Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Grain growth and austenite decomposition in two niobium containing line pipe steels Robinson, Isaac Dalton Gordon 2016

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

Item Metadata

Download

Media
24-ubc_november_2016_robinson_isaac.pdf [ 5.8MB ]
Metadata
JSON: 24-1.0308782.json
JSON-LD: 24-1.0308782-ld.json
RDF/XML (Pretty): 24-1.0308782-rdf.xml
RDF/JSON: 24-1.0308782-rdf.json
Turtle: 24-1.0308782-turtle.txt
N-Triples: 24-1.0308782-rdf-ntriples.txt
Original Record: 24-1.0308782-source.json
Full Text
24-1.0308782-fulltext.txt
Citation
24-1.0308782.ris

Full Text

Grain Growth and Austenite Decompositionin Two Niobium Containing High StrengthLine Pipe SteelsbyISAAC DALTON GORDON ROBINSONB.A.Sc, The University of British Columbia, 2013A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTSFOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES(Materials Engineering)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)August 2016©Isaac Dalton Gordon Robinson, 2016AbstractNiobium is a common microalloying element for line pipe steels for the promotion of fine fer-rite microstructures following hot rolling. During construction via welding, niobium carbonitridesformed during coiling may dissolve depending on peak temperatures reached. This has a stronginfluence on the microstructure and mechanical properties of the heat affected zone. The primaryobjective of the current work is to study the effect of niobium microalloying on heat affected zonemicrostructures. Two steel compositions were selected for this study, possessing a high (0.091 wt%)niobium content and differing (0.028 and 0.058 wt%) carbon contents.To study the effect of niobium on grain growth, thermal histories were designed in the contextof continuous heating and isothermal holding for a range of heating rates and holding temperatures.Laser ultrasonics were used to observe in-situ grain growth during these histories and experimentalresults were confirmed by ex-situ metallography using an appropriate etchant. Niobium is shown tobe effective in restricting grain growth behavior at temperatures when precipitates are stable. Anincrease in heating rate was found to reduce overall grain growth (from 48 to 25 µm between 10 and1000°C/s), eventually reaching a limiting behavior at the maximum heating rate (i.e. 1000°C/s).To study the effect of niobium on austenite decomposition, thermal histories were designedto produce two grain sizes (5 and 35 µm) and two states of niobium (in and out of solution).The latter was investigated through precipitation experiments. Cooling rates were varied between3 and 30°C/s and mechanical dilatometry was used to measure the phase transformation. Asexpected, an increase in cooling rate, carbon content, and prior austenite grain size results in lowertransformation temperatures. The effect of niobium in solution was strongly dependent on prioraustenite grain size.Microstructures formed by austenite decomposition were characterized via metallography andiiABSTRACTmicrohardness. The microstructural constituents include ferrite, bainite, and M/A. Samples repre-sentative of each microstructure were selected to quantify M/A fraction through etching and pointcounting. Niobium was shown to have a strong effect on hardness when the prior austenite grainsize was large (up to 70 HV).iiiPrefaceAll experimental design, heat treatments, metallography, experimental analysis, and thermody-namic calculations were conducted by the author at the University of British Columbia. Sampleslisted in Chapter 4 were designed by Thomas Garcin and fabricated by the Materials Engineeringmachine shop. Analysis of laser ultrasonics waveforms and dilatometry were carried out using theCTOME program suite (LUMFILE and PROFILE, respectively) developed by Thomas Garcin.Thermodynamic calculations listed in Chapter 5 were carried out using the CALPHAD softwareThermocalc v4.0 (‘tcfe7’ database), developed by Thermo-Calc Software Inc.A portion of the experimental work discussing grain growth and austenite decomposition withreference to a steel with a lower niobium content will be presented and published later this year.(I.D.G. Robinson, T. Garcin, W.J. Poole, M. Militzer ”The Effect of Niobium on Austenite De-composition in High Strength Line Pipe Steels”, 2016 10th International Pipeline Conference, InPress)ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.1 Microstructures in Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.1.1 Ferrite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.1.2 Bainite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.1.3 Martensite-Austenite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82.2 Development of Line Pipe Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.2.1 The Role of Niobium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112.3 Girth Welding of Line Pipes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182.3.1 Heat Affected Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.3.2 Welding of Nb-based HSLA Steels . . . . . . . . . . . . . . . . . . . . . . . . 213 Scope and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24vTABLE OF CONTENTS4.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244.2 Grain Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254.2.1 Laser Ultrasonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264.2.2 Grain Growth Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294.3 Austenite Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314.3.1 Dilatometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314.3.2 Nb(C,N) Reprecipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324.3.3 CCT Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334.4 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.1 Grain Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.1.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.1.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525.2 Niobium Precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565.2.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565.2.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.3 Austenite Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615.3.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615.3.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746 Summary and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 806.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 806.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82viList of Tables1.1 Overview of line pipe composition of recent steel grades between 2008 and 2015 [3] . 24.1 Given composition of studied alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245.1 Overview of grain size distribution data for continuous heating grain growth trials . 515.2 Overview of grain size distribution data for isothermal holding grain growth trials . 515.3 Overview of grain growth distribution data for quenched samples for CCT reheatingfollowing solutionization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515.4 Equilibrium volume fractions of Nb(C,N) precipitates at selected isothermal holdingtemperatures and percentage relative to ambient equilibrium fraction [41] . . . . . . 525.5 Composition of previously characterized steel [5] . . . . . . . . . . . . . . . . . . . . 545.6 M/A volume fractions for analyzed CCT samples; all samples have a prior austenitegrain size of 35 µm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71viiList of Figures2.1 Iron-carbon phase diagram over a range of carbon contents relevant to steel produc-tion [7]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.2 Continuous-cooling transformation diagram for X70 line pipe steel following heatingto 1300°C for variety of cooling rates; F, P and B are ferrite, pearlite, and bainiterespectively [10]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3 Illustration of upper and bainite formation and morphology [20] . . . . . . . . . . . . 72.4 Example of a rolling schedule for TMCP of steel . . . . . . . . . . . . . . . . . . . . 102.5 Formation enthalpies of common microalloy precipitates in steels [42] . . . . . . . . . 122.6 Precipitation-time-temperature (PPT) diagram highlighting the effect of strain onprecipitation start (Ps) and finish (Pf ) times on a Nb-V-Ti steel measured by elec-trical resistivity [45] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132.7 Experimentally measured recrystallization stop temperatures of steels as function ofmicroalloying content [31] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.8 Effect of niobium in solution on a) ferrite start temperature and b) bainite starttemperature in a low carbon (0.021%C) steel following simulated finish rolling, datafrom [66]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172.9 Peak temperature and related t8−5 measurements following single torch and dualtorch welding [74]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.10 Microstructures forming in a weld area for a single welding pass and correspondingFe-C phase diagram [42] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204.1 As-received microstructures of selected compositions a) 0.028%C b) 0.058%C . . . . 254.2 Sample dimensions for bar specimens a) tubular specimens b) . . . . . . . . . . . . 264.3 Experimental setup for laser ultrasonics [57] . . . . . . . . . . . . . . . . . . . . . . . 27viiiLIST OF FIGURES4.4 Calibration of b-parameter from experimental grain size measurements [86] . . . . . 294.5 Thermal history for continuous-heating grain-growth trial . . . . . . . . . . . . . . . 304.6 Thermal history for isothermal holding grain-growth trial . . . . . . . . . . . . . . . 304.7 Schematic dilatometry curve of austenite-to-daughter phase transformation . . . . . 324.8 Thermal history for dissolution and reprecipitation of Nb(C,N) . . . . . . . . . . . . 344.9 Thermal history for austenite decomposition at a) small and b) large prior austenitegrain sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354.10 Process to measure prior austenite grain size from a micrograph a) Original micro-graph b) Hand outlining c) Threshold specification d) Area count measurement . . . 374.11 Process to measure M/A fraction through thresholding, top left to lower right, a)Original figure b) Grayscale and filtering c) Threshold specification d) Area countmeasurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384.12 An example hardness indent and caliper measurement from CCT sample . . . . . . . 385.1 Laser ultrasonic measurement of grain growth during continuous heating as a func-tion of heating rate for compositions with carbon contents of of a) 0.028%C b)0.058%C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415.2 Laser ultrasonic measurement of grain growth during isothermal holding as a functionof holding temperature for compositions with carbon contents of a) 0.028%C b)0.058%C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425.3 Micrographs of etched 0.028% C samples from continuous heating grain growth trials(a), c), e)) and grain diameter histograms (b), d), f)); heating rates of 10° C/s (a),b))100° C/s (c), d)) and 1000° C/s (e), f)) . . . . . . . . . . . . . . . . . . . . . . . . 445.4 Micrographs of etched 0.058% C samples from continuous heating grain growth trials(a), c), e)) and grain diameter histograms(b), d), f)); heating rates of 10° C/s (a),b)) 100° C/s (c), d)) and 1000° C/s (e), f)) . . . . . . . . . . . . . . . . . . . . . . . 455.5 Micrographs of etched 0.028% C samples from isothermal grain growth trials(a),(c),(e)and grain diameter histograms(b),(d),(f); holding temperature of 950° C (a), b))1150° C (c), d)) and 1250° C (e), f)) . . . . . . . . . . . . . . . . . . . . . . . . . . . 46ixLIST OF FIGURES5.6 Micrographs of etched 0.058% C samples from isothermal grain growth trials(a), (c),(e)) and grain diameter histograms(b), (d), (f)); holding temperature of 950° C (a),b)) 1150° C (c), d)) and 1250° C (e), f)) . . . . . . . . . . . . . . . . . . . . . . . . . 475.7 Micrographs of etched samples from isothermal grain growth trials (a),(c) and grainsize histograms (b),(d); 0.028% C (a), b)) and 0.058% (c), d)), holding temperatureof 1050° C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485.8 Micrographs of etched 0.028% C (a), b)) and 0.058% (c), d)) samples quenched frompeak temperatures prior to CCT tests (a),(c)) and grain size histograms (b),(d));peak temperature of 950° C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495.9 Micrographs of etched 0.028% C (a), b)) and 0.058% (c), d)) samples quenched frompeak temperatures prior to CCT tests (a),(c)) and grain size histograms (b),(d));holding temperature of 1250° C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505.10 Thermodynamic stability of phases in a) 0.028%C b) 0.058%C composition, fromThermocalc 4.0 [41] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535.11 Grain growth behavior in continuous heating as a function of composition for heatingrates of a) 10°C/s b) 100°C/s; 0.034%Nb steel data from [5] . . . . . . . . . . . . . . 555.12 Comparison of grain size distribution self-similarity for isothermal holding trials asa function of temperature for compositions with carbon contents of a) 0.028%C b)0.058%C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.13 Grain growth behavior in isothermal holding as a function of composition for holdingtemperatures of a) 1050°C b) 1150°C c) 1250°C; 0.034%Nb steel data from [74] . . . 585.14 Dilatometry curve for samples aged at selected times at 900°C for compositions witha carbon content of a) 0.028%C b) 0.058%C . . . . . . . . . . . . . . . . . . . . . . . 595.15 Transformation start temperature as a function of aging time at 900°C for steels withcarbon contents of a) 0.028%C and 0.058%C . . . . . . . . . . . . . . . . . . . . . . 605.16 Aging curve constructed from Ts data as a function of aging time at 900°C for steelswith carbon contents of 0.028%C and 0.058%C . . . . . . . . . . . . . . . . . . . . . 615.17 Example of recoalesence in 0.058% C alloy, Nb in solution, 35 µm prior austenitegrain size, 30°C/s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62xLIST OF FIGURES5.18 Austenite decomposition curves of CCT tests, observing transformation dependencyon niobium content for a) 0.028%C and a prior austenite grain size of 5 µm b)0.058%C and a prior austenite grain size of 5 µm c) 0.028%C and a prior austenitegrain size of 35 µm d) 0.058%C and a prior austenite grain size of 35 µm . . . . . . 635.19 Austenite decomposition curves of CCT tests, observing transformation dependencyon carbon content for a) niobium precipitated and a prior austenite grain size of5 µm b) niobium in solution and a prior austenite grain size of 5 µm c) niobiumprecipitated and a prior austenite grain size of 35 µm d) niobium in solution and aprior austenite grain size of 35 µm . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645.20 Austenite decomposition curves of CCT tests, observing transformation dependencyon prior austenite grain size for a) niobium precipitated and 0.028%C b) niobiumprecipitated and 0.058%C a) niobium in solution and 0.028%C b) niobium in solutionand 0.058%C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655.21 Nital 2% etched microstructures produced by austenite decomposition of 0.028% Calloy from a 5 µm average prior austenite grain size; shown as niobium precipitated(a),(c),(e) or in solution (b),(d),(f) with cooling rate of 3° C/s (a), b)) 10° C/s (c),d)), 30° C/s (e), f)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 675.22 Nital 2% etched microstructures produced by austenite decomposition of 0.058% Calloy from a 5 µm average prior austenite grain size; shown as niobium precipitated(a),(c),(e) or in solution (b),(d),(f), with cooling rate of 3° C/s (a), b)), 10° C/s (c),d)), 30° C/s (e), f)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 685.23 Nital 2% etched microstructures produced by austenite decomposition of 0.028% Calloy from a 35 µm average prior austenite grain size; shown as niobium precipitated(a),(c),(e) or in solution (b),(d),(f), with cooling rate of 3° C/s (a), b)), 10° C/s (c),d)), 30° C/s (e), f)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 695.24 Nital 2% etched microstructures produced by austenite decomposition of 0.058% Calloy from a 35 µm average prior austenite grain size; shown as niobium precipitated(a),(c),(e) or in solution (b),(d),(f), with cooling rate of 3° C/s (a), b)), 10° C/s (c),d)), 30° C/s (e), f)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70xiLIST OF FIGURES5.25 Le Pera etched microstructures for selected samples produced by CCT a) irregularferrite (0.028%C, Nb precipitated, 10°C/s cooling) b) upper bainite (0.028%C, Nb insolution, 10°C/s cooling) c) lower bainite (0.058%C, Nb in solution, 30°C/s cooling);all samples have a prior austenite grain size of 35 µm . . . . . . . . . . . . . . . . . . 725.26 Microhardness measurements for CCT samples plotted as function of cooling ratefor a prior austenite grain size of a) 5 µm b) 35 µm . . . . . . . . . . . . . . . . . . . 735.27 M/A volume fractions in samples produced by CCT as a function of Ts, adaptedfrom [27] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 765.28 Microhardness measurements for CCT samples plotted as function of Ts for a prioraustenite grain size of a) 5 µm b) 35 µm . . . . . . . . . . . . . . . . . . . . . . . . . 785.29 Continuous cooling diagrams for selected compositions, observing transformationdependency on niobium content for a) 5 µm prior austenite grain size and 0.028%Cb) 5 µm prior austenite grain size and 0.058%C c) 35 µm prior austenite grain sizeand 0.028%C d) 35 µm prior austenite grain size and 0.058%C . . . . . . . . . . . . 