Investigations on Decreased High Temperature Ductility of Different Continuously Cast Steel Grades

Continuous casting of premium steel grades requires a process with a high degree of precision and the knowledge about the mechanical behavior of the steel at temperatures above 800 °C. Herein, several origins of effects which lead to unwanted impairment of the hot strand shell like segregations, size, amount, kind, and distribution of precipitates as well as porosities from a metallurgical point of view are dealt. The systematic description of potential defect reasons helps to predict harmful operation parameters in context with the chemical composition of steel grades. A compilation of results from experiments at Department of Ferrous Metallurgy of RWTH Aachen University is complemented by a literature review. It is focused on the high temperature ductility and the underlying mechanisms inside the solidifying steel. Finally, potential measures to adjust the continuous casting process to prevent defects are elaborated.

aforementioned test temperature, and in situ, whereby the liquidus temperature was exceeded, e.g., at IEHK's continuous casting (CC) simulator. Nowadays, only the latter two methods were used, whereby only the in situ strategy maps effects such as grain coarsening, precipitation behavior, or hot cracking according to the CC process. [1] In Austria at MU Leoben testing methods as the "mold simulator" and the "submerged split-chill tensile test" (SSCT) were developed to determine and describe the ductility behavior under CC similar conditions. [2,3] In addition, there were various high temperature bending methods which investigate the strain, the frequency of crack formation, and the maximal crack length. [4][5][6][7] Torsion and compression tests were also used to determine critical strains. [8,9] To investigate mechanical loads during solidification as well as hot tearing, the high temperature bending simulator was developed at the IEHK. This enabled the deformation of the ingot shell with liquid core, whereas force and displacement data were recorded. Alloy-specific characteristic values for the compressive load were determined in hot compression tests.

Reduction of Cross-Sectional Area and Ductility Influencing Factors
To prevent internal and surface cracking in continuous casting of steel, it is essential to understand the mechanisms in the steel during solidification and cooling. Complex interactions between alloy composition, and external conditions like cooling rate and strain rate are proven to have a great impact on each steel, individually.
Hot tensile tests to obtain the ductility are inter alia evaluated in terms of the relative reduction of area RA (%) of the specimen's cross section. Figure 1 shows the effect of changes in the steel itself differentiated in temperature ranges on the RA according to the recent results of research at the IEHK. Two [10,11] or three, [12] respectively, low ductility zones LDZ of composition-depended emphasis can be identified. LDZ1 describes the temperature range before the steel reaches its maximum ductility at the point T BD , the temperature of the brittle-ductile transformation. LDZ1 is mainly influenced by the degree of solidification, as this determines whether the dendritic crystals are already firmly connected. At the zerostrength temperature, the adjacent dendrites meet this criterion for the first time resulting in the material being able to bear a small load but being still too weak to deform before failing. There is controversial information on T sol being either higher than the zero-ductility temperature (ZDT), [10,[13][14][15] or slightly lower. [2] For peritectic steel, grades were even found that the ZDT equals the ZST if body-centred cubic δ-Fe is the first phase. The calculated f s at this point was given with 0.8. [16] Cracks, that occur in the LDZ1 and show a dendritic structure at the fracture surface, are referred to as "hot tearing." [17][18][19] The LDZ2 defines the second temperature range where the ductility is impaired by mechanisms inside the steel, starting with the ductile-brittle transition at T DB . In case, the deterioration starts because of the formation of secondary MnS or low melting FeS, it is referred to as LDZ2a, whereas LDZ2b is induced by the precipitation of nitrides. The material is the weakest when the transition from austenite to ferrite starts at primary austenite grain boundaries. A thin layer of ferrite makes the steel particularly vulnerable. [20] After the phase transition, the materials' ductility recovers. Just like the evolving of ferrite, other phase transformations are impairing the ductility.
For the interpretation of these results, it must be noted that the testing method [21][22][23] and the testing parameters have a significant influence on the materials' behavior. Variation of parameter Figure 1. Schematic representation of RA curve for high temperature ductility, depending on: temperature, individual composition, formed precipitates, and phase transformations, representation based on data from literature. [2,10,11,58,60] www.advancedsciencenews.com sets [20,[24][25][26][27][28] in hot ductility testings thus provide important information on possible adjustments of the casting process to match the materials' properties. Figure 2 shows the qualitative influence of various alloying elements. Ratios and products of alloying elements, that form a joint precipitated which affects the ductility are considered. Nitrogen alone is mostly found to have no harmful impact, but Al-and Nb nitrides have. [29,30] ratios Mn/S, Ti/N, and B/ N are included, because a stochiometric ratio of these alloying elements, prevent the formation of precipitates that have a strong negative impact on the hot ductility, if exceeded. A sufficient Mn content protects the formation of low melting (Fe,Mn)S, whereas TiN and BN bind the nitrogen resulting in inhibited AlN formation at lower temperatures.

