UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Austenite decomposition of a HSLA-Nb/Ti steel and an A1-TRIP steel during continuous cooling Lottey, Kulwinder Kaur 2004

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2005-0079.pdf [ 11.45MB ]
Metadata
JSON: 831-1.0078775.json
JSON-LD: 831-1.0078775-ld.json
RDF/XML (Pretty): 831-1.0078775-rdf.xml
RDF/JSON: 831-1.0078775-rdf.json
Turtle: 831-1.0078775-turtle.txt
N-Triples: 831-1.0078775-rdf-ntriples.txt
Original Record: 831-1.0078775-source.json
Full Text
831-1.0078775-fulltext.txt
Citation
831-1.0078775.ris

Full Text

AUSTENITE DECOMPOSITION OF A HSLA-Nb/Ti STEEL AND AN Al-TRIP STEEL DURING CONTINUOUS COOLING By Kulwinder Kaur Lottey B. Eng. & Mgt, McMaster University, Hamilton, Canada, 2002 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T F O R T H E D E G R E E OF M A S T E R OF APPLIED S C I E N C E in T H E F A C U L T Y OF G R A D U A T E STUDIES D E P A R T M E N T OF M A T E R I A L S E N G I N E E R I N G We accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH C O L U M B I A November 2004 © Kulwinder Lottey, 2004 Abstract The ability to control and predict the mechanical properties of hot rolled steel depends strongly on the thermomechanical processes which the steel undergoes. The phase transformation that occurs on the run-out table of a hot strip mill is a critical processing step which significantly influences the final microstructure, and thus, the properties of the hot rolled steel. This work examines the austenite-to-ferrite phase transformation of a high-strength low-alloy steel (HSLA-90) microalloyed with niobium and titanium and a transformation-induced plasticity steel alloyed with aluminium (Al-TRIP). The austenite decomposition kinetics have been investigated using a Gleeble 3500 thermomechanical simulator equipped with a dilatometer. The effect of cooling rate on the austenite decomposition kinetics were quantified for both steels with initial microstructures comprised of austenite. For the Al-TRIP an additional initial microstructure consisting of a mixture of austenite (67%) and ferrite (33%) phases was also investigated. It was shown that accelerated cooling lowers the transformation temperatures in addition to refining the resulting ferrite grains. Microstructural analyses revealed a decrease in polygonal ferrite fraction with accelerated cooling and the formation of acicular products for the HSLA-90 steel. A transition from high temperature products such as, polygonal ferrite and pearlite, to low temperature products such as, bainite and martensite, was seen for the Al-TRIP with accelerated cooling where there was an increase in bainite and martensite fractions. The effect of initial austenite grain size and a pancaked austenite microstructure was investigated for the HSLA-90 steel. Increasing the austenite grain size resulted in a shift to lower transformation start temperatures and an associated decrease in the polygonal ferrite fraction. However, accelerated cooling and smaller austenite grain sizes resulted in refining the ferrite grains. Additional ferrite grain refinement was obtained with the prior deformation of the initial austenite microstructure which increased the ferrite nucleation rate by introducing additional nucleation sites both on the austenite grain boundary and within the deformed grains at crystallographic defects. The transformation start temperatures and polygonal ferrite fraction were significantly increased by the retained strain. A previously developed sequential transformation model was applied to describe the austenite-to-polygonal ferrite transformation which consisted of sub-models to predict the transformation start temperature, ferrite growth and ferrite grain size under continuous cooling conditions. The first sub-model predicted transformation start temperature by combining corner nucleation of ferrite with early growth. The subsequent ferrite growth was described using a model that employed the Avrami equation (or J M A K model) which was adapted to non-isothermal transformations, i.e. continuous cooling, by using the Scheil equation of additivity with an Avrami exponent of n = 0.85 and a rate constant b which depends exponentially on temperature; the effect of austenite grain size on the transformation kinetics was captured with a suitable grain size exponent m. The ferrite grain size was predicted by employing a model that expressed ferrite grain size as a function of initial austenite grain size and transformation start temperature. The combined effect of austenite grain size and retained strain was incorporated into the transformation start, transformation kinetics and ferrite grain size models by employing an effective grain size, i.e. deff = dy exp(- e). i i i . Table of Contents Abstract ii Table of Contents : vi List of Figures vii List of Tables xi List of Symbols xii Acknowledgements xvi Chapter 1 Introduction 1 1.1 Background 1 1.2 Scope and Objectives 3 Chapter 2 Literature Review 5 2.1 Austenite Decomposition on the Run-out Table 5 2.1.1 Microstructures Developed from Austenite Decomposition 5 2.1.2 Isothermal Transformation of Austenite-to-Ferrite 8 2.1.3 Isothermal Transformation Kinetics: JMAK Approach 15 2.2 Factors Influencing Austenite Decomposition Behaviour 15 2.2.1 Chemistry 15 2.2.2 Cooling Rate...! 19 2.2.3 Initial Austenite Microstructure 20 2.2.3.1 Austenite Grain Size 20 2.2.3.2 Effect of Deformation 21 2.3 Modelling of Microstructural Evolution Under Run-Out Table Conditions 24 2.3.1 Austenite-to-Ferrite Transformation Kinetics Modelling 24 2.3.1.1 Transformation Start Temperature 24 2.3.1.2 Application of Isothermal Transformation Kinetics to Non-Isothermal Conditions 26 2.3.1.3 Ferrite Growth Kinetics: JMAK Approach with Additivity 27 2.3.2 Polygonal Ferrite Grain Size Model 29 2.3.3 Comparison of Model Parameters for Previously Investigated Steels 30 2.3.4 Other Models to Describe the Overall Austenite Decomposition 31 Chapter 3 Experimental Methods 33 iv 3.1 Materials ,33 3.2 Experimental Equipment 33 3.2.1 Gleeble 3500 Thermomechanical Simulator 33 3.3 Experimental Methodology 34 3.3.1 Austenite Grain Sizes 34 3.3.2 Double-Hit Compression Tests 34 3.3.3 Controlled Cooling Tests 36 3.3.4 Deformation-Transformation Tests ..' 39 3.4 Microstructural Investigation 40 3.5 Hardness Measurements 42 Chapter 4 Experimental Results: HSLA-90 Steel 43 4.1 Initial Austenite Grain Size Conditions 43 4.2 Continuous Cooling Transformation Results 44 4.2.1 Effect of Cooling Rate on Transformation Kinetics 44 4.2.2 Effept of Austenite Grain Size on Transformation Kinetics 46 4.2.3 Microstructural Evaluation for CCT Tests 48 4.2.4 Hardness Results for CCT Tests 51 4.3 Effect of Retained Strain on Austenite Decomposition 52. 4.3.1 Double-hit Compression Tests 52 4.3.2 Effect of Retained Strain and Accelerated Cooling on Transformation Kinetics 55 4.3.3 Microstructural Results for Pancaked Austenite and Accelerated Cooling 58 4.4 Effect of Applied Strain on Hardness 60 Chapter 5 Modelling of Austenite-to-Ferrite Transformation: HSLA-90 Steel 62 5.1 Modelling of Austenite-to-Ferrite Transformation Kinetics 62 5.1.1 Transformation Start Temperature 62 5.1.2 Austenite-to-Ferrite Transformation Kinetics 64 5.2 Modelling of Polygonal Ferrite Grain Size 70 Chapter 6 Experimental Results: Al-TRIP Steel 73 6.1 Austenitization Tests 73 6.2 Quench-in Tests 77 v 6.3 Continuous Cooling Transformation Results 81 6.3.1 Single Austenite Phase Initial Condition 81 6.3.2 Two Phase Initial Condition 89 Chapter 7 Conclusions and Future Work 94 7.1 Conclusions • 94 7.2 Future Work 96 References 97 vi List of Figures Figure 1.1 Schematic of thermomechanical rolling of steel strip in a hot strip mill 2 Figure 2.1 Growth sequence for the formation of a ferrite grain [16] 8 Figure 2.2 Schematic of: a) a time-temperature-transformation (TTT) diagram; b) fraction transformed as a function of time and temperature [5] 9 Figure 2.3 Fe-Fe3C phase diagram showing the equilibrium concentrations for proeutectoid ferrite formation from austenite [17] 10 Figure 2.4 Diffusion field of carbon during growth of proeutectoid ferrite 11 Figure 2.5 Heterogeneous nucleation sites in fully recrystallized austenite [19] 12 Figure 2.6 Schematic of the pillbox critical nucleus model for semicoherent nucleation of proeutectoid ferrite at austenite grain boundaries [21] 12 Figure 2.7 Thickening and lengthening of grain boundary allotriomorph [26] 13 Figure 2.8 The effect of niobium and titanium in solution on the transformation start temperature (A r3) [36] 18 Figure 2.9 Solubility products of niobium, aluminium, vanadium and titanium nitrides and carbides [39] 18 Figure 2.10 Effect of cooling rate on the transformation start temperature for an Al-TRIP steel i.e. Fe-0.2C-l.47Mn-2.18Al (wt%) [40] 20 Figure 2.11 Change in geometry of an austenite grain: a) before rolling; b) after rolling with a reduction p [52] 22 Figure 2.12 Ratio of surface area to rolling reduction p [52] 23 Figure 2.13 Continuous cooling transformation shown as a sum of short time isothermal steps [55] 26 Figure 3.1 A typical double-hit compression test deformed to a strain of 0.5 and at a strain rate of Is"1 35 Figure 3.2 Schematic thermal path used for the CCT tests 36 Figure 3.3 Schematic of CCT test set-up: (a) full jaw set; (b) set-up of specimen 37 Figure 3.4 Schematic of a typical dilation response during continuous cooling transformation 38 vii Figure 3.5 Fraction transformed data for several CCT tests repeated on the HSLA-90 steel for austenitization of 2min at 950°C and a cooling rate of 5°C/s 39 Figure 3.6 Specimen geometry for deformation-transformation tests 39 Figure 3.7 Schematic thermal path used for the deformation-transformation tests 40 Figure 4.1 Austenite microstructure after reheating to 1150oC with a holding time of 120s and water quenching 44 Figure 4.2 Effect of cooling rate on the austenite decomposition kinetics for an austenite grain size of 14pm 45 Figure 4.3 Effect of cooling rate on transformation start temperature for an austenite grain size of 14pm 46 Figure 4.4 Comparison of austenite decomposition kinetics for various austenite grain sizes at a cooling rate or approximately 60°C/s 47 Figure 4.5 Effect of austenite grain size and cooling rate on transformation start temperature 47 Figure 4.6 Microstructure for continuous cooling tests for an austenite grain size of 14um at: a) l°C/s; b) 56°C/s; c) 127°C/s; d)172°C/s 48 Figure 4.7 Microstructure for continuous cooling tests for an austenite grain size of 32pm at: a) l°C/s; b) 17°C/s; c) 130°C/s 49 Figure 4.8 Effect of cooling rate and austenite grain size on polygonal ferrite grain size. 50 Figure 4.9 Effect of cooling rate and austenite grain size on hardness 51 Figure 4.10 The dependency of hardness on transformation start temperature for various austenite grain sizes 52 Figure 4.11 Double-hit compression test for a deformation temperature of 900oC strain of 0.5 and strain rate of ls-1 53 Figure 4.12 Double-hit compression test for a deformation temperature of 850°C, strain of 0.5 and strain rate of 1 s"1 54 Figure 4.13 Comparison of softening at various interhit times, deformation temperatures and austenite grain sizes for a strain of 0.5 and strain rate of Is"1 55 Figure 4.14 Effect of strain on the transformation kinetics for an austenite grain size of 14pm and cooling rates of approximately 63-88°C/s 57 V l l l Figure 4.15 Effect of cooling rate on the transformation kinetics for an austenite grain size of 53pm and an applied strain of 0.5 57 Figure 4.16 Effect of retained strain on transformation start temperature for an austenite grain size of 53um 58 Figure 4.17 Microstructural evolution of polygonal ferrite with accumulation of retained strain with an austenite grain size of 14pm: a) e=0, CR=56°C/s; b) e=0.25, CR=63°C/s; c) e=0.5, CR=88°C/s 59 Figure 4.18 Ferrite grain refinement for an austenite grain size of 14pm with the application of accelerated cooling and presence of retained strain 60 Figure 4.19 The combined effects of cooling rate and retained strain on hardness for an austenite grain size of 14pm 61 Figure 4.20 The dependency of transformation start temperature on hardness for strain and strain-free cases for an austenite grain size of 14pm 61 Figure 5.1 Model predictions of transformation start temperature without deformation. 63 Figure 5.2 Transformation start model predictions incorporating effect of strain 64 Figure 5.3 Experimental \nb expressed as a linear function of temperature for various cooling rates, an austenite grain size of 32 pm and n = 0.85 65 Figure 5.4 Derivation of model parameters for the grain size-modified Avrami equation. 67 Figure 5.5 Application of the grain size-modified JMAK model 67 Figure 5.6 Application of the grain size-modified JMAK model incorporating the effective austenite grain size for an austenite grain size of 14pm and a cooling rate of approximately 25°C/s 68 Figure 5.7 Comparison of measured and predicted temperatures for 20% fraction transformed 69 Figure 5.8 Comparison of measured and predicted temperatures for 80% fraction transformed 69 Figure 5.9 Comparison of the measured and predicted ferrite grain sizes 71 Figure 5.10 Comparison of measured and predicted ferrite grain size 72 ix Figure 6.1 Section through the phase diagram at 1.5Mn-l .6A1 (wt.%) for the Fe-C-Mn-Al system as predicted by THERMOCALC 73 Figure 6.2 Examples of the resulting microstructures: a) 950°C; b) 1050°C; and c) 1167°C 74 Figure 6.3 Austenite grain growth of the Al-TRIP steel with a holding time of 2min 76 Figure 6.4 Transformation kinetics from measured dilation for slow cooling with the corresponding quenched-in microstructure with quench temperatures of: a) 1000°C; b) 800°C; c) 750°C; and d) 650°C 78 Figure 6.5 Effect of cooling rate on the austenite decomposition kinetics for the single austenite phase condition 82 Figure 6.6 Effect of cooling rate on transformation temperatures for the single austenite phase condition 82 Figure 6.7 Microstructures for continuous cooling tests for the single austenite phase at: a) l°C/s; b) 20°C/s and c) 160°C/s 84 Figure 6.8 Comparison of measured transformation start temperatures and estimated average martensite start temperature ranges for various cooling rates 87 Figure 6.9 Model predictions for transformation start temperature for austenite phase... 88 Figure 6.10 Effect of cooling rate on the austenite decomposition kinetics for the two phase initial condition 90 Figure 6.11 Effect of cooling rate on transformation temperatures for the two phase initial condition 90 Figure 6.12 Microstructures for continuous cooling tests for the two phase initial condition at: a) l°C/s; b) 15°C/s and c) 166°C/s 92 x List of Tables Table 2.1 Classification of Phase Transformation Mechanisms 5 Table 2.2 Values of n and m for Various Transformation Modes 28 Table 2.3 Steel Chemistries (in wt%) and A e3 Temperatures 30 Table 2.4 Model Parameters for the JMAK Model for Other Steel Chemistries 31 Table 3.1 Chemical Compositions of Steels Investigated (wt.%) 33 Table 3.2 Results for using a Tint Etchant to Distinguish Various Phases in a TRIP Steel 41 Table 4.1 Heat Treatment Schedules and Measured Austenite Grain Sizes 43 Table 4.2 Deformation-Transformation Test Schedule 55 Table 5.1 Model Parameters for the Linear Expression of \nb 66 Table 5.2 Model Parameters for Predicting the Final Ferrite Grain Size 70 Table 6.1 Ferrite Fractions from Microstructural Analysis and THERMOCALC Predictions 75 Table 6.2 Chemical Composition of Al-TRIP Steel (wt%) 76 Table 6.3 Comparison of Ferrite Fraction by Microstructural Analysis and Dilation Measurement 79 Table 6.4 Measured Austenite and Ferrite Grain Sizes 80 Table 6.5 Effect of Cooling Rate on the Phase Fractions for the Single Austenite Phase Initial Condition 85 Table 6.6 Effect of Cooling Rate on the Phase Fractions for the Two Phase Initial Condition 93 xi List of Symbols a Symbol denoting ferrite ar Thermal expansion coefficients for austenite ap Thermal expansion coefficients for product phases y Symbol denoting austenite e True strain o> Reload yield point <jm Maximum stress immediately before reloading G0 True flow stress T Isothermal incubation time <j) Cooling rate yj Equilibrium growth angle of a grain boundary allotriomorph co Parabolic rate constant for thickening S Model parameter for ferrite grain size A Temperature dependant model parameter in Avrami equation Ae3 Austenite-to-ferrite equilibrium temperature Aferrue Total area measured for ferrite phase A,otai Total area measured b Rate constant in Avrami equation xii B Temperature dependant model parameter for ferrite growth ca Equilibrium carbon concentration in ferrite cY Equilibrium carbon concentration in austenite c0 Carbon concentration in bulk material c* Limiting carbon concentration for nucleation C Parameter for ferrite grain size model parameter M Carbon concentration remaining in austenite following transformation at remaining & g l y e n c 0 0 l j n g r a t e da Ferrite grain size dY Austenite grain size dy (T) Extrapolated dilation from austenite region deff Effective austenite grain size dm (T) Measured dilation dp (T) Extrapolated dilation from product region Ada Measurement error in ferrite grain size D Diffusion coefficient of carbon in austenite D' Temperature independent model parameter for ferrite growth E Model parameter in ferrite grain size model F Polygonal ferrite fraction Fs Fraction softening xiii GL Lengthening rate m Exponential constant in grain-size modified Avrami equation Model parameter function of initial microstructure in ferrite grain size model Mp Number of preferred sites for corner nucleation of ferrite Ms Martensite transformation start temperature n Time exponent in Avrami equation n' Exponential parameter for ferrite grain size model parameter M N Number of ferrite grains per area N' Number of ferrite grains Error in number of ferrite grains per area from the ferrite grain size measurements p Rolling reduction q Exponential model parameter for ferrite grain size r'. Radius of curvature r* Radius of growing ferrite nuclei R Pre-exponential parameter for ferrite grain size model parameter M Rf Radius of spherical growing ferrite grain S Half-thickness of planar ferrite S°.b Grain surface area of austenite before rolling Sg.h Grain surface area of austenite after rolling xiv Time Holding time following initial deformation and prior to second deformation Temperature Deformation temperature Temperature of no recrystallization Nucleation temperature Transformation start temperature Model parameter for limiting carbon concentration (c*) in transformation start temperature model Model parameter for limiting carbon concentration (c*) in the transformation start temperature model Model parameter for limiting carbon concentration (c*) in the transformation start temperature model Total ferrite fraction transformed Fraction transformed Total bainite fraction transformed Temperature independent model parameter for ferrite growth xv Acknowledgements I would like to express my sincere thanks to Professor M . Militzer, who supervised me during the course of this work, for his guidance, constant encouragement and many valuable discussions. Many thanks to Fateh Fazeli and all the members of the research group for their advice, expertise and many stimulating discussions