E ﬀ ect of Intercritical Annealing and Austempering on the Microstructure and Mechanical Properties of a High Silicon Manganese Steel

: High Silicon Austempered steels (AHSS) are materials of great interest due to their excellent combination of high strength, ductility, toughness, and limited costs. These steel grades are characterized by a microstructure consisting of ferrite and bainite, accompanied by a high quantity retained austenite (RA). The aim of this study is to analyze the e ﬀ ect of an innovative heat treatment, consisting of intercritical annealing at 780 ◦ C and austempering at 400 ◦ C for 30 minutes, on the microstructure and mechanical properties of a novel high silicon steel (0.43C-3.26Si-2.72Mn wt.%). The microstructure was characterized by optical and electron microscopy and XRD analysis. Hardness and tensile tests were performed. A multiphase ferritic-martensitic microstructure was obtained. A hardness of 426 HV and a tensile strength of 1650 MPa were measured, with an elongation of 4.5%. The results were compared with those ones obtained with annealing and Q&T treatments.


Introduction
Nowadays, one of the most important objectives of steel producers and researchers in metallurgy is to bring to the market materials with improved properties and performance, high strength-to-weight ratios, and low costs. To achieve these objectives, high-alloyed steels, such aluminum and titanium alloys are not preferred due to the high cost of their raw material; on the contrary, the use of high strength steels (AHSS) is strongly recommended. Great attention is devoted to the study of multiphase steels due to their interesting mechanical characteristics, and both TRIP (TRansformation Induced Plasticity) steels, which belong to the second generation of Advanced High Strength Steels and High Silicon Austempered Steels are considered as more promising.
High silicon austempered steels are attractive grades for their particular combination of mechanical properties and ausfererritic microstructure, which is a mixture of ferrite and high carbon enriched austenite [1,2]. This particular microstructure leads to better mechanical performance, in terms of strength, hardness, and impact toughness in comparison with austempered ductile irons [1].
Moreover martensite, formed after quenching, speeds up the kinetic of bainitic transformation and the process of austenite stabilization [22,23].
Oliveira et al [21], analyzing the behavior of a MnSi steel after austempering and Q&P, observed that in the first case a microstructure consisting in a mixture of plate and granular bainite, martensite, and retained austenite is present. After Q&P a martensite matrix with bainitic islands and austenite films/blocks is instead formed.

Novelty of the Work
In this paper, a new dual phase high silicon steel with a novel composition was investigated. The effect of intercritical annealing and austempering on microstructure and mechanical behavior was studied The samples were characterized by optical microscopy (OM, LEICA DMRE, Leica Microsystems S.r.l., Milan, Italy), scanning electron microscopy (SEM, LEICA™ Cambridge Stereoscan LEO 440, Leica Microsystems S.r.l., Milan, Italy, EBSD, AMETEK BV, Tilburg, The Netherlands), transmission electron microscopy (TEM, JEOL JEM 200CX, Jeol Ltd, Tokyo, Japan), X-ray diffraction (XRD, Bruker D8 Advance, Karlsruhe, German) techniques, and mechanical tests. The results were compared with those obtained after annealing and Q&T treatments.

Materials and Methods
The material used in this study was a high silicon manganese steel, produced by Magnitogorsk Nosov State Technical University of Magnitogorsk (Russia). The chemical composition is presented in Table 1. The steel was prepared melting high purity raw material in an induction furnace. Liquid metal was cast, and the produced ingots were forged, and water cooled to room temperature

Heat Treatments
Heat treatment setup was composed by a Carbolite tubular electrical furnace and Nabetherm 3000 muffle electrical furnace. Tubular furnace was used for austenitization during quenching and for tempering the samples. It was also used for the intercritical annealing step. The other furnace was employed for austempering treatment. An air-cooling system was realized for the cooling step from the biphasic region to the Isothermal Bainitic transformation temperature. The final cooling to room temperature, from the IBT temperature, was performed in water. The samples' temperatures were recorded during the heat treatments with a K-thermocouple.
Heat treatments were designed using JmatPro software. A theoretical phase diagram was built to study the biphasic austenite-ferrite region, to evaluate the volume fraction of the two microstructural constituents and to determine austenite composition as a function of temperature ( Figure 1). The obtained information permitted us to choose the temperatures to perform the IA step, highlighted by the red line in Figure 1b. Theoretical CCT and TTT curves ( Figure 2) were built to define required cooling rates in the transition between annealing and the second austempering phase. This operation was necessary to avoid ferrite and perlite precipitation during treating cycles. Finally, holding times in the bainitic region were set thanks to the relative start and end transformation curves.  We decided to perform the intercritical annealing with the parameters reported in Table 2.   The heat treatments were performed starting from a martensitic microstructure, because martensite allows for the faster recrystallization processes of ferrite during intercritical annealing and higher performance after the treatment, according to other works [4,24,25]. Furthermore, Kim et al. [4,25] demonstrated that fine microstructure, resulting from a rapid cooling, favors austenite formation and its carbon enrichment when material is re-heated.