Segregation
Segregation is caused by different solubilities of alloying elements in the liquid steel and the solid matrix. Rejection of elements from the solid phase leads to an increase in the residual melt and consequently to a positive microsegregation. [31][32][33] The primary solidification in steel castings can be columnar dendritic, equiaxed dendritic, or globular, depending on the boundary conditions. [34][35][36][37] During dendritic solidification, the microsegregated melt gets captured between the dendritic arms. As globulites have a round shape and no branches, this structure provides bigger coherent area. Caused by this geometric effect, a mesosegregation, which is larger than microsegregation, can be found ( Figure 3). Figure 2. Qualitative representation of the influence of increasing contents of alloying elements on the hot ductility, blue: improvement, red: deterioration; data from literature. [13,20,25,[28][29][30]53,[61][62][63][64][65][66][67][68] summarized by Unterberg. [69]  The line scans in Figure 4 clearly show the relationship between the as-cast structure and the resulting microsegregation distribution and intensity.
This can in turn have a major impact on the microstructure development, as Figure 5 underlines by showing the microstructure after reheating and a hot compression test of a high manganese steel. Thus, different concentrations are present in a relatively small sample area, which significantly influence the microstructure and its transformation during cooling and deformation, creating narrow bands of adjacent different phases. The brownish areas represent the former dendrite stems in which a temperature-induced martensitic substructure is present, which is only possible due to the decrease in Mn and C as well as an increase in Al concentration. This effect can be explained by Equation (1) for martensite start temperature from Hollomon et al., [38] which shows the impact of the elements. [39] The remaining light blue/purple areas are to be interpreted as Mn-rich austenite.
Segregation, which is rather caused by the casting process than by the solidification structure, is called macrosegregation. The process leads to fluid flows that concentrate the microsegregated melt in the middle part of the casting semi. [34,40,41] This mechanism can be suppressed or mitigated either by applying mechanical or thermal soft reduction or a change of the as-cast structure by electromagnetic stirring. [37,[42][43][44][45]