79xiiAcknowledgementsI would like to thank EVRAZ Steel, Transcanada Pipelines and the Natural Sciences and En-gineering Research Council of Canada for the provision of funding for the pipeline project, as wellas information on the compositions studied.I would like to thank my supervisor, Dr. Warren J. Poole, for his support, guidance and inputthroughout my thesis. I would also like to thank Dr. Matthias Militzer for his excellent advice inmeetings as well as Dr. Thomas Garcin for his invaluable assistance in the use of the Gleeble andsoftware I used to acquire and analyze results. I am fortunate for the opportunity to have workedwith all of them.I would like to thank all my colleagues in the Microstructure Engineering group, especially mylabmates in AMPEL 349, Jenny, Mahsa, Mojtaba, Keiji, Maddie, Brian, and Morteza. I couldn’thave made it through my thesis without your advice and companionship. As well, a special thanksto Ross McLeod and the technicians for their work ethic and attention to detail in fabricating thesamples used in my work.Finally, I would like to thank my family for their continued love and support throughout mydegree and, indeed, the whole of my life. Love you all!xiiiChapter 1IntroductionOil and natural gas are important resources for transportation, power production and heatgeneration. Canada alone contains the world’s third largest oil reserves and development hasaccelerated in recent years. Worldwide energy use is projected to rise another third by 2040,primarily driven by increased demand in the Middle East, India, China, and other developingnations [1].Oil and gas deposits in North America commonly require transport from remote locations.Pipelines are a safe and economical method of transportation of oil and gas as compared to railand truck transport [2] and are the most common method of transportation in the United Statesand Canada. There are a number of technical challenges in pipeline design, construction andoperation to be addressed by pipeline engineers. For example, it is more cost effective to increasethe throughput of oil and gas by increasing the flow rate. To increase the flow rate, it is necessaryto either increase the pipe diameter or the operating pressure. As another example, reducing thewall thickness of the pipe is advantageous as less material is used per length of pipe, resultingin reduced production and construction costs. In colder geographies, such as the high Arctic,pipelines undergo strain from unstable permafrost heaving, temperature fluctuations, etc. In allthree scenarios, higher stresses are produced in the pipe during operation and steels must be selectedwith enhanced mehanical properties that meet these more rigorous demands.Steel is the material of choice for oil and gas pipelines because of its high strength and relativelylow cost. Line pipe steels are designed to to meet the requirements of strength, ductility andweldability. These steels contain a low carbon content and a variety of microalloying elements1INTRODUCTIONunder a total of 1 wt%. Table 1.1 provides an overview of the range of line pipe steel compositionsin recent years. API X70/80 steels (where the number signifies the yield strength in ksi) have beenpreviously developed for use in an industrial context and X100/120 are in development for use infuture pipeline construction.Table 1.1: Overview of line pipe composition of recent steel grades between 2008 and 2015 [3]Alloy, Wt% C Si Mn Ni Cr Mo Al Ti V NbMaximum 0.09 0.4 1.9 0.44 0.25 0.3 0.05 0.023 0.06 0.09Minimum 0.028 0.11 1.51 0.01 0.01 0 0.003 0.001 0 0.003Average 0.06 0.25 1.64 0.07 0.10 0.10 0.03 0.02 0.03 0.06During the construction of pipelines, in-field girth welding by gas metal arc welding (GMAW)is used to join pipe segments together. When steel is welded, the microstructure undergoes asignificant change causing the mechanical properties of the weld zone to differ from those of the basematerial. This is known as the heat affected zone (HAZ). The final properties of the HAZ depend onthe spatial distribution of peak temperatures, the cooling rate from the peak temperature, and thechemistry of the steel in question. Multiple welding passes are commonly required to completelyjoin two sections. For this reason, dual-torch welding is likely to be advantageous, as the totalnumber of welding passes required to join sections of pipe is reduced, resulting in large savingsfrom reduced labor costs. However, the simple cost savings are offset by the complexity of thetemperature distributions in the weld zone. It is important to optimize the chemistry of the nextgeneration of line pipe steels to improve the properties of the HAZ when pipe sections are welded.Research suggests that niobium, one such microalloy useful in steelmaking, has significant, thoughvarying, effects on the mechanical behavior of microstructures within the HAZ.This study concerns the effect of the condition of niobium on the grain growth and austenitedecomposition behavior of two selected line pipe steels with different carbon contents. This projectis part of a much larger effort to develop an integrated microstructure/mechanical property modelfor the HAZ zone in Nb microalloyed line pipe steels [4] [5].2Chapter 2Literature ReviewThe literature on metallurgy of microalloyed low-carbon steel grades, specifically line pipegrades, has been reviewed, focusing on commonly observed microstructures, the effects of niobiummicroalloying, and welding metallurgy.2.1 Microstructures in SteelThe iron-carbon phase diagram displays the regions of phase stability in the iron-carbon binarysystem as a function of temperature and carbon content. This is shown in Figure 2.1. At roomtemperature, pure iron has a body-centered cubic (BCC) atomic structure referred to as ferrite,or α. This phase is stable upon heating at temperatures up to 912°C (1185 K), above which irontransforms to a face-centered cubic (FCC) atomic structure referred to as austenite or γ. Austeniteis stable until 1394°C (1667 K), when the structure once more transforms to BCC for a smalltemperature region known as delta ferrite, or δ. Above 1538°C (1811 K), iron transforms to theliquid phase. As carbon is added to iron, it can be observed from the phase diagram that thetemperature at which ferrite transforms to austenite is lowered, forming a region where austeniteand ferrite co-exist. For a selected carbon content, in this dual phase region, the temperature atwhich the steel is completely ferritic is known as the Ae1 temperature and the temperature at whichthe steel is completely austenitic is known as the Ae3 temperature. The introduction of carbonproduces graphite coexisting with ferrite below Ae1. However, graphite is not observed in industrialconditions. Instead, metastable iron carbide (Fe3C), or cementite, is formed [6]. For this reason,3LITERATURE REVIEWFigure 2.1: Iron-carbon phase diagram over a range of carbon contents relevant to steel production[7].cementite usually takes the place of graphite in the iron-carbon phase diagram.Due to the non-equilibrium nature of metallurgical processes in the vast majority of industrialsteel production scenarios, the microstructures formed upon cooling to room temperature are farmore complex than those predicted in the iron-carbon phase diagram. For example, the rate ofcooling from temperatures above Ae3 has a strong effect on which phases are formed and theirmorphology. This is commonly illustrated by continuous-cooling transformation (CCT) diagramsconstructed for particular steel compositions. As shown schematically in Figure 2.2, the microstruc-ture formed from a single composition may vary dramatically with cooling rate. In addition, theaustenite grain size prior to transformation and the overall composition of the steel have a strongeffect on austenite decomposition.Austenite decomposition is defined as the transformation of austenite to its subsequent daughterphases. According to the iron-carbon phase diagram, decomposition will begin as temperature dropsbelow Ae3. However, due to kinetic considerations, the onset of transformation may be depressedby over 100°C from that predicted by the phase diagram. The transformation may occur through4LITERATURE REVIEWeither diffusion or martensitic reaction; the diffusion of interstitial and substitutional atoms ordiffusionless transformation, respectively [8]. These two modes are often distinguished from eachother by the velocity of the interfacial front, the extent of transformation as functions of temperatureand time [9], and the composition profile around the interface. Several particular microstructurescommon in as-received and HAZ microstructures in the steels being studied are discussed in thefollowing sections.Figure 2.2: Continuous-cooling transformation diagram for X70 line pipe steel following heating to1300°C for variety of cooling rates; F, P and B are ferrite, pearlite, and bainite respectively [10].2.1.1 FerriteFerrite begins nucleating at relatively low undercoolings from Ae3, growing from austenite graincorners due to the favorable nucleation conditions. The growth of ferrite is controlled by carbondiffusion, the intrinsic mobility of the γ/α boundary, and further substitutional alloying additions(eg. Mn, Nb, Mo). The morphology of ferrite formed at low undercoolings is equiaxed and thegrains have a low dislocation density. The equilibrium ferrite fraction is dependent on the carboncontent, and following completion of the ferrite transformation, the remaining austenite transformsto pearlite, a lamellar structure of ferrite and cementite. As undercooling decreases further belowthe Ae3 temperature, the driving force for the formation of ferrite increases and ferrite begins toform at alternative nucleation sites such as austenite grain edges and faces [11].5LITERATURE REVIEWAs undercooling decreases still further, ferrite begins to nucleate from austenite grain interi-ors, growing from the nucleation sites in random crystallographic orientations [12]. The resultingmicrostructure is known as acicular ferrite [13]. In this microstructure, metallography shows thatgrain boundaries are irregular, the grain size is considerably finer and the dislocation density ishigh [14] [13]. Irregular ferrite is a highly desirable microstructure for line pipe steels, posessing anexcellent combination of strength, toughness, corrosion resistance and weldability [15], propertiesthat follow from its low carbon content, fine grain size, high number of high angle grain boundariesand the interlocking nature of the ferrite plates; all of which inhibit crack propagation.In addition to undercooling, the prior austenite grain diameter D has been shown to have astrong effect on the morphology of the subsequent ferrite following austenite decomposition. Asferrite preferentially forms at grain boundary corners, edges and faces when intragranular sites arenot present, a higher density of grain boundaries will result in a higher density of ferrite nuclei,with site density proportional to 1/D3, 1/D2 and 1/D depending on nucleation at corners, edgesor faces respectively. Therefore, prior austenite grain size refinement is important to subsequentferrite grain size refinement [16].2.1.2 BainiteBainite forms at intermediate temperatures, between high temperature transformation productssuch as ferrite or pearlite and low temperature transformation products such as martensite [11]. Thenature of the transformation and morphology of bainite has been a strongly debated subject sinceits initial classification by Davenport and Bain in 1930 [17]. The initial observations were formedfor isothermal treatments at intermediate temperatures, although bainite is commonly reportedin continuous cooling. In general, there are two schools of thought on the formation of bainite.The first school argues that formation is through completely diffusionless shear; the second schoolhypothesizes ferrite nucleates and grows through short-range (paraequilbrium) diffusion [11]. Morein-depth discussion on the arguments in favor of either school can be found elsewhere [17] [18] [19].The morphologies of bainite are commonly categorized as either upper or lower bainite. Eventhough the two were originally observed as microstructures forming in isothermal holding, thegeneral description has been helpful as a standard. Both morphologies consist of aggregates offerrite plates (or laths), referred to as sheaves, nucleating at prior austenite grain boundaries and6LITERATURE REVIEWFigure 2.3: Illustration of upper and bainite formation and morphology [20]growing inward. In upper bainite, which forms at higher temperatures [11], carbon in the saturatedferrite plates is partitioned into the remaining austenite, forming carbides between each ferriteplate. In lower bainite, which forms at lower temperatures, carbon cannot completely partitioninto austenite and therefore carbides precipitate both between and inside formed ferrite plates.This is illustrated in Figure 2.3. Bramfitt and Speer devised a morphology classification system forcontinuously cooled bainite [21], as they observed that microstructures showed substantial differenceeven with relatively constant transformation temperatures. Morphologies were separated into typesI (granular, or irregular ferrite and M/A), II (upper bainite), and III (lower bainite) depending onthe cooling path and the subsequent location of carbide. They also clarified that either cementite,austenite, or martensite may form between or within ferrite plates. The latter two are possiblein steels with higher Si or Al contents (which destabilize the cementite phase) and in low carbonsteels, such as HSLA steels.The effect of prior austenite grain size on the transformation behaviour of bainite has varieddepending on the particular study, although in general it is found that finer austenite grain sizeslead to more rapid transformation behaviour [22] and higher transformation temperatures. Singhand Bhadeshia found the plate thickness in bainite decreases with transformation temperature [23]7LITERATURE REVIEWand that the subsequently finer spacing results in a significant improvement on mechanical prop-erties [24].2.1.3 Martensite-AusteniteMartensite-austenite (or M/A) is a mixed microstructural feature consisting of high carbonmartensite and austenite retained following austenite decomposition. Martensite is a body-centeredtetragonal (BCT) microstructure formed from carbon-enriched austenite through displacive shear.During continuous cooling of low carbon steels, sections of austenite located between ferrite lathsbecome isolated and carbon is partitioned to them. As the martensitic reaction does not go to100% completion, a certain amount of austenite is retained in the final microstructure. As a result,in low carbon steels it is common to observe M/A in bainite in lieu of cementite [25]. A high degreeof shape change occurs during M/A formation and strain is produced in the structure, creating alarge dislocation density [8]. The Ms, or the temperature at which martensite is formed, can beestimated using empirical formulas based on steel chemistries. Ishida [26], for example, suggeststhe following phenomenological formula, where concentrations are in wt%Ms = 545 - 71C + Al + 7Co - 14Cr - 15Cu - 23Mn - 8Mo - 6Nb - 13Ni - 4Si + 3Ti - 4V (2.1)The morphology has been studied across a range of continuously cooled samples by Reichertet.al [27] for a range of prior grain sizes and cooling rates and found in several configurations:block-like ’islands’ 0.5 to 3 µm in diameter; elongated ‘necklace’ particles 5 to 10 µm in lengthand 0.5 to 2 µm in width; connected or nearly connected M/A particles; or a dual phase mixtureof M/A connected with carbide. A phenomenological model was developed for an X80 steel topredict the fraction of M/A formed as a function of transformation start temperature [27]. It wasfound that the amount of M/A formed is maximized at intermediate temperature ranges, whenupper bainite is formed. Additionally, at intermediate transformation start temperatures, it wasfound that the ‘necklace’ morphology takes place along prior austenite grain boundaries, which isimportant to fracture properties.The effect of M/A on bulk mechanical properties is still under debate. Martensite has tradition-8LITERATURE REVIEWally been reported as having a hard, brittle microstructure [8] and therefore it is intuitive that thepresence of M/A would lead to lower toughness. However, Ohya reported that fracture sites weremost associated with bainitic ferrite plates posessesing different crystallographic orientations andthat M/A was not likely to embrittle HAZ microstructures [28]. On the other hand, Li found thatM/A had a strong effect on fracture behaviour [29]. There are several reasons why M/A could havean adverse effect on the mechanical properties: for example, M/A could act as a crack nucleationpoint due to its brittleness or as a stress concentrator in the surrounding matrix [30].2.2 Development of Line Pipe SteelsLine pipe steels have traditionally included both plain carbon and microalloyed steels for pro-duction. The latter are used to meet the design requirements of more modern pipelines, as the needfor steels with increasingly enhanced mechanical properties rises. Gladman published a comprehen-sive review on common alloying elements and their effects in steelmaking [31] Alloying is often usedto enhance the properties of steel, including carbon, manganese, copper, chromium and silicon.However, the increase in elements