during the course of this research. Thanks are also extended to my fellow graduate students for their constant support and friendship. I would also like to thank all faculty and staff members in the Department of Materials Engineering for their assistance during the course of this work. 1 would like to especially thank Mr. Ross McLeod, Carl Ng and David Torok for preparing samples used in the experiments, this work could not have been completed without their assistance. I am also grateful to the Natural Sciences and Engineering Research Council of Canada for financial assistance and Dofasco Inc. and IPSCO Inc. for supplying the materials. And finally 1 would like to thank my family and friends who have been a source of wisdom, understanding and constant strength throughout my studies, this work could not have been completed without them. xvi i Chapter 1 Introduction 1.1 B a c k g r o u n d There is an increasing demand for high quality steel with superior properties, such as improved strength and toughness, for use in a diverse range of applications like construction components, automotive parts and in the pipeline industry. Increased competition, advances in new technologies and the continuing pressure of environmental regulations has led to a rapid introduction of novel high strength steels in the evolving steel markets. Optimization of the thermomechanical processing the steel undergoes for the purpose of controlling the resulting microstructures and thus, improving the mechanical properties, is one approach for meeting these current market demands. The main industrial processing route for producing commercial sheet steel, as a hot rolled product, is through a hot strip mill. A schematic diagram of a typical hot strip mill process is shown in Figure 1.1 which consists of four operation stages: 1) reheating; 2) rough rolling and finish rolling; 3) cooling on the run-out table; and 4) coiling. A reheat furnace is employed for initially heating up a steel slab to a temperature of approximately 1250°C, which ensures dissolution of soluble precipitates formed during casting and facilitates the subsequent deformation processes. Employing a reversing or tandem roughing mill for the multi-pass deformation process, the steel slab is reduced to a transfer bar with a thickness of 20-50mm. The final deformation process occurs in the tandem finishing mill at temperatures in the range of 1100-850°C further reducing the transfer bar to a steel strip with a thickness ranging from approximately 1-1 Omm. Exiting the finishing mill the rolled steel strip enters the run-out table where it undergoes accelerated cooling in the range of 10-150°C/s by a combination of water sprays and air cooling between the spray banks. The final processing step is coiling of the hot band on the down coiler. Chapter 1 INTRODUCTION 2 H&t strip mill Meatus S o u g h i n g Finish Ruling Cooling m m Figure 1.1 Schematic of thermomechanical rolling of steel strip in a hot strip mill [1]. The microstructural phenomena associated with each of the respective four operation stages are as follows: 1 ) austenite grain growth; 2) recrystallization and recovery; 3) austenite-to-ferrite phase transformation; and 4) precipitation. Significant austenite grain growth usually occurs during the initial reheating of the steel slab, any additional grain growth may occur in the roughing mill, where the steel slab is still at an elevated temperature, during the interpass times between the roughing rolls. The microstructural evolution of the austenite during the hot deformation processes at the roughing and finishing mills is characterized by work hardening of austenite followed by recovery and recrystallization. The deformed austenite then undergoes a phase transformation on the run-out table and decomposes into ferrite and other transformation products. Finally, subsequent precipitation of carbides, nitrides and/or carbonitrides takes place in ferrite during cooling of the coiled hot band. The last three processing steps namely, controlled rolling, cooling and coiling, significantly influence the final microstructure and thus, the mechanical properties of the hot rolled steel. Further insight in developing advanced hot rolled steels with improved mechanical properties can be achieved by examining the transformation behaviour and final microstructure produced under run-out table conditions. The most important parameters for the continuous cooling transformation that occurs on the run-out table include initial austenite grain size, retained strain, chemical composition and applied cooling rate. Chapter I INTRODUCTION 3 In recent years, there has been a particular focus on the development of process models to describe the observed microstructural evolution of hot rolled steels under run-out table conditions. These process models are required for reducing costs by optimizing the production of high quality advanced steels by providing a link between process parameters and final properties of the hot band. 1.2 Scope and Objectives Significant research had been undertaken at the Centre for Metallurgical Process Engineering at the University of British Columbia as a part of the Advanced Process Control program, to develop a process model capable of predicting the microstructural changes during the hot rolling of steel and to relate the final microstructure to the mechanical properties of the steel. The Advanced Process Control program was funded by the American Iron and Steel Institute (AIS1) and the Department of Energy (DOE) and was conducted in collaboration with a number of steel companies in the United States and Canada. These studies are now continued for advanced high strength steels (AHSS) and contribute to building fundamental metallurgical knowledge for developing novel processing routes for multiphase steels. The aim of this research is to apply models to describe the kinetics of austenite decomposition and microstructural evolution during continuous cooling under run-out table process conditions for two novel steels. The two high strength steels investigated in this study are: a high-strength low-alloy Nb/Ti microalloyed steel (HSLA-90) and an aluminium alloyed transformation-induced plasticity (Al-TRIP) steel. Initially, the kinetics and microstructure were characterized through experiments as follows: a) Experimental quantification of the effects of initial austenite grain size, cooling rate and retained strain on the austenite decomposition kinetics b) Determination of resulting microstructure including quantification of ferrite grain size, polygonal ferrite fraction and other transformation products c) Hardness measurements Chapter 1 INTRODUCTION In addition, previously developed models were applied to experimental results describe the following: a) Transformation start temperature b) Polygonal ferrite growth using an Avrami approach and adopting additivity c) Polygonal ferrite grain size 5 Chapter 2 Literature Review 2.1 Austenite Decomposition on the Run-out Table 2.1.1 Microstructures Developed from Austenite Decomposition Commercial steels are produced using heat treatment schedules where austenite is continuously cooled through the transformation temperature range on the run-out table. As a result, the final microstructure is a mixture of many transformation products formed by different transformation mechanisms. The solid state reaction of austenite decomposition can be classified into two distinct categories as listed in Table 2.1, which have different time-temperature relationships with respect to the transformation kinetics UP]. Table 2.1 Classification of Phase Transformation Mechanisms Type of Transformation Mechanism Phases Formed ~ . r c . , Nucleation and Ferrite Diffusional „ , n .. Growth Pearl ite Diffusionless Shear Martensite *Note: Bainite has been omitted since its transformation is still controversially discussed The types of ferrite formed under continuous cooling conditions are varied and complex in nature, ln addition, the type of ferrite formed is highly dependant on the cooling rate which is reflected by the morphology of the ferrite formed. Some of the different types of ferrite formed are listed below as a function of decreasing transformation temperature [3]: Polygonal Idiomorphic/Quasi-polygonal Widmanstatten Acicular Chapter 2 L I T E R A T U R E R E V I E W 6 One of the primary transformation products of a hypoeutectiod steel that is subjected to continuous cooling at slow cooling rates will be polygonal ferrite which forms by diffusional means nucleating at high energy sites such as grain boundaries and growing towards the center of the austenite grain. The growing ferrite will form disk shaped faces along both sides of the austenite grain boundaries which link up into grain boundary allotriomorphs and eventually form equiaxed polygonal ferrite grains [3][4]. Idiomorphic or quasi-polygonal ferrites on the other hand, grow with the development of semi-coherent interfaces advancing into the austenite grain using ledge motion [5]. The resulting ferrite has irregular curved boundaries or planar segments forming the quasi-polygonal grains. Widmanstatten ferrite is a plate-like or lath product which typically nucleates from austenite grain boundaries (primary Widmanstatten ferrite) or grain boundary ferrite (secondary Widmanstatten ferrite) and grows into the interior of the austenite grain for large undercoolings [6]. Two major morphologies of Widmanstatten ferrite are classified as sideplates and sawteeth which can be formed as either primary or secondary products [7]. Acicular ferrite also forms plate-like products at large undercoolings similar to those seen in Widmanstatten ferrite, however they nucleate inside the austenite grain on isolated dislocations or at small angle austenite boundaries and grow in many directions in star shaped or more complex configurations [8][9]. In the binary iron-carbon alloy, pearlite transformation occurs from slow cooling eutectoid austenite that contains 0.8 wt.%C forming a lamellar phase which is a mixture of cementite embedded in ferrite [9][10]. Pearlite nucleation frequently occurs on austenite grain boundaries however, proeutectoid ferrite and cementite can also be used as nucleation sites. The semi-spherical pearlite nuclei grow into the austenite grain by diffusion where the final morphology depends on the degree of undercooling. Bainite transformation occurs in a temperature range, i.e. ~250-600°C, below that of pearlite and is characterized classically by a non-lamellar mixture of ferrite laths and a carbide phase where the later forms from carbon enriched residual austenite usually as Chapter 2 L I T E R A T U R E REVIEW .7 cementite [11]. The morphology of bainite changes as the transformation temperature is lowered due to a higher undercooling where two types of bainite are formed: upper and lower bainite. At a higher transformation temperature, upper bainite is formed consisting of thick ferrite laths with cementite precipitates on the lath boundaries. Lower bainite formed at lower transformation temperatures, is characterized by finer ferrite laths with smaller cementite precipitates inside these laths. In the case of TRIP steels, a carbide free bainite can be formed where residual carbon enriched austenite can be present. The mechanism of bainite transformation has been an area of much debate and controversy for a number of years, continuing into the present time, from which two different view points on the mechanism of transformation have emerged [11][12][13][14]. In the first approach, Hultgern [14] proposed that the transformation proceeds by carbon diffusion controlled edgewise growth of parallel Widmanstatten ferrite plates. Conversely, Zener suggested a displacive mechanism of transformation, that was later delineated mainly by Bhadeshia, where the transformation has a shear nature in which neither iron nor substitutional elements can diffuse and resembles the martensitic transformation [11]. The bainitic ferrite grows with supersaturation with respect to carbon and partitions following growth. In this case, at sufficiently large undercoolings, when diffusional transformation of ferrite or pearlite becomes impossible, it is expected that there is a sufficient driving pressure for a diffusionless shear mode of transformation. In the diffusional approach, bainitic ferrite is assumed to grow with equilibrium carbon concentration which is in contrast to the displacive approach. Finally, the diffusionless martensite transformation occurs at very high undercoolings by the cooperative movement of atoms less than one interatomic spacing. The cooling rate must be sufficient to suppress the higher temperature diffusion controlled ferrite, pearlite and bainite transformations in order for martensite to form. The martensitic microstructure can be easily recognized by the transformation of each austenite grain into thin plates or lath structures in a similar way to Widmanstatten sideplates [9]. Chapter 2 LITERATURE REVIEW 8 The production of low carbon steels involves continuous cooling through the transformation temperature range on the run-out table resulting in a final microstructure that can be a mixture of many transformation products. Ferrite is the primary phase produced under run-out table cooling conditions where the remaining austenite is enriched by carbon rejected from ferrite to form secondary phases such as pearlite and bainite depending on the degree of undercooling. 2.1.2 Isothermal Transformation of Austenite-to-Ferrite In order to attain a better understanding of the mechanisms involved in the austenite-to-ferrite transformation an initial study of the isothermal transformation process can be useful. The proeutectoid ferrite transformation has been reviewed extensively by Reynolds et al. [8] and Aaronson [15]. The isothermal transformation of proeutectoid ferrite is a diffusionally controlled process characterized by a sequential transformation process consisting of incubation, nucleation, growth and impingement, shown in Figure 2.1. Impingement Coa lescence Figure 2.1 Growth sequence for the formation of a ferrite grain [16]. Chapter 2 L I T E R A T U R E REVIEW 9 The kinetics of an isothermal process can be conveniently represented by plotting the fraction transformed as a function of time and temperature, i.e. a TTT diagram, shown schematically in Figure 2.2 a) and can be translated into the fraction transformed over time, as shown in Figure 2.2 b). Figure 2.2 Schematic of: a) a time-temperature-transformation (TTT) diagram; b) fraction transformed as a function of time and temperature [5]. As an example, consider the Fe-FesC phase diagram for a given transformation temperature T\ with an initial carbon content c0, using the Lever rule the equilibrium carbon concentration in austenite and ferrite are determined to be cr and ca, as shown in Figure 2.3. Chapter 2 L I T E R A T U R E REVIEW 10 Ca C 0 C , Wt% Carbon Figure 2.3 Fe-Fe3C phase diagram showing the equilibrium concentrations for proeutectoid ferrite formation from austenite [17]. These equilibrium concentrations are reached by the rejection of carbon from ferrite into the remaining austenite as dictated by the solubility limit at temperature Ti from the phase diagram. For large carbon levels, carbon accumulates at the austenite:ferrite interface, as shown in Figure 2.4, and the rate at which the carbon at the interface can diffuse into the remaining austenite determines the rate at which the interface can move [5]. However, for much lower carbon levels the change in crystal structure between austenite and ferrite can proceed at a higher rate than carbon diffusion so that interface mobility plays a larger role than the diffusion process [1]. In low carbon steels, the transformation will most probably be determined by the interaction between the interface mobility and carbon diffusion so that a mixed-mode may be operational [18]. Chapter 2 L I T E R A T U R E REVIEW 11 1 cy -1 y a Distance Figure 2.4 Diffusion field of carbon during growth of proeutectoid ferrite. Nucleation Kinetics In a solid state phase transformation such as that of proeutectoid ferrite, there are two possible means for initiating nucleation: homogeneous nucleation and heterogeneous nucleation. In general, nucleation will be dominated by the formation of a critically sized nucleus that has the minimum activation barrier; heterogeneous nucleation has a lower activation energy for nucleation then homogeneous nucleation [5]. As a result, nucleation of proeutectoid ferrite generally initiates heterogeneously in austenite where the most favourable nucleation sites are grain corners, grain edges or grain boundaries/surfaces, as shown schematically in Figure 2.5 [19]. In this case, the creation of a nucleus releases some free energy by the destruction of a defect i.e. grain corners, grain edges and/or grain boundaries, which lowers the activation energy barrier. A decrease in the transformation temperature i.e. increased undercooling, activates nucleation sites from grain boundaries to grain edges to grain corners and finally other less preferable sites are activated. Chapter 2 L I T E R A T U R E REVIEW 12 A G r a i n B o u n d a r y G r a i n E d g e G r a i n C o r n e r Figure 2.5 Heterogeneous nucleation sites in fully recrystallized austenite [19]. Some of the earliest studies undertaken to understand and develop feasible models for the nucleation kinetics of proeutectoid ferrite at austenite grain boundaries have been from Cahn [20] and Russell [21]. Enomoto and Aaronson [22] have shown agreement between experimentally observed and calculated nucleation rates using classical heterogeneous nucleation theory in conjunction with a semicoherent pillbox nucleation model, shown in Figure 2.6, for the allotriomorph proeuctectiod ferrite nuclei. SEMI-COHERENT INCOHERENT / CX I T 7 7 SEMI-COHERENT Figure 2.6 Schematic of the pillbox critical nucleus model for semicoherent nucleation of proeutectoid ferrite at austenite grain boundaries [21]. This model is associated with a maximum driving pressure for nucleation and minimum activation barrier. Enomoto and Aaronson [23][24], extended this model by incorporating critical nucleation at grain edges, which have low energy interfaces, by assuming a critical nucleus shape of an equilateral trigonal prism. This nucleation model by Enomoto and Aaronson was extended to conditions of continuous cooling by Militzer et al. [25] where they proposed a model approach to describe transformation start. In their analysis Chapter 2 LITERATURE REVIEW 13 they found that nucleation temperature is relatively independent of cooling rate and that early growth of the nuclei has to be taken into account to describe the transformation start. Growth of Ferrite In a typical solid state phase transformation the nucleated phase will grow at the expense of the old phase by the migration of the interphase boundary, which is driven by difference in the free energies of the two phases. Growth of ferrite proceeds with carbon diffusing away from the advancing ferrite and enriching the austenite with carbon. The growth kinetics of a grain boundary allotriomorph can be separated into thickening kinetics and lengthening kinetics, as shown schematically in Figure 2.7. Lengthening Figure 2.7 Thickening and lengthening of grain boundary allotriomorph [26]. For thickening, the low mobility semicoherent plate-like interfaces migrate using a ledge mechanism and the incoherent interfaces are highly mobile [5]. If the interface of a grain boundary allotriomorph is considered to be incoherent than a diffusion model can be applied to predict a parabolic thickening rate as follows [27], S = ootU2 - (2.1) where S is the half thickness, co is a parabolic thickening rate constant and t is growth time. If a linear carbon gradient in the remaining austenite is assumed, the parabolic rate constant can be expressed as [17], Chapter 2 L I T E R A T U R E REVIEW 14 where D is the diffusivity of carbon in austenite, c 0 is the average carbon bulk concentration, cy is the equilibrium carbon concentration in austenite and ca is the equilibrium carbon concentration in ferrite. This expression assumes that: 1) the interphase boundary is planar and disordered; 2) the kinetics are volume diffusion controlled; and 3) the diffusivity of carbon in austenite is composition invariant. Lengthening kinetics are also assumed to be diffusion controlled thus migration of the planar interface is carbon controlled, in addition, a ledge mechanism is activated for plate-like interfaces [5]. Simonen et al. [28] have observed that lengthening occurs by the ledge mechanism. It has been shown that if the side plates are considered to be curved the resulting grain boundary allotriomorph lengthens at a constant rate Gt, as follows [4], G, = D (cr-c0) 4r'(c o - c j sin 0 (2.3) where r' is the radius of curvature and t// is the equilibrium growth angle as shown in Figure 2.7. The shape that develops during growth is determined by the relative migration rates of the interfaces. It has been shown that lengthening occurs much faster then thickening and that the average aspect ratio between the half thickness and half length is 1/3 [29][30]. As a result, a thin layer of allotriomorph ferrite outlines the austenite grain boundary and growth of the ferrite will occur into the volume of the grain as the cooling rate is increased. Impingement Impingement of the growing ferrite grains can be classified into two distinct groups: soft impingement and hard impingement [1]. Soft impingement occurs at first when ferrite continues to grow by diffusion of carbon into the remaining austenite until diffusional fields from separately nucleated ferrite begin to overlap and interfere with Chapter 2 L I T E R A T U R E REVIEW 15 each other's growth. Eventually, the growing ferrite grains begin to impinge on each other which prohibits any further growth in that direction i.e. hard impingement. As a result, the growth rate decreases to zero as the process of impingement of the transforming regions progresses. 