Bainite in Steel: A Century of Microstructural Discovery, Mechanistic Debate, and Dynamic Transformation

Bainite in Steel: A Century of Microstructural Discovery, Mechanistic Debate, and Dynamic Transformation C. Garcia-Mateo ([email protected].es) National Center for Metallurgical Research (CENIM-CSIC) Madrid, Spain May 2025 This research is funded by the European Union under the RFCS project WarP – Grant Nº 101112425. First and foremost……… First and foremost……… Happy Birthday, Silesian University! Eighty years of academic excellence, innovation, and community—a remarkable milestone. It’s an honour to be part of this celebration and to contribute to the legacy of such a distinguished institution. First and foremost……… I would also like to express my gratitude to Professor Adam Grajcar—for the kind invitation to speak today, and for many years of fruitful scientific collaboration and personal friendship. It is an honour to work alongside you. INDEX ❑Bainite historical context ➢Discovery and Early Characterization ➢Emerge of competing Theories ➢Characterization through time ➢Evolution of theories through time ❑Explaining Bainitic transformation ➢Transformation Mechanisms ➢Displacive & Diffusionless transformation. ➢Evolution of the transformation ➢Morphologies ➢Thermodynamics. To-Line ➢Understanding the C journey in bainite ❑Assessment of the contributing factors to the scale of bainite ➢Ways and means to strengthen austenite ➢Bainite growth to figure out some contributions. ➢Models based on static inputs/contributing factors ➢Beyond the models. Case Study ➢Dynamic approach. Plate thickness evolution with transformation. INDEX ❑Bainite historical context ➢Discovery and Early Characterization ➢Emerge of competing Theories ➢Characterization through time ➢Evolution of theories through time ❑Explaining Bainitic transformation ➢Transformation Mechanisms ➢Displacive & Diffusionless transformation. ➢Evolution of the transformation ➢Morphologies ➢Thermodynamics. To-Line ➢Understanding the C journey in bainite Steel Early Characterization (Pre-1930s) Fe-C Phase Diagram Sir William Chandler Roberts-Austen (1897) published the first complete Fe-C diagram, synthesizing earlier 19thcentury studies on carbon's role in steel. Ferrite and Cementite •These were long-established constituents of steel, forming the basis of microstructures like pearlite. Ferrite (α-iron) body-centered cubic phase, while cementite (-Fe₃C) is an iron carbide. Pearlite •Identified as a lamellar mixture of ferrite and cementite, formed during slow cooling of austenite. By the 1920s, finer pearlite variants (later termed "troostite") were recognized. Martensite •A hard, needle-like phase formed by rapid quenching of austenite. Its transformation mechanism was studied extensively in the early 20th century. Edgar C. Bain Discovery and Early Characterization (1920s-1940s) In the 1930s, E.S. Davenport and Edgar Bain discovered a new steel microstructure that: •Occupied a transformation temperature range between pearlite and martensite. Initially, they provisionally called it "martensitetroostite" as it appeared intermediate between the then-known martensite phase and what was called troostite (later identified as fine-pearlite). •This microstructure was subsequently named "bainite" by Bain's colleagues at the United States Steel Corporation, though scientific acceptance of the term was slow, with metallurgical textbooks as late as 1947 failing to mention bainite by name. Microstructures in a eutectoid steel: (a) Pearlite formed at 720 °C (b) bainite formed at 290 °C (c) bainite formed at 180 °C (d) martensite. Optical microscopy (OM) was the primary characterization tool