Establishing a Methodology for Thermo-Mechanical Refurbishment of Nickel Based Superalloy Aeroengine Components

MISSING IMAGE

Material Information

Title:
Establishing a Methodology for Thermo-Mechanical Refurbishment of Nickel Based Superalloy Aeroengine Components
Physical Description:
1 online resource (197 p.)
Language:
english
Creator:
Mendoza, Alvaro G Jr
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Materials Science and Engineering
Committee Chair:
Fuchs, Gerhard E
Committee Members:
Dempere, Luisa A
Holloway, Paul H
Patterson, Burton Roe
Arakere, Nagaraj K

Subjects

Subjects / Keywords:
bladed -- blisk -- grain -- heat -- integrally -- jet -- nickel -- refurbishment -- repair -- rotor -- size -- superalloy -- thermomechanical -- treatment -- turbine
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre:
Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Increasing the efficiency of aeroengines requires a compromise between performance, weight and cost. Integrally bladed rotors (IBR) or combined discs and blades are used in the engine increase performance while saving weight. Cost becomes an issue, as damaged IBRs are discarded, due to the challenges of in-situ repair. Current IBR repair strategies involve additive manufacturing with a similar alloy. For this study, the repair processes examined were plasma powder deposition (PPD) and Laser Engineered Near-Net Shape (LENS), these methods were used to deposit powder metallurgy IN-100 nickel based superalloy on to a damaged IN-100 IBR airfoil. The IBR is a fine grained forged and heat treated component, whereas the PPD and LENS repaired airfoils were directionally solidified coarse grained structures. Post-deposition thermal and thermo-mechanical processing methods were investigated in order to refine the repaired structures towards the higher strength fine grain structure of the IBR basemetal. Initially a wide range of thermal treatments was investigated, including temperatures as low as 1093°C to as high as 1232°C, just below incipient melting temperature. Along with temperature, the thermal treatment test matrix included variations in heat treatment time from 30 minutes to 2 hours to determine the optimal heat treatment schedule. Additional thermal strain was added using a ratcheting heat treatment. Thermal treatments alone were not sufficient to refine the as-deposited structure to a fine grain condition. Post deposition thermo-mechanical processing was pursued to provide the deformation needed to promote recrystallization of the coarse grained deposited structure. The deformation provided the added driving force for recrystallization but only when the deformed alloy was heat treated above the ?’-solvus of 1177°C for 1 hour. This condition yielded the best microstructure of all of those examined in this study, a relatively fine-grained, equiaxed, homogeneous microstructure. Different levels of deformation were used to varying degrees of success from 4 to 10%. This data was used to define a process window to repair the IBR. Overall the combination of optimum levels of heat treatment and deformation resulting from the exploration of the factors involved in the refurbishment process lead to the desired microstructure for a refurbished airfoil.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Alvaro G Jr Mendoza.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
Local:
Adviser: Fuchs, Gerhard E.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044093:00001


This item is only available as the following downloads:


Full Text

PAGE 1

ESTABLISHING A M ETHODOLOGY FOR THERMO-MECHANICAL REFURBISHMENT OF NICKEL BASED SUPERALLOY AEROENGINE COMPONENTS By ALVARO G. MENDOZA JR. A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012 1

PAGE 2

2012 Alv aro G Mendoza Jr. 2

PAGE 3

To my family, friends and colle agues thank you for the support 3

PAGE 4

ACK NOWLEDGMENTS I would like to thank everyone who has helped, supported or influenced me because this undertaking would not have been possible alone. First and foremost, I want to thank my family who has always been supportive of me even if they could never understand quite what I was doing. They would nod politely and commiserate with me regardless. My pa rents, Al and Hilary have always supported me no matter what I chose to do and have alwa ys had a confidence in me that inspired me to do more than I thought I could do and my brothers, Mario and Robie for always being a welcome distraction when I was overwhelmed. I would like to thank my gi rlfriend Alex Butler who has been there every step of the way, being my biggest cheerleader even w hen we were living on opposite sides of the country. I cant really expre ss how much that means to me. I would also like to thank the High Te mperature Alloys group, past and present. Andrew Wasson, Brandon Wilson and Krish na Ganesan for leading by example and teaching me the ropes, Phil Draa and Max Kaplan who provide sounding boards and support, Carlos Inguanzo and Gowri Balasubram anian who assisted me in the lab in tasks large and small. I would especially like to thank Dr. Fu chs, my advisor. He has over the years influenced my education and who I am as an engi neer at every turn, through classes, lab work and discussions in his office. He kindled my love of metallurgy in his intro class and has been a font of knowledg e and practical advice since. All the staff at MAIC here at UF includi ng but not limited to Dr. Amelia Dempere, Rosabel Ruiz and Wayne Acree have assisted me at some point during this process and have my thanks for their aid. 4

PAGE 5

TABL E OF CONTENTS page ACKNOWLEDG MENTS..................................................................................................4 LIST OF TABLES............................................................................................................9 LIST OF FI GURES ........................................................................................................11 LIST OF ABBREVI ATIONS...........................................................................................16 ABSTRACT ...................................................................................................................20 CHAPTER 1 INTRODUC TION....................................................................................................22 2 BACKGRO UND...................................................................................................... 25 Introduction to Superallo ys.....................................................................................25 Microstructure and Strengt hening ...........................................................................26 Crystal Structure and Microstruc ture................................................................26 Chemistry and St rengthening...........................................................................27 Performance vs. Cost.............................................................................................28 Materials Selection and Deposition Pr ocesse s.......................................................30 IN-100: Characteristics, Fabrication and Superp lasticit y..................................30 Deposition Pr ocess..........................................................................................32 LENS Proc ess..................................................................................................34 Post-Deposition Processi ng....................................................................................35 Recrystallization: Kineti cs and Residual Stresses............................................35 Thermo-Mechanical Processi ng.......................................................................37 3 EXPERIMENTAL PROCEDUR ES..........................................................................46 Materials.................................................................................................................46 Heat Treat ments.....................................................................................................47 Mechanical Deformation.........................................................................................50 Mechanical Te sting.................................................................................................51 Sample Prepar ation................................................................................................51 Characteri zation ......................................................................................................53 Light Optical Microscopy..................................................................................53 Scanning Electron Microscopy.........................................................................53 X-Ray Diffraction ..............................................................................................53 Quantificat ion..........................................................................................................54 5

PAGE 6

4 RESULTS THERMAL TREATME NTS..................................................................61 Development of H eat Treatm ents...........................................................................61 1 PPD Heat Treat me nt Ma trix...............................................................................63 st2 Heat Treatm ent M atrix......................................................................................64 ndAs-Deposit ed ....................................................................................................64 Sub-solv us........................................................................................................65 Super-Solvus....................................................................................................66 Summary ..........................................................................................................67 3 Heat Treatm ent M atrix.......................................................................................67 rdVerifica tion ........................................................................................................ 68 Directed Air Coolin g.........................................................................................68 Water Coo ling...................................................................................................70 Summary ..........................................................................................................70 LENS Heat Tr eatment s...........................................................................................70 Non-Ratchet Tr eatments..................................................................................71 Ratcheted Treat ments......................................................................................72 Summary ..........................................................................................................72 Summary of Factors................................................................................................73 Re-evaluating Methodology....................................................................................74 5 RESULTS FEASIBI LITY ST UDY.........................................................................92 Adding Defo rmation................................................................................................92 Proof of Concept Testing........................................................................................92 X-Ray Diffraction Verifica tion..................................................................................93 Justification fo r Testin g.....................................................................................93 X-ray Diffracti on Result s...................................................................................93 Use of Rolling as Deformation Method...................................................................94 Rolling Feasibility Study High Defo rmation..........................................................95 As-Rolled Sa mples...........................................................................................95 Sub-Solvus Heat Treatm ents...........................................................................96 Super Solvus Heat Treatm ents........................................................................97 Summary ..........................................................................................................98 Rolling Feasibility Study Low Defo rmation...........................................................98 As-Rolled Sa mples...........................................................................................99 Super Solvus Heat Treatm ents........................................................................99 Super Solvus and Ratcheting Treatm ents......................................................100 Summary ........................................................................................................101 Materials Change and Optimizati on......................................................................101 6 RESULTS THERMO-MECHA NICAL PROC ESSING .........................................118 Rolling Verifi cation................................................................................................118 Low Deformati on Level ...................................................................................119 Medium Deformati on Level .............................................................................120 High Deformati on Level ..................................................................................121 6

PAGE 7

Summary ........................................................................................................ 122 Alternate Methods of Deformati on........................................................................122 Shot peeni ng..................................................................................................122 Laser Shock Peening.....................................................................................123 Low Plasticity Burnishi ng................................................................................125 Summary ........................................................................................................126 In-situ Deformati on Proces s...........................................................................127 Confirmati on.........................................................................................................128 7 DISCUSSI ON.......................................................................................................150 Microstructural Evaluatio n.....................................................................................150 Dendritic Stru cture..........................................................................................150 Grain Struct ure...............................................................................................151 Twinning .........................................................................................................151 Carbides and Impurities ..................................................................................152 Segregation and Hom ogenization ..................................................................153 Porosity ..........................................................................................................155 Summary ........................................................................................................156 Recrystallization Factors.......................................................................................156 Time...............................................................................................................157 Temperature...................................................................................................158 Purity ..............................................................................................................159 Prior Deforma tion...........................................................................................159 Prior Grain Size..............................................................................................160 Summary ........................................................................................................161 Recrystallization Factors and t he Effect on Gr ain Size .........................................161 Overview ........................................................................................................162 Time and Temperatur e Effect s.......................................................................162 Deformation Level..........................................................................................164 Ancillary Effects..............................................................................................165 Comparison to AST M Grain Si ze...................................................................167 Summary ........................................................................................................167 8 CONCLUSION S ...................................................................................................184 Closing Remarks..................................................................................................184 Future Wo rk..........................................................................................................186 APPENDIX A Rolling Force Ca lculatio n......................................................................................187 B Low Plasticity Burnis hing Hardness Data.............................................................189 C Grain Size Data for PP D and LENS Sa mples.......................................................191 7

PAGE 8

LIST OF REFE RENCES .............................................................................................192 BIOGRAPHICAL SKETCH ..........................................................................................197 8

PAGE 9

LIST OF TABLES Table page 2-1 Roles of element s in superalloys ........................................................................41 2-2 Composition of Modified IN -100.........................................................................42 3-1 IN-100 Com positio n............................................................................................56 3-2 Initial heat tr eatment ma trix................................................................................57 4-1 IN-100 Heat Treatm ent window values ...............................................................75 4-2 Examples of standard nickel based superalloy solu tion heat treatments............75 4-3 1 Heat Treatm ent M atrix...................................................................................77 st4-4 2 HT Ma trix......................................................................................................78 nd4-5 3 Heat Treatm ent M atrix...................................................................................83 rd4-6 LENS verificati on sample matrix.........................................................................87 5-1 Pre-treatments fo r rolling trials.......................................................................... 106 5-2 Actual deformation level for 8% RIA ro lling study .............................................106 5-3 Test matrix for 8% RIA rolling study.................................................................107 5-4 Test matrix for low defo rmation level rolling study ............................................112 5-5 Actual deformation level fo r 4-6% RIA rolling study ..........................................112 6-1 Deformation data for LE NS rolling study..........................................................129 6-2 LENS 4% roll ing matr ix.....................................................................................129 6-3 LENS 8% Rol ling Matr ix...................................................................................133 6-4 Sample matrix fo r LSP samp les.......................................................................141 6-5 Sample matrix fo r LPB samp les.......................................................................144 6-6 In-situ deformation sample ma trix.....................................................................147 6-7 Average hardness for different in-situ heat tr eatments.....................................149 7-1 Legend for grain si ze analysi s..........................................................................176 9

PAGE 10

7-2 ASTM grain size number and t he corresponding gr ain area ............................181 7-3 Estimated ASTM grain size numbers for the most promising results sorted by increasing grai n area........................................................................................182 B-1 Inside edge length hardness tr ace....................................................................189 B-2 Outside edge length hardness trace.................................................................189 B-3 Center section hardness trac es........................................................................190 B-4 Width tr aces......................................................................................................190 C-1 Grain size data fo r PPD samp les......................................................................191 C-2 Grain size data fo r LENS samp les....................................................................191 10

PAGE 11

LIST OF FIGURES Figure page 2-1 Schematic of Jet E ngine .....................................................................................39 2-2 Turbine inlet temperatures in Roll-Royce civilian eng ines 19402010................39 2-3 Nickel based superalloy microstruc ture..............................................................40 2-4 Schematic of gamma and gamma prime crystal structures................................40 2-5 Effect of Temperatur e on IN-100 pr operties .......................................................41 2-6 Temperature Capability of Superallo ys (C) .......................................................42 2-7 Room temperature properties of IN -100 based on proce ssing hist ory................43 2-8 SEM image of the IN100 powder at 500x..........................................................43 2-9 Cross-section of laser we ld deposited st ructure.................................................44 2-10 The LENS pr ocess.............................................................................................44 2-11 Surface residual ratio vs. annea ling time for carbon st eel..................................45 3-1 Layer build during dep osition pr ocess................................................................56 3-2 LENS and PPD samples....................................................................................56 3-3 Heat-up Rate Diagr am........................................................................................57 3-4 Schematic of ratcheting heat treatment showing the sample temperature versus time.........................................................................................................58 3-5 SEM photomicrographs of deposited (PW 1074) samples etched with various etchant s..............................................................................................................59 3-6 Powder XRD Schemat ic .....................................................................................60 3-7 Schematic of calculation for grain size analysis..................................................60 4-1 Effect of composition on the -solvus temperatures and incipient melting temperatures for several well known s uperallo ys...............................................76 4-2 Grain growth of IN-100 (P/M) as a function of annealing temperature for powders produced by different methods and cast and wrought process............77 4-3 As-Deposited PPD Sample at 50x (1 Matrix)....................................................78 st 11

PAGE 12

4-4 Samples P1 and P1 -R at 50x .............................................................................79 4-5 Samples P2-30, P2 and P2-120 at 50x..............................................................80 4-6 Sample P3 at 50x...............................................................................................81 4-7 Sample P4..........................................................................................................81 4-8 Sample P5 at 100x.............................................................................................82 4-9 Sample P6 at 250x.............................................................................................82 4-10 Sample P6-30 and P6-R both at 50x..................................................................83 4-11 Verification micrographs of Sample P6 from the 2 HT Matrix..........................84 nd 4-12 Micrographs of P1 R/6-30 DA C ...........................................................................84 4-13 Micrographs of P1R/6 DA C................................................................................85 4-14 Micrographs of P4 R/6-30 DA C...........................................................................85 4-15 Micrographs of P4R/6 DA C................................................................................86 4-16 Micrographs of P1R/6-30 WC.............................................................................86 4-17 Micrographs of P1R/6 WC..................................................................................87 4-18 Dendrite reorientation at layer bands in powder pl asma deposit sample............88 4-19 Micrographs of non-ratchet samp les with no RXN at 50x...................................89 4-20 Micrographs of non-ratchet samp les that exhi bited RX N....................................90 4-21 Micrographs of ratc heted samp les......................................................................91 5-1 Example of Vickers hardness ind ents...............................................................103 5-2 LOM Micrographs of RXN hardness i ndents in LENS ma terial at.....................103 5-3 SEM Micrographs of RXN har dness indents in LENS material at 650x............104 5-4 Micrograph showing the bulk of the LE NS hardness samp les at 50x...............104 5-5 Full x-ray diffrac tion patte rn..............................................................................105 5-6 Close-up of XRD Pattern ..................................................................................105 5-7 Decrease in strength due to over aging ............................................................. 106 12

PAGE 13

5-8 As-rolled microstructures with 8% RIA at 100x .................................................107 5-9 Microstructures heat treated at 1121 C with 8% RIA at 100x...........................108 5-10 Microstructures heat treated at 1149 C with 8% RIA at 100x...........................109 5-11 Microstructures heat treated at 1177 C with 8% RIA at 100x...........................110 5-12 Microstructures heat treated at 1204 C with 8% RIA at 100x...........................111 5-13 Microstructures of As-Rolled samples with 5% RIA..........................................113 5-14 Microstructures heat treated at 1177 C with 5% RIA........................................114 5-15 Microstructures heat treated at 1204 C with 5% RIA........................................115 5-16 Microstructures heat treated at 1177 C with 5% RIA and Ratcheting treatments at 100x............................................................................................116 5-17 Microstructures heat treated at 1204 C with 5% RIA and Ratcheting treatments at 100x............................................................................................117 6-1 Micrographs of low deformation level samp les in the as-rolled condition.........130 6-2 Micrographs of low deformation level samples heat treated at 1177C............131 6-3 Micrographs of low deformation level samples heat treated at 1204C............132 6-4 Micrographs of medium deformation level samples as -rolled...........................134 6-5 Micrographs of medium deformation leve l samples heat treated at 1149C.....135 6-6 Micrographs of medium deforma tion leve l samples heat treated at 1177C.....136 6-7 Micrographs of medium deformation leve l samples heat treated at 1204C.....137 6-8 Micrographs of high deformation level samples heat treated at 1204C...........138 6-9 As-deposited sample with no treatment after shot peeni ng at 50x...................139 6-10 Micrographs of shot peened and heat treated sa mples....................................140 6-11 Micrographs of as-deformed LSP samples at 50x............................................141 6-12 Micrographs of LSP samples heat tr eated for 15 minutes at 50x.....................142 6-13 Micrographs of LSP samples heat treated for 1 hour at 50x.............................143 6-14 Higher magnification image of intergranular cracking in LSP sample (100x)....143 13

PAGE 14

6-15 Micrographs of LPB samples heat treated at 1177C ....................................... 144 6-16 Micrographs of LPB samples heat treated at 1204C .......................................145 6-17 Average hardness values by location of high deformation LPB sample...........146 6-18 Width hardness traces for high def ormation level L SP sample ........................146 6-19 Residual stress comparison for LPB, LSP and Shot peened samples.............147 6-20 In-situ deformed sample heat treated fo r 1 hour...............................................148 6-21 In-situ deformed sample heat treated for 30 minutes.......................................148 6-22 In-situ deformed sample heat treated for 15 minutes.......................................149 6-23 In-situ deformation hardness after heat treatment ............................................149 7-1 Dendrite structure evoluti on with temper ature..................................................169 7-2 Effect of temperature on the dendrite structure at 50x......................................170 7-3 Schematic of a twin boundar y...........................................................................171 7-4 EDS spectra of a pr imary carbide.....................................................................171 7-5 As-deposited microstructure before heat treatment featuri ng impurities and porosity at 100x................................................................................................172 7-6 Microstructures of transit ion region at 50x........................................................172 7-7 Illustration of the effect of cold work and annealing temperature on Cu35Zn Brass................................................................................................................173 7-8 Relationship between annealing ti me and te mper atures..................................173 7-9 Changes in annealing temper ature and deformation in high purity copper.......174 7-10 The effect of additions of chromi um and molybdenum on recrystallization temperatur e......................................................................................................174 7-11 Recrystallized grain size as a functi on of prior plasti c deformati on...................175 7-12 Recrystallized grain size as a functi on of prior plastic deformation with recrystallization temperatur es...........................................................................175 7-13 Grain area vs. Deformation Lev el plot that shows all t he samples examined..176 14

PAGE 15

15 7-14 The result of heat treatm ents on the base metal control samples at various temperatures a fter 1 hour .................................................................................177 7-15 Grain size as a function of time for identical sa mples with varying heat treatment ti mes.................................................................................................177 7-16 Grain area for low deformation condi tion graphed against processing condition s.........................................................................................................178 7-17 Grain area for medium deformation condition graphed against processing condition s.........................................................................................................178 7-18 Grain areas for low and m edium deforma tion...................................................179 7-19 Ratcheted and non-ratcheted LENS and PPD samples under identical condition s.........................................................................................................179 7-20 The relationship between pre-treatm ents and temperature/ratcheting treatment s........................................................................................................180 7-21 Grain area versus deformation for low and medium deformation conditions....180 7-22 Graphical relationship between grain ar ea and ASTM grain size number........182 7-23 Graphical representation of Table 7-3..............................................................183

PAGE 16

LIST OF ABBREVIAT IONS Terms and Phrases AA Argon atomized AC Air cooling AS As-deposited ASTM American Society for Testing and Materials BSE Backscattered electron C&W Cast and wrought CAD Computer aided design DAC Directed air cooling DHP Dissolved hydrogen process DS Directional solidification DTA Differential thermal analysis EDM Electrodischarge machining EDS Energy dispersive spectroscopy FCC Face centered cubic GB Grain boundary HAZ Heat affected zone HRC Rockwell C hardness HT Heat treatment HV Vickers hardness I/M Ingot metallurgy IBR Integrally bladed rotor JC Jet cooling 16