Microstructural Study
The microstructure of the samples was analyzed along the cross section of the specimens. To perform the analysis an optical microscope, LEICA DMRE (Leica Microsystems S.r.l., Milan, Italy), and a scanning electron microscopy (SEM) LEICA™ Cambridge Stereoscan LEO 440 (Leica Microsystems S.r.l., Milan, Italy) were used. The preparation of the samples was carried out following the standard technique. The specimens were cut with SiC discs lubricated with a mixture of water and oil, mounted in phenolic resin, grinded with SiC papers (320, 500, 800, 1200 grit), polished with clothes and 6 µm and 1 µm polycrystalline diamond suspensions. EBSD analysis occurred after the samples were further polished with 200 nm and 40 nm silica colloidal particles suspension. In order to reveal the microstructure of the samples they were etched with Nital 2.
Phase identification and phase quantification were carried out through X-ray diffraction with a Bruker D8 Advance (XRD, Bruker D8 Advance, Karlsruhe, German), operating at 40 kV and 40 mA and a Cu radiation tube (Kβ radiation was filtered by mean of nickel filter on the tube side). The investigated angular range was between 40 • and 105 • , steps scan of 0.025 • and counting time of 3 s. The obtained patterns were analyzed using High Score Plus software in order to identify the constituent phases. Volume fraction calculation of the phases was performed through Rietveld analysis on the same software.
Electron Backscattered diffraction (EBSD) analysis was performed after the heat treatments in order to study in detail the obtained microstructure. In particular, the identification and quantification of the phases were carried out. Moreover, phase distribution, orientation, and the presence of textures were investigated. For EBSD investigations, we used a FEI QUANTA 205 FEG SEM (Thermo Fisher Scientific, Hillsboro, OR, USA), equipped with AMETEK EBSD (AMETEK BV, Tilburg, The Netherlands) system and OIM Analysis™, operating at 20 kV. The analyses were performed with the following parameters: Scan step: 200 nm, area:~150 × 200 µm 2 ; confidence Index >5% and Scan step: 100 nm; area:~35 × 40 µm 2 , confidence Index >5%.
Transmission electron microscopy (TEM) was also performed to complete the characterization of RA on the samples using a JEOL JEM 200CX (Jeol Ltd, Tokyo, Japan) operating at 160 kV. The preparation of the thin foils was realized by mechanical grinding until thickness of 70 µm, followed by mechanical punching to obtain 3 mm diameter specimens. The final polishing and etching were performed electrochemically using a twin-jet polisher STRUERS TENUPOL-3 (Struers S.A.S., Milan, Italy), with 95% acetic acid (CH 3 COOH) and 5% perchloric acid (HClO 4 ) solution, at 45 V and room temperature [26,27].

Mechanical Tests
Vickers micro-hardness, with a Leitz™ DURIMET (Leica Microsystem S.r.l., Milan, Italy) hardness tester, were conducted on each sample performing three indentations, with a 300 g load.
The treated samples were also subjected to tensile tests according to ASTM A370 ε 19´1 . The tests were carried out on dog-bone samples, at strain rate of 5 × 10 -3 s -1 with a MTS tensile test machine (MTS System Corporation, Eden Prairie, MN, USA), using a maximum force of 50 kN. The displacement was measured through the crosshead movement and the force by the machine load cell.

