7. Conclusions ...... 69

8. Suggestions for Future Work ...... 71

References ...... 73

Appendix A – Technical Drawings ...... 79

Appendix B – Ti-5111 Hot Compression Testing ...... 83

viii

Appendix C – MS&T 2008 Conference Proceeding ...... 90

Appendix D – TMS 2009 Conference Proceeding ...... 99

LIST OF FIGURES

Figure Page

Figure 2.1: Unit cells of (left) and (right) phases [10] ...... 4

Figure 2.2: Schematic of a Pseudo-binary Section of a β Isomorphous Phase Diagram .... 6

Figure 2.3: Ti-6Al-4V Processing Map [18]...... 9

Figure 2.4: FSW process from the original patent [25] ...... 11

Figure 2.5: Schematic of Advancing and Retreating Sides ...... 12

Figure 2.6: Linkage of FSP Attributes [28] ...... 13

Figure 2.7: Titanium FSW/FSP regions, redrawn according to [33] ...... 14

Figure 4.1: As-received microstructure (44x magnification) ...... 22

Figure 4.2: EWI GTC Accustir machine ...... 23

Figure 4.3: Close-up view of the GTC Accustir machine during processing of Ti-5111 . 24

Figure 4.4: Thermocouple placement in configuration 1 (left) and 2 (right)...... 26

Figure 4.5: Schematic of thermocouple preparation ...... 28

Figure 4.6: Schematic of transformation start and finish as determined by SS-DTA ...... 29

Figure 4.7: Modified hot torsion specimen ...... 31

Figure 4.8: Torsion sample chamber and quenching system [7] ...... 32

Figure 4.9: An example of the - phase transformation determination from dilatometry (Run 1) ...... 35

Figure 5.1: Photomacrographs of the BT-1 (top), AT-1 (middle) and AT-2 (bottom) panels ...... 37

Figure 5.2: BT - configuration 1 thermal histories ...... 38 x

Figure 5.3: AT - configuration 1 thermal histories ...... 39

Figure 5.4: AT - configuration 2 thermal histories ...... 40

Figure 5.5: Thermocouple displacement on the retreating side of AT - Configuration 1 (hole 5) – (7.5x magnification) ...... 43

Figure 5.6: - transformation on-cooling for Hole 5, AT - configuration 1, as determined using SS-DTA ...... 44

Figure 5.7: Regions observed in the Ti-5111 FSP panel (AT- configuration 1) ...... 46

Figure 5.8: Ti-5111 BM microstructure (116x magnification) ...... 47

Figure 5.9: Ti-5111 TZ microstructure, AT - configuration 1 (116x magnification) ...... 47

Figure 5.10: Tool wear present in the BT SZ, SEM photomicrograph (BSE detector) .... 49

Figure 5.11: AT SZ microstructure, SEM photomicrograph (BSE detector) ...... 50

Figure 5.12: BT SZ microstructure, SEM photomicrograph (BSE detector) ...... 50

Figure 5.13: Photomicrograph of the fine ribs in the AT SZ (14,400x) ...... 51

Figure 5.14: IPF map with an IQ overlay of the BM, TZ and SZ in BT- configuration 1, as determined through EBSD...... 52

Figure 5.15: Equal area projection pole figures for the BT SZ, as determined through EBSD ...... 54

Figure 5.16: Equal area projection pole figures for the AT SZ, as determined through EBSD ...... 54

Figure 5.17: Ti-5111 Vicker’s hardness map -500 gram load (top) and photomicrograph of indented region (bottom) ...... 56

Figure 6.1: Resulting microstructures from the Gleeble® hot torsion test ...... 59

Figure 6.2: Optical photomicrographs of samples 3-5, tested at 940 °C, 200 RPM ...... 60

Figure 6.3: Schematic and shear strain equation associated with the scribed line method ...... 61

Figure 6.4: Photomacrograph of a tested Ti-5111 torsion specimen showing strain localization ...... 61

Figure 6.5: IPF map with an IQ overlay of the center gauge section showing an adiabatic shear band, as determined through EBSD ...... 62

Figure 6.6: SEM photomicrographs of torsion sample 3 (left) and the AT - configuration 1 SZ (right) ...... 63

Figure 6.7: Equal area project pole figures for torsion sample 3 (top) and AT-1 SZ (bottom) as determined through EBSD...... 64

Figure 6.8: Optical photomicrographs of the CCT samples and BM (4.2x magnification) ...... 67

Figure 6.9: Ti-5111 CCT Diagram ...... 68

Figure A.1: Modified Gleeble® torsion specimen [6] ...... 80

Figure A.2: FSP backing plate (dimensions given in inches) ...... 81

Figure A.3: FSP side plates (dimensions given in inches)...... 82

Figure B.1: Schematic of compression test setup ...... 84

Figure B.2: Example of indent displacement for strain determination in sample 1 ...... 86

Figure B.3: Optical photomicrographs of the hot compression test samples ...... 88

Figure C.1: Optical Photomicrographs of the Fully Lamellar BM Structure (left) and the TZ between the BM and Above-Transus SZ (right) ...... 93

Figure C.2: Backscattered SEM Photomicrographs of the Lamellar Alpha Grains of the Above-Transus SZ (left) and Fine Equiaxed Alpha Grains of the Below-Transus SZ (right) ...... 94

Figure C.3: EBSD Map of the Fully Lamellar Alpha Grains ...... 94

Figure C.4: Stir Zone Microtextures – a) Alpha Pole Figure for Above-Transus SZ, b) Beta Pole Figure for Above-Transus SZ, c) Alpha Pole Figure for Below-Transus SZ, d) Beta Pole Figure for Below-Transus SZ………………………………...... 95

Figure C.5: In-Situ Beta-to-Alpha Phase Transformation On-Cooling, Analyzed using SS-DTA...... 96 xii

Figure D.1: In situ β-to-α phase transformation on-cooling, analyzed using SS- DTA...... 103

Figure D.2: Optical photomicrographs of the lamellar BM (left) and the TZ between the BM and above-transus SZ (right) ...... 104

Figure D.3: Backscattered electron SEM photomicrographs of the lamellar α grains in the above-transus SZ (left) and equiaxed α grains in the sub-transus SZ (right) ...... 105

Figure D.4: Equal area projection pole figures describing the retreating-side SZ microtexture for the above- and below-transus FSP panels ...... 105

Figure D.5: Backscattered electron SEM photomicrographs of a torsion simulation sample (left) and the SZ of the above-transus FSP panel (right) ...... 106

xiii

LIST OF TABLES

Table Page

Table 2.1: Alloying Elements and their Effective Role in Titanium Alloys ...... 5

Table 4.1: Chemical composition of Ti-5111 (weight percent) ...... 21

Table 4.2: Thermocouple details for configuration 1 ...... 26

Table 4.3: Thermocouple details for configuration 2 ...... 27

Table 4.4: Friction stir processing parameters ...... 29

Table 4.5: Testing parameters of hot torsion specimens ...... 33

Table 5.1: Thermocouple characteristics for BT - configuration 1 ...... 41

Table 5.2: Thermocouple characteristics for AT - configuration 1 ...... 41

Table 5.3: Thermocouple characteristics for AT - configuration 2 ...... 42

Table 5.4: β-to-α transformation on-cooling, determined through SS-DTA ...... 45

Table 6.1: Results of the Ti-5111 CCT analysis ...... 65

Table B.1: Testing parameters of the hot compression samples ...... 85

Table B.2: Results of the compression strain and strain rate analyses ...... 87

Table D.1: FSP Parameters for Above and Below Transus Conditions ...... 102

xiv

LIST OF ABBREVIATIONS

A Advancing (location) AT Above-transus (processing) BCC Body centered cubic BM Base material BSE Back scattered electron BT Below-transus (processing) C Center (location) CAMM Center for the Accelerated Maturation of Materials CCT Continuous cooling transformation CP Commercially pure DRC Dynamic recovery DRX Dynamic recrystallization DSI Dynamic Systems Inc. DTA Differential thermal analysis EBSD Electron backscatter diffraction EWI Edison Welding Institute FEG Field emission gun FSP Friction stir processing FSW Friction stir welding GMAW Gas metal arc welding GTAW Gas tungsten arc welding GTC General Tool Company HAZ Heat affected zone HCP Hexagonal close packed HSLA High strength low alloy xv IPF Inverse pole figure IQ Image quality MCU Mobile conversion unit ONR Office of Naval Research PCBN Polycrystalline cubic boron nitride (tool) R Retreating (location) RPM Revolutions per minute SE Secondary electron SEM Scanning electron microscope SS-DTA Single sensor differential thermal analysis SZ Stir zone TMAZ Thermo-mechanically affected zone TZ Transition zone

xvi

CHAPTER 1

INTRODUCTION

Friction stir processing (FSP), a modification of friction stir welding (FSW), can be used to modify a material’s microstructure and grain size. The significant reduction of grain size improves mechanical properties, including increasing the yield stress and resistance to fatigue crack initiation. The lower temperature associated with friction stir welding generally leads to lower residual stresses and distortion compared to fusion welds. While most titanium alloys are weldable by fusion processes, several disadvantages are associated with melting these alloys. Fusion welding can result in a significant increase in oxygen and an increase in grain size leading to reduced ductility, as well as distortion and high residual stresses [1]. Postweld heat treatments may also be necessary following fusion welding, which requires additional time and expense [2]. Employing FSW/FSP can alleviate these problems while simultaneously providing the benefits of a refined microstructure. In titanium alloys, the refined microstructure produced by FSW/FSP is considerably affected by the phase field in which processing occurs, specifically if deformation is introduced above or below the beta transus temperature. However, current friction stir production schemes are limited by the trial and error approach in determining adequate processing parameters. This is due to a lack of experimental data on the relationship between high temperature deformation behavior and processing parameters. The vested interest of the defense agencies (the Navy and Air Force) and NASA, in friction stir processing of titanium alloys necessitates a deeper understanding of this complicated process. 1 The relationship between processing parameters, microstructural evolution, and material behavior during processing (strain, strain rate and temperature) can be used to develop and validate titanium friction stir models. An accurate model of the process will save both time and money by eliminating the trial and error approach to parameter optimization. Furthermore, a model may facilitate alloy design and heat treatments to take full advantage of friction stir processing. The focus material of this study, Ti-5111 (5Al-1Sn-1Zr-1V-.8Mo), is a near- titanium alloy with an equilibrium beta transus of approximately 980 °C [3]. The US Navy developed this alloy as a lower cost alternative for Ti-6Al-4V for structural applications in ships and submarines [4]. The desirable properties of Ti-5111 include intermediate strength in combination with excellent toughness, corrosion and room temperature creep resistance [3,5]. A transition from fusion welding to FSW of Ti-5111 will result in decreased gas shielding and fume generation as well as an increase in the structural integrity and mechanical properties of components. These advantages could result in significant cost savings for the Navy. Physical simulation of the friction stir process can help to elucidate the relationship between strain, strain rate, temperature and microstructure. The controlled testing environment provided by the Gleeble® 3800 thermo-mechanical simulator equipped with the torsion mobile conversion unit (MCU) has been used to successfully simulate the microstructural regions observed in steel FSW [6,7]. In this study, friction stir processing was performed both above and below the beta transus temperature of Ti-5111, as confirmed through microstructural examination and in situ phase transformation analysis using single sensor differential thermal analysis (SS-DTA). The stir zone simulations were then attempted with hot torsion tests using the torsion MCU and hot compression tests using the pocket jaw MCU. The high strains and strain rates imposed by the Gleeble®, as well as the steady temperature gradient, allowed for simulation of the above-transus stir zone under certain simulation conditions.

CHAPTER 2

2 BACKGROUND

2.1. Titanium Metallurgy The main properties of titanium alloys are high specific strength, low density and excellent corrosion resistance [8]. Titanium alloys are used in the aircraft industry, biomedical applications and components in chemical processing equipment, to name a few. However, the high cost of titanium alloys and their high affinity for oxygen during traditional welding practices has limited their use [8].

2.1.1. Equilibrium Phases Pure titanium is often compared to iron, in that it undergoes an allotropic phase transformation at 882 °C transforming from the body centered cubic (BCC) β phase to hexagonal close packed (HCP) phase upon cooling. In titanium alloys, this transformation temperature, the beta transus, is defined as the completion of the transformation to on heating, or inversely, the start of the transformation to on cooling [9]. The unit cells of the and phases are shown in Figure 1, including the room temperature lattice parameters [10].

Figure 2.1: Unit cells of (left) and (right) phases [10]

2.1.2. Alloying Additions Alloying elements affect the properties and phase balance of titanium alloys differently. These elements are classified by their effect on the / transformation temperature of pure titanium [10]. stabilizers increase the transformation temperature above 882 °C, while stabilizers decrease the transformation below 882 °C, widening the stable phase region. stabilizers can be further classified as isomorphous or eutectoid stabilizers, depending on the transformation from the phase field. Other elements are considered neutral because they do not have a significant effect on the / transformation temperature. Collings discussed the usefulness of classifying a multi- component titanium alloy in terms of its equivalent aluminum and molybdenum content [11]. The Rosenberg aluminum equivalent and Molchanova molybdenum equivalent are as follows, where [x] indicates the weight percent of element “x” [11]. In these equations, the potency of each element with respect to stabilization of the or phase is evident.

[Al]eq = [Al] + [Zr]/6 + [Sn]/3 + 10[O] [11]

[Mo]eq = [Mo] + [Ta]/5 + [Nb]/3.6 + [W]/2.5 + [V]/1.5 + 1.25[Cr] + 1.25[Ni] + 1.7[Mn] +1.7[Co] +2.5[Fe] [11]

Table 2.1 contains the typical alloying additions and their effective role [10].

Table 2.1: Alloying Elements and their Effective Role in Titanium Alloys

stabilizer Neutral Stabilizer Isomorphous Eutectoid Al V Fe Zr O Mo Mn Sn N Nb Cr Hf C Ta Ni Cu Si H

2.1.3. Titanium Alloy Classifications Titanium alloys can be classified into four main categories based on their alloying content and microstructure. The classifications are α, near- , α+β and β alloys [9]. Figure 2.2 is a schematic of a pseudo-binary section through a β isomorphous phase diagram [10]. Each alloy classification will be discussed further in the next sections, with near- alloys included in the + alloy section.

Figure 2.2: Schematic of a Pseudo-binary Section of a β Isomorphous Phase Diagram [10]

2.1.3.1. Alloys Alloys classified as “ alloys” include commercially pure (CP) titanium and are typically not 100% phase due to the addition of up to 2% -stabilizers. They contain a small amount, usually 2-5 volume percent phase [10]. As mentioned previously, the phase exhibits a HCP structure. alloys demonstrate the best corrosion and creep resistance and are the most weldable titanium alloys [12]. However, they exhibit low tensile strength and poor formability. Due to their primarily single phase structure, alloys will not develop better mechanical properties through heat treatments [12]. Chemical processing plants and power generation plants are typical applications for titanium alloys due to their excellent corrosion resistance.

2.1.3.2. + Alloys + alloys have compositions from the phase boundary to the intersection of the Ms, or martensite start, line with room temperature (see Figure 2.2). Therefore, + alloys can form martensite with very fast cooling rates [10]. They typically contain 4-6% -stabilizers, as well as -stabilizers to solid solution strengthen the phase. + alloys are the most widely used group due to the “workhorse” alloy, Ti-6Al-4V. +

6 alloys can be heat treated and are typically stronger than or alloys and have the best combination of strength and ductility. + alloys also exhibit good resistance to fatigue crack growth. Application examples of + alloys include structural components of military and commercial aircraft, fan and compressor blades in gas turbine engines, and risers on offshore oil and gas rigs [13].