PAGE 17

LENS Laser Engineered Net Shaping LOM Light optical microscopy LPB Low plasticity burnishing LSP Laser shock peening MAIC Major Analytical Instrumentation Center P/M Powder metallurgy PERC Particle Engineering Research Center PPB Prior particle boundaries PPD Plasma powder deposition PS Partial solution RC Rockwell C hardness REP Rotating electrode process RIA Reduction in area RXN Recrystallization SDAS Secondary dendrite arm spacing SE Secondary electron SEM Scanning electron microscopy SFE Stacking fault energy SO Super overage SX Single crystal TET Turbine entry temperature TIP Thermally induced porosity TMP Thermo-mechanical processing 17

PAGE 18

WC Water cooling XRD X-ray diffraction Elements and Compounds Al Aluminum B Boron C Carbon Co Cobalt Cr Chromium HCl Hydrochloric acid HNO3 Nitric acid H2O Water M23C6 Metal carbide M6C Metal carbide MC Metal carbide Mo Molybdenum NaCl Sodium chloride (rock salt) Nb Niobium Nd:YAG Neodymium-doped yttrium aluminum garnet Re Rhenium Ta Tantalum Ti Titanium TiC Titanium Carbide V Vanadium 18

PAGE 19

19 W Tungsten Zr Zirconium Units C degrees Celsius F degrees Fahrenheit g Grams hrs Hours in Inches min Minutes nm Nanometers sec Seconds m Micrometer or micron V Volts wt. % Weight percent

PAGE 20

Abstract of Dissertation Pr esented to the Graduate School of the University of Fl orida in Partial Fulf illment of the Requirements for t he Degree of Doctor of Philosophy ESTABLISHING A METHODOLOGY FOR THERMO-MECHANICAL REFURBISHMENT OF NICKEL BASED SUPERALLOY AEROENGINE COMPONENTS By Alvaro G. Mendoza Jr. August 2012 Chair: Gerhard Fuchs Major: Materials Science and Engineering Increasing the efficiency of aeroengines requires a compromise between performance, weight and cost. Integrally bladed rotors (IBR) or combined discs and blades are used in the engine increase perform ance while saving weight. Cost becomes an issue, as damaged IBRs are discarded, due to the challenges of in-situ repair. Current IBR repair strategies involve additive manufacturing with a similar alloy. For this study, the repair processes ex amined were plasma powder deposition (PPD) and Laser Engineered Near-Net Shape (LENS) these methods were used to deposit powder metallurgy IN-100 nickel based supera lloy on to a damaged IN-100 IBR airfoil. The IBR is a fine grained forged and heat treated component, whereas the PPD and LENS repaired airfoils were directionally solidified coarse grai ned structures. Postdeposition thermal and thermo-mechanical processing methods were investigated in order to refine the repaired st ructures towards the higher strength fine grain structure of the IBR basemetal. Initially a wide range of thermal trea tments was investigated, including temperatures as low as 1093C to as high as 1232C, just below incipient melting temperature. Along with tem perature, the thermal treatm ent test matr ix included 20

PAGE 21

21 variations in heat treatment time from 30 minutes to 2 hours to determine the optimal heat treatment schedule. Additional therma l strain was added using a ratcheting heat treatment. Thermal treatments alone were not sufficient to refine the as-deposited structure to a fine grain condition. Post deposition thermo-mechanical proc essing was pursued to provide the deformation needed to promote recrystalliz ation of the coarse grained deposited structure. The deformation provided the added dr iving force for recrystallization but only when the deformed alloy was heat treated above the -solvus of 1177C for 1 hour. This condition yielded the best microstructure of all of those examined in this study, a relatively fine-grained, equiaxed, homogeneous microstructure. Different levels of deformation were used to varying degrees of success from 4 to 10%. This data was used to define a process window to repair the IBR. Overall the combination of optimum levels of heat treatment and deformation resulting from the explorati on of the factors involved in the refurbishment process lead to the desired microstructure for a refurbished airfoil.

PAGE 22

CHA PTER 1 INTRODUCTION Since the birth of the jet engine in the early 20th centur y, there has been a constant push for hotter, faster and lighter engines. This has led to some compromises for performance sake. Nickel based superallo ys are heavy but necessary to push the engine to higher temperatures, which leads to better efficiency. Integrally bladed rotors (IBR) are used in the compressor section of some aircraft gas turbines for increased performance. In particular, t hese IBRs offer a weight reduc tion due to the simplification of the rotor assembly. However, the simplicit y of the IBR does have a potential limitation when it comes to repair of damaged airfoils. Since the airfoils are integral to the rotor, repair of individual airfoils is much more complex. In a traditional rotor assembly, the rotor assembly is disassembled, a damaged ai rfoil is simply removed from the disk and a new airfoil inserted and the rotor returned to service. For an IBR that experienced airfoil damage beyond original engine manufact urer (OEM) blend limit s, the whole rotor needs to be replaced pending an aerospace approv ed repair process. Due to the high cost of manufacturing an IBR, airfoil repair rather than re placement would be the most cost effective course of action. Repairing the airfoils involves re moving the damaged area and depositing new blade material using methods like plasma powder deposition or the Laser Engineered Net Shaping (LENS) process, and then a postweld heat treatment scheme is needed to give the new blade section properties similar to the original material. This heat treatment is important because the asdeposited structure is coarse and has properties, which do not match the original blade section. The desired structure is fi ne-grained, giving the best mix of strength, fatigue resistance and toughness. 22

PAGE 23

Repairing these parts, though, presents a unique technical challe nge. The ability to repair the damaged airfoil with OEM properti es in the area of t he repair has not been demonstrated for nickel superalloy IBRs. Theref ore, a program was initiated to develop not only a means to repair a damaged nickel s uperalloy IBR airfoil, but to also address the ability to produce a microstructure in the repaired region that is similar to the base metal (i.e., IBR). In order to produce an airf oil repair with microstructures and properties similar to the base metal, several thermal treatment s and deformation processing steps were considered to refine the as-deposited microstruc ture by recrystallization. Initial effort involved a wide range of therma l cycles that ranged in heat tr eat temperatures from the standard IN-100 treatment to as high as tem peratures just below incipient melting temperature of the superalloy. Along with temperatur e, the thermal treat ment test matrix included variations in heat treatment time to determi ne the optimal heat treatment schedule. After these parameter s were investigated, the focus of the research was narrowed to the times and temperatures, whic h resulted in microstructural refinement approaching the baseline wrought metal microstruc ture. After a significant effort in post deposition thermal treatment, it was determined that the the rmal treatments alone were not shown to significantly refine and homogenize the microstructure by inducing recrystallization of the deposited material. Therefore, (mechanical) deformation processing was added to the processing scheme to provide more energy for recrystallization (RXN). Prior to deformati on processing, the as-deposited material was also given thermal processing to improve the workability of the deposit. Findings in examining the microstructures of heat-treated samples show that refinement of the 23

PAGE 24

24 microstructure is possible with residual stress in the material providing the energy for recrystallization of new grains to occur. The samples from the various heat treat ment schemes were characterized by optical microscopy, scanning electron micros copy and x-ray diffraction to determine the amount of RXN exhibited by t he samples and to determine the final microstructure of the deposit following thermo-mechanical processi ng. Additional analysis in the form of hardness testing and grain size analysis, wa s used to confirm the findings from the examination of the microstructure. The lessons learned from the investigati on, as reinforced during every step of analysis, was that for sufficient recrysta llization and homogenization to occur, the material must undergo thermo-mechanical proc essing using relatively high deformation, temperatures above the -solvus, and heat treatments of sufficient length to achieve a microstructure similar to the original airf oil. Additional work is needed to refine the process window for the thermo-mechanical processing of the deposit.

PAGE 25

CHA PTER 2 BACKGROUND Introduction to Superalloys In the early part of the 20t h century the jet engine was developed. This type of engine worked by combining an internal com bustion engine with a rotating turbine as seen in Figure 2-1. Air is pulled into the front of the engine via the intake and rotating fans. Several stages of fans compress the ai r intake to high pressures before fuel is injected. The air/fuel mixture combusts, which increases the temperature and volume and as the hot air expands, driving the turb ine blade section and continues to expand out of the nozzle at the rear of the engine. This expansion provides thrust as well as rotating the turbine to power the compre ssor and fan sections of the engine [1]. The turbine blade section is called the hot section of the engine because it is the sole location of the combusted gases and in modern engines; it is not unusual to see gas temperatures around 1600C [2]. Since V on Ohain and Whittles first jet engine in 1940 there has been a 700C increase in turbine entry temperature (TET) which had the unintended consequence of increas ingly high temperatures in the compressor section of the engine [3]. This increase in TET can be seen in Figure 2-2. The Junkers Jumo 004B, the jet engine in the Messerschmitt Me 262, used aluminum alloy compressor blades where in a modern military engine this sa me alloy would be at risk of melting [4]. These modern engines are run at higher temperatures and rotational speeds, creating new materials challenges for engine designers. This research involves the materials in the back of the compressor section, which sees temperatures significant enough to warrant the use of nickel-based superalloys. 25

PAGE 26

The environment (temperature and stress) that turbine blades endure during service necessitates the careful materials se lection and processing that allows for safe operation. Turbines in service face gas temp eratures close to the typical melting range of most metallic alloys, corrosive environments and substances including, but not limited to, hydrocarbons (fossil fuels) and sea air. Thermal and mechanical cycles that impose fatigue cycles on the structure and rotations speeds that place incredible stresses on the blades are also part of the service c onditions. These issues, when summed, require the use of nickel-based supera lloys. This alloy system is c apable of high creep strength, corrosion resistance, and microstructural stabili ty up to 80% of the melting temperature [5]. Microstructure and Strengthening Nickel-based superalloys derive their balanc e of properties from the following. Crystal Structure Microstructure Composition and Alloying Processing and Heat Treatment All of these things work together synergistically to result in a material, which can be used in gas turbines. Crystal Structure and Microstructure Nickel based superalloys are at their core precipitation st rengthened alloys that have what is commonly referred to as a brick and mortar structure seen in Figure 2-3. This structure feature the cuboidal phase (brick) and the surrounding area is the matrix (mortar) [5]. In typical nickel-based superalloys, the phase is a L12 ordered structure or NaCl (rock salt) structure. Gamma prime is an ordered, coherent precipitate in the form 26

PAGE 27

of Ni3(Al,Ti,Nb,Ta) [6]. These Ni-Ti-Al precipitates provide significant creep and low temperature strength. It is the most important part of the alloy system, as it increases in strength as temperature rises without being embrittled [7 ]. The mechanisms for this strengthening are discussed later in this section. The other phase that appears in the microstructure is the phase otherwise known as the matrix. The matrix has an FC C Ni structure and is an austenitic structure with significant alloying additions. The matrixs contribution to strengthening is in the solid solution strengthening with the addition of qualifying elements [8]. Representations of both and phases can be seen in Figure 2-4. Chemistry and Strengthening Nickel-based superalloys usually cont ain 12-15 elemental additions and each plays specific roles including solid soluti on strengthening and precipitation hardening (Table 2-1). Solid solution strengthening pr ovided by elements like Co, Cr, Fe, Mo, W, Ta, and Re act to strengthen the phase. This solid solution str engthening provides significant low temperature strength, but this effect dec reases as temperatures increase, because diffusion becomes easier allowing the elements to diffuse in the ma trix, though even at high temperatures the heavier elements lik e Re will still be effe ctive. At higher temperature the precipitates, provide the creep strength nec essary to function in the engine. As noted, gamma prime is the intermetallic phase that strengthens the alloy. The lattice mismatch between precipitate and matrix is very limited, which contributes to the long-term stability of the alloy system [5, 8]. These precipitates make up a significant 27

PAGE 28

portion of the alloy, in modern alloys it c an be in excess of 60% volume fraction '. The service temperatures of the airfoils woul d normally mean that precipitates would overage and lose strength but the low misfit strain as well as smaller particles in the matrix from the second aging heat treatment prevent precipitate growth and maintain its strength and structure. These smaller particles are called secondary phases and provide strengthening to the alloy [9]. The effect of te mperature on these alloys can be seen in Figure 2-5. Performance vs. Cost Superalloys have been in a constant state of technological evolution since their inception, from the original wrought alloys to the cast alloys and then directionally solidified (DS) and single crystal (SX) alloys as seen in Figure 2-6. This constant improvement has been responsible for the increases in performance of the gas turbine in the last 60 years. In SX alloys, the microstructure reflects the fact that there are no grain boundaries as the whole blade is a continuous single crystal. In contemporary engines, single crystal alloys are used in the hot section of the engines, as they are the only materials able to withstand the operating conditions. The increased performance in the hot section has driven an increase in turbine entry temperature (TET) leading to increases in overall performance. This performance has an unexpected effect on the front of the turbine in that these increased temperat ures are being pushed into the compressor section. The compressor section in high performance and military engines now require nickel-based superalloys to handle the high tem peratures and high stresses but, as they are still at temperatures lower than the turb ine section, less advanced alloy systems and 28

PAGE 29

polycrystalline materials can be used to save co st. The alloy used in this study is IN100. In order to produce a hi gher performance engine, designers have used more tools than just the materials in the engine to ac hieve their goals. Most commercial engines use mechanical fastening techniques to assemble the blades and discs; however for lighter weight and better performance, t hese components should be one piece, also known as an integrally bladed rotor (IBR) IBRs take the usual disc and blades and make them into one piec e by a forging process. This process accomplishes goals su ch as saving weight, increasing strength, and decreased complexity [10]. Saving weight is important as it allows the IBR to spin faster and gives the engine better thrust to weight ratio; wh ile increasing the strength allows the structure to be abl e to spin at faster rpms without damage and wear. Also the number of parts decreases complexity in t he compressor section. These improvements come at a cost, for if damaged, the entire IBR has to be replaced even if the damage is small. High replacement costs provides motiva tion to the manufactures to develop a solution to damage to the IBR. The proposed re search is an attempt to explore potential methods to repair the damaged IBRs so that they can be put back into service after damage/wear without a loss in the properties or perform ance. One potential repair process consists of cutting the worn/dam aged part of the compressor blade, and then rebuilding it with a deposited metal. The newly repaired blade is machined into final shape and then heat treated to re gain its former properties. 29

PAGE 30

Materials Selection and Deposition Processes The material selected for the IBR was IN-100, an alloy commonly used in IBRs. This material was chosen for several reasons, most importantly it s strength and ability for superplastic forming. IN-100: Characteristics, Fabr ication and Superplasticity IN-100 is a polycrystalline nickel based super alloy that is now used primarily as a disk alloy [10]. It is used for disks present ly because of its high temperature strength and ability to form small grain size. The IN -100 used for the IBR is a modified low carbon variant with the compos ition seen in Table 2-2. The low carbon alloy is used because it in creases the medium temperature creep and tensile properties. These pr operties improve since, wit h lower carbon content, the alloys propensity to form TiC at the prio r particle boundaries decreases. The titanium carbides have the effect of slowing gr ain growth by pinning the boundaries [11]. However, the most important pr operty of IN-100 for the fabric ation of IBRs is its ability for superplastic forming. Superplastic behavior is observed when a material resists necking during tensile testing resulting in large tensile strains wit hout failing [12]. The properties necessary for superplastic behavior are a high gamma prime volume fraction, to stabilize the matrix grain size, and a fine grain structure [10]. Re ichman was one of the first to show that argon atomized IN-100 becomes superplasti c when consolidated us ing hot extrusion [13]. This processing route creates ultrafine grains (on the order of microns) that allow for this behavior [14]. Initially the proce ss was used to create a blade that could be compared to its cast and wrought counter parts, but it was discovered that the 30

PAGE 31

superplastic material exhibit ed better stress rupture properti es. Hot extrusion breaks the particles up through shearing, allowing fo r less debris at the prior particle [14]. Pratt & Whitney developed Gatorizing, wh ich is a method of superplastic forming that, can create full parts with close toleranc es [14]. Patented in 1970, the process can create compressor disks from a powder bill et of IN-100 [15]. Gatorizing begins by creating a ultrafine recrystallized part using a deformation process like hot extrusion. The powder is extruded at a te mperature below the recrysta llization temperature or the solvus temperature. The strain and the adiabatic heat fr om the extrusion process push the part above the recrystallization temper ature, creating a fi ne-grained structure [10]. This ultrafine-grained material can exhi bit superplasticity in forging operations at slow strain rates, then the net shape co mpressor disk can be further heat treated providing grain coarsening wh ich improves the high temper ature properties as seen in Figure 2-7. The figure shows the room temper ature properties of IN-100 where the asextruded condition shows the highest strength; however at high temperature the grain coarsened sample would be t he most creep resistant. This process can also be applied to the f abrication of IBRs. In IBR fabrication the superplastic forming causes the material to flow throughout the di e, creating one solid piece that is both compressor disk and blades. Before the Gatorizing process and IBR fabr ication begins, it is necessary to have the correct powder to get the desired fine grain size. Though there are many methods to create the powder, the most popul ar is inert-gas atomization [10]. In this process, the molten alloy is poured through a nozzle where an inert gas (argon in this case) is blasted at the stream. The particles cool quickly and the size can be tailored by the 31

PAGE 32

atomization parameters. The resulting powder particles are then compacted by extrusion. It has been shown that the us e argon atomized powder results in good properties after the extrusion process [14]. It was for these reas ons that the powder used in this study was an argon atomized ve rsion of IN-100. The powder as seen in Figure 2-8, contained different si ze particles to facilitate complete filling and density of the final part. Since the IBR is being used in the compre ssor section, creep strength is the most critical factor, so a fine-gr ained structure is desirable becaus e of the Hall-Petch effect (Equation 2-1). (2-1) This relationship shows that as grain size (d) decreases, room temperature strength ( y) increases [16]. For this reason, argon at omized powder will allow for a more refined structure because of the growth of MC ty pe carbides on the prior particle boundaries [14]. These carbides restrict grain coarsening in post weld heat treatments by pinning the grain boundaries. However, the most ap pealing reasons to use argon atomization from an industrial standpoint are its ease of mass production and low-cost processing [10]. Deposition Process Superplastic behavior and Gatorizing are impor tant parts of t he IBR fabrication process, but may complicate the repair pr ocess. When an IBR comes out of service because of wear or damage, one method incl udes the removal of the effected area 32

PAGE 33

(usually the leading edge and then virgin materi al is deposited to rebuild the blade. The microstructure formed is similar to other deposition techniques, as seen in Figure 2-9 In this process, powder is fed into a defocused laser beam that passes across a substrate. The laser melts the powder, which solidifies on the substrate to build up the structures. This builds a 3D layered structure, which can be machined to form the repaired blade [17]. As each layer is deposited, heat from the source (l aser) escapes into the substrate/base metal which acts as a heat sink. This allows dendritic growth along the [001] easy growth direction for FCC materials. As dendrites grow, a process called coring occurs. Coring is a type of microsegr egation where certain solutes are rejected from the solidifying dendrite arms into the interdendritic liquid, causing segregation [18]. When the source passes over the previous ly deposited layer, which has solidified, partial re-melting of the deposited layers allow columnar grains to grow. These columnar structures form becaus e the imprecise natur e of the source melts the previous layer, allowing the dendrites to continue gr owing in roughly the same orientation. The partial re-melting of the layers also cause the formation of layer ba nds that indicate the depth of re-melting and a change in dendrite orientation. Hussein et. al. have shown through studies involving laser deposition of Waspaloy, that as the number of layers increase, the cooling rate decreases due to distance from the heat sink (base metal) [17]. This was demonstrated by measuring the secondary dendrite arm spacing (SDAS) in each layer. The SDAS was reported to increase with distance from the base. This increase in SDAS also correlates with an increase in solidification time or slower rate of solidification. The changes in solidification times from layer to layer and the repeated 33

PAGE 34

heating and cooling cycles due to heat flow from more recent layers to the base, can also buildup the thermal strain within the mate rial. This strain is leveraged to produce the recrystallization and thus grain refinement. The change in solidification times has other consequences as evidenced by the change in hardness between layers. With more solidification time, more gamma prime nucleates in the layer, thus making a higher volume fraction and increasing the hardness [17]. The deposited material is homogenized by post-deposition heat treatments, which must include enough time and temperature to allow for nucleation and growth of primary gamma prime as well as the growth of secondary gamma prime along the entire length of the deposit. This pr ocess, along with the grain refinement, can lead to good high temperature pr operties for the repaired blade. LENS Process The LENS or Laser Engineered Net Shaping was developed at Sandia National Labs as a rapid prototyping/manufacturi ng technique; however in has been used especially in the defense industry as a repai r technique [19]. Parts in military service often experience aggressive environments and stress resulting in shortened lifetimes. Also these parts are often made of advanced materials that offer limited repair/refurbishment options. For example the Anniston Army Depot use LENS systems to quickly replace or repair turbine co mponents for the LENS process [20]. LENS offers several potential benefits ov er the PPD process. The LENS process is computer aided design (CAD) controlled which allows for fi ner control of the deposition. As the for the deposit ion process itself, due to a more targeted heat sources and subsequently finer weld bead, there is also a smaller heat affected zone (HAZ). The 34