such as carbon increase the hardenability of steel, potentiallyresulting in cracking following processing or welding. Line pipe steels, therefore, posess low carboncontents (0.03 - 0.10 wt%), manganese contents of 0.3 - 2 wt% and a total microalloying level of<1% [31]. Microalloying elements include niobium, vanadium, molybdenum, boron and titanium,referred to by their significant contributions to steel properties even when present as small fractionsof the composition.The primary goal of thermo-mechanical processing (TMCP) of microalloyed line pipe steelsis the formation of fine grained ferrite in conjunction with fine microalloy (V, Nb, Ti, Mo, etc.)precipitates [32] [33]. TMCP is carried out in four stages, illustrated as an example in Figure2.4. The first stage following reheating consists of rough rolling where the steel is rolled abovethe recrystallization temperature to reduce the austenite grain size through static recrystallization;during this, new strain free grains are formed. The second stage consists of finish rolling, wherethe steel is rolled below the recrystallization temperature but above the Ae3 temperature. Thisproduces ‘pancake’ shaped austenite grains with deformation bands in the grain interior, which actas further nucleation sites for ferrite [16]. In the third stage, the steel is continuously cooled with9LITERATURE REVIEWhigh pressure water jets to refine the ferrite grain size further by achieving higher undercooling forincreased ferrite nucleation sites [6]. Continuous cooling also develops transformation strengtheningvia the formation of acicular ferrite and bainite. Following this stage, the coil of steel is slowlycooled, during which time precipitates form in ferrite to provide additional strengthening. Followingcoiling, the steel is formed into pipes by cold forming the plate into a helix and using submerged-arcwelding (SAW) along the seam, a process known as spiral welding [34]. With larger diameter pipes,such as those used in offshore developments, an alternative process known as UOE is employed,named after the latter three steps in the method. Plates are first crimped by dies along the edgesinto circular arcs, then formed into a U-shape by rollers in three-point bending (U-pressing). Theplate is then pressed into a near O-shape (O-pressing) and welded with SAW along the seam insideand outside the pipe to form it. Finally, the pipe is mechanically expanded from the interior toproperly size and improving its circularity. [35]Figure 2.4: Example of a rolling schedule for TMCP of steel10LITERATURE REVIEW2.2.1 The Role of NiobiumNiobium was first used as a microalloying agent in steels in 1958, and the first symposium heldto discuss the effect of niobium in a wide range of steels was held in 1981 [36]. Since then, niobiumhas been a common alloying and microalloying addition in a variety of steels.Niobium in carbon steels forms carbonitride (Nb(C,N)) precipitates during finish rolling andcoiling in TCMP processing [37]. As seen in Figure 2.5, niobium is less soluble in austenite thanvanadium and molybdenum and more soluble than titanium. In addition, niobium carbides andnitrides have varying solubilities, the latter having lower solubility. There is a strong interactionbetween Nb and Ti microalloying additions. For example, titanium preferentially bonds with nitro-gen to form titanium nitride (TiN) and the remainder of nitrogen will be distributed into Nb(C,N).Similarily, if not enough nitrogen is present, titanium remaining after TiN formation will prefer-entially bond with carbon to form titanium carbide (TiC) before the remainder of the carbon willform NbC, although in general this is usually avoided due to TiC generally forming coarse precip-itates. When a large variety of microalloying elements are present, complex carbonitrides, such as(Nb,Mo)(C,N) or (Ti,Nb)(C,N) may form, as the elements have solubility with each other’s respec-tive precipitates [38]. Because of this, a large range is present in the calculation of the solubilityproduct of niobium precipitates in the literature, and this is reflected in recent software [36] [39].Gladman suggests [31], in the case of stoichiometric niobium carbide, the solubility product islog([Nb][C]) = 2.26− 6770T(2.2)where [Nb] and [C] are the respective concentrations of niobium and carbon in solution in austeniteand T is the temperature in K. In the study of Martin et. al [40], however, the equation for thesolubility of niobium carbonitrides is modified with the concentration of nitrogen [N] to yieldlog([Nb][C +12N14]) = 2.26− 6770T(2.3)As an alternative to empirical solubility products, recent CALPHAD (Computer Coupling ofPhase Diagrams and Chemistry) software (e.g. Thermocalc, Matcalc, OpenCalphad, etc.) hasbeen highly useful in efficient and realistic prediction of the thermodynamic stability of coexisting11LITERATURE REVIEWphases. These programs work on the principle of minimization of experimentally developed Gibbsfunctionals represented at user-defined temperatures, pressures and compositions. When energyis minimized throughout the entire system (known as global minimization), this results in overallequilibrium of the system, and from this phase fraction and composition can be calculated [41].Figure 2.5: Formation enthalpies of common microalloy precipitates in steels [42]Dissolution and precipitation of Nb(C,N) may be studied through several methods, includ-ing transmission electron microscopy [43], electrolytic dissolution [44], electrical resistivity [45] andthrough its effect on transformation temperature [46]. The rates of Nb(C,N) precipitation are rapidin ferrite, even at low (500°C) temperatures [36]. Thus, coiling at temperatures where ferrite isstable results in a large number of fine precipitates due to the high driving force and relatively rapidformation rates. The rates of niobium precipitation are slow in unstrained austenite due to preci-pation occuring heterogeneously at dislocations and subgrain boundaries [47]. Precipitation ratesare far more rapid in strained austenite, as illustrated in Figure 2.6, as shear bands are produced,allowing for an increased number of precipitation nucleation sites [48]. The kinetics can be affectedfurther by other alloying elements contained in the steel. For example, manganese has been shown12LITERATURE REVIEWto retard precipitation by increasing the solubility of niobium in austenite [49]. Molybdenum hasbeen shown to quicken precipitation due to its solubility within the carbontride [48] [47].Figure 2.6: Precipitation-time-temperature (PPT) diagram highlighting the effect of strain on pre-cipitation start (Ps) and finish (Pf ) times on a Nb-V-Ti steel measured by electrical resistivity [45]The state of niobium as either a precipitate or in solution will lead to Zener pinning or solutedrag, respectively. Cahn studied the phenomena of solute drag in a variety of materials [50]. Thepremise of the theory is that in a mixed system, a solute profile gradient exists in advance of amoving interphase boundary. This results in a drag pressure on the moving boundary, delayingits movement. The effect is highly non-linear depending on the velocity of the boundary and thebehavior is separated into separate regimes. At high solute concentrations or low driving forces, theboundary drags the solute along and the solute drag effect is large. At low solute concentrations orhigh driving forces, the boundary achieves ‘breakaway’ from the solute and the solute drag effect issmall. It was found that solute drag is proportional to grain size when the impurity preferentiallysegregates to boundaries [51], the effect increasing when grain sizes are large.Microalloying steels with niobium provides two primary advantages. First, the addition of nio-bium has been shown to impede static and dynamic recrystallization [52] [53]. This takes placethrough a combintation of precipitate pinning when Nb is precipitated as carbonitride and solute13LITERATURE REVIEWdrag when Nb is in solution. Zurob et. al [54] developed a numerical model to predict recrystal-lization behaviour in Nb-steels with coupled solute drag and Zener pinning. Their model predictsthat the addition of 0.1% Nb completely inhibits recrystallization at 1000°C for 2000 s. Literaturegenerally indicates that precipitates are more effective than solute in the inhibition of recrystal-lization [16]. However, a numerical model by Hutchinson et al. [55] predicts that that niobiumin solution is more effective on impeding recrystallization in ferrite and at lower temperatures inaustenite than in precipitate form. Further, Figure 2.7 shows that the recrystallization-stop tem-perature, the temperature below which recrystallization does not occur for a given holding time, ishighly increased by niobium compared to other common microalloying elements. This allows forhigher finishing rolling temperatures while still impeding recrystallization, subsequently allowingfor reduced rolling loads with no reduction in properties.Figure 2.7: Experimentally measured recrystallization stop temperatures of steels as function ofmicroalloying content [31]Niobium has also been shown to inhibit grain growth in line pipe steels at elevated temperatures[56] [57]. The kinetics of grain coarsening were studied for isothermal holding in various HSLA steelsbearing a number of microalloying precipitates [58]. Niobium had a significant effect on refining theaustenite grain size, increasing the grain-coarsening temperature by as much as 200°C with additionsof 0.11 wt%. Non-homogeneous grain size distributions were found between 1100 and 1150°C aftertwo-hour holding times, indicating the onset of abnormal grain growth. This phenomenon is present14LITERATURE REVIEWin materials containing second phases where the partial dissolution of precipitates can occur [59]and has been observed in reheating of HSLA slabs in conjunction with inhomogeneity in precipitatevolume density [60].Grain growth is driven by the subsequent reduction in grain boundary energy via the reductionin grain boundary area. A number of models have been constructed for the prediction of graingrowth behaviour in the presence of a second phase. Zener’s solution is a simple numerical modelwith good effectiveness at the prediction of grain growth [61] [57]. The velocity of a movingboundary v is given byv = M∆P (2.4)where M is the mobility of the boundary and ∆P is the driving pressure moving the boundary andthe rate of grain growth dDdt = 2v, where D is the grain diameter. The driving pressure for graingrowth is proportional to the curvature of the grain, given by∆Fg =γαD(2.5)where γ is the grain boundary energy, and α is a geometric constant. For three-dimensional graingrowth, α = 4. This pressure is reduced by the Zener pinning pressure provided by second-phaseparticles such as microalloy precipitates. In the case of spherical precipitates,∆Fp =fγβr(2.6)where f is the current volume fraction of a particular precipitate family and r is the radius of theparticular precipitate. When n precipitate families are present, for example in an Nb-Ti steel, thetotal pinning pressure is the sum of each precipitate group’s pinning pressure. The mobility ofgrain boundaries is commonly given as an Arrhenius equation,M = M0e−QRT (2.7)such that15LITERATURE REVIEWdDdt= Moe−QRT γ(αD−n∑i=1fiβri) (2.8)It can be observed from the equation that when the driving pressure term is equal to the pinningpressure term, grain growth will halt indefinitely for a given temperature. This is known as thelimiting grain size, whenDlim =αn∑i=1fiβri(2.9)This solution assumes coarsening of precipitates does not occur at high temperatures i.e. r is aconstant. Some low solubility precipitates, such as TiN, may coarsen over time [62] and the limitinggrain size may thus increase slightly over time for extended holding times at high temperatures.Secondly, niobium has been shown to strongly affect austenite decomposition via the reduction intransformation rates [43] [63] and depression in transformation temperatures [16] [46]. This resultsin finer ferrite grains, production of acicular ferrite [64] and bainitic microstructures. As stated inSection 2.1.1, irregular ferrite in literature is a highly desirable microstructure and it is of interestto produce this in line pipe steels. Although large amounts of niobium can thermodynamically actas ferrite stabilizers [65] to raise transformation temperatures, the amount of niobium present inline pipe steels is <0.1 wt% and thus would not have an extensive thermodynamic effect to this end.Rather, niobium in solution acts to depress transformation temperatures kinetically, likely throughsolute drag. For example, the effect of niobium in solution on transformation temperatures werestudied for a range of niobium contents in a low carbon steel following simulated finish rolling [66], asshown in Figure 2.8. The ferrite and bainite start temperatures were reduced as niobium contentincreased, though the effect saturates beyond a certain concentration of Nb. There exists somedebate in literature of the precise mechanics of how niobium delays austenite decomposition. Forexample, atom probe microscopy by Maruyama et. al [67] showed that niobium prefers to segregateto grain boundaries. It was theorized that in this state, niobium would reduce the surface energyof the boundary, reducing the viability of ferrite nucleation [68]. Alternatively, niobium could havean interaction with carbon, reducing the rate at which it partitions from the growing ferrite toaustenite.16LITERATURE REVIEWThe role of microalloy precipitates has a less well defined role on austenite decomposition. Itwas found that inclusions formed in the welding of a low carbon steel act as nucleation sites forthe formation of acicular ferrite during austenite decomposition [69]. This would imply inclusionsand precipitates act to raise the transformation temperatures and promote ferritic microstructures.However, microscopy showed that acicular ferrite only forms on inclusions with diameters greaterthan 200 nm, as larger inclusions provide more favorable nuclation conditions due to the reductionin critical nuclei size [70]. Given that rolling schedules are designed to form niobium microalloyprecipitates far smaller (<50nm) [71] in line pipe steels, it does not seem likely they act as effec-tive nucleation sites, though it is possible that coarser precipitates such as TiN could act in thismanner. It is possible that niobium precipitates act to slow interphase boundaries during austenitedecomposition via pinning pressure in the same manner as grain growth. However, little conclusiveevidence has been shown in literature to confirm this. For example, Brechet et. al [52] did notobserve an effect from niobium precipitates on interphase mobility during ferrite formation.640660680700720740760780T s,˚C0 1 2 3 4 5 6 7 8 9 10Nb content, 10-2wt%5 ˚C/s10 ˚C/s15 ˚C/s20 ˚C/s580600620640660680700720T s,˚C0 1 2 3 4 5 6 7 8 9 10Nb content, 10-2wt%5 ˚C/s10 ˚C/s15 ˚C/s20 ˚C/sb)a)Figure 2.8: Effect of niobium in solution on a) ferrite start temperature and b) bainite starttemperature in a low carbon (0.021%C) steel following simulated finish rolling, data from [66].17LITERATURE REVIEW2.3 Girth Welding of Line PipesIn the following, a summary will be presented on the application of the physical metallurgy ofmicroalloyed steels to pipeline construction.To construct a pipeline, the lengths of pipe are welded end to end in the field. Gas-metal arcwelding (GMAW) is the most common method of welding in pipeline construction. GMAW iscommonly performed with a single weld torch in conjunction with inert gas protection to stabilizethe arc and protect the weld from oxidation [70]. Submerged arc welding (SAW) is a commonlyused variant where flux is deposited on the weld surface in advance of the welding torch, producinga higher heat input. This allows for slower cooling rates and further penetration of heat intothe workpiece [42], although extended grain growth behavior is present for the longer times atelevated temperatures [72]. A welding pass can be thought of as a thermal cycle consisting of arapid heating to a peak temperature followed by an immediate cooling to ambient temperatures.The peak temperature of a location in the line pipe near the welding torch and cooling rate aredependent on a number of parameters including heat input, distance from torch tip and the velocityof the torch. Detailed solutions for heat flow in welding and the shape of heat affected zones aredescribed in the literature [42].A large proportion of the cost in pipeline production is due to welding and non-destructiveinspection of pipes, such that an increase in welding productivity significantly reduces the totalcost. To increase the productivity of welding, tandem wire or dual torch GMAW may be used byrequiring less total passes. In the former, a second electrode feeds into the same weld pool as thelead torch and in the latter the second electrode is set 50 to 200 mm behind the first to producetwo distinct weld pools. If spacing is sufficiently wide, multiple temperature peaks are reached.Measurements of cooling rate and peak temperatures reached in a pipe section during a weldingscenario have been made by installing thermocouples at selected distances from a weld line prior toa welding scenario and measuring the temperature during welding [73] [74]. An example of coolingrates as functions of peak temperature is shown in Figure 2.9. In the context of welding, coolingrates are often expressed as the measured time during which the steel cools from 800°C to 500°C, orthe ∆t8−5; the average cooling rate may be derived