Near- alloys are typically considered a type of + alloys. They contain a volume percent of less than 10. Ti-5111 is an example of a near- alloy. These alloys were designed for high temperature applications and have good strength and high resistance to creep and fatigue crack growth. Please refer to section 2.2 for more information on Ti-5111.

2.1.3.3. Alloys alloys are typically all within the metastable range in Figure 2.2. alloys do not form martensite when quenched, but instead form the metastable phase, from which can be precipitated. alloys can achieve very high strengths and are the most easily formed titanium alloys due to their BCC crystal structure. Another advantage of alloys is their higher hydrogen tolerance [14]. However, alloys exhibit poor creep resistance. alloys are becoming more popular in the aerospace industry due to their higher strength compared to the alloy, Ti-6Al-4V. The Boeing 777 aircraft is a prime example of alloy applications. The landing gear, nut clips, springs and parts of the engine are all made from alloys [14]. Other applications of alloys include helicopter rotor heads, oil and gas drilling parts, and biomedical applications [14].

2.2. Ti-5111 Ti-5111 is a near- titanium alloy that was developed in a US Navy sponsored program for structural applications in ships and submarines [4]. Ti-5111 is known for its combination of intermediate strength, excellent toughness and room temperature creep resistance [5]. An excellent corrosion resistance and a very high resistance to stress corrosion cracking and hydrogen embrittlement are also exhibited [4,15].

The nominal composition of Ti-5111 is 5Al-1Sn-1Zr-1V-0.8Mo, in weight percent, with no more than 0.11 weight percent oxygen [3]. Aluminum, tin and zirconium increase the strength of the alloy through solid solution strengthening, while zirconium also allows silicide to distribute uniformly in the matrix. Vanadium and molybdenum, - stabilizers, increase the strength as well, but molybdenum also improves the toughness. Silicon at a weight percent of .06-.14 enhances the creep resistance, strength and toughness of Ti-5111 [3]. A study conducted by Baxter et al. explored the possible joining methods of Ti-5111. Both Gas Metal Arc Welding (GMAW) and Gas Tungsten Arc Welding (GTAW) were tested, as well as diffusion bonding. While GMAW and GTAW produced solid joints, these welds exhibited lower toughness than the parent material. Furthermore, the diffusion bonding was successful at 950 degrees °C for 4 hours, but this is not practical for manufacturing naval ships and submarines [4]. Fonda et al., from the Naval Research Laboratory, conducted a FSW study on Ti- 5111 [16]. The results of this study will be reviewed in section 2.4.3.3.

2.3. High Temperature Deformation of Titanium Alloys One effective way to refine the grain size of an alloy, and thus improve the strength as known from the Hall-Petch relation is through deformation [17]. The flow stresses of titanium alloys, however, are significantly affected by temperature and strain rate. Small changes in working conditions can lead to significantly different results: material instability, globularization, dynamic recovery or dynamic recrystallization, to name a few. Seshacharyulu et al. [18] created a processing map for Ti-6Al-4V with a lamellar starting structure, as shown in Figure 2.3. The map illustrates the outcomes associated with varying the strain rate and temperature during hot working [18].

Figure 2.3: Ti-6Al-4V Processing Map [18]

Globularization is defined as the breaking up of the lamellar microstructure, which has been suggested to occur when the platelets are sheared and the phase penetrates to complete the separation [19-20]. In dynamic recovery (DRC), the dislocations created during working are annihilated, while dynamic recrystallization (DRX) leads to grain nucleation and growth during the application of deformation [21]. Bruschi et al. [22] conducted hot compression tests on Ti-6Al-4V and found that approximately 95% of the plastic work introduced by significant plastic deformation was nearly instantaneously converted into heat which led to material softening. Thus, high

9 temperatures and strain rates eliminated the stabilizing effects of work hardening and caused material instability [22]. Nicolaou et al. [23] investigated the formation of cavities during hot torsion testing of Ti-6Al-4V, and found that they most frequently formed along prior grain boundaries and at triple points. It was determined through EBSD analysis that the presence of hard and soft grain orientations led to the anisotropic formation of cavities as well as local dynamic globularization [23]. Furuhara et al. [24] discovered that as the strain rate decreases and the deformation temperature increases in and + alloys, the size and fraction of recrystallized grains increases. In alloys, DRC was found to be the dominant process, with DRX occurring partly along the grain boundaries. In + alloys, however, continuous DRX dominated. It was reported that the application of DRX in titanium alloys could be useful in lowering flow stresses, improving superplasticity and obtaining fine grained microstructures [24].

2.4. Friction Stir Welding

2.4.1. Background Friction stir welding is a solid state joining processes that was developed and patented by W. M. Thomas et al. at The Welding Institute (TWI) in 1991 [25]. In friction stir welding (FSW), a pin tool harder than the work-piece is plunged into the joint and rapidly rotated. Rotational friction causes the work-pieces to plasticize and “stir” together. Figure 2.4 is an illustration of the process from the original patent.

Figure 2.4: FSW process from the original patent [25]

The process is considered self-regulating due to the fact that the rising temperature in the work-piece causes the metal to soften and the torque to decrease, which imparts less heat due to mechanical work. This phenomenon stabilizes the temperature and prevents melting of the metal [26]. Based on the direction of the tool rotation, the sides of the friction stir weld are classified as advancing or retreating. The advancing side is defined as the half-plate where the direction of rotation is the same as that of welding. The retreating side is designated as the half-plate where the rotation vector opposes the welding direction [27]. Figure 2.5 is a schematic of a FSW showing the advancing and retreating sides.

Advancing Side

Tool Tool Travel Rotation

Retreating Side

Figure 2.5: Schematic of Advancing and Retreating Sides

2.4.2. Friction Stir Processing Friction stir processing (FSP) is an adaptation of friction stir welding. FSP incorporates the advantages of FSW without joining two work-pieces. Figure 2.6 shows the different attributes of FSP and how they are interrelated [28]. FSP has very distinct advantages that can be exploited in its applications. For example, friction stir processing refines the grain size which can increase the part’s strength or resistance to fatigue crack initiation [29]. Furthermore, FSP can be used to eliminate fusion weld defects and improve weld properties [28]. It can also be used to repair castings and modify fabricated structures [28].

Figure 2.6: Linkage of FSP Attributes [28]

2.4.3. Friction Stir Welding of Titanium Alloys Since 2000, several FSW and FSP studies have been performed on titanium alloys. The most common alloy studied was the work-horse titanium alloy, Ti-6Al-4V, however, both the and alloys are represented by CP titanium and TiMet 21S, respectively. The following sections will discuss the tools utilized, the weld regions observed, and the important findings of these studies.

2.4.3.1. Tooling The FSW tool, comprised of a shoulder and a protruding pin, provides the downward force, frictional heating and mechanical stirring necessary to plastically deform and join the material. Titanium alloys require more advanced tool materials than aluminum alloys, which employ tool steels. This is due to the higher melting 13 temperatures and flow stresses of titanium alloys. Polycrystalline cubic boron nitride (PCBN), which is an effective tool material for high temperature materials such as steels and nickel-base alloys, is not suitable for titanium alloys. PCBN reacts with titanium to form titanium nitrides and borides within the weld [30]. Successful tool materials reported for joining titanium alloys include sintered TiC [1], pure tungsten or its alloys [2,16,31-33]. -juhas,31,32,33, 15-fonda].

2.4.3.2.Weld Regions Pilchak et al. aspired to standardize the terminology of titanium FSW and FSP regions by proposing the schematic of a transverse cross-section in Figure 2.7 [33].

Figure 2.7: Titanium FSW/FSP regions, redrawn according to [33]

The stir zone, SZ, is a region of fine-grained microstructure that was directly stirred by the FSW pin tool. It experiences the highest deformation and temperature during processing. The microstructure of the SZ is significantly affected by the phase field in which processing occurs, specifically if deformation is introduced above or below the beta transus temperature. The transition zone, TZ, is a region of deformed microstructure between the SZ and heat affected zone, HAZ. The deformation introduced to this region is insufficient to cause recrystallization. Another name for the TZ is the TMAZ, or thermo-mechanically

14 affected zone, because the region is influenced both by mechanical deformation from the tool and from the heat of the process [34]. The HAZ does not experience microstructural changes due to deformation, so any alteration of the structure is caused by the FSW/FSP thermal cycle. The base metal, BM, is the original work-piece and its microstructure is not affected in any way by the friction stir welding process. While it may experience a change in thermal history during welding, the temperature increase is insufficient to change the microstructure. Fonda et al did not observe a HAZ in the FSW study on Ti-5111. Only three regions were observed: the unaffected BM, the fine-grained SZ and a narrow band of deformation between the two (TZ) [16].

2.4.3.3. FSW/FSP Studies Information on titanium FSW/FSP studies will be organized according to alloy. As the bulk of experimentation has been performed on Ti-6Al-4V, those studies will be reported first. The remaining studies on CP titanium, Ti-5111 and TiMet21S will follow. Threadgill reported in 1996 that encouraging preliminary FSW results were obtained on titanium alloys [35]. Lienert described the successful FSW of 0.2 inch thick plates of Ti-6Al-4V by researchers at the Edison Welding Institute and the Air Force Research Laboratory, Wright-Patterson Air Force Base in Ohio. The welds contained no defects and exhibited 100% joint efficiency with regards to yield strength and tensile strength [36]. Juhas et al. analyzed Ti-6Al-4V FSW obtained from the Edison Welding Institute and found that the stir zone microstructure was a mix of equiaxed and colonies of laths bound by ribs [2]. The presence of small, equiaxed grains with a low dislocation density in the stir zone and TMAZ of mill-annealed and -annealed Ti-6Al-4V plates indicated the occurrence of DRX during FSW, according to Ramirez et al. [31]. Thermocouple data obtained by Li et al. revealed that temperatures could exceed 1000 °C during FSW of Ti-6Al-4V [37]. Jones et al. reported that high temperatures combined with oxygen rich environments could lead to case formation on the FSW 15 surface, thus pointing out the importance of adequate shielding [38]. Another concern associated with the high temperatures experienced during welding is the tool life. For titanium alloys, the heat generated at the tool does not easily conduct away to soften the next area to be welded, thus the tool experiences both high temperatures and high forces. Russell at al. proposed the idea of stationary shoulder FSW, where a stationary shoulder provides a sliding containment mechanism during the process, to provide better control and uniformity of the heat profile [39]. More recent studies on Ti-6Al-4V focus on FSP, as opposed to FSW. Ma et al. noted that the refinement of cast Ti-6Al-4V microstructure through FSP led to an increased yield strength and fatigue crack initiation resistance [29]. Pilchak et al. pointed out that the increased fatigue crack initiation resistance of the stirred region in combination with the superior fatigue crack growth resistance of the as-cast lamellar microstructure, make this process beneficial for aerospace applications [33]. Electron backscatter diffraction, EBSD, analysis in a follow-up study revealed a transformation texture in the phase governed by the Burger’s relationship when processed above the beta transus and a shear texture when processed sub-transus [40]. Lee et al. investigated the microstructural evolution of CP titanium FSW [1]. A considerable number of twins were found within the stir zone, and could be explained by the lack of slip systems in HCP materials. The regions observed within the weld also coincided with those of Ti-6Al-4V FSW as described by Pilchak et al. [33]. As mentioned previously, Fonda et al. analyzed FSW of Ti-5111 and did not observe a HAZ region. The only regions observed were the unaffected BM, the fine-grained SZ and the TZ between the two [16]. EBSD analysis in a subsequent study by Fonda et al. showed shear deformation and deformation twinning of the phase in Ti-5111 near the edges of the FSW, while a strong texture associated with the Burger’s orientation relationship occurred near the tool [41]. A study on TiMet 21S, a alloy, found that defect-free welds could be made with Tungsten-Rhenium tools that showed little to no wear after 30 feet of welding [42]. EBSD analysis of the welds determined that a shear texture was observed at the center of

16 the stir zone. This shear texture could be rotated to correspond to previously published BCC torsion textures [28].

2.4.4. FSW and FSP Thermal History Acquisition In situ FSW thermal histories have been successfully acquired using neutron diffraction [43] and thermocouples [6-7,44-45]. A data reduction method developed and validated by Woo et al. employed the measurement of lattice distortion by in situ neutron diffraction to determine thermal cycles in FSW of an aluminum alloy [43]. Several methods have been employed for using thermocouples to measure the thermal cycles experienced during FSW. Hoffman et al measured the thermal histories in Al 6061-T6 using type K thermocouples embedded in steel channels set directly below the stirred region [38]. Lambrakos et al utilized a different approach by drilling thermocouple holes directly into the bottom of a 6061 aluminum plate. Type K thermocouples were inserted into the holes and packed with silver conductive paste to ensure adequate thermal conductivity [39]. Norton also installed thermocouples into holes drilled directly into steel plates in his ferrous FSW study. However, he welded type K thermocouples into the holes at various depths, as well as on the top and bottom surfaces of the plates [6]. Sinfield took this idea one step further by utilizing type K welded wire and sheathed thermocouples as well as eroding thermocouples within the stir zone [7].

2.4.4.1.Single Sensor Differential Thermal Analysis (SS-DTA) Differential thermal analysis (DTA) is a technique for recording the difference in temperature between a sample and reference material as the two are heated or cooled in a controlled environment [46]. DTA is a valuable resource in determining phase transformations, but it is limited due to the necessity of a reference sample. Single sensor differential thermal analysis (SS-DTA) was developed at the Ohio State University as a method of evaluating phase transformations during welding and heat treatment. Instead of utilizing a reference sample, the SS-DTA software calculates a reference curve to fit the in situ collected thermal data. Deviations from the reference curve indicate the beginning and end of phase transformations [47-48]. The accuracy of 17

SS-DTA has been confirmed using classical DTA techniques and dilatometry [49]. SS- DTA was successfully used to determine phase transformations during FSW of steels, as reported by Sinfield [7].

2.5. Gleeble ® Simulations Gleeble® simulations allow for samples to be tested in a controlled environment with precise control of temperature, strain and strain rate. The determination of which combinations of these variables produce the microstructures observed in FSW/FSP can aid in the validation of predictive models.

2.5.1. Background The Gleeble® system was first developed by Nippes et al in the 1950’s [50]. Today, the Gleeble® is commercially produced by DSI, Inc. The Gleeble® is a high- temperature, high strain-rate machine that uses resistance heating to heat samples at rates up to 10,000 °C/sec [51]. The machine utilizes closed loop control of mechanical and thermal operations. The thermal signal is obtained from thermocouples percussion welded to the center of the sample. The feedback control is then used to ensure the sample follows the pre-programmed thermal cycle to simulate the weld [52].