PAGE 35

lower heat input and faster process leads to finer microstructures than can be seen in other repair processes [19]. The LENS process, as seen in Figure 210, works by coupling a high power laser (Nd:YAG or Fiber) with a computer control in an argon atmosphere. The laser creates a melt pool into which powder is injected; t and the laser in scanned across the sample in order to build the desired shape [19]. Post-Deposition Processing Recrystallization: Kinetics and Residual Stresses The most important part of the grain refinement process is recrystallization. The post-deposition treatment provides thermal ener gy to give the structure the impetus to recrystallize. The energy stored in the strain fi elds of the samples is the driving force for the nucleation and growth of recrystallized grains. During recrystallization new stressfree grains grow at the site of the former strain fields. This energy is stored in the form of residual stresses, put into the materials by the proc essing route. However, the amount of energy stored in the material is finite, depending on several factors such as materials, deformation, te mperature, etc. [21]. The post-deposition heat treatments for this system are based on the annealing process with the following steps: recovery, recrystallization and grain growth. Recovery can be defined as a process that occurs befor e the nucleation of stress-free grains [21]. As soon as new grains nucleate, recrysta llization has begun. Grain growth is the process whereby the average grain size of the sample increases after it has been recrystallized. For the purpose of this project, recrystallizat ion is the crucial process. Recovery is undesirable because it doesnt actually result in any refinement of the sample [21]. Similarly, grain growth will cause the grain size to increase, decreasing the 35

PAGE 36

strength of the materials as per the Hall-Petch relations hip. Optimal heat treatment design will therefore limit recovery and eliminate grain growth. This can be accomplish ed by analyzing the kinetics of re crystallization of the deposited material. Being a thermally controlled process, recrystallization requires minimum activation energy to begin the process. It has been shown experimentally through calorimetry that recovery and recrystalliz ation are competing processes and that recovery has lower activation energy. Asc ending to higher temperatures activates recrystallization and skipping the recovery stage altogether [21]. Empirical evidence shows that the amount of recr ystallization in a material is a function of the time and temperature of the heat tr eatment and the time and temper ature are inversely related [22]. This can be seen in Figure 2-11, whic h demonstrates that as time and temperature of heat treatments increase t he amount of residual stress in the materials decreases, showing that RXN has also increased. In a practical sense this means that it is not possible to achieve recrystallization with a relatively low temperat ure and short timeframe without significant residual strain and stress, so either low temperature or short time can be used. The two prerequisites for activation of the recrystallization process are residual stress in the material and a base thermal energy. Burke and Turnbull, in their Rules for Recrystallization, stated that there is a minimum amount of deformation necessary for recrystallization and that the greater the degree of deformation, the greater the recrystallization [23]. 36

PAGE 37

Thermo-Mechanical Processin g With the ultimate goal of inducing RXN in the workpiece in order to get a grain structure that matches the rest of the IBR, it is desirabl e to increase the distortional energy available for RXN. To this end, ther mal processing (heat tr eatments) may not be sufficient to provide this energy, thus me chanical deformation could be added to the process route. Before any mechanical processing, the onl y residual stress in the airfoils comes from the deposition process and is a result of cooling stress. The size of the airfoil and the heat involved in deposition cause rapid cooling to occur in the layers as they are deposited and re-melted with successive passes; the rapidly cooled structures leave compressive stress on the surface and tensile st resses in the core of the airfoil [24]. After the deposition process even more stor ed energy can be put into the material by mechanical deformation or cold work. The cold work in this case is preferable to hot working for several reasons. In order to refi ne the grain structure, which is the primary objective of this project, the working tem perature of the process must be greater than the RXN temperature in order to have produce dynamic recrystallization of the microstructure [25]. This can be problematic as it is difficult to maintain the high temperature need for RXN during working in both laboratory and industrial settings with such small workpieces. Poor temperatur e and deformation control will lead to nonuniform properties and microstr ucture. Most importantly, hot working such a small airfoil in-situ is just not feasible. These disadvan tages leave cold working as the only real option. Cold working besides being easier to do a llows for a better surface finish, which can be critical if coatings are necessary dow n the line. In addition for the relatively low 37

PAGE 38

38 deformation levels and thin section, cold wo rking provides more accurate dimensional control and better strength and hardness values [25]. The total deformation level imparted dur ing deformation is of higher importance than the manner of the cold working and th ermo-mechanical processing (TMP). Where rolling can be a valuable tool in a research setting, allowing for controllable, uniform deformation, this proves nearly impossible to do to an IBR. Therefore, methods of in-situ deformation such as shot peening were included in this investigation.

PAGE 39

Figure 2-1. Schematic of Jet Engine [Reprinted with permission from Roger C. Reed, The Superalloys, 2006 p. 3, Fig. 1-3] Figure 2-2. Turbine inlet temperatures in Roll-Royce civilian engines 1940-2010 [Reprinted with permission from Roger C. Reed, The Superalloys, 2006, p. 5, Fig. 1-5] 39

PAGE 40

Figure 2-3. Nickel based superalloy microstructure [Reprinted with permission from Harshad .K.D.H Bhadeshia, Nickel Based Superalloys University of Cambridge Website] Figure 2-4. Schematic of gamma and gamma prime crystal structures [Reprinted with permission from Harshad. K.D. H Bhadeshia, Nickel Based Superalloys University of Cambridge Website] 40

PAGE 41

Table 2-1. Roles of elements in superalloys [Reprinted with permission from Superallo ys ASM Handbooks, 2003, Table 1(b)] 0 200 400 600 800 1000 1200 020040060080010001200 Temperature (oC)Strength (MPa) UTS 0.2% YS Figure 2-5. Effect of Tem perature on IN-100 properties [Reprinted with permission from Superalloys ASM Handbooks, 2003, Table 7] 41

PAGE 42

Figure 2-6. Temperature Capabi lity of Superalloys (C) [Reprinted with permission from Roger C. Reed, The Superalloys, 2006 p. 5, Fig. 1-5] Table 2-2. Compositio n of Modified IN-100 [Reprinted with permission from Gernot H. Gessinger, Pow der Metallurgy Superalloys, 1984, p. 133, Table 5.1] 42

PAGE 43

Figure 2-7. Room temperature properties of IN-100 based on processing history [Reprinted with permission from Gernot H. Gessinger, Pow der Metallurgy Superalloys, 1984, p. 159, Figure 5.21] Figure 2-8. SEM image of the IN-100 powder at 500x 43

PAGE 44

Figure 2-9. Cross-section of laser weld deposited structure [Reprinted with permission from N.I.S. Hussein Microstructure formation in Waspaloy multilayer builds following direct metal depos ition with laser and wire, 2008, p. 263, Figure 3] Figure 2-10. The LENS process [Reprinted with permission from Neil Calder and Martin Hedges: Near Net Shaped Rapid Manufacturing and Repair by LENS, p 13-2, Figure 2] 44

PAGE 45

45 Figure 2-11. Surface residual ratio vs. annealing time for carbon steel [Reprinted with permission from J. Hoffman et. al: Relaxation of Re sidual Stresses of Various Sources by Annealing Residual Stresses in Scienc e and Technology, p.696]

PAGE 46

CHA PTER 3 EXPERIMENTAL PROCEDURES The purpose of the investigation described below was to evaluate the use of either heat treatment alone or to combine deformation and heat treatment to alter the microstructure and properties of a repaired airfoil in an IBR. Materials The material used in this investigati on was a powder metallurgy (PM) alloy, which is essentially IN-100, whose compositi on has been modified for PM processing (Table 3-1). IN-100 is a polycrystalli ne nickel-based superalloy and can be used in the latter stages of the compressor, where high perform ance (high TET) is needed. Two different conditions of IN-100 were evaluated in this study. The first materi al condition was the Gatorized IN-100, which was the base metal for the IBR, and was, t herefore, considered the Baseline material in this study. T he deposited, or repair material was the second condition examined in this study. The baseline material was machined from a Gatorized forging that was fully heattreated by Pratt & Whitney. The first depos ited material was fabricated by Pratt & Whitney using a proprietary plasma deposition process (PPD) to build up layers on a IN100 (Baseline) substrate as seen in Figur e 3-1. The as-deposited, or starting, microstructure exhibited large columnar grains and residual den dritic structure, while the desired final microstructure (baseline) was characterized as a fine-grained microstructure with small equia xed grains. In general, the baseline microstructure would be expected to result in high strength and duc tility, in comparison to the coarse, assolidified microstructure in the as-deposit ed material [10]. Th is equiaxed structure should be similar to microstructure of the intact portion of the blade and rotor. 46

PAGE 47

The second part of the study utilized the same powder metallurgy alloy as the first part however the deposition process changed to the Laser Engineered Net Shaping (LENS) process. This process uses a laser in stead of plasma as the heat source to melt the powder and deposit it on the substrate. This process results in a sample that has smaller layers than the plasma process (Fi gure 3-2). Due to this difference the LENS material shows a more homogenized structure and smaller dendrites. Heat Treatments In the initial portion of this program heat treatments with rapid heating/cooling rates were used to develop internal strains. These internal strains were intended to induce recrystallization (RXN), which would refine the microstructure and achieve the desired finer grain size. In order to develop internal strains dur ing the heat treatment, it was necessary to control the heating and coolin g rates of the samples. Fast heating rates would be expected to develop greater levels of stra in because the more rapid heating/cooling rates would prevent the relieving of the inter nal strains developed in the sample. Slower heating rates would be reduce the internal strains developed in the samples during heating/cooing that occurs dur ing the deposition process and would be undesired as it would reduce the overall amount of energy available for RXN [26]. To achieve a rapid heating rate, the samples are charged into a pre-heated furnace already at the appropriate heat treatment temperature. Since it was the ultimate goal of this project to develop a heat treatment sequence for industry, the heating rate used in this study had to be evaluated, compared to the heating rate expected in industry. Using several K-type thermocouples, the temperature at specific time intervals were noted and graphed (Figure 3-3). It was determined that 47

PAGE 48

the heat-up rate of the labor atory furnace did not surpass the capabilities of the proposed industrial method. Ther efore, the rapid heating rate used in the laboratory was considered the baseline for all subsequent thermal cycles. During the thermal processing, the cooling rates were also important in some heat treatments because rapid cooling can provide addi tional thermal stra in/energy available for RXN [26]. In order to eval uate the effect of cooling rate on the microstructure of the heat-treated deposits, several different cooli ng techniques were examined. The quickest cooling rate was obtained using a water quench; in this case the sample was dropped in a metal bucket containing approximately 1 gall on of room temperature water. Though water cannot reach the cooling rate of iced brine, water was judged to best approximate the maximum cooling rate achievable by in dustry, The next quickest cooling method examined was (air) jet cooling. In this case, the sample was held underneath an outlet for laboratory compressed air. Fan air coo ling or directed air cooling (DAC) which utilized a desk fan to blow air over the spec imen was also evaluated. However, the fan air cooling did not appear to provide any increased cooling rate and was only examined in a limited number of experi ments. The cooling method most commonly used, which also providing the slowest cooling rate was ai r cooling in still laboratory air. During air cooling, the sample was remo ved from the furnace and plac ed in a ceramic tray. These are just several of the many ways samples c an be cooled, but they were chosen to best replicate possible cooling rate s achievable in industry. The heat treatments that were used in th is study were intended to recrystallize the microstructure of the deposited material. The thermal cycles were selected to take advantage of the strain in the material from the deposition pr ocess and from the 48

PAGE 49

heating/cooling cycles, to induce RXN. The heat treatment test matrices included examining the effects of bot h heat treatment times and tem peratures. Table 3-2 shows the initial sample matrix. Heat treatment temperatures ranged from 1093C to 1232C. This temperature range, therefore, included both sub-solvus and super-solvus temperatures with the solvus of around 1177 C. Additionally, the effect of heat treatment hold time was investigated by including 30 minute and 120 minute heat treatments with the s tandard 60 minute treatments per formed on all the samples. Many recrystallization treatm ents are longer than even 2 hour s; however the longer the treatment the greater chance of inducing grain growth on the original blade segment. This resulted in the impetus to examine shorter treatment s. Subsequent heat treatment matrices or iterations were selected based on the results up to that point. The results from the initia l heat treatment matrices clearly indicated that a single thermal cycle was not sufficient to devel op strains in the deposit to initiate recrystallization. Based on these initial resu lts from the single heat treatment cycle, ratcheting treatments were developed to evaluate thermal cycling. The thermal ratcheting was intended to devel op greater strains in the sample, which in turn would cause recrystallization in the samples (Figure 3-4). These ratc heting heat treatment utilized two short term (15-minute) thermal cycles with rapid heating followed by fast cooling before the standard 1 hour heat treatment. The high temperature heat rate, along with the fast (typically water cooli ng), combined to add ther mal strain to the structure. 49

PAGE 50

All the heat treatments were per formed in a Carbolite CWF 1300 air box furnace at the University of Florida, monitored by two external K-type t hermocouples to ensure the accuracy and precision of the heat treatment temperature. Mechanical Deformation Based on the results of the thermal treat ments, it was determined that additional methods of imparting strain in the samp les needed to be evaluated. Therefore, the effect of mechanical deformation in concert with thermal treatments on the recrystallization behavior of the deposits was examin ed. Different amounts of deformation were used to help establis h a processing window for subsequent experiments. A feasibility study was undertaken using a rolling mill to provide controllable, uniform deformation for test coupons. These c oupons were rolled initially to either a small amount (low) deformation condition (4-6% RIA) or a greater amount (high) of deformation condition (8-10 % RIA). Befo re rolling could commence, rolling force calculations were performed to determine if the intended deformation would exceed the threshold of the equipment. The results of th is calculation can be seen in Appendix A. All rolling trials were performed on a two high Fenn rolling mill at the University of Florida. Since the surface of the as rece ived deposit samples was made up of multiple layers, the surface was not flat and was not suitable for rolling. Therefore, the asreceived deposit samples were machined to produce flat surfaces and a thickness of 0.060, which was the same thickness as the baseline metal samples. This allowed for the base metal sample to be used to set the rolls for the deposited samples as well as simulate the final airfoil thick ness after repair. With the rolls set, the samples were lightly lubricated using Mobil gear oil and sent through the mill. The samples were then 50

PAGE 51

measured by micrometer for final dimensions. The measurements were later used to determine the reduction in area or deformation level of the samples. The succes s of the thermo-mechanical processing led to the in vestigation of other more industrially applicable deformation methods. Other manners of deformation were investig ated including low plasticity burnishing (LPB), shot peening, laser shock peening (LSP) and a proprietary in-situ deformation process during deposition. These deformation processes were performed out of house. Mechanical Testing For the process of heat treatment and recrystalliza tion evaluation, hardness testing was used to determine deformation depth profiles and relative hardness values for comparing results of thermo-mechanical processing routes. Hardness testing was performed on a Buehl er Micromet II Vickers Hardness testi ng machine using 1kg/ft. load and a loading time of 15 seconds. Values were converted to Rockwell C using the ASTM standard E140 [27]. Sample Preparation The base metal and deposit samples were sectioned to produce heat treatment samples using two techniques. Smaller samples were sectioned using an Allied diamond saw with oil cooling and larger pieces were sectioned at Precision Tool & Engineering (Gainesville, FL). If the samples were of limited thickness, they were electrodischarge machined (EDM) to minimize sample loss, at Advanced Manufacturing Techniques in Clifton Park, New York. The samples were then mounted longitudinally in clear epoxy resin for metallographic sample preparation. Polishing consisted of two stages, performed on a LECO Vari/Pol VP-55 polishing wheel with water lubrication. The first stage, grinding, used Buehler SiC abr asive papers of ascending grits from 240 51

PAGE 52

to 1200. The second stage, polishing, used mi crofiber polishing clot hs and polishing suspensions consisting of alumina and water. The alumina used for polishing was in a decreasing size: 14.5, 5, 3, 1 and 0.3 m. The mounted samples were cleaned between steps in water or methanol in an ultrasonicator bath following the final grinding and polishing steps. After the sample was polished, the surf ace was etched in order to observe the desired microstructural features. Initially se veral etchants were used to examine the grain structure: AG21 (25 mL C3H6O3 acid + 15 mL HNO3 + 1 mL HCl) PW #17 (100 mL H2O + 100 mL HNO3 + 100 mL HCl + 3 g MoO3.H2O). Kallings Waterless R eagent II (50mL HCl + 50mL C2H2OH + 3g CuCl2) Neither of these etchants produced sufficient grain contrast (Figur e 3-5) to see the desired features; so another type of etching was needed. Pratt & Whitney provided the specifications for an electrolytic etch using a 50/50 solution of hydrochloric acid (HCl) and methanol In this case, the etching solution was placed in a glass beaker with the mounted sample in it and stainless steel bolt (positive). A voltage of 1 to 5 V was applied to the sample via a metal rod (negative). The rod was touched to the surface of the sample, while the stainless steel bolt was moved around the surface of the sample unt il a green haze formed on the sample. The green haze indicates a proper etching c ondition. Then the etched microstructure was checked under a laboratory light optical microscope. For the LENS samples, Kallings Waterless Reagent II was used to great effect to bring out the grain microstructure. The etchant was swabbed onto the sample surface using a cotton swab for between 10-40 seconds. 52

PAGE 53

Characterization The samples were characterized using a variety of techniques to determine the efficacy of the thermo-mechani cal treatments in homogenizing the sample and refining the grain size. Light Optical Microscopy Light optical microscopy was used to ex amine the microstructure of each of the samples. Due to the relative coarseness of the microstructures, light optical microscopy provided the best grain contrast for the as-deposited and heat treated samples. Three light optical microscopy systems were used in this investigation: a LECO Metallograph with SPOT camera, an Olympus BX60 also with a SPOT camera and Leica DM2500 with ProgRes Capture 2. 7 software. Samples were imaged at magnifications of 50x, 100x, 200x and 500x. Scanning Electron Microscopy The scanning electron microscope (SEM) was also used in cases when higher magnification imaging (>500x) was needed, to examine the / microstructure and for compositional information. Most of the images were taken in secondary electron (SE) mode for microstructural and t opographical analysis. Energy dispersive spectroscopy (EDS) was used to gather semi-quantitative compositional information about phases, carbides, inclusions etc. The SEM used in this study was a JE OL 6400 with EDS and BSE detectors for compositional data. X-Ray Diffraction Additionally, x-ray diffra ction (XRD) was used to evaluate the qualitative differences between strain conditions of sample s in different states (i.e. as-deposited, 53

PAGE 54

machined deposit and annealed depos it). The XRD sy stem used in this study was an APD 3720 Powder Diffractometer. Powder X-ray diffraction is a technique where a sample is bombarded with x-rays as the detector is rotated through angles (2 ) from roughly 0-180 while the sample maintains the same orientation see Figure 36 [28]. The output is graphed intensity vs. 2 Quantification As an additional measure to judge the efficacy of the thermo-mechanical processing on the final microstructure it was necessary to quantify the grain size of the grains that have recryst allized from the initia l dendritic structure. There are many established grain size methods for equiaxed grains including Heyn lineal intercept, Planimet ric (Jeffries) and circular in tercept methods. [29] For nonequiaxed and multiphase materials these sa me techniques can be applied with some variations [30]. None of these methods c an be applied consistently to the micrographs from this study because not all the samples ex hibit a full grains structure, one where the grains are all touching and a grain boundary becomes an intersection point for measurement. Additionally the size of t he grains varies widely between samples, depending on the processing route, that there are not enough gr ains in each micrograph to get accurate measurements. Automatic analysis techniques such as t he use of intensity hi stograms on the SEM were tested. The differences in grains contrast and the inability to consistently detect grain boundaries made the technique insufficien t. For the same reasons the standard methods proved ineffective, automatic analysis methods that determined grain size were deemed inadequate for this application. These difficulties necessitated coming up 54

PAGE 55

55 with a novel method that allowed for adequate characterization of the grain sizes that resulting from the processing methods. The method that allowed for the ease of measurement and usable results was to treat the grains as ellipses, due to the non-equiaxed nature, and measuring the lengths of the major and minor axes. The lengths were measured with the aid of the free software, ImageJ, wherein the scale was used as the basis for measurements made on the screen. The measurements were taken at 100x fo r the sake of consistency, but other magnifications were used if the grains we re particularly large or small. Three micrographs were used per sample with 5-10 grains measured from each micrograph. Also, grains that were not fully encompassed by the micrograph were assumed to be twice as large as seen in the micrograph. This number comes from the ASTM standard in that when the line terminates in the grain it is counted as having half an intersection. The grains chosen for measurement had clearly defined boundaries, and were representative of the size di stribution within the micrograph. For samples with a variety of large and small grains, measurements were taken of grains that would show this distribution. With measurements of t he major and minor axes, other data points such as the aspect ratio of the grains and the grain area (area of the ellipse) as seen in Figure 3-7. Additionally, using grain area numbers from the ASTM standard, the sizes of the RXN grains were compared the ASTM grain size numbers commonly used in industry, just for comparison purposes. The me asurement of individual gr ains cannot be put into ASTM numbers [30].