from this. From the figure, it can be observedthe cooling rates are lower in the trail torch due to the increased heat input.18LITERATURE REVIEWFigure 2.9: Peak temperature and related t8−5 measurements following single torch and dual torchwelding [74].2.3.1 Heat Affected ZoneDuring welding, a temperature gradient exists through the material extending from the weldpool of the material and spanning outwards. The heat affected zone (HAZ) pertains to the regionof a material during welding whose peak temperatures are not high enough to transform it toliquid state, but whose properties and microstructure are affected by the heat input [42] [70]. Thishas multiple definitions depending on the material considered. In the case of steels, the bordersof the heat affected zone are regarded as slightly below the Ae1 temperature on the Fe-C phasediagram and the melting temperature. The HAZ can be divided into several subsections to bedescribed in brief and displayed in Figure 2.10. The coarse-grained heat-affected zone (CGHAZ)lies just below the melting temperature (i.e. directly adjacent to the weld pool) and, roughlyspeaking, above the temperature at which precipitates dissolve (e.g. Nb(C,N)). As precipitatescoarsen and dissolve, austenite grain boundaries are unpinned and grow to large sizes. Combinedwith the effect of niobium in solution on reducing transformation temperatures, microstructuresformed in this region are commonly martensite or bainite, depending on the particular weldingparameters and steel. Consequently, the highest values of observed hardness will be formed in the19LITERATURE REVIEWCGHAZ. The fine-grained heat-affected zone (FGHAZ) is located above the Ac3 temperature (thatis, the temperature at which austenite is formed upon heating, a few degrees higher than Ae3 [75])and below the temperature at which precipitates dissolve. Although the reversion to austenite iscomplete, grains are pinned at small sizes and this favors the formation of ferrite and acicular ferrite.The intercritical heat-affected zone (ICHAZ) is located below the Ac3 temperature and above theAc1 temperature. In this region, the base microstructure partially transforms to austenite anddecomposes upon cooling. Although fine grained ferrite is formed, carbon enriched austenite ispartially formed, which can lead to substantial fractions of M/A after cooling [76]. Below theAc1 lies the subcritical heat-affected zone (SCHAZ); effects in this region are less pronounced asaustenite is not formed and decomposed; however, high carbon regions of the microstructure, suchas M/A, may be tempered and microalloy precipitates, such as molybdenum carbide (Mo2C) maybe dissolved or precipitated.Figure 2.10: Microstructures forming in a weld area for a single welding pass and correspondingFe-C phase diagram [42]20LITERATURE REVIEW2.3.2 Welding of Nb-based HSLA SteelsAs stated in Section 2.2.1, niobium is typically precipitated as Nb(C,N) after hot rolling andcoiling of steels. Given the high rates of heating in welding, the prediction on the dissolution ofbehavior precipitates is likely more complex than those predicted by equilibrium calculations. Assuch, numerical models have been constructed to predict the dissolution temperature of precipitatesduring heating with a combination of thermodynamics and kinetics [42] Alternatively, precipitatesmay undergo coarsening at elevated temperatures, reducing their effectiveness at pinning grainboundaries by reducing the r term in the Zener equation (i.e. Equation 2.8). Heating rate has beenshown to have a strong effect on the grain growth behaviour in continuous heating [71] and duringisothermal holding following reheating [77]. Non-isothermal grain growth was reviewed extensivelyby Mishra and Debroy [78] on both experimental work and a variety of modeling approaches.In addition to its effect on grain growth, niobium once dissolved in solution will lead to lowertransformation temperatures due to solute drag. In combination with large austenite grain sizes,this will lead to the formation of upper and lower bainite, significantly changing the mechani-cal properties of the HAZ. However, it is possible that niobium precipitates that are dissolvedupon heating will re-precipitate following cooling in welding scenarios with lower cooling rates (e.g.<10°C/s). For example, Fossaert showed that the cooling rate in the intermediate range (980to 750°C) had a significant effect on the critical cooling rate following this to form a martensiticmicrostructure [79]. The M/A content has been studied with metallography in a number of simu-lated CGHAZ from steels with a range of C and Nb contents by El-Kashif and Koseki [80]. Theyreported that an increase in C and Nb resulted in a higher fraction of M/A following austenitedecomposition. However, the niobium content was limited to a maximum of 0.03% and thereforeis not representative of behaviour at higher Nb contents (0.09%).The literature reports mixed interpretations on the effect of Nb on HAZ mechanical properties.Hattingh and Pienaar reported that Nb had beneficial effects on toughness at intermediate (0.03- 0.06%) Nb contents for lower heat inputs in simulated HAZ microstructures but had negativeeffects on toughness in higher heat inputs [81]. This was attributed to grain refinement in theformer cases and additional hardenability from niobium in solution in the latter. Further, hardnessmeasurements on an X80 steel were shown to have high dependency on prior austenite grain size21LITERATURE REVIEWwhen niobium was in solution [82]. Consequently, it is possible that when niobium dissolves in theCGHAZ, the subsequent microstructures have higher strength and lower toughness, but the twoare not necessarily directly correlated. Wei et al. researched the mechanical properties of simulatedwelded HAZ microstructures of two high (0.07 and 0.10%) Nb steels and the higher Nb steel wasshown to have higher tensile strength and lower toughness for all peak temperatures and heat inputsstudied [83]. Simulated welds were produced to study the effect of niobium on HAZ toughness. Asfound, increasing Nb to 0.08% resulted in a slight loss of HAZ toughness measured in Charpy andCTOD tests. A post weld heat treatment increased the hardness extensively, though details on thetreatments were not mentioned [84]. It was speculated by the author that the increase in hardnesswas due to precipitation hardening. Similar results were found by El-Kashif and Koseki [80].22Chapter 3Scope and ObjectivesIn this study, the experimental work was conducted on two line pipe steels, the compositions ofwhich are listed in Section 4.1. Austenite grain growth trials were conducted in conjunction withlaser ultrasonics for metallurgy (LUMet) to study in-situ grain growth as a function of heating ratein continuous heating and isothermal holding temperature. These measurements were compared toex-situ metallography. Niobium re-precipitation from solution was characterized by the comparisonof austenite decomposition kinetics during continuous cooling as a function of holding time at aselected aging temperature. CCT tests were conducted in a range of 3 to 30°C/s and the kineticsof austenite decomposition were measured using dilatometry. Microstructures were observed withoptical microscopy and characterized by microhardness measurements.The objectives of this research project are as follows:1. To quantify austenite grain growth in two high Nb line pipe steels (0.028%C, 0.091%Nb and0.058%C, 0.091%Nb)2. To determine the kinetics of niobium precipitation at 900°C by measuring transformationstart temperature3. To construct CCT diagrams accounting for the effect of cooling rate, prior austenite grainsize, carbon content, and niobium solute content on austenite decomposition23Chapter 4Methodology4.1 MaterialsThe steels studied were supplied by EVRAZ Inc. NA with the composition shown in Table4.1. The as-received material was present in the form of a section of pipe from which samples weremachined in the longitudinal direction. Initial metallography on the steels show that both compo-sitions are present as fine irregular ferrite with fine M/A islands, as shown in Figure 4.1. Previoustransmission electron microscope (TEM) studies on a similar steel showed that the microalloyingelements are precipiated as TiN, MoC, and Nb(C,N) particles [71].Alloy, Wt% C Mn Si Nb Ti Mo N S FeL-294L 0.058 1.65 0.26 0.091 0.016 0.292 0.0076 0.0009 BalanceL-294M 0.028 1.70 0.31 0.091 0.017 0.297 0.0078 0.0019 BalanceTable 4.1: Given composition of studied alloysSamples for the experiments were machined in one of two geometries; the first was a rectan-gular strip geometry used for the grain growth kinetics and re-precipitation experiments, and thesecond was a hollow tubular geometry used for continuous cooling experiments to study austenitedecomposition. The dimensions of the samples are shown in Figure 4.2.24METHODOLOGY(a)(b)Figure 4.1: As-received microstructures of selected compositions a) 0.028%C b) 0.058%C4.2 Grain GrowthHeat treatment experiments were conducted using the Gleeble 3500 thermo-mechanical simu-lator under high vacuum ( 1E-3 Pa) atmosphere with pre-programmed thermal histories i.e. usingresistive heating and natural cooling or quenching with He gas, the temperature of the samplecan be regulated via thermocouple feedback and high heating (up to 1000°C/s) and cooling (upto 300°C/s) rates can be achieved. For 1150 and 1250°C isothermal grain growth trials, S-type(Pt/Pt-10 % Rh) thermocouples were used for their stability at high temperatures for extendedtime periods; the remainder of the trials used K-type (chromel/alumel) thermocouples for ease ofuse and response time.25METHODOLOGYFigure 4.2: Sample dimensions for bar specimens a) tubular specimens b)4.2.1 Laser UltrasonicsChapter 2 summarized the importance of austenite grain size on the austenite decompositionbehavior. Austenite grain size is difficult to measure using traditional metallography, due to the ex-tensive time required for analysis, errors introduced from under/overetching of the microstructure,and measurement bias. In the current study, laser ultrasonics will be used as an additional tool tostudy the grain growth behavior. The following provides a brief review of the current literature.Laser ultrasonics is a novel, non-destructive technology for rapid, accurate in-situ observationsof metallurgical processes, in particular for this study, the austenite grain growth [85] [86] [87] [88].In the case of grain growth, the measurement is made by the study of the attenuation of theultrasound signal. Ultrasound signal attenuation through a polycrystalline metal is controlled byseveral phenomena, including scattering by internal inhomogenities (most notably grain boundaries,but also inclusions, precipitiates, etc.), absorption through internal friction and diffraction. Theattenuation αTot of a signal can be represented byαTot = αSc + αIF + αD (4.1)where αSc is attenuation from scattering phenomena, αIF is scattering due to internal friction, and26METHODOLOGYαD is attenuation due to signal diffraction. αSc has been found to vary with the ratio of grainsize to signal wavelength, and therefore can be used to find the average grain size of a metallicsample through analysis of the attenuation of a signal. A sample configuration of laser ultrasonicsapparatus is displayed in Figure 2.11. An ultrasound signal is generated in a sample by the ablationof a surface layer by a high energy pulse laser. The ablation produces an ultrasound wave that passesthrough the material and reflects off either surface several times. A lower energy and frequencylaser aimed at the spot location is used to monitor the displacement on the surface of the sampleafter the initial pulse. The signal waveform produced will include multiple echoes of successivelydecreasing amplitudes. To measure the attenuation due to grain size, the amplitudes of eithertwo successive echo numbers of the same waveform or the same echo number of waveform and areference waveform are compared. As amplitude A is dependent on a range of collected frequencies,this is usually represented as a spectrum A(f). The total attenuation α of a signal passing from apoint x1 to a successive point x2 is defined as:Figure 4.3: Experimental setup for laser ultrasonics [57]αTot(f, T ) =20x2 − x1A(x1, f, T )A(x2, f, T )(4.2)where A(xn, f, T ) is the amplitude spectra being recorded. The complexity of the solution of thisformula can be reduced by several assumptions. First, αD is a complicated parameter to evaluate,27METHODOLOGYdependent on frequency, generation/detection area, velocity and propagation distance. This canbe removed empirically by using the single echo method i.e. comparing the same echo numberfrom a current and reference signal of the same sample. This sets the propagation distance (samplethickness) and velocity as equivalent between current and reference, as well as generation/detectionarea if the lasers have equivalent spot size. In this case, diffraction is equivalent between the currentand reference samples. Second, it is useful when the reference sample has a fine grain size, reducingαSc in the reference signal to 0. Third, temperature has a linear effect on the amplitude (as theshape of the waveform is not influenced by temperature, assuming the two amplitudes measuredare from the same sample) when the wave is at x = 0. Finally, αIF can be assumed to contributea frequency-independent amount to the total attenuation. From this, Equation 2.11 becomes20xA(∆x, f, Tref )A(∆x, f, T )=20log(c)∆x+ αIF (T ) + αSc(f, T ) (4.3)where ∆x is the total propagation distance of the waveform between echoes (in the case listedin Figure 2.11 where generation and detection are on the same surface, this is equal to twice thesample thickness). This can be simplified toαTot = a+ bfn (4.4)whereb = K(T )Dn−1 (4.5)and K(T) is a parameter dependent on temperature and texture, D is the average grain size and nis a parameter dependent on the wavelength to grain size ratio, usually taken as 3 in experiments.A material must be calibrated in order to find a and b at multiple D values to generate a fit curveto be used for analysis. n is found to vary between 0 and 4, depending on the ratio of signalwavelength to grain size. An example ultrasonics calibration of steel from literature is shown inFigure 2.12.28METHODOLOGYFigure 4.4: Calibration of b-parameter from experimental grain size measurements [86]4.2.2 Grain Growth TestsIn order to study the austenite grain growth kinetics of the two steels, thermal histories weredesigned to observe the grain growth kinetics via laser ultrasonics. Ultrasound waveforms in thematerial were produced by a short (5ns) energetic (150mJ) Nd:YAG laser pulse with a green(532nm) wavelength; the reflection was detected using a laser interferometer with a longer (50 us)less energetic (70 mJ) Nd:YAG laser with an infrared (1064nm) wavelength. The second echo inthe signal is selected for purposes of analysis.Two separate methodologies were taken for studying grain growth kinetics; the first set of ex-periments were used to study the evolution of grain size during continuous heating, shown in Figure4.3. Samples were heated in the Gleeble at either 10, 100 or 1000°C/s from ambient temperaturesto 1350°C and held for 1 s before quenching with He gas (100 - 200°C/s) to ambient temperatures.During heating, laser ultrasonics were used at an acquisition rate between 10 and 50Hz to trackthe evolution of grain size during heating. The second set of experiments were used to study theevolution of grain size during the heating to and holding at an isothermal temperature to study thelimiting grain size achieved at a number of different temperature, shown in Figure 4.4. Sampleswere heated at 100°C/s from ambient temperatures to either 950, 1050, 1150, and 1250°C.29METHODOLOGY˚˚˚˚Figure 4.5: Thermal history for continuous-heating grain-growth trial˚˚˚˚˚Figure 4.6: Thermal history for isothermal holding grain-growth trial30METHODOLOGYFollowing this, the samples were held for between 15 and 25 minutes to approach a limiting grainsize before being quenched with He gas to ambient temperatures. Laser ultrasonics were used at anacquisition between 50 and 0.5 Hz to realistically capture early and late stage grain growth. Signalprocessing and analysis was carried out through the software LUMFILE from the software packageCTOME (Computational Tools For Metallurgy) [89] developed by Dr. Thomas Garcin. Samplesproduced were characterized via metallography to verify the results from laser ultrasonics.4.3 Austenite Decomposition4.3.1 DilatometryDilatometry is an indirect method to study austenite decomposition. During austenite decom-position, there is a difference in the molar volume of the parent and daughter phases due to thediffering atomic structure (ie. FCC vs BCC). Therefore, by measuring the dimensional change ofsteel during this, one can observe and quantify phase transformations. In context of a transfor-mation during continuous cooling of steel, the daughter phase may be any combination of ferrite,bainite, or martensite. Dilation can be physically measured with several methods, but the mostconvinient is to use a mechanical transducer attached to the sample during a phase transformation;the rods touching the sample have a low enough coefficient of thermal expansion (CTE) com-pared to the sample being measured, giving as little effect on measurement as possible via theirown expansion and contraction. It is worth noting that carbon redistribution between ferrite andaustenite will alter the molar volume of the latter as the transformation progresses, altering