2.5.2. Hot Torsion Tests The Gleeble® 3500 and 3800 models have the capability of using a torsion system Mobile Conversion Unit (MCU). The torsion MCU can achieve a maximum rotation speed of 1500 rpm and a maximum torque cell rating of 56 Nm (500in-lb), while maintaining a constant temperature gradient [53]. The temperature can be controlled either by thermocouples attached to the surface of the sample or by using a pyrometer. The thermocouples offer a reliable means of thermal control as long as they remain attached to the sample. The pyrometer does not rely on direct connection, but the accuracy may suffer due to oxidation and changes in the emissivity [54]. The strain and strain rate in the torsion specimens can be controlled by the programmed number of revolutions and RPMs. It has been shown in two ferrous alloy studies that hot torsion tests can be used to correlate strain, strain rate and temperature to

18 microstructure evolution in FSW [6-7]. These correlations can be combined in the future to aid in the creation of friction stir welding and processing models. These models have the ability to significantly reduce the parameter optimization phase of a project.

2.5.2.1.Hot Torsion Test Samples The DSI, Inc. hot torsion specimen is designed with hollow shoulders and a solid, reduced diameter gauge section. These features allow for a consistent current density throughout the sample and improved temperature uniformity in the gauge length [50]. In order to achieve the cooling rates necessary to simulate friction stir welding in the Gleeble®, a modified hot torsion test sample was designed by Norton [6]. The sample is hollow through the gauge section to allow for internal helium quenching. Refer to Appendix A, Figure A.1, for a detailed mechanical drawing of the modified torsion specimen.

2.5.3. Hot Compression Tests The Gleeble® 3800 is capable of exerting up to 20 tons of static force in compression, at displacement rates up to 2000 mm/second [55]. Forrest et al used compression testing in a Gleeble® 1500 to simulate FSW of HSLA-65 [56]. A short, cylindrical sample was tested at temperatures, strains and strain rates predicted by computational modeling. The simulated microstructures were compared with the FSW microstructures to validate both the Gleeble® simulations and the model results.

CHAPTER 3

OBJECTIVES

Determining the relationship between processing parameters, material behavior and microstructural evolution can aid in the development of accurate friction stir models. These models have the potential to eliminate the current trial and error approach in production schemes, saving both time and money. The primary objectives of this investigation are:

1. To assess the effect of above- and sub-transus friction stir processing on the microstructure and microtexture in Ti-5111

2. To determine the thermal history and associated - phase transformation during friction stir processing of Ti-5111 by using imbedded thermocouples and single sensor differential thermal analysis.

3. To simulate the microstructures observed in friction stir processing using a Gleeble® 3800 thermo-mechanical system

The strain, strain rate and temperature data with correlating microstructures will be integral in the validation of titanium friction stir processing models.

CHAPTER 4

4 MATERIAL, EQUIPMENT AND EXPERIMENTAL PROCEDURES

The following sections outline the as-received material, primary equipment and experimental procedures utilized in this study.

4.1 Material – Ti-5111 Plates of 2 inch thick, -annealed Ti-5111 were provided by the Office of Naval Research (ONR). The material came from heat H6073 and was melted, ultrasonically inspected and pickled by TIMET®. The material certification sheet quantifies the tensile strength and yield strength as 118 and 103 ksi, respectively. The “5111” designation corresponds to the nominal weight percent composition of 5Al-1Sn-1Zr-1V-0.8Mo. The detailed chemical composition of the provided material is given in Table 4.1.

Table 4.1: Chemical composition of Ti-5111 (weight percent)

Ti Al Sn Zr V Mo C Fe H N O Y Si 90.47 5.2 1.01 1.28 0.98 0.8 0.007 0.03 0.0058 0.01 0.09 0.0014 0.11

The as-received material exhibits a fully lamellar microstructure with an average prior- grain diameter of 200 to 500 μm, as determined through averaging PaxIt! optical microscopy measurements. A published equilibrium -transus temperature for Ti-5111 is 980 °C [3]. As a near- alloy, Ti-5111 contains less than 10 volume percent of , which 21 is present as ribs between the laths. Phase fraction analysis performed on the material received indicated the presence of 9.28 volume percent . Figure 4.1 is an optical photomicrograph of the as-received base material.

Figure 4.1: As-received microstructure (44x magnification)

4.2 Equipment

4.2.1. Optical Microscope Optical microscopy was performed using a Nikon Epiphot binocular microscope equipped with a PAXcam digital camera. The Pax-it! Image Management System (Version 6) was used for digital image capturing and microstructural measurements.

4.2.2. Scanning Electron Microscopes Scanning electron microscopy was performed in the Ohio State University Material Science and Engineering department at the Campus Electron Optics Facility. A Quanta200 tungsten-source scanning electron microscope (SEM) equipped with a TSL

Electron Backscatter Diffraction (EBSD) and phase identification system was used for grain orientation determination. A Sirion field-emission gun (FEG) SEM was employed for high-resolution imaging. Both the secondary electron (SE) and back scattered electron (BSE) detectors were employed.

4.2.3. Friction Stir Machine Friction stir processing was performed at the Edison Welding Institute using an Army-funded, position-controlled General Tool Company (GTC) Accustir machine. The machine is capable of performing at a maximum spindle speed of 1000 rev/min and a maximum travel speed of 40 in/min at the load capacity of 7500 lbf. The process schedule is executed through R2 Control Software™. Figure 4.2 is a photograph of the gantry-style machine with the control console.

Figure 4.2: EWI GTC Accustir machine

Figure 4.3 is a close-up view of the GTC Accustir during Ti-5111 FSP, with key components highlighted.

Figure 4.3: Close-up view of the GTC Accustir machine during processing of Ti-5111

4.2.4. Gleeble 3800 System A Gleeble® 3800 thermo-mechanical simulator, located within the Ohio State University Welding Engineering program, was utilized in this study. Tests were performed using both the Torsion Mobile Conversion Unit (MCU) and Pocket Jaw MCU.

4.3 Experimental Procedures

4.3.1. Friction Stir Processing The following sections outline the procedures employed in the preparation and execution of friction stir processing.

4.3.1.1. Thermal history acquisition system In order to determine the effect of friction stir processing on the beta transus temperature and to ensure processing both above and below the transus, the thermal data from eight channels were collected using an instruNet 100 data acquisition board. The thermal histories were recorded by the instruNet software at a 400 Hz sampling rate. Thermocouple extension cables bridged the gap between the embedded thermocouples

24 and the acquisition board, all of which were individually grounded. The acquisition system was started after the plunge step was complete and the tool began to traverse.

4.3.1.2. Thermocouple Placement The thermocouples were imbedded in the plates at various distances below the surface to capture thermal histories in the center, advancing and retreating sides of the SZ as well as in the base material. Holes were drilled from the bottom of the plate and special backing and side plates were designed to allow for the thermocouple wires to connect to the acquisition system extension cables. Refer to Appendix A, Figures A.2 and A.3, for detailed drawings of the backing and side plates, respectively. Two configurations of thermocouples were utilized. Configuration 1 measured temperatures along the centerline at various depths for the above and below transus processing parameters. Configuration 2 allowed for thermal history acquisition of the center, advancing and retreating sides at one consistent depth and was used only for FSP in the above transus condition. Figure 4.4 contains schematics of the two configurations while Tables 4.2 and 4.3 contain additional dimensions and information. All thermocouple holes measure 1.8 mm (0.071 in.) in diameter. For both configurations the thermocouples were installed no less than 3 in. (7.62 cm) after the tool plunge to allow for establishment of steady state conditions.

1 2

1 3 2 3 4 5 4 5 6 6 7 7 8 8

Figure 4.4: Thermocouple placement in configuration 1 (left) and 2 (right)

Table 4.2: Thermocouple details for configuration 1

Configuration 1 Plate Dimensions: 8 inches x 4 inches x 0.3 inches Distance from Depth Below Distance from TC # Location Plunge (in.) Surface (in.) Centerline (in.) 1 C 3.0 0.125 2 C 3.5 0.100 3 C 4.0 0.075 4 C 4.5 0.050 5 C 5.0 0.050 6 C 5.5 0.075 7 C 6.0 0.100 8 C 6.5 0.125 26

Table 4.3: Thermocouple details for configuration 2

Configuration 2 Plate Dimensions: 11 inches x 4 inches x 0.3 inches Distance from Depth Below Distance from TC # Location Plunge (in.) Surface (in.) Centerline (in.) 1 A 3.5 0.05 0.375 2 R 3.5 0.05 0.375 3 C 4.5 0.05 4 A 5.5 0.05 0.375 5 R 5.5 0.05 0.375 6 C 6.5 0.05 7 A 7.5 0.05 0.375 8 R 7.5 0.05 0.375

4.3.1.3. Thermocouple Types and Preparation Two types of wire thermocouples were used during processing: Types K and C. The type K (Chromel-Alumel) 0.01 inch diameter insulated thermocouple wires were rated for temperatures up to 1250 °C. The type C (Tungsten-5% Rhenium/ Tungsten-26% Rhenium) 0.01 inch diameter bare wire thermocouples were rated for temperatures up to 2315 °C and were chosen in case the initial trials neared the temperature limits of the Type K wires. Color coded Teflon tubing was used to identify and insulate the positive and negative leads of the type C thermocouples. A type K negative wire was used to ground each thermocouple pair. The thermocouples were prepared using an Omega capacitance discharge TL- Weld – Thermocouple and Fine Wire Welder. The tips of the positive and negative leads were threaded through a small ceramic tube for insulation and then welded together using Argon shielding to avoid oxidation. This thermocouple junction was then capacitance discharge welded into the bottom of the thermocouple hole in the FSP plate. The ground wire was welded to the FSP plate next to the thermocouple hole, as shown in Figure 4.5. Channels in the backing plate and side plates allowed the thermocouple wires to pass under the FSP plate to the acquisition system without being pinched. 27

Figure 4.5: Schematic of thermocouple preparation

4.3.1.4.Processing Parameters The Ti-5111 panels were processed to a depth of 0.1 inches with the position- controlled GTC AccuStir machine. Pin tools of a refractory-based material with different dimensions were used to achieve the necessary heat inputs for the two processing conditions: above and below the beta transus temperature. The tools were redressed prior to each processing pass. Argon shielding was employed to avoid oxidation of the panels. Initial trials were run with a trial-and-error approach to determine the parameters necessary for both processing conditions. Tool rotation and travel speed were varied independently, with the assumption that a higher tool rotation and lower travel speed would lead to higher peak temperatures during processing. SS-DTA and metallographic analysis on the initial plates led to the selection of the final processing parameters, found in Table 4.4.

Table 4.4: Friction stir processing parameters

Tool Rotation Travel Speed FSP Condition rev/min in/min mm/sec Above Transus 450 3 1.3 Below Transus 300 7 3.0

4.3.2. Single Sensor Differential Thermal Analysis The in situ collected thermal histories were analyzed using SS-DTA [47-49,57]. The software (version 5c) fit a reference curve to the data using a 3-term 2nd order polynomial on-heating and a 5-term Rosenthal equation on-cooling. A deviation between the reference curve and thermal history is caused by the heat of reaction from the α-β phase transformation. The beginning and ending of deviation (δT) indicate the start and finish temperatures of the transformation, as shown in Figure 4.6.

1 – Transformation Start 2 – Transformation Finish

Figure 4.6: Schematic of transformation start and finish as determined by SS-DTA 29

4.3.3. Characterization Characterization of the FSP panels was performed on transverse cross sections. The sections were mounted in conductive Bakelite and ground through 800 grit silicon carbide paper. A colloidal silica vibratory polisher was used to achieve the final 0.02 m finish. Kroll’s reagent (2 ml HF – 10 ml HNO3 – 88 ml H2O) was selected to etch the optical and scanning electron microscopy samples.

4.3.3.1. Metallography The base material and FSP regions were photomicrographed optically using a Nikon Epiphot binocular microscope at magnifications between 25 and 1000x. Differences in microstructure, as well as the location of the thermocouples within the stirred profile were determined optically. To resolve the finer details of the microstructure, the samples were imaged using a Quanta200 tungsten-source SEM and a Sirion FEG SEM with high-resolution capabilities. The typical operating parameters for SEM imaging include a 15 kV accelerating voltage, spot size of 5 and a working distance of 12 mm or less. Both SE and BSE detectors were utilized.

4.3.3.2. EBSD Electron backscatter diffraction (EBSD) was performed on unetched samples using a Quanta200 SEM to compare the microtextures associated with the base material and FSP regions. All EBSD scans were collected on the retreating side of the FSP panel near the plate surface. The scans were analyzed using EDAX/TSL/OIM Analysis 5 software (version 5.2). The data sets were rotated so the TD direction corresponded to the advancing side and RD to the tool travel direction. Equal area projection pole figures and inverse pole figure (IPF) maps with image quality (IQ) overlays were generated to compare the microtextures.

4.3.3.3. Hardness Mapping A hardness map was performed on a transverse section of the above-transus FSP panel using a Leco LM 100AT Microhardness tester with a 500 gram load. Over three

30 thousand Vicker’s microhardness points were mapped across the transverse section to determine the difference in hardness between the base metal and FSP regions. Contour plots of the data were generated using Igor Pro (version 6.03A2) software.

4.3.4. Gleeble Simulations Two separate tests were conducted on the Gleeble® 3800 thermo-mechanical simulator: hot torsion and continuous cooling transformation (CCT). The first test was designed to simulate the microstructure observed in the stir zone and determine the relationship between strain, strain rate, temperature and microstructure. The second test, the construction of the CCT diagram, was conducted to determine the effect of cooling rate on the - transformation. It allowed for the separation of cooling rate and deformation effects on the depression of this transformation during FSP.

4.3.4.1. Hot Torsion Hot torsion tests were performed on the Ti-5111 base material using a Gleeble® 3800 thermo-mechanical simulator equipped with a torsion system MCU. The temperature, strain and strain rate were varied to simulate the microstructure observed in the SZ of the processed panels. Modified torsion samples, designed by Norton [6], were utilized to allow internal Helium quenching to match the cooling rates experienced during FSP. Figure 4.7 shows the modified torsion sample. For a more detailed drawing, refer to Appendix A, Figure A.1.

Figure 4.7: Modified hot torsion specimen

In preparation for testing, the sample chamber was evacuated to 2.1x10-2 torr then back purged twice with Argon to avoid oxidation. Figure 4.8 illustrates the chamber and hot torsion test setup, including the quenching system [7].

Figure 4.8: Torsion sample chamber and quenching system [7]

Two type-C 0.01 inch diameter bare wire thermocouples were percussion welded to the samples using a DSI Thermocouple Welder. Omegabond® “600” high temperature chemical set cement was used to reinforce the joint between the thermocouples and the specimen. The thermocouples measured the temperature in the center of the sample as well as ¼ inch offset. The center thermocouple provided closed loop control during testing. A two color optical pyrometer was also used to measure the temperature in the center of the sample when it exceeded 650 °C.

Peak testing temperatures of 940 °C and 1000 °C were selected based on the observed peak temperatures in FSP. The samples were heated to the peak temperatures at a rate of 50 °C/second then stabilized for 5 seconds. After stabilization, the samples were rotated one revolution at various RPMs then internally and externally quenched with Helium to cool at a rate similar to FSP. Table 4.5 includes the testing parameters of the fifteen torsion test specimens.

Table 4.5: Testing parameters of hot torsion specimens

100 RPM 200 RPM 400 RPM 600 RPM 940 °C 1, 2 3, 4, 5 6 7 1000 °C 8, 9 10, 11, 12 13, 14 15

Initial trials showed that torsion testing at temperatures below the beta transus would not be possible due to the high flow stresses at these testing temperatures. The samples sheared in the center of the gauge section with no apparent recrystallization prior to breaking. It was also discovered that measurement of the strain gradient, as determined in ferrous alloys by Norton and Sinfield [6-7], would not be applicable to the Ti-5111 samples due to deformation localization. The scribed-line method used in those studies, in which the angular displacement of lines scribed across the gauge section prior to testing and after testing is converted into strain incrementally to determine a strain gradient across the sample, was not performed [6-7]. This lack of strain and strain rate data for Ti-5111 hot torsion tests will be discussed further in Section 6.1.