2.1.3 Isothermal Transformation Kinetics: J M A K Approach For isothermal conditions, temperature, which is one of the most fundamental physical parameters, is kept constant in order to study the transformation behaviour of the steel. Kolmogorov [31], Johnson and Mehl [32] and Avrami [33][34]conducted some of the earliest research in the development of a semi-empirical equation to describe the isothermal kinetics of diffusional phase transformation kinetics. The general form is known as the Johnson-Mehl-Avrami-Kolmogorov (JMAK) or Avrami equation and can be given as follows, * = l-expH>/") • (2.4) where X is the fraction transformed, n is dependant on the nucleation conditions and the shape of the growing particle and b is a temperature dependant variable related to the magnitude of the nucleation and growth rate. The kinetics of transformation is an interplay between nucleation and growth kinetics of the new phase, as shown in Figure 2.2 b) by the fraction transformed over time. 2.2 Factors Influencing Austenite Decomposition Behaviour 2.2.1 Chemistry The addition of alloying elements has been found to overcome the deficiencies of plain carbon steels and has resulted in improved material properties of steel. The thermodynamic stability of phases is altered with the addition of alloying elements to pure iron, which leads to a wide variety of microstructures and mechanical properties obtained as a result of austenite decomposition. For HSLA steels, the base chemistry consists of carbon and manganese where the principle microalloying elements are Chapter 2 L I T E R A T U R E REVIEW 16 niobium, vanadium and titanium. The base chemistry for a TRIP steel is also carbon and manganese but contains additional alloying elements such as silicon and/or aluminium. In the presence of substitutional alloying elements, three different modes of transformation are possible, described as follows [9], 1. Orthoequilibrium at the austenite:ferrite interface with growth controlled by diffusion of the alloying element which occurs during very slow cooling. The austenite:ferrite interface migrates with the alloy fully partitioning between the austenite and ferrite phases; this process may take several hours to reach equilibrium. 2. Paraequilibrium at the austenite:ferrite interface with growth controlled by volume diffusion of carbon. No partitioning occurs i.e. the substitutional sublattice is configurationally frozen; only the chemical potentials of carbon are in equilibrium across the boundary. 3. No partition local equilibrium (NPLE) occurs at high cooling rates where the austenite:ferrite interface may still migrate with an equilibrium interface, however slow partitioning of the alloy element causes the redistribution to only occur locally at the interface resulting in a solute spike. Growth is still controlled by volume diffusion of carbon [29]. For run-out table cooling, paraequilibrium and NPLE are the typical modes of transformation. As a result, the rate for diffusionally controlled transformations is retarded by most substitutional alloy elements due to: 1) the thermodynamic effect of alloying elements; 2) ternary diffusion interactions and 3) solute drag effect at the austenite:ferrite interface due to the slow diffusing alloying elements [31]. Carbon is an efficient austenite stabilizer and, in general, retards the transformation kinetics by shifting the TTT curves to increasingly longer times as the Chapter 2 L I T E R A T U R E REVIEW 17 carbon content is increased [9]. As a result, non-equilibrium transformation products such as bainite and martensite can be produced. The driving pressure for austenite decomposition at any temperature is reduced by increasing the carbon content due to a lowering of the A e 3 temperature. In general, low carbon steels are comprised of carbon contents up to 0.25 wt%. The addition of manganese to low carbon steel produces several important changes. Like carbon, manganese acts as an austenite stabilizer and effectively expands the temperature range where stable austenite can form. The A e 3 temperature is substantially lowered which can enhance ferrite grain refinement and by increasing the manganese content a transition from a polygonal ferrite-pearlite microstructure to a ferrite-bainite microstructure can be attained [19]. Niobium and titanium have been found to significantly influence the austenite decomposition kinetics [36][37][38]. An example of the effect of niobium and titanium in solution on the transformation start temperature is shown in Figure 2.8, which indicates that both niobium and titanium have a strong effect on delaying the start of the proeutectoid ferrite transformation, and that niobium is more effective for this delay than titanium. In addition, niobium and titanium have different affinities for carbon and nitrogen in austenite as shown in Figure 2.9, so that precipitates of carbides, nitrides and carbonitrides are formed for both niobium and titanium. Coarse grain boundary precipitates have been found to be potent nucleation sites for ferrite formation due to the large mismatch with austenite which provides a high energy interface suitable for nucleation [37]. In addition, the formation of these coarse precipitates locally reduces the niobium and carbon contents in solution resulting in the effect of solute drag to be diminished and further promoting ferrite nucleation. Chapter 2 L I T E R A T U R E REVIEW 18 800 £ 750| a a £ 700| to 0-5'C/s (Do=100um) 0 0 02 004 0 06 0 08 soluble atom (wt °/o) Figure 2.8 The effect of niobium and titanium in solution on the transformation start temperature (A r3) [36]. 2000 r 10.0 1300 1100 900 800 T E M P E R A T U R E , * C . Figure 2.9 Solubility products of niobium, aluminium, vanadium and titanium nitrides and carbides [39]. Chapter 2 L I T E R A T U R E REVIEW 19 Both silicon and aluminium used in TRIP steels have been found to be ferrite stabilizers and effectively expand the temperature range where stable ferrite can form. Silicon is the conventional alloying element in TRIP steels however, high silicon levels are required and are not well suited for the galvanization process route in industry. As a result, aluminium is used to either fully or partially replace silicon content due to the improvement in coatability of the steel during the galvanization process. In addition, replacing the silicon content in TRIP steels by aluminium enhances the formation of polygonal ferrite in the final microstructure [40]. 2.2.2 Cooling Rate Cooling rate is one of the key parameters which, can be adjusted on a run-out table due to its strong effect on the kinetics of austenite decomposition and resulting microstructure [41]. A wide variety of transformation products i.e. polygonal ferrite, acicular ferrite, bainite and/or martensite, can be obtained in H S L A and TRIP steels by changing the cooling conditions during the continuous cooling phase transformation. Austenite decomposition is a thermally activated process where the formation of the new product phase requires time to initiate nucleation and to continue into the growth stage. As a result, by increasing the cooling rate the available time at any given temperature to start the nucleation and growth processes is decreased, and thus, shifts the transformation start to lower temperatures [25][41] [42][43]. Increasing the cooling rate also produces a finer ferrite grain size since there is a greater difference in free energies of the austenite and growing ferrite which leads to the activation of more potential nucleation sites [5]. At very high cooling rates the transformation rate decreases due to a reduction in the diffusivity of carbon. In addition to producing a finer ferrite grain size, an increase in the cooling rate can also form low temperature secondary products such as bainite and martensite due to the suppression of the transformation start temperature. A n example of this can be seen from a recent study by Manohar et al. [40] on the continuous cooling transformation behaviour of an aluminium TRIP steel i.e. Fe-0.2C-l.47Mn-2.18Al (wt%). The effect of Chapter 2 L I T E R A T U R E REVIEW 20 cooling rate on the suppression of the transformation start temperature can be clearly seen in Figure 2.10. Cooling Rate (K s ') Figure 2.10 Effect of cooling rate on the transformation start temperature for an Al-TRIP steel i.e. Fe-0.2C-l.47Mn-2.18Al (wt%) [40]. By controlling the final microstructure through cooling the • desired - final mechanical properties of the steel can be achieved. In this way, accelerated cooling on the run-out table is an important thermal treatment which can effectively control the final microstructure and mechanical properties of the steel. 2.2.3 Initial Austenite Microstructure The initial austenite microstructure plays an important role in the transformation behaviour for run-out table cooling. The two microstructural features of austenite which are important for the transformation process include the austenite grain size and the degree of pancaking. 2.2.3.1 Austenite Grain Size The heterogeneous nucleation of ferrite occurs predominately on austenite grain boundaries therefore, a decrease of the austenite grain size leads to an increase in the grain boundary area per unit volume i.e. there now exists a greater surface area for potential nucleation sites. As a result of increasing the number of potential nucleation Chapter 2 L I T E R A T U R E REVIEW 21 sites, transformation starts at higher temperatures and produces higher temperature transformation products such as polygonal ferrite [41][44]. A finer ferrite grain size is produced since the number of ferrite formed increases due to more available potential nucleation sites where impingement of the growing ferrite grains will occur earlier due to the decrease of the austenite grain size. It is observed that with a decrease in the austenite grain size, the transformation rate increases due to the increase in the ratio of nucleation rate to growth rate [45] [46]. In addition, formation of polygonal ferrite is depressed with an increase in the austenite grain size and formation of non-polygonal microstructures is promoted [41]. 2.2.3.2 Effect of Deformation A deformed austenite microstructure is characterized by elongated austenite grains i.e. pancaked grains, formed by deforming at a temperature below the recrystallization stop temperature, Tnr. Numerous studies have been undertaken [47][48][49] which show that the austenite decomposition is strongly affected by deformation and that the retained strain in austenite contributes to the refinement of ferrite. The increase in the ferrite nucleation rate per unit volume of austenite due to the pancaked structure can be accounted for by: 1) the formation of additional nucleation sites; 2) an increase in the nucleation rate per austenite grain surface area due to surface irregularities; and 3) an increase in the austenite grain surface by elongation of the grains. In an undeformed austenite grain the predominate nucleation sites for ferrite are at the austenite grain boundaries however, by deforming the austenite grain nucleation can occur not only on the austenite grain boundaries but also within the grain. The types of additional nucleation sites observed for a deformed austenite include: annealing twins, dislocations, sub-grains and deformation bands [50]. Kozasu etal. [51] represented the overall interfacial area of austenite grain boundaries and deformation bands available for nucleation in an effective austenite interfacial area. As a result of the increase in the density of available nucleation sites the transformation may occur at higher temperatures. Chapter 2 L I T E R A T U R E REVIEW 22 In addition to increasing the available nucleation sites, deformation also alters the originally smooth austenite grain boundaries into irregular curving boundaries by a non-uniform slip mechanism occurring inside the grain. These irregular austenite grain boundaries are associated with high energies and can act as additional potential nucleation sites [19]. If it is assumed that the initial austenite grain has a spherical geometry with a normalized unit radius before the deformation process, as shown in Figure 2.11, then after deformation the geometry changes to an ellipsoid. Figure 2.11 Change in geometry of an austenite grain: a) before rolling; b) after rolling with a reduction p [52]. The additional austenite grain boundary surface area formed due to deformation can be calculated where a grain with unit radius will have a grain boundary surface area as given by, After deformation with a reduction of p the grain boundary surface area can be given as follows [52], (a) (2.5) - p 2 sin 2 8 -dO x \-x\\~p) x2(\-p)6 (2.6) V J Chapter 2 L I T E R A T U R E REVIEW 2 3 Figure 2.12 is a plot of the ratio of the surface area before and after deformation i.e. q - S h IS"h, as a function of p which indicates that by applying a large reduction of 0.7 the austenite grain boundary surface area can be increased twofold. strain E (=-ln(1-p)) Rotting Reduction (p) Figure 2.12 Ratio of surface area to rolling reduction p [52]. Lacroix et al. [53] suggested a simple model which combines the effect of strain and austenite grain size on the austenite decomposition by employing the ellipsoid geometry shown in Figure 2.11b). Further it was suggested that a thin layer of ferrite instantly covers all austenite grain boundaries which grow in the direction of the smaller axis. Consequently, the ferrite growth rate is determined by the growth in this axial direction, represented by 1-p, and can be captured by employing an effective austenite grain size defined as follows [53], deff =drexp(-e ) = dr(\-p) (2.7) where dr is the volumetric austenite grain size and s is the applied strain. Chapter 2 L I T E R A T U R E REVIEW 24 2.3 Modelling of Microstructural Evolution under Run-Out Table Conditions Predicting the microstructural evolution under continuous cooling conditions as found on the run-out table is a challenging task since in addition to cooling conditions, the austenite decomposition is affected by chemistry, austenite grain size and retained strain. A process model for the austenite-to-ferrite transformation during continuous cooling is presented, which consists of three parts: 1) prediction of the transformation start temperature; 2) ferrite growth model; and 3) polygonal ferrite grain size prediction. 