themeasurement [90]. This is most evident in higher carbon steels.An example of dilation plotted against temperature can be seen in Figure 2.13. As the sampleis cooled while in the austenite phase field, the material continuously contracts at a linear rate withtemperature according to the coefficient of thermal expansion. Upon nucleation and growth ofthe daughter phase(s), the molar volume changes in proportion to the amount of each phase in thesample. When completely transformed, the material again contracts at a continuous linear rate withtemperature, although at a different rate due to the different CTE. The degree of transformationin continuous cooling can be extracted by defining the linear portions of the curve. From this, thevolume fraction of transformed phase X can be inferred from the amount of deviation from either31METHODOLOGYinterpolated line on the curve, known as the tie line equation, where x(T ) is the dilation of thesample at temperature T and xγ(T ) and xα(T ) are the inferred dilations of austenite and daughterphases, respectively. Therefore,˚x(T)x(T)x(T)Figure 4.7: Schematic dilatometry curve of austenite-to-daughter phase transformationX(T ) =x(T )− xγ(T )xα(T )− xγ(T ) (4.6)The transformation start (Ts) and transformation finish temperatures (Tf ) are defined as thepoints at which the curve starts and stops diverging from linearity. This is generally taken at,respectively, 5% and 95% completion of the phase transformation.4.3.2 Nb(C,N) ReprecipitationIn order to study the precipitation kinetics of Nb(C,N) in the studied alloys, thermal historieswere designed in a method to study the precipitation kinetics in the two steels. As stated in Section32METHODOLOGY2.2.1, it is known that Nb in solution will retard the transformation of austenite to ferrite via solutedrag; therefore, if all other microstructure and processing variables are held constant, the amountof Nb in solution can be inferred from the transformation temperatures. Experiments by Fazel et.al [46] showed that in particular, the Ts of austenite decomposition could be correlated with theNb solute content in the steel following an aging period in austenite. A similar approach will betaken in this study.Heat treatments were carried out in the Gleeble 3500 and temperature monitored by K-typethermocouples. The thermal history is shown in Figure 4.5. All samples were first heated to1300°C to dissolve Nb in solution before quenching with He to minimize re-precipitation of Nb(C,N).Samples were then heated to 1250°C for 15 seconds to produce a large grain size before naturalcooling to 900°C followed by holding at 900°C for various selected times between 0 s and 45 min.It was found in the literature that 900°C is a reasonable temperature for precipitating Nb(C,N) inaustenite [45] and this was selected as an temperature to re-precipitate Nb from solution. Followingisothermal holding at 900°C, samples were naturally cooled (30°C/s) to 800°C to minimize furtherreprecipitation of Nb(C,N) before cooling at 1°C/s to ambient temperatures. Dilatometry was usedto monitor the transformation kinetics with an acquisition high enough to give 2 points per degree(2Hz). Transformation curves were developed through the software PROFILE from CTOME [89]to find the Ts.4.3.3 CCT TestingIn order to study the effect of niobium in and out of solution on austenite decomposition,thermal histories were designed in order to control grain size, niobium in solution or precipitated.and cooling rate as independent variables. It can be considered that the amount of C in the steelswould be considered a fourth parameter based on the otherwise composition similarity betweenthe two alloys. The thermal histories are shown in Figure 4.6 a) and b) for small and large prioraustenite grain sizes respectively. All samples were first heated to 1300°C to fully dissolve all Nbinto solution before quenching to prevent re-precipitation of Nb(C,N). To select either a smallor large prior austenite grain size, samples were re-heated to, respectively 950°C for 5 secondsor 1250°C for 15 seconds before natural cooling the samples to 900°C. To confirm the grain sizefor this treatment, samples were quenched following these peak temperatures and the grain size33METHODOLOGY˚˚˚˚˚˚˚Figure 4.8: Thermal history for dissolution and reprecipitation of Nb(C,N)found through metallography. Two extreme conditions were selected with regards to amount ofNb in solution; either Nb remained completely solutionized from the prior treatment or Nb wasnearly completely precipitated in austenite before austenite decomposition. This was done byeither cooling austenite directly from 900°C or by holding the sample for 30 minutes at 900°C toreprecipitate Nb(C,N) from solution followed by cooling. Grain growth is very slow at 900°C andno significant increase in grain size is expected to occur [82]. Following either of these scenarios,samples were cooled at either 3, 10 or 30°C/s to ambient conditions and dilatometry was taken tomeasure the austenite decomposition. The acquisition rate for dilatometry during cooling was highenough to give 3 points per degree (10, 30 and 100 Hz depending on cooling rate). As in Section4.2.2, transformation curves were developed through the software PROFILE. Samples producedwere charaterized by metallography.4.4 CharacterizationIn order to characterize the samples produced, metallography was conducted on grain growthsimulation and austenite decomposition samples. Samples produced from the grain growth simu-34METHODOLOGY˚˚˚˚˚ ˚˚˚˚(a)˚˚˚˚˚˚˚˚˚(b)Figure 4.9: Thermal history for austenite decomposition at a) small and b) large prior austenitegrain sizeslations were tempered in a tube furnace back filled with argon at 500°C for 24 hours to segregateimpurities to the grain boundaries to provide better contrast during etching. All samples were cut35METHODOLOGYwith a water-cooled precision-saw at the centerline of the sample and mounted with a PneumetI mounting press in green phenolic to provide better handling on the sample for the purposes ofmetallography. The mounted samples were ground on a Phoenix 4000 sample preparation systemin successive order, with 320, 600 and 1200-grit silicon carbide papers and subsequently polishedwith 6 µm and 1 µm diamond suspension to produce a flat scratchless surface.To delineate the prior austenite grain boundaries, the samples were etched at 60°C for 5 to 8minutes with an appropriate grain boundary etchant (4 parts 4% saturated aqueous picric acid,1 part Greenworks detergent). The samples were photographed a number of times with a NikonEpiphot 3000 between 200x and 1000x with enough micrographs to measure a stastically significantnumber of prior austenite grains (>300). The austenite grain boundaries were further delineatedby highlighting the boundaries by hand before scanning the outlines. ImageJ v1.48 software wasused for analysis. A threshold was applied on the outlines and an area count method was used tomeasure the area A of each grain. The process is illustrated in Figure 4.7. From this, the equalarea diameter D can be found by simple geometry, i.e.D =√4Api(4.7)To investigate the microstructures of the CCT samples, the samples were etched at ambienttemperature between 10 and 30 seconds with Nital (2% nitric acid in ethanol). The samples werephotographed at 500x magnification to qualitatively observe the microstructure. Ferritic regionsare etched white, whereas carbides and M/A are etched dark.To reveal and quantify the M/A fraction of the various structures, three samples representativeof irregular ferrite, upper bainite and lower bainite were selected to be etched for 50 secondswith Lepera’s etchant [91] (equal parts 4% picric acid in ethanol and 1% sodium metabisulfatein water). M/A is highlighted in white while the matrix is tinted brown. The samples werephotographed at 500x magnification enough times to provide a statistical treatment. M/A particlesin the photographs were highlighted by applying a threshold in ImageJ and using an area countmethod to measure the total volume fraction of M/A in each picture. The process is illustrated inFigure 4.8.Vickers microhardness measurements were also taken of the CCT samples to further characterize36METHODOLOGY(a) (b)(c) (d)Figure 4.10: Process to measure prior austenite grain size from a micrograph a) Original micrographb) Hand outlining c) Threshold specification d) Area count measurementthe microstructures produced in terms of mechanical strength. Samples were ground to 1200-grit toprovide a smooth enough surface and indented with a pyramidal diamond head with a 1-kgN load.Enough indents (i.e. 6 indents per sample) were made to provide a statistically relevent measureof hardness of each sample [92]. An example indent is shown in Figure 1.8. The dimensions of theindents were measured by microscope at 500x magnification as shown below and Vickers hardnesswas calculated byHV = 1.854Fd2(4.8)Where F is the force used and d is the arithmetic mean of the two diagonals.37METHODOLOGY(a) (b)(c) (d)Figure 4.11: Process to measure M/A fraction through thresholding, top left to lower right, a)Original figure b) Grayscale and filtering c) Threshold specification d) Area count measurementFigure 4.12: An example hardness indent and caliper measurement from CCT sample38Chapter 5Results and DiscussionThis chapter reviews the experimental results from this study and provides a discussion ofthese results with respect to the current literature. First, laser ultrasonics grain growth resultswill be shown with the corresponding grain size distributions measured by metallography. Second,austenite decomposition behavior results will be presented for different precipitation conditions.Third, the austenite decomposition results as a function of cooling rate, prior austenite grain size,and niobium solute content is shown for the low and high carbon steels. Finally, metallographyand microhardness measurements will be presented for the continuous cooling tests.5.1 Grain Growth5.1.1 ResultsThe laser ultrasonic results for measuring austenite grain growth in the continuous heating trialson the lower and higher carbon steels are shown in Figure 5.1 a) and b). The apparent decrease ingrain size observed by laser ultrasonics from 900 to 1000°C does not correspond to a real decrease inthe mean grain size in the initial stage of austenite grain growth (as can be observed in Figure 5.2for isothermal holding). The careful observation of the raw waveforms collected in this temperaturerange show a significant saturation of the echoes (i.e. the amplitude of light received in the laserdetector was too large and lead to the distortion in amplitude of the signal measured). As themeasurement of grain size is directly linked to the frequency dependance of the attenuation, itintroduces an artifact in the predicted grain size. At higher temperatures, the reflectivity of the39RESULTS AND DISCUSSIONsample surface decreases, reducing the amount of light collected in the detector and allowing forthe subsequent waveforms collected above 1000°C to be measured reliably. The variation of grainsize during the continuous heating experiments is therefore only discussed between 1000°C andabove. The laser detector was later adjusted to better control the light intensity collected duringsuch investigations.At 1000°C, it is observed that for all cases, the initial austenite grain size is approximately 5µm. For both steels, the increase in grain size with temperature is most pronounced at 10°C/s.reaching a final grain size of approximately 48 µm for both 0.028%C and 0.058%C at 1350°C. At100°C/s, the grain growth is less pronounced, reaching a final grain size of approximately 35 µm at1350°C. Grain growth is lowest with 1000°C/s, both producing a final grain size of approximately25 and 30 µm for 0.028%C and 0.058%C respectively.The laser ultrasonic results for isothermal holding trial results are shown in Figure 5.2 a) andb), where t = 0 is defined as when the sample reaches the selected isothermal holding temperature.For the lower %C steel, shown in Figure 5.2a), holding at 950°C yields a grain size of approximately11 µm, showing a small amount of grain growth. For holding at 1050°C, grain growth is delayed forapproximately 100 s before grain growth begins, reaching a limiting grain size of approximately 37µm, showing sigmoidal grain growth behavior over the holding time. For holding at 1150°C, graingrowth is initially rapid and slows down after 1000 s, reaching a limiting grain size of 35 µm; somecoarsening is observed after this. Finally, at 1250°C, grain growth is even more rapid than 1150°Cand slows down to reach a large limiting grain size of approximately 55 µm, although the noise inthe curve makes the determination of the limiting grain size difficult.The results for austenite grain growth in the higher %C steel are shown in Figure 5.2b). Holdingat 950°C yields a grain size of approximately 11 µm, once again showing a small amount of graingrowth. For a holding at 1050°C, grain growth is very slow, showing a small increase in grainsize over the 1500s holding period with a final grain size of 17 µm. For holding at 1150°, graingrowth is initially rapid and slows down after 500 s to a limiting grain size of 29 µm after 900s. At1250°C, grain growth is rapid and slows down to reach a large limiting grain size of approximately65 µm, although once again the noise in the curve makes the determination of the limiting grainsize difficult. It is possible for these two holding temperatures, no limiting grain size is reached dueto coarsening of TiN particles.40RESULTS AND DISCUSSION˚˚˚˚(a)˚˚˚˚(b)Figure 5.1: Laser ultrasonic measurement of grain growth during continuous heating as a functionof heating rate for compositions with carbon contents of of a) 0.028%C b) 0.058%CThe micrographs of samples produced from the grain growth trials are shown in Figures 5.3 - 5.4for continuous heating and Figures 5.5 - 5.7 for isothermal holding. In addition, samples quenchedfrom the peak temperatures of 950 and 1250°C following a solutionization treatment were etched41RESULTS AND DISCUSSION˚˚˚˚(a)˚˚˚˚(b)Figure 5.2: Laser ultrasonic measurement of grain growth during isothermal holding as a functionof holding temperature for compositions with carbon contents of a) 0.028%C b) 0.058%Cto find the austenite grain size prior to cooling in the CCT trials. These are shown in Figures5.8 - 5.9. The etched prior austenite grain boundaries are delineated in black. The histograms ofaustenite grain diameter follow a log-normal relationship and this function has been fit to the data.42RESULTS AND DISCUSSIONThe average grain size of each distribution is found by taking the arithmetic average of grain sizearea and computing the equivalent area EQAD by simple geometry, ie.EQAD =√4AAvepi(5.1)The measured grain size through metallography is shown in Table 5.1, where N is the number ofgrains measured in each case. In addition, the mean of the log-normal fits is displayed as calculatedbyE[x] = eµ+σ22 (5.2)where µ and σ are, respectively, the peak and width of the grain size distribution. The measuredmean grain size measured with laser ultrasonics data and experimental measurement shows gener-ally good agreement with each other, although this is less so for a holding temperature of 1250°Cfor both compositions and in the case of 1050°C for the 0.028%C composition. This is possibly dueto larger grains in the distribution which skew the LUMet measurements to higher values. The1050°C isothermal holding samples show long right-skews in their distributions; this is especiallypronounced in the low C alloy. This is also potentially an indicator of abnormal grain growth inthese trials.43RESULTS AND DISCUSSION(a) (b)(c) (d)(e) (f)Figure 5.3: Micrographs of etched 0.028% C samples from continuous heating grain growth trials(a), c), e)) and grain diameter histograms (b), d), f)); heating rates of 10° C/s (a), b))100° C/s (c),d)) and 1000° C/s (e), f))44RESULTS AND DISCUSSION(a) (b)(c) (d)(e) (f)Figure 5.4: Micrographs of etched 0.058% C samples from continuous heating grain growth trials(a), c), e)) and grain diameter histograms(b), d), f)); heating rates of 10° C/s (a), b)) 100° C/s (c),d)) and 1000° C/s (e), f))45RESULTS AND DISCUSSION(a) (b)(c) (d)(e) (f)Figure 5.5: Micrographs of etched 0.028% C samples from isothermal grain growth trials(a),(c),(e)and grain diameter histograms(b),(d),(f); holding temperature of 950° C (a), b)) 1150° C (c), d))and 1250° C (e), f))46RESULTS AND DISCUSSION(a) (b)(c) (d)(e) (f)Figure 5.6: Micrographs of etched 0.058% C samples from isothermal grain growth trials(a), (c),(e)) and grain diameter histograms(b), (d), (f)); holding temperature of 950° C (a), b)) 1150° C(c), d)) and 1250° C (e), f))47RESULTS AND DISCUSSION(a) (b)(c) (d)Figure 5.7: Micrographs of etched samples from isothermal grain growth trials (a),(c) and grainsize histograms (b),(d); 0.028% C (a), b)) and 0.058% (c), d)), holding temperature of 1050° C48RESULTS AND DISCUSSION(a) (b)(c) (d)Figure 5.8: Micrographs of etched 0.028% C (a), b)) and 0.058% (c), d)) samples quenched frompeak temperatures prior to CCT tests (a),(c)) and grain size histograms (b),(d)); peak temperatureof 950° C49RESULTS AND DISCUSSION(a) (b)(c) (d)Figure 5.9: Micrographs of etched 0.028% C (a), b)) and 0.058% (c), d)) samples quenched frompeak temperatures prior to CCT tests (a),(c)) and grain size histograms (b),(d)); holding temper-ature of 1250° C50RESULTS AND DISCUSSIONTable 5.1: Overview of grain size distribution data for continuous heating grain growth trialsComposition Thermal History N EQAD, µm E[x], µm LUMet