4.3.4.2. CCT Diagram In order to generate a CCT diagram for Ti-5111, the - transformation during cooling in the Gleeble® 3800 pocket jaw MCU was determined using SS-DTA and dilatometry. All tests were run in a 1x10-6 torr vacuum environment to prevent oxidation of the samples. One type-C 0.01 inch diameter bare wire thermocouple was percussion welded to the center of each bar-shaped, ¼ inch round diameter sample for thermal control. The samples were heated to 1000 °C at a rate of 100 °C/sec to simulate the heating experienced during FSP. After a 5 second stabilization period at 1000 °C, the samples were cooled at rates between 1 °C/sec and ~47 °C/sec. The sample cooled at 1 °C/sec was control-cooled and the transformation was determined solely through dilatometry. The remaining samples were free-cooled at rates determined by the free span, which was varied from 0.5 in. (12.7 mm) to 1.5 in. (38.1 mm), and both SS-DTA and dilatometry were utilized for - transformation determination. The details of SS-DTA were described previously in section 4.3.2. Dilatometry analysis was performed using Microsoft Excel, in which linear trend lines were fit to the curves both before and after the phase transformation. The deviation from the trend lines indicated the start and finish temperatures of the transformation. An example of this analysis technique is included in Figure 4.9.

7.00E-02

6.00E-02

5.00E-02

4.00E-02

3.00E-02

2.00E-02 Dilatometry(mm) 1.00E-02

0.00E+00 0 200 400 600 800 1000 1200 Transformation Start = 850 °C -1.00E-02 Transformation Finish = 845 °C

-2.00E-02 Temperature (C)

Figure 4.9: An example of the - phase transformation determination from dilatometry (Run 1)

The start and finish temperature results of the SS-DTA and dilatometry analyses were compared and plotted against the cooling curves. A logarithmic time scale allowed the wide range of data to be more easily compared.

CHAPTER 5

5 FRICTION STIR PROCESSING - RESULTS AND DISCUSSION

The following sections will provide and discuss the results associated with FSP of Ti-5111, namely macro examination, thermal histories, SS-DTA, metallographic analysis, microtexture examination and hardness analysis. Throughout this chapter, BT and AT are used to designate “below-transus” and “above-transus” processing conditions, respectively. Furthermore, schematics of the thermocouple hole numbers for configuration 1 and 2 can be found in Figure 4.3.

5.1 Macro Examination Ti-5111 was processed above and below the beta transus temperature. Figure 5.1 contains photomacrographs of the three panels analyzed in this study, with the tool travel direction left to right. It should be noted that the tool shoulder diameter was smaller for the BT condition, allowing for a lower heat input. All photomacrographs were taken after the flash was removed by grinding.

Figure 5.1: Photomacrographs of the BT-1 (top), AT-1 (middle) and AT-2 (bottom) panels

Flash was present in all processing attempts, and was more prevalent in the below-transus panel presumably due to the higher flow stresses at the lower processing temperature. The above-transus panels exhibited more oxidation, despite argon shielding, as a result of the higher temperatures achieved during processing The FSW study on TiMet 21S by Loftus et al. [42] also noted the presence of flash and oxidation after welding. They believe that using a concave shoulder tool, instead of a flat shoulder tool, as well as purging the weld with argon for 1 hour would eliminate both issues.

5.2 Thermal Histories The thermal histories collected during processing are provided in Figures 5.2-5.4. The thermal data for any channel that successfully collected the heating cycle is included, regardless of whether the cooling cycle could be successfully analyzed using SS-DTA.

Thermocouple characteristics, including peak temperature, depth, location and the T97 on cooling follow in Tables 5.1-5.3. The locations are abbreviated using A (advancing), R

(retreating) and C (center). The T97, or time required to cool from 900 – 700 °C, is not included for any thermocouple channel that did not reach 900 °C or those that exhibited a noisy cooling history.

BT - Configuration 1 300 RPM - 7 IPM 900 Hole 1 - Type K - d=0.125 in. (C) 800 Hole 2 - Type K - d=0.100 in. (C) Hole 4 - Type C - d=0.050 in. (C) 700 Hole 6 - Type C - d=0.075 in. (C) Hole 7 - Type K - d=0.100 in. (C) 600 Hole 8 - Type K - d=0.125 in. (C)

500

400

Temperature(C) 300

200

100

0 80 90 100 110 120 130 140 150 160 170 Time from Start of Tool Traverse (s)

Figure 5.2: BT - configuration 1 thermal histories

AT - Configuration 1 450 RPM - 3 IPM

1200 Hole 1 - Type K - d=0.125 in. (C) Hole 3 - Type C - d=0.075 in. (C) 1000 Hole 5 - Type C - d=0.050 in. (C) Hole 6 - Type C - d=0.075 in. (C) Hole 7 - Type K - d=0.100 in. (C) 800 Hole 8 - Type K - d=0.125 in. (C)

600

Temperature(C) 400

200

0 50 70 90 110 130 150 170 190 210 230 250 Time from Start of Tool Traverse (s)

Figure 5.3: AT - configuration 1 thermal histories

AT - Configuration 2 450 RPM - 3 IPM

1200 Hole 1 - Type K - d=0.050 in. (A) Hole 2 - Type K - d=0.050 in. (R) 1000 Hole 3 - Type C - d=0.050 in. (C) Hole 4 - Type K - d=0.050 in. (A) Hole 5 - Type K - d=0.050 in. (R) 800 Hole 7 - Type K - d=0.050 in. (A) Hole 8 - Type K - d=0.050 in. (R)

600

Temperature(C) 400

200

0 20 40 60 80 100 120 140 160 180 200 Time from Start of Tool Traverse (s)

Figure 5.4: AT - configuration 2 thermal histories

Table 5.1: Thermocouple characteristics for BT - configuration 1

BT - Configuration 1 Distance below Surface Peak Temperature TC # Type in. (mm) (°C) 1 K 0.125 (3.18) 686 2 K 0.100 (2.54) 650 4 C 0.050 (1.27) 884 6 C 0.075 (1.91) 845 7 K 0.100 (2.54) 697 8 K 0.125 (3.18) 680 All thermocouples along the centerline

Table 5.2: Thermocouple characteristics for AT - configuration 1

AT - Configuration 1 Distance below Surface Peak Temperature TC # Type in. (mm) (°C) 1 K 0.125 (3.18) 795 3 C 0.075 (1.91) 982 5 C 0.050 (1.27) 940 6 C 0.075 (1.91) 946 7 K 0.100 (2.54) 809 8 K 0.125 (3.18) 915 All thermocouples along the centerline

41 Table 5.3: Thermocouple characteristics for AT - configuration 2

AT - Configuration 2 Peak Temperature TC # Type Location (°C) 1 K A 946 2 K R 1023 3 C C 980 4 K A 960 5 K R 980 7 K A 975 8 K R 1001 All thermocouples 0.050 in. (1.27 mm) below the surface

The thermocouples that extended inside the SZ exhibit a double-peak in the collected thermal history. This is likely due to the shearing and displacement of the thermocouple. The peak temperatures collected by these thermocouples may be lower than what is actually experienced, but are included as a comparison for those outside the SZ. The temperatures confirm the findings of the study by Li et al. [37] that determined the peak temperature in titanium alloy FSW/FSP can exceed 1000 °C. Figure 5.5 shows the displacement of the thermocouple in the retreating side of the AT – Configuration 2 panel (hole 5). The Type K thermocouple was swept along the SZ boundary toward the outer edge of the processed zone and deposited with a different orientation than when it was welded into the plate.

Figure 5.5: Thermocouple displacement on the retreating side of AT - Configuration 1 (hole 5) – (7.5x magnification)

Eroding thermocouples, which were shown to provide reliable cooling data before and after shearing by Sinfield [7], were not used in this study due to size limitations. The smallest probe diameter for Nanmac’s pencil probe eroding thermocouple is 0.125 in. (3.18 mm) [58], which is too large to accurately measure the temperature within a 0.1 in. (2.54 mm) deep processed zone in a 0.3 in. (7.62 mm) thick plate without significantly altering the heat and material flow in that region. The AT – Configuration 2 panel seemingly exhibited higher peak temperatures at the retreating side than at the advancing side. However, this was found to be due to a slight misalignment of the tool over the thermocouple holes. The retreating side holes were closer to the SZ edge than those on the advancing side, thus a higher peak temperature was collected. It is assumed that if the tool was properly aligned, the 43 advancing side would reach a higher temperature, as would be expected based on other studies [7,59].

The calculated T97 values for both configurations of the AT processing condition ranged from 8.2 – 5.7 sec, which correspond to cooling rates of 24.4 – 35.1 °C/sec. This range of cooling rates was taken into consideration when creating the Ti-5111 CCT diagram, which can be found in Section 6.2.

5.3 SS-DTA The collected thermal cycles with smooth cooling curves were analyzed using SS- DTA. A 5-term Rosenthal equation was fit to the in situ collected data on cooling. Figure 5.6 shows the on-cooling results of SS-DTA for the thermocouple in hole 5 of the AT-configuration 1 panel.

Figure 5.6: - transformation on-cooling for Hole 5, AT - configuration 1, as determined using SS-DTA

Results of the on-cooling SS-DTA can be found in Table 5.4. The start temperature of the -to- transformation on-cooling can also be considered the non- equilibrium beta transus temperature.

Table 5.4: -to- transformation on-cooling, determined through SS-DTA

Start of β→α Finish of β→α FSP Plate TC Hole Location Peak Temp. Transformation Transformation °C °C °C 5 C 940 862 808 AT-1 8 C 915 845 752 3 C 980 889 802 4 A 960 858 788 AT-2 5 R 980 868 730 7 A 975 862 807 8 R 1001 880 782 BT-1 6 C 845 -- --

The transformation start temperatures for the AT panels range from 845 to 889 °C. As compared to a predicted equilibrium beta transus temperature of 988 °C using the titanium database in JMatPro (version 4.1), these values are depressed by approximately 100 °C. No transformation was evident in the BT panel, as was expected. In an effort to determine whether either the extreme deformation or the non- equilibrium cooling rate was responsible for the depression of the transformation temperature, a CCT diagram was constructed to separate the two effects. The results of this test yielded comparable -to- transformation temperatures and can be found in Section 6.2. The transformation on-heating could not be determined due to extreme noise in the thermocouple signals. The FSP tool was directly over the thermocouple holes during this transformation, thus the signal noise prevented SS-DTA from providing an accurate analysis of the on-heating transformation behavior.

5.4 Metallographic Analysis Metallographic analysis of transverse sections of the processed panels revealed three distinct regions: BM, TZ and SZ. This is in agreement with the study by Fonda et

45 al. [15]. This section will describe and discuss the microstructure observed in these regions, through optical and scanning electron microscopy. Figure 5.7 is a transverse cross-section of AT-configuration 1 (center photomacrograph of Figure 5.1) with the three regions labeled. The BT panels exhibited the same regions, with a slightly different processed zone profile due to the different tool dimensions.

Figure 5.7: Regions observed in the Ti-5111 FSP panel (AT- configuration 1)

As previously discussed in Section 4.1, the -annealed BM exhibits a fully lamellar structure. The TZ, which experiences both heat and deformation from processing, is characterized by distortion of the lamellae and the beginning of recrystallization. The TZ is narrow in the AT and BT conditions: varying from 25-50 m. Figures 5.8 and 5.9 are optical photomicrographs of the BM and TZ from AT – configuration 1, respectively. It should be noted that the BM and TZ for the BT condition exhibit the same microstructural characteristics, the only difference being observed in the SZ.

Figure 5.8: Ti-5111 BM microstructure (116x magnification)

Figure 5.9: Ti-5111 TZ microstructure, AT - configuration 1 (116x magnification)

Phase fraction analysis was conducted on the BM by taking high-resolution photomicrographs in a Sirion FEG SEM using the BSE detector. The phase fraction was obtained through use of an image processing and stereological toolkit under development by the CAMM (Center for the Accelerated Maturation of Materials) research center which utilizes conventional stereological algorithms. Because several photomicrographs were analyzed and the results were averaged, the area fraction was assumed to be an accurate representation of the volume fraction. The Ti-5111 BM used in this study contained 9.28 volume % , which is within the typical value of less than 10 volume % for near- alloys. Pilchak et al. [33] noted the same microstructural characteristics in the TZ of FSP Ti-6Al-4V that were present in Ti-5111. Deformed lamellae as well as recrystallized equiaxed primary grains were observed. Bands of refractory-based inclusions from tool wear were observed throughout the BT SZ, as shown in Figure 5.10. Pilchak et al. [33] also observed tool contamination in FSP of Ti-6Al-4V. Energy dispersive spectrometry (EDS) analysis was performed by Pilchak which confirmed the presence of tungsten-rich particles. One issue with tungsten-based tool wear in FSW/FSP of titanium alloys is that tungsten is a -stabilizer and the debris may cause a pocket of metastable within the SZ.

Figure 5.10: Tool wear present in the BT SZ, SEM photomicrograph (BSE detector)

The SZ for the AT and BT conditions have very different microstructures. The AT SZ is comprised of 10-20 m grains of transformed containing lamellar , with at the prior- grain boundaries. This microstructure is due to recrystallization above the beta transus temperature. The BT SZ consists of 1 m equiaxed primary with a small volume fraction of . Figures 5.11 and 5.12 are SEM photomicrographs of the AT and BT SZ, respectively, taken with the BSE detector. Similar SZ microstructures were observed in the AT and BT conditions of FSP Ti-6Al-4V by Pilchak et al. [60].

Figure 5.11: AT SZ microstructure, SEM photomicrograph (BSE detector)

Figure 5.12: BT SZ microstructure, SEM photomicrograph (BSE detector)

The change in phase fraction from the BM to the SZ could not be determined in the BT panel due to the bands of tool wear in the SZ. The refractory-based particles would interfere with accurate phase fraction analysis based on color contrast. The phase fraction in the AT SZ was not determined due to the extremely fine ribs. Figure 5.13 is a high-resolution SEM photomicrograph on an unetched transverse cross section of the AT SZ.

Figure 5.13: Photomicrograph of the fine ribs in the AT SZ (14,400x)

One issue preventing an accurate phase fraction analysis of the AT SZ is that the light ribs are visible in only one orientation in Figure 5.13. The ribs are not evident in the orientation in the lower right corner of the Figure 5.13. Furthermore, a very large number of photomicrographs at 14,400x magnification would need to be analyzed to provide an accurate representation of the fraction of and present.

5.5 Microtexture Analysis Microtextural analyses of the processed regions were performed using EBSD on a Quanta 200 tungsten-source SEM. Figure 5.14 is an IPF map of BT-configuration 1, near the surface of the retreating side with an IQ overlay. The SZ and TZ are labeled.

Figure 5.14: IPF map with an IQ overlay of the BM, TZ and SZ in BT- configuration 1, as determined through EBSD

Figure 5.13 illustrates that not only do the lamellae bend as they approach the TZ, but the orientation of the parent grain also changes. This is especially obvious on the left side of Figure 5.14, as the parent grain deforms from the 2110 (green) direction to the 1010 (blue) direction as it nears the SZ.