PAGE 56

Table 3-1. IN-100 Com position [Reprinted with permission from Gernot H. Gessinger, Pow der Metallurgy Superalloys, 1984, p. 133, Table 5.1] Figure 3-1. Layer build during deposition process [Reprinted with permission from N.I.S. Hussein Microstructure formation in Waspaloy multilayer builds following direct metal depos ition with laser and wire, 2008, p. 263, Figure 3] Figure 3-2. LENS and PPD samples 56

PAGE 57

Figure 3-3. Heat-up Rate Diagram Table 3-2. Initial he at treatment matrix 57

PAGE 58

Figure 3-4. Schematic of ratcheting heat tr eatment showing the sample temperature versus time 58

PAGE 59

A) B) C) Figure 3-5. SEM photomicrographs of deposited (PW1074) samples etched with various etchants. A) AG21 B) PW #17 and C) Kallings Reagent 59

PAGE 60

60 Figure 3-6. Powder XRD Schematic [Reprinted with permission from Identification of Compounds and Phases Using X-Ray Powder Diffraction, ASM Handbooks, ASM, 2003, Figure 8] Figure 3-7. Schematic of calculation for grain size analysis

PAGE 61

CHA PTER 4 RESULTS THERMAL TREATMENTS This section represents the beginning of the study from the development of the first heat treatment to the ev aluation of the factors in the refurbishment process and finally the evaluation of the results from performing only thermal treatments on both the PPD and LENS samples. Development of Heat Treatments The first heat treatment matrix was development with an eye toward industrial feasibility and evaluation of different times and temperatures The matrix was defined by the heat treatment window or the difference between the solvus and incipient melting temperatures [31, 32]. The incipient melti ng temperature is defi ned as the temperature where the retained eutectic that is in the interdendritic region melts [33]. This melting causes a decrease in properties and changes the microstructure including the formation of porosity n additional eutectic regions. For IN-100, as with most alloys, the heat treatment window varies with t he processing route and the c ondition of the material. A sample of the same alloy prepared by i ngot metallurgy (I/M) and powder metallurgy (P/M) routes will have different heat treatment windows due to differences in composition and processing [34 ]. Composition changes also impacts this heat treatment window (Figure 4-1), and it c an be seen that unmodified IN-100 has a tight window. The same is true of an alloy that is cast (l arge grains and segregated structure) compared with wrought (small grains and more homogenous structure). Since the original PPD material is a combination of as-cast solidified and wrought structures and its composition varies from traditional IN-100, literature values for the solvus and 61

PAGE 62

incipient melting temperatures were used as guidelines. Table 4-1 shows heat treatment window values from various sources in the literature. The heat treatment methodology was first conceived along the lines of a standard nickel-base superalloys heat treatment with s eparate solution and aging steps to get the optimal microstructure as seen in Table 42 [35]. The proposed heat treatment (HT) was to be just the solution heat treatment with aging to follow. Standard IN-100 heat treatments from t he literature were considered at but ultimately not included due to the difference in starting conditions and the PPD and LENS materials. The baseline heat treatment for IN-100 is 1121C (2050 F) as indicated in the specifications for the powder version of IN-100 used in the deposition process. This and other temperatures were considered based on the starting condition of the P/M product. Figure 4-2 shows the grain size of samples of P/M IN-100 based on the type of powder used: argon atomized (AA), rotating electr ode process (REP), and dissolved hydrogen process (DHP) and cast and wrought (C&W) for comparison. This shows that for the argon atomized IN-100 used in this study, he at treatments can be performed at higher temperatures than traditional methods due to increased homogeneity and the prior particle boundary TiC particles pinning the gr ain boundaries [36]. Te mperatures were chosen right up to the incipient melting tem perature, which was considered the upper limit of the heat treatment window. The standard nickel based superalloy heat treatment schedule calls for long soaks at temperature for full solutioning; however for this study with the object of producing industrially feasible refurbishment procedure, such long HT times are not appropriate. From an industrial standpoint the longer the HT and the higher the 62

PAGE 63

temperature the more expens iv e the treatment becomes. More importantly for the current study, these treatm ents must be done in-situ due to the nature of IBRs. Its impossible to separate the secti ons of the airfoil, repaired and original, from each other. The original section can be insulated to prevent it from experiencing the same temperatures as the repaired section but it will still experience a thermal cycle. Therefore shorted treatments of 1hour were considered the standard for the initial matrices with the addition of shorter (30 minutes) and longer (2 hours) cycles to be evaluated as well. The result ant heat treatment matrix can be seen in Table 4-3. 1st PPD Heat Treatment Matrix The first matrix was desi gned to explore different ti mes (30 min., 1 hr. and 2 hours) as well as different temperatur es 1093C to 1232C (2000F to 2250F). All samples were air cooled after being remov ed from the furnace. The first round of treatments were performed according to pr otocols laid out in the experimental procedures section; however did not yield any conclusive result other than the need for a fast heat-up rate to effect any meaningful recrystallizati on. The micrographs from this first matrix showed very little impr ovement from the as-deposited dendritic microstructure. This indicated that the samp les must see rapid heatup in order to make the most of the inherent cyclic thermal strain. Having the samples heat up with the furnace as it heated up did little for refinemen t as it relieved the residual stresses without any real refinement or RXN [37, 38]. These samples provided very little recrystallization data, they did allow for the development of etching and microscopy techni ques. Initially several etchants were used to show the desired features until settling on the electrolytic etching procedure for the PPD material. Additionally, images of the samples we re taken on both scanning 63

PAGE 64

electron microscopes and optical microscopes in order to ascertain which would deliver the best grain contrast. Ultimately the opt ical microscope provided the best grain contrast that allowed for the evaluation of the efficacy of the heat treatment procedures. As a result of this preliminary matrix it was decided to repeat the procedure with some changes including speeding up the heat-up rate. 2nd Heat Treatment Matrix Apart from the change in heating rate, the second matrix added ratcheting heat treatments (thermal cycling) in order to increase the amount of strain energy available for recrystallization, aiming to produce a finer grain size [39]. The ratcheting treatments consisted of two shorter thermal cycles wit h rapid heating and cooling rates. In this case, the samples were quenched in water. For this iteration, ratcheting treatments were added to the highest and lowest tem peratures as seen in Table 4-4. As-Deposited The as-deposited sample seen in Fi gure 4-3 featured a coarse dendritic structure. The structure appeared to be heavily segregated as well. This dendritic structure is the result of the deposition process, wher e the powder is heated and deposited on the base metal that represents the airfoil. Here the dendrites grow along the [001] direction, which can be seen running roughly vertical in the micrographs. The properties of samples with the as-deposited microstructure would be inferred those of the fine-grained, homogenized base metal of the blade before replacement, thus additional steps would be needed to devel op a microstructure mo re similar to the Gatorized microstructure. 64

PAGE 65

Sub-solvus The first heat treatment at 1093 C was the lowest to be tested and is 28 degrees Celsius below the baseline IN-100 heat treatment and the results can be seen in Figure 4-4. For the nonratcheted sample (Sample P1) the structure remained mostly dendritic; however the beginning of refinement can be seen. As for the ratcheted sample (Sample P1-R), it had the same coar se dendritic structure as Sample P1 but due to the extra thermal strain it appeared to exhibit a limited amount of refinement. Refinement in these cases can be seen by the appearance of lar ge grains in the micrograph that were not seen in the as-deposited sample. As the standard heat treatment for IN -100, it was expec ted that the 1121C treatment would be more effect ive at refining the microstructure; however the increase in temperature was only slightly more effect ive then the previous treatment (Figure 4-5). The microstructure in Sample P2 remained dendritic if seemingly finer than previous samples. In addition, some new grains appear at this temper ature but they are relatively large in size. The effect of time is also investigated at this temperature, wit h treatments of 30 and 120 minutes in addition to the standard 60 minute sample. The microstructure observed the 30 minute heat tr eatment (Sample P2-30) exhi bited less refinement than was observed after the 60 minute treatm ent and the 120 minute (Sample P2-120) exhibited the greatest amount of refinement. However, ev en with longer times the heat treatments at this tem perature did not achieve the desired results, indicating that this temperature is still below the recrysta llization temperature for the deposit. 65

PAGE 66

Sample P3 (Figure 4-6) showed much the same structure of the previous samples with a mostly dendritic microstructure; however in this sample the degree of segregation appear to be reduced com pared to previous samples. Super-Solvus By heat treating the post-weld struct ure at temperatur e at or above the solvus, there was RXN over a large scale showin g that temperature were approaching the recrystallization temperature for the level of residual strain and heat treatment lengths of 1 hour. This sample (Sample P4) heat treated at 1177 C is the first in this matrix to show apparent homogenization of the st ructure. The homogenizati on was not complete but can be seen in Figure 4-7B. Also apparent is t hat the majority of the microstructure consisted of large columnar grains, with a few dendritic regions, which can be seen at a magnification of 50x in 4-7A. Sample P5, heat treated at 1204 C, showed further homogenization of the previously segregated st ructure, as would be expected fr om kinetics. The nature of the grains was still mostly columnar but appear to be less elongated. The highest temperature tested was 1232 C and was just below the incipient melting temperature, wher e retained eutectic could have an adverse effect on the properties [c40]. Despite this narrow window, these samples and especially Sample P6 had the best microstructure of the ent ire matrix with seemingly complete homogenization and mostly equiaxed grains (F igures 4-9 and 4-10). Even the 30 minute treatment (Sample P6-30) at this temperature resulted in microstructural refinement and mostly equiaxed structure; however it is less refined then the 60 minute treatment. 66

PAGE 67

On the other hand, Sample P6-R, which experienced a ratcheting treatment at 1232 C, experienced severe cracking. The sa mple showed large intergranular cracks beginning at the edges of the sample and extending inward. Though it had a homogenized and equiaxed struct ure the cracks are clearly a problem and seem to result from thermal shock due to the quenching of the specimen during the high temperature ratchet. Piearcey an d Terkelsen also reported a similar result where high strength alloys such as IN-100 will experie nce grain boundary cracking where lower strength alloys will favor grain deformations in response to stresses [41]. Summary In general, as the temperat ure of the heat treatment increased so did the amount of homogenization and refinement, to the po int, above the solvus temperature, of recrystallizing the microstructure to produc e mostly equiaxed grains. Also as the temperature increased, homogenization of the previously segregated deposited structure was observed. Heat treatment time was also observed to have an effect on the samples tested at different time intervals. The longer the heat treatment the more time diffusion had to reduce the segregation and to allow for recrystallization in the test coupons. 3rd Heat Treatment Matrix With the lesson learned from the previous matrices, the 3rd heat treatment matrix (Table 4-5) was developed to investigate se veral factors seen in the last matrix including: heat treatment time temperature of the ratchet, c ooling rate during the ratchet and solution heat treatment temperature. All the treatments use a solution heat treatment of 1232 C, as it showed the most promis ing results, and though time was 67

PAGE 68

investigated in the last matrix more dat a was needed to fully understand the effect of time of the heat treatments. Additionally, the ratcheting tr eatment from the previous trials used water-cooling; however as this ma y be impractical for industrializ ation of this process and provided perhaps too high of a cooling rate, other methods were examined. Finally, the results from 1232 C /1 hour heat treatment were verified by replicating the sample. Verification Using the same procedure as the previous Sample P6, a new sample was prepared to verify the favorabl e results from the previous matrix. The microstructure observed in the sample was again refined and with a mostly equiaxed grain structure as was seen in the specimen (Figure 4-11). This re sult confirmed that this temperature was indeed above the RXN temperature for t he level of residual stress and time. Directed Air Cooling In testing the effect of cooling rate on the RXN of the deposited samples, two ratcheting treatment temperat ures were used. Although the temperatures were below 1232 C, both low (1093 C) and high (1177C) temperatures were examined from the previous matrix. Samples were tested using the same ratchet treatm ent as sample P1-R except a fan was used to cool the samples between thermal cycles; hereafter designated directed air cooling (DAC). The intent was to investigate the effect of a lower cooling rate than water quenching but faster t han the previously used still air cooling. After the ratcheting cycle, the sample underwent solution heat treatments at 1232 C for both 30 and 60 minutes. 68

PAGE 69

Sample P1R/6-30, which was given a 30 minute heat treatment, exhibited large columnar grains throughout the sample (Figure 4-12), but did not appear to be fully solutioned. The sample given the 60 minute heat treatment (Sample P1R/6) exhibited the same columnar grains structure; however with smaller grains and even some equiaxed grains (Figure 4-13). The presence of columnar grains in both samples indicated that perhaps by using a low temper ature ratchet and slower cooling rate, the stored energy (residual stress) available for RXN was either insufficient or was being used in the recovery process prior to the main heat treatment. Using the same procedure as t he low temperature ratchet (1093C), the temperature was increased to 1177 C to put more thermal strain in the material before the solution heat treatment. It should be not ed that the maximum ratchet temperature was close to the solvus which was empirically dete rmined in this study to be around 1177 C. Heat treatment time was again varied so that the difference between 30 and 60 minutes could be investigated. In Sample P4 R/6-30 (Figure 4-14) t he grains seen were smaller than in the samples with the low temperature ratchet, but were somewhat elongated. While the higher ratcheting tem perature appeared to result in a greater degree of homogenization, the grains were closer to equiaxed with an aspect ratio seeming to approach 1. The 60 minute treatm ent, as seen in Figure 4-15, resulted in more refined, homogenous structure than th e 30 minute one, with finer grains, more equiaxed grains and a more refi ned, solutioned structure. 69

PAGE 70

Water Cooling Only the low temperature ratchet condi tion was evaluated with water cooling because it was determined that it would not be feasible to use water cooling in industry. When samples that used the same ratcheti ng treatment were compared, the water cooled samples showed more refined structures than the air cooled. As seen in the air cooled samples, the 60 minute treatment, Figure 4-16 showed more equiaxed grains and more structural refinement than the shor ter treatments (Figure 4-17). However, the samples exhibited more columnar grains, ra ther than the desired equiaxed grains that would match the original airfoil. Summary Results from the 3rd matr ix followed the same trends observed in the previous matrices. Lower temperatures in the ratchet ing treatment lead to microstructures that were not fully solutioned, though faster cooli ng rates mollified this effect. The higher temperature ratchet proved to aid RXN mo re than the lower temperature ratchet, and also moving to a more moderate temperature in the ratcheting treatment (1204 C to 1177 C) and lowering the cooling rate elimin ated the cracking issues seen at the higher temperature. The promis ing results of the 1204 C/1hour solution heat treatment from Sample 6 were also verified in this matrix. LENS Heat Treatments After initial evaluation of the PPD pr ocess as a method of repair, the LENS process was also put forth as an alternat e deposition technique. There was less LENS material available, so samples were prepared in order to verify the trends seen in the PPD samples, without replicating the ent irety of the PPD sample matrices. 70

PAGE 71

The samples tested were a mix of ratchet and non-ratchet samples that replicated samples from earlier matrices al ong with new test tr eatments as seen in Table 4-6. Non-Ratchet Treatments The LENS samples that were process ed with standard heat treatments exhibited microstructures that appeared similar to th e microstructures in the PPD samples; however the LENS material showed somewhat finer, more homogenous grain structures overall. In addition, the LE NS material appeared to have a cleaner microstructure, owing to the method of deposition. The as-deposited LENS sample exhibit ed the typical dendritic microstructure observed in the PPD material; however the layer bands were closer together and the dendrites were aligned mostly in the depositi on direction in the LENS samples, whereas the PPD samples had shown significant re-m elting at the layer bands and the dendrites more randomly oriented (Figure 4-18). Sa mple O1, which was processed at 1093C exhibited a mostly dendritic structure; how ever with areas of large grains and some smaller grains. These grains are not really recrystallized grains but dendritic areas that have begun to reorganize. Sample O3 was very similar to O1 in that the structure was dendritic with scattered large grain areas. A dditionally the dendrites did not follow the deposition direction but some areas shifted or ientation. These sample micrographs can be seen in Figure 4-19. Samples O5 and O6 replicated the most successful samples from the original matrices, the highest temperatures tested. T he LENS samples again were similar to the PPD samples in that both showed a recrys tallized grain structure that also appeared to be refined (Figure 4-20). Similar to the previ ous results for the PPD samples, the higher 71

PAGE 72

temperature (O6) heat treatment resulted in a final grain size than the lower temperature heat treatment (O 5). These results supported t he previous findings that samples must be heat treated above their solvus temperature for any significant RXN. Ratcheted Treatments Ratcheting treatments were also evaluated ev en though previous results indicated that the ratcheting heat treatments were of limited succe ss. Different times, ratcheting temperatures and post treatment temperatures were used in the LENS samples. O1R/6-30 samples was given a lower temperat ure ratchet with water cooling combined with a short, high temperatur e post-treatment. This samp le showed large and medium sized grains with a dendritic character. The dendrites in this sample were not fully homogenized with the chemical segregation remaining and s howing up faintly in the micrograph. This indicated that the heat tr eatment was too short and did not allow for sufficient diffusion to o ccur. Samples O4R/5-30 and O4R/6 both used a higher temperature ratchet with varied cooling rates (jet cool and air cool respectively) and different lengths of the post treatment ( 30 minutes and 1 hour). These samples were similar in appearance with het erogeneous microstructures, a wide range of grain sizes and most importantly, faint layer bands. The micrographs of the ratcheted samples can be seen in Figure 4-21. The layer bands, wh ich were the result of the deposition process and the segregation from dendritic solidification, repres ented the different passes made during the deposition process. These layer bands indicate that through insufficient time, temperature or both that the sample is not fully homogenized. Summary Overall the LENS samples replicated the behavior of their PPD counterparts with samples heat treated at temper atures under the solvus, showing little or no RXN, while 72

PAGE 73

samples heat treated above the solvus showed signific ant RXN. The ratchet samples all exhibited some recrystallization; howev er none of these samples were fully homogenized. The LENS samples exhibited micr ostructures that were more refined and more homogenous than the PPD samples. Summary of Factors During this phase of the investigation, many factors were examined that contributed to the potential re crystallization of the test sa mples: time, temperature, ratcheting treatments and cooling rates. Each of these factors was determined to have different effects on the resulting microstructure. Time did not play an import ant part because it was held to a relatively small window of 30 min, 1 hr. and 2 hrs. U nder the constraints, the 1 hour treatment was shown to produce the optimal micros tructure when coupled with the correct temperature. Temperature proved to be the most important factor ; in addition the optimal recrystallization temperature also became relatively fixed and was tied to the gamma prime solvus. It was shown that below the so lvus temperature little to no refinement occurred, while there was significant refinement above this temperature. The importance of the ratcheting tem perature and the cooling rate was also shown. Using too low a temper ature regardless of the cooli ng rate was not successful in producing a recrystallized microstructure, while too high a temperature resulted in severe cracking in the structure. Moderat e temperatures were more effective but seemingly not sufficient to produ ce the desired microstructure. 73

PAGE 74

74 Re-evaluating Methodology At this point, in the project, It was determined that therma l processing alone was not sufficient to recrystallize the deposit to produce the desired microstructure. There was not enough residual stress in the material to leverage for recrystallization; so additional residual stress had to be added. A feasibility study was then started using a hardness tester to add micro deformation to a small area which was heat treated and examined. The feasibility study was expanded to a larger scale by using a rolling mill to add a uniform and controllable amount of deformation to the test coupons. Though rolling is not the method that would be used in industrial processing, the idea of the investigation was to determine if it was po ssible to reach the optimal microstructure. Additionally thermal pre-treatments were added to the test coupons to increase the effectiveness of the rolling proc ess. Both of these pre-defo rmation heat treatments were used to modify the gamma prime size to re duce the strength of the alloy and allow for increased deformation. The partial so lution process puts a portion of into solution, which increases the ductility and decreases the strength of the material, while the super overage treatment coarsens the precipitate to also reduce the strength of the sample.