2.3.1 Austenite-to-Ferrite Transformation Kinetics Modelling 2.3.1.1 Transformation Start Temperature The transformation start temperature can be predicted with a model that combines corner nucleation of ferrite with early growth, which was originally proposed for plain carbon steels [25]. The model assumes that early.growth of corner ferrite nucleated at TN is controlled by the carbon diffusion in austenite where TN can be determined at a slow cooling and has been found by Militzer et al. [25] to be independent of cooling rate. The radius of the growing spherical ferrite grain, Rf, can be calculated by, dRf dT C v _ Cn 1 -^°LL = DJL <L_L (2.8) dT dt cy - ca Rf assuming steady-state growth conditions along the grain boundaries where D is the diffusivity of carbon in austenite, c0 is the average carbon concentration and c^and ca are the equilibrium carbon concentrations in austenite and ferrite, respectively. Integrating Equation 2.8 for continuous cooling, T= TN -<j> t, yields: (2.9) where <fi is a constant cooling rate. Chapter 2 L I T E R A T U R E REVIEW 25 Nucleation occurs preferentially at grain corners for low cooling rates where the number of preferred sites for corner nucleation, Mp, was estimated to be approximately two per grain [25]. Growth of early corner nuclei determines the quantity of remaining available austenite boundary area for additional nucleation at higher cooling rates for the continuous cooling transformation. As the temperature decreases during continuous cooling, nucleation sites are shifted from preferred grain corners to edges and grain boundaries. Eventually, the entire grain boundary is covered by a layer of the ferrite nuclei, i.e. site saturation, at which time nucleation ceases and measurable transformation starts. Ferrite nucleation cannot take place at those boundary areas which are already covered by ferrite or in the vicinity of the growing ferrite grain, where the increase in the carbon concentration is associated with a reduction in the driving pressure for nucleation. Thus, nucleation rate becomes a function of the distance from the growing ferrite [25]. A limiting carbon concentration, c*, can be introduced so that nucleation can only occur in those areas where the carbon concentration remains below c*. By employing c*, a nucleation limiting radius around the growing ferrite can be defined as a function of the ferrite radius as follows, c — c C -c (2.10) When Mp(r*)2 - d1', further nucleation becomes impossible and for a constant cooling rate, (j), the transformation start temperature, Ts, can be estimated from [25], 2 ( c - c n ) r * - r - 7 0 L Co ,1/2 • N DCy C° dT c r ~ c a (2.11) which relates the effect of cooling rate and austenite grain size to the transformation start temperature. Chapter 2 L I T E R A T U R E REVIEW 26 2.3.1.2 Application of Isothermal Transformation Kinetics to Non-Isothermal Conditions Continuous cooling transformations are widely used in industrial processing and it was recognized that isothermal transformation information could be used to investigate the continuous cooling transformation behaviour of steel. Scheil [54] proposed a method for describing nucleation based on the incubation time associated with the isothermal transformation. It was assumed that for nucleation, an isothermal incubation time, T(T), at temperature T is required so that for a time step dt at each temperature a fraction of incubation time equal to dt/t(T) will be consumed. Thus, for continuous cooling, austenite uses a fraction of incubation time and when this sum of fractions equals unity incubation is complete and transformation can begin, i.e., Modelling the transformation kinetics for isothermal conditions can be extended to non-isothermal conditions based on the additivity principle proposed by Cahn [20]. The additivity principle involves the sum of short time isothermal holdings at successive temperatures, as schematically shown in Figure 2.13 [55]. Figure 2.13 Continuous cooling transformation shown as a sum of short time isothermal steps [55]. (2.12) A ti A 1; A t j Chapter 2 L I T E R A T U R E REVIEW 27 The additivity principle proposed by Cahn [20] for isokinetic, i.e. when nucleation rate is proportional to growth rate, reactions states that a reaction is additive whenever the transformation rate is a function only of fraction transformed, X, and temperature, T, as follows, ^L = f(X,T) (2.13) dt Further, Christian [1] suggested that a transformation will only be additive if the transformation rate can be expressed independently as two functions, as follows, dt G(X) where H(T) is a function only of temperature and G(X) is a function only of fraction transformed. Lusk and Jou [56] confirmed that the solution given by Christian is required for a reaction to be additive. 2.3.1.3 Ferrite Growth Kinetics: J M A K Approach with Additivity As a result of nucleation usually occurring heterogeneously at grain boundaries, edges and corners, the austenite grain size has a direct effect on the transformation kinetics, as previously discussed. An Avrami equation can be adapted to employ the principle of additivity by using Equation 2.4 and taking the derivative with respect to time, as follows, f C2-.5) { 1 1 ( 1 \ [ \ - x ) l l n ( l - * ) J Chapter 2 L I T E R A T U R E REVIEW 28 where the criterion for additivity is satisfied if b=f(T) and n=constant. Consequently, the Avrami equation was modified by Umemoto et al. [58] to incorporate the austenite grain size into the following form, X = 1 - exp d (2.16) r J where m is a model parameter dependant on the growth geometry, b and n are the same as previously defined. For run-out table cooling conditions, it has been proposed that b can be expressed as a function of undercooling below the A e 3 temperature as follows [42], b = exp(b1{TAe3-T) + b2) (2.17) where 6/,and bj are constants which depend on the steel chemistry. The values of n and m and the transformation modes were experimentally determined from isothermal transformations and are summarized in Table 2.2. For ferrite formation in low carbon steels typical values of n have been reported in the range of 0.8-1.2 [41][42][44][59][60]. Table 2.2 Values of n and m for Various Transformation Modes Transformation n m Nucleation Pearlite 4 2 Edge nucleation Nucleation and Growth Ferrite 1 1 Surface nucleation near site saturation Bainite 4 0.6 Grain boundary, intergranular, nucleation and growth The parameter b, which is a function of temperature, can be expressed using any number of functions from experimental results. Pandi [41] used a linear relation to express Inb as a function of temperature where two model parameters adequately described the transformation rate, whereas Nakata and Militzer [44] employed two separate functions of Inb to characterize the early and later stages of the transformation Chapter 2 L I T E R A T U R E REVIEW V) for an HSLA steel. Further, Hawbolt et al. [60] employed a polynomial function for Inb to characterize the transformation rates in their studies. Numerous studies [25][59][60] have been conducted with the successful adoption of the principle of additivity to the austenite-to-ferrite transformation, among the earliest researchers include Hollomon et al. [61] and Manning et al. [62]. Leblond and Devaux [63] extended the approach in terms of formation of two or more phases as products of austenite decomposition. Further, Suehiro et al. [64] and Umemoto et al. [55] extended their overall transformation models to continuous cooling by employing the additivity principle. 2.3.2 Polygonal Ferrite Grain Size Model Prediction of the ferrite grain size offers valuable information regarding the material properties of the steel, as a result, they are commonly integrated into process models. Ferrite grain size depends strongly on the start of transformation for conditions of site saturation where no additional nucleation takes place below the transformation start temperature. Studies have shown that no significant ferrite grain growth occurs after the completion of transformation [41][65]. Accordingly, the resulting ferrite grain size is related to the additional nucleation at grain edges and grain boundaries taking place during the early stages in the continuous cooling transformation. . The ferrite grain size can be expressed as a function of the transformation start temperature in the form suggested by Suehiro et al. [64], as follows, • ( , -A da = Fexp M T _ where da is the ferrite grain size quantified as an equivalent area diameter (EQAD) in pm, Ts is the transformation start temperature measured in K, F is the final polygonal ferrite fraction, measured by optical metallography, E is a fitting parameter and M is a Chapter 2 L I T E R A T U R E REVIEW 30 function of the initial austenite microstructure. In the absence of retained strain, Mean be expressed as a function of austenite grain size as follows, M=Zdvy (2.19) where H and q are model parameters specific to the steel [42]. 2.3.3 Comparison of Model Parameters for Previously Investigated Steels The previously described process model for the austenite-to-ferrite transformation during continuous cooling has been applied to a variety of low carbon steels investigated by other researchers. However, limited results are available for advanced high strength steels. The steel chemistries and associated A e 3 temperatures for some H S L A steels and a dual phase steel are summarized in Table 2.3. Table 2.3 Steel Chemistries (in wt%) and A e 3 Temperatures Steel Grade A e 3 CC) C Mn A l N Si V Mo Nb Ti S P HSLA-50 [42] 857 0.07 0.76 0.053 0.0067 0.014 - - 0.023 0.013 - -HSLA-65 [66] 840 0.062 1.24 0.040 0.0070 0.051 0.001 0.008 0.063 0.002 0.004 0.007 HSLA-80 [42] 843 0.07 1.35 0.044 0.0070 0.140 - - 0.086 0.047 - -Dual Phase 822 0.060 1.86 0.043 0.0070 0.077 - 0.155 - 0.011 0.004 0.015 [66] The transformation start temperature was modelled using the model parameter c* and TN in reference to Equation 2.11. However, for the HSLA-50 steel a new equation was formulated to improve the accuracy for predicting the. transformation start temperature where c* was represented as follows [42], Chapter 2 L I T E R A T U R E REVIEW c* = (x* +xrldr + Axexp(-0.0003(7; -Tf 2 ) )c 0 31 (2.20) where x*, xr and Ax are model parameters. The model parameters for the transformation start temperature in addition to the ferrite growth kinetics and ferrite grain size for these steels are summarized in Table 2.4. Table 2.4 Model Parameters for the J M A K Model for Other Steel Chemistries Steel Grade T N C O X* xr Ax m b, b 2 q n HSLA-50 [42] 800 1.23 0.85 0.15 1.3 0.026 -0.44 50.7 0.037 0.9 HSLA-65 [66] 819 1.8 0 0 2.9 0.018 20.8 30.2 0.056 0.9 HSLA-80 785 2 0 0 1.3 0.035 -3.6 * * 0.9 [42]** Dual Phase 732 1.3 0 0 1.3 0.020 15.4 45.2 0.019 0.9 [66] * no grain size dependence recorded. ** control rolled with s > 0.6. It is evident from Table 2.4 that a range of values can be suggested for each model parameter for which the process model can be applied. 2.3.4 Other Models to Describe the Overall Austenite Decomposition There are a number of alternative models which have been developed to describe the overall austenite decomposition kinetics, where some are extensions of the present J M A K model. Some examples include, a model by Umemoto et al. [55] that incorporates a criteria for the transition between sequential formation of different transformation products, as well as the transformation from work-hardened austenite of which the latter is an important feature of the microstructure as it enters the run-out table. Jones and Bhadeshia [6] have successfully adopted the J M A K model for the overall transformation Chapter 2 L I T E R A T U R E REVIEW 32 kinetics to deal with the simultaneous formation of allotriomorphic ferrite, Widmanstatten ferrite and pearlite. 33 Chapter 3 Experimental Methods 3.1 Materials Experiments were carried out on a high-strength low-alloy Nb/Ti steel (HSLA-90) and an aluminium alloyed transformation-induced plasticity steel (Al-TRIP) with the chemical compositions (wt.%) given in Table 3.1. IPSCO Inc. supplied the HSLA-90 steel as a hot rolled transfer bar (an end product of the roughing mill operation) and the AI-TRIP steel was supplied by Dofasco Inc. as a lab cast forged bar. Table 3.1 Chemical Compositions of Steels Investigated (wt.%) Grade C Mn Al N Si Cu Mo Ni Nb Ti S P H S L A -90 0.05 1.65 0.027 - 0.025 0.29 0.196 0.16 0.071 0.021 0.004 0.01 A l -TRIP 0.19 1.42 1.68 0.0032 - - - - - - - -3.2 Experimental Equipment 3.2.1 Gleeble 3500 Thermomechanical Simulator A Gleeble 3500 thermomechanical simulator was used to perform the reheating and deformation tests in order to establish suitable austenite microstructures for subsequent continuous cooling of the specimen to determine the kinetics of austenite decomposition. Temperature control of the specimen during thermal cycling was achieved by spot welding a Chromel-Alumel or Platinum-Platinum 10%Rhodium thermocouple onto the center surface of the specimen. Temperature feedback control and direct resistive heating of the system provided accurate heating and cooling of the specimen. Controlled cooling was achieved by employing resistive heating in combination with Helium gas cooling. The chamber was evacuated to a pressure below Chapter 3 E X P E R I M E N T A L METHODS 34 3mTorr for each test in order to minimize decarburization and oxidation of the specimen. Feedback control of thermal and mechanical systems in addition to data acquisition was performed by the computer interface software Quick-Sim. 3.3 Experimental Methodology 3.3.1 Austenite Grain Sizes Austenite grain growth was studied with reheating tests in order to establish suitable austenite grain sizes which reflected grain sizes at the exit of the finishing mill. A rectangular specimen geometry with a length of 15mm, width of 6mm and 3mm thickness was used for the HSLA-90 steel. For the Al-TRIP steel, a tubular geometry was employed which was 20mm in length with an 8mm outer diameter and a 1mm wall thickness. Specimens were heated at 5°C/s to various austenitizing temperatures and held for a specified time to obtain the desired austenite grain sizes. Specimens were subsequently quenched, using water or Helium gas, to room temperature in order to obtain a martensitic microstructure. 3.3.2 Double-Hit Compression Tests Double-hit compression tests were conducted in order to establish conditions for a pancaked austenite microstructure where sufficiently slow recrystallization allows the applied strain to be retained. Solid cylindrical specimens, with a gauge length of 15mm and a diameter of 10mm, were austenitized to obtain the desired initial austenite grain size and cooled to a deformation temperature above the A e 3 temperature. After holding at this temperature for a given time ranging from 15-60s the specimen was deformed to a prescribed strain, unloaded immediately and after a specified time (interhit time) deformed again using the same strain rate. Different strain conditions and a constant strain rate were selected in order to reflect industrial process conditions. The interhit time, deformation temperature and deformation holding time was established for. conditions where the degree of static softening was minimized i.e no recrystallization conditions. Chapter 3 E X P E R I M E N T A L M E T H O D S 3 5 The fraction of softening, Fs, which takes place during the unloading of the specimen can be estimated as follows, (3.1) where o0 and ar are the yield points related to the first and second deformations, respectively, and crm is the flow stress immediately before unloading occurs, as shown in Figure 3.1. Yield points were determined using the 0.2% offset method, i.e. s = 0.002. In this method, the elastic modulus, E, was determined from the linear elastic portion of the stress-strain curve, i.e. E is the slope of the linear portion of the stress-strain curve. Using this slope a parallel line to the linear portion of the stress-strain curve at a = Oand s = 0.002 was drawn. The stress at which this line intersects with the stress-strain curve is called the yield stress. Temperature, force, stroke and mid-length dilation were recorded as a function of time for each specimen undergoing testing using a C-strain gauge placed at mid-plane for the dilation measurements. Experimental data obtained from each double-hit compression test was converted to true stress and true strain, an example of which is shown in Figure 3.1. CD 300 250 200 w _ 150 w S 100 50 0 Cm ¥ <y0 T d e f = 850°C *interhit~ 45seC dy=14u.m 0.0 0.2 0.6 0.8 0.4 True Strain Figure 3.1 A typical double-hit compression test deformed to a strain of 0.5 and at a strain rate of Is"1. Chapter 3 E X P E R I M E N T A L METHODS 36 3.3.3 Controlled Cooling Tests Continuous cooling transformation (CCT) tests were conducted to dilatometrically quantify the austenite decomposition kinetics as a function of initial austenite microstructure and cooling rate. Tubular specimens, 20mm in length with an 8mm outer diameter and a 1mm wall thickness, were austenized to the established conditions to obtain the desired austenite grain size. Following austenitization, the specimens were immediately cooled by employing three types of cooling procedures: resistive heating (< 15°C/s), air cooling (~ 20°C/s) and resistive heating in combination with Helium gas cooling (> 30°C/s) . Investigated cooling rates ranged from l°C/s to 179°C/s. The cooling rate was calculated by taking an average of + 2 0 ° C at the A e 3 temperature for the HSLA-90 steel and over the range of transformation for the Al-TRIP steel. Figure 3.2 shows a schematic thermal path used for the C C T tests. Temperature and mid-length dilation was recorded as a function of time for each specimen undergoing testing using a crosswise dilatometer. The schematic diagram of a specimen held in place by tubular stainless steel holders and copper anvils inside stainless steel jaws is shown in Figure 3.3. The specimen was allowed to expand during heating without deformation by placing a spring between the jaws and load cell. Hold at T„ t, • Time Figure 3.2 Schematic thermal path used for the C C T tests. Chapter 3 E X P E R I M E N T A L METHODS 37 (a) Copper Anvils Stainless Steel Jaws n (\ i ' "*-•> Specimen i • \ / Specimen Holder U Spring < Load Cel l Thermocouples Hel ium Out Measurement Figure 3.3 Schematic of CCT test set-up: (a) full jaw set; (b) set-up of specimen. When austenite decomposes during cooling there is a measurable increase in atomic volume from austenite (FCC crystal structure) to ferrite (BCC crystal structure) and other product phases. This is a result of the different crystal structures for example, the FCC crystal structure is close packed i.e. high density, whereas a BCC crystal structure has a lower density. Volume fraction transformed, X(T), was calculated using the dilation measurements, dm(T), and employing the Lever Rule as follows, X(T) = dm(T)-dy(T) dp(T)-dr(T) (3.2) where dy(T) = Iy +ayT dp(T) = Ip+apT Chapter 3 E X P E R I M E N T A L METHODS 38 are the extrapolated dilations from the austenite and product phase regions, and ar and ap are the thermal expansion coefficients for the austenite and product phases, respectively. Figure 3.4 shows a schematic of a dilation response as a function of temperature, subsequent austenite-to-ferrite transformation kinetic curves can be derived from this data. VP -2 5 m / Di 9 7 Temperature T Figure 3.4 Schematic of a typical dilation response during continuous cooling transformation. In addition, a series of quench-in tests were conducted for the Al-TRIP steel at a slow cooling rate of l°C/s in order to validate the measured dilation data using an analysis of the quenched-in microstructure. Figure 3.5 shows the fraction transformed for several repeated CCT tests where specimens were austenitized at 950°C for 2min and cooled at 5°C/s using the HSLA-90 steel. An error of +4°C was estimated for the transformation start temperature and an error of + 8°C was estimated for 50% fraction transformed from these measurements. Chapter 3 E X P E R I M E N T A L METHODS 39 1.0 o.o -\ . . . . . . .—:—I 560 580 600 620 640 660 680 700 720 Temperature, °C Figure 3.5 Fraction transformed data for several C C T tests repeated on the HSLA-90 steel for austenitization of 2min at 950°C and a cooling rate of 5°C/s. 3.3.4 Deformation-Transformation Tests The influence of retained strain i.e. deformation under no recrystallization conditions, and accelerated cooling on the austenite decomposition behaviour and the resulting microstructure was investigated using deformation-transformation tests. Specimens were 120mm in length, with a diameter of 10mm; the gauge length of the working zone was 10mm in length with a 6mm diameter, as shown in Figure 3.6. V - . . . . 