Measurement, µm0.028%C10°C/s, 1350°C peak 400 39 40 48100°C/s, 1350°C peak 728 29 30 351000°C/s, 1350°C peak 939 25 27 250.058%C10°C/s, 1350°C peak 342 41 40 48100°C/s, 1350°C peak 583 33 33 351000°C/s, 1350°C peak 724 29 32 30Table 5.2: Overview of grain size distribution data for isothermal holding grain growth trialsComposition Thermal History N EQAD, µm E[x], µm LUMet Measurement, µm0.028%C950°C, 15min 470 6.8 7.5 111050°C, 25min 1378 25 17 371150°C, 25min 591 31 31 351250°C/s, 15min 456 39 39 550.058%C950°C, 15min 498 6.6 7.3 111050°C, 25min 908 13 11 171150°C, 15 min 846 23 22 291250°C/s, 25min 369 47 48 65Table 5.3: Overview of grain growth distribution data for quenched samples for CCT reheatingfollowing solutionizationComposition Thermal History N EQAD, µm E[x], µm0.028%C950°C, 5s 462 5.0 4.31250°C, 15s 498 36 410.058%C950°C, 5s 470 4.9 4.51250°C, 15s 446 39 4351RESULTS AND DISCUSSION5.1.2 DiscussionIn order to understand the grain growth behavior with reference to the microalloy precipitatestabilities, computational thermodynamic software was used to the model phase stability of thesteels and their associated precipitates. The compositions of the steels were used as inputs in theCALPHAD software Thermocalc (version 4.0) and the Gibbs energy functionals contained in the’tcfe7’ database were used to describe the system. The temperature was then varied to predict theequilbrium phase stabilities between room temperature and up to the melting point of steel. Themost relevent sections of the calculations are shown in Figure 5.10.It was found that the Ae1 and Ae3 temperatures of the two steels are close (3°C higher and 12°Clower, respectively, in the case of the higher %C steel). Three types of precipitates (respectively, Mo,Ti and Nb-rich) were predicted to exist at ambient temperatures, consistent with the TEM studieson similar steels previously mentioned. Mo-rich precipitates are predicted to dissolve at relativelylow temperatures, below the Ae1. Ti-rich precipitates dissolve at high temperatures, i.e. near themelting temperature. Finally, Nb-rich precipitates dissolve at intermediate temperatures, abovethe Ae3 and below the melting temperature. Because of this, it will be assumed that in austenite,Mo will be completely dissolved and Ti will be nearly completely precipitated; depending on thepeak temperature reached, Nb in austenite may be completely precipitated, partially dissolved, orcompletely dissolved. It can be seen that the alloy with higher carbon concentration has a largervolume fraction of Nb(C,N) precipitates and these precipitates are stable to higher temperatures.The equilibrium volume fraction and percent of Nb(C,N) precipitates stable at the temperaturerelative to the initial amount is shown in Table 5.4.Table 5.4: Equilibrium volume fractions of Nb(C,N) precipitates at selected isothermal holdingtemperatures and percentage relative to ambient equilibrium fraction [41]Temperature0.028%C 0.058%CVf % Stable Precipitates Vf % Stable Precipitates950°C 8.71E-4 80.1 9.89E-4 90.81050°C 5.17E-4 47.6 7.71E-4 70.81150°C 0 0 3.01E-4 27.61250°C 0 0 0 0As the samples are heated to higher temperatures in the austenite phase region, grain growth52RESULTS AND DISCUSSION˚(a)˚(b)Figure 5.10: Thermodynamic stability of phases in a) 0.028%C b) 0.058%C composition, fromThermocalc 4.0 [41]behavior is enhanced by the dissolution of Nb(C,N) precipitates, the dissolution of TiN precipitatesand an increased grain boundary mobility. As in literature [71], it is observed in the laser ultrasonicstrials that an increase in heating rate results in less grain growth. This is due to the reduced time53RESULTS AND DISCUSSIONat elevated temperatures for the higher heating rates. There is little difference between 100 and1000°C/s, implying that a limiting heating rate is reached at some heating rate between this, wherethe final grain size is no longer dependent on heating rate.Grain growth in continuous heating is compared as a function of composition in Figure 5.11,including previous measurements from laser ultrasonics on a lower Nb and Ti composition of linepipe steel [5] for 10 and 100°C/s heating (data on grain growth at 1000°C/s was not present forthe lower Nb steel). The composition of the previously characterized steel is listed in Table 5.5.At 10°C/s, the lower Nb steel has considerably larger grain size at temperatures above 1050°Ccompared to both high Nb steels. At 100°C/s, the three compositions have similar grain growthbehavior. In general, the two high Nb steels have approximately equivalent grain growth behavior.Table 5.5: Composition of previously characterized steel [5]Alloy, Wt% C Mn Si Nb Ti Mo N S FeMGP 0.063 1.70 0.10 0.034 0.012 0.24 0.005 0.0009 BalanceFor isothermal holding (see Figure 5.2), differences between the low and high carbon steelscan be observed. At 950°C, both alloys have approximately a similar fraction of stable Nb(C,N)precipitates pinning grain boundaries and thus little if any grain growth can be observed. At 1150and 1250°C, both alloys have normal grain growth before reach a limiting grain size determinedby the Nb(C,N) and TiN precipitate fraction stable at these temperatures. At 1050°C, differingbehavior is shown for the two compositions i.e. the higher %C steel having very slow growth andthe lower %C steel having sigmoidal-shaped grain growth behavior before reaching a limiting grainsize. This sigmoidal shape has been observed in literature for measurement of grain growth insuperalloys [88] and carbon steels [93] when abnormal grain growth was present. In general thehigher %C steel has slightly lower grain growth behavior due to the increased stability of Nb(C,N)precipitates, as can be observed in the solubility product, though at 1150 and 1250°C, this is lessobvious. In the latter case, it is observed the grain size in the lower %C case is slightly lowercompared to the higher %C alloy, but as in the case of grain growth through continuous heating,these two curves are likely equivalent.Despite possessing long tails in the grain size distribution histograms of isothermal holding at1050°C, the two compositions do not show two frequency peaks as in traditional bimodel grain size54RESULTS AND DISCUSSION˚˚(a)˚˚(b)Figure 5.11: Grain growth behavior in continuous heating as a function of composition for heatingrates of a) 10°C/s b) 100°C/s; 0.034%Nb steel data from [5]distributions observed in abnormal grain growth [60]. To further compare the grain size distribu-tions, the logarithm of the cumulative density function of each distribution was normalized by eachdistribution’s average grain size to compare each distribution’s self similarity, as shown in Figure55RESULTS AND DISCUSSION5.12. It can be observed that both samples held at 1050°C have quite different curves from theother selected curves and therefore show possible evidence of abnormal grain growth. It can be ob-served that the ratio of the largest grain in the distribution Dmax to the EQAD of the distributionis much larger in the case of 1050°C holding (>3) than in the other holding temperatures (2-2.5).Therefore, it is likely that grain growth in these conditions is abnormal.The grain growth during isothermal holding at different temperatures is shown in Figure 5.13comparing measurements using the lower Nb composition [57]. In the case of holding at 1150°C,an increase in Nb yields a lower limiting grain size from the increased content of stable Nb(C,N)precipitates. In addition, it takes an increased amount of time for the limiting grain size to bereached. In the same way, in the case of holding at 1250°C, the three steels have similar limitinggrain sizes, although it takes longer for the higher Nb steels to reach their limiting grain sizes. Atthis temperature, Nb(C,N) is not stable and TiN is responsible for pinning grain boundaries; asthe lower niobium steel has a lower titanium content, this is to be expected. In the case of holdingat 1050°C, the lower Nb steel shows normal grain growth, unlike the studied alloys.5.2 Niobium Precipitation5.2.1 ResultsAs described in Section 4.2.2, the steels were solutionized, reheated and aged at 900°C priorto dilatometry to gather information on the precipitation behavior of Nb(C,N). From the mea-surements observed in Figure 5.9 and noted in Table 5.3, the prior austenite grain size of the twosteels are considered to be similar, i.e. '35 µm. The dilatometry curves of austenite decompositionfrom samples following a given precipitation time at 900°C are shown in Figures 5.14 a) and b).When niobium is fully solutionized, the Ts is low, at 613 and 599°C for low and high %C steelsrespectively. With increased holding time at 900°C, the Ts increases before saturating at a value of710 and 690°C for the low and high %C steels, respectively.No significant increase in Ts is found between 30 and 45 minutes. In both steels, the Ts increasesapproximately 90°C between 0 and 45 minutes of aging time.56RESULTS AND DISCUSSION˚˚˚˚(a)˚˚˚˚(b)Figure 5.12: Comparison of grain size distribution self-similarity for isothermal holding trials as afunction of temperature for compositions with carbon contents of a) 0.028%C b) 0.058%C5.2.2 DiscussionBased on previous studies [46], the increase in Ts following a holding time at 900°C is assumedto be related to the reduction of niobium in solution. As the holding time at 900°C is increased,57RESULTS AND DISCUSSION˚(a)˚(b)˚(c)Figure 5.13: Grain growth behavior in isothermal holding as a function of composition for holdingtemperatures of a) 1050°C b) 1150°C c) 1250°C; 0.034%Nb steel data from [74]Nb(C,N) is increasingly precipitated and Nb in solution reduced, thereby decreasing the effect ofsolute drag on austenite decomposition. For both steels, it can be estimated that a thirty minuteholding at this temperature is sufficient to reprecipitate niobium from solution, as there is no58RESULTS AND DISCUSSION˚(a)˚(b)Figure 5.14: Dilatometry curve for samples aged at selected times at 900°C for compositions witha carbon content of a) 0.028%C b) 0.058%Csignificant increase, in Ts after this. The Ts as a function of holding time at 900°C are displayedin Figure 5.15. The difference in Ts between these two steels can be attributed to the differencein C content reducing the Ae3 temperature. A simple estimate of precipitation kinetics can be59RESULTS AND DISCUSSIONconstructed by the assumption that niobium in solution has a linear effect on Ts and the data canbe normalized asf(t) =Ts(t)− Ts(0)Ts(45)− Ts(0) (5.3)where f(t) is the volume fraction of precipitates formed after a given aging time, Ts(t) is the Tsfor a given aging time and Ts(0) and Ts(45) are, respectively, the Ts after holding for 0 and 45minutes. The result of this analysis is displayed in Figure 5.16, showing that the two compositionshave similar precipitation kinetics. However, this formulation may not be completely precise asthe effect of niobium in solution on Ts may not be linear with solute concentration, as reviewed inChapter 2.˚Figure 5.15: Transformation start temperature as a function of aging time at 900°C for steels withcarbon contents of a) 0.028%C and 0.058%C60RESULTS AND DISCUSSIONFigure 5.16: Aging curve constructed from Ts data as a function of aging time at 900°C for steelswith carbon contents of 0.028%C and 0.058%C5.3 Austenite Decomposition5.3.1 ResultsBased on the results for austenite grain size measurement presented in Section 5.1, the selectedheat treatments at 950 and 1250°C following the solution treatment produced austenite grain sizesof approximately ' 5 and ' 35 µm, respectively. It was found that for 30°C/s cooling rates whenniobium was in solution and an prior austenite grain size of 35 µm, the heat of transformation waslarge enough to cause the experimental cooling curve to deviate away from the programmed coolingrate, as illustrated in Figure 5.17. This could not be suppressed by the He quench, and thereforeimposed an upper limit on cooling rate in the study. Although this deviation would make the totaltransformation curve less than ideal, the Ts for these cases are accurate, as they occur prior to heatgeneration.In addition to removing niobium from solution, the precipitation treatment will remove carbonfrom solution to form the carbonitride. Thermocalc equilibrium calculations show that the equilib-rium fraction of carbon in austenite in the alloys is 71.2 and 85.7% of the total carbon in the alloyin the 0.028 and 0.058 wt% compositions, respectively. This would have an effect on the austenite61RESULTS AND DISCUSSION˚Figure 5.17: Example of recoalesence in 0.058% C alloy, Nb in solution, 35 µm prior austenite grainsize, 30°C/sdecomposition via reducing the amount of carbon controlling the phase transformation (resultingin a carbon content of 0.020 and 0.050 wt%, respectively).The transformation curves are shown in Figures 5.18 through 5.20 for the selected CCT testsas functions of the prior austenite grain size, niobium solute content and cooling rate. In all cases,it is found that an increase in cooling rate leads to a decrease in the transformation temperatures.However, the amount Ts decreases with cooling rate is a function of the austenite grain size andniobium content. The largest decrease in Ts when the the cooling rate increased from 3 to 30°C/sis 94°C in the case of a 35 µm austenite grain size, Nb in solution and 0.058%C.The effect of niobium in and out of solution on austenite decomposition is shown in Figure5.18. In general, it is found that niobium in solution decreases the transformation temperatures,though the effect is also dependent on the prior austenite grain size. For example, the effect ofniobium in solution is less significant in the case of small austenite grain size, the largest decreaseof Ts being 31°C (in the case of 0.028%C, 30°C/s), and the majority being '20°C. In some cases,(eg. 0.028%C, 5 µm austenite grain size, 3°C/s) the Ts are close enough to be equivalent, althoughthe Tf is lower in the case of niobium in solution. The effect of niobium in solution is much morepronounced in the case of large prior austenite grain size, the largest decrease in Ts being 115°C,62RESULTS AND DISCUSSION˚˚˚˚˚˚˚(a)˚˚˚˚˚˚˚(b)˚˚˚˚˚˚˚(c)˚˚˚˚˚˚˚(d)Figure 5.18: Austenite decomposition curves of CCT tests, observing transformation dependencyon niobium content for a) 0.028%C and a prior austenite grain size of 5 µm b) 0.058%C and a prioraustenite grain size of 5 µm c) 0.028%C and a prior austenite grain size of 35 µm d) 0.058%C anda prior austenite grain size of 35 µmin the case of 30°C/s cooling, and the average being '60°.The effect of carbon on austenite decomposition is shown in Figure 5.19. In general, it is found63RESULTS AND DISCUSSION˚˚˚˚˚˚˚(a)˚˚˚˚˚˚˚(b)˚˚˚˚˚˚˚(c)˚˚˚˚˚˚˚(d)Figure 5.19: Austenite decomposition curves of CCT tests, observing transformation dependencyon carbon content for a) niobium precipitated and a prior austenite grain size of 5 µm b) niobiumin solution and a prior austenite grain size of 5 µm c) niobium precipitated and a prior austenitegrain size of 35 µm d) niobium in solution and a prior austenite grain size of 35 µmthat increased carbon decreases the transformation temperatures. As for the case of niobium, theeffect is less pronounced with 5 µm prior austenite grain size; indeed, in several cases (eg, niobium64RESULTS AND DISCUSSION˚˚˚˚˚˚˚(a)˚˚˚˚˚˚˚(b)˚˚˚˚˚˚˚(c)˚˚˚˚˚˚˚(d)Figure 5.20: Austenite decomposition curves of CCT tests, observing transformation dependencyon prior austenite grain size for a) niobium precipitated and 0.028%C b) niobium precipitated and0.058%C a) niobium in solution and 0.028%C b) niobium in solution and 0.058%Cprecipitated, and 10°C/s cooling) the wt%C is shown to have minimal if any effect on Ts, thoughthe Tf is reduced in these cases. In the case of 35 µm prior austenite grain size, carbon has agreater effect on transformation start temperature; this effect is roughly similar for niobium in and65RESULTS AND DISCUSSIONout of solution, the largest decrease in Ts with carbon content being 50°C in the case of niobiumin solution and a cooling of 30°C/s, and the majority being on average '30°C.The effect of prior austenite grain size is shown in Figure 5.20. In general, it is found thatincreased austenite grain size decreases the transformation temperatures. In general, the austenitegrain size had the largest effect in decreasing transformation temperatures. Ts was decreased bya minimum of 54°C in the case of 30°C/s cooling, niobium precipitated and 0.028%C of and amaximum of 153°C in the case of 30°C/s cooling, niobium in solution and 0.058%C. As in theprevious cases, the effect is pronounced for niobium in solution and larger carbon concentrations.The microstructures revealed by Nital etching following the CCT tests are illustrated in Figures5.21 - 5.24; ferrite appears as white and carbon rich phases, such as retained M/A, and grainboundaries appear dark.For austenite decomposition conditions leading to high transformation temperatures (ie. lowcarbon content, niobium precipitated, low cooling rate, and a small prior austenite grain size) themicrostructure is characterized primarily by irregular ferrite, although the morphology varies withthe specific Ts and the prior austenite grain size. At the highest transformation temperatures,for a prior austenite grain size of 5 µm and 3°C/s cooling rates, the ferrite grains is relativelylarge and secondary phase is equiaxed, forming at ferrite grain corners. At lower transformationtemperatures, the ferrite grain size is correspondingly finer, and appears as decreasingly equiaxed.For a prior austenite grain size of 35 µm and Nb completely in precipitate form, the ferrite grainsare correspondingly large, secondary phase appears coarser, and no retained prior austenite grainboundary is observed, instead showing irregular boundaries. A representative example of this isshown in 5.23 c).For conditions with lower transformation temperatures in samples with a large prior austenitegrain size, the microstructures are characterized by bainite. This can be delineated into upperand lower bainite depending on the morphology of the microstructure. A representative exampleof upper bainite is shown in Figure 5.23 d). Here it can be observed that the austenite grainboundary is strongly delineated by secondary phase and islands of secondary phase are present ingrain interiors in roughly equiaxed islands. At lower Ts, lower bainite is primarily formed. A goodexample of this is shown in Figure 5.24 f). Once again, it can be observed that the prior austenitegrain boundaries are still delineated, but in this case secondary phase is present in grain interiors66RESULTS AND DISCUSSIONin long, fine laths.(a) (b)(c) (d)(e) (f)Figure 5.21: Nital 2% etched microstructures produced by austenite decomposition of 0.028% Calloy from a 5 µm average prior austenite grain size; shown as niobium precipitated (a),(c),(e) orin solution (b),(d),(f) with cooling rate of 3° C/s (a), b)) 10° C/s (c), d)), 30° C/s (e), f))67RESULTS AND DISCUSSION(a) (b)(c) (d)(e) (f)Figure 5.22: Nital 2% etched microstructures produced by austenite decomposition of 0.058% Calloy from a 5 µm average prior austenite grain size; shown as niobium precipitated (a),(c),(e) orin solution (b),(d),(f), with cooling rate of 3° C/s (a), b)), 10° C/s (c), d)), 30° C/s (e), f))68RESULTS AND DISCUSSION(a) (b)(c) (d)(e) (f)Figure 5.23: Nital 2% etched microstructures produced by austenite decomposition of 0.028% Calloy from a 35 µm average prior austenite grain size; shown as niobium precipitated (a),(c),(e) orin solution (b),(d),(f), with cooling rate of 3° C/s (a), b)), 10° C/s (c), d)), 30° C/s (e), f))69RESULTS AND DISCUSSION(a) (b)(c) (d)(e) (f)Figure 5.24: Nital 2% etched microstructures produced by austenite decomposition of 0.058% Calloy from a 35 µm average prior austenite grain size; shown as niobium precipitated (a),(c),(e) orin solution (b),(d),(f), with cooling rate of 3° C/s (a), b)), 10° C/s (c), d)), 30° C/s (e), f))70RESULTS AND DISCUSSIONThe three microstructures mentioned (i.e. Figure 5.23 c), Figure 5.23 d), and 5.24 f) ) asrepresentative for different microstructure (i.e. irregular ferrite, upper bainite, and lower bainite)were selected for the study of M/A fraction by Le Pera etching. The resulting microstructures areshown in Figure 5.25. When etched with LePera, the matrix appears brown, while retained M/Aappears white. The irregular ferrite sample shows a moderate fraction of randomly distributedM/A with a blocky morphology with no evidence of the prior austenite grain boundary. The upperbainitic sample shows a high fraction of M/A, with blocky islands of ferrite inside the prior austenitegrain boundary and elongated ‘necklaces’ outlining prior austenite grain boundaries; it is uncertainfrom these micrographs whether each necklace segment is a single particle or a number of smaller,closely fitting M/A particles. The lower bainitic sample shows a low fraction of M/A, the majorityof the consituent inside a few grain interiors as long laths, and occasionally a ‘necklace’ segmentalong a prior austenite grain boundary. The volume fractions of M/A within each microstructureare listed in Table 5.6.Table 5.6: M/A volume fractions for analyzed CCT samples; all samples have a prior austenitegrain size of 35 µmMicrostructure CCT Conditions Ts, °C M/A Volume %Irregular Ferrite 0.028%C, Nb precipitated, 10 °C/s 647 6.5 ± 0.47Upper Bainite 0.028%C, Nb in solution, 10 °C/s 577 11.1 ± 0.99Lower Bainite 0.058%C, Nb in solution, 30 °C/s 487 1.7 ± 0.48The microhardness values of the various samples produced from CCT are summarized in Figure5.26, where each measurement has an uncertainty of at most ±5 HV. In all cases, the hardnessincreases with increased cooling rate, increased carbon content and niobium in solution. Whenniobium is precipitated, there is a small increase in hardness for all cooling rates when the prioraustenite grain size is increased from 5 to 35 µm, up to 25 HV in the case of 0.058%C, Nb insolution and 3°C/s . As in the case of austenite decompostion, niobium in solution has a strongereffect on increasing hardness when the austenite grain size is large. In the case of a small austenitegrain size, the difference is at most 20 HV, whereas in the case of a large austenite grain size, thereis an extensive increase in hardness for all cases, up to an increase of 70 HV. It is observed thatthe relation between hardness and cooling rate is near-logarithmic, as observed in prior studies ofmicrostructures produced by continuous cooling [34].71RESULTS AND DISCUSSION(a)(b) (c)Figure 5.25: Le Pera etched microstructures for selected samples produced by CCT a) irregularferrite (0.028%C, Nb precipitated, 10°C/s cooling) b) upper bainite (0.028%C, Nb in solution,10°C/s cooling) c) lower bainite (0.058%C, Nb in solution, 30°C/s cooling); all samples have a prioraustenite grain size of 35 µm72RESULTS AND DISCUSSION˚(a)˚(b)Figure 5.26: Microhardness measurements for CCT samples plotted as function of cooling rate fora prior austenite grain size of a) 5 µm b) 35 µm73RESULTS AND DISCUSSION5.3.2 DiscussionThe decomposition behavior of austenite is shown to have a strong dependence on the prioraustenite grain size, niobium solute content, carbon content, and cooling rate. An increase in thecarbon content results in a decrease in transformation temperatures. Ferrite growth is a diffusionalprocess, dependent on the movement and partitioning of carbon across the interphase boundary.In addition, carbon is an austenite phase stabilizer, further decreasing the transformation temper-atures. The effect on bainite is similar, as the transformation involves the nucleation and growthof ferrite laths from prior austenite grain boundaries, redistributing carbon between the laths. Itis worth noting there are several cases (as seen in Figure 5.19 a) and b)) when austenite grain sizeis small, there is negligable reduction in Ts when carbon content is increased, but in these cases adecrease in Tf is present, indicating the effect of carbon.The increase in prior austenite grain size has a significant and well documented effect on austen-ite decomposition by setting the initial number of ferrite nuclei, as ferrite preferentially forms atgrain boundary corners, edges and faces, a higher density of grain boundaries will result to a higherdensity of ferrite nuclei. The site density is proportional to 1/D3, 1/D2 and 1/D depending on nu-cleation at corners, edges or faces respectively. This will lead to higher transformation temperatureswhen the austenite grain size is small and lower transformation temperatures when the austenitegrain size is large. As stated, this variable had the strongest effect on decreasing the transformationtemperatures. In future work, investigation of austenite decomposition from a third austenite grainsize may be useful.The cooling rate has importance in the kinetics of austenite decomposition. As the transfor-mation of austenite to the corresponding daughter phases is thermally activated, the newly formedproduct need time to nucleate and grow. Therefore, increasing the cooling rate will reduce thetime these products take to grow and form, reducing the transformation temperatures. In all casesin the study, an increase in cooling rate will lead to a corresponding decrease in transformationtemperatures. In general, the decrease in transformation temperatures from an increase in coolingrate is more significant for small austenite grain size than for large.Finally, the niobium solute content is shown to have an effect on transformation temperaturesdue to solute drag. It is observed in this study that for small prior austenite grain sizes, the74RESULTS AND DISCUSSIONreduction in transformation temperatures by the inclusion of niobium in solution is small, and inseveral cases there is no reduction in Ts. However, for large prior austenite grain sizes, the reductionin transformation temperatures from the dissolution of nioium is large, up to 115°C in the case of0.058% C and 30°C/s cooling. This is potentially due to the segregation of niobium to austenitegrain boundaries; the density of grain boundaries is inversely related to the prior austenite grainsize and therefore the concentration of solute at a given boundary will be larger for larger grainsizes. This would have significance in relating to austenite decomposition in the CGHAZ, whereprior austenite grain sizes are expected to be larger.For a small prior austenite grain size, irregular ferrite is produced in all cases, although themorphology varies due to the Ts of the steel. At the highest values of Ts, comparitively large ferriteis produced, likely through nucleation at austenite grain corners. As the Ts is reduced, the higherdegree of undercooling activates a larger number of ferrite nuclei to form along austenite grainboundary corners, edges and faces. In the same way, an increase in carbon content and niobiumsolute content result in a decrease in Ts and therefore the grain size is smaller in these cases aswell. However, a dedicated quantitative study of ferrite grain following cooling would be useful toconfirm this in future work.For a large prior austenite grain size, a larger variety of microstructures are produced.Theaustenite decomposition products can be roughly correlated with the Ts for each condition. Forconditions with high Ts, ferrite is able to nucleate and grow from the grain boundaries to producecoarse ferrite plates, retaining M/A as the transformation reaches completion. As Ts is reduced,upper bainite is formed following austenite decomposition. At still lower Ts, lower bainite is pro-duced. Given the high fraction of M/A in upper bainite, as well as the presence of M/A alongprior austenite grain boundaries in ’necklace’ formations, there is a concern that this will poten-tially promote intergranular cracking during plastic deformation due to the brittleness of M/Aparticles [29] [30]. This would be worth further investigation.Previous experimental work [45] on a low Nb-containing line pipe steel showed that the amountof M/A observed in a microstructure formed following continuous cooling is dependent on the Ts.This relation is shown in Figure 5.25, and illustrates that for Ts > 620°C, the M/A fraction is4-7%. For Ts <620°C, the M/A fraction increases, reaching a maximum of 13% at 586°C coincidingwith the formation of upper bainite. As Ts decreases further, lower bainite is formed and the M/A75RESULTS AND DISCUSSIONcontent of the steel decreases, reaching a minimum of 1-3% at 525°C when the microstructure iscompletely lower bainite. The measurements of M/A observed in the selected microstructures inthis work are included in the original figure with the original measurements of M/A fraction in theseparate steel. These limited results appear consistent with previous measurements.˚Figure 5.27: M/A volume fractions in samples produced by CCT as a function of Ts, adaptedfrom [27]Turning to the case of hardness measurements, it was found that niobium in solution contributesto the hardness via transformation strengthening and refinement of the ferrite grain size as thetransformation temperatures are lowered. It is uncertain how much solution strengthening andprecipitation hardening is present in these cases. For the smaller prior austenite grain size of 5 µm,the condition of Nb in or out of solution does not affect hardness values much, as the transformationtemperatures between the two cases are very similar, as can be seen in Figure 5.18 and 5.19. Fora large prior austenite grain size of 35 µm, the hardness is low when Nb is precipitated, despitethe lower transformation temperatures, due to the coarse grain size reducing strength. As seen inmicrostructures at this temperature range (ie. Figure 5.23/24 a)/c)/f)), coarse ferrite is producedand the microstructure will posess low hardness values. When Nb is in solution, however, thehardness shows an extensive increase, as upper and lower bainite microstructures are produced,microstructures with a much higher strength. As expected, higher carbon amounts increase the76RESULTS AND DISCUSSIONhardness in all cases. This is likely due to transformation strengthening, although it is possible thisis through solution strengthening. It is uncertain whether lower bainite is an undesirable structure;although posessing a high hardness value, this is not necessarily indicative of a low toughnessmicrostructure, especially given its low M/A fraction. Fracture toughness testing, such as CharpyV-notch (CVN) would be useful to examine this question.Figure 5.28 shows the relation between hardness and transformation start temperatures for thetwo selected austenite grain size, showing that lower Ts values generally lead to higher hardnessvalues. This relation has been observed in prior study on an HSLA steel [9].Finally, Figure 5.29 shows CCT diagrams of the alloys for each selected carbon content and prioraustenite grain size, with accompanying hardness values. These were constructed by displaying theTs and Tf data along a specified cooling path.It is noted that a greater number of cooling rateswould be useful in the refinement of these CCT diagrams.77RESULTS AND DISCUSSION˚(a)˚(b)Figure 5.28: Microhardness measurements for CCT samples plotted as function of Ts for a prioraustenite grain size of a) 5 µm b) 35 µm78RESULTS AND DISCUSSION˚˚˚˚(a)˚˚˚˚(b)˚˚˚˚(c)˚˚˚˚(d)Figure 5.29: Continuous cooling diagrams for selected compositions, observing transformation de-pendency on niobium content for a) 5 µm prior austenite grain size and 0.028%C b) 5 µm prioraustenite grain size and 0.058%C c) 35 µm prior austenite grain size and 0.028%C d) 35 µm prioraustenite grain size and 0.058%C79Chapter 6Summary and Future Work6.1 Summary1. Niobium precipitated as Nb(C,N) in as-received line pipe steels has a significant effect at re-stricting grain growth during continuous heating. This is especially true at lower heating rates(i.e. 10°C/s). At higher heating rates, there is an decrease in total grain growth upon reachinga peak temperature due to the reduced time at elevated temperatures. Limiting grain growthbehavior is approached at the highest heating rates (1000°C), when grain growth becomesindependent of heating rate. Grain growth studied by laser ultrasonics were confirmed byex-situ metallography and showed good agreement.2. Niobium precipitated as Nb(C,N) in as-received line pipe steels has a significant effect atrestricting grain growth At higher holding temperatures, there is an increase in grain growthdue to the increased solubility of Nb(C,N). Abnormal grain growth behavior was noted duringholding in a temperature region when precipitates are partially stable. This was confirmedthrough comparison of self-similarity between the studied grain size distributions. At tem-peratures in excess of the dissolution temperature of Nb(C,N), TiN is responsible for therestriction of grain growth.3. The precipitation kinetics of Nb(C,N) were inferred by comparing the Ts of austenite decom-position following a selected holding time. After a holding of 30 min at 900°C, it was foundthe Ts no longer decreased with holding time and therefore, it is assumed precipitation is80SUMMARY AND FUTURE WORKcomplete. It is possible, the kinetics of niobium precipitation may be constructed by such amethod.4. The increase in cooling rate, prior austenite grain size and carbon content has a significanteffect on austenite decomposition by decreasing transformation temperatures. The effect ofniobium in and out of solution was found to have a variable effect with regards to prioraustenite grain size selected. For small austenite grain sizes, niobium in solution has a slighteffect on decreasing austenite decomposition temperatures. However, at large grain sizes,niobium in solution has a significant effect (up to 120°C for Ts) on decreasing transformationtemperatures.5. Microstructures formed by continuously cooling from small prior austenite grain sizes areirregular ferritic. As transformation temperatures are decreased, the ferrite grain sizes arerefined. Microstructures formed from large prior austenite grain sizes range from coarse ferriteto upper and lower bainite. The fraction of M/A was found to be highest in upper bainiticmicrostructures6. Hardness measurements are strongly influenced by the presence of niobium in solution, par-ticularily when the prior austenite grain size is large.6.2 Future Work1. Using the experimental laser ultrasonics results derived from the study, grain growth behaviorin these (or similar) steels may be used as inputs in modelling grain growth in a HAZ. Insuch a model, parameters of a simulated welding pass (or passes) could be used to constructa thermal history of a location within the HAZ, and the model could find the correspondinggrain growth.2. The technique presented to quantify niobium precipitation as a function of transformationtemperature could be compared to a more well-established method (i.e. electrolytic dissolu-tion, electrical resistivity, etc.) to confirm its predictive capacity.3. A quantitative analysis of the microstructures produced by continuous cooling would be81SUMMARY AND FUTURE WORKuseful. Phase fraction of ferrite, upper bainite and lower bainite could be found throughoptical metallography to better refine the CCT curves. Alternatively, EBSD could be usedto more precisely characterize the microstructures.4. Mechanical testing of the simulated microstructures produced in austenite decompositionwould give better indication of the mechanical properties of upper and lower bainitic mi-crostructures forming in the CGHAZ than microhardness testing. This could be achievedthrough tensile testing or Kahn tear tests.5. Using the dilation results from the study, austenite decomposition behavior in these steels (orsimilar) steels may be used as inputs in modelling microstructure development in a HAZ. Insuch a model, parameters of a simulated welding pass (or passes) could be used to constructa thermal history of a location within the HAZ, and the model could find the correspondingtransition temperatures and phase fractions of microstructural constituents in the resultinglocation.82Bibliography[1] U.S. Energy Information Admnistration, “Short-Term Energy Outlook.” U.S. Energy Infor-mation Administration, Retrieved February 3, 2016. http://www.eia.gov/forecasts/steo.