The microtextures present in the BT and AT SZ were compared using equal area projection pole figures, which are found in Figures 5.15 and 5.16, respectively. In these figures, the TD direction corresponds to the advancing side and RD to the tool travel direction. As noted previously, all scans were performed on the retreating side of the SZ near the plate surface. The intensity of each orientation, as shown in the bar to the right of each plot is defined as “multiples of the probability in a uniform distribution.” If each orientation was equally likely and no texture was present, the value would be “1” uniformly. Therefore, the intensity number indicates how many times more likely it is to occur in the analyzed dataset than in a random distribution; a higher intensity denotes a more textured microstructure.

Figure 5.15: Equal area projection pole figures for the BT SZ, as determined through EBSD

Figure 5.16: Equal area projection pole figures for the AT SZ, as determined through EBSD 54

The maximum intensity of the poles in the BT SZ is approximately 3.4x, which is less than half that of the AT SZ at approximately 7.2x. This is due to the BT SZ exhibiting a shear deformation texture, while the AT SZ exhibits a texture associated with the transformation of -to- on cooling. The higher intensity is due to the 12 possible variants for each orientation, as dictated by the Burgers orientation relationship [61]:

The pole figures for each processing condition are very similar, with maximum intensities of approximately 2x. The pole figures for both processing conditions exhibit shear deformation textures. The results of the EBSD analysis coincide with the findings of Pilchak et al. and Fonda et al. who both noted a strong texture associated with the Burgers orientation relationship in the AT SZ of Ti-6Al-4V and Ti-5111, respectively [36-37]. Pilchak et al. also noted a shear texture in the BT SZ of Ti-6Al-4V [36], which further substantiates the results.

5.6 Hardness Mapping Vicker’s hardness mapping was performed on a Leco LM 100AT Microhardness tester with a 500 gram load. Results of the hardness map on the AT – configuration 2 panel, as well as a photomicrograph of the indented transverse section can be found in Figure 5.17. Only two regions were evident in the hardness map: BM and SZ. This verifies the previous hardness mapping results on Ti-5111, performed by Fonda et al. [37], which also showed very similar hardness values for the regions.

Figure 5.17: Ti-5111 Vicker’s hardness map - 500 gram load (top) and photomicrograph of indented region (bottom)

56 The BM hardness ranged from 250-340 VHN due to the large, anisotropic grains. The SZ exhibited a more uniform hardness of approximately 300 VHN. It was assumed that similar results would be obtained from the BT panel, as the fine grain size in the SZ would exhibit a more uniform, slightly higher hardness than the BM.

CHAPTER 6

6.GLEEBLE® SIMULATIONS - RESULTS AND DISCUSSION

The results of the Gleeble® 3800 hot torsion simulations and CCT diagram construction for Ti-5111 are presented and discussed in the following sections.

6.1 Hot Torsion Simulations Hot torsion tests were performed using the modified Gleeble® torsion sample designed by Norton [6]. The samples were heated at 50 °C/sec to a peak temperature of 940 °C or 1000 °C, rotated one revolution at one of four RPM, then internally and externally quenched with Helium to achieve a T97 cooling rate of 8.2 seconds, which is in the range of T97 observed during FSP. The samples were sectioned longitudinally and analyzed using optical microscopy. Figure 6.1 contains photomicrographs of the 15 samples tested.

Figure 6.1: Resulting microstructures from the Gleeble® hot torsion test

Tests of identical parameters did not produce repeatable microstructures. For example, Samples 3-5 which were tested at a peak temperature of 940 °C and 200 RPM, exhibited very different microstructures. Sample 3 recrystallized, sample 4 showed evidence of lamellar distortion and sample 5 exhibited partial recrystallization. Figure 6.2 contains optical photomicrographs of samples 3-5.

Figure 6.2: Optical photomicrographs of samples 3-5, tested at 940 °C, 200 RPM

This lack of repeatability is due, in part, to the localization of deformation in the center of the gauge section. The localization caused large strain variability over the tested parameters. Another disadvantage of the deformation localization is the inability to determine the strain gradient across the sample using the scribed line method, as employed in ferrous alloy studies performed by Norton and Sinfield [6,7]. In the scribed line method, a line is scribed longitudinally across the gauge section both prior to testing (original line) and after testing (reference line). The angular displacement of the original line is measured incrementally across the gauge section and then converted to angular strain to determine the strain gradient. Figure 6.3 is a schematic of the scribed line method. The equation used to determine the shear strain is also included. 60

Figure 6.3: Schematic and shear strain equation associated with the scribed line method

Figure 6.4 is a photomacrograph of a Ti-5111 hot torsion specimen showing the strain localization in the center of the gauge section. The dotted red line indicates the location of the original line scribed across the sample prior to testing.

Figure 6.4: Photomacrograph of a tested Ti-5111 torsion specimen showing strain localization

The initial localization of strain may be influenced by the large anisotropic BM grains or the formation of adiabatic shear bands. An "adiabatic shear band" is a strain localization phenomenon attributed to plastic instability due to thermal softening during adiabatic or quasi-adiabatic deformation [62]. Once the strain becomes localized in the center of the gauge section, it weakens and continues to deform while the surrounding areas do not experience any further strain [63]. Adiabatic shear bands in Ti-6Al-4V were characterized by narrow bands of extremely fine grains with poorly defined boundaries and the presence of voids or cracks [62]. Figure 6.5 is an inverse pole figure (IPF) map with an image quality (IQ) overlay of the center gauge section of a torsion specimen, as determined through EBSD analysis. A narrow adiabatic shear band with extremely fine grains and voids is present in the center of the center of the image, as indicated by the white arrows.

Figure 6.5: IPF map with an IQ overlay of the center gauge section showing an adiabatic shear band, as determined through EBSD

Despite the inability to measure the strain present in the torsion specimens, one sample did simulate the microstructure observed in the AT SZ. Sample 3, which was tested at a peak temperature of 940 °C and revolved at 200 RPM exhibited a lamellar structure with a transformed grain size of 10-20 m and at the prior grain boundaries. Figure 6.6 contains SEM photomicrographs of sample 3 (left) and the AT – configuration 1 SZ (right) using the BSE detector.

Figure 6.6: SEM photomicrographs of torsion sample 3 (left) and the AT - configuration 1 SZ (right)

The only microstructural difference between sample 3 and the AT SZ is the finer lamellae and grain boundary in the torsion specimen. The control thermocouple in sample 3 detached during the application of torsion, thus the sample was cooled faster than the cooling profile observed in FSP. EBSD analysis was performed on the torsion sample 3. Figure 6.7 contains the equal area projection pole figures for the center of the gauge section in sample 3 (top) as compared to the AT- Configuration 1 SZ. In the torsion sample pole figures, TD corresponds to the torsion axis.

Figure 6.7: Equal area project pole figures for torsion sample 3 (top) and AT-1 SZ (bottom) as determined through EBSD

The maximum intensity of sample 3 is approximately 4x, which is lower than that observed in the AT SZ. Both exhibit transformation textures, due to the peak temperature exceeding the beta transus temperature, with simple shear characteristics. However, the stronger intensity and increased texture displayed in the AT SZ points to a higher level of strain experienced during processing.

6.2 Continuous Cooling Transformation Diagram A CCT diagram was constructed for Ti-5111 using the Gleeble® 3800 pocket jaw MCU. Samples were heated to a peak temperature of 1000 °C, at 100 °C/sec, stabilized for 5 seconds, then cooled at various rates. Samples 1-5 were tested twice, corresponding to runs 1-10. These samples were free-cooled at rates dependent on the sample free-span, and both SS-DTA and dilatometry were used to analyze the - phase transformation. Sample 6 was run at a slow, controlled cooling rate of 1 °C/sec, and analyzed by dilatometry only. Table 6.1 contains the results of the SS-DTA and dilatometry analysis. The cooling rate, in °C/sec, is applicable to the range of 900-700 °C.

Table 6.1: Results of the Ti-5111 CCT analysis

Cooling Rate SS-DTA Dilatometry Analysis

Sample Run T97 (sec) °C/sec Start (°C) Finish (°C) Start (°C) Finish (°C) 3 6 4.28 46.73 846 819 3 5 4.34 46.08 848 827 848 830 4 8 6.28 31.85 838 827 4 7 6.49 30.82 839 828 2 4 9.75 20.51 846 833 2 3 9.9 20.2 846 845 846 842 5 10 14.06 14.22 846 840 5 9 14.43 13.86 850 841 1 2 20.24 9.88 848 846 850 846 1 1 20.51 9.75 851 845 850 845 6 11 200 1 883 853

The start temperature indicates the beginning of the diffusional transformation of to on cooling. The finish occurs when this transformation is exhausted. It should be noted, however, that the phase does not completely transform to ; a small volume percent remains at room temperature.

Several runs could not be analyzed using dilatometry. This may be due to slipping of the contact bar on the surface of the sample or compression of the surface during the sample’s first run. The SS-DTA and dilatometry analysis results are precise within the tolerance range of the thermocouples: ~ 8 °C [64]. The same strong depression of the - transformation that occurred during processing is evident in these results, indicating that the non-equilibrium temperature profile, not the deformation experienced, is responsible for this phenomenon. Figure 6.8 contains optical photomicrographs of the 6 samples tested, as well as the BM for comparison. The cooling rate (in the range of 900-700 °C) is included. It should be noted that the faster a sample was cooled, the finer and less distinct the lamellar microstructure appears. This is due to the diffusional -to- transformation being suppressed. The and stabilizers do not have enough time to fully partition to their respective phases. Figure 6.9 is the completed CCT diagram for Ti-5111. The x-axis has a logarithmic scale to better display the data. The outlying data from Run 3 was excluded from this diagram, and the trend lines were drawn in by hand as a representation of the transformation start and finish temperatures. The CCT diagram is a useful graphical representation of the effect of cooling rate on the - transformation.

Figure 6.8: Optical photomicrographs of the CCT samples and BM (4.2x magnification)

Figure 6.9: Ti-5111 CCT Diagram 68

CHAPTER 7

CONCLUSIONS

Friction Stir Processing 1. Ti-5111 was processed both above and below the beta transus, with centerline peak temperatures of 982 and 884 °C, respectively.

2. The on-cooling -to- transformation, or non-equilibrium beta transus temperature, was depressed by more than 100 °C relative to equilibrium, as calculated using JMatPro.

3. Three regions were observed in transverse sections of the processed panels: base material, transition zone and stir zone.

4. The above-transus stir zone exhibited a lamellar structure with prior β grain size of 10-20 μm, with along the prior grain boundaries.

5. The sub-transus stir zone was characterized by 1 μm equiaxed α grains, and contained bands of tool debris.

6. The above-transus stir zone exhibited a moderate intensity, transformed-α microtexture, governed by the Burgers orientation relationship, and a deformed-β microtexture with shear characteristics.

69 7. The sub-transus stir zone exhibited shear deformation textures for both the α and β phases.

8. Vicker’s microhardness mapping showed a hardness difference between two regions: the 250-340 VHN base material and a more uniform, slightly harder 300 VHN stir zone.

Gleeble® 3800 Simulations 9. Hot torsion testing produced a matching microstructure and microtexture to the above-transus stir zone, but the results were not repeatable.

10. Strain localization during hot torsion testing prevented the measurement of a strain gradient across the sample.

11. The hot torsion sample should be redesigned for titanium alloys to take into consideration the low thermal conductivity and ease in adiabatic shear band formation of these alloys.

12. A continuous cooling transformation diagram revealed that the depression of the -to- transformation observed during processing is caused by the non- equilibrium temperature profiles, not the extreme deformation.

CHAPTER 8

SUGGESTIONS FOR FUTURE WORK

While Ti-5111 was processed both above and below the beta transus temperature, the results were not optimal. Oxidation and flash were present in all panels. Furthermore, the high flow stresses experienced during sub-transus processing caused a non-uniform weld surface and bands of tool debris within the stir zone. It is recommended that parameter optimization studies be performed for both conditions. Determining the correct shielding gas flow may prevent oxidation in the above transus panels. Furthermore, pre-heating the plate could help to lower the flow stress in the below transus condition and lead to a more uniform processed surface. The tool material and design, as well as parameter selection could potentially eliminate the presence of flash in the processed zones. It is also suggested that multiple-pass processing be performed to explore the effects of re-processing the edges of the previous passes on both the microstructure and microtexture. The low thermal conductivity of titanium alloys may cause pre-heating of the plate in multiple-pass FSP, leading to different microstructures and properties than those observed in single-pass FSP. The relationship between arc welding and FSP should also be explored. In application, FSP may be performed over existing arc welds to improve the properties. Conversely, repair welding of processed structures may be necessary, thus knowledge of the interactions between arc welding and FSP would be beneficial. It is suggested that both single-pass and multiple-pass GTA weld and FSP interaction studies be performed.

71 With regard to the physical simulation of Ti-5111 FSP, the modified Gleeble® torsion sample developed for steels was not suitable for this alloy. A suggestion for future work is to redesign the torsion specimen to take into account the characteristics of titanium alloys that may have caused the issues encountered: poor thermal conductivity and the ease of adiabatic shear band formation. A solid sample may help to eliminate the material instability experienced during the application of torsion. However, with the elimination of internal Helium quenching, the sample length may need to be reduced and copper grips employed to simulate the cooling rates experienced during FSP. A sample designed specifically for titanium alloys may provide more repeatable results and determine the necessary strain, strain rate and temperature relationships for validation of FSW/FSP models. Finally, hot ductility testing of Ti-5111 and Ti-6Al-4V is suggested. Understanding the elevated temperature behavior of both alloys may help to elucidate the differences in FSP qualities.

REFERENCES

1. Lee, W.-B., Lee, C.-Y., Chang, W.-S., Yeon, Y.-M., & Jung, S.-B. (2005). Microstructural investigation of friction stir welded pure titanium. Materials Letters. 59 (26), 3315.

2. Juhas, M. C., Viswanathan, G. B., and Fraser, H. L. Microstructural evolution in Ti Alloy Friction Stir Welds. (Paper presented at the 2nd International Symposium on Friction Stir Welding, Gothenburg, Sweden, 27-29 June 2000).

3. Czyryca, E. J., Wells, M. E., & Tran, K. Titanium Alloy Ti-5111 for Naval Applications. (Paper presented at the RINA International Conference - Advanced Marine Materials: Technology and Applications, London, 9-10 October 2003), 41-49.

4. Baxter, D., & Wallis, I. Joining of Titanium Alloy 5111. (Paper presented at the 10th World Conference on Titanium, Hamburg, 13-18 July 2003), 651-658.

5. Been, J., & Faller, K. (1999). Titanium: History, Processing, Applications - Using Ti- 5111 for Marine Fastener Applications. JOM : the Journal of the Minerals, Metals & Materials Society. 51 (6), 21.

6. Norton, S. J. “Ferrous Friction Stir Weld Physical Simulation” (Ph.D. Thesis, The Ohio State University, 2006).

7. Sinfield, M. F. “Advancements in Physical Simulation and Thermal History Acquisition Techniques for Ferrous Alloy Friction Stir Welding” (M.S. Thesis, The Ohio State University, 2007).