PAGE 75

Table 4-1. IN-100 Heat Treatment window values Solvus Temp. (C) Incipient Melting Temp. (C) Window (C) Source IN-100 (I/M) 1235 1241 6 I N-100 (P/M) 1177 N/A N/A ** I N-100 (P/M) 1169-1171 N/A N/A *** Note: Reprinted with permission from D.L. Sponseller: Superalloys 2008 Conference Proceedings, 2008, p. 259. ** M.M Allen, R.E. Anders on and J.A. Miller: Process fo r fabricating integrally bladed bimetallic rotors, US Patent 4,479,293, 1984. *** Kai Song: "Grain growth phenom ena in P/M Ni-base superalloys", 2005 Table 4-2. Examples of standard nickel based superalloy solution heat treatments [Reprinted with permission from G.L. Erickson: A New, Third Generation, Single Crystal, Casting Superalloy. Journal of Mate rials, TMS, Warrendale, PA, 47, April 1995, p.36] 75

PAGE 76

Figure 4-1. Effect of composition on the -solvus temperatures and incipient melting temperatures for several well known superalloys [Reprinted with permission from D.L. S ponseller: Superallo ys 2008 Conference Proceedings, 2008, p. 259] 76

PAGE 77

Figure 4-2. Grain growth of IN-100 (P/M) as a function of annealing temperature for powders produced by different methods and cast and wrought process [Reprinted with permission from J.M. Larson: In Modern Developments in Powder Metallurgy, 1974, p. 537] Table 4-3. 1st Heat Treatment Matrix 77

PAGE 78

Table 4-4. 2nd HT Matrix Figure 4-3. As-Deposited PPD Sample at 50x (1st Matrix) 78

PAGE 79

A) B) Figure 4-4. Samples P1 and P1-R at 50x A) P1 and B) P1-R 79

PAGE 80

A) B) C) Figure 4-5. Samples P2-30, P2 and P2-120 at 50x A) P2-30, B) P2, C) and P2-120 80

PAGE 81

Figure 4-6. Sample P3 at 50x A) B) Figure 4-7. Sample P4 A) 50x and B) 100x 81

PAGE 82

Figure 4-8. Sample P5 at 100x Figure 4-9. Sample P6 at 250x 82

PAGE 83

A ) B) Figure 4-10. Sample P630 and P6-R both at 50x A) P6-30 and B) P6-R Table 4-5. 3rd Heat Treatment Matrix 83

PAGE 84

A ) B) Figure 4-11. Verification microgr aphs of Sample P6 from the 2nd HT Matrix 50x (A) and 100x (B) A) B) Figure 4-12. Micrographs of P1R/6-30 DAC 50x (A) and 100x (B) 84

PAGE 85

A ) B) Figure 4-13. Microgr aphs of P1R/6 DAC 50x (A) and 100x (B) A) B) Figure 4-14. Micrographs of P4R/6-30 DAC 50x (A) and 100x (B) 85

PAGE 86

A ) B) Figure 4-15. Microgr aphs of P4R/6 DAC 50x (A) and 100x (B) A) B) Figure 4-16. Micrographs of P1R/6-30 WC 50x (A) and 100x (B) 86

PAGE 87

A ) B) Figure 4-17. Microgr aphs of P1R/6 WC 50x (A) and 100x (B) Table 4-6. LENS verification sample matrix 87

PAGE 88

Figure 4-18. Dendrite reorient ation at layer bands in powder plasma deposit sample 88

PAGE 89

A) B) C) Figure 4-19. Micrographs of non-ratc het samples with no RXN at 50x A) As-deposited B) O1 C) O3 89

PAGE 90

A) B) Figure 4-20. Micrographs of non-ratc het samples that exhibited RXN A) O5 B) O6 90

PAGE 91

91 A) B) C) Figure 4-21. Micrographs of ratcheted samples A) O1R/6-30 B) O4R/5-30 C) O4R/6

PAGE 92

CHA PTER 5 RESULTS FEASIBILITY STUDY Adding Deformation Until this point the goal was to induce re crystallization in the simulated airfoils without adding mechanical deformation; however as seen through the results of the first phase of the study. Simple thermal processing, even with the added step of ratcheting treatments were not sufficient to refine the grain sizes of the deposit. Therefore mechanical deformation of the PPD samp les was added in the next phase. The PPD samples were used because of a dearth of LENS material at this point in the study. Initially, though, a proof of concept test wa s performed before completing a full test matrix. Proof of Concept Testing In order to test the addition of defo rmation on a small scale, a Rockwell hardness tester was used to make 3 indents with varyi ng loads, similar to the indents in Figure 51 [42]. This testing was performed on a LENS sa mple, as it was a small test piece that had already been ground flat. The procedure imparted a large amount of local deformation in several spots, then the samp le was heat treated at 1232C and the oxide removed by gentle polishing. The samples were examined by LOM and SEM and the results can be seen in Figure 5-2 and 5-3 respectively. The result of this testing shows that while the sample itself underwent RXN but the areas around the indents showed a much smaller grain structure than seen in the other areas (Figure 5-4). 92

PAGE 93

X-Ra y Diffracti on Verification Justification for Testing X-ray diffraction experiments were performe d to obtain a qualitative comparison of the amount of residual stress by examining the width of the peaks that appear in the graph [43]. For the characteristic peaks appear at around 50 for the (001) plane [44]. These experiments were intended to determine the texture that appears after a sample has undergone mechanical deformation and subsequent RXN. It has been shown in polycrystalline iron that samples where gr ains have the same orientation predeformation tend to form a [111] texture upon recrystallizati on despite being initially oriented in a different direction [45]. T herefore XRD can be used to determine if RXN has occurred. Samples in several conditions were exam ined to determine the effect of heat treatments on residual stress in the material and degree of recrystallization. These conditions include the untreated base metal, as-deposited PPD material, machined PPD material and annealed PPD material. These conditions were chosen because each represented different expected levels of residual stress and the results can be seen in Figure 5-5. X-ray Diffraction Results PW 1074 base metal under powder diffraction, showed strong, narrow peaks at (111), (200) and (220). This is common fo r materials with small equiaxed grains. Whereas the as-deposited PPD material disp layed strong peaks at (200) and (220) but not (111). This result was expected because of (100) growth direction common for nickel-based superalloy materials. These resu lts indicated a lack of RXN, which would 93

PAGE 94

be expected in the as-deposited c ondition. The peak width of this sample was consistent with the other specimens with little to no residual stress. The machined PPD material was machined in order produce flat surface for rolling. Machining produced surface deformation, whic h added stress to the material, as shown in the broadness of all the peaks. The peaks fo r this sample are all lower intensity and broader than both the as-depos ited and annealed materials. This sample showed a weak (111) peak, which signified a sm all amount of recrystallization. The unstressed sample was subjected to a stress-relief anneal. When tested in the under x-ray diffraction, it showed a strong (111) peak and a weaker (200) peak, which showed that during the anneal some RXN occurr ed to lead to mostly [111] texture with little of the initial [001] texture being left. Th is (200) peak is most narrow of the tested PPD materials, confirming the decrease in re sidual stress when compared to the rest of the samples tested (Figure 5-6). Overall the x-ray diffraction experiment seemed to confirm the result seen during the proof of concept testing, that even a small amount of deformation caused RXN in the deposited material after the heat treatment. Use of Rolling as Deformation Method Based on the above results, additional test ing was performed to evaluate thermomechanical testing. The method of deforma tion used was rolling due to the ease of adding uniform and controllable deformation and the availability of the facilities. Also added to the testing matrix were sample pr e-treatments, which we re designed to allow the sample to be more easily deformed. Two pre-deformation heat treatments were employed: partial solution (PS) and super ov erage (SO). The partial solution treatment was performed for 1 hour at 1121C with air cooling in order to partially solution the to 94

PAGE 95

decrease the strength of the alloy [46]. The super ov erage pre-treatment was designed to increase the size of the precipitates to increase the mismatch between and and lose the strength that derives from coherent precipitates as seen in Figure 5-7 [47, 48, 49]. These samples from these condi tions along with the as-deposited condition were rolled according to the procedure laid out in the Experimental procedures chapter and then analyzed using metallography and other characterization techniques. Rolling Feasibility Study High Deformation The investigation of the effect of in creased deformation on the recrystallization of the samples began with testing of samples wit h a greater amount of deformation. Using a rolling mill, 7.5-8.5% deformation (Table 52) was imparted to the PPD samples in pre-treated and as-deposited conditions. Th ese samples then underwent recrystallizing heat treatments at the same te mperatures as the previous matrices, except the highest temperature (1232 C) due to the industrial concerns mentioned before. The test matrix can be seen in Table 5-3. The as-rolled condit ion indicates that no heat treatment was given to these samples after rolling. This condition was added to look for dynamic RXN, and then determine the extent of RXN durin g subsequent heat treatments. The asdeposited condition was rolled then given the same heat treatments as the pre-treated samples to examine the effect of the pretreatments on rolling and recrystallization heat treatments. As-Rolled Samples The as-rolled samples had no post treatment and functioned as a control to verify the effect of post deformation heat treatments. In addition, samples of the fine-grained IN-100 that represented the original airfoil or the substrat e for the deposition process were included. These samples were intended to show the effect of the heat treatments 95

PAGE 96

on the microstructure of the airfoi l adjacent to the deposit after similar thermal cycles during the post deposition heat treatment. T he base metal for the as-rolled sample showed very fine equiaxed grains as woul d be expected from Gatorized powder metallurgy stock. The rest of the samples tested (as-deposited, partial solution and super overage) all showed similar microstructures, coarse dendritic structures as seen in the other deposited unrecrystallized samples in this study (Figure 5-8). As expected, the samples that were pre-treated showed a few large grains, but not the desired recrystallized structure. The heat treatments are divided into categories of sub-solvus and super-solvus below for clarity. Sub-Solvus Heat Treatments The samples heat treated at 1121 C showed little difference over the as-rolled condition, that is no large scale RXN, but generally more refinement and some large grains. The as-rolled base metal in sample PM-1-2 showed enlarged but still equiaxed grains. This seems to indicate no RXN had occurred but the sample exhibited instead grain growth. Compared to the as-rolled condi tion, sample PM-2-2 or the as-deposited condition exhibited a more re fined dendritic structure with some moderate size grains. The partially solutioned sample showed a finer dendritic struct ure with apparent lines that formed between weld layers. Also eviden t were some large grains, but again not a significant amount of RXN. T he heat treatment on the SO samp le left this structure with retained dendrites and many large grains. T he microstructures can be seen in Figure 59. The samples that were underwe nt heat treatments at 1149 C, exhibited similar but more refined structures than the 1121 C samples; however only a limited amount of 96

PAGE 97

RXN was observed. The microstructures can be seen in Figure 5-10. The as-rolled sample, PM-1-3, maintained its equiaxed st ructure but showed larger grains than sample PM-1-2. The as-deposited sample retained dendrites and exhibited s ome columnar grains. The PS sample showed more refined dendritic structure than after the 1121 C treatment, while for the SO, the post-treat ment left this sample with most large grains with a few areas of reta ined dendrites similar to PM-4-2. Overall these samples, like those in previous test matrices without added deformation, did not show any RXN and ve ry little refinement when heat treated at samples below the solvus. Super Solvus Heat Treatments As with previous test matrices, the sa mples heat treated at around and above the solvus temperature, began to exhibit RXN, indicating that the recrystallization temperature was reached. The amount of RXN typically increased with increasing temperature. The samples heat treated at 1177 C showed significant re finement over the subsolvus treatments as seen in Figure 5-11. The base metal sample showed increasingly large grains compared to previous samples PM-1-2 and PM-1-3, further indicating that there is no RXN of the already fine grains but instead grain growth. The as-deposited sample exhibited a refined grain structur e with elongated grains, whereas the PS and SO samples showed refined grain structures with a mixed character of large and small grains and a mostly equiaxed structure. Continuing the trend seen thus far, the samples heat treated at 1204C exhibited more refinement than the lower temperatur es. It should be noted that the base metal 97

PAGE 98

sample giv en the 1204 C heat treatment had a much larger grain size than the same samples treated at the lower temperature. Significant grain growth had taken place, which may limit the use of these high heat treatment tem peratures in practice. The sample with no pre-treatment showed a ho mogenized microstructure however the grains still had an elongated nat ure. The samples that we re given pre-treatment exhibited a more refined mi crostructure. The post-treatment was very effective at refining the structure and producing an equiaxed grain structure in t he partial solutioned sample, while the super overage seemed to show the most promising results with a homogenized structure and finer equiaxed grains than seen in previous samples. The microstructures can be seen in Figure 5-12. Summary The most important result from this testing matrix is that the added deformation did increase the amount of re crystallization in the sample. As seen from previous results, the lower temperatur e post-treatments did little to promote RXN, only when near the solvus temperature (>1177 C) did significant refinement and RXN occur. Also the super overage pre-treatment se emed more effective overall than the partial solution, yielding more refined, homogeneous microstr uctures; however both pre-treatments proved effective. Rolling Feasibility Study Low Deformation After establishing with the high deforma tion testing that the material can be successfully recrystallized to near optimal grai n size, it became necessary to try a lower deformation condition to determine the thres hold for the desired microstructure (Table 5-4). Samples with the same pre-treatments were rolled to lower deformation rates of 98

PAGE 99

about 4.5 -7% (Table 5-5), then subjected to the same pos t-rolling heat treatments and analyzed. With the high deformation condition confirming that sub-solvus temperatures were not sufficient to induce RXN, these te mperatures were dropp ed from the testing; however the ratcheting treatments were included to gauge their effectiveness at augmenting the deformation added mechanically. As-Rolled Samples The as-rolled samples for the low deformati on rolling trial appeared very similar in structure to the high deformati on condition as seen in Figur e 5-13. The as-rolled sample displayed small equiaxed grains expected of powder metallurgy product, while the asdeposited condition featured a dendritic stru cture. The partial solution and super overage pre-treatment samples exhibited a dendritic structure with some large grains evident, similar to the previous matrix. Super Solvus Heat Treatments The samples heat treated at 1177C (Figure 5-14), appeared similar to the high deformation matrix with one major difference, the character of the grains. Where the high deformation condition exhibited most ly equiaxed grain, the low deformation condition features mostly elongated grains. Th is is true of all pre-treatment conditions. The super overaged sample showed the most equiaxed character while the asdeposited showed the least. The as-rolled sample continued the trend of displaying that exposure to the post deposit he at treatments would cause grain growth in the original airfoil segments. Similar to previous test matrices, the high temperatur e heat treatments at 1204 C exhibited better microstruc tures than the lower temperature heat treatment. This 99

PAGE 100

temperature brought a more desirable ratio of equiaxed to elongated grains in the deposited s amples with and witho ut pre-treatments. As seen before the as-rolled samples showed more grain growth when co mpared with the lower temperature heat treatment. The microstructure s can be seen in Figure 5-15. Super Solvus and Ratcheting Treatments The samples heat treated at 1177 C with a ratchet at 1177 C showed results similar to the non-ratchet treatment (Fi gure 5-16). The as-rolled samples exhibited larger grains than the non-ratchet sample along with a preponderance of twinning. The as-deposited and partial solution samples displayed a mixed equiaxed and elongated grain character. Unlike previous sample gr oups, the super overage samples seemed to show a more elongated grains structure t han the other samples, but still retained a mixed grain character. The ratcheted samples, which were heat treated at the higher temperature, 1204 C, once again displayed the best microstr uctures as seen in Figure 5-17. The asrolled test sample appeared similar to the non-ratcheted sample but the ratchet treated sample had larger grains and more twinni ng. This comparison held true for both the high and low temperature conditions. The effect of the higher temperature on the as-deposited and partial solution was the same as seen in lower temperature exce pt the ratchet samples appeared to have a larger percentage of equiaxed to elongated grains; however the super overage sample seemed to have smaller grains than the s uper overaged samples that did not have the ratcheting treatment. 100

PAGE 101

Summary As with previous heat treatm ents at the higher temperatur es yielded better results than lower temperatures and with all the deposited samples, the grain size and proportion of equiaxed grain to elongated gr ains was increased with the ratcheting treatments. In this matrix the super overage pre-treatm ent also yielded the best microstructure with greater amounts of refinement and homogen ization. Also important is the effect seen in the base metal samp les, no recrystallization occurred, even with ratcheting, only grain growth was observed. Materials Change and Optimization After the feasibility study using PPD mate rial was performed the primary method of deposition changed from PPD to LENS. The LENS material from initial testing seemed superior to the PPD deposits because there were smaller layers and less remelting between layers. The LENS material al so seemed cleaner with fewer inclusions in the microstructure and better deposited with seemingly less porosity. Though this change did not seriously affect the study it di d require some verification that the results and trends from the addition of mechanical deformation would be the same for the LENS material. The result of the feasibilit y study was that it was show n that it was possible to recrystallize and refine the microstructure of the PPD deposited samples. This essentially answered the first question, could this be done. With that out of the way it, the essence of the study became the optimization of the pr ocess. The questions that needed answering were 101

PAGE 102

102 Does the LENS material behave the same as the PPD material? What method will allow for in-situ deformation? How much deformation can be added? Are the pre-treatments necessary? With the increase in deformati on will shorter tr eatments work? The experiments in the last phase of t he study were designed to answer these questions; however there was no doubt that thermo-mechanical processing was the method required to achieve the objectives of the project.

PAGE 103

Figure 5-1. Example of Vickers hardness indents [Reprinted with permission from Vickers Hardness Testing, ASM Handbooks, 2003, Figure 21] A) B) Figure 5-2. LOM Micrographs of RXN hardness indents in LENS material at A) 100x and B) 200x 103

PAGE 104

A ) B) Figure 5-3. SEM Micrographs of RXN hardness indents in LENS material at 650x Figure 5-4. Micrograph showing the bulk of the LENS hardness samples at 50x 104

PAGE 105

Figure 5-5. Full x-ray diffraction pattern Figure 5-6. Close-up of XRD Pattern 105

PAGE 106

Table 5-1. Pre-treatments for rolling trials Temperature (C) Time (Hours) Cooling Method Secondary Cooling Partial Solution 1121 1 Air Cool N/A Super Overage 1038 1 Furnace Cool to 538C Air cool to Room Temperature Figure 5-7. Decrease in strength due to overaging [Reprinted with permission from Robert E. Reed-Hill and Reza Ab baschian, Physical Metallurgy Principles, 1994, p.524, Figure 16-10] Table 5-2. Actual deformation level for 8% RIA rolling study Sample ID RIA (%) As-Rolled 8.5 As-Deposited 8.1 Partial Solution 7.8 Super Overage 7.7 106

PAGE 107

Table 5-3. Test matrix for 8% RIA rolling study A) B) C) D) Figure 5-8. As-rolled microstruc tures with 8% RIA at 100x A) PM-1-1 B) PM-2-1 C) PM-3-1 D) PM-4-1 107

PAGE 108

A ) B) C) D) Figure 5-9. Microstructures heat treated at 1121 C with 8% RI A at 100x A) PM-1-2 B) PM-2-2 C) PM-3-2 D) PM-4-2 108

PAGE 109

A ) B) C) D) Figure 5-10. Microstructures heat treated at 1149 C with 8% RI A at 100x A) PM-1-3 B) PM-2-3 C) PM-3-3 D) PM-4-3 109

PAGE 110

A ) B) C) D) Figure 5-11. Microstruc tures heat treated at 1177 C with 8% RI A at 100x A) PM-1-4 B) PM-2-4 C) PM-3-4 D) PM-4-4 110

PAGE 111

A ) B) C) D) Figure 5-12. Microstruc tures heat treated at 1204 C with 8% RI A at 100x A) PM-1-5 B) PM-2-5 C) PM-3-5 D) PM-4-5 111

PAGE 112

Table 5-4. Test matrix for low deformation level rolling study Table 5-5. Actual deformation level for 4-6% RIA rolling study Pre-Treatment RIA (%) As-Rolled 5.0 As-Deposited 6.9 Partial Solution 5.6 Super Overage 4.8 112

PAGE 113

A ) B) C) D) Figure 5-13. Microstructures of As-Rolled samples with 5% RIA A) PL-1-1 at 1000x and B) PL-2-1 C) PL-3-1 D) PL-4-1 AT 100x 113

PAGE 114

A ) B) C) D) Figure 5-14. Microstructures heat treated at 1177 C with 5% RIA A) PL-1-3 at 200x B) PL-2-3 C) PL-3-3 D) PL-4-3 at 100x 114

PAGE 115

A ) B) C) D) Figure 5-15. Microstructures heat treated at 1204 C with 5% RIA A) PL-1-4 at 200x B) PL-2-4 C) PL-3-4 D) PL-4-4 at 100x 115

PAGE 116

A ) B) C) D) Figure 5-16. Microstructures heat treated at 1177 C with 5% RIA and Ratcheting treatments at 100x A) PL-1-3/4R B) PL-2-3/4R (C) PL-3-3/4R (D) PL-4-3/4R 116

PAGE 117

117 A) B) C) D) Figure 5-17. Microstructures heat treated at 1204 C with 5% RIA and Ratcheting treatments at 100x A) PL-1-4/4R B) PL-2-4/4R (C) 4PL-3-4/4R (D) PL-1-4/4R

PAGE 118

CHA PTER 6 RESULTS THERMO-MECHANICAL PROCESSING With the paradigm shift in the investigat ion from strictly thermal processing to thermo-mechanical processing (TMP), it bec ame apparent that significant amounts of RXN could be induced in the workpieces for the purpose of repair. Thermo-mechanical processing combines heat treatment and deformation/cold working to refine the microstructure [50]. The goal of this combi nation approach is to accomplish a specific microstructural goal, by using a variety of mechanical deformation techniques and subsequent heat treatments [10]. The TM P route depends greatly on the final microstructure, which for this study is a fine grained homogenous microstructure. In the case of this study, cold work was added to the material using an array of methods then the samples subsequently heat treated above the recrystallization temperature to cause RXN of the microstructure. Rolling Verification With the results from t he thermal treatments verifi ed, the next step was to add deformation to the LENS samples and evaluate the potential of using thermomechanical processing. Like the PPD samples, the LENS samples were cold rolled after being pre-treated to allow for easier deforma tion. Additionally, several deformation levels were used 4-6% (Low), 8-10% (Medium) and 10-15% (High), the actual deformation levels can be seen in Table 6-1. These levels helped to establish the deformation window, or the minimum necessa ry deformation and the effects of higher levels of deformation, beyond the previous maximum level of 8%. The new matrices tested the same times and temperatures that were effective before; however, temperatures far below the solvus were dr opped because of their ineffectiveness. The 118