4 •—• 10 mm 120 mm Figure 3.6 Specimen geometry for deformation-transformation tests Chapter 3 E X P E R I M E N T A L METHODS 40 The specimens were austenitized at the established conditions, cooled to the deformation temperature and held for a specified time followed by deformation at the prescribed strains and constant strain rate. Immediately after deformation the specimens were cooled with cooling rates ranging from l°C/s to 60°C/s (air cooling). Figure 3.7 shows a schematic thermal path used for the deformation-transformation tests. Helium gas cooling, to obtain higher cooling rates, was not employed due to the introduction of thermal gradients at the cross section of the sample. Cooling rates were calculated at +20°C of the A e 3 temperature. Hold at Ti, t, Time Figure 3.7 Schematic thermal path used for the deformation-transformation tests 3.4 Microstructural Investigation Specimens to be used for metallographic examination were cut at the thermocouple position using a silicon carbide cut-off wheel and spray coolant, to protect the specimens from overheating during cutting. The surfaces of interest were cold mounted in either an acrylic resin or a polymer resin and then progressively ground using silicon carbide papers to a 600 grit finish followed by final polishing employing 6pm and lpm diamond solutions. Prior austenite grain boundaries for the HSLA-90 steel were revealed using an etchant with the following chemistry, Chapter 3 E X P E R I M E N T A L METHODS 41 2 g picric acid 1 mL hydrochloric acid 1 g dodecylbenzenesulfonic acid 100 mL distilled water Etching was accomplished by heating the etchant to 60°C and immersing the specimens in the solution for 3-4min. Ferrite and other transformation products from the C C T and deformation-transformation specimens for the HSLA-90 steel were revealed by tempering at 200°C for 2hrs and etching using a l%Nital solution for l-2min. For the Al-TRIP steel, it was sufficient to use a 2% Nital solution for 30-60s to reveal the prior austenite grain boundaries on account of the austenite grain boundaries being outlined by grain boundary ferrite. In addition, a 2% Nital solution for 30-60s was employed for the quench-in tests for purposes of quantifying the ferrite fraction. In order to reveal the variety of microstructures formed for the C C T tests, a tint etchant consisting of 10% Sodium metabisulfite in distilled water was employed, with an etching time of 5-15s, to reveal ferrite grain boundaries in addition to colorizing the ferrite, pearlite, austenite, bainite and martensite phases. Table 3.2 summarizes the etching results of the tint etchant used to distinguish the various phases found in the final microstructure of the Al-TRIP steel. Table 3.2 Results for using a Tint Etchant to Distinguish Various Phases in a TRIP Steel Phase Etching Result Austenite White Ferrite Off-white or light grey Bainite Dark blue Martensite Dark brown An optical microscope and scanning electron microscope (SEM) was utilized to obtain micrographs of the specimen's cross-sectional area. Grain boundaries were Chapter 3 E X P E R I M E N T A L METHODS 42 outlined using a felt-tip marker on transparencies that were laid over micrographs for subsequent analysis. Quantitative analysis of the mean austenite grain size, mean polygonal ferrite grain size, mean polygonal ferrite fraction and fractions for the other transformation products were carried out utilizing a Clemex Image Analysis System. The mean austenite and polygonal ferrite grain sizes were calculated using Jeffries' planimetric method, the exact details of this method are given in ASTM Standard El 12-88. Using this method, grains completely within the measurement field were counted once and grains intersecting the perimeter of the measurement field were counted as a half grain. For statistically relevant results, several random areas were chosen as measurement fields and at least 50 grains were present in each field. Also, for each specimen a minimum of 300 grains were counted. The mean grain area from Jeffries' method was-' used to calculate the mean equivalent area diameter (EQAD). For determining the polygonal ferrite grain size, the area from the transformation products other than ferrite needed to be subtracted from the total measurement field area so that only the ferrite area was used for the grain size calculations. Polygonal ferrite fraction was quantified by dividing the measured ferrite area by the total measurement field area. The same procedure was undertaken to quantify the fractions for the other transformation products. 3.5 Hardness Measurements A Micromet 3 microhardness tester was used to perform hardness measurements on the CCT and deformation-transformation specimens. Measurements were taken in the same cross-sectional plane as the microstructural analysis was performed. A random selection of measurement fields was made while avoiding areas near the specimen surface due to oxidation and decarburization. Five measurements were made per specimen using a diamond indenter. 43 Chapter 4 Experimental Results: HSLA-90 Steel 4.1 Initial Austenite Grain Size Conditions Austenite grain growth was studied using isothermal tests in order to establish a suitable range of austenite grain sizes that would reflect grain sizes at the exit of the finishing mill of a hot strip mill. It was necessary to chose conditions, for subsequent transformation tests, that produced the desired average grain size in combination with a normal grain size distribution. T H E R M O C A L C software was employed to determine the A e 3 temperature, which was found to be 821°C for the HSLA-90 steel. A summary of the selected heat treatment cycles for application to C C T and deformation-transformation tests are specified in Table 4.1. Specimens were heated to the desired holding temperature at a rate of 5°C/s and water quenched following the specified holding time. An example of the resulting austenite microstructure is shown in Figure 4.1. The austenite grain size is reported as an equivalent volume diameter, which was obtained by multiplying the measured equivalent area diameter (EQAD) by 1.2 i.e. dvoi = \.2dEQAD, as discussed by Giumelli et al. [67]. Table 4.1 Heat Treatment Schedules and Measured Austenite Grain Sizes Holding Temp. (°C) Holding Time (s) Measured Austenite Grain Size (EQAD), d E Q A D (u.m) Measured Volumetric Austenite Grain Size, d v o i (pm) 950 120 12 14 1050 120 27 32 1150 120 44 53 Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 44 Some variation in the calculated EQAD of the austenite grain sizes from different field measurements was observed for a given specimen which could be accounted by the etching process, inhomogeneity in the steel, etc. From this, an overall experimental error associated with the austenite grain size measurements was estimated to be approximately 5%. T S 50 | im Figure 4.1 Austenite microstructure after reheating to 1150°C with a holding time of 120s and water quenching. 4.2 Continuous Cooling Transformation Results 4.2.1 Effect of Cooling Rate on Transformation Kinetics An example of continuous cooling transformation kinetics is shown in Figure 4.2, for various cooling rates and an austenite grain size of 14pm. The transformation kinetics is the result of a complex combination of thermally activated nucleation and growth processes. Accelerating the cooling rate from the austenite region shifts the transformation kinetics to lower transformation temperatures as shown in Figure 4.2. For this diffusional transformation, an increase in the cooling rate decreases the time available for ferrite Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 45 grains to nucleate and grow. Therefore, at any given temperature the time available for nucleation and growth is reduced resulting in a decrease in the transformation temperatures, a trend which is seen for all the austenite grain sizes. "D CD E CO c CO 1 h -c o o co 1 .0 0.8 0.6 0.4 0.2 |A A 0.0 A A A A A A * 1°C/s o 20°C/s • 56°C/s A 172°C/s A A A O O O O O O O O 9> 1 t 450 500 550 600 650 700 750 T e m p e r a t u r e , °C Figure 4.2 Effect of cooling rate on the austenite decomposition kinetics for an austenite grain size of 14pm. Transformation is often characterized by the transformation start temperature, Ts, which is defined as the temperature of 5% fraction transformed. A non-linear relationship between the cooling rate and transformation start temperature is evident from the transformation kinetics in Figure 4.2. The effect of cooling rate on transformation start temperature is more clearly illustrated in Figure 4.3. For example, the change in Ts from a cooling rate of l°C/s to 56°C/s is 111°C while from 56°C/s to 172°C/s it is only 34°C. The transformation finish temperature is defined as the temperature when 95% transformation has occurred. Trends for the effect of cooling rate on transformation finish temperatures are similar to those seen for transformation start temperatures, as shown in Figure 4.3. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 46 O o <D i_ -•—' CD i CU D. E cu l-•c 03 CO c o '•4—' CC E i o 4— c CO 800 700 600 500 H 400 5% fraction transformed 95% fraction transformed 0 20 40 60 80 100 120 140 160 180 200 Cooling Rate, ° C / s Figure 4.3 Effect of cooling rate on transformation start temperature for an austenite grain size of 14um. 4.2.2 Effect of Austenite Grain Size on Transformation Kinetics The influence of austenite grain size on the continuous cooling transformation kinetics is illustrated in Figure 4.4, without deformation. There is a clear shift of the transformation kinetics to lower transformation temperatures with an increase in the initial austenite grain size. A larger initial austenite grain size provides less boundary surface area per unit volume, thus reducing the number of available nucleation sites for ferrite and as a result transformation occurs at lower temperatures. In addition, the carbon diffusion distance is greater for a larger initial austenite grain size requiring additional time for diffusion, thereby lowering the transformation temperature. The shift to lower transformation start temperatures for an increase in the initial austenite grain size, shown in Figure 4.5, is associated with the delay of transformation start for larger austenite grains since, comparatively they provide fewer nucleation sites. Similar trends were observed for transformation finish temperatures. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 47 1.0 i ; 1 0.0 J 1 r - , , < , , , , , 1 420 440 460 480 500 520 540 560 580 600 620 640 Temperature, °C Figure 4.4 Comparison of austenite decomposition kinetics for various austenite grain sizes at a cooling rate or approximately 60°C/s . O 800 6 Z3 - t — • CO 750 \ CL o 650 CO C O c o ro £ c CO 600 550 500 450 400 • A d y=14pm • d y=32pm • d y=53pm 0 20 40 60 80 100 120 140 160 180 200 Cooling Rate, °C/s Figure 4.5 Effect of austenite grain size and cooling rate on transformation start temperature. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 48 4.2.3 Microstructural Evaluation for C C T Tests The resulting microstructures were investigated for the CCT tests where ferrite fraction and ferrite grain size were quantified. Figure 4.6 illustrates the effect of cooling rate on the resulting microstructure for cooling rates ranging from 1 to 172°C/s and an austenite grain size of 14um. A mainly polygonal ferrite-pearlite microstructure was obtained for the slowest cooling rate, a further increase in the cooling rate to 56°C/s resulted in ferrite grain refinement a decrease of the ferrite phase which can be associated with the decrease in the transformation start temperature. At cooling rates greater than 56°C/s, a microstructure consisting predominately of non-polygonal transformation products such as bainite and acicular ferrite was formed. Thus, an increase in the cooling rate leads to a decrease in the polygonal ferrite fraction. This was observed for cooling rates of l°C/s and 56°C/s, where the measured polygonal ferrite fraction was approximately 90% and disappeared completely for cooling rates greater than 56°C/s. Figure 4.6 Microstructure for continuous cooling tests for an austenite grain size of 14pm at: a) l°C/s; b) 56°C/s; c) 127°C/s; d)172°C/s. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 49 Increasing the austenite grain size to 32um shifted the transition to non-polygonal transformation products to a lower cooling rate i.e. 17°C/s, as illustrated in Figure 4.7 . A measured polygonal ferrite fraction of approximately 90% was observed only for cooling rates less than 17 °C/s, at larger cooling rates non-polygonal transformation products, such as bainite and acicular ferrite, were observed. A larger austenite grain size of 53um, produced non-polygonal transformation products for the whole range of cooling rates investigated, i.e. 1 to 179°C/s. The larger austenite grain size increases the diffusion distance and as a result an increase in the time required for the redistribution of carbon. Therefore, a greater degree of supercooling is achieved that encourages the formation of non-polygonal transformation products under the same cooling conditions. \ a) ; : : i V v -••..Yi^if. n^M v » * i ,t ' 50 pm • .50 urns c) ••: . • .'. 50 pmf.Jjsv* Figure 4.7 Microstructure for continuous cooling tests for an austenite grain size of 32pm at: a) l°C/s; b) 17°C/s; c) 130°C/s. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 50 The effect of the austenite grain size and cooling rate on the final ferrite grain size was also quantified. Significant refinement of the final microstructure and in particular the final ferrite grain size was observed for an increase in the cooling rate and a decrease in the initial austenite grain size, as illustrated in Figure 4.8. £ Cooling Rate, C /s Figure 4.8 Effect of cooling rate and austenite grain size on polygonal ferrite grain size. The final ferrite grain size depends mainly on the number of available nucleation sites. The driving pressure for ferrite nucleation and transformation is directly dependant on the magnitude of the applied undercooling. As the degree of undercooling is increased by continuous cooling more nuclei are formed with preferred nucleation sites initially at grain corners, followed by grain edges and grain boundaries, respectively. The larger nucleation density achieved through undercooling yields a greater number of grains and as a result a finer ferrite grain size. Furthermore, the grain boundary area per unit volume is inversely related to the austenite grain size, therefore, the number of available nucleation sites increase with a decrease in the austenite grain size resulting in a finer ferrite grain size. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 51 4.2.4 Hardness Results for C C T Tests Results from hardness measurements carried out on the CCT specimens are shown in Figure 4.9. The increase in cooling rate and/or initial austenite grain size gave rise to a significant increase in the hardness of the steel. For example, an increase in the cooling rate from 1 to 179°C/s increased the hardness by 90 HV for an austenite grain size of 53pm and for an increase in the austenite grain size from 14 to 53pm, the hardness increased by 58 HV for a cooling rate of 127°C/s. 300 n , 140 "I ' 1 ' ' r - ' 1 1 — i 1 0 20 40 60 80 100 120 140 160 180 200 Cool ing Rate, °C/s Figure 4.9 Effect of cooling rate and austenite grain size on hardness. The observed increase in hardness obtained by increasing the cooling rate is related to a decrease in the transformation temperatures for which non-polygonal products, i.e. harder phases are formed. In addition, decreasing the austenite grain size also results in an increase in hardness; this is due to the refinement of ferrite grains with a decrease in the austenite grain size. This relationship between transformation start temperature and hardness for various cooling rates and austenite grain sizes is shown in Figure 4.10. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 52 300 800 Figure 4.10 The dependency of hardness on transformation start temperature for various austenite grain sizes. 4.3 Effect of Retained Strain on Austenite Decomposition 4.3.1 Double-hit Compression Tests Double-hit compression tests were conducted in order to establish conditions that would introduce retained strain in the initial austenite microstructure which reflects the microstructure at the exit of the finishing mill. By deforming the initial austenite microstructure under no recrystallization conditions, the effect of retained strain on the transformation behaviour and resulting microstructure can be studied. Factors which affect restoration kinetics in addition to temperature, strain, strain rate and interhit time include chemical composition and austenite grain size. In order to maximize the amount of retained strain in the specimen, deformation was applied at temperatures slightly above the calculated A e 3 temperature. Two deformation temperatures namely, 900°C and 850°C, were chosen for investigation which are consistent with temperatures found at the exit of the finishing mill. Specimens were held for 30s at the deformation temperatures for homogenization purposes. A low constant strain rate of Is"1 was chosen to provide sufficient control of the deformation Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 53 process. Strain levels of 0.25 and 0.5 and interhit times between 15-60s were selected for study. A typical stress-strain curve for a double-hit compression test is illustrated in Figure 4.11 for an applied strain of 0.5, interhit time of 15s and a deformation temperature of 900°C. Under these deformation conditions, a significant amount of softening i.e. approximately 72%, was observed during unloading which indicates that recovery and recrystallization occurred during the interhit time. However, decreasing the deformation temperature to 850°C and employing the previous deformation conditions, resulted in a decrease in the softening to approximately 20%, as shown in Figure 4.12, which indicates that only recovery has taken place during the unloading. This retardation of recrystallization is a result of reduced grain boundary mobility due to the solute drag effect by Nb and Ti and pinning effects by fine precipitates of Nb and Ti carbonitrides at grain boundaries. 300 -1 1 250 \ 0.0 0.2 0.4 0.6 0.8 True Stra in Figure 4.11 Double-hit compression test for a deformation temperature of 900°C strain of 0.5 and strain rate of Is"1. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 54 300 250 ro P 150 oo CD 100 50 0 interhit ^ ^ S f in terh i t - ^5S T d e f = 850°C d y =14ujn 0.0 0.2 0.4 0.6 0.8 True Stra in Figure 4.12 Double-hit compression test for a deformation temperature of 850°C, strain of 0.5 and strain rate of Is"1. Furthermore, increasing the interhit time provides additional time for recovery and recrystallization to occur, which is shown in Figure 4.12 for interhit times of 15 and 45s, and thus, increases the degree of softening. The effect of interhit time and austenite grain size on the softening behaviour is shown in Figure 4.13. It can be seen that increasing the interhit time and/or deformation temperature results in a significant increase in softening. There is a substantial decrease in softening with an increase in the austenite grain size from 14u,m to 53 pm, where for softening up to approximately 20% it is assumed that primarily only recovery takes place [66]. In addition, it was found that for a strain of 0.25, softening could be attributed to only recovery taking place. As a result, a deformation temperature of 850°C was chosen for subsequent deformation-transformation tests, in addition to the holding time of 30s before applying strains of 0.25 and 0.5 at a constant strain rate of Is"1. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 55 C £Z CD * ± O co 100 80 -60 40 20 0 0 d y =14um;T d e f = 900°C d=14urn ;T d e f = 850°C d y=53um; T d e f = 900°C 10 20 30 40 50 60 Interhit Time, s Figure 4.13 Comparison of softening at various interhit times, deformation temperatures and austenite grain sizes for a strain of 0.5 and strain rate of Is"1. 