Project B4: Heat treatment of high Mn steels



Prof. Dr. rer. nat. Molodov (Institut für Metallkunde und Metallphysik, RWTH Aachen)

Dr.-Ing. Barrales-Mora (Institut für Metallkunde und Metallphysik, RWTH Aachen)

The aim of the project is to develop an optimal method to produce the desired microstructure with respect to grain size, grain shape, grain boundary character, and texture by heat treatment. In project B4, a robust simulation methodology will be developed, which enables the prediction of microstructure and texture evolution during the heat treatment of deformed Fe-Mn-C (Fe-Mn-Al-C) alloys.

During hot forming and heat treatment of steels, thermally activated processes, such as recrystallization and grain growth, are initiated. These processes may be desirable or undesirable depending on the application. Despite intensive studies and numerous theoretical approaches, the scientific problems of the relatively simple phenomenon ‘grain growth’ are still unresolved, even in the two-dimensional case. The reason for this unsatisfactory situation is the strong anisotropy of grain boundary properties, such as energy and mobility, and the complexity of the grain boundary network, which is composed of triple lines and quadruple points. The properties of these microstructural features can have also a significant influence on kinetics, topology, and texture. In this sub-project, the behavior of such complex systems will be simulated with a 3D-Vertex model, which is able to reproduce the physical mechanisms and the topology including anisotropy and complexity of the grain boundary network.



In contrast to grain growth, recrystallization involves the movement of isolated grain boundaries. This physical process is severely affected by the segregation of atoms of alloying elements and precipitation of second-phases. In turn, grain boundary mobility that depends on the character of the grain boundary is also influenced by these drag effects. For accurate recrystallization simulations, it is necessary to consider and model accurately these factors. Additionally, information on the recrystalli­zation nucleation and the local distribution of stored energy (dislocation density distribution) are calculated with specially developed simulation tools (cf. sub-project A7). The evolution of the recrystallized microstructure can be subsequently simulated by means of a cellular automaton model for recrystallization.

Consequentially, the influence of recrystallization and grain growth on macroscopic properties can be determined. Characteristics of the system Fe-Mn-Al-C due to the chemical composition can be considered as the respective input data is determined, for example by projects A3, A7, C1, and C8, and is then implemented in the simulation setups.

Modelling is only reasonable if the real conditions and processes can be reflected and predicted correctly. Therefore, experimental validation of the simulation results and experimental determination of material properties are absolutely essential. Accordingly, detailed microstructure analyses are necessary that are also carried out within the scope of the project. These include optical microscopy, X-ray diffraction (XRD), scanning and transmission electron microscopy (SEM/TEM) to study the macro/micro texture and microstructure as well as hardness measurements and orientation microscopy (EBSD) to determine the recrystallization kinetics.



Simulation of recrystallization

During the 1st research period of the SFB 761 first recrystallization simulations by means of a three-dimensional cellular automaton have already been performed. By a systematic variation of the recrystallization nuclei placed with certain orientations at preferred nucleation sites, good agreement between the experimental data and simulation results was achieved.
The simulations in the 2nd research period were carried out using an advanced version of the cellular automaton that was developed in the 1st period. Initially, the deformed microstructure that was calculated by CP-FEM simulations in project A7 was transferred to the grid of the cellular automaton. By transferring the CP-FEM data the orientations of the deformation texture and the orientation-dependent dislocation densities, which serve as the driving force for recrystallization, were implemented. The number of recrystallization nuclei and their orientations were determined by EBSD measurements on specimens annealed to achieve incipient recrystallization. In order to separate the possible effects on microstructure and texture evolution as accurately as possible, the simulations were performed for cases of low prior deformation (before the onset of shear band formation during cold rolling) in order to avoid nucleation at grain boundaries, which is the dominant nucleation mechanisms at high rolling degrees. Recrystallization nuclei were then allocated to grain boundary cells depending on the dislocation density differences between neighboring deformed grains.
The comparison of the experimental and simulated results of the recrystallization of a 30% cold-rolled and subsequently annealed (700 °C) Fe-28wt.-%Mn-0.28wt.-%C TWIP steel is shown in Fig. 1. The heterogeneous distribution of recrystallization nuclei at grain boundaries was successfully represented by the model. Furthermore and in contrast to simulations with spatially irregularly distributed nuclei, the setting of this nuclei distribution resulted in a slow-down of the recrystallization kinetics at higher annealing times due to previous impingement of grains. This behavior was observed in the experiment as well. In addition to the good agreement between experimental and simulated kinetics, the grain size as well as the texture after recrystallization could be calculated correctly, which again confirms the reliability of the experimentally determined input data.

Fig. 1
Fig. 1: Comparison of the experimental studies with the results from recrystallization simulations. Distribution of the recrystallized grains at the beginning of recrystallization: a) simulation and b) experiment (EBSD mapping); fully recrystallized microstructure: c) simulation and d) experiment; e) recrystallization kinetics.



Experimental investigation of the microstructure and texture evolution during heat treatment

During the 1st research period, the recrystallization kinetics and the ε-martensite transformation during annealing in TRIP/TWIP and TWIP steels were studied. In turn, the experimental work in the 2nd research period aimed to determine the relationship between microstructure and texture development during recrystallization and grain growth.

Essentially, two alloys were studied (V19: Fe-28Mn-0.28C, V43: Fe-23Mn-1.5Al-0.3C). The alloys were first cold rolled with thickness reductions in the range between 10% and 80% in project B2 and subsequently heat treated in B4. After cold rolling both alloys showed microstructures and textures typically developed in fcc materials with a low stacking fault energy (SFE). During plastic deformation dislocation slip, mechanical twinning and at high deformation degrees shear band formation were activated. As a result of twinning, a brass-type texture was developed at medium rolling degrees (30-50%), which consists mainly of the S, Brass, Goss, and Copper twin texture components. At high rolling degrees (60-80%) a weak γ-fiber (<111>//ND) developed due to the formation of shear bands.

