Grain Size Control ALSO BY THE AUTHOR The Physical Metallurgy of Microalloyed Steels Grain Size Control T.GLADMAN Formerly Head of Physical Metallurgy, Swinden Laboratories, British Steel Corporation, and British Steel Professor of Metallurgy, University of Leeds, UK. @ MANEY FOR THE INSTITUTE OF MATERIALS, MINERALS AND MINING 10M Communications Ltd is awholly-owned subsidiary of The Institute of Materials, Minerals &Mining (10M3). B0802 First published for 10M3 in 2004 by Maney Publishing I Carlton House Terrace London SWI Y 5DB, UK © 10MCommunications Ltd 2004 All rights reserved. No part of this publication may be reproduced, stored in aretrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying or otherwise, without the written consent ofthe copyright holder. Requests for such permission should be addressed to Maney Publishing. ISBN I 904350 22 4 Statements in the volume reflect the views of the author and not those of the Institute or publisher. Typeset by Old City Publishing, USA Printed and bound in the UK by the Charlesworth Group Contents Preface ix Acknowledgemnts xiii Chapter 1 The Nature of Grain Boundaries 1 1.1 Historical Views of Metallic Structures 1 1.2 Modem Approaches to Po1ycrystalline Materials 4 1.3 Grain Boundary Energy 6 1.4 Grain Boundary Energy and Interfacial Tension 10 1.5 Grain Boundary Forces 11 1.6 Grain Boundary Segregation 15 1.7 Summary 17 References 18 Chapter 2 Granular Arrays - Topology, Grain Shape, Stereology, and Grain Size 21 2.1 Topology 21 2.1.1 General Principles 21 2.1.2 Surface Energy and Space Filling Effects 23 2.1.3 Implications for Grain Shape 24 2.1A Experimental Assessment of Grain Shape 27 2.2 Grain Size 29 2.2.1 Stereological Principles 30 2.2.2 Stereo10gical Aspects ofVarious Polyhedra 35 (a) Mean Areal Intersection 35 (b) Mean Linear Intercept 37 2.3 Experimental Measurement of Grain Size 40 2.3.1 The Comparator Chart Method 40 2.3.2 The Heyn Intercept Method 41 2.3.3 Jeffries Planimetric Method 42 2.3.4 Relationships Between Grain Size Measures 43 References 44 Chapter 2Exercises 45 Chapter 3 Grain Growth 47 3.1 Establishing aGranular Structure 47 3.2 Normal Grain Growth 48 3.2.1 Analytical Treatments 49 3.2.2 Topological Models 56 3.2.3 Computer Based Models 58 3.2.4 Other Approaches 60 3.3 Abnormal Grain Growth 61 3.4 Growth Velocity and Impurity Effects 62 References 65 VI Grain Size Control Chapter 4 Inhibition and Control of Grain Growth 67 4.1 Introduction 67 4.2 Second Phase Particles 68 4.2.1 The Zener Relationship 69 4.2.2 Granular Arrays with Small Volume Fractions of Uniformly Sized Randomly Distributed, Second Phase Particles 71 (a) Growing Tetrakaidecahedra 71 (b) Shrinking Tetrahedra 74 4.2.3 Granular arrays with Small Volume Fractions ofUniformly Sized Grain Boundary Particles 76 (a) Growing Tetrakaidecahedra 77 (b) Shrinking Tetrahedra 78 4.2.4 Comparison of Grain Boundary Particles and Randomly Distributed Particles 79 4.2.5 The Effects of Mixed Particle Sizes 80 4.3 Free Surface Effects 84 4.3.1 Thermal Grooving 85 4.3.2 Variations in Free Surface Energy 88 4.3.3 Substrates and Capping of Foils 92 4.3.4 Secondary Recrystallisation in Foils or Thin Strip 94 4.4 Thermal and Mechanical Cycles 95 4.4.1 Thermal Cycling ofAllotropic Materials 96 44.2 Thermomechanical Treatments 100 44.3 Conventional Controlled Rolling of Steel 102 References 108 Chapter 4Exercises 109 Chapter 5 Second Phase Particle Contents and Particle Coarsening III 5.1 Introduction 111 5.2 Particle Solubility 112 5.2.1 The Solubility Product 112 5.2.2 Using Simple Solubility Product 115 5.2.3 Calculating the Particle Contents for Simple Compounds 118 5.2.4 Solubility Data for Other Grain Refining Additives in Steel 119 5.2.5 Multiple Additions of Grain Refining Additives- Mutually Exclusive Compounds 123 5.2.6 Mutually Soluble Components 124 5.3 Conversions to Volume Fractions 129 5.4 Particle Size and Particle Coarsening 131 5.4.1 Precipitation 131 54.2 Ostwald Ripening 133 54.3 Implications ofthe Wagner Equation for Particle Selection 137 5.4.4 Calculation of Particle Growth Rates 140 References 140 Chapter 5Exercises 141 CONTENTS va Chapter 6 Applications in Grain Size Control 143 6.1 Steel Normalising Treatments 143 6.1.1 Fine Grained Normalised Steels 145 6.1.2 The Grain Coarsening Temperature 147 6.1.3 Secondary Recrystallisation 148 6.1.4 The Grain Size atVery High Temperatures inthe Austenite Range 150 6.2 The Effects ofAluminium Content on Coarsening 151 6.2.1 The Solute Content 152 6.2.2 Effects of Prior Treatment 154 6.2.3 General Observations onAluminium Treatment 157 6.3 Titanium Nitride Technology 158 6.4 Hot-rolled Products 164 6.4.1 Control of the Starting Grain Size 164 6.4.2 Inhibition of Grain Growth inHot Working 166 6.4.3 Retarded Recrystallisation 166 6.5 Secondary Recrystallisation ofCold Worked andAnnealed Strip 168 6.6 The Use of Oxide Particles 169 References 174 Answers to Excersies 177 Index 181 Preface IMPORTANCE OF GRAIN STRUCTURES IN METALS There are very few mechanical properties that are unaffected by either the grain size or the grain characteristics in polycrystalline metals. Two that readily spring to mind are the elas- tic modulus and the specific density. These properties depend on inter-atomic forces and the atomic structure of the particular metals. Most other properties show varying levels of sensitivity to grain size or grain orientations. Probably the best known effect of grain size is that expressed by the Hall-Petch rela- tionship, which states that the yield strength of a polycrystalline metal is related to the inverse of the square root of the grain size, and very significant increases in strength can be attained by maintaining a very small grain size. Similar effects are seen in other strength properties such as the ultimate tensile strength, and the fatigue strength. The effect of grain size on the ductile cleavage transition temperature for b.c.c. metals is very pronounced. The transition temperature