4.3.2 Effect of Retained Strain and Accelerated Cooling on Transformation Kinetics The effect of retained strain in combination with cooling rate and austenite grain size on the transformation kinetics, resulting microstructure and hardness was investigated. Details on the experimental conditions employed for the deformation-transformation tests are given in Table 4.2. Table 4.2 Deformation-Transformation Test Schedule Volumetric Deformation Strain Strain Cooling Rates Austenite Grain Temperature Rate Size, dY (pm) (°C) CS"') (°C/s) 14 850 1 0.25 1,5,25,63 850 1 0.50 1,5,25,88 32 850 1 0.25 1, 5,25, 84 850 1 0.50 1,5,25,84 53 850 1 0.25 1,5,25,98 850 1 0.50 1,5,25,78 Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 56 An example of the accumulation of retained strain on the transformation kinetics, for cooling rates from 63-88°C/s and an austenite grain size of 14pm, are shown in Figure 4.14. As illustrated in Figure 4.14, an increase in the applied strain prior to accelerated cooling significantly shifts the transformation to higher temperatures. Similar observations were made for other austenite grain sizes and cooling rates. The difference in transformation behaviour between the deformed and strain-free cases was the result of an increase in available nucleation sites in the deformed microstructure. Pancaking the initial austenite grain by deformation increases the grain boundary surface area per unit volume, increases the irregularity of the grain boundaries and introduces defects i.e. deformation bands inside the initial austenite grain, these factors all contribute to increasing, the density of available nucleation sites for ferrite and as a result transformation occurs at lower temperatures. The combined effect of retained strain and accelerated cooling on the transformation kinetics was also investigated. An example of the resulting transformation kinetics are shown in Figure 4.15, for a constant strain of 0.5 and an austenite grain size of 53pm. Similar trends were seen for other austenite grain sizes and applied strains. These results further confirm the behaviour observed for the undeformed conditions where it was seen that increasing the cooling rate shifts the transformation to lower temperatures. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 1.0 57 CD E co c CO o CO 0.8 0.6 § 0.4 0.2 0.0 8 = 0 s = 0.25 s = 0.5 500 520 540 560 580 600 620 640 660 680 Temperature, °C Figure 4.14 Effect of strain on the transformation kinetics for an austenite grain size of 14pm and cooling rates of approximately 63-88°C/s . 500 520 540 560 580 600 620 640 660 680 Tempera ture , ° C Figure 4.15 Effect of cooling rate on the transformation kinetics for an austenite grain size of 53pm and an applied strain of 0.5. Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 58 Increasing the amount of retained strain in the austenite microstructure effectively increased the transformation start temperature, as illustrated in Figure 4.16. For example, the 75 can be increased by approximately 40 °C by applying a strain of 0.5 at a cooling rate of approximately 60°C/s, however this effect is diminished as the cooling rate is decreased. co c CO £ 400 4 , , , , , , , , , 1 0 20 40 60 80 100 120 140 160 180 200 Cooling Rate, ° C / s Figure 4.16 Effect of retained strain on transformation start temperature for an austenite grain size of 53pm. 4.3.3 Microstructural Results for Pancaked Austenite and Accelerated Cooling The evolution of the ferrite microstructure with the application of strain up to 0.5 is shown in Figure 4.17 for a cooling rate of approximately 60°C/s with an austenite grain size of 14pm. Ferrite is shown as dark regions whereas the other transformation products are shown as white regions in the SEM micrographs. As mentioned previously, a mainly polygonal ferrite-pearlite microstructure was obtained without deformation, the application of 0.25 and 0.5 strain significantly refined the microstructures and thus, decreased the ferrite grain size. However, there was no significant effect of retained strain on the amount of polygonal ferrite fraction i.e. measured ferrite fraction was relatively constant at an average of 0.9 with an estimated error of +0.02 for all experiments. On the other hand, an increase in the applied strain shifted the transition from polygonal to quasi-Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 59 polygonal microstructures to higher cooling rates. For example, for an austenite grain size of 14u,m, polygonal ferrite was present up to a cooling rate of 56°C/s, while at a strain of 0.5 this cooling rate was increased to 88°C/s. Thus, the presence of retained strain is a prerequisite to produce a predominantly polygonal ferritic microstructure under run-out table cooling conditions. Figure 4.17 Microstructural evolution of polygonal ferrite with accumulation of retained strain with an austenite grain size of 14pm: a) 8=0, CR=56°C/s; b) s=0.25, CR=63°C/s; c) s=0.5, CR=88°C/s. A finer polygonal ferrite grain size was achieved with the application of deformation due to an increase in available nucleation sites for ferrite grains, as discussed previously. An example of the measured polygonal ferrite grain sizes at various cooling rates and an austenite grain size of 14pm is given in Figure 4.18. Significant ferrite grain refinement of approximately 2pm was achieved with the application of accelerated cooling and deformation for the smallest austenite grain size of 14pm. For example, the Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 60 finest ferrite grain size of 2.1um was attained with the largest strain of 0.5 and accelerated cooling at 88°C/s while a ferrite grain size of 4um was measured from slow cooling at l °C/s with no deformation. A measurement error of approximately 10% was estimated for the ferrite grain size measurement. Cool ing Rate, °C/s 100 Figure 4.18 Ferrite grain refinement for an austenite grain size of 14um with the application of accelerated cooling and presence of retained strain. 4.4 Effect of Applied Strain on Hardness The combined effects of applied strain and accelerated cooling on hardness are illustrated in Figure 4.19 for an austenite grain size of 14pm. The successive application of strain led to a decrease in hardness, for example, applying a strain of 0.25 at a cooling rate of approximately 60°C/s led to a decrease in hardness by 22 H V . Also, it was previously shown that accelerated cooling resulted in an increase in hardness for a constant strain, which is verified by the results shown in Figure 4.19. Similar trends were observed for the other austenite grain sizes. The associated decrease of hardness with an increase of strain can be attributed to an increase in the transformation start temperatures, as illustrated in Figure 4.20. As Chapter 4 E X P E R I M E N T A L RESULTS: HSLA-90 61 shown previously, the accumulation of retained strain leads to higher transformation temperatures and the formation of a progressively more polygonal ferrite microstructure i.e. a softer product phase. 240 > X </) (/) CD \ ro i CD o 220 ] 200 140 0 = 0 • e = 0.25 • s = 0.5 20 Figure 4 40 60 80 100 120 140 Cool ing Rate, °C/s 19 The combined effects of cooling rate and retained strain on hardness for an austenite grain size of 14pm. > X CO w CD c ~o CD X CO L-CD o 240 220 200 180 > 160 140 s = 0 s = 0.25 s = 0.5 580 600 620 640 660 680 700 720 740 760 Transformation Start Temperature, °C Figure 4.20 The dependency of transformation start temperature on hardness for strain and strain-free cases for an austenite grain size of 14pm. a Chapter 5 Modelling of Austenite-to-Ferrite Transformation: HSLA-90 Steel 5.1 Modelling of Austenite-to-Ferrite Transformation Kinetics In order to describe the austenite-to-ferrite transformation kinetics a sequential transformation model was applied which consisted of sub-models to predict the transformation start temperature, ferrite growth and ferrite grain size, as previously described in Chapter 2. 5.1.1 Transformation Start Temperature The transformation start temperatures without any retained strain were predicted using Equation 2.11, where the diffusivity of carbon in austenite was calculated using the Agren approach [68]. The carbon concentrations in austenite, cr, and ferrite, ca, were calculated using THERMOCALC assuming orthoequilibrium and c0 is the average carbon concentration. The model parameter TN was estimated to be TN = 787°C from a cooling rate of l°C/s. The second model parameter c* was represented as follows, c* = (2.5 + 2.5exp(-0.0002(r/v-7;)17))co (5.1) This model approach gives an accurate description of the experimentally observed transformation start temperatures, as illustrated in Figure 5.1. Some deviation from the model was observed, not shown in Figure 5.1, for the larger austenite grain size at higher cooling rates where a complex microstructure was obtained for these conditions. However, for run-out table cooling conditions, where the aim is to have a microstructure containing 80% ferrite or more, these conditions are irrelevant. Chapter 5 M O D E L L I N G : HSLA-90 63 250 0 ~\ r • 1 • 1 10 1 10 2 10 3 10 4 10 5 (pd y 2, °Cs-Vm2 Figure 5.1 Model predictions of transformation start temperature without deformation. The present model was also applied to transformation start temperatures with retained strain where the existing expressions for the model parameters TN and c* were employed. The effect of retained strain was incorporated in the model be means of an effective austenite grain size, defined as follows, deJf = dy exp(-f) (5.2) where dY is the volumetric austenite grain size and s is the applied strain. Model predictions and observed transformation start temperatures for various strains and austenite grain sizes are compared in Figure 5.2. As seen in this Figure, the proposed model provides an accurate description of the observed transformation start temperatures for all of the investigated strain conditions. Chapter 5 M O D E L L I N G : HSLA-90 64 Figure 5.2 Transformation start model predictions incorporating effect of strain. 5.1.2 Austenite-to-Ferrite Transformation Kinetics The subsequent ferrite growth was modelled using an Avrami approach (i.e. J M A K model) and adopting the additivity rule, the details of which are described in Chapter 2. In order to model the ferrite growth, any non-polygonal and secondary transformation products and the associated kinetics were excluded from the analysis. Thus, the analysis was applied to transformation kinetics that had a minimum polygonal ferrite fraction of 85%, determined from microstructural analysis, by normalizing the polygonal ferrite fraction transformed at each temperature increment with respect to the total fraction transformed. Employing the previously described J M A K ' model with the additivity rule, Equation 2.16 can be transformed as follows, h = (dXY < v dt , In 1 - X 0 - 4 (5.3) Chapter 5 M O D E L L I N G : HSLA-90 65 where the model parameters b and n were determined using experimental data. The additivity rule requires that only b be a function of temperature whereas the exponent n must be temperature independent. Thus, using C C T results n was varied in the range of 0.8 - 1.2 to determine a relationship for b which was independent of cooling rate. A best fit constant value of n = 0.85 was determined for which the parameter b could be expressed solely as a function of temperature for all cooling rates. Figure 5.3 illustrates an example of the experimental \x\b as a function of temperature using C C T data for an austenite grain size of 32 pm and n = 0.85. Previous work has shown that as a first approximation, lnfr can be assumed to increase linearly with temperature for the whole range of cooling rates at a given austenite grain size, as follows, \x\b = BT + Y (5.4) where B and Fare the model parameters summarized in Table 5.1 for the C C T results. c 1 0 -1 -2 -3 -4 -5 Inb = -0 .04T + 22.4 1°C/s 5°C/s 17°C/s 540 560 580 600 620 640 660 680 700 720 Tempera tu re , °C Figure 5.3 Experimental \nb expressed as a linear function of temperature for various cooling rates, an austenite grain size of 32 pm and n = 0.85. Chapter 5 M O D E L L I N G : HSLA-90 Table 5.1 Model Parameters for the Linear Expression of \nb 66 dyiyrm) B Y 14 -0.04 24.2 32 22.4 As shown in Table 5.1, the slope B, is independent of austenite grain size however, the intercept Y, decreases as the austenite. grain size increases. Using the grain size-modified Avrami equation, Equation 2.16, the effect of austenite grain size was incorporated in the model as follows, X = 1 - exp ' At^ V dr J b = — or ln b - ln A - m ln dY d: y (5.5) where A is the temperature dependant factor and Equation 5.4 can be re-written as follows, \nb = BT + D'-m\ndy where Y = D'-m\ndy (5.6) Using data presented in Table 5.1, the values for m and D' were determined as shown in Figure 5.4. Thus, \nb can be expressed as follows, \nb = -0.047/ + 29.7 - 2.2 ln dv (5.7) As illustrated in Figure 5.5, the grain size-modified J M A K model provides a reasonable description of the observed transformation kinetics. Chapter 5 M O D E L L I N G : HSLA-90 25 22 -I . , ,—• , 1 2.4 2.6 2.8 3.0 3.2 3.4 lnd y Figure 5.4 Derivation of model parameters for the grain size-modified Avrami equati 560 580 600 620 640 660 680 Tempera ture , °C Figure 5.5 Application of the grain size-modified JMAK model. Chapter 5 M O D E L L I N G : HSLA-90 68 The effect of retained strain was included in the model by employing the effective austenite grain size, defined in Equation 5.2, to obtain the following relation for \x\b, In 6 = BT + D'-m\n{dy(-s)\ (5.8) An example of the model predictions and experimental data are shown in Figure 5.6 for an austenite grain size of 14pm and a cooling rate of approximately 25°C/s . The model predictions show a reasonable agreement with the experimental data, similar results were observed for all other cases. This agreement is clearly shown in Figure 5.7 and Figure 5.8 which compares the measured and predicted temperatures at a ferrite fraction of 20% and 80%, for a range of predominately polygonal ferrite microstructures including the retained strain cases. Figure 5.6 Application of the grain size-modified J M A K model incorporating the effective austenite grain size for an austenite grain size of 14pm and a cooling rate of approximately 25°C/s . Chapter 5 M O D E L L I N G : HSLA-90 Figure 5.7 Comparison of measured and predicted temperatures for 20% fraction transformed. 560 580 600 620 640 Temperature C Figure 5.8 Comparison of measured and predicted temperatures for 80% fraction transformed. Chapter 5 M O D E L L I N G : HSLA-90 70 5.2 Modelling of Polygonal Ferrite Grain Size The ferrite grain size, which is essentially determined at the start of transformation assuming nucleation site saturation, is one of the most important microstructural features due to its influence on the final mechanical properties of the steel. The ferrite grain size model applied to the experimental data is subsequently a function of the transformation start temperature, in addition to austenite grain size and ferrite fraction transformed, as given in Equation 2.18. The model was applied only to those cases with a predominately polygonal ferrite microstructure i.e. non-polygonal products produced for austenite grain sizes of 14 and 32pm without retained strain at higher cooling rates, also all cases for an austenite grain size of 53pm were excluded from the analysis. Further, for an austenite grain size of 14pm, cases for the slowest cooling rate of l°C/s were excluded from the analysis since these results could not be accurately described using this model and further work is needed to include this data in the model. As discussed previously the model parameter Ti is a function of the initial austenite grain size, thus the effect of retained strain needs to be considered when employing Equation 2.19. Once again an effective austenite grain size can be employed such that M can be written as a function of retained strain and austenite grain size as follows, where C, R and N are model parameters, e is the applied strain and dr is the austenite grain size (EQAD) in pm. The model parameters C, R, n' and E are summarized in Table 5.2 and the .final ferrite fraction, F, was approximated as 0.9 from microstructural measurements. Table 5.2 Model Parameters for Predicting the Final Ferrite Grain Size M = C + Rdrexp(-n's) (5.9) C R n E 13.4 0.068 6.3 10 000 Chapter 5 M O D E L L I N G : HSLA-90 71 The model accurately predicts the ferrite grain size for the various austenite grain sizes and strain conditions investigated, as illustrated in Figure 5.9. A complete comparison of the model predictions and measured ferrite grain sizes can be clearly seen in Figure 5.10. E n CD" N C O _ 6 CD H—' I— CD s 0 0.25 0.5 14 u,rn o ® • 32 u.m • • • Mode l o 620 640 660 680 700 720 740 760 Transformat ion Start Tempera ture , °C Figure 5.9 Comparison of the measured and predicted ferrite grain sizes. Chapter 5 M O D E L L I N G : HSLA-90 1.5 2.0 2.5 3.0 3.5 4 .0 4 .5 5.0 M e a s u r e d Ferr i te G r a i n S i z e , u.m Figure 5.10 Comparison of measured and predicted ferrite grain size. Chapter 6 Experimental Results: Al-TRIP steel 6.1 Austenitization Tests T H E R M O C A L C software was employed in order to establish suitable reheat temperatures to be employed for subsequent transformation tests. Figure 6.1 shows a section through the phase diagram at 1.5Mn-1.6Al (wt.%) for the Fe-C-Mn-Al system predicted by T H E R M O C A L C . As shown in Figure 6.1, for a carbon concentration of 0.2 wt.% a single austenite phase (FCC) was not predicted, instead a mixed phase with austenite and ferrite (FCC and B C C ) is expected for all reheat conditions. For a temperature 1167°C a maximum austenite fraction of 0.87 and a ferrite fraction of 0.13 was expected. An isothermal test was conducted in order to verify this prediction, where a C C T specimen was heated at a rate of 5°C/s to 1167°C with a holding time of 2min followed by a Helium gas quench. This test revealed a fully martensitic structure indicating the presence of only austenite at this reheat condition. As such, further isothermal tests were conducted for a temperature range of 1 2 5 0 - 9 5 0 ° C to analyze the predictions by T H E R M O C A L C . 2000 \ .Liquid+FCC+BCC BCC+FCC+Cementite BCC+Cementite 10 12 14 16 18 20 4 6 WEiGHT_FRACTION C Figure 6.1 Section through the phase diagram at 1.5Mn-1.6Al (wt.%) for the Fe-C-Mn-Al system as predicted by T H E R M O C A L C . Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 74 A microstructural analysis was completed on the isothermal tests. A fully martensitic microstructure was revealed for a reheat condition of 1167°C which indicates the presence of only austenite as shown in Figure 6.2c). At a lower reheat temperature of 1050°C, small amounts of ferrite (white) were present throughout the microstructure as seen in Figure 6.2b). The ferrite fraction increased with a lowering of the reheat temperature to 950°C as seen in Figure 6.2a), in addition, dark needle-like structures, where some of the needles were seen grouped together, were observed throughout the resulting microstructure which were not present for a reheat temperature of 1050°C. For this analysis, these dark needle-like structures were considered to be part of the austenite phase fraction. Figure 6.2 Examples of the resulting microstructures: a) 950°C; b) 1050°C; and c) 1167°C. A summary of the ferrite fractions quantified by a microstructural analysis of the isothermal tests is shown in Table 6.1. At temperatures ranging from 1250°C to 1117°C a Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 75 fully austenitic phase is present, however, at a temperature below 1117°C a two phase region exists consisting of austenite and ferrite. This would also suggest that the A e 3 temperature is in the range of 1117-1050°C. An experimental error for the measured ferrite fraction at 950°C and 1050°C was estimated to be 15% of ferrite fraction, i.e. errors arose from the outlining of the fine ferrite structures. Table 6.1 Ferrite Fractions from Microstructural Analysis and THERMOCALC Predictions Reheat Temperature (°C) Microstructural Analysis THERMOCALC Predictions 1250 0 0.19 1217 0 0.15 1167 0 0.13 1117 0 0.15 1050 0.04 0.20 950 0.33 , 0.35 A comparison of the measured ferrite fraction and THERMOCALC predictions is presented in Table 6.1. The discrepancy between predicted and measured ferrite fractions, approximately 0.15 ferrite fraction, was confirmed for other reheat temperatures besides the previously discussed reheat temperature of 1167°C. A consistently larger ferrite fraction was predicted by THERMOCALC. Furthermore, International Plasma Labs Ltd. conducted an independent chemical analysis to verify the given chemical composition of the Al-TRIP steel from Dofasco Inc., as shown in Table 6.2. A higher aluminium content was reported by International Plasma Labs Ltd., which suggests that an even larger fraction of ferrite should be present at a given temperature based on the THERMOCALC database. Thus, this would suggest a limitation of THERMOCALC for predicting phase diagrams for large levels of Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 76 aluminium, which is most likely due to the lack of available thermodynamic data in the databases used for calculations. Table 6.2 Chemical Composition of Al-TRIP Steel (wt%) C Mn Al* Dofasco Inc. 0.19 1.42 1.68 International Plasma Labs Ltd. 0.18 1.45 1.80 *The aluminium content is reported as the as-soluable aluminium, i.e. aluminium in solution, for both chemical analyses. Figure 6.3 shows the austenite grain size of the Al-TRIP steel at elevated temperatures for a holding time of 2min where the initial structure was fully austenitic with the exception of 1050°C where the austenite fraction was 0.96. As expected, an increase in the reheat temperature results in an increase in the austenite grain size, i.e. austenite grain growth. 60 T E A 55 -20 \ , , , , , 1 1000 1050 1100 1150 1200 1250 1300 Temperature, °C Figure 6.3 Austenite grain growth of the Al-TRIP steel with a holding time of 2min. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 77 There was some variation in the calculated EQAD of the austenite grain sizes from different fields of measurements for a given specimen, where an experimental error for the austenite grain size measurement was estimated to be 10%. Possible measurement errors could have arisen from not being able to fully distinguish some of the austenite grain boundaries. 6.2 Quench-in Tests A series of quench-in tests were undertaken with the purpose of validating the dilatometry measurements taken during continuous cooling; a slow cooling rate of l°C/s was chosen for this study. Specimens were austenitized at 1167°C for 2min followed by cooling to the desired quench temperature, i.e. 1000°C, 800°C, 750°C and 650°C, and were subsequently quenched using Helium gas to room temperature. Figure 6.4 illustrates the transformation kinetics from measured dilation during slow cooling along with the corresponding quenched-in microstructures at the various quench temperatures. The microstructural evolution of ferrite can be clearly seen in Figure 6.4 where an increase in the ferrite fraction is seen as cooling progresses. For quench temperatures of 1167°C and 1000°C a fully austenitic microstructure was present; Figure 6.4a) shows the austenite microstructure for the quench temperature of 1000°C. At a temperature of 800°C the austenite begins to transform to ferrite and continues transforming as seen for temperatures of 750°C and 650°C shown in Figure 6.4b)-d). The ferrite fractions were also measured for each quench temperature, .as shown in Table 6.3, which shows the increase in ferrite fraction as the transformation progresses. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 78 Figure 6.4 Transformation kinetics from measured dilation for slow cooling with the corresponding quenched-in microstructure with quench temperatures of: a) 1000°C; b) 800°C; c) 750°C; and d) 650°C. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 79 A comparison of the measured ferrite fraction by microstructural analysis and from the dilation measurement was completed, as shown in Table 6.3. It is evident from Table 6.3 that there is reasonable agreement between the two methods of quantifying the ferrite fraction, and thus, it can be concluded that the dilation measurements accurately describe the transformation behaviour. Table 6.3 Comparison of Ferrite Fraction by Microstructural Analysis and Dilation Measurement Quench Temperature (°C) Ferrite Fraction -Microstructural Analysis Ferrite Fraction -Dilation Measurement 1167 - -1000 800 0.32 0.33 750 0.47 0.51 650 0.65 0.72 In addition to measuring ferrite fraction, the ferrite grain sizes were also quantified, as shown in Table 6.4. Variations in the calculated EQAD of the ferrite grain sizes from different field measurements were observed for a given specimen which could be due to some indistinct ferrite grain boundaries resulting in an associated experimental error given in Table 6.4 as Ada. An increase in the ferrite grain size was observed as the specimen was progressively cooled which could be the result of nucleation, growth, coalescence or a combination of these processes. As a first approximation to evaluate which of these processes might be occurring, the number of ferrite grains per area was calculated for each temperature as follows, N = N' lolal 4F Jtdl (6.1) Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 80 where N' is the number of ferrite grains, Alolal is the total measured area, F is the ferrite fraction andJ ais the ferrite grain size (EQAD). As shown in Table 6.4, there was an increase in the number of ferrite grains per area as the temperature decreased from 800°C to 650°C by approximately 13% which would suggest that nucleation and growth occurs. The error for the number of ferrite grains per area, AN, can be estimated from the maximum and minimum number of ferrite grains per area, Nmax and Nmin, as follows, N ,1 n{da-AdJ and N- = 4 F n (da+AdJ (6.2) as shown in Table 6.4. The error for the number of ferrite grains per area is approximately 20% using this approach. Since this error is greater than the 13% increase seen in the number of ferrite grains per area with a decrease in temperature, than as a first approximation the number of ferrite grains per area can be taken as constant, i.e. only growth is occurring. In addition, from microstructural observations intergranular nucleation was not confirmed. However, a further detailed statistical analysis needs to be completed, i.e. more ferrite grains need to be counted for the results to be more statistically relevant, in order to make any further conclusions. Table 6.4 Measured Austenite and Ferrite Grain Sizes Quench Temperature Measured Austenite Grain Size (EQAD) Measured Ferrite Grain Size (EQAD) -da ± Ada Number of Ferrite Grains per area - TV Range of Number of Ferrite Grains from Ferrite Grain Size Measurement Error Nmax Nmin (°C) (pm) (pm) (grains/mm ) (grains/mm ) (grains/mm2) 1167 42 - - - -1000 • 43 - - - -800 - 22 ± 2.2 868 1072 717 750 - 26 ± 2 . 6 885 1093 732 650 - 29 ± 2 . 9 984 1215 813 Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 81 6.3 Continuous Cooling Transformation Results The austenite decomposition of the Al-TRIP steel was studied using two different initial conditions: 1) a fully austenitic phase with reheating at 1167°C and 2) a mixed phase consisting of 33% ferrite and 67% austenite achieved by reheating at 950°C. An isothermal holding time of 2min was employed at the reheat temperature subsequently followed by cooling. 6.3.1 Single Austenite Phase Initial Condition Transformation Kinetics for Single Austenite Phase The continuous cooling transformation kinetics for the single austenite phase are presented in Figure 6.5, for a range of cooling rates with a mean initial austenite grain size of 42pm (EQAD). The transformation kinetics are shifted to lower temperatures as the cooling rate is accelerated from the austenite region due to a reduction in the available time for the nucleation and growth processes at any given temperature. In particular, there is a significant effect of cooling rate on transformation start temperatures, which is seen clearly in Figure 6.6. However, for the transformation finish temperatures there seems to be little effect of cooling rate which could be associated with the martensitic transformation at these low temperatures. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 82 1.0 -o 0.8 <u E *§ 0.6 c co .2 0 4 o CO 0.2 0.0 • A • ° O _ A A M . O • A O o A A • A O * 1 °C/s o 20 °C/s • 60 °C/s A 120°C/s • 160°C/s o A • A A • A • * • A O O O O i 200 300 400 500 600 700 800 900 1000 Tempera ture , °C Figure 6.5 Effect of cooling rate on the austenite decomposition kinetics for the single austenite phase condition. 1200 j -o o CD 1000 -Z3 «+—* CO CD o 800 -E Q) h- 600 -C o rmal 400 -<> Cfl c 200 -CD f— 0 0 50 5% fraction t ransformed 9 5 % fraction transformed 100 150 200 Cool ing Rate, °C/s Figure 6.6 Effect of cooling rate on transformation temperatures for the single austenite phase condition. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 83 Microstructural Evaluation for C C T Tests for Single Austenite Phase The resulting microstructures were investigated for the CCT tests where the corresponding fractions for the various phases and polygonal ferrite grain sizes were quantified. Figure 6.7 illustrates the effect of cooling rate, ranging from l°C/s to 160°C/s, on the resulting microstructure for transformation from the single austenite phase. A mainly polygonal ferrite microstructure was obtained for the slow cooling rate of l°C/s where the other transformation products consisted of fine pearlite and bainite structures shown in Figure 6.7a). A ferrite-bainite microstructure was obtained for a cooling rate of 20°C/s, shown in Figure 6.7b), where the fraction of bainite increased in comparison with cooling at l°C/s; this is associated with the decrease in the transformation start temperature. At the largest cooling rate of 160°C/s, shown in Figure 6.7c), ferrite almost completely disappears and a microstructure consisting predominately of bainite and martensite was obtained; similar microstructures were also observed for cooling rates of 60°C/s and 120°C/s. For the slower cooling rates of l°C/s and 20°C/s, it was possible to quantify the mean ferrite grain sizes (EQAD), which were found to be 31um and 14um, respectively. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 84 b) la i f * : "IfiiS' -• Fern Figure 6.7 Microstructures for continuous cooling tests for the single austenite phase at: a) l°C/s; b) 20°C/s and c) 160°C/s. The effect of cooling rate on the volume fraction of the various phases was also quantified as shown in Table 6.5. Ferrite was present for both cooling rates of l°C/s and 20°C/s, however, an increase in the cooling rate from l°C/s to 20°C/s resulted in a shift from pearlite to bainite as the secondary phase which, corresponds to the shift seen in the transformation kinetics. For these complex microstructures, the criteria used to quantify bainite was a defined packet of lath structures, some examples are outlined in Figure 6.7a) and b); pearlite can be seen as the dark regions in Figure 6.7a). A 12% measurement error was estimated for the ferrite and bainite phase fractions at a cooling rate of 20°C/s due to the challenges in distinguishing the bainite structure, whereas for the cooling rate of l°C/s the measurement error was lower than this. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 85 At larger cooling rates, i.e. 60°C/s to 160°C/s, there was a shift to a predominately bainite-martensite microstructure which, is consistent with the low transformation temperatures required for the formation of these phases seen in the transformation kinetics. For these cooling rates, the two types of ferrite observed were Widmanstatten (primary sawteeth) and allotriomorph ferrite which were easily distinguishable from bainite and martensite. The bainitic and martensitic phases were differentiated by the tint etching procedure as discussed in Chapter 3. The complexity of the bainite and martensite microstructures provided a challenge for clear identification of phase boundaries, as a result, the measurement error associated with the measured phase fractions was estimated to be 20%. Table 6.5 Effect of Cooling Rate on the Phase Fractions for the Single Austenite Phase Initial Condition Measured Phase Fractions Cooling Rate Ferrite Pearlite Bainite Martensite Retained Austenite 1 0.76 0.16 0.05 - 0.03 20 • 0.58 • 0.04 0.34 - 0.04 60 0.08 0.74 0.18 -120 0.10 0.18 0.72 -160 0.07 0.15 0.78 -In order to evaluate the transformation finish temperatures with the martensitic transformation some estimates of the martensite start, Ms, temperature can be made using empirical relationships. Eldis [69] and Andrews [70] have proposed empirical relationships as a function of steel chemistry, i.e., M ( = 531 - 391.2C - 43.3M? - 21.8M - 16.2Q- (Eldis) (6.3) Ms = 539 - 423C - 30.4M? - 17.7 Ni - 12.K> - 115/ - 7Mo (Andrews) (6.4) Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 86 For the present analysis it is essential to account for the carbon enrichment in austenite such that carbon does not represent the nominal carbon concentration of the steel but the carbon concentration remaining in the austenite phase during ferrite formation. The carbon and manganese concentrations for the Al-TRIP steel were within the given ranges for which these equations are applicable. In addition, neither relationship accounts for the effect of aluminium on Ms which, introduces a source of error in the predicted Ms temperatures. The carbon concentration remaining in austenite was determined for each cooling rate from the following relation [9], Cr = s. (6.5) r a m m m " s { \ - x a - x B ) where c0 is the initial carbon concentration, Xa is the ferrite fraction determined from microstructural analysis and Xti is the bainite fraction determined from microstructural analysis; it is assumed that bainite forms as a carbide-free bainite. The estimated Ms temperatures calculated from both Equations 6.1 and 6.2 predict within + 8°C of each other; an average Ms temperature from these equation is shown in Figure 6.8 along with the transformation finish temperatures. As shown in Figure 6.8, for a cooling rate of l °C/s , the calculated Ms range was 73-78°C which is less than the measured transformation finish temperature of 333°C, in addition the Ms range calculated for a cooling rate of 20°C/s was 172-174°C which is also below the measured transformation finish temperature of 307°C. However for these cooling rates no martensite was observed in the microstructure, and thus, Ms temperatures are not applicable to these cooling conditions. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 87 600 O o g) 500 400 \ CD i_ Z2 -*—* CD CD CL E CD H C o '•4—* CD E o M— CO c CO 100 0 • 9 5 % fraction t ransformed • Est imated M c • • • 0 50 100 150 200 Figure Cool ing Rate, C / s 6.8 Comparison of measured transformation start temperatures and estimated average martensite start temperature ranges for various cooling rates. For a cooling rate of 60°C/s, the calculated Ms range was 50-57°C which is lower then the measured transformation finish temperature of 292°C, in addition, martensite (approximately 18%) was observed in the resulting microstructure. For this case, it is possible that the entire transformation was not captured, i.e. measurements were taken during cooling until approximately 100°C. However, a simple dilation test would be difficult to analyze below the estimated Ms range of 50-57°C since a linear portion at the finish of transformation is required for the analysis. In addition, the assumption that the bainite is a carbide-free bainite may not be valid. For the last cooling rates of 120°C/s and 160°C/s, the estimated Msranges were 368-384°C and 376-392°C, respectively, which are above the measured transformation finish temperatures of 287°C and 236°C, respectively. Thus, according to the estimated temperatures, martensite transformation has already begun which may suggest from this analysis that the measured transformation finish temperatures coincide with the martensite finish temperatures. This simple analysis suggests that for the cases where martensite was present some correlation may exist between the measured transformation finish temperatures and those of martensite Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 88 transformation temperatures, however, further investigations are required which, are beyond the scope of this work. The transformation start temperature model, as described in Chapter 2, was applied to the results for the single phase austenite phase condition. The transformation start temperatures were predicted using Equation 2.11, where the diffusivity of carbon in austenite was calculated using the Agren approach [68]. The carbon concentrations in austenite, cr, and ferrite, ca, were calculated using THERMOCALC assuming orthoequilibrium and c0 is the initial carbon concentration. The model parameter TN was estimated to be TN = 890°C from a slow cooling rate of l°C/s and the second model parameter c* was determined to be c*/c = 1.2. This model approach gives a reasonable trend of the experimentally observed transformation start temperatures, as illustrated in Figure 6.9, however, the trend is not as good as seen for the HSLA-90 steel; this maybe due to the questions raised with THERMOCALC. Cooling rates of 120°C/s and 160°C/s, where complex microstructures were obtained, were excluded from this analysis since for run-out table cooling conditions the aim is to have a microstructure consisting of 40-60% ferrite, so that these conditions were irrelevant. 0 20000 40000 60000 80000 100000 120000 cpdy2, °Cs-1um2 Figure 6.9 Model predictions for transformation start temperature for austenite phase. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 89 6.3.2 Two Phase Initial Condition Transformation Kinetics for the Two Phase Condition Similar effects of cooling rate on the continuous cooling transformation kinetics were seen for the two phase initial condition, as shown in Figure 6.10. In this case, the transformation of bainite and martensite phases were clearly distinguishable as evidenced by the bumps in the kinetics curve, in particular for cooling rates of 60°C/s , l20°C/s and 166°C/s. It is interesting to note that for this initial condition, the effect of cooling on the transformation start temperature is less pronounced whereas the transformation finish temperature is significantly affected by cooling rate, as illustrated in Figure 6.11. The transformation finish temperatures would suggest that at slow cooling rates of l °C/s and 15°C/s the final microstructure would consist of high temperature transformation products such as ferrite and pearlite. This is followed by a transition, as seen in the decrease in the transformation finish temperature from a cooling rate of 15 to 60°C/s , to low temperature transformation products such as martensite for the cooling rates between 60 and 166°C/s. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 90 CD CO CD - C C L CU ' c CD H—* CO < T3 CD E o CO c CO c o o CO 1.0 0 .8 0 .6 0 .4 0.2 0 .0 A » A A A * 1 u C / s o 1 5 ° C / S • 6 0 ° C / s A 1 2 0 ° C / S • 1 6 6 ° C / s A A A * A « m OA o , > A . \ A • ^ , Q 2 0 0 4 0 0 6 0 0 8 0 0 Temperature, C Figure 6.10 Effect of cooling rate on the austenite decomposition kinetics for the two phase initial condition. 1 2 0 0 O CD 1 0 0 0 -4—' CO I CD Q . E CD r -C o CD E CO c CO 8 0 0 6 0 0 4 0 0 2 0 0 0 .0 A 5% fraction transformed • 9 5 % fraction transformed 5 0 100 1 5 0 2 0 0 Cooling Rate, C/s Figure 6.11 Effect of cooling rate on transformation temperatures for the two phase initial condition. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 91 From Figure 6.11, it is evident that transformation start temperature is not as sensitive to cooling rate as seen for the single austenite phase. Both ferrite nucleation and growth contribute to the transformation start temperature as a function of cooling rate. As a result, if one of the processes is eliminated the sensitivity of transformation start temperature to cooling rate should be reduced which is observed for this case. Thus, if only growth of ferrite occurs than the transformation start model is not applicable to the two phase initial condition since an assumption of the model is that both ferrite nucleation and early growth of corner ferrite occurs. Further analysis is required to develop new models that account for only growth of ferrite, in addition, the role of alloying elements will need to be taken into account. Microstructural Evaluation for C C T Tests for the Two Phase Condition Figure 6.12 illustrates the effect of cooling, with rates ranging from 1 to 166°C/s, on the resulting microstructure for the case of the two phase initial condition. The resulting microstructure for the slowest cooling rate, shown in Figure 6.12a), consisted mainly of polygonal ferrite and a fine pearlite phase (dark regions). Further, white needle-like structures randomly oriented within the ferrite grains were observed throughout the microstructure; these were assumed to be acicular ferrite. The pearlite fraction decreases with an increase in the cooling rate to 15°C/s, in addition, bainite was observed; bainite was assumed to be a defined packet of laths, examples of which are shown in Figure 6.12b). Further increasing the cooling rate to a range above 15°C/s produced low temperature transformation products such as bainite and martensite with the disappearance of pearlite, which is consistent with the shift to lower transformation temperatures seen in the transformation kinetics, i.e. Figure 6.10. As an example, the resulting microstructure for the largest cooling rate is shown in Figure 6.12c) where the dark regions were a mixture of bainite and martensite and were identified by employing the same procedures as described for the single phase condition. In addition, there were dark needle-like structures distributed throughout the microstructure, which were also present in the quenched-in microstructure. As before, a Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 92 defined packet of these dark needle-like structures were assumed to be bainite, examples of this are shown in Figure 6.12c). Figure 6.12 Microstructures for continuous cooling tests for the two phase initial condition at: a) l°C/s; b) 15°C/s and c) 166°C/s. A summary of the measured phase fractions for the two phase initial condition is given in Table 6.6. Ferrite is reported as the combination of the initial ferrite fraction and any new ferrite which has formed during cooling. Chapter 6 E X P E R I M E N T A L RESULTS: Al-TRIP S T E E L 93 Table 6.6 Effect of Cooling Rate on the Phase Fractions for the Two Phase Initial Condition Measured Phase Fractions Cooling Rate Ferrite (Old and New) Pearlite Bainite Martensite Retained Austenite 1 0.78 0.21 - - 0.01 15 0.87 0.03 0.06 - 0.04 60 0.75 - 0.08 0.12 0.05 120 0.53 - 0.10 0.37 -166 0.36 - 0.18 0.46 For the slowest cooling rate of l°C/s, the polygonal ferrite, acicular ferrite and pearlite phases were clearly distinguishable for measurement. For a cooling rate of 15°C/s, pearlite (dark regions) and retained austenite (white regions) where easily identified, as shown in Figure 6.12b). In addition, bainite was considered to be a defined packet of laths, examples of which are shown in Figure 6.12b), and the remaining was assumed to be ferrite. Due to the challenges in quantifying the bainite phase, i.e. outlining bainite, a measurement error was estimated to be 20%. The phase fractions for cooling rates ranging from 60 to 166°C/s were also quantified. For the large dark regions seen in Figure 6.12c), which are a mixture of bainite and martensite, tint etching was employed to distinguish between these two phases. In addition, to those areas identified as bainite within the large dark regions, packets of the dark needle-like structures were also identified as bainite, as shown in Figure 6.12c); the remaining was assumed to be ferrite. The complex nature of these microstructures provided a challenge for identifying the various phases; a measurement error associated with these phases was estimated to be 25%. As summarized in Table 6.6, an increase in cooling rate results in a shift to low temperature transformation products, i.e. bainite and martensite, which is consistent with the associated transformation kinetics. However, further analysis of these complex microstructures is required to accurately identify and quantify the various phases present. Chapter 7 Conclusions and Future Work 7.1 Conclusions The austenite decomposition kinetics and resulting microstructures during continuous cooling for two high strength steels were investigated, i.e. a Nb/Ti microalloyed H S L A steel (HSLA-90) and an aluminium alloyed transformation-induced plasticity (Al-TRIP) steel. A previously developed sequential transformation model was applied to describe the austenite-to-polygonal ferrite transformation. Based on the experimental results and model predictions of the transformation kinetics and resulting microstructure, the following conclusions can be made: HSLA-90 1. The austenite decomposition is heavily influenced by the initial austenite grain size and cooling rate. For a given cooling rate, an increase in austenite grain size results in lower transformation start temperatures resulting in a decrease in polygonal ferrite fraction Further, for a given austenite grain size, accelerated cooling lowers the transformation start temperature with an associated decrease in the polygonal ferrite fraction in addition to refining the resulting ferrite grains. Accelerated cooling in combination with smaller austenite grain sizes further refines the ferrite grains. 2. Deformation in a temperature range between T„r and Ae3 leads to an increase in nucleation rate of allotriomorph ferrite due to an increase in potential nucleation site density resulting in additional ferrite grain refinement. In addition, the transformation start temperatures were significantly increased by retained strain. The range of cooling rates over which polygonal ferrite was present increases with retained strain. Chapter 7 CONCLUSIONS AND F U T U R E W O R K 95 3. The effect of processing variables indicates that the hardness increases with austenite grain size and cooling rate as a result of the decrease in transformation start temperature where low temperature transformation products are formed, i.e. bainite and martensite, which are harder phases. An increase in the applied strain results in lower hardness values due to an increase in transformation start temperatures which produce softer phase products such as, polygonal ferrite. 4. A transformation start model that combines corner nucleation of ferrite with early growth was employed to describe the transformation start temperatures of allotriomorph ferrite. The combined effect of austenite grain size and retained strain was incorporated into the model by employing an effective grain size. There was good agreement between the measured and predicted transformation start temperatures. 5. The JMAK model was used to describe the subsequent ferrite growth during continuous cooling by adapting the additivity rule. The effect of austenite grain size on the transformation rate was incorporated into the model by employing a grain-size modified Avrami equation and for cases of retained strain the effective grain size approach was employed. The model accurately predicted the measured ferrite growth. 6. The polygonal ferrite grain size was modelled as a function of transformation start temperature and initial austenite grain size using a semi-empirical approach. For the cases of retained strain, the effective grain size approach was once again employed. The model predictions and measured ferrite grain sizes showed good agreement. Al-TRIP 7. A series of isothermal tests were conducted to quantify the initial microstructure. Decreasing the austenitization temperature reveals a transition from a single austenite phase to a mixed phase consisting of austenite and ferrite where the Chapter 7 CONCLUSIONS AND FUTURE W O R K 96 austenite fraction decreases, i.e. as a result ferrite fraction increases, as the temperature is lowered. A comparison of the measured phase fractions with predictions from T H E R M O C A L C showed the T H E R M O C A L C predictions to be consistently 15% lower for the austenite fraction. This would suggest a limitation of T H E R M O C A L C for predicting phase diagrams for large levels of aluminium. 8. A series of quench-in tests for a slow cooling rate at a fully austenitic initial condition was conducted. Ferrite and austenite fractions predicted by the dilation measurements were in good agreement with the measured phase fractions. 9. Accelerated cooling decreases the transformation start temperatures for a fully austenitic initial condition, and results in a shift from a ferrite-pearlite to a bainite-martensite microstructure. However, accelerated cooling for a mixed phase initial condition has a less pronounced effect on the transformation start temperature since there is no ferrite nucleation, i.e. only ferrite growth. In addition, accelerated cooling rate decreases the transformation finish temperatures which results in a transition from a ferrite-pearlite to a bainite-martensite microstructure. 7.2 Future Work 1. Polygonal ferrite has been quantified, however other transformation products for both steels need to be analyzed further to quantify them in terms of morphology, kinetics of formation and structure. 2. The overall transformation model needs to be applied to the Al-TRIP steel in order to describe the transformation start temperatures, ferrite growth and the ferrite grain size. 3. Additional initial austenite conditions, i.e. austenite grain sizes and deformation, need to be studied to quantify their effects on the transformation kinetics, resulting ferrite grain size and microstructure for the Al-TRIP steel. 17 References [I] J.W. Christian, The Theory of Transformation in Metals and Alloys, 1st ed., Pergamon Press, Oxford, U K , 1965. [2] W. D. Callister, Materials Science and Engineering: A n Introduction. 3 r d ed., John Wiley and Sons Inc, New York, USA, 1994. [3] D. Dunne, Materials Forum. Vol. 23, 1999, 63.. [4] G.R. Speich, T . M . Scoonover, Microstructure and Properties of H S L A Steels, ed. A.J . DeArdo, T M S , Pittsburgh, PA, U S A , 1988, 263. [5] D.A. Porter and K . E . Easterling, Phase Transformations in Metals and Alloys, 2 n d ed., Chapman and Hall, London, U K , 1992. [6] S.J. Jones and H. Bhadeshia, Acta Metallurgica, Vol . 45, No. 7, 1997, 2911. [7] A . Dube, H. Aaronson and R. Mehl, Revue de Metallurgie. Vol . 55, 1958, 201. [8] W. Reynolds, M . Enomoto and H. Aaronson, Phase Transformations in Ferrous Alloys, ed. A . Marder and J. Goldstein, The Metallurgical Society of A1ME, Philadelphia, U S A , 1984, 155. [9] R. Honeycombe and H . Bhadeshia, Steels: Microstructure and Properties. Edward Arnold, London, U K , 1995. [10] N . Ridley, Phase Transformations in Ferrous Alloys, ed. A . Marder and J. Goldstein, The Metallurgical Society of A I M E , Philadelphia, U S A , 1984, 201. [II] H. Bhadeshia, Bainite in Steels, The Institute of Materials, London, U K , 1992. [12] R. Hehemann, K. Kinsman and H. Aaronson, Metallurgical Transactions 3A, 1972, 1077. [13] M . Hillert and G. Purdy, Scripta Materialia. Vol. 43, 2000, 831. [14] A . Hultgren. Transactions A S M , Vol. 39. 1947,915. [15] H . Aaronson, Decomposition of Austenite by Diffusional Processes, ed. V. Zackay and H . Aaronson, Interscience Publishers, N Y , U S A , 1960, 387. [16] G . Speich, L . Cuddy, G . Gordon and A, DeArdo, Phase Transformations in Ferrous Alloys, ed. A . Marder and J. Goldstein, The Metallurgical Society of A I M E , Philadelphia, U S A , 1984, 341. REFERENCES 98 17] A.K. Sinha, Ferrous Physical Metallurgy, Butterworths, Stoneham, MA, 1988. 18] S. te Velthuis, Ph.D. Thesis, Delft University of Technology, Delft University Press, The Netherlands, 1999. 19] I. Tamura, C.Ouchi, T. Tanaka and H. Sekine, Thermomechanical Processing of High Strength Low Alloy Steels, Butterworths, London, UK, 1998, 21. 20] J. Cahn, Acta Metallurgica. Vol. 4, 1956, 449. 21] K. Russell, Phase Transformations, American Society for Metals, Metals Park, OH, USA, 1970, 219. '22] M.Enomoto and H. Aaronson, Metallurgical Transactions A, Vol. 17A, no.8, 1986,1381. 23] M. Enomoto, W.F. Lange III and H. Aaronson, Metallurgical Transactions A, Vol. 17A, no.8, 1986, 1399. 24] M. Enomoto, ISLJ, Vol. 70, no. 14, 1984, 1648. '25] M. Militzer, R. Pandi and B. Hawbolt, Metallurgical and Materials Transactions A, Vol. 27A, no. 6, 1996, 1547. 26] H. Aaronson, C. Laird and K. Kinsman, Phase Transformations, ASM, Metals Park, OH, USA, 1968, 313. [27] C. Zener, Journal of Applied Physics. Vol. 20, 1949, 250. [28] E. Simonen and H. Aaronson, Metallurgical Transactions A, Vol. 4, 1973, 1239. [29] J. Bradley, J. Rigsbee and H. Aaronson, Metallurgical Transactions A, Vol. 8A, 1977,323. [30] J. Bradley and H. Aaronson, Metallurgical Transactions A, Vol. 12A, 1981, 1729. [31] A. Kolmogorov, Izv. Akademii Nauk USSR Ser. Matemat, Vol. 1, 1937, 355. [32] W. Johnson and R. Mehl, Transactions AIME, Vol. 135, 1939, 416. [33] M. Avrami, Journal of Chemical Physics, Vol. 7, 1939, 1103. [34] M. Avrami, Journal of Chemical Physics, Vol. 9, 1941, 177. REFERENCES 99 [35] J. Gilmour, G. Purdy and J. Kirkadly, Metallurgical Transactions, Vol . 3A, 1972, 3213. [36] S. Okaguchi, T. Hasimoto and H . Ohtani, T H E R M E C ' 8 8 . ed. I. Tamura, IS1J, Vol. 1, 1988, 330. [37] A. Singh, D. Ramakrishna and S. Gupta, Zeitschrift fur Metallkunde, Vol. 79, 1988, 180. . • ' [38] L.Collins and W. Liu, Phase Transformation During the Thermal/Mechanical Processing of Steel, ed. B. Hawbolt and S. Yue, The Metallurgical Society of CIM, 1995, 419. [39] F. Oberhauser, F. Listhuber and F. Wallener, Mircoalloying'75, ed. M . Korchynski, Union Carbide Corporation, New York, USA, 1977, 665. [40] P. Manohar, K. Kunishige, T. Chandra and M . Ferry, Materials Science and Technology, Vol . 18, 2002, 856. [41] R. Pandi, Ph.D. Thesis, The University of British Columbia, Vancouver, B C , Canada, 1998. [42] M . Militzer, B. Hawbolt and R. Meadowcroft, Metallurgical and Materials Transactions, Vol . 31 A , 2000, 1247. [43] M . Militzer, B. Hawbolt and R. Meadowcroft, Phase Transformation During the Thermal/Mechanical Processing of Steel, ed. B. Hawbolt and S. Yue, The Metallurgical Society of CIM, 1995,445. [44] N . Nakata and M . Militzer, Mechanical Working and Steel Processing Conference Proceedings, Vol. XXXVIII , ISS, Warrendale, PA, USA, 2000, 813. [45] B. Hawbolt, B. Chau and K. Brimacombe, Mathematical Modelling of Hot Rolling of Steel, ed. S. Yue, The Metallurgical Society of C I M , Canada, 1990, 424. [46] R. Pandi, M . Militzer, B. Hawbolt and R. Meadowcroft, 37 t h Mechanical Working and Steel Processing Conference Proceedings, Vol . XXXIII, ISS, Warrendale, PA, USA, 1995, 635. [47] E. Essadiqi, M . Akben and J. Jonas, Phase Transformations in Ferrous Alloys, ed. A. Marder and J. Goldstein, The Metallurgical Society of A I M E , Philadelphia, USA, 1984, 391. REFERENCES 100 L. Collins, J. Barry and J. Boyd, Phase Transformations in Ferrous Alloys, ed. A . Marder and J. Goldstein, The Metallurgical Society of A I M E , Philadelphia, USA, 1984,397. R. Priestner, Phase Transformation During the Thermal/Mechanical Processing of Steel. ed.B. Hawbolt and S. Yue, The Metallurgical Society o f C I M , 1995,211. M . Umemoto, H. Ohtsuka and I. Tamura, Journal of the Iron and Steel Institute of Japan. Vol . 70, no. 6, 1984, 557. I. Kozasu, C. Ouchi, T. Sampei and T. Okita, Microalloving'75, Union Carbide Corportation, New York, N Y , U S A , 1977, 100. M . Umemoto, H . Ohtsuka and I. Tamura, Transactions ISIJ. Vol . 23, 1983, 775. S. Lacroix, Y . Brechet, M . Veron, D. Quidort, M . Kandel and T. lung, Austenite Formation and Decomposition, ISS and T M S , Warrendale, PA, 2003, 367. E.Scheil, Arch. Eisenhuttenwesen, Vol. 8, 1935, 565. M . Umemoto, A . Hiramatsu, A . Moriya, T. Watanabe, S. Nanba, N . Nakajima, G. Anan and Y . Higo, ISIJ International. Vol. 32, 1992, 306. M . Lusk and H. Jou, Metallurgical and Materials Transactions, Vol . 28A, 1997, 287. M . Umemoto, K. Horiuchi and I. Tamura, Transactions ISIJ, Vol . 23. 1983, 690. M . Umemoto, N . Komatsubara and I. Tamura, Journal of Heat Treating. Vol . 1, 1980, 57. B. Hawbolt, B. Chau and K. Brimacombe, Metallurgical Transactions. Vol. 14A, no.9, 1983, 1803. B. Hawbolt, B. Chau and K. Brimacombe, Metallurgical Transactions, Vol . 16A, 1985, 565. J. Holloman, L . Jaffe and M . Norton, Transactions A I M E . Vol. 167, 1946, 419. G. Manning, and C. Lorig, Transactions A I M E . Vol. 167, 1946, 442. J. Leblond and J. Devaux, Acta Metallurgica. Vol. 32, 1984, 137. M.Suehiro, K. Sato, Y. Tsukano, H . Yada, T. Senuma and Y . Matsumura, Transactions ISIJ. Vol . 27, 1987, 439. REFERENCES 101 [65] M . Militzer, B. Hawbolt and R. Meadowcroft, H S L A '95, ed. L . Guoxun, H. Stuart, Z. Hongtao and L . Chenggji, China Science and Technology Press, Beijing, China, 1995,271. [66] P. Petkov, M.A.Sc . Thesis. The University of British Columbia, British Columbia, Canada, 2004. [67] A . Giumelli, M . Militzer and B. Hawbolt, ISIJ International. Vol . 39, 1999, 271-280. [68] J. Agren, Scripta Metallurgica. Vol. 20, 1986, 1507. [69] J. Barralis and G. Maeder, Collection Scientifique E N S A M . 1982, 270. [70] K. W. Andrews, Journal of the Iron and Steel Institute. 203, Vol . 7, 1965, 721. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0078775/manifest

Comment

Related Items