[2] D. Furchtgott-Roth and K. P. Green, “Intermodal safety in the trans-port of oil.” Fraser Institute, October 2013, Retrieved May 4, 2016.https://www.fraserinstitute.org/sites/default/files/intermodal-safety-in-the-transport-of-oil-rev3.pdf.[3] R. Lazor Private Communication, 2016.[4] W. J. Poole, M. Militzer, and T. Garcin, “An integrated model to predict microstructure andmechanical properties in the heat affected zone for X80 linepipe,” in Proceedings Of The 9thInternational Pipeline Conference, 2012.[5] T. Garcin, M. Militzer, W. Poole, and L. Collins, “Microstructure model for the heat-affectedzone of X80 linepipe steel,” Materials Science and Technology, vol. 32, pp. 708–721, 2016.[6] D. A. Porter and K. Easterling, Phase Transformations in Metals and Alloys. Chapman &Hall, 1992.[7] T. Massalski and H. Okamoto, Binary Alloy Phase Diagrams, vol. 1. ASM International,2nd ed., 1990.[8] H. Bhadeshia and R. Honeycombe, Steels: Microstructure and Properties. Butterworth-Heinemann, 2011.[9] P. Petkov, “Austenite decomposition of low carbon high strength steels during continuouscooling,” Master’s thesis, University of British Columbia, 2003.[10] H. Li, J.-L. Liang, Y.-F. Feng, and D.-X. Huo, “Microstructure transformation X70 pipelinesteel weld heat-affected zone,” Rare Metals, vol. 33, no. 4, pp. 493–498, 2014.[11] G. Krauss, Steels: Processing, Structure and Performance. ASM International, second ed.,2015.[12] F. Xiao, B. Liao, D. Ren, Y. Shan, and K. Yang, “Acicular ferritic microstructure of a low-carbon Mn-Mo-Nb microalloyed pipeline steel,” Materials Characterization, vol. 54, pp. 305–314, 2005.[13] K. Wu, “Three-dimensional analysis of acicular ferrite in a low carbon steel containing tita-nium,” Acta Materialia, vol. 54, pp. 569–574, 2006.[14] H. Bhadeshia, Bainite In Steels. Institute of Materials, 2nd ed., 2001.83BIBLIOGRAPHY[15] F.-R. Xiao, B. Liao, Y.-Y. Shan, and G.-Y. Qiao, “Challenge of mechanical properties of anacicular ferrite pipeline steel,” Materials Science & Engineering, A, vol. 431, pp. 41–52, 2006.[16] I. Tamura, H. Sekine, and T. Tanaka, Thermomechanical Processing of High-Strength Low-Alloy Steels. Butterworth-Heinemann, first ed., 1988.[17] L. Fielding, “The bainite controversy,” Materials Science and Technology, vol. 29, no. 4,pp. 383–399, 2013.[18] A. Borgenstam and M. Hillert, “Bainite in the light of rapid continuous cooling information,”Metallurgical and Materials Transactions A, vol. 27A, pp. 1501–1512, 1996.[19] H. Aaronson, W. Reynolds, G. Shiflet, and G. Spanos, “Bainite viewed three different ways,”Metallurgical and Materials Transactions A, vol. 21A, pp. 1343–1380, 1990.[20] M. Takahashi and H. Bhadeshia, “Model for transformation from upper to lower bainite,”Materials Science and Tech, vol. 6, pp. 592–603, 1990.[21] B. Bramfitt and J. Speer, “A perspective on the morphology of bainite,” Metallurgical andMaterials Transactions A, vol. 21A, pp. 817–829, 1990.[22] L. Y. Lan, C. L. Qiu, D. W. Zhao, and X. H. Wang, “Effect of austenite grain size on isothermalbainite transformation in low carbon microalloyed steel,” Materials Science and Technology,vol. 27, no. 11, pp. 1657–1663, 2011.[23] S. Singh and H. Bhadeshia, “Estimation of bainite plate-thickness in low-alloy steel,” MaterialsScience & Engineering A, vol. 245, pp. 72–79, 1998.[24] Y. R. Im, B. J. Lee, Y. J. Oh, J. H. Hong, and H. C. Lee, “Effect of microstructure onthe cleavage fracture strength of low carbon Mn-Ni-Mo bainitic steels,” Journal of NuclearMaterials, vol. 324, pp. 33–40, 2004.[25] N. Takayama, G. Miyamoto, N. Kamikawa, H. Nako, T. Maki, and T. Furuhara, “Forma-tion of martensite austenite constituent in continuously cooled Nb-bearing low carbon steels,”Materials Science Forum, vol. 638-642, pp. 3080–3085, 2010.[26] K. Ishida, “Calculation of the effect of alloying elements on the Ms temperature in steels,”Journal of Alloys and Compounds, vol. 220, pp. 126–131, 1995.[27] J. M. Reichert, T. Garcin, M. Militzer, and W. J. Poole, “Formation of martensite/austenite(M/A) in X80 linepipe steel,” in Proceedings Of The 9th International Pipeline Conference,2012.[28] K. Ohya, J. Kim, K. Yokoyama, and M. Nagumo, “Microstructures relevant to brittle fractureinitiation at the heat-affected zone of weldment of a low-carbon steel,” Metallurgical andMaterials Transactions A, vol. 27A, pp. 2574–2582, 1996.[29] Y. Li and T. Baker, “Effect of morphology of martensite-austenite phase on fracture of weldheat affected zone in vanadium and niobium microalloyed steels,” Materials Science andTechology, vol. 26, no. 9, pp. 1029–1040, 2014.[30] C. Davis and J. King, “Cleavage initiation in the intercritically reheated coarse-grained heat-affected zone: Part i. fractographic evidence,” Metallurgical and Materials Transactions A,vol. 25A, pp. 563–573, 1994.84BIBLIOGRAPHY[31] T. Gladman, The Physical Metallurgy Of Microalloyed Steels. London: Institute of Materials,2002.[32] D. Matlock and J. Speer, “Microalloying concepts and application in long products,” MaterialsScience and Technology, vol. 25, no. 9, pp. 1118–1125, 2009.[33] C. Chen, H. Yen, F. Kao, W. Li, C. Huang, J. Yang, and S. Wang, “Precipitation hardening ofhigh-strength low-alloy steels by nanometer-sized carbides,” Materials Science & Engineering,A, vol. 499, pp. 162–166, 2009.[34] S. S. Sohn, S. Y. Han, J. H. Bae, H. S. Kim, and S. Lee, “Effects of microstructure and pipeforming strain on yield strength before and after spiral pipe forming of API X70 and X80 steelsheets,” Materials Science & Engineering, A: Structural Materials, vol. 573, pp. 18–26, 2013.[35] M. Hernyk, S. Kyriakides, and A. Onoufriou, “Effects of the UOE/UOC pipe manufacturingprocesses on pipe collapse pressure,” International Journal of Mechanical Sciences, vol. 49,pp. 533–553, 2007.[36] A. J. DeArdo, “Niobium in modern steels,” International Materials Reviews, vol. 48, no. 6,pp. 371–402, 2003.[37] M. Charleux, W. Poole, M. Militzer, and A. Deschamps, “Precipitation behavior and its effecton strengthening of an HSLA-Nb/Ti steel,” Metallurgical and Materials Transactions A, 2001.[38] A. Craven, K. He, L. A. J. Garvie, and T. N. Baker, “Complex heterogeneous precipitationin titanium-niobium microalloyed Al-killed HSLA steels - I. (Ti,Nb)(C,N) particles,” ActaMaterialia, vol. 48, pp. 3857–3868, 2000.[39] R. Sharma, V. Lakshmanan, and J. Kirkaldy, “Solublity of niobium carbide and niobiumcarbonitride in alloyed austenite and ferrite,” Metallurgical and Materials Transactions A,vol. 15A, pp. 545–553, 1984.[40] D. S. Martin, F. G. Caballero, C. Capdevila, and C. G. de Andres, “Austenite grain coarseningunder the influence of niobium carbonitrides,” Materials Transactions, vol. 45, no. 9, pp. 2797–2804, 2004.[41] J. Andersson, T. A. Helander, L. Hglund, P. Shi, and B. Sundman, “Thermo-Calc and DIC-TRA, Computational tools for materials science,” Calphad, vol. 26, pp. 273–312, 2002.[42] K. Easterling, Introduction To The Physical Metallurgy Of Welding. Butterworth-Heinemann,2nd ed., 1992.[43] J. C. Cao, Q. Y. Liu, Q. L. Yong, and X. J. Sun, “Effect of niobium on isothermal transfor-mation of austenite to ferrite in HSLA low-carbon steel,” Journal of Iron and Steel ResearchInternational, vol. 14, no. 3, pp. 51–55, 2007.[44] A. Rivas, E. Vidal, D. Matlock, and J. G. Speer, “Electrochemical extraction of microalloycarbides in Nb-steel,” Revista de Metalurgica, vol. 44, no. 5, 2008.[45] J.-S. Park, Y.-S. Ha, S.-J. Lee, and Y.-K. Lee, “Dissolution and precipitation kinetics ofNb(C,N) in austenite of a low-carbon Nb-microalloyed steel,” Metallurgical and MaterialsTransactions A, vol. 40A, pp. 560–568, 2009.85BIBLIOGRAPHY[46] F. Fazeli and M. Militzer, “Effect of Nb on austenite formation and decomposition in an X80linepipe steel,” International Journal Of Iron and Steel Research, vol. 1, pp. 658–663, 2011.[47] J. Jang, Y.-U. Heo, C.-H. Lee, H. Bhadeshia, and D.-W. Suh, “Interphase precipitation in Ti-Nb and Ti-Nb-Mo bearing steel,” Materials Science and Technology, vol. 29, no. 3, pp. 309–313,2013.[48] J.-G. Jung, J.-S. Park, J. Kim, and Y.-K. Lee, “Carbide precipitation kinetics in austenite ofa Nb-Ti-V microalloyed steel,” Materials Science & Engineering A, vol. 528, pp. 5529–5535,2011.[49] H. Zurob, G. Zhu, S. Subramanian, G. Purdy, C. Hutchinson, and Y. Brechet, “Analysis of theEffect of Mn on the Recrystallization Kinetics of High Nb Steel: An Example of Physically-based Alloy Design,” ISIJ International, vol. 45, no. 5, pp. 713–722, 2005.[50] J. Cahn, “The impurity-drag effect in grain boundary motion,” Acta Materialia, vol. 10,pp. 789–798, 1962.[51] A. Michels, C. Krill, H. Ehrhardt, R. Birringer, and D. Wu, “Modelling the influence of grain-size-dependent solute drag on the kinetics of grain growth in nanocrystalline materials,” ActaMaterialia, vol. 47, pp. 2143–2152, 1999.[52] Y. Brechet, C. Hutchinson, H. Zurob, and C. Sinclair, “Effect of Nb on ferrite recrystallizationand austenite decomposition in microalloyed steels,” Steel Research International, vol. 78,pp. 210–215, 2007.[53] S. Hansen, J. V. Sande, and M. Cohen, “Niobium carbonitride precipitation and austeniterecrystallization in hot-rolled microalloyed steels,” Metallurgical and Materials TransactionsA, vol. 11A, pp. 387–402, 1980.[54] H. Zurob, Y. Brechet, and G. Purdy, “A model for the competition of precipitation andrecrystallization in deformed austenite,” Acta Materialia, vol. 49, pp. 4183–4190, 2001.[55] C. Hutchinson, H. Zurob, C. Sinclair, and Y. Brechet, “The comparative effectiveness of Nbsolute and NbC precipitates at impeding grain-boundary motion in Nb steels,” 59, vol. 59,pp. 635–637, 2008.[56] S. J. Lechuk, “A study of austenite grain growth in a Ti-Nb HSLA steel,” Master’s thesis,University Of British Columbia, 1996.[57] M. Maalekian, R. Radis, M. Militzer, A. Moreau, and W. J. Poole, “In situ measurement andmodelling of austenite grain growth in a Ti/Nb microalloyed steel,” Acta Materialia, vol. 60,pp. 1015–1026, 2012.[58] L. Cuddy and J. Raley, “Austenite grain coarsening in microalloyed steels,” Metallurgical andMaterials Transactions A, vol. 14A, pp. 1989–1995, 1983.[59] O. N. Dogan, G. Michal, and G. Kwon, “Pinning of austenite grain boundaries by AlN pre-cipitates and abnormal grain growth,” Metallurgical and Materials Transactions A, vol. 23A,pp. 2121–2129, 1992.[60] D. Chakrabarti, C. Davis, and M. Strangwood, “Development of Bimodal Grain Structuresin Nb-containing High-Strength Low-Alloy Steels during Slab Reheating,” Metallurgical andMaterials Transactions A, vol. 39A, p. 1963, 2008.86BIBLIOGRAPHY[61] J. Moon, J. Lee, and C. Lee, “Prediction for the austenite grain size in the presence of growingparticles in the weld HAZ of Ti-microalloyed steel,” Acta Materialia, vol. 459, pp. 40–46, 2007.[62] X. L. Wan, K. M. Wu, G. Huang, R. Wei, and L. Cheng, “In situ observation of austenitegrain growth behavior in the simulated coarse-grained heat-affected zone of Ti-microalloyed,”International Journal of Minerals, Metallurgy and Materials, vol. 21, no. 9, pp. 878–885, 2014.[63] Y. Chen, D. Zhang, Y. Liu, H. Li, and D. Xu, “Effect of dissolution and precipitation ofNb on the formation of acicular ferrite/bainite ferrite in low-carbon hsla steels,” MaterialsCharacterization, vol. 84, pp. 232–239, 2013.[64] J. H. Yang, Q. Y. Liu, D. B. Sun, and X. Y. Li, “Microstructure and transformation charac-teristics of acicular ferrite in high niobium-bearing microalloyed steel,” Journal of Iron andSteel Research, vol. 17, no. 6, pp. 53–59, 2010.[65] E. C. Bain, The Alloying Elements in Steel. 1939.[66] Z. Zhong, H. Saka, T. Kim, E. Holm, Y. Han, and X. Xie, “Continuous cooling transformationtemperatures and microstructures of niobium bearing microalloyed steels,” Materials ScienceForum, vol. 475-479, pp. 65–68, 2005.[67] N. Maruyama, G. Smith, and A. Cerezo, “Interaction of the solute niobium or molybdenumwith grain boundaries in alpha-iron,” Materials Science & Engineering A, vol. 353, pp. 126–132, 2003.[68] P. Yan and H. Bhadeshia, “Austenite-ferrite transformation in enhanced niobium, low carbonsteel,” Materials Science and Technology, vol. 31, no. 9, pp. 1066–1076, 2015.[69] T.-K. Lee, H. Kim, B. Kang, and S. Hwang, “Effect of inclusion size on the nucleation ofacicular ferrite welds,” ISIJ International, vol. 40, no. 12, pp. 1260–1268, 2000.[70] S. Kou, Welding Metallurgy. John Wiley & Sons, 2nd ed., 2003.[71] K. Banerjee, M. Militzer, M. Perez, and X. Wang, “Nonisothermal austenite grain growthkinetics in a microalloyed X80 linepipe steel,” Metallurgical and Materials Transactions A,vol. 41A, no. 6, pp. 3161–3172, 2010.[72] B. Hutchinson, J. Komenda, G. S. Rohrer, and H. Beladi, “Heat affected zone microstructuresand their influence on toughness in two microalloyed HSLA steels,” Acta Materialia, vol. 97,pp. 380–391, 2015.[73] S. Murugan, P. Kumar, B. Raj, and M. Bose, “Temperature distribution during multipasswelding of plates,” Internation Journal Of Pressure Vessels And Piping, vol. 75, pp. 891–905,1998.[74] M. Gaudet, The Tensile Properties and Toughness of Microstructures Relevant to the HAZ ofX80 Linepipe Steel Girth Welds. PhD thesis, University Of British Columbia, 2015.[75] J. Dossett and H. Boyer, Practical Heat Treating. ASM International, 2nd ed., 2006.[76] A. Guo, R. Misra, J. Liu, L. Chen, X. He, and S. Janstro, “An analysis of the microstructureof the heat-affected zone of an ultra-low carbon and niobium-bearing acicular ferrite steelusing EBSD and relationship to mechanical properties,” Materials Science & Engineering, A,vol. 527, pp. 6440–6448, 2010.87BIBLIOGRAPHY[77] M. Militzer, A. Giumelli, E. B. Hawbolt, and T. R. Meadowcroft, “Austenite grain growthkinetics in Al-killed plain carbon steels,” Metallurgical and Materials Transactions A, vol. 27A,pp. 3399–3409, 1996.[78] S. Mishra and DebRoy, “Non-isothermal grain growth in metals and alloys,” Materials Scienceand Technology, vol. 22, no. 3, pp. 253–278, 2006.[79] C. Fossaert, G. Rees, T. Maurickx, and H. Bhadeshia, “The effect of niobium on the hard-enability of microalloyed austenite,” Metallurgical and Materials Transactions A, vol. 26A,pp. 21–30, 1995.[80] E. El-Kashif and T. Koseki, Alloy Steel - Properties and Use, ch. 4, pp. 87–108. InTech, 2011.[81] R. J. Hattingh and G. Pienaar, “Weld HAZ embrittlement of Nb containing C-Mn steels,”Internation Journal Of Pressure Vessels And Piping, vol. 75, pp. 661–677, 1998.[82] R. Tafteh, “Austenite Decomposition in an X80 Linepipe Steel,” Master’s thesis, The Univer-sity Of British Columbia, 2011.[83] C. X. Wei, L. Bo, Q. G. Ying, G. Yu, W. Xu, and X. Fu-ren, “Effect of Nb on mechanicalproperties of HAZ for high-Nb X80 pipeline steels,” Journal of Iron and Steel Research, vol. 20,no. 12, pp. 53–60, 2013.[84] P. Mitchell, P. Hart, and W. Morrison, “The effect of microalloying on HAZ toughness,” inMicroalloying’95 Conf. Proc, pp. 149–162, 1995.[85] S. Sarkar, A. Moreau, M. Militzer, and W. Poole, “Evolution of austenite recrystallization andgrain growth using laser ultrasonics,” Metallurgical and Materials Transactions A, vol. 39A,pp. 897–907, 2008.[86] D. Levesque, S. Kruger, G. Lamouche, R. Kolarik, G. Jeskey, M. Choquet, and J.-P. Monchalin,“Thickness and grain size monitoring in seamless tube making process using laser-ultrasonics,”in Advances in Signal Processing for Non Destructive Evaluation of Materials, 2005.[87] M. Militzer, M. Maalekian, and A. Moreau, “Laser-ultrasonic austenite grain size measurementin low-carbon steels,” Materials Science Forum, vol. 715-716, pp. 407–414, 2012.[88] T. Garcin, J. H. Schmitt, and Matthias, “Application of laser ultrasonics to monitor mi-crostructure evolution in INCONEL 718,” Journal of Alloys and Compounds, vol. 670, pp. 329–336, 2016.[89] T. Garcin, “Computational Tools For Metallurgy (CTOME),” Retrieved March 24, 2016.http://ctome.org.[90] T. A. Kop, J. Sietsma, and S. Van Der Zwaag, “Dilatometric analysis of phase transformationsin hypo-eutectoid steels,” Journal of Materials Science, vol. 36, pp. 519–526, 2001.[91] F. Lepera, “Improved Etching Technique for the Determination of Percent Martensite in High-Strength Low Alloy Steels,” Metallography, vol. 12, pp. 263–268, 1979.[92] “”Standard Test Method for Knoop and Vickers Hardness of Materials”, E384-10e2,” 2011.[93] M. Dubois, M. Militzer, A. Moreau, and J. F. Bussiere, “A new technique for the quantitativereal-time monitoring of austenite grain growth in steel,” Scripta Materialia, vol. 42, pp. 867–874, 2000.88

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:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0308782/manifest

Comment

Related Items