PAGE 119

highest temperature (1232C) was dropped since it was not practi cal for industrial scaleup. The temperatures, used in this portion of the work were just above and just below the solvus temperature. Additionally the as -rolled samples examined in the feasibility studies were not evaluated; the as-deposited samples shown in this chapter were heat treated the same as the pre-treated samples. Low Deformation Level The low deformation level (Table 6-2) was used to determine the minimum amount of the deformation need to cause RXN of the samples. As seen with the thermal treatments, the LENS samples behaved similarl y to the TMP route. The samples, which were in the as-rolled condition (Figure 6-1) showed dendritic micr ostructures with some areas that might indicate dynamic RXN, but these areas were not widespread. The pretreated samples exhibited a greater degree of RXN than t he as-deposited sample with the super overage treatment re sulting in the finest, most homogenous microstructure remaining the most effective. These samples as expected, displayed the characteristics of a segregated microstructure namely t he presence of the layer bands from the deposition process. These features implied remnant s egregation even after heat treatment. The samples heat treated at 1177C, were similar to the past samples heat treat at this temperature with a generally mixed grai n characteristic of large and small grains as seen in Figure 6-2. The as-deposited and heat treated sample showed the appearance of retained dendrites and extensiv e twinning. The pre-treated samples displayed more elongated grain character with grains scattered around the sample, not a continuous structure. Unlike the prev ious results, the super overaged sample exhibited the largest grains. 119

PAGE 120

Similar to previous results, an increas ing heat treatment temperature (1204C) exhibited smaller grains than the lower heat treatment temperature. These samples also featured a mixed grain charac ter with both large and small grains in the as-deposited samples and medium and small grains in the partial solution. The super overage sample again had large elongated grains with some smaller grains, too (Figure 6-3). Overall the low deformation level did not achieve the desired results but did correlate well with the PPD samples under the same conditions. It would seem that this level of deformation is too low for recr ystallization to the optimal grain size. Medium Deformation Level The medium deformation level tested wa s the same as maximum level used in the feasibility study on the PPD material. The sample matrix can be seen in Table 6-3. The as-rolled condition showed similar re sults as seen before in Figure 6-4, with a mostly dendritic structure with some RX N having occurred. Though there seemed to be some grains evident, the grains were elongated in the rolling direction and were dendritic in nature. The dendrit ic nature of the grains c ould be seen by the wavy edges instead of straight lines in recrystallized grains [47]. These samples also showed the layer bands evident in structur es that are not homogenous. Samples heat treated below the solv us at 1149C, once again confirmed the trend of showing little RXN, more than the as-rolled samples but less than those processed at higher temperatures. The sample s, as seen in Figure 6-5, featured areas of RXN with elongated grains but still a mo stly dendritic charac ter. This dendritic character along with the faint but still pres ent layer bands reinforced the need to heat treat samples above the solvus. 120

PAGE 121

Unlike previous results, the sample s processed at 1177C, which was above the solvus exhibited marked recrystallization. The pre-treated samples showed mixed grains characters with compar atively large and small grai ns, whereas the as-deposited samples showed only large grains. The pre-tr eated samples also featured twins in the microstructure. All of the specimens at this temperature were homogenous as can be seen in Figure 6-6. The highest temperature tested, 1204C, again exhibited microstructures with the greatest degree of RXN (Figure 6-7). The as -deposited samples showed some grains; however the pre-treated samples displayed an equiaxed grain structure with smaller grains than the as-deposited samples. The super overaged samples presented what could be a duplex structure [26]. The duplex gr ain structure is a mixture of fine and coarse grains that could result from parti al coarsening of fine gr ain structures. This could be evidence of grain growth possible at higher temperatures. High Deformation Level For the highest deformation level tested, the specimens were given between 10 and 14% reduction in area to establish the high end of the deformation window. The samples were heat treated at only the highes t temperature because it had consistently been the most effective at RXN. The as-deposited samples showed very li ttle RXN and a mostly dendritic structure, as well as layer bands. The pre-treated samples both exhibited a homogenized structure with a mixture of grain sizes. The grains were mostly equiaxed in character with some elongated also as can be seen in Figure 6-8. 121

PAGE 122

Summary The recrystallization behavior observed in the LENS samples was similar to the PPD samples tested under similar conditions. Once again it was necessary to process the samples above the solvus temperature for any significant amount of RXN to occur and the pre-rolling heat treatments seemed to result in finer, more uniform final microstructures. Also higher deformation le vels proved more effective than the low deformation level since lower deformation levels resulted in microstructures with elongated and larger grains. Alternate Methods of Deformation With the LENS processed material re sponding better to the thermo-mechanical processing than the PPD material, the study was continued with em phasis on the LENS material. The first step in this evaluation of potentially viable met hods to implement the process in industry was to develop a method of deformation, which could be used directly on the IBR without destructive disa ssembly. The following surface improvement methods were evaluated to impart deformation in the IBR blade. Shot peening Shot peening is a surface improvement treatment that involves the impacting a surface using shot of various materials like glass or metal and is designed to put compressive stress in the surface of the ma terial [51, 52]. This process deforms the surface of the materials plastically without damaging the workpiece. This process was applied to the LENS samples to evaluate it s use as the mechanism of deformation. The as-deposited sample was evaluated wi thout heat treatment to ensure that there was no difference when compared to the other deformed samples. As seen in Figure 6-9, the sample evidenced a dendritic microstructure with layer bands similar to 122

PAGE 123

the previous as-deposited samples. The shot peened samples whic h were heat treated at 1204C, both displayed few large grains with layer bands evident (Figure 6-10). The microstructures indicated that very littl e RXN had occurred reflecting that the shot peening had not imparted the minimum level of deformation necessary to recrystallize the microstructure. Laser Shock Peening Laser shock peening or LSP is another surf ace improvement treatment that has results similar to shot peening, namely, plastic deformation of the surface. LSP however uses lasers to heat an ablative material applie d to surface of the workpiece. The laser heats the ablative and then the re sulting shockwave is channeled toward the surface by water surrounding the part [53, 54]. In this manner, deforma tion is applied in small sections all over the surfac e of the part. Proce ss parameters for LSP include spot size or the area where each small section is deformed, the ener gy imparted with each shock which is a function of laser power and number of passes. Each of these factors comes together to determine the am ount of deformation imparted. Following previous procedures, the samples were heat treated at 1204C for either 15 minutes or an hour after the deformation process, to begin to re-investigate the shorter heat treatments with higher deformation levels. There were no pre-treatments applied to these samples. The sample matrix can be seen in Table 6-4. Once again as-deformed samples were examined to establis h baseline behavior for the LSP samples (Figure 6-11). Similar to previous results, the LEN LSP samples exhibited fully dendritic structures; however an extensive amount cracking was observed. Similar to the cracking observed in the ratcheted samples (Figure 4-10B), the LSP samples exhibited cracking along the d endrites. This cracking appeared to be due 123

PAGE 124

to excessiv e deformation or heating was appli ed in a very short time period, causing them to crack along interdendritic region. This area is ty pically brittle, due to the segregation, coarse structure, and presence of carbides [40]. The samples heat treated for 15 minutes exhibited retained dendritic characters (Figure 6-12). There was also RXN, which re sulted in a mix of fine and large grains oriented along the prior dendrite areas. Samp les 1-15 and 3-15 showed some cracking at the edges of the samples but not to the scale in the as -deformed samples. It seemed for this deformation level that 15 minutes or the amount of deformation was not sufficient to establish large scale recrystallization. The samples which were heat treated fo r 1 hour showed homogenized structures free of any remnant dendritic character along with a duplex recrystallized grain size. However the most evident feature in these samples was the return of the widespread cracking seen in Figure 6-13. The cracking was similar in appearance to the asdeformed samples but the cracking occurr ed mostly along the grain boundaries (Figure 6-14). As before, the intergranular cracking indicated that the deformation applied by LSP was probably due to too much energy bei ng imparted to the sample. The samples, which had lower energy and bigger spot size seemed to exhibit a decreased amount of cracking. The higher energy concentrated on a smaller spot had the effect of increasing the cracking seen in samples. Additionally the samples, which underwent 4 passes instead of 2, cracked more frequently than the other samples. The amount of cracking in the material as a result of this deformation process combined with the relatively little RXN that occurred indicated that this process was unsuitable for use in the repair of IBR components. 124

PAGE 125

Low Plasticity Burnishing Low plasticity burnishing or LPB is anot her surface treatment where a free rolling ball supported by a fluid bearing is moved across a material with just enough force to plastically deform the surfac e of the sample [55]. Samples were deformed using the LPB process to a low and high deformation level and then heat treated at 1177C and 1204C as seen in Table 6-5. These samples were characterized by metallography to examine the microstructure and by hardness testing to determine the penetration of the deformation. The samples that were heat treated after receiving the lower amount of deformation (Figure 6-15) exhibited a homogenized structure with very few large grains. In the samples receiving higher amounts of deformation, the heat treatments produced microstructures with a mix of grains sma ller than the low deformation condition and some large elongated grains. There was al so some twinning evident in 13.80% reduction in area (RIA) sample. The samples processed at the hi gher temperature, showed a homogenous structure after both the low and high deformation levels. These microstructures were composed of a mixed grain character of rela tively large and small grains. However the high deformation sample showed a more equi axed grain morphology than the other condition, as seen in Figure 6-16. In addition to metallographic examinations, hardness traces on the high deformation sample was performed to determine the depth of the cold work imparted to the sample [56]. To examine the depth of penetration hardness traces were taken both across the width and along the interior and exte rior lengths of the samples. The average hardness data by location on the sample is presented in Figure 6-17. As this figure 125

PAGE 126

shows, the resulted in a significant har dness increase ov er the front and back areas, which did not see any deformation, HRC 52 co mpared to 44 for the front and 42 for the back. Additionally the top and bottom areas of the samples corresponding to the surfaces, which underwent the LPB process, were harder than the center section where the deformation did not penetrate. The lack of penetration can especially be seen in the width hardness traces that have been graphed against the dist ance from the edge of the samples in Figure 6-18. The depth profiles dem onstrated that the deformation and thus the hardness dropped precipitously (over 10 HRC) even 0.25 mm from the edge of the sample. The hardness values for all the trac es done on the LPB samples can be seen in Appendix B. The lack of even penetration of cold wo rk combined with the insignificant RXN seen in the microstructures of the heat treated samples conf irmed the unsuitability of using LPB as the method of cold working the workpieces. Summary The examination three of the common surface improvement techniques in nickel based superalloys did not identify any pot ential methods of inducing RXN when combined with the heat treatm ents that with deformation hav e proved effective in the rolling trials. As seen in Figure 6-19, L PB was the only process where the highest residual stress levels reach even 0.5 millimeter s into the samples. The other methods, after an initial superficial spike in resi dual stress, exhibited residual stress that decreased rapidly with distance. For depth of penetration, LSP seemed to be the best technique as even 1.4 millimeter s into the sampled there was some residual stress [57]. With sample widths between 1.0 and 1.5 mm, these deformation techniques would not 126

PAGE 127

provide the consistent level of deforma tion throughout the samples that has proven effective in the rolling trials. Empirically, each technique was unsuitabl e for a different r eason. Shot peening did not provide enough energy to produce RX N, while LSP provided too much energy and caused severe cracking of the s pecimens before any heat treatment was performed. Finally LPB proved inadequate due to the lack of penetration of the deformation energy. In-situ Deformation Process In the process of the examining possible in-situ deformation methods, the industrial sponsor who provi ded the LENS material develope d a proprietary method for evaluation. The method involved in-situ cold/ hot working of the samples, which imparted an unknown amount of deformati on during deposition. These samples were then heat treated at 1204C, the most effective and industrially feasible temperature from the investigations. Heat treatments of 15, 30 and 60 minutes were used based on the desire to decrease the exposure of the unrepaired part of the IBR to high temperatures to reduce the risk of grain growth (Table 6-6). The samples heat treated for an hour performed as expected with homogenized and relatively fine grained microstructures as seen in Figure 6-20. The 30 minute treatment (Figure 6-21) had so mewhat larger grains than the hour heat treatment along with faint layer bands indicating that t he homogenization of the sample was not complete. The 15 minute heat treatment exhibi ted smaller grains than the other samples which had longer heat treatm ents but like the 30 minute tr eatment showed the layer bands that indicated that the sample was not fully homogenized (Figure 6-22). 127

PAGE 128

128 Hardness traces were performed on these samples to gauge the difference in the heat treatments in recrystallizing the microstr ucture, the results are presented in Figure 6-23 with the spot locations reflecting the di stance from the substrate. The graph shows that as the length of t he heat treatment increases the hardness of the sample decreases. Clearly (Table 6-7) recrystallizat ion had occurred with the transformation of the microstructure at the expense of the dislocation density in the deformed samples. Overall this in-situ process proved to be effective in conjunction with heat treatments developed in this study. Confirmation Through the testing of varying defo rmation methods along with the time and temperature of the post-def ormation heat treatment, thermo -mechanical processing has proven to be effective at inducing RXN in the deposited samples. The different processing parameters proved to be important in achieving the optimal microstructure and through much iteration, have been tested.

PAGE 129

Table 6-1. Deformation data for LENS rolling study Table 6-2. LENS 4% rolling matrix 129

PAGE 130

A) B) C) Figure 6-1. Micrographs of low deformation le vel samples in the as-rolled condition A) OL-1-A B) OL-2-A C) OL-3-A all at 50x 130

PAGE 131

A) B) C) Figure 6-2. Micrographs of low deformati on level samples heat treated at 1177C A) OL-1-3 B) OL-2-3 C) OL-3-3 all at 100x 131

PAGE 132

A) B) C) Figure 6-3. Micrographs of low deformati on level samples heat treated at 1204C A) OL-1-4 B) OL-2-4 C) OL-3-4 all at 100x 132

PAGE 133

Table 6-3. LENS 8% Rolling Matrix 133

PAGE 134

A) B) C) Figure 6-4. Micrographs of medium deformation level samples as-rolled A) OM-1-A B) OM-2-A C) OM-3-A all at 50x 134

PAGE 135

A) B) C) Figure 6-5. Micrographs of medium deforma tion level samples heat treated at 1149C A) OM-1-2 B) OM-2-2 C) OM-3-2 all at 50x 135

PAGE 136

A) B) C) Figure 6-6. Micrographs of medium deforma tion level samples heat treated at 1177C A) OM-1-3 B) OM-2-3 C) OM-3-3 all at 100x 136

PAGE 137

A) B) C) Figure 6-7. Micrographs of medium deforma tion level samples heat treated at 1204C A) OM-1-4 B) OM-2-4 C) OM-3-4 all at 100x 137

PAGE 138

A) B) C) Figure 6-8. Micrographs of high deformati on level samples heat treated at 1204C A) AS B) PS C) SO at 50x 138

PAGE 139

Figure 6-9. As-deposited sample with no treatment after shot peening at 50x 139

PAGE 140

A) B) Figure 6-10. Micrographs of shot peened and heat treated samples A) PS B) SO both at 50x 140

PAGE 141

Table 6-4. Sample matrix for LSP samples A) B) C) D) Figure 6-11. Micrographs of asdeformed LSP samples at 50x A) 1-AS B) 2-AS C) 3-AS D) 4-AS 141

PAGE 142

A ) B) C) D) Figure 6-12. Micrographs of LSP samples heat treated for 15 minutes at 50x A) 1-15 B) 2-15 C) 3-15 D) 4-15 142

PAGE 143

A ) B) C) D) Figure 6-13. Micrographs of LSP samp les heat treated fo r 1 hour at 50x A) 1-60 B) 2-60 C) 3-60 D) 4-60 Figure 6-14. Higher magnification image of intergranular cracking in LSP sample (100x) 143

PAGE 144

Table 6-5. Sample matrix for LPB samples A) B) Figure 6-15. Micrographs of L PB samples heat treated at 1177C A) OB-L-3 B) OB-H-3 both at 50x 144

PAGE 145

A) B) Figure 6-16. Micrographs of L PB samples heat treated at 1204C A) OB-L-4 B) OB-H-4 both at 50x 145

PAGE 146

Figure 6-17. Average hardness values by location of high deformation LPB sample Figure 6-18. Width hardness traces for high deformation level LSP sample 146

PAGE 147

Figure 6-19. Residual stress comparison for LPB, LSP and Shot peened samples [Reprinted with permission fr om William Beres: FOD /HCF Resistant Surface Treatments, p.57, Figure 8] Table 6-6. In-situ deformation sample matrix 147

PAGE 148

Figure 6-20. In-situ deformed sample heat treated for 1 hour OIS-60 at 50x Figure 6-21. In-situ deformed sa mple heat treated for 30 minutes OIS-30 at 50x 148

PAGE 149

149 Figure 6-22. In-situ deformed sa mple heat treated for 15 minutes OIS-15 at 50x Figure 6-23. In-situ deformati on hardness after heat treatment Table 6-7. Average hardness for di fferent in-situ heat treatments

PAGE 150

CHA PTER 7 DISCUSSION In this chapter the result s of the various phases of th is study: thermal treatments, feasibility study and thermo-mechanical processing will be compar ed. The evaluation will be based on the previously presented micr ostructures and grain size as well as ancillary techniques such as XRD and hard ness testing. The results will be examined and compared to what is understood about RXN and microstructures from the literature. Microstructural Evaluation Throughout the study, the microstructure of the samples have been one of the primary methods of determining the effi cacy of the heat treat ments and later the addition of deformation, mostly by ex amining the apparent grain structure and microstructure. There are a num ber of microstructural featur es that are important to understand what occurs during TMP and the vi ability of the simulated repaired part. Dendritic Structure The as-deposited structur e seen throughout this study regardless of deposition process was a coarse dendritic structure. During deposition the dendrites grow along the [001] fast growth direction, with each su ccessive layer, re-melting the ends of the dendrites and continuing the gr owth process. The resulting deposition is a coarse grained, fully dendritic struct ure that is unsuitable fo r repair purposes. During heat treatment this structures begins to c hange, and the segregation decreases as the dendrites become columnar grains as shown by Figure 7-1, which depicts solidification near mold walls during casting [47]. These columnar grains retain their dendritic character until the recrystallization temper ature is reached, this sequence can be seen in Figure 7-2, where PPD samples heat tr eated for 1 hour at 1093C, 1121C, 1149C 150

PAGE 151

and 1177 C are shown. This sequence showed that until the RXN temperature of 1177C was reached the struct ure remained dendritic. Grain Structure Above the solvus, which was empirically shown to be the recrystallization temperature, the dendritic as -deposited structure evolves in to a microstructure with varying grain sizes depending on the processing route. The constant throughout the study was the character of grains. Below the solvus temperature there were few recrystallized grains; areas of reorganized dendrites, with features that appear to be grain boundaries but are actually ce ll boundaries and the precursors to the recrystallized grain boundaries. When recrystallization occurred over t he entire samples the grain structure reflected this, showing the decreased size that is emblematic of recrystallized grains and showing a microstructure that resembled the blade before repair. Twinning Beginning in the thermo-mechanical def ormation phase of the study, twins began to appear in the microstructure. Twinning is a mechanism of deformation where slip is not involved, but where atoms move by shear so that the structure is identical to that across the twin boundary as in Figure 7-3 [58]. Mechanical or deformation twinning is the second most important deformation mechani sm behind slip and they are competing processes. Twinning was more likely to occu r at high strain rates and low temperatures, which explained twinning was more likely to form during rolling and in-situ deformation than by other mechanisms [50]. These me chanical twins can be seen in the asdeformed condition prior to heat treatment as narrow bands, which exist solely within grains [25]. 151

PAGE 152

In samples that have been subjected to the recrystallization heat treatment, annealing twins are prevalent. These features are common in FCC metals like nickel superalloys. Annealing twins are broad bands, which are in contrast to the narrow bands of mechanical twins, which occur afte r annealing heat treatment s as a result of stacking faults in crystals [50, 59]. It wa s also thought that annealing twins originated from the mechanical twins that occur in deformed materials that had not been heat treated, where the twins grew rather than being destroyed during annealing heat treatments [25]. The twinning confirmed the effect of deformation on the microstructure creating twins and the effect of the annealing heat treat ments on growing the twins. Babiak and Rhines showed the importance of twins on the grain size determination in that without taking annealing twins into account, the Hall-Pe tch relationship did not hold up [60]. In this study, twins were not taken into account for grain size determination as not all of the samples exhibited twinning and the grain size s were used solely as a method of comparing the efficacy of recrystallization heat treatments and deformation processes. Carbides and Impurities Carbides in nickel based superalloys are typically formed at the grain boundaries and fall into primary and secondary carbides. Pr imary or MC carbides form fi rst after solidification at grain boundaries and in the matrix from elements including: vanadium, niobium, tantalum, ti tanium and hafnium [10]. At lower temperatures, pr imary carbides transform into more stable M23C6 carbides, which typically form as blocky, discontinuous precipitates on the grain boundaries, where they inhibit grain boundary sliding. These usually chromium rich carbid es can also have a detrimental effect in lowering the ductility as continuous grain boundar y films [10]. 152