8. Lütjering, G., & Williams, J. C. (2003). Titanium. Berlin: Springer. (Chapter 1)

9. Boyer, H. E., & Gall, T. L. (1984). Metals handbook. Metals Park, Ohio: American Society for Metals. (Page 9-2)

10. Lütjering, G., & Williams, J. C. (2003). Titanium. Berlin: Springer. (Chapter 2)

11. Welsch, G., Boyer, R., & Collings, E. W. (1994). Materials properties handbook: Titanium alloys. Materials Park, OH: ASM International.

12. Donachie, M. J. (1982). Titanium and titanium alloys. Metals Park, Ohio: American Society for Metals.

13. Lütjering, G., & Williams, J. C. (2003). Titanium. Berlin: Springer. (Chapter 5)

14. Lütjering, G., & Williams, J. C. (2003). Titanium. Berlin: Springer. (Chapter 7)

15. Been, J., Davis, D., & Aylor, D. Corrosion Property Evaluation of Ti 5111 in Marine Environments. (Paper presented at the National Association of Corrosion Engineers Annual Conference, Lviv, Ukraine, 6-9 June 2000)

16. Fonda, R. W., Knipling, K. E., & Moon, D. W. (2007). Microstructural Evolution in Ti 5-1-1-1 Friction Stir Welds. In R. S. Mishra, M. W. Mahoney, T. Lienert, K. V. Jata (Eds.) Friction Stir Welding and Processing IV (pp. 295-302). Warrendale, PA: TMS.

17. Huang, J. S., Huang, L., Liu, B., Zhang, Y. h., Zhang, W., He, X. Y., et al. (2007). Simulation of hot compression of Ti-Al alloy. Intermetallics. 15 (5-6), 700-705.

18. Seshacharyulu, T., Medeiros, S. C., Frazier, W. G., & Prasad, Y. V. (2002). Microstructural mechanisms during hot working of commercial grade Ti-6Al-4V with lamellar starting structure. Materials Science and Engineering A -Lausanne-. 325 (1-2), 112 - 125.

19. Margolin, H., and Cohen, P. (1980) Titanium 80: Science and Technology, ed. H. Kimura and O. Izumi (Warrendale, PA: TMS, 1980), 1555–1561.

20. Weiss, I., Froes, F. H., Eylon, D., and Welsch, G. E. (1986) Modification of Alpha Morphology in Ti-6Al-4V by Thermomechanical Processing. Metallurgical and Materials Transactions A. 17 (11), 1935-1947.

21. Nes, E., Marthinsen, K., & Brechet, Y. (2002). On the mechanisms of dynamic recovery. Scripta Materialia. 47 (9), 607-611.

22. Bruschi, S., Poggio, S., Quadrini, F., & Tata, M. E. (2004). Workability of Ti-6Al-4V alloy at high temperatures and strain rates. Materials Letters. 58 (27-28), 3622-3629.

23. Nicolaou, P. D., Miller, J. D., & Semiatin, S. L. (2005). Materials Processing - Cavitation during Hot-Torsion Testing of Ti-6Al-4V. Metallurgical and Materials Transactions. A, Physical Metallurgy and Materials Science. 36 (12), 3461.

24. Furuhara, T., Poorganji, B., Abe, H., & Maki, T. (2007). Dynamic recovery and recrystallization in titanium alloys by hot deformation. JOM. 59 (1), 64-67.

25. Thomas, W. M., Nicholas, E. D., Needham, J. C., Murch, M. G., Temple-Smith, P., and Dawes, C. J. (1991) Friction Stir Butt Welding. International Patent No. PCT/GB92/02203.

26. Mishra, R. S., & Mahoney, M. W. (2007). Friction stir welding and processing. Materials Park, Ohio: ASM International. (Page 39)

27. Nandan, R., DebRoy, T., & Bhadeshia, H. K. (2008). Recent advances in friction-stir welding - Process, weldment structure and properties. Progress in Materials Science. 53 (6), 980-1023.

28. Mishra, R. S., & Mahoney, M. W. (2007). Friction stir welding and processing. Materials Park, Ohio: ASM International. (Page 130)

29. Ma, Z. Y., Pilchak, A. L., Juhas, M. C., & Williams, J. C. (2008). Microstructural refinement and property enhancement of cast light alloys via friction stir processing. Scripta Materialia. 58 (5), 361-366.

30. Zhang, Y., Sato, Y. S., Kokawa, H., Park, S. H., & Hirano, S. (2008). Stir zone microstructure of commercial purity titanium friction stir welded using pcBN tool. Materials Science and Engineering A. 488 (1-2), 25-30.

31. Ramirez, A. J., & Juhas, M. C. (2003). Microstructural Evolution in Ti-6Al-4V Friction Stir Welds. Materials Science Forum. 426/432, 2999-3004.

32. Reynolds, A. P., Hood, E., & Tang, W. (2005). Texture in friction stir welds of Timetal 21S. Scripta Materialia. 52 (6), 491-494.

33. Pilchak, A. L., Juhas, M. C., & Williams, J. C. (2007). Microstructural Changes Due to Friction Stir Processing of Investment-Cast Ti-6Al-4V. Metallurgical and Materials Transactions. A, Physical Metallurgy and Materials Science. 38 (2), 401.

34. Microstructure Classification of Friction Stir Welds. (2007). Retrieved May 10, 2009, from TWI Web site: http://www.twi.co.uk/content/fswqual.html

35. Threadgill P L. (1996) Solid state joining of titanium alloys. Titanium World, 65-67.

36. Lienert, T. J. (2002). Friction Stir Welding Developed for Titanium Alloys. Advanced Materials and Processes.

37. Li, T., Trapp, T., Stotler, T., Coffee, J. Friction Stir Welding of Ti-6Al-4V Alloy. (Paper presented at AeroMat 2006, Seattle, Washington, 15-18 May 2006)

38. Jones, R. E., & Loftus, Z. S. (2006). Friction Stir Welding of 5mm Titanium 6Al-4V. Materials Science and Technology – Association for Iron and Steel Technology-. 6, 119- 134.

39. Russell M.J., Blignault C., Horrex N.L., & Wiesner C.S. (2008). Recent developments in the friction stir welding of titanium alloys. Welding in the World. 52 (9- 10), 12-15.

40. A. L. Pilchak et al., “A Comparison of Friction Stir Processing of Investment Cast and Millannealed Ti-6Al-4V,” Welding in the World, 52 (9/10) (2008), Research Supplement.

41. Fonda, R.W., et al., Microstructure of Ti 5111 Friction Stir Welds. (Presentation at the 136th TMS Annual Meeting, Orlando, Florida, 25 February – 1 March 2007)

42. Loftus, Z., Takeshita, J., Reynolds, A., and Tang, W. An Overview of Friction Stir Welding Beta 21S Titanium. (Paper presented at the 5th International Conference on Friction Stir Welding, Metz, France, 14-16 September 2004)

43. Woo, W., Feng, Z., Wang, X.-L., Brown, D.W., Clausen, B., An, K., Choo, H., Hubbard, C.R., and David, S. A. (2007) In situ neutron diffraction measurements of temperature and stresses during friction stir welding of 6061-T6 aluminium alloy. Science & Technology of Welding & Joining, 12 (4), 298–303.

44. Hofmann, D. C., and Vecchio, K. S. (2007) Thermal history analysis of friction stir processed and submerged friction stir processed aluminum. Materials Science & Engineering A, 465 (1-2), 165–175.

45. Lambrakos, S. G., Fonda, R. W., Milewski, J. O., & Mitchell, J. E. (2003). Analysis of friction stir welds using thermocouple measurements. Science and Technology of Welding and Joining. 8, 385-390.

46. Bhadeshia, H. K. D. H. Thermal Analysis Techniques. Retrieved May 11, 2009, from University of Cambridge, Materials Science & Metallurgy Web site: http://www.msm.cam.ac.uk/phasetrans/2002/Thermal1.pdf

47. Alexandrov, B. T. and Lippold, J.C. Methodology for In Situ Investigation of Phase Transformations in Welded Joints. (IIW. Doc. IX-2114-04, 57th Annual Assembly of IIW, Osaka, Japan, July 2004).

48. Alexandrov, B. T., et al., A New Methodology for Studying Phase Transformations in High Strength Steel Weld Metal. (Paper presented at the 7th International Trends in Welding Research Conference, Pine Mountain, Georgia, 16-20 May 2005).

49. Alexandrov, B. T., et al., (2007) Single Sensor Differential Thermal Analysis of Phase Transformations and Structural Changes during Welding and Postweld Heat Treatment. Welding In the World. 51 (11/12), 48-59.

50. Nippes, E. F., Savage, W. F., Bastian, B. J., Mason, H. F. and Curran, R. M., (1955) An Investigation of the Hot Ductility of High Temperature Alloys. Welding Journal. 183s-196s.

51. Nicolaou, P. D., Bailey, R. E. and Semiatin, S. L. (2003) Hot-Tension Testing. Handbook of Workability and Process Design. Materials Park, Oh: ASM International, 68-70.

52. Campbell, R. D. and Walsh, D. W., (1994) Gleeble Testing. Retrieved July 16, 2007, from the ASM Handbook Online, Volume 6 Web site: http://products.asminternational. org/hbk/do/highlight/content/V06/D12/A01/S0033146.html

53. Gleeble® Application Note 020 Hot Torsion Testing on a Gleeble System. Retrieved April 13, 2009 from the DSI Web site: http://www.gleeble.com/products/appnote020.htm

54. Operating Instructions for Hot Torsion Testing. (2001) Gleeble 3500/3800 Options Reference Manual. (2001) DSI. 10-5.

55. Gleeble 3800 System. Retrieved April 13, 2009 from the DSI Web site: http://www.gleeble.com/3800.htm

56. Forrest, D. R., Nguyen, J. P., Posada, M., and DeLoach, J. J. Simulation of HSLA- 65 Friction Stir Welding. (Paper presented at the 7th International Trends in Welding Research Conference, Pine Mountain, Georgia, 16-20 May 2005).

57. Alexandrov, B. T., Lippold, J. C., and Norton, S. J. (2006) Method and Device for Investigation of Phase Transformations in Metals and Alloys. United States Patent 7473028.

58. Pencil Probe “Eroding” Thermocouple – E12 Series. Retrieved January 17, 2009 from Nanmac Web site: http://nanmac.com/handbook/productindex.htm

59. Mishra, R. S., and Ma, Z. Y. (2005). Friction Stir Welding and Processing. Materials Science and Engineering R, 50, 1-78. (page 17)

60. Pilchak, A., Juhas, M., & Williams, J. (2008). IIW-1888-08 (ex-doc. III-1435r1-07) A comparison of friction stir processing of mill annealed and investment cast Ti-6Al-4V. Welding in the World. 52 (9/10), 60-68.

61. Germain L, Gey N, & Humbert M. (2007). Reliability of reconstructed beta- orientation maps in titanium alloys. Ultramicroscopy. 107 (12), 1129-35.

62. Grebe, H. A., Pak, H.-R., and Meyers, M. A. (1985). Adiabatic shear localization in titanium and Ti-6 pct Al-4 pct V alloy. Metallurgical and Materials Transactions A. 16 (5), 761-775.

63. Zener, C. and Hollomon, J. H. (1943). Effect of Strain Rate on Plastic Flow of Steel. Journal of Applied Physics. 15, 22-32.

64. Thermocouple Tolerances. Retrieved May 1, 2009 from Omega Web site: http://www.omega.com/toc_asp/frameset.html?book=Temperature&file=TC_GEN_SPE CS_REF

APPENDIX A

TECHNICAL DRAWINGS

Figure A.1: Modified Gleeble® torsion specimen [6]

Figure A.2: FSP backing plate (dimensions given in inches)

2 3/32

¼-0.01 3

0.1 3 / 1 1 / 6 2 1 / 2 1 / 2 1 / 2 1 / 2 1 1 1 / 2 1 / 2 1 / 2 1 / 2 1 / 2

Figure A.3: FSP side plates (dimensions given in inches)

APPENDIX B

TI-5111 HOT COMPRESSION TESTING

Hot compression tests were performed on Ti-5111 using the Gleeble® 3800 thermo-mechanical simulator. Compression data is more desirable for FSP modelers than hot torsion data due to the simpler applied stress fields. The following sections outline the experimental procedures and the results of the hot compression tests.

Experimental Procedures

Hot compression tests were performed on Ti-5111 base material using a Gleeble® 3800 thermo-mechanical simulator equipped with a pocket jaw mobile conversion unit (MCU). Bar-shaped, ¼ inch round diameter samples cut to a length of 2 15 /16 inches were utilized. The length of the sample was chosen so the sample ends were flush with the jaws and in contact with the C-clamp backing plate while maintaining a ½ inch free span. This loading setup and free span length ensured that the samples would not slip or buckle during the application of compression. Figure B.1 is a schematic of the sample setup.

Figure B.1: Schematic of compression test setup

1 Five indents placed /8 inch apart were made across the ½ inch sample free span prior to testing using a Wilson Rockwell manual hardness tester (150 kg load). A Nikon

SMZ1000 binocular microscope was used to take photomicrographs of the samples before and after testing; the change in distance between the indents after compression was used to determine the strain and strain rate for each sample. One type-C 0.01 inch diameter bare wire thermocouple was percussion welded to the center of each sample for thermal control. For three samples, one of each of the peak 1 temperatures tested, another type-C thermocouple was welded /8 inch away from the center to determine the thermal gradient. The initial trials included 9 samples tested at 3 different peak temperatures and 3 varying stroke rates, as demonstrated in Table B.1. The samples were heated at 100 °C/sec to the peak temperature, stabilized for 1 second, 1 compressed /8 of an inch at the specified stroke rate, then free cooled. All tests were run in a 1x10-6 torr vacuum environment to prevent oxidation of the samples.

Table B.1: Testing parameters of the hot compression samples

Compression Stroke Rate Peak 6.25 in/sec 12.5 in/sec 25 in/sec Temperature °C 158.8 mm/sec 317.5 mm/sec 635 mm/sec 850 1 2 3 950 4 5 6 1050 7 8 9

After testing, the strain and strain rate for each sample was determined. Metallographic samples were produced from axial sections mounted in conductive Bakelite and polished to a 0.02 μm finish. The samples were etched using Kroll’s reagent

(2 mL HF – 10 mL HNO3 – 88 mL H2O) and analyzed using optical microscopy.

Results and Discussion Metallographic analysis of the samples revealed that uniform compressive strain was confined to the center ¼ inch of each specimen, and this region exhibited a uniform microstructure. Because the strain occurred in the gauge section, the strain was 1 determined by measuring the change in distance between the indents placed /8 inch away from the sample center. An example of the relative displacement of indents used for strain determination in sample 1 is shown in Figure B.2.

Figure B.2: Example of indent displacement for strain determination in sample 1

The strain rate was then calculated by relating the measured strain to the time of applied compression. The strain and strain rate for each sample is found in Table B.2.

Table B.2: Results of the compression strain and strain rate analyses

Peak Temp. Strain Strain Rate Sample °C in./in. in./in./sec 1 850 0.28 14 2 850 0.33 33 3 850 0.19 37 4 950 0.33 17 5 950 0.3 30 6 950 0.2 41 7 1050 0.32 16 8 1050 0.3 30 9 1050 0.28 56

The measured strain varied between 0.19 and 0.33 for all tests. A longer free span and more crosshead displacement is necessary to achieve higher strains. However, buckling considerations limit the free span length. A sample with a higher diameter:length ratio may be beneficial in achieving higher strains without sample buckling. Figure B.3 shows photomicrographs of the 9 compression samples. Sample 1-3 were tested at a peak temperature of 850 °C and did not exceed the α-β transformation temperature. The thermal histories of samples 4-6, tested at 950 °C, and samples 7-9, tested at 1050 °C, showed evidence of the thermal effect associated with the α-β transformation on cooling.