Precipitates
Precipitates are secondary, nonmetallic phases which form due to a local exceedance of the solubility product. Thus, they are directly depending on the local concentrations and therefore Reproduced with permission. [71] Copyright 2017, ASMET.  from the chemical composition (internal parameter), cooling rate (external parameter). Figure 7 shows exemplary two quantifications of influencing factors on the RA in the temperature range of the second ductility minimum LDZ2. An increase in the particle size clearly leads to a great improvement of the hot ductility and indicates that the finer the particles are, the more significant is their influence on the ductility. [46,47] The right diagram shows how a reduction of the total amount of precipitating sulfur benefits the ductility at a testing temperature of 1000 C, same applies to an increase in the Mn content. The higher Mn contents were found to shift the size distribution of the MnS to bigger particles, whereas significantly lowering the amount of particularly small sulfides. [48] That in turn fits well with the results from the left diagram and gives an idea of the complex interrelationships of various influencing variables. The fact, that especially small precipitates are impairing the materials ability to resist loads or stresses, leads to a special sensitivity to surface and subsurface cracks in materials that are likely to produce a high number of small precipitates, like for example microalloyed steels. The particle size of the precipitates like TiN, AlN, Nb(CN), and MnS are dependent on the cooling rate. With a high cooling rate, the particle size decreases and it is  . Examples for quantified influencing factors on the RA in LDZ2, left: increasing particle size in Ti-containing steels beneficial for hot ductility, [46,47] right: decreasing amount of precipitated S and increasing Mn content improve the ductility in low C steel with 0.03-0.05 wt% C. [48] www.advancedsciencenews.com l www.steel-research.de steel research int. 2021, 92, 2100323 easier for cracks to propagate along the chain of precipitates on the grain boundaries. Small precipitates at grain boundaries result in a pinning of the grain boundaries and thus lower the ductility. This effect is particularly pronounced in microalloying steels, especially for Nb(CN) precipitates. [22,23] Continuous casting employs a high cooling rate and a cyclical thermal load on the surface of the strand. Combined, the effects lead to an increased risk for transverse corner cracks in microalloyed steels. [47,49] An example for a surface crack propagating along primary austenite grain boundaries in a microalloyed steel is shown in Figure 8. Small-scale precipitations were detected with SEM-EDX near the crack, which was filled with an oxide layer during cooling.
Thermodynamic calculation of the phase fraction of precipitates depending on the temperature can help to understand the reasons for an impaired hot ductility, as shown in Figure 9. Especially in the LDZ2, the drop of the RA is related to the formation of first AlN and second NbC, in particular due to the former explained effect of grain boundary pinning.
To monitor and validate the effects of the chemical composition, segregation, and the resulting precipitation behavior, the IEHK's slow solidification furnace is used to produce large crystals and nonmetallic inclusions. Figure 10 shows precipitates of MnS and AlN as a result of this slow solidification process. Depending on the Al content present, the precipitation sequence results and the first formed phase serves as a heterogeneous nucleation surface for the subsequent phase. Thus, the chemical composition of the melt influences the chronological order when the solubility product log K [%Mn][%S] ¼ À17 026/T þ 5.161 or log K [%Al][%N] ¼ À12 950/T þ 5.58 is exceeded and thus the respective phase is precipitated. [50,51] The solubility product is used  . RA from hot tensile tests of a microalloyed steel grade and the related Thermo-Calc © calculations; the temperature of precipitate formation can be compared with the partially impaired ductility value. Reproduced with permission. [25] Copyright 2015, Metall Mater Trans B. Figure 10. SEM-EDX analysis Â1000 of a) X30MnAl17-1 and b) X30MnAl17-5; slow solidification process with a solidification speed of v ¼ 15 mm h À1 . [72] www.advancedsciencenews.com l www.steel-research.de steel research int. 2021, 92, 2100323 as a base for thermodynamic calculation of steels to determine the precipitation sequence.
By etching the matrix, the structure of the precipitates in twophase high manganese steels has been shown in three dimensions by SEM. In Figure 11, the nitrides exhibit a shell-like, fibrous growth; small punctiform precipitates (<1 μm) can be seen around the complex precipitate. These dot-like precipitates are AlN in the initial stage. Due to the diffusion-driven depletion of [Al] and [N] of the residual melt forming this complex and due to the interdendritic separated areas during subsequent solidification, the precipitates do not have the possibility to coarsen and agglomerate with the complex precipitate. [52]

Phase Transformations
Investigation of tensile strength peritectic steels show, that γ-Fe has a higher tensile strength than δ-Fe. [16] In Figure 12, the influence of phase transformation in peritectic steels on the hot ductility behavior and the maximum tensile force are compared for different carbon contents. All compositions have in common, that the maximum tensile strength is strongly influenced by the present phase. If γ-Fe is absent, F max is low, but as soon as γ-Fe forms, a significant increase in the maximum tensile strength can be noted. The second observation concerns the effect of the phase-transformation on the reduction of area. In temperature ranges, where more than one phase is present, the ductility never reaches its maximum, and the ZDT is always lower than the T sol . [13] The softness of δ-Fe is not hindering the steel from reaching its maximum ductility, but the δ-γ-transformation leads to a local strain concentration in the δ-phase.
In LDZ2, the γ-α-phase transformation start with the precipitation of fine ferrite films at prior austenite grain boundaries. These thin layers of ferrite strongly impair the ductility; material recovers when ferrite films are thicker. The lower ductility results from the ferrite being the softer phase, wherefore the strain locally concentrates there. Thin bands of ferrite may occur at higher temperatures than Ar 3 and reach temperatures up to Ae 3 due to the transformation being strain induced. [53]