PAGE 153

Powder metallurgy versions of IN-100, es pecially the high carbon variant (0.18 wt% carbon), formed TiC primary carbides (MC ) on the prior parti cle boundaries, which inhibited grain boundary, motion such as slid ing [10]. The low carbon variant, which was used in this study (0.07 wt% carbon), was created to allow for larger grain sizes to improve creep strength at high temperatures, but overall prov ided better properties such as ductility and toughness owing to the decrease in brittle carbide phases. This variant also forms TiC primary carbides as seen in Fi gure 7-4 in the EDS spectra of a carbide in the PPD deposit material. In a practical sense the purpose of carbides in this study was to pin the grain boundaries in the recrystallized condition and slow down grain growth. The use of the low carbon variant reduces t he amount of Zener drag of the carbides on the grain boundaries. Zener drag is the force exert ed on the grain boundaries by particles like carbides, which, retarded grain growth during heat treatment [16, 61]. Impurities inherent in the microstructure as seen in Figure 7-5, can have the same effect as carbides in preventing ex cessive grain growth. These microstructural features, carbi des and impurities, coupled with keeping optimizing the heat treatments to limit time and temperatur e to levels necessary for RXN, assisted in reaching a minimum gr ain size for recrystallized samples. Segregation and Homogenization The deposition processes, because of t he rapid melting and cooling of the deposited material causes segregation in the sa mples, as shown in Waspaloy by Moat et. al. [62]. This segregati on like coring in large cast ingots, is caused by dendritic freezing where as the dendrites grow lighter elements are reje cted into the interdendritic liquid and makes up the last material to solidify. 153

PAGE 154

Cored structures are undesir able because of their effe ct on the properties of a material, which inc lude brittleness, and inhomogeneity of mechanical and physical properties. The cored nature of the microstructure can most easily be seen by the layer bands that formed by the re-melting of the dendrites at the start of each layer [17]. These layer bands are evident in the as-deposited microstructure both before heat treatment and post heat treatment if either time or temperat ure were not sufficient to allow for homogenization. To remove the segregated nature of the microstructure, homogenizing heat treatments are perform ed. These are heat tr eatments at such te mperatures where diffusion is rapid enough to allow homogeni zation of the microstructure. The temperature must be high enough to allow for diffusion but below the solidus to prevent melting of the alloy. The time necessary for homogenization depended on several factors including the extent of coring, temper ature and diffusion rate [63]. In the case of the IN-100 deposits, the tem perature had to be relatively high due to the level of segregation and the presence of heavy elements like v anadium and molybdenum, which required more energy for diffusi on. It was also possible to combine homogenization and recrystallization into one process. The addition of cold work, which is necessa ry to recrystallize the microstructure, allows for annealing heat treat ments at time and te mperatures to remove compositional gradients in the samples. This was acco mplished because RXN increases the diffusion rates and accelerates homogenization [63]. These heat treatments permitted for the resolution of the two issues for repair of IBRs: the segregation during deposition and the dendritic as-deposited microstructure. 154

PAGE 155

The temperatures used during the proce ss of recrystallizing the sa mples were sufficient to homogenize the microstructures, but the longer times (1 hour) resulted in what appeared to be more completely hom ogenized structures with possible grain growth while shorter times (<30 minut es) showed evidence of incomplete homogenization but seemingly fully recrystallized grain structure. For current processing parameters, these two goals of homogenization and recrystallization were competing processes and mechanical testing is required to determine which condition is more significant to the repair of the material. Porosity For the sake of discussion, the deposition process can be thought of as a casting where each successive layer behaves like the walls of a casting mold. The major difference in this analogy is the rapid cooling rate during deposition due to the small workpiece sizes. In casting, as in the deposition, porosity can be caused by several issues. Shrinkage porosity is the result of the 36% volume difference between the sample at temperature and after cooling from liquid to solid, and for dendritic structures form when the dendrite arms restrict the flow of th e interdendritic liquid and pores at the base of these arms are formed [ 16]. The volume of the shrinkage must be accounted for and this porosity can be reduced by control of processing parameters. The type of porosity that was more import ant to the study was thermally induced porosity (TIP). TIP occurs when the atomizin g gases, such as argon in the case of the powder metallurgy processed IN-100 used, become trapped within hollow powder particles, which become part of the consol idated structure [64]. Upon heat treatment, the building of pressure in the closed stru ctures causes expansion of the pores, so 155

PAGE 156

samples may appear to have small pores and after heat treatment may exhibit large porosity [10]. To identify possible TIP, sampl es were examined unetched before and after heat treatment conditions Figure 7-6 showed that the amount and size of porosity in the transition area from base metal to deposit remained the same despite 1 hour at 1232C. With TIP not a factor in the final microstr ucture the porosity that was evident could be reduced or prevented by optimization of the process parameters during deposition. Summary The most important microstruc tural features to determine the viability of the final repaired part were the grain structure and the amount of homogenization. The optimal part would show a fine recrystallized grain size and a fully homogenized microstructure. The other structures seen in these sa mples are of secondary importance because though porosity is detrimental for performance, optimization of processing parameter can limit the inherent porosity and it was shown that TIP is not a factor for this material and heat treatment. While twinning, carbides and impurities certainly affect the final performance of the repaired microstructure; they were overshadowed by the importance of grain structure and a homogenous microstructure. Recrystallization Factors With the main goal of the study being the recrystallization and refinement of the as-deposited structure to more closely match the original material, the factors affecting RXN became very important as they influen ce the final grain size along with other properties as seen in Figure 7-7. These fa ctors that were examined included: heat treatment time and temperature, purity, prior deformation and prior grain size [65]. Each of the factors affects the am ount of recrystallization that occurred in the sample. Some 156

PAGE 157

of the factors are fixed, but others were changed during the c ourse of the investigation and through the analysis of the sample matrices and the interaction of the RXN factors, the best possible methodology can be determined. Time The annealing time or the length of the annealing heat tr eatment is inextricably linked to the annealing temperatur e in that generally as the time increases the minimum RXN temperature decreases. It can be seen in Figure 7-8 that as the annealing time decreases from 10 hours to 1 hour, the corresponding RXN temperature will increase. This occurs because at its core, recrystalliz ation is a nucleation and growth process [65]. New grains nucleate at areas of high di slocation density turning the stored energy from working the material into these new gr ains. This process takes time, with the time necessary dependent on the temperat ure of the heat treatments. In this study several HT times were examined: 0.25, 0.50, 1 and 2 hours. These are relatively short for annealing times; however this was done to minimize the effect of the heating on the base metal attached to the deposit. In general, the goal of this work was to minimize grain growth as well as keep the costs low. For the purely thermal heat treatments, the 0.5 hour samples showed le ss RXN than their 1 hour counterparts in the form of less homogenous and more dendritic micr ostructures. Additionally the 2 hour treatment did not show subst antial improvement over the 1 hour treatments to justify its use. For this application, having treatm ents that were too long could be counterproductive, as once RXN is complet ed, the grains, which recrystallized, will begin to grow, producing less desirable sample s. This was investigated once TMP was being used to refine the samples. The addition of mechanical deformation allowed for 157

PAGE 158

the exploration of lower annealing times su ch as 0.25 and 0.50 hours. These shorter treatments proved as effectiv e as the longer treatments, if there was sufficient deformation imparted to the sample. Temperature Temperature started out as t he main factor under invest igation, as along with the time it had the most potential effects on th e base metal. Throughout the investigation temperatures ranging from 1093 C to 1232C were inv estigat ed in search of the RXN temperature or the temper ature where in 1 hour RXN is 95% complete [58]. The temperature was limited on the high end initially by the incipient melting temperature and later by industrial concerns. On the low end, 1093C, was a common heat treatment temperature for IN-100. Du ring the research it was demonstrated consistently that at temperature lower than 1177C, almost no RXN occurred. It was then determined that this was the minimu m RXN temperature for times 1 hour and shorter and at deformation levels below about 15%. Above this temperature recrystallization of the dendritic structure seemed to occur readily under many different times and deformation conditions. It was possible that with higher deforma tion and or longer times that this temperature could be lowered as is demonstrated with copper in Figure 7-9; however the conditions tested represent what was thought to be industrially feasible. Initially the temperature of the heat treatments was treated as a variable factor, but the experimental results repeatedly indicated that temperatures lower than 1177C would not be effective, while 1232C was eliminated due to the concern over excessive grain growth in the base metal. These temperat ures (1177C and 1204C) proved to effective 158

PAGE 159

over all the deformation levels with no samp les that retained their previously dendritic microstructure. Purit y This factor was basically fixed from the beginning of the study as the alloy, IN-100 was chosen by the industrial sponsors but purity of the material does have an effect on the recrystallization temperat ure. Higher purity materials have lower RXN temperatures than materials that are alloyed [65]. As nickel based superalloys are heavily alloyed, their RXN temperatures are higher than that of pure nickel, which is 600C. IN-100 which is roughly 55% nickel should have a higher RXN temperature which was borne out through the experiments. Detert had shown the effect of varying the amount of chromium and molybdenum on t he RXN temperature of nicke l alloys [66]. This can be seen in Figure 7-10, where adding 10 wt% Cr increases the temperature about 100C and 10 wt% increases the temperature by 200C The presence of however; aids recrystallization because hard second phase particles have been shown to act as originat ors of stress concentrators and localized strain during deformation [67]. Prior Deformation As has been stated previously, prior deformation or cold working the sample has the effect of increasing the amount of stor ed energy available to aid RXN. The more cold work in the sample the easier it is to recrystallize the sample. Thermal treatments at first were used to take advantage of the energy from the thermal strain of the deposition process. The heat treatments were then augmented with ratcheting treatments with t he goal of increasing the stored energy and thus aiding 159

PAGE 160

RXN. Thou gh these samples saw some recr ystallization, they did not achieve the desired fine grain size. Shifting to thermo-mechanical proc essing and adding cold work through mechanical deformation instead of ratchet ing treatments allowed for significant increases in energy available for RXN. By using rolling as the means of imparting the deformation on the samples, controllable and measureable levels of cold work were added to the samples, which allowed for the evaluation of the critic al deformation levels for RXN. The critical level for deformation is the minimum deformation level below which there is not enough strain energy to nucl eate new grains [39, 68]. This was demonstrated with data from recrystallization studi es of aluminum sheet in Figure 7-11. Figure 7-12, also highlights the importance of deformation level in the RXN of alpha brass; however it is also shown that the temperature of recrysta llization does not matter as long it is high enough to induce RXN. To achieve the wide scale RXN to reach the smallest possible grain size, it appeared to be necessary to deform the sample to at least 4%; however full results from grain si ze analysis will be presented later in this chapter. Prior Grain Size This is another fixed factor as the deformation process used set the microstructure before any thermo-mechanical treatments. Bo th the PPD and LENS processes resulted in microstructures that were coarse and dendritic with no grains. The differences between the two processes resulted in the LE NS samples having finer structures with dendrites running only in the deposition direct ion due to the decreas ed amount of remelting that occurred between layers. The LENS process because it used a more 160

PAGE 161

precise heat source resulted in finer structu res with smaller layers; this had the effect of increasing the thermal strain in the samp les and having a less segregated structure. While the prior grain size of the sa mples did not have an effect on the RXN temperature, as is usually the case, the LENS samples seemed to produce better results based on the differences in the deposition processes. Summary Throughout the study of the sample microstr ucture it was shown that overall the two most important factors in inducing enough RXN to approach the optimal grain size were temperature and deformation level. If there was sufficient cold work and the sample was heat treated above the solvus for a reasonable amount of time then RXN occurred in some manner; however the goal of the project was to determine a methodology to achieve comparable grain sizes to the base metal. Thus, it was necessary to compare the resu lts of the experiments across sample matrices to isolate the effect of the different factors, which was done using data measured from the micrographs. Recrystallization Factors and the Effect on Grain Size In order to isolate the importance of the factors on the overall methodology, quantitative measurements were used to add to the lessons learned from the microstructures to form a complete picture. Grain size analysis as detailed in the Experimental Procedures chapter was performed on all the sa mples that saw sufficient RXN to create a grain structure. The data gener ated was in the form of major and minor axes in microns since the grains were frequently not equiaxed. These measurements were used to get an overall grain area for each sample in square microns and an aspect ratio (major/minor axis) that corresponded to the elongation character of the grains in 161

PAGE 162

the sample. Aspect ratios closer to one refl ected equiaxed characters while the further away from 1 indicated elongated characters. If a sample had an aspect ration below 1 it signified elongation in the direction perpendicular to the rolling/deposition direction. Values which were higher than 1 as seen in some of the rolled samples corresponded to elongation along the rolling direction. T he breadth of the data precludes the analysis of factors with figures that include all the specimens. The figures included will contain those samples, which illustrated the effect of specific factors on the process. Overview An overview of all the samples analyzed in this grain size study can be seen in Figure 7-13. The number of samples precludes examining the sample for trends however some observations can be made from this figure based on the grouping of the samples. While there seemed to be a lar ge scattering of samples there is a large grouping of samples between 4 and 10% deformation and grain areas under 50,000 um2. This seemed to indicate that this ar ea can be reached reliabl y with the processing routes examined in this study, so 50,000 um2 was established as the cutoff for a sample to be considered having been effectively recr ystallized. Though there are other samples under this cap including in-situ, high def ormation rolling and some with only thermal treatments, the majority of these samples were ro lled indicating that uniform deformation was the key to their success. Also none of the ratcheted samples were under the cutoff line, indicating that the ex tra heat treatments were not effective in adding strain energy but this will be examined later. Time and Temperature Effects As was discussed previously, the temperature the samples were heat treated at was of great importance to the level of RXN seen. Looking at its effect on control 162

PAGE 163

samples is olated the effect of temperature. These control samples were base metal samples that were rolled to the same def ormation level as the deposited samples, in this case 8.5 % RIA, and then heat treated for 1 hour at the tem peratures that have been evaluated in the study. The results of the analysis can be seen in Figure 7-14. The figure indicated that the fine grains structure of the base metal did not see any effect of the temperature until above 1149 C, where the grains grew rapidly. This supports the results from all of the different processi ng routes examined duri ng the course of the study, that there is a mi nimum temperature that needed to be reached in order to activate RXN. The temperature shown in this figure corresponded with the point where the sample microstructures began to show evidence of recrystallization, which, showed that samples must be heat treated at tem peratures least 1177C to see any large scale RXN of the structure. The effect of time is not as easily is olated as temperature, for with enough time a low temperature heat treatment could be effective for recrystallization; however as this was not industrially desirable, the time scale for heat treatment s was limited to an hour or less. It was initially t hought that the lower treatments were not as effective at recrystallizing the samples as the longer treatments, but with the addition of mechanical deformation, the shorter heat treatments bec ame more viable. The in-situ deformed samples in Figure 7-15 show the benefit of thermo-mechanical processing over thermal processing for RXN as the other samples shown featured a ratcheting treatment. The effect of increasing lengths of heat treatments in this figure showed that increased time could lead to grain growth in the sample s; however the lessons learned from the microstructures of these specimens was that shorter times led to incomplete 163

PAGE 164

homogeniz ation of the samples. This could be seen throughout the samples matrices in the form of unsolutioned par ticles and layer bands from the deposition process. Ideally the repair methodology would result in a microstructure that is fully recrystallized and fully homogenized; however with these competing goals it might be necessary to choose which has the biggest effe ct on the properties of the final part. In this case, mechanical testing would be needed to determine the most effective final processing route. Alternativ ely, the temperature could be increased however the increases must be limited due to concerns for possible incipient melting and grain growth in the base metal. Again these are choices that must be made in order to optimize the methodology, but each change has trade-offs that must be investigated. Deformation Level The level of deformation added to the sa mples before heat treatment has proven to be another important factor in the recrystallization seen in the final microstructure. As was seen in Figure 7-13, the most consistent samples had deformation levels between 4 and 10%, these samples were examined and plotted by their deformation level for both low and high levels in Figures 7-16 and 7-17. The low deformation condition shows a marked difference between the LENS and PPD deposits, with the LENS being significantly under the cutoff poi nt indicating finer grain sizes, while the PPD pr ocess exhibited coarser struct ures and thus were generally unsuitable. Also while the LENS samples remain relatively consistent for all temperatures evaluated, the PPD sample s clearly showed the effect of high temperature with the higher te mperature treatment s having better results than the lower temperatures. 164

PAGE 165

In Figure 7-17, the samples defo rmed between 8 and 10% showed similar behavior for both the LENS and PPD materials, neither markedly better than the other; however the sample group as a whole in under the cutoff po int. A dditional deformation resulted in both LENS and PPD materials exhi biting similar RXN and this can be seen because the scale of the graph is different than Figure 7-16. When the rolled samples (between 4 and 10% deformation) were compared base on heat treatment temperature (Figure 7-18), it was seen that while the higher temperature (1204C) exhibited better overall results in reaching lower grain areas, the lower temperature (1177C) showed more consistent results across different deformation levels. Also important to note is that the LENS samples, as a whole, performed better than the PPD samples. This confirms what was seen in the microstructures, that the LENS samples showed overall a more homogeneous structure with less porosity and inclusions. This will be further discussed below. Ancillary Effects Time, temperature and deformation level have proven to be the most important factors in the investigation; however thei r effects were not the only ones that were observed during analysis of all of the samples. In Figure 7-19, the grain area of two sets of samples is plotted against their aspect ratio. These samples are both LENS and PPD processed samples, one that was ratcheted and one that was not. None of the samples had any added deformation. This figure further demonstrated t hat the LENS samples exhi bited finer, more equiaxed microstructures than the PPD samples under the same conditions. Additionally, the ratcheting treatments did not pr ovide any significant refinement over one step RXN heat 165

PAGE 166

treatments. It is important to note that while the non-ra tcheted samples both exhibited similar aspect ratios, the ratcheted samples showed as pect ratios significantly greater than 1, being elongated in the deposition direction. The relationship between pre-treatments, temperature and ra tcheting treatments was shown in Figure 7-20. In this graph it can be seen that higher heat treatment temperatures performed significantly better than the lower temperatures including the ratcheting treatments. It is al so important to note that the as-deposited condition exhibited better or similar results to samples given the pre-treatments. This is indicates that there is no substantial benef it to pre-treating of the mate rial to facilitate deformation. Further, removing the pre-treat ments from the processing route would decrease the risk of grain growth in the base metal and dec rease the number of thermal cycles the deposit and adjacent base me tal would be exposed to. The ability to eliminate t he pre-treatments was also demonstrated by Figure 7-21. Each condition has 2 data points, each representing a 1177C and 1204C recrystallization heat treatment after deforma tion. This figure showed the effect of varying both pre-treatments and deformation level on grain area. It can be seen that higher deformation resulted in more samples under the 50,000 um2 cutoff. These results, when combined with earlier result wit h similar thermal cycles, indicates that the deformation level controls the amount of RXN in the sample. Higher deformation levels also seemed to lessen the effect of heat treat ment temperature, as the exhibited similar grains sizes. It can also be seen that in the lower deformation condition, the asdeposited samples generally exhibited finer grai n sizes than the pre-treated; while in the higher deformation condition all the samples ma intained similar grain areas. Therefore 166

PAGE 167

the pre-treatments did not provide a tangible benefit when weighed against the possib ility of causing grain growth in the bas e metal, the cost of the pre-treatment and the potential for errors to be made during additional heat treatments. Comparison to ASTM Grain Size Though by necessity the procedure to quantif y the results from this study deviated from the industry standard ASTM grain size numbers [69 ], the results could still be compared using data from ASTM Standard E 112 [29]. The relationship between the grain size number and grain area in um2 can be seen in Table 7-2. The data from this study was re-plotted (Figure 7-22) and the relationship to ASTM grain size was determined. Samples were sorted by their grain area and then all those that didnt fall under the cutoff point of 50,000 um2 were culled. The results were presented in Table 7-3. None of the deposited samples approached the finegrained base metal even after grain growth (Samples C1 to C5); however the bes t samples shown exhibited a grain size of almost 4.0 (Figure 7-23). Though not the same as the base me tal, these microstructures appear to be fine enough to be considered a vi able path for the repai r of the damaged IBR components. However it will be necessary to also confirm the viability by mechanical testing of the deposited material. Summary Through quantitative analysis of grain area and comparison between conditions the most favorable conditions for recryst allization of the deposited samples were determined and the values for all the samples evaluated in this study can be seen in Appendix C. 167

PAGE 168

168 The most important factor seen was deformation, withou t the minimum amount of deformation of around 4% the grain size did not consistently reach quantities, which were deemed acceptable. Higher levels of deformation ensured that the samples consistently reach finer more equiaxed grain sizes. The manner of deformation was important insomuch as it had to be more or less uniform, penetrating throughout the samples. The other critical factor in reaching the final grain size that was acceptable was the temperature of the post deformation recrystallizing treatment. As demonstrated heat treatment temperatur es must be above the solvus for recrystallization to occur. The benefit of higher temperatures can be s een in increased homogenization and the elimination of segregat ion, while lower temperatures seemed to be more effective for recrystallization. The ancillary effects observed proved that the pre-deformation treatments and ratcheting treatments did not pr ovide significant benefit to justify the cost and risk of grain growth in the base metal. Also the LENS deposition process has been clearly shown to be superior to the PPD process in terms of cleanliness and homogenization.