available to metallurgists. Bain and Davenport made several fundamental observations after isothermal holding experiments, where austenitized steel specimens were rapidly quenched to specific intermediate temperatures (250– 550°C) and held to allow transformation : Discovery and Early Characterization (1920s-1940s) •C-curve kinetic behavior in time-temperature-transformation (TTT) diagrams. They established the temperature range for bainite formation: approximately 125-550°C, depending on alloy content. •They identified two distinct morphologies/ranges: 'upperrange' bainite forming at higher temperatures and 'lower-range' bainite forming near the martensite start temperature. Bain, E.C. (1939). Functions of the Alloying Elements in Steel. American Society for Metals. Davenport, E.S., & Bain, E.C. (1930). Transactions of the Metallurgical Society of AIME, 90, 117-154. Crystallographic Understanding and Nanostructured Bainite (2000s-2010s) Electron backscatter diffraction (EBSD) coupled with SEM in the late 1990s and early 2000s revolutionized our understanding of bainite's crystallographic nature. •Though EBSD allowed researchers to map crystallographic orientations across larger areas, providing statistical data on orientation relationships between phases. •A significant discovery during this period was the confirmation that bainitic ferrite can have a tetragonal crystal structure (BCT) rather than the cubic (BCC) structure traditionally associated with ferrite. This tetragonality results from carbon supersaturation in bainitic ferrite and supports the displacive transformation theory. •The early 2000s also saw the development of nanostructured bainitic steels, characterized by: •Extremely fine bainitic ferrite plates (tens of nanometers thick). •Retained austenite with two distinguishable morphologies: thin films between ferrite plates and larger blocks. •High dislocation densities (~5.1±2.7×1014 m-2), similar to martensitic structures. •Excellent combinations of strength, toughness, and ductility. Modern Analytical Techniques (2010s-Present) The most recent decade has seen the application of atom probe tomography (APT) to bainitic steels, allowing three-dimensional mapping of elements at the atomic scale. This has provided unprecedented insights into carbon distribution within bainitic structures, clarifying the nature of carbon supersaturation in bainitic ferrite and carbon enrichment in retained austenite. In-situ Xray and Neutron Difraction can track lattice parameter changes in both ferrite and austenite, providing direct evidence of carbon partitioning and its effect on stabilizing retained austenite. b f b 1920s-1930s: Initial Observations Early speculations: Initial hypotheses suggest a hybrid mechanism combining aspects of diffusioncontrolled and shear transformations 1950s-1960s: Emergence of Competing Theories Displacive Theory Foundations •TEM observations reveal plate morphology and dislocation structures resembling martensite •Concept of invariant-plane strain shape deformation proposed •Theory formalized: Bainite grows via shear mechanism without carbon diffusion during transformation Diffusive Theory Development •Alternative hypothesis suggests carbon diffusion controls growth rate •Proposes similarity to Widmanstätten ferrite formation •Argues surface relief effects don't preclude diffusion control Two theories about bainite transformation mechanisms. Evolution through time. 1970s-1980s: Experimental Differentiation Key advances through TEM/SEM. Displacive evidence: •Discovery of carbon supersaturation in bainitic ferrite •Observation of incomplete reaction phenomenon (transformation stops at T₀ curve) •Mechanical stabilization effects unique to displacive transformations Diffusive counterpoints: •Cementite precipitation patterns suggest carbon migration •Temperature-dependent transformation rates cited as diffusion evidence 1990s: Quantitative Modelling Bhadeshia's kinetic model (1982-1992) successfully predicts: •Autocatalytic nucleation behavior •Temperature dependence of plate thickness •Incomplete transformation phenomenon Model strongly supports displacive mechanism with subsequent carbon partitioning Two theories about bainite transformation mechanisms. Evolution through time. 2000s: Crystallographic Validation •Tetragonal distortion in bainitic ferrite (δ ≈ 1.0008) •Kurdjumov-Sachs orientation relationships with austenite Both features match martensitic (displacive) transformation signatures 2010s: Atomic-Scale Confirmation Atom probe tomography (APT) shows: •Carbon supersaturation (0.1-0.3 wt%) in fresh bainitic ferrite •Delayed carbide precipitation (seconds-minutes after transformation) •Carbon gradients at α/γ interfaces consistent with post-transformation partitioning Two theories about bainite transformation mechanisms. Evolution through time. 2020s: Current Status Displacive mechanism dominates for ferrite formation •Displacive Theory: Fundamentally relies on a displacive mechanism, similar to martensite. This means a shear-like transformation with coordinated atom movements and little or no diffusion during the initial growth step. Carbon diffusion occurs after the initial ferrite formation. •Diffusional Theory:Initially, it considered a ledge propagation growth mechanism. However, the newer considerations accept a mixed mechanism that is reconstructive and displacive growth of the bainitic ferrite Key Points: •The core difference lies in whether atom movement is primarily coordinated (displacive) or diffusioncontrolled during the initial growth of the ferrite. •Both theories acknowledge the role of carbon diffusion in the overall bainite transformation process, particularly concerning carbide precipitation and the incomplete reaction. Fielding, L.C.D. The Bainite Controversy. Mater. Sci. Technol. (United Kingdom) 2013, 29 , 383–399, doi:10.1179/1743284712Y.0000000157. Two theories about bainite transformation mechanisms. Evolution through time. INDEX ❑Bainite historical context ➢Discovery and Early Characterization ➢ Emerge of competing Theories ➢Characterization through time ➢Evolution of theories through time ❑Explaining Bainitic transformation ➢Transformation Mechanisms ➢Displacive & Diffusionless transformation. ➢Evolution of the transformation ➢Morphologies ➢Thermodynamics. To-Line ➢Understanding the C journey in bainite Diffusional Diffusionless & Displacive Displacive H.K.D.H Bhadeshia. Bainite in Steels. The Institute of Materials (2001) London. 3d Edition Explaining Bainite (Personal point of view and experience) Bs Ms Bainite. No-Equilibrium . Atomic Mechanisms of Transformation Pattern of atoms changes during transformation, a disciplined motion of atoms (less than one inter-atomic spacing) necessarily leads to a change in the shape of the transformed region, and like any deformation, such changes cause strains in the surrounding material. Austenite Displacive Transformation Atomic correspondence IPS shape change with a significant shear component Diffusionless Austenite Austenite Interface Interface ReconstructiveTransformation No atomic correspondence No shape change with shear component Possible composition change Ferrite Ferrite Austenite Interface Austenite Ferrite FCC BCC/BCT FCC Bainite. Displacive transformation : Invariant-plane strain shape deformation with large shear component (S) + small dilatation normal to the plane (d) Bainite. Nucleation sites and growth evolution At transformation temperature Cementite (𝜃) can be avoided by the introduction of Si +  Carbide free Bainite (CFB) ++b •Silicon has a limited solubility in cementite (Fe3C), retards the nucleation and growth of cementite. Bainite. Nucleation sites and growth evolution At room temperature +  Si rich steels No Si +b+ ++b At room temperature Bainitic ferrite (b) +2nd phases  + ’ At transformation temperature ++   + Two main morphologies of austenite ❑Blocks between the sheaves of bainite ❑Thin films, in between the plates Guo, Y.; Feng, K.; Lu, F.; Zhang, K.; Li, Z.; Hosseini, S.R.E.; Wang, M, doi:10.1016/j.apsusc.2015.08.132. Thermodynamics Nucleation : Under paraequilibrium condition (only C diffuses) Bhadeshia, H. K. D. H. (2001). Bainite in Steels. Garcia-Mateo, C. Mater. Sci. Eng. A () pN GG    →+  1 3.5463 ( ) 3499.4 N G T K J mol− =− GN defines the minimum free energy change necessary in any steel, in order to nucleate bainite. It is a universal nucleation function : Thermodynamics Growth : Diffusionless growth. No change of chemical composition from the parent phase to the product phase. Bhadeshia, H. K. D. H. (2001). Bainite in Steels. Garcia-Mateo, C. Mater. Sci. Eng. A Thermodynamics Growth : Diffusionless growth. No change of chemical composition from the parent phase to the product phase. Bhadeshia, H. K. D. H. (2001). Bainite in Steels. Garcia-Mateo, C. Mater. Sci. Eng. A Displacive transformation : Invariant-plane strain shape deformation with large shear component. Stored energy of the ferrite due to the displacive mechanism of transformation 400 J/mol = GSB Thermodynamics Growth : Diffusionless & Displacive Bhadeshia, H. K. D. H. (2001). Bainite in Steels. Garcia-Mateo, C. Mater. Sci. Eng. A Stored energy of the ferrite due to the displacive mechanism of transformation 400 J/mol = GSB Free energy/ J mol-1 Temperature BS MS → G  N G + → G SB G N G 𝐺𝑁 𝛼′represents the critical value of the free energy change Δ𝐺𝛾→𝛼 𝑀𝑠    →  Ns GMG needed before the athermal, diffusionless nucleation and growth of martensite becomes possible Thermodynamics. Critical Transformation T () pN GG    →+  40 Does 100% of the austenite transform into bainite? How far the transformation can proceed? NO. As austenite gets richer and richer in C there is a limit where it is thermodynamically impossible for the transformation to proceed. Incomplete reaction phenomena, T0 line () pN GG    →+  Cbulk 𝑥𝛾𝛼 ′′ 𝑥𝛼𝛾 ′′ Carbon content  T1 𝐺𝛾 𝐺𝛼 𝐺𝛼+ 𝐺𝑠 𝐴𝑒1 ′′ 𝑇0 ′𝐴𝑒3 ′′ 𝑥𝑇0 ′ T1 Gibbs free energy Temperature Bhadeshia and Christian (1990) Metall. Trans. A, 21 A(4):767–797. Incomplete reaction phenomena, T0 line The lower the transformation temperature, the further the transformation can proceed ➔ higher % C in austenite is allowed in austenite; in other words ➔ higher fractions of bainitic ferrite are obtained. TVb ҧ𝑥 Stone, H. J., M. J. Peet, H. K. D. H. Bhadeshia, P. J. Withers, S. S. Babu and E. D. Specht (2008). Proc. R. Soc. A Q2Does all the  receive the same amount of C? A-Initial pattern consisted of narrow peaks exclusively from austenite. B-Their intensity decreased as transformation progressed. In addition, broad ferrite peaks appeared along with a second set of broad austenite peaks at lower 2 angles than the initial ones. The latter is caused by the partitioning of carbon from the bainite into the residual austenite, thereby leading to an increase in its lattice parameter. A B In-Situ Observations in a synchrotron beam line, 2nd set of experiments , higher resolution. In-Situ Observations in a synchrotron beam line, 2nd set of experiments , higher resolution. Q2Does all the  receive the same amount of C? Block austenite Ferrite Thin films A C Austenite films entrapped between neighboring subunits of bainitic ferrite Blocks of residual austenite located between the sheaves of bainite 11 at.%C 12 at.%C 12 at.%C 7 at.%C Fe-4.3C-2.8Si-1.2Mn-1.3Cr (at.%); 200 oC, 10 days Ex-Situ Observations in a 3D Atom probe tomography Q2Does all the  receive the same amount of C? Nanoscale Austenite films Sub-micron Austenite block 6 at.%C Q2Does all the  receive the same amount of C? A2No Size/location is paramount Tracking the C in its “journey” from  ➔  Q3-Does all the C escapes from ? F.G. Caballero Acta (2008) F.G. Caballero Scripta (2012) 0.00 0.25 0.50 0.75 1.00 1.25 150200250300350400450500550 Carbon content in ferrite, at.% Transformation temperature, ºC Medium Carbon Low Silicon Steel Medium Carbon High Silicon Steel High Carbon High Silicon Steel PE with Austenite PE with Cementite Carbon Supersaturation in Ferrite APT values away from defects or clusters Q3-Does all the C escapes from ? F.G. Caballero Acta (2008) F.G. Caballero Scripta (2012) 0.00 0.25 0.50 0.75 1.00 1.25 150200250300350400450500550 Carbon content in ferrite, at.% Transformation temperature, ºC Medium Carbon Low Silicon Steel Medium Carbon High Silicon Steel High Carbon High Silicon Steel PE with Austenite PE with Cementite APT values away from defects or clusters Q3-Does all the C escapes from ? Carbon Supersaturation in Ferrite Why? Analysing XRD Spectra A tetragonal phase with the space group I4/mmm was introduced as initial structural model in the Rietveld refinement. Since lower values of all residuals indicate a better fit, it is suggested that bainitic ferrite did not present a cubic structure, but a tetragonal structure . cubic tetragonal Garcia-Mateo, C et al. Low Temperature Bainitic Ferrite: Evidence of Carbon Super-Saturation and Tetragonality. Acta Mater. 2015, 91 , 162–173, doi:10.1016/j.actamat.2015.03.018. Transf. T/ ºC Time/ h a/ Ǻc/ Ǻc/a C/ wt.% 220 24 2.857 2.880 1.0087 0.19 250 14 2.856 2.878 1.0078 0.17 300 5 2.857 2.877 1.0072 0.16 350 4.5 2.859 2.876 1.0059 0.13 Quench (´+)---- 2.856 2.932 1.0266 0.587 Tetragonality. XRD values c/a = 1+ 0.045 C (wt.%) Lose of tetragonality as transformation temperature increases 50 nm     C = 0.75 at. %  Reasonable explanation?? 220 ºC for 7 days. Austenite: a=b=c=0.361 nm / 3.61 Ǻ Bainite: a=b=0.287 nm / 2.87 Ǻ c=0.290 nm / 2.90 Ǻ c/a=1.010 Measuring the lattice parameters of bainitic ferrite by the distance between the phase contrast peaks. Tetragonality. TEM / HR-TEM Garcia-Mateo, C et al. Acta Mater. 2015, 91 , 162–173 Its growth is displacive & difussionless Austenite Displacive Transformation Atomic correspondence IPS shape change with a significant shear component Diffusionless Austenite Austenite Interface Interface ReconstructiveTransformation No atomic correspondence No shape change with shear component Possible composition change Ferrite Ferrite Austenite Interface Austenite Ferrite How Bainite Grows Because there is plastic deformation there are dislocations/twins that help to relax the strain (mainly in austenite). The local increase in dislocation density caused by the yielding of the austenite, halts the movement of the glissile semi-coherent interface ➔ each plate only achieves a limited size (<PAGS) and impedes its thickening….. F.G. Caballero et al. / Acta Materialia 59 (2011) 6117 How Bainite Grows Ways and means to strengthen austenite 𝜎(𝑌𝑆)= 𝜎𝐹𝑒 + 𝜎𝐼𝑛𝑡𝑒𝑟𝑠𝑡𝑖𝑡𝑖𝑎𝑙𝑠 (𝐶,𝑁) + ෌ ሶ 𝑖𝑘𝑖 𝜎s𝑠 𝑖+ 𝐾ത 𝐿−1/2 + 𝑘𝑃𝐿−1 + 𝐶 𝑝0.5 Strengthening ➔ impede dislocation movement ❑Interstitials (C,N) have strong effect ❑Dislocations (e.g. deformed austenite) ❑Small austenite is stronger than bigger austenite On the course of the transformation, plates collide and impede further lengthening or thickening → hard impingement Bowing of transformation interface at strong pinning points, particularly prominent in regions identified by arrows V. Ruiz-Jimenez Materials 2021 Hu et. Al. Mater. Letters 2014 Chang, L. C. and H. K. D. H. Bhadeshia (1995). Mater. Sci. Technol. How Bainite Grows Identified important factors (so far) affecting the scale Strength of austenite  (wt.%, L, 𝝆,T) 𝜎(𝑌𝑆)= 𝜎𝐶ℎ𝑒𝑚𝑖𝑐𝑎𝑙 𝐶𝑜𝑚𝑝 + 𝐾ത 𝐿−1/2 + 𝐶 𝑝0.5 + 𝑇𝑒𝑚𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒 Hard impingement  (∆𝐺𝛾→𝛼) ∆𝐺𝛾→𝛼 = 𝑓(𝐶ℎ𝑒𝑚𝑖𝑐𝑎𝑙 𝑐𝑜𝑚𝑝𝑜𝑠𝑖𝑡𝑖𝑜𝑛, 𝑇𝑒𝑚𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒) Dislocation density  (𝑻) 𝑝 = 𝑓(𝑇) S.B. Singh, Mater. Sci. Eng. A 245 (1998) 72–79 Description of bainite scale by ANN model Bainitic ferrite plate thickness Dislocation density 𝝆 (𝑻) not included Yang et. al. Mater. Sci. Eng. A 748 (2019) 16-20 Bainitic ferrite plate thickness Dislocation density 𝝆 (𝑻) not included 1E+14 1E+15 1E+16 050 100 150 Dislocation density/ m-2 t/ nm Sub Nano (c) T  dynamic recovery may take place  which in turn eases  t Trapping of C in dislocations as Cotrell in the vicinity of the interface might exert an extra contribution  as T  Strong correlation𝝆 and  t Dislocation density 𝝆 (𝑻) effect exists Dislocation density 𝝆 (𝑻) effect exists S.H. He et al. / Acta Materialia 135 (2017) As transformation progresses, is the plate size smaller? Lin, S. et al.. Mater. Charact. 2022, Transformation Transformation Transformation Dislocation density Whole system is dynamic Whole system is dynamic 0 0.005 0.01 0.015 0.02 0.025 0 10000 20000 30000 40000 50000 60000 DRCL/ % s-1 Time/ s 900°C-280°C 860°C-280°C Transformation Transformation rate 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0 5000 10000 15000 20000 RCL (%) DRCL (%/s) time (s) tmax,Vmax Incubation period Fast transformation up to the Max. Transformation kinetics considerations Typical stages during bainitic transformation THE SYSTEM IS DYNAMIC Deceleration of the transformation, very sluggish Logic and metallurgical data seem to indicate that as transformation progresses, plate size should be smaller Transformation Beginning End Why weren't we successful? Thickness of a plate can increase even after lengthening has halted until the chemical force is exhausted by the accumulation of strain energy, thermoelastic equilibrium, or by the presence of adjacent parallel plates, hard impingement. *Chang, L.; Bhadeshia, H.K.D.H. Mater. Sci. Technol. 1995, 11 , As transformation progresses, is the plate size smaller? ➔ NO Calculations were made considering the parameters at each stage (DoT) → C (∆G,YS) THE SYSTEM IS DYNAMIC Calculations were made considering the parameters at each stage (DoT) → C (∆G,YS) THE SYSTEM IS DYNAMIC Max. rate of transformation (grey band) Calculations were made considering the parameters at each stage (DoT) → C (YS,∆G) •Theory still predicts that plate thickness becomes thinner as transformation progresses. •The thickening of the plates is greater after the Max. rate of transformation (grey band) THE SYSTEM IS DYNAMIC Experimental measurements at each DoT are considering all the plates ➔ those formed at previous stages and the actual stage ➔ is an accumulative measurement. THE SYSTEM IS DYNAMIC Experimental measurements at each DoT are considering all the plates ➔ those formed at previous stages and the actual stage ➔ is an accumulative measurement. THE SYSTEM IS DYNAMIC Calculate the thickness of the plates formed at a certain stage (DoT), t*b i.e. not considering what it was formed at previous stages