Figure B.3: Optical photomicrographs of the hot compression test samples

Samples 1-3 exhibit lamellar crimping due to the application of compression below the α-β transformation temperature. A fine, basket-weave structure is present in samples 7-9, as a result of the peak temperature exceeding the α-β transformation temperature. Sample 4-6 display a mixed microstructure of lamellar crimping and a fine, basket-weave structure. The resulting microstructures were more similar to the transition zone in Ti-5111 FSP than the stir zone. This is likely due to the total strain being too low to cause recrystallization. As mentioned previously, a longer free span and more compression would help to achieve higher strains, which may cause recrystallization. However, the

88 sample geometry may need to be redesigned with a larger diameter, to avoid buckling at longer free spans.

APPENDIX C

MS&T 2008 CONFERENCE PROCEEDING

Rubal, M., Lippold, J., & Juhas, M. (2008). Friction Stir Processing of Ti-5111. Materials Science and Technology. 4, 2341-2348.

Friction Stir Processing of Ti-5111

M. Rubal, J. Lippold Welding Engineering, The Ohio State University, Columbus, Ohio, United States of America M. Juhas Materials Science and Engineering, The Ohio State University, Columbus, Ohio, United States of America

Keywords: Friction Stir Processing, Titanium, Microstructure Evolution

Abstract

Microstructure evolution during friction stir processing (FSP) of Ti-5111 performed both above and below the beta-transus temperature was investigated. Each processed panel was instrumented with thermocouples to record the thermal histories experienced within the stir zone and in the adjacent heat-affected zone. Temperature profile analysis was performed using single sensor differential thermal analysis (SS- DTA) which allowed determination of the beta transus. The FSP microstructures were characterized using light and scanning electron microscopy. The microtexture of the stir zones were compared to the base metal using electron backscatter diffraction (EBSD). Friction stir processing both above and below the beta transus resulted in extreme grain refinement – reducing the 200-500 micron base material grains to 1-20 microns depending on processing temperature. The above-transus stir zone exhibited a transformation texture in the alpha phase and a deformation texture in the beta phase. Both the alpha and beta phase exhibited deformation textures in the below-transus stir zone. The introduction of large amounts of strain during FSP depressed the beta transus to more than 100 °C below equilibrium in the panel processed above the transus.

Introduction

Ti-5111 is a near-alpha titanium alloy, containing less than 10 volume percent of beta in equilibrium [1], which was developed in a US Navy sponsored alloy development program for structural applications in ships and submarines [2]. Ti-5111 is known for its combination of intermediate strength and excellent toughness, corrosion and room temperature creep resistance [2,3]. The nominal composition of Ti-5111 in weight percent is Ti-5Al-1Sn-1Zr-1V-.8Mo with no more than .11 weight percent oxygen, and its beta transus is 980 °C [4].

91 The large grain size of titanium alloys can be refined using FSP, a modification of friction stir welding (FSW). FSW is a solid state joining process that was developed by W. M. Thomas et al at The Welding Institute (TWI) in 1991 [5]. In FSW, a pin tool harder than the work-piece is plunged into the joint and rapidly rotated. Rotational friction causes the work-pieces to plasticize and “stir” together. FSP can be used to alter the surface microstructure without the presence of a joint. Several FSW and FSP studies have been performed on CP titanium and Ti-6-4 [6-8] which characterized the friction stir regions as follows: stir zone (SZ), thermo- mechanical affected zone (TMAZ), heat affected zone (HAZ) and base material (BM). Fonda et al at the Navy Research Laboratory studied the microstructural evolution in Ti-5111 FSW, however, and noted only the SZ, BM and a slight region of deformation between the two [9]. This region of deformation that did not experience full recrystallization is referred to as the transition zone (TZ) [10]. Several techniques can be used to evaluate the microstructural changes that occur during welding and processing, one of which is SS-DTA. SS-DTA measures the heat effects that accompany phase transformations and compares in-situ recorded weld thermal histories with software-calculated reference curves [11,12]. The accuracy of SS-DTA has been confirmed using classical DTA techniques and dilatometry [13].

Experimental Procedures

Beta-annealed plates of Ti-5111 were friction stir processed above and below the beta transus at the Edison Welding Institute (EWI). A position-controlled GTC AccuStir machine supplied by the Army was utilized. Eight thermocouples were installed in each plate prior to processing to record the thermal histories at a sampling rate of 1000 s-1 experienced in the center-line of the stir zone at various depths. EWI supplied two refractory-based variable penetration tools (VPT) with different dimensions to achieve the heat inputs necessary for the two conditions. The above- transus panel was processed with a tool rotation of 450 rev min-1 and a tool travel speed of .13 cm sec-1 (3 in. min-1). The processing conditions below the beta-transus were 300 rev min-1 and .30 cm sec-1 (7 in. min-1). The panels were shielded with argon during processing to avoid oxidation. Transverse sections of the processed zones were mounted in conductive Bakelite and ground using 240, 320, 400, 600 and 800 grit silicon carbide paper. The final polish was achieved by using a vibratory polisher and .02 micron colloidal silica for 24 hours. Optical and scanning electron microscopy samples were etched using Kroll’s reagent (2 mL HF – 10 mL HNO3 – 88 mL H2O). An FEI Quanta scanning electron microscope (SEM) was used to analyze the SZ for both conditions. Electron backscatter diffraction (EBSD) was performed on unetched samples using EDAX software to determine the microtexture associated with the BM and two processed conditions. The thermal histories collected during processing were analyzed using SS- DTA. A reference curve was fit to the data which allowed the beta transus to be determined on-cooling.

92 Results and Discussion

Optical and Scanning Electron Microscopy

The results of the metallographic analysis corresponded to the Ti-5111 FSW study performed by Fonda et al. Only three regions were observed within the processed panels – BM, TZ and SZ. More information on each zone follows.

Base Material and Transition Zone

The BM exhibits a fully lamellar structure due to beta annealing. The beta phase is present as ribs between the alpha laths and the prior-beta grain size varies between 200 and 500 microns. Figure C.1 contains optical photomicrographs of the BM (left) and transition between the BM and above-transus SZ (right). The TZ is characterized by a distinct distortion of the lamellae and the beginning of recrystallization.

Figure C.1 Optical Photomicrographs of the Fully Lamellar BM Structure (left) and the TZ between the BM and Above-Transus SZ (right)

Above and Below-Transus Stir Zones

The above-transus SZ is characterized by transformed beta containing lamellar alpha, with alpha at the prior beta grain boundaries. The grains are a product of dynamic recrystallization above the beta transus temperature and are approximately 10-20 microns in diameter. In contrast, the SZ of the panel processed below the beta-transus exhibits finer grains, approximately 1 micron in diameter, of equiaxed primary alpha with a small volume fraction of beta. Bands of refractory- based inclusions from tool wear were observed within the below-transus SZ. Figure C.2 contains backscattered scanning electron photomicrographs of the SZ for both processing conditions.

Figure C.2 Backscattered SEM Photomicrographs of the Lamellar Alpha Grains of the Above-Transus SZ (left) and Fine Equiaxed Alpha Grains of the Below-Transus SZ (right)

Electron Backscatter Diffraction

The microtextures associated with the BM, above-transus SZ and below- transus SZ were determined using EBSD. All scans were performed on transverse sections of the processed panels. TD corresponds to the advancing side and RD is the tool travel direction. The pole figures are displayed as equal area projections. Figure C.3 is an EBSD map of the BM grains, showing the orientation differences between the alpha colonies. Due to the large grain size of the BM, pole figures are not included as the sampling size did not provide an accurate indication of the microtexture.

Figure C.3 EBSD Map of the Fully Lamellar Alpha Grains

The alpha and beta pole figures for the above-transus and below-transus SZ are included in Figure C.4. The EBSD scans were collected on the retreating side of the SZ.

94 a) b)

c) d) Figure C.4 Stir Zone Microtextures – a) Alpha Pole Figure for Above-Transus SZ, b) Beta Pole Figure for Above-Transus SZ, c) Alpha Pole Figure for Below-Transus SZ, d) Beta Pole Figure for Below- Transus SZ

The alpha pole figure for the above-transus SZ (Fig. C.4a) shows a transformation texture with one main intensity, or one preferential orientation, on the basal plane. The maximum intensity associated with this pole figure is significantly higher than the others due to the limited orientations of alpha transforming from beta. The beta pole figure for this processing condition (Fig. C.4b) reflects a deformation texture. Both pole figures for the below-transus SZ (Fig. C.4c and C.4d) reflect deformation textures as no transformation from beta-to-alpha occurred. The similarity between the beta pole figures of both conditions should be noted.

Single Sensor – Differential Thermal Analysis

The thermal histories collected during processing were analyzed using SS- DTA. The software used a 5-term Rosenthal equation to fit a reference curve to the data on-cooling. The heat of reaction from the alpha-beta phase transformation caused a deviation between the reference curve and thermal history. The maximum and minimum deviation (δT) indicates the locations of the transformation start and finish. Due to the limited amount of smooth, complete cooling curves, only two thermal histories were analyzed from the panel processed above the beta transus. Figure C.5 shows the start and finish temperatures of the beta-to-alpha transformation on-cooling for a thermocouple located within the stir zone.

95 840 880

Transformation Start 820 860 (862 C) Transformation Finish (808 C) 800 840

780 Temp (deg.C) Temp Temp (deg.C) Temp 820

760 800

740 780 -20 -10 0 10 20 -20 -15 -10 -5 0 T (deg.C) T (deg.C) Figure C.5 In-Situ Beta-to-Alpha Phase Transformation On-Cooling, Analyzed using SS-DTA

The transformation start temperature coincides with the beta transus temperature. For this thermocouple, the beta transus during FSP is 118 °C lower than the equilibrium transus. The other thermal history analyzed yielded a beta transus of 845 °C. The difference in results can be attributed to the varying amount of strain in each thermocouple location and the effect of the plasticized material sticking or slipping past the tool. As expected, the panel processed below the beta transus did not exhibit any thermal affects associated with the alpha-beta phase transformation.

Conclusions

After analyzing the Ti-5111 panels that were friction stir processed above and below the beta-transus, the following conclusions were made.

1. The initial 200-500 micron wide prior-beta grains in the base material were refined to 1 micron equiaxed alpha grains when processed below the beta transus and 10-20 micron lamellar alpha grains when processed above the beta transus.

2. The above-transus stir zone exhibited a transformed-alpha and deformed-beta microtexture. On the other hand, the below-transus stir zone exhibited deformation textures for both alpha and beta due to the lower processing temperature and the lack of a beta-to-alpha transformation.

3. The beta transus during processing was determined by analyzing in-situ thermal histories using SS-DTA. The non-equilibrium and highly-strained conditions associated with FSP depressed the equilibrium beta transus by more than 100 °C.

96 Acknowledgements

The authors would like to thank the Office of Naval Research for their support, as well as the Edison Welding Institute for performing the friction stir processing. We also appreciate the help provided by the Ohio State Univerisity Welding and Joining Metallurgy Group. A special thanks goes out to Prof. J. C. Williams and Adam Pilchak of the Materials Science and Engineering Department at the Ohio State University for discussions on titanium metallurgy and guidance with the EBSD analysis.

References

[1] G. Lutjering,̈ and J. C. Williams, Titanium, Springer, 2003, p 33

[2] D. Baxter, and I. Wallis, Joining of Titanium Alloy 5111, Proceedings from the 10th World Conference on Titanium, 2003, p 651-658

[3] J. Been and K. Faller, Using Ti-5111 for Marine Fastener Applications, JOM, Vol 51 (No. 6), 1999, p 21-24

[4] E. J. Czyryca, M. E. Wells, and K. Tran, Titanium Alloy Ti-5111 for Naval Applications, Proceedings from the RINA International Conference - Advanced Marine Materials: Technology and Applications, 2003, p 41-49

[5] W. M. Thomas, E. D. Nicholas, J. G. Needham, M. G. Murch, P. Temple-Smith, and C. J. Dawes, Improvements Relating to Friction Welding, (Patent No. 2,123,097) UK, 1991

[6] W. Lee, C. Lee, W. Chang, Y. Yeon, and S. Jung, Microstructural Investigation of Friction Stir Welded Pure Titanium, Materials Letters, Vol 59 (No. 26), 2005, p 3315-3318

[7] R. E. Jones, and Z. S. Loftus, Friction Stir Welding of 5 mm Titanium 6Al-4V, MS&T – AIST, Vol 6, 2006, p 119-134

[8] M. C. Juhas, G. B. Viswanathan, and H. L. Fraser, Microstructural Evolution in Ti Alloy Friction Stir Welds, Proceedings from the 2nd International Symposium on Friction Stir Welding, 2000

[9] R. W. Fonda, K. E. Knipling, C. R. Feng, and D. W. Moon, Microstructural Evolution in Ti 5-1-1-1 Friction Stir Welds, Friction Stir Welding and Processing IV, TMS, 2007, p 295-302

[10] A. L. Pilchak, M. C. Juhas, and J. C. Williams, Microstructural Changes Due to Friction Stir Processing of Investment-Cast Ti-6Al-4V, Met. and Mat. Trans. A, Vol 38 (No. 2), 2007, p 401-408

[11] B. T. Alexandrov, and J. C. Lippold, Methodology for In-situ Investigation of Phase Transformations in Welded Joints, IIW, Doc. IX-2114-04. 2004

[12] B. T. Alexandrov, and J. C. Lippold, A New Methodology for Studying Phase Transformations in High Strength Steel Weld Metal, Proceedings from the 7th International Trends in Welding Research Conference, 2005

[13] B. T. Alexandrov, and J. C. Lippold, Single Sensor Differential Thermal Analysis of Phase Transformations and Structural Changes during Welding and Postweld Heat Treatment, Welding In the World, Vol 51 (No. 11/12), 2007, p 48-59

APPENDIX D

TMS 2009 CONFERENCE PROCEEDING

Rubal, M., Lippold, J., and Juhas, M. (2009). Physical Simulation of Friction Stir Processed Ti-5111. In R. S. Mishra, M. W. Mahoney & T. J. Lienert (Eds.) Friction Stir Welding and Processing V (pp. 21-28). Warrendale: The Minerals, Metals & Materials Society (TMS).

PHYSICAL SIMULATION OF FRICTION STIR PROCESSED TI-5111

M. Rubal1, J. Lippold1, M. Juhas2

1Welding Engineering Program, The Ohio State University; 1248 Arthur E. Adams Dr.; Columbus, OH 43221, USA 2Dept. of Material Science and Engineering, The Ohio State University; 2041 College Rd.; Columbus, OH 43210, USA

Keywords: Friction Stir Processing, Titanium, Hot Torsion

Abstract

Friction stir processing (FSP) of Ti-5111 was performed above and below the beta transus temperature allowing for investigation of the microstructural evolution in both conditions. Each processed panel was instrumented with thermocouples to record the thermal histories in the stir zone and adjacent heat-affected zone. Single sensor differential thermal analysis (SS-DTA) was used to determine the beta transus temperature during processing. The FSP microstructures were characterized using optical and scanning electron microscopy, while the microtextures of the FSP regions were compared using electron backscatter diffraction (EBSD). FSP produced extreme grain refinement in both processing conditions – reducing the 200-500 m base material grains to 1-20 m. The microstructures observed in the FSP panels were simulated using a Gleeble® 3800. The strain and strain rate data may be used to verify FSP modeling programs of titanium to reduce the parameter selection phases of future friction stir projects.