PAGE 169

Figure 7-1. Dendrite structure evolution with temperature [Reprinted with permission from David A. Porter and Kenneth E. Easterling, Phase transformations in metals and alloys, 2009, p.235, Figure 4.42] 169

PAGE 170

A ) B) C) D) E) Figure 7-2. Effect of temperatur e on the dendrite st ructure at 50x A) 1093C B) 1121C C) 1149C D) 1177C E) 1204C 170

PAGE 171

Figure 7-3. Schematic of a twin boundary [Reprinted with permission from Daniel Henkel and Alan W. Pense, The Structure and Properties of Engineer ing Materials, p. 90, Figure 4-6] Figure 7-4. EDS spectra of a primary carbide 171

PAGE 172

Figure 7-5. As-deposited microstructure befor e heat treatment feat uring impurities and porosity at 100x A) B) Figure 7-6. Microstructures of transition region at 50x A) Before B) after heat treatment 172

PAGE 173

Figure 7-7. Illustration of the effect of cold work and annealing temperature on Cu35Zn Brass [Reprinted with permission from Davi d R. Askeland and Pradeep P. Phule, The Science and Engineering of Ma terials, p.336, Figure 7-18] Figure 7-8. Relationship betw een annealing time and temperatures [Reprinted with permission from Davi d R. Askeland and Pradeep P. Phule, The Science and Engineering of Ma terials, p.338, Figure 7-19] 173

PAGE 174

Figure 7-9. Changes in anneal ing temperature and deformati on in high purity copper [Reprinted with permission from Daniel Henkel and Alan W. Pense, The Structure and Properties of Engineer ing Materials ,p. 105, Figure 4-12] Figure 7-10. The effect of additions of chromium and molybdenum on recrystallization temperature [Reprinted with permission from K. Dete rt, Influence of Alloying Additions on Primary Recrystallizati on in Nickel Alloys, Recrystallization, Grain Growth and Textures Seminar Proceedings, p.134, Figure 1] 174

PAGE 175

Figure 7-11. Recrystallized grain size as a function of prior plastic deformation [Reprinted with permission from Michael F. Ashby and David R.H. Jones: Engineering Materials 2: an introduction to microstructures, processing and design, 1986, p. 141, Figure 14-11] Figure 7-12. Recrystallized grain size as a function of prior plastic deformation with recrystallization temperatures [Reprinted with permission from Robert E. Reed-Hill and Reza Ab baschian, Physical Metallurgy Principles, 1994, p.248, Figure 8-21] 175

PAGE 176

Table 7-1. Legend for grain size analysis Figure 7-13. Grain area vs. Deformation Level plot t hat shows all the samples examined 176

PAGE 177

Figure 7-14. The result of heat treatments on the base metal control samples at various temperatures after 1 hour Figure 7-15. Grain size as a function of ti me for identical samples with varying heat treatment times 177

PAGE 178

Figure 7-16. Grain area for low deformation condition graphed against processing conditions Figure 7-17. Grain area for medium deformation condition graphed against processing conditions 178

PAGE 179

Figure 7-18. Grain areas for low and medium deformation Figure 7-19. Ratcheted and non-ratcheted LE NS and PPD samples under identical conditions 179

PAGE 180

Figure 7-20. The relationship between pr e-treatments and tem perature/ratcheting treatments Figure 7-21. Grain area versus deformation for low and medium deformation conditions 180

PAGE 181

Table 7-2. ASTM grain size num ber and the corresponding grain area [Reprinted with permission from ASTM E112: Standard Method for Determining Average Grain Size ASTM International, 2007.] 181

PAGE 182

Figure 7-22. Graphical relationship between grain area and ASTM grain size number Table 7-3. Estimated ASTM grain size numbers for the most promising results sorted by increasing grain area 182

PAGE 183

183 Figure 7-23. Graphical repr esentation of Table 7-3

PAGE 184

CHA PTER 8 CONCLUSIONS Closing Remarks The results presented here show that the in itial goal of the project, to develop a methodology that would result in an airfoil that had a microstructure similar to the undamaged section of the part, was achieved. Initially the effects of a variety of heat treatments, featuring times from 30 minutes to 2 hours and temperatures from 1093C to 1232C, on the coarse, asdeposited microstructure was determined. Addi tionally the effect of supplemental thermal strains in excess of the inherent t hermal strain from the deposition process was explored by ratcheting heat treatments. In these ratcheting heat treatments the samples were placed in a furnace at temperature then rapidly cooled to induce thermal strains to aid in recrystallization. The use of onl y thermal treatments provided insufficient recrystallization to meet the goal. The higher temperatures above the solvus were the only temperatures that exhibi ted refinement of the dendritic structure. Shorter heat treatments (30 minutes) were less effective than the longer heat treatments (1 hour). Even the additional thermal strain provi ded by the ratcheting heat treatments did no provide significant benefit over the recrysta llization heat treatment by itself. At a high temperature (1232C) the ratcheting treatment proved detriment al to the microstructure resulting in substantial intergranular cracking. When this approach proved ineffective, mechanical deformation was examined using a test case involving hardness indent er that provided local deformation. The sample was then heat treated to induce recr ystallization, and the areas with the applied local deformation were compared to the bulk with a larger incidence of recrystallized 184

PAGE 185

grains seen around the indent. With the results from the test case it was determined that thermo-mechanical processing c ould induce full recrystallization in the samples and result in a microstructure close to that of the original airfoil. The thermo-mechanical processing was per formed using rolling as the means of mechanical deformation. Samples of both PPD and LENS deposition processes were rolled to 2 levels of deformation low (4-6%) and high (8-10%), then heat treated at times and temperatures that re sulted in recrystallization in other sample matrices (1 hour and >1177C). The result of the rolling study was significant recr ystallization of the microstructure. Recrystallization was seen in all deformation levels low and high. With the feasibility of TMP established, alternate methods of deformation that could be applied to the IBR in-situ were expl ored. The methods tested included surface improvement techniques such as shot-peening, low plasticity burnishing (LPB) and laser shock peening (LSP). These methods we re inadequate for the purpose of recrystallization compared to the results seen in the ro lling study; however another proprietary in-situ deformation method proved effective. In addition to microstructural characte rization, grain size analysis was undertaken to examine the effects of processing param eters on the degree of re crystallization. The processing parameters were examined included deformation level, deformation method, time and temperature. The correlation among these factors and their effect of grain size were analyzed and used to form the final processing window. The most important factors were found to be time, te mperature and deformation level. To achieve the optimum level of recr ystallization the sample must be deformed in a uniform manner to a minimum level of 4% and subsequently heat treated for at least 185

PAGE 186

186 15 minutes up to an hour at te mperatures above 1177C or the solvus temperature. Within this window the individual proce ss parameters can be varied for optimal recrystallization. Future Work The groundwork for this process has been la id out; however it is still possible to optimize the parameters to fu rther refine the structure. The parameter with the most variability is the method of deformation. Ma ny have been tried and it becomes important to appraise the most feasible method t hat will allow for the maximum amount of deformation in the repaired airfoil. Te mperature was another important process parameter and the effect of higher temperatur es on the homogeneity and recrystallization of the samples. Samp les must be heat treated at both 1177C and 1204C under the same deformation condition to understand the effect of temperature. Additionally the result of decreasing the heat treatment time and its effect on the final recrystallization must be investigated. By working with industry to further refine the process an optimal process following the guideline laid out from this investigati on can be created. Additionally, the viability of the repaired microstruc tures must be examined by mechanical testing of components repaired using the methodology outlined in this study. With homogenization and recrystallizati on being competing goals, the efficacy of both fully recrystallized and fully homogenized structures under loads that simulate service must be assessed.

PAGE 187

APPENDIX A ROLLING FORCE CALCULATI ON Before rolling, the rolling forces must be calculated to ensure that the forces exerted during rolling do not surpass the capacit y of the rolling mill. For the low frictional condition, equation A-1 is used. (A-1) Where w is the width of the strip in this ca se 0.625 inches, Y is the average flow stress for IN-100 is 145 ksi [10]. (A-2) Substituting equation A-2 into A-1 yields equation A-3 (A-3) Here R is the radius of the rolls (3 inches) and h is the difference in height before and after rolling. For an 8% pass using equati on A-4 to put the reduction per pass into A-5. (A-4) (A-5) The results from A-5 are put into A-3 along with the given values to determine the final value. 187

PAGE 188

188 This value is far below the maximum load of 100,000 lbs. so it is safe to do 8% reduction in area passes on the Fenn rolling mill.

PAGE 189

APPENDIX B LOW PLAS TICITY BURNISHING HARDNESS DATA Table B-1. Inside edge length hardness trace Table B-2. Outside edge length hardness trace 189

PAGE 190

190 Table B-3. Center section hardness traces Table B-4. Width traces

PAGE 191

APP ENDIX C GRAIN SIZE DATA FOR PPD AND LENS SAMPLES Table C-1. Grain size data for PPD samples Table C-2. Grain size data for LENS samples 191

PAGE 192

LIST OF REFERENCES [1] T.A. Kolakowski and D.H. Maxwell: The Crystallography of Cast Turbine Airfoils, Quest, 1980 [2] R.C. Reed: The Superalloys: Fundamentals and Applications Cambridge University Press, 2006 [3] N.A. Cumpsty: Jet Propulsion: A Simple Gu ide to the Aerodynamics and Thermodynamic Design and Performance of Jet Engines Cambridge University Press, Cambridge, 1997. [4] C.B. Meher-Homji: Anslem Franz and the Jumo 004, Mechanical Engineering September 1997, p. 88. [5] R.F.Decker: Strengthening mechani sms in nickel-base superalloys, Steel Strengthening Mechan isms Symposium Zurich, Switzerland, May 1969, p. 1-24 [6] H.K.D.H Bhadeshia: Nickel Base S uperalloys, University of Cambridge. [7] Nickel-Base Alloys, ASM Handbooks, ASM International, 2003. [8] E. W. Ross and C. T. Sims: Nickel-Base Alloys, Superalloys II C. T. Sims, N. S. Stoloff, and W. C. Hage l ed., Wiley, New York, New York, 1987, pp. 97-133. [9] F.L. Versnyder and M.E. Shank: The De velopment of Columnar Grain and Single Crystal Temperature Materials through Directional Solidification, Source Book on Elevated Temperature Material s Metals Park, Ohio, 1979. [10] G.H. Gessinger: Powder Metallurgy of Superalloys Butterworth & Co., London,1984. [11] S.H. Reichman and J.W. Smythe: New Developments in Superalloy Powders, Modern Developments in Powder Metallurgy, Hauser. ed, Vol. 5, Plenum Press, 1970, p. 73. [12] K. Bowman: Mechanical Behavior of Materials John Wiley & Sons, Inc., Hobken, NJ, 2004. [13] S.H. Reichman and J.W. Smythe: S uperplasticity in P/M In-100 Alloy, International Journal of Powder Metallurgy Vol. 6, 1970, p. 65. [14] N.J Grant, L.N. Moskomitz and R.M. Pelloux: Properties of IN-100 Processed by Powder Metallurgy, Superalloys 1972 Conference Proceedings TMS, 1972, p Z1 Z-26. 192

PAGE 193

[15] R.L. Athey and J. B. Moore: Development of IN100 Powder Metallurgy Discs for Advanced Jet Engine Application, 18th Sagamore Army Materials Research Conference August, 1971 [16] R. Abbaschian and R.E. Reed-Hill: Physical Metallurgy Principles PWS Publishing Co., Boston, MA, 1992. [17] N.I.S. Hussein, J. Segal, D.G. McCartney, and I.R. Pashby: Microstructure formation in Waspaloy multilayer builds following direct metal deposition with laser and wire, Materials Science and Engineering A 497, 2008, p. 260-269. [18] D.R. Poirer and G.H. Geiger: Transport Phenomena in Materials Processing TMS Publications, Warrendale, PA, 1994. [19] N.Calder and M. Hedges : Near Net Shaped Rapid Ma nufacturing and Repair by LENS, Unclassified NATO-OTA N Report, May, 2006, 13-1-13-14. [20] R.P. Mudge and N.R. Wald: Laser E ngineered Net Shaping Advances Additive Manufacturing and Repair, Welding Journal January, 2007. [21] J.G. Byrne: Recovery, Recrystallization and Grain Growth The Macmillan Company, New York, NY, 1965. [22] J. Hoffman et. al: Rel axation of Residual Stresses of Various Sources by Annealing Residual Stresses in Science and Technology E. Acherauch and V. Hauk, ed., Vol. 2, DG M Informationsgellschaft mbH, 1987, p. 695-702. [23] J.E.Burke and D.Turnbull: Progress in Metal Physics, v.IIIB, ,1952, p.220. [24] G. Sachs and K.R. Van Horn: Practical Metallurgy: Applied Physical Metallurgy and the Industrial Processing of Ferrous and Non-Ferrous Metals and Alloys ASM, Cleveland, OH, 1940. [25] M.C. Smith: Principles of Physical Metallurgy Harper, New York, NY, 1956. [26] M.A. Grossman: Principles of Heat Treatment United States St eel Corporation, Cleveland, OH, 1935. [27] ASTM E140: Standard Hardness Conversion Tables for Metals Relationship Among Brinell Hardness, Vickers Hardness, Rockwell Hardness, Superficial Hardness, Knoop Hardness, and Scleroscope Hardness, ASTM International, 2007. [28] Identification of Compounds and Phases Using XRay Powder Diffraction, ASM Handbooks ASM International, 2003. 193

PAGE 194

[29] ASTM E112: Standar d Method for Determining Average Grain Size, ASTM International, 2007. [30] R.T. DeHoff and F.N. Rhines: Quantitative Microscopy McGraw-Hill, New York, NY, 1968. [31] G.L. Erickson: The Developm ent and Applications of CMSX-10, Superalloys 1996 Conference Proceedings D. L. Anton, A. D. Cetel, D. J. Deye, R. D. Kissinger, M. V. Nathal, T. M. Pollock, and D. A. Woodford, ed., TMS, 1996, p. 35-44. [32] E.C. Caldwell, F.J. Fela and G.E. Fuchs: Segregation of Elements in High Refractory Content Single Cryst al Nickel Based Superalloys. Superalloys 2004 Conference Proceedings K.A. Green, H. Harada, T. E. Howson, T.M. Pollock, R.C. Reed, J.J. Schirra, and S, Wa lston, ed., TMS, 2004, p. 811-818. [33] S.R. Hegde a ,, R.M. Kearsey, and J.C. Beddoes a: Designing homogenization solution heat treatments for single crystal superalloys, Materials Science and Engineering A, 527, 2010. [34] D.L. Sponseller: Superalloys 2008 Conference Proceedings D. L. Anton, A. D. Cetel, D. J. Deye, R. D. Kissinger, M. V. Nathal, T. M. Pollock, and D. A. Woodford, ed., TMS, Warr endale, PA, 2008, p. 259. [35] G.L. Erickson: A New, Third Genera tion, Single Crystal, Casting Superalloy. Journal of Materials TMS, Warrendale, PA, 47, April 1995, p.36. [36] J.M. Larson: In Modern Developments in Powder Metallurgy H.H. Hauser, ed., Vol. 8, Plenum Press, New York, NY, 1974, p. 537 [37] W.D. Callister: Fundamentals of materials scienc e and engineering: an integrated approach John Wiley & Sons, Hoboken, NJ, 2005. [38] R.A. Flinn: Fundamentals of Metals Casting Addison-Wesley Publishing Co., Reading, MA, 1963. [39] M.F. Ashby and D.R.H. Jones: Engineering Materials 2: an introduction to microstructures, processing and design Pergamon Press, Oxford, England, 1986. [40] K.L. Gasko, G.M.Janowski, and B.J. Pletka: The Influence of Eutectic on the Mechanical Properties of Conv entionally Cast MAR-M247, Materials Science and Engineering A 104, 1988, p.1-8. [41] J. Piearcey and B.E.Terkelsen: Trans. Metall. Soc. AIME, 239, 1967. 194

PAGE 195

[42] Vickers Hardness Testing, ASM Handbooks, ASM International, 2003. [43] B.D. Cullity: Elements of X-Ray Diffraction Addison-Wesley Publishing Co., Reading, MA, 1978. [44] G.P. Dinda a A.K. Dasgupta, and J. Mazumd er: Laser aided direct metal deposition of Inconel 625 superalloy: Microstructural evolution and thermal stability Materials Science and Engineering A 509, 2009, p. 98-104. [45] H. Inagaki: Formation of [111] Recrystallization Textur e in Polycrystalline Iron Transactions ISIJ 24, 1984, p. 266-274. [46] M.C. Mataya and G. Krauss: J. Applied Metalworking Vol. 2, 1981, p. 28. [47] D.A. Porter and K.E. Easterling: Phase Transformations in Metals and Alloys CRC Press, Boca Raton, FL, 2004. [48] J.J. Jonas and S.L. Semiatin: Formability & workability of metals ASM International, Metals Park, OH, 1984. [49] J.M. Silcock, T.J. Heal and H.K. Hardy: J. Inst. Met., 239, p.1953-1954. [50] K.K Chawla and M.A, Meyers: Mechanical Metallurgy Principles: Principles and Application Prentice Hall, Inc., E nglewood Cliffs, NJ, 1984. [51] The Making, Shaping and Treating of Steel W.T. Lankford, N.L. Samways, R.F. Craven and H.E. McGannon, ed ., United States Steel, 10th Edition, 1985. [52] K. G. Budinski and M.K Budinksi: Engineering Materials Properties and Selection Pearson Prentice Hall, Upper Saddle River, NJ, 2005. [53] M.J. Shepard, et al: Introduction of Compressive Residual Stresses in Ti-6Al-4V Simulated Airfoils via Laser Shock Processing, Journal of Materials Engineering and Performance 2001, Vol. 10, No. 6, pp. 670-678. [54] A. Clauer, J.H. Holbrook, and B.P. Fa irland: Effect of Laser Induced Shock Waves on Metals, Shock Waves and High Strain Rate Phenomena in Metals Plenum Publishing Corp, New York, NY, 1981, pp. 675-703. [55] W. Beres: FOD/HCF Resistant Su rface Treatments, Unclassified NATO-OTAN Report RTO-TR-AVT-094, 5-1-5-12. [56] P.S. Prevey et. Al.: Method and Syst em for improving a parts resistance to stress induced failure. US Patent 7,219,044, 2007. 195

PAGE 196

196 [57] P.S. Prevey et. Al.: The Effect of Low Plasticity Burnishing (LPB) on the HCF Performance and FOD Resistance of Ti-6Al-4V, 6th National Turbine Engine HCF Conf erence Proceedings Jacksonville, FL, 2001. [58] D. Henkel and A.W. Pense: The Structure and Prope rties of Engineering Materials McGraw-Hill, New York, NY, 2002. [59] R.H. Hertzberg: Deformation and Fracture Mechanics of Engineering Materials John Wiley and Sons, 1989. [60] W.J. Babiak and F.N. Rhines: Trans AIME, 218, 21, 1960. [61] L.Wang, G.Xie, J. Zhang and L.H. Lou: On the role of carbides during the recrystallization of a directionally solidified nickel-base superalloy Scripta Materialia Elsevier, 55, 2006, 457-460. [62] R. Moat, M.Ka radge, A.J. Pinkerton, A. Descham ps, F. Bley, L. Li et. Al.: Intergranular precipitati on variations in laser deposited Waspaloy due to compositional inhomogeneities, TMS Annual Meeting Proceedings TMS, Warrendale, PA, 2006. [63] M.C. Smith: Alloy Series in Physical Metallurgy Harper Brothers Publishing, New York, NY, 1956. [64] X. Zhao, J. Chen, X. Lin and W. Huang: Study on microstructure and mechanical properties of lase r rapid forming Inconel 718 Materials Science and Engineering A Elsevier, 478, 2008, 119-124. [65] D.R. Askel and and P.P. Phule: The Science and Engineering of Materials Thomson Brooks/Cole, Pacific Grove, CA, 2003. [66] K. Detert: Influence of Alloying Additions on Primar y Recrystallization in Nickel Alloys Recrystallization, Grain Growth and Textures Seminar Proceedings American Society for Metals Metals Park, OH, 1966. [67] W.C. Leslie, J.T. Michalak and F.W. Aul: Iron and Its Dilute Solid Solutions Interscience, New York, 1963 [68] J.S. Smart and A.A. Smith: Trans. AIME 152, 103, 1943. [69] L.H. Van Vlack: Elements of Material Science and Engineering Addison-Wesley, 1989.

PAGE 197

BIOGRAPHICAL SKETCH Alvaro G. Mendoza Jr. was born at St. Ma rys hospital in West Palm Beach, Florida to Alvaro and Hilary Mendoza. He liv ed in Jupiter, FL and attended Palm Beach Day School until continuing his education at The Benjamin School before graduating in 2003. He matriculated at the University of Fl orida in Gainesville and began his studies in Materials Science and Engineering. Choos ing to specialize in metals, he graduated cum Laude in 2007 with a Bachelor of Science degree. He decided to continue his education at t he University of Florida and during his graduate career worked with Dr. Fu chs on high temperature alloys for use in turbines for jet engines. He received his Ph.D. in 2012 and upon graduat ion he will begin employment in Savannah, GA wo rking in the forging industry. 197