Porosity
Porosity also should be considered evaluating hot ductility test results. Before testing, the specimen rod is melted completely in a zone. In this process, the diameter is increasing up to the diameter of the crucible. This increased volume at constant mass could lead to higher amounts of porosity ( Figure 13) which is reducing the force bearing area and therefore the maximum reachable ductility. Porosity in general is known to reduce the mechanical properties, e.g., the Young's modulus and shortens the fatigue life. In continuous casting, porosity in the strand shell increases the breakout probability, too. [54][55][56][57] Different types of porosity can be defined. Figure 14 shows a classification of porosity origins. A main differentiation is made between gas-and solidification-induced porosity.
Gas porosity may form by the recombination of dissolved gases or can be induced by stopper gas at the SEN in CC, such as argon or nitrogen. Figure 15 shows a shrinkage pore with a MnS precipitation. This picture illustrates that even if pores are closed in further processing, e.g., by hot rolling, the precipitates Figure 11. SEM-EDX analysis of X30MnAl17 3, Â1000, complex (AlN) and Mn(S,Se) precipitation, and (AlN) in the initial stage of the formation process; slow solidification process with a solidification speed of v ¼ 15 mm h À1 . Figure 12. Influence of phase transformation at high temperatures in peritectic steels on ZST, ZDT, RA max , and maximum tensile force F max (kN), that is needed to keep a constant strain rate. Reproduced with permission. [13] Copyright 1980, Wiley-VCH.
www.advancedsciencenews.com l www.steel-research.de steel research int. 2021, 92, 2100323 they contain remain and represent a flaw in the matrix. Precipitates even might be sucked into the pore caused by a local vacuum during pore formation. This can lead to a higher concentration of precipitates by movement of microsegregated melt seen in Figure 15b. The figure also presents different kinds of shrinkage porosity. Figure 15a shows shrinkage porosity in the solid phase (sharp edges) and in Figure 15b, shrinkage porosity is highlighted where some liquid has been present at final solidification (round edges). Figure 16 shows two more examples of shrinkage porosity. Figure 16a shows an interdendritic pore (round shapes) and Figure 16b in the upper left side shows shrinkage at the grain boundaries.
To determine and understand which porosity leads to which weaknesses further research needs to be carried out.

Conclusions and Measures Against Decreased Ductility
To avoid decreased ductility ranges in critical CC plant positions, internal and external parameters need to be adjusted. The ductility of steel grades is influenced by various factors, such as steel composition, grain boundary development, thermal history, and strain rate. [58] The described internal phenomena have to be systematically coordinated with the process to achieve the optimum properties of the casting semi.
Considering the chemical composition first, some alloying elements have a more detrimental impact than others. For example, ductility can be significantly improved by adjusting the Al, Nb, or N content, as well as by setting the optimum Mn/S ratio. [25]    These elements can lead to an increased rate of fine precipitates, which are often the origin of cracks. An example of a damaging nitride is shown in Figure 8. However, sulfides or carbides can also act as crack initiators.
The effect of small changes of phase fractions during cooling can be estimated using thermochemical calculations, as shown in Figure 9. As well as the precipitates, phase transformations impair the RA curve at characteristic areas.
Knowledge of the precipitation formation in the as-cast structure and the areas of phase transformation of the specific steel compositions are an important input to individually set the optimum external parameters in casting process. This consideration is crucial to control the materials' properties with a tailored cooling strategy. Therefore, considering the temperature history is an essential step, especially of the surface region. Bending and unbending points are particularly critical areas of the continuous caster. [59] Combined, these factors ensure that the material can resist the stresses even in regions with high strains. Less harmful precipitates in terms of formation temperature, size, number, and shape can be achieved by adjusting the process parameters. [47] In the end, the overall process should be safely controlled by analyzing the solidification properties of the casted material at different temperatures.
It has to be noted, that the adjustment of parameters can, at the same time, have beneficial effects on the LDZ1 by increasing the ZDT and lead to a worse ductility in the LDZ2. [27]

Summary and Outlook
Various methods to analyze high temperature mechanical behavior were summarized. By these investigations the effects of steel composition and as-cast structure on ductile or brittle temperature ranges are determined. Segregation, precipitation of nonmetallic particles, phase transformation, and porosity are origins of brittleness during solidification and cooling. Inhomogeneities, especially fine-dispersed precipitates and proximate phases with different lattices and plasticities, are detrimental to the high temperature ductility.
It was shown, that complex interrelationships between internal parameters weaken the material. The extend of this impairment can be amplified or mitigated by external parameters. For this reason, the possibilities of setting individual CC process parameters for crack-sensitive steels were discussed.
The precipitation kinetics can be understood by use of the institutes' slow solidification furnace, since an extremely large particle size is obtained. For high Mn steel grades, nucleation and growth of AlN and MnS precipitates became apparent.
Adaptation of the process to the material and reduction of defects are continuous challenges and need a deep understanding of multiple influences. In this context prediction of the high temperature mechanical behavior based on thermodynamic and mathematical modeling of the ductility, together with experiments and computation of solidification can support adjusting the CC process.