is negatively related to the inverse of the square root of the grain size, so that in order to avoid brittle cleavage fracture, the requisite low transition temperatures are attained by refining the grain size. The beneficial effect of fine grain sizes on the impact temperature makes grain refinement the only strengthening mechanism that will simultaneously improve toughness. The effects of grain size on creep properties aremore complex, though no less significant, and depend quite fundamentally on the creep conditions and the predominant creep mecha- nism involved. At relatively high stresses (or low temperatures) diffusion controlled disloca- tion creep predominates, and the creep rate isvirtually independent of grain size.At relative- ly low stresses (or high temperatures) creep occurs by a bulk diffusion process involving transport ofatoms from transverse boundaries to lengthways boundaries, and the creep rate is inversely related to the square of the grain size. Thus, smaller grains allow faster creep rates and reduce the creep strength. At intermediate stresses (or temperatures), creep occurs pre- dominantly by aprocess ofgrain boundary diffusion known asCoble creep, andthe creep rate is inversely related to the cube of the grain size. In order to give the extended lives required in creep service, stresses and temperatures are commonly set at levels in or near to the bulk diffusion regime and therefore fine grain sizes arenot sought for creep resisting applications. The grain sizecanalso influence ductility atthehigh strainrates associated with metal form- ingifthedeformation temperature isinacriticaltemperature range, duetotheformation ofcav- itiesontransverse boundaries. Athigher temperatures, rapid diffusion prevents cavitation, whilst atlowertemperatures dislocation deformation predominates. Thiseffectisalsoseenatlowertem- peratures when creepdeformation ratesareimposed. Intergranular cohesion canbereduced quite dramatically by the segregation of impurities to grain boundaries, and can result inbrittle inter- granular fracture. Inmaterial embrittled inthisway,aductile/intergranular transition temperature may be observed, thetransition temperature increasing with increasing impurity concentration. IX x Grain Size Control Other properties may depend onother grain characteristics such asgrain orientation rather than grain size. The occurrence of a preponderance of grains of a certain orientation is described as atexture, and certain properties are known to depend onthe type oftexture pro- duced. Aclassic example of this isthe dependence of magnetic characteristics on the texture of thin strip for the high-silicon magnetic transformer steels. Development of a strong Goss texture can result in dramatic improvements of the efficiency of transformers. These textures have been developed traditionally by using a secondary recrystallisation treatment which encourages the abnormal growth of specific grains of favourable orientation. Another exam- ple of texture effects is to be found in sheet metal-forming of both ferrous and non-ferrous metals. Suitably textured material canbeproduced having superior sheet metal pressing prop- erties, and also minimising the in-plane variability of strain that gives rise to earing inpressed sheet, which intum reduces the waste from edge trimming. There are literally thousands of publications illustrating the effects of grain size and grain structure on the properties of metals, but these are outside the scope of the present book. It should be clear from the above remarks, however, that the grain size can produce dramatic effects on the properties of metals, and indeed many alloy specifications call for specific levels of grain size control, or for property levels that can only be attained by such acontrol. The same can be said for texture control. DEVELOPMENTS IN THE PAST FIFTY YEARS It is almost fifty years since McLean (1957) published his book on 'Grain Boundaries in Metals'. This book provided an excellent account of the understanding of many matters concerning grains, grain boundaries, grain size, and associated phenomena existing at that time. Indeed, the understanding of grain structure atthat time was very well advanced, inthat many of the phenomenological features, associated with grains and grain/growth had been derived by the many physicists, chemists, and mathematicians, who had been attracted by the very catholic field of metallurgy, and which were soundly based on thermodynamic or other scientific argument. Aparticularly interesting case relates tothe effects of small second phase particles in inhibiting grain growth. The effect of spherical particles had been derived by Zener, using classical surface energy physics for the interaction force and a simple collapsing bubble asthe driving force forboundary motion, giving theclassical Zener equation, R = 4r/3f where the grain radius, R, isrelated to the particle radius, r, and the volume fraction ofpar- ticles, f.McLean rightly concluded that large inclusions inmetals areprobably unimportant inrestricting grain growth, "but inclusions too small to be seen may limit the grain size, as, perhaps, in fine grained steels". This was indeed aprophetic statement, because atthis time, the only solid fact was that additions of aluminium to steel would produce grain refinement and there was little indication ofthe nature of the growth inhibiting features, particularly as growth eventually occurred despite the continued presence of aluminium and it's stable compounds. Itisimportant torealise that electron microscopy, invented inthe early 1930's, was just starting to be used in studies of steel in the mid 1950's, thereby opening up the study of grain refining particles, especially when combined with electron diffraction and, at a later stage, energy dispersive and wavelength dispersive analytical techniques for these small particles. Also, theories of Ostwald ripening were not well developed. The original