The correlation between microstructure and texture evolution during recrystallization was subsequently investigated by EBSD and X-ray texture measurements. At the beginning of annealing, a slight sharpening of the macrotexture due to recovery took place. With the onset of primary recrystallization the texture components of the deformation texture were retained, but their intensity decreased successively. This was accompanied by an increase of the volume fraction of randomly oriented grains. In addition to the described randomization of the macrotexture, new texture components with orientations along the so-called α-fiber (<110>//ND) were formed preferentially. The reasons for this texture development could be obtained by using EBSD data. It has been found that nuclei at grain boundaries and triple junctions formed preferentially with orientations of the deformed matrix phase (see Fig. 2a and b). Furthermore, the formation of annealing twins facilitated the randomization of the texture as well as the formation of the α-fiber texture components A and rotated Goss due to the formation of those annealing twins in Brass- and Goss-oriented nuclei. This is illustrated in Figure 2c. The red arrows indicate the preservation of deformation texture components in the recrystallized grains, whereas the green arrows represent the formation of new texture components as a result of the formation of annealing twins.

Fig. 2
Fig. 2: a) EBSD band contrast mapping of the partially recrystallized alloy V19, b) nucleation at grain boundaries (inset of a)) with misorientation angle θ between the deformed matrix grains 1 and 2 and the neighboring grain, c) microtexture of the deformed matrix and the grain boundary nuclei.




Fig. 3
Fig. 3: a) Correlation between texture development and volume fraction growth of deformation twins in relation to the rolling degree, b) Optimization of mechanical properties by heat treatment.




In an additional study, the results of the microstructure and texture investigations of deformed and annealed samples were used to develop a method that enables to determine the optimal rolling and annealing parameters for a specific processing route by texture analysis. It is known that, for a suitable combination of cold rolling and subsequent recovery annealing, the yield strength-ductility ratio of TWIP steels can be manipulated, if the mechanically induced twins are thermally stable. First, alloy V43 was specifically chosen based on the SFE maps calculated in project A5 and then rolled in project B2 with varying rolling degrees. The samples with the highest fraction of mechanical twins and the lowest fraction of shear bands were determined by texture analysis. The optimal annealing time for the transition between recovery and the beginning of recrystallization was also set by analyzing texture data. The reliability of the texture data was verified using deformation simulations (correlation of the fraction of deformation twins with texture data) in project A7 (see Fig. 3a). The optimization of the mechanical properties, which should be achieved through the processing route, was finally validated in project C2 (Fig. 3b).




Previous Phase

The objective of this subproject is the development of optimal processes for microstructure tailoring during thermal treatment with particular focus on grain size, grain topology, grain boundary character and texture.

During hot-working and thermal treatment of steels, thermally activated phenomena are triggered, such as recrystallization and grain growth. These phenomena can be desired or undesired depending on the application, since they alter the microstructure and thus, the properties of materials. In order to tailor materials by the modification of their microstructure, it is necessary to understand the underlying physics of these phenomena. Despite intensive studies and many theoretical approaches, the comprehensive description of an apparently simple phenomenon, such as grain growth, is still unresolved, even in the two-dimensional case. The reason for this unsatisfactory situation is the strong anisotropy of grain-boundary properties such as structure, mobility and energy. The complication increases also if the topology of real polycrystals is considered where the boundaries intersect in triple lines and quadruple junctions which can exert a substantial influence on the evolution of the microstructure. One part of this subproject is the simulation, by means of a 3D-Vertex Model, to study the behavior of such complex system. This model is capable of reproducing in computer simulations the fundamental physics underlying the mechanisms of grain growth, properly accounts the topological network of the polycrystal and considers correctly the anisotropic properties of the grain boundaries.

In contrast to grain growth, during recrystallization the movement of isolated grain boundaries takes place and, for this reason, the segregation and precipitation of second phases must be considered since they can strongly affect the boundary motion. For the modeling of recrystallization, it is necessary to obtain the information on the nucleation mechanisms of recrystallized grains and the local distribution of the stored energy (dislocation density distribution). Both can be taken into account by means of particular simulation tools developed for this purpose. The grain boundary velocities determined in subproject A6 depend, of course, on the local distribution of the driving forces and are influenced by dragging factors (second phases, impurity atoms, etc). Since the information delivered by A6 relates only to the microscopic scale, a correlation with the mesoscopic scale needs to be accomplished.  For this purpose, a cellular automata model is utilized, which is capable of predicting the evolution of the microstructure and texture development.

Generally, the macroscopic effects of the processes recrystallization and grain growth that take place during thermal treatments will be determined. The chemistry-driven properties of the system Fe-Mn-C will also be considered since the data delivered by A2 are obtained in conjunction with subprojects A1 and A2 (ab-initio simulation) which address the influence of chemistry of the material properties.

Modeling is only meaningful if it delivers or predicts correctly the real behavior of physical systems. Experimental validation of the model is, for this reason, necessary. Moreover, experiments provide also important natural data for the models (e.g. information about the nucleation mechanisms). Accordingly, detailed microstructure investigations will be indispensable, which are accomplished in the context of this subproject. For this, different experimental techniques will be utilized, optical microscopy, X-ray-diffractometry (XRD), and scanning electron microscopy (SEM). Additionally, hardness measurements and orientation microscopy (EBSD) will be used for a determination of recrystallization kinetics.