Introduction

Ti-5111 (5Al-1Sn-1Zr-1V-.8Mo) is a near- titanium alloy with an equilibrium beta transus of approximately 980 °C [1]. The US Navy developed this alloy as a lower cost alternative for Ti-6Al-4V for structural applications in ships and submarines [2]. The desirable properties of Ti-5111 include intermediate strength in combination with excellent toughness, corrosion and room temperature creep resistance [2,3].

Friction stir processing (FSP), a modification of friction stir welding (FSW), can be used to reduce the grain size of titanium alloys. This refined structure improves the mechanical properties, including increasing the yield stress and resistance to fatigue crack initiation [4]. A previous FSW study was conducted on Ti-5111 by Fonda et al. at the Naval Research Laboratory. A microstructural analysis revealed the welds contained only three zones: the stir zone (SZ), base material (BM) and a narrow

100 region of deformation between the two [5]. This deformed region, which experiences only partial recrystallization, is referred to as the transition zone (TZ) [6]. The stir zone microstructure is significantly affected by the phase field in which processing occurs, specifically if deformation is introduced above or below the beta transus temperature. This was explored in Ti-6Al-4V FSP by Pilchak et al. [6] A goal of this study is to determine the effect of processing above and below the beta transus temperature on the microstructures observed in Ti-5111 FSP panels.

Single sensor differential thermal analysis (SS-DTA), developed at the Ohio State University, is one of several techniques capable of evaluating phase transformations during welding and processing. SS-DTA is unique from classical differential thermal analysis (DTA) in that it does not require a reference sample; instead, a software- calculated reference curve is fit to the in situ collected thermal data. Deviations from the reference curve indicate the start and finish temperatures of a phase transformation [7,8]. Both classical DTA techniques and dilatometry have been used to confirm the accuracy of SS-DTA [9].

Electron backscatter diffraction (EBSD) is an important tool in determining the degree of microtexture present in friction stir welds. Several texture studies have been conducted on titanium FSW and FSP panels, including the near- Ti-5111,  alloy TIMET 21S and the - alloy Ti-6Al-4V. Fonda et al. observed shear deformation and deformation twinning of the  phase in Ti-5111 FSW near the edges of the weld, while a strong texture associated with the Burgers orientation relationship occurred near the tool [10]. A shear texture was detected in TIMET 21S FSW by Reynolds et al. which could be rotated to correspond to a previously published BCC torsion texture. [11]. Finally, a study on Ti-6Al-4V FSP by Pilchak et al. revealed a moderate intensity (~5x) transformation texture in the  phase governed by the Burgers relationship when processed above the beta transus and a shear texture when processed sub-transus [12].

The Gleeble® is a thermo-mechanical simulator, first developed by Nippes et al. and currently produced by DSI, Inc. [13]. The 3800 model has an optional torsion mobile conversion unit (MCU) capable of imposing high strains and strain rates on a sample while maintaining a steady temperature gradient. Previous studies on steels have proved that the microstructures observed within the FSW regions can be simulated using Gleeble® torsion tests [14,15]. Determining the strain, strain rate and temperature combinations capable of producing a particular microstructure could be used to validate titanium FSW and FSP models.

101 Experimental Procedures

Beta-annealed plates of Ti-5111 were friction stir processed under two conditions: above and below the beta transus temperature. The processing was performed at the Edison Welding Institute (EWI) using an Army-supplied, position-controlled GTC AccuStir machine. Variable penetration tools (VPT) of a refractory-based material with different dimensions were used to achieve the necessary heat inputs. Table D.1 contains the processing parameters for the above and below transus conditions. Argon shielding was used during processing to avoid oxidation of the panels.

Table D.1 FSP Parameters for Above and Below Transus Conditions Tool Rotation Travel Speed FSP Condition rev/min in/min mm/sec Above Transus 450 3 1.3 Below Transus 300 7 3.0

Each panel contained eight thermocouples imbedded at various depths along the centerline. These thermocouples recorded the thermal histories experienced in the SZ and heat affected zone (HAZ) at a sampling rate of 1000 sec-1. SS-DTA analysis was performed on the in situ collected data to determine the on-heating and on-cooling beta transus temperatures.

The processed zones were cut into transverse sections, mounted in conductive Bakelite and ground through 800 grit silicon carbide paper. A colloidal silica vibratory polisher was used to achieve the final .02 m finish. Kroll’s reagent (2 ml HF – 10 ml HNO3 – 88 ml H2O) was selected to etch the optical and scanning electron microscopy samples. The SZ and TZ for both conditions were analyzed using an FEI Quanta scanning electron microscope (SEM). Electron backscatter diffraction (EBSD) was performed on unetched samples to compare the microtextures.

Hot torsion tests were performed on the Ti-5111 base material using a Gleeble® 3800 thermo-mechanical simulator equipped with a torsion system mobile conversion unit (MCU). The temperature, strain and strain rate were varied to simulate the microstructures observed in the SZ and TZ of the processed panels.

Results and Discussion

Single Sensor Differential Thermal Analysis

The in situ collected thermal histories were analyzed using SS-DTA. The software fit a reference curve to the data using a 3-term 2nd order polynomial on-heating and a 5- term Rosenthal equation on-cooling. A deviation between the reference curve and thermal history is caused by the heat of reaction from the - phase transformation.

102 The maximum and minimum deviation (T) indicates the start and finish temperatures of the transformation.

The sub-transus processed panels did not exhibit any thermal effects associated with the - phase transformation. The data collected during above-transus processing, however, contained a positive thermal effect on-heating, which could correspond to the beta transus or the superposition of the transus and dynamic recrystallization. This effect occurred while the tool was directly over the thermocouple, though, and the signal noise prevented accurate analysis. Two cooling curves collected during above-transus processing produced smooth thermal histories that exceeded the beta transus. Figure D.1 shows the analyzed cooling curve for a thermocouple located within the stir zone.

840 880

Transformation Start 820 860 (862 C) Transformation Finish (808 C) 800 840

780 Temp (deg.C) Temp Temp (deg.C) Temp 820

760 800

740 780 -20 -10 0 10 20 -20 -15 -10 -5 0 T (deg.C) T (deg.C) Figure D.1. In situ -to- phase transformation on-cooling, analyzed using SS-DTA

The transformation start temperature in Figure D.1 corresponds to the beta transus temperature. The on-cooling beta transus during processing was determined to be 862 °C at one thermocouple location within the SZ and 845 °C for the other analyzed thermocouple. These are depressed by more than 100 °C compared to the published equilibrium transus. Thus, the effect of non-equilibrium heating and extreme deformation experienced during FSP can significantly lower the beta transus. The difference in the two analyzed beta transus temperatures can be attributed to the varying amount of strain experienced by each thermocouple and the effect of the plasticized material sticking or slipping past the tool.

SS-DTA analysis of a furnace test on Ti-5111 base material heated at 5 °C/min and slow cooled produced a clear indication of the on-heating beta transus at 1006 °C and the on-cooling transus at 960 °C. These values are similar to the published beta transus of approximately 980 °C.

103 Optical and Scanning Electron Microscopy

Three regions were observed within the processed panels: base material (BM), transition zone (TZ) and stir zone (SZ) which confirms the findings of the FSW study performed by Fonda et al. on Ti-5111 [5].

The -annealed BM exhibits a fully lamellar microstructure with a prior- grain size of 200 to 500 m. The small volume fraction of  is present as ribs between the  laths. The TZ is characterized by distortion of the  lamellae and the beginning of recrystallization near the SZ. Figure D.2 contains optical photomicrographs of the BM and TZ in the panel processed above the beta transus.

Figure D.2. Optical photomicrographs of the lamellar BM (left) and the TZ between the BM and above-transus SZ (right)

The above-transus SZ exhibits 10-20 m grains of transformed  containing lamellar , with  at the prior- grain boundaries. This microstructure is due to recrystallization above the beta transus temperature. The sub-transus SZ, however, consists of 1 m equiaxed primary  with a small volume fraction of . It should be noted that bands of refractory-based inclusions from tool wear were observed within the sub-transus SZ. Backscattered electron SEM photomicrographs of the SZ for both processing conditions are in Figure D.3.

104

Figure D.3. Backscattered electron SEM photomicrographs of the lamellar  grains in the above-transus SZ (left) and equiaxed  grains in the sub-transus SZ (right)

Electron Backscatter Diffraction

The microtextures associated with the SZ for both processing conditions were determined using EBSD. All scans were performed on transverse sections of the FSP panels on the retreating side of the SZ. The TD direction in the pole figure corresponds to the advancing side and RD is the tool travel direction. Figure D.4 contains the  and  pole figures for the above and below-transus SZ, displayed as equal area projections.

Above-transus  Above-transus 

Below-transus  Below-transus  Figure D.4. Equal area projection pole figures describing the retreating-side SZ microtexture for the above- and below-transus FSP panels

The  pole figure for the above-transus SZ shows a transformation texture with one main intensity on the basal plane. The maximum intensity associated with this pole figure is more than double the others. This is due to the limited orientations of  transforming from , as dictated by the Burgers orientation relationship. The  pole figure for the above-transus processing condition reflects a deformation texture. Both

105 pole figures for the below-transus SZ exhibit deformation textures as no transformation from -to- occurred. The similarity between the  pole figures of both conditions should be noted.

Gleeble® Torsion Simulation

A Gleeble® 3800 equipped with a torsion MCU was used to simulate the microstructure of the above-transus SZ. Thermal histories obtained during processing were input into the test program and the strain and strain rate were varied. The torsion samples were quenched with argon both on the sample surface and through the hollow interior to achieve the cooling rates experienced during processing.

The below-transus SZ cannot be simulated using Gleeble® torsion tests due to the high flow stress of the material at the processing temperature. The simulation samples sheared at the center prior to recrystallization.

In the samples used to simulate the above-transus SZ, strain localization was observed at testing temperatures below 1000 °C, which prevented the determination of a strain gradient as in the studies performed by Norton and Sinfield [14,15]. However, this localization allowed for full recrystallization of the sample microstructure. Figure D.5 shows backscattered electron SEM photomicrographs of a torsion sample and the SZ of the above-transus FSP panel. The torsion sample was tested at a peak temperature of 940 °C and experienced one full revolution at a speed of 200 rev/min.

Figure D.5. Backscattered electron SEM photomicrographs of a torsion simulation sample (left) and the SZ of the above-transus FSP panel (right)

Both the torsion sample and the above-transus SZ of the FSP panel contain lamellar  grains with  at the prior  grain boundaries. However, the  lamellae and grain boundary  are coarser in the SZ microstructure. This is due to detachment of the

106 control thermocouple during testing and the torsion sample cooling faster than the FSP panel. The grain size for both is approximately 10-20 m.

Once a method is developed for determining the strain experienced in the sample, the data collected from the simulations can be used to verify a model and reduce the parameter selection phase for titanium FSW and FSP.

Conclusions

After analysis of the Ti-5111 FSP panels and the initial Gleeble® 3800 simulations, the following conclusions were made.

1. The on-cooling beta transus during processing was determined using SS-DTA. The non-equilibrium and highly-strained processing conditions depressed the transformation by over 100 °C, compared to the equilibrium transus.

2. The FSP panels contained only three microstructural regions: the base material, transition zone and stir zone. The above-transus stir zone was characterized by 10-20 m lamellar  grains and the below-transus stir zone exhibited 1 m equiaxed  grains.

3. Processing above the beta transus caused a transformed- and deformed- microtexture. The intensity of the  microtexture was higher due to the Burgers orientation relationship. The sub-transus stir zone exhibited deformation textures for both the  and  phases.

4. Initial Gleeble® 3800 torsion simulations were able to match the microstructure of the above-transus stir zone. However, a method for determining the strain experienced in the sample must be established.

Acknowledgements

The authors would like to thank the Office of Naval Research, Cognizant Program Officer Johnnie Deloach, for support of this project. Our gratitude is also expressed to Brian Thompson and Seth Shira of the Edison Welding Institute for performing the friction stir processing. Special appreciation goes to Prof. Jim. Williams and Adam Pilchak of the Department of Materials Science and Engineering at the Ohio State University for helpful discussions on microtexture and titanium metallurgy.

107

References

1. E. J. Czyryca et al., “Titanium Alloy Ti-5111 for Naval Applications” (Paper presented at the RINA International Conference - Advanced Marine Materials: Technology and Applications, London, 9-10 October 2003), 41-49.

2. D. Baxter et al., “Joining of Titanium Alloy 5111” (Paper presented at the 10th World Conference on Titanium, Hamburg, 13-18 July 2003), 651-658.

3. J. Been et al., “Using Ti-5111 for Marine Fastener Applications,” JOM, 51 (6) (1999), 21-24.

4. Gerd Lütjering and James C. Williams, Titanium (Berlin: Springer, 2003), 191.

5. R. W. Fonda et al. “Microstructural Evolution in Ti 5-1-1-1 Friction Stir Welds,” Friction Stir Welding and Processing IV, ed. R. S. Mishra et al. (Warrendale, PA: TMS, 2007), 295-302.

6. A. L. Pilchak et al., “Microstructural Changes Due to Friction Stir Processing of Investment-Cast Ti-6Al-4V,” Met. and Mat. Trans. A, 38 (2) (2007), 401- 408.

7. B. T. Alexandrov et al., “Methodology for In-situ Investigation of Phase Transformations in Welded Joints” (IIW Doc. IX-2114-04, International Institute of Welding, 2004).

8. B. T. Alexandrov et al., “A New Methodology for Studying Phase Transformations in High Strength Steel Weld Metal” (Paper presented at the 7th International Trends in Welding Research Conference, Pine Mountain, Georgia, 16-20 May 2005).

9. B. T. Alexandrov et al., “Single Sensor Differential Thermal Analysis of Phase Transformations and Structural Changes during Welding and Postweld Heat Treatment,” Welding In the World, 51 (11/12) (2007), 48-59.

10. R. W. Fonda et al., “Microstructure of Ti 5111 Friction Stir Welds” (Presentation at the 136th TMS Annual Meeting, Orlando, Florida, 25 February – 1 March 2007).

11. A. P. Reynolds et al., “Texture in Friction Stir Welds of Timetal 21S,” Scripta Materialia, 52 (6) (2005), 491-494.

12. A. L. Pilchak et al., “A Comparison of Friction Stir Processing of Investment Cast and Mill-annealed Ti-6Al-4V,” Welding in the World, 52 (9/10) (2008), Research Supplement.

108 13. E. F. Nippes et al., “An Investigation of the Hot Ductility of High Temperature Alloys,” Welding Journal, 4 (1955), 183s-196s.

14. S. J. Norton, “Ferrous Friction Stir Weld Physical Simulation” (Ph.D. Thesis, The Ohio State University, 2006).

15. M. F. Sinfield et al., “Physical Simulation of Friction Stir Weld Microstructure of a High-Strength, Low Alloy Steel (HSLA-65)” (Paper presented at the 7th International Symposium on Friction Stir Welding, Awaji Island, Japan, 20-22 May 2008).

109