Microstructure
The microstructure of the as-received material is shown in Figure 4a-c. The as-received structure shows a predominant martensitic matrix with some allotriomorphic ferritic colonies (F A ) formed at the prior austenitic grain boundaries during quenching. From the SEM micrograph (Figure 4b,c) it can be further observed the presence of idiomorphic ferrite (F I ), nucleated inside the original austenite grains (GB), present before the quenching. Figure 5a-c reports the evolution of the microstructure caused by the annealing treatment at 870 • C for 10 min. From OM and the SEM micrograph (Figure 5a,b) it is possible to observe the presence of pearlitic islands (P). In the red box, shown in Figure 5c at higher magnification, eutectoidic islands with different lamellae orientation, surrounded by ferritic (F) grains, can be distinguished. This is the typical microstructure of a medium carbon steel after annealing treatment [28].    [29,30]. In Figure 7c idiomorphic ferrite (F I ) at the center of the grain is shown.  After Quenching, Intercritical Annealing, and Austempering treatments, the samples show a different microstructure compared to that of the previous ones. It is possible to observe (Figure 8a,b) a dual phase microstructure, characterized by the presence of ferrite surrounded by a martensitic matrix and retained austenite. Ferrite is formed during the intercritical annealing at 780 • C. The ferrite presents in microstructure, visible in Figure 8, is a combination of different types and morphologies. A certain volume fraction of ferrite is formed during the intercritical annealing in the biphasic region. It is also possible to observe allotriomorphic and idiomorphic ferrite, as result of the partial transformation of austenite during the first cooling phase. The formation of these kinds of ferrite contributes strongly to the carbon enrichment of austenite. The low carbon solubility of ferrite forces carbon partitioning in austenite, increasing its stability and the possibility to retain it at room temperature. Martensite derives from austenite formed during heating in biphasic regime; austenite partially transforms into ferrite during the first heating step and into martensite during the last cooling phase. Bainite is not present during the austempering phase of the treatment and it is not visible in the microstructure, even if expected from the cooling transformation curve. This phenomenon could be related to the evolution of the curves and transformation temperatures caused by carbon partitioning during the heat treatment steps. It can be supposed that carbon partitioning leads to a decrease of Bs. FEG-SEM micrograph (Figure 9), taken in backscattered electron mode, allows us to see in detail the microstructure of the specimen subjected to austempering treatment. In the red box Figure 9b, it is clearly visible lath martensite in light grey and ferrite in black. This type of martensite morphology was confirmed in a research work by Ahmad and colleagues, which demonstrated that with intercritical annealing a lath morphology is acquired [31].

X-ray Diffraction
One of the main goals of this thermal cycle is to achieve a multiphase material with a sufficiently high carbon content in the retained austenite which promote its retention at room temperature and leads a significant volume fraction of RA. Figure 10 reports the X-ray diffraction patterns of the specimens subjected to annealing, Q&P, and Austempering heat treatment. X-ray pattern of annealed sample (Figure 10a) shows BCC iron peaks in the investigated angle range. Moreover, as 2θ increases it is possible to resolve Kα1 and Kα2 line peaks (indicated with the arrows). This could be related with the presence of coarse grains, which allows a better peak definition and high pattern resolution.
A slight difference in the peak position in the X-ray patterns can be observed, which could be attributed to the different dimension of the cell, caused by the different heating cycles.
Asymmetrical broadening of ferrite diffraction peaks towards low 2θ angles can be observed in the austempered samples (Figure 10b). Those peaks are related to the presence of tetragonal martensite. The presence of a displacement between ferrite and martensite peaks is due to the high carbon content of martensite and the strong distortion of the cell [32].
The presence of martensite in samples can be related to room temperature cooling that causes austenite transformation. High carbon austenite, which is not transformed in bainite during austempering, transforms in martensite when cooled down.
The X-ray diffraction technique allowed us to confirm the presence of retained austenite in the austempered samples, with a volume fraction similar to the one found in several TRIP steels and austempered high silicon steels [4]. The results of the phase quantification, reported in Table 3, demonstrate that the treatment cycle favors the process of carbon partitioning, allowing austenite retention at room temperature.

Electron BackScattered Diffraction
In order to better analyze the microstructural features of the austempered samples, EBSD analysis was performed. The main goal of the investigation was to understand distribution retained austenite, and its morphology and size.
No preferential orientation was observed in the material subjected to the austempering treatment ( Figure 11). The absence of texture and preferred grain orientation can be explained considering that: a) the material was not subjected to deformation or forming process during heat treatment, b) a pretreatment, which produces recrystallization, was performed.  Figure 12 shows the detailed EBSD analysis on smaller sample areas. It can be noted also in this case that each phase has randomly oriented grains (Figure 12a). In Figure 12b, a phase distribution map is shown: red zones represent FCC iron, green ones refer to tetragonal martensite and retained austenite (FCC iron) is colored in yellow. A matrix of ferrite can be distinguished, with uniformly distributed austenitic and martensitic islands. Austenite is located at ferrites grain boundaries close to the martensitic islands, with a grain size of 1-5 µm. Such dimension of the FCC iron refer to high stable austenite with high carbon content. These islands should exhibit a strain induced martensitic transformation during plastic deformation at high strains [4,9,10,[33][34][35][36][37][38][39].

Transmission Electron Microscopy
A further microstructural investigation was carried out by TEM on the austempered samples (Figure 13a,b) in order to deeply analyze the morphology of the martensitic-austenitic regions formed during the heat treatment and shown with the other techniques.  Figure 13a shows the presence of martensitic zones (light grey islands) and films of retained austenite (black islands). A detail of a retained austenite film between martensitic region is shown in Figure 13b. SAED was performed in the correspondence of the white circle zone, and the resulting diffraction pattern, indexed according to [40], confirm the presence of austenite. Therefore, it is possible to state that the treatment allows the austenite retention, due to the carbon partitioning.
EBSD and TEM analysis show that retained austenite has two morphologies: films between martensitic islands and blocks. The obtained retained austenite is characterized by different carbon content and different stability, in particular RA located near the martensitic islands is richer in carbon and more stable, according to [41]. Blocky austenite, which is poor in carbon, transforms earlier than lath austenite during the deforming process, as evidenced by [9].

Microhardness Test
Microhardness results are reported in Table 4. It is possible to observe that the material, when annealed, has lower hardness compared to the samples subjected to the other treatments. Moreover, material subjected to austempering treatment has higher hardness than the sample treated by Q&T. The lower hardness of the Q&T sample could be explained considering the effect of the tempering treatment on martensite. The subsequent heating of martensite, below austenitization temperature, favors stress release, carbon diffusion, and carbide precipitation.

Tensile Tests
Mechanical properties were evaluated in terms of yield strength (YS), ultimate tensile strength (UTS), and fracture strain. The specimens, after annealing treatment and Q&P, showed the highest uniform elongation (Figure 14 and Table 5), while after austempering treatment the sample exhibited poor elongation, without evident yielding and plastic deformation. In detail, in the austempered sample, a fracture strain of approximately 4.5% and UTS of 1650 MPa were recorded. The brittle nature of the material could be explained considering the amount of martensite and its high dislocation density which limits further plastic deformation and the low strength of the martensite-ferrite cohesion.
A similar behavior was observed by Ahmad and colleagues [31]. The presence of a slight plastic deformation can be attributed to the presence of a considerable amount of ferrite. Figure 14. Engineering stress-strain curves. Blue curve with blue circle is related to austempered samples, red curve with diamonds referred to the Q&P treatment and green curve with squares representing the annealed samples. The fracture surface of specimens after annealing ( Figure 15) shows the typical features of a ductile fracture, confirming the mechanical behavior evidenced in the stress-strain curves. It also is possible to observe in Figure 15b the fracture surfaces of pearlitic grains [42].   Brittle failure could be observed in the case of austempered specimen ( Figure 17), with the development of cleavage facets [31], highlighted in the figure by circles.

Austempering Treatment
The austempering treatment was performed in correspondence with the bainitic region nose in order to minimize the soaking time due to the fast kinetics ( Figure 1). In fact, the time required to complete the transformation is the shortest and the transformation is also anticipated, reducing the overall time of the heat treatment [1].

Microstructure and Mechanical Properties
The annealing treatment, as shown in Figure 5, produced a microstructure consisting of pearlite and ferrite, as expected. The material is characterized by an equiaxed grain structure with eutectoidic islands randomly oriented, while quenching and tempering allowed us to obtain tempered martensite (Figure 7). The microstructure obtained after these two treatments agree with the ones reported in literature.
The austempering treatment produced a completely different microstructure; in fact, the observation carried out on the samples with optical and scanning electron microscopes (Figure 8) revealed the presence of a multiphase microstructure. This microstructure consists of ferrite, martensite, and retained austenite.
During soaking at IBT temperatures a significant fraction of austenite stabilized due to the carbon in FCC iron. After the heat treatment, retained austenite showed two morphologies: i) films between martensite laths, and ii) blocks according to [9,33,[35][36][37][38]43,44]. The authors also observed that the austenite carbon content was not homogeneous, in particular, it was higher in the first type of retained austenite compared to the blocky austenite.
No upper or lower bainite was present in the microstructure, as expected. As a matter of fact, carbon partitioning, which takes place during the treating cycles, especially during annealing and the first cooling phase, may modify bainite start temperature (Bs). In detail, the increase in the carbon content in austenite can reduce Bs, as predicted in some models, as those reported by Lee [45] Concerning the mechanical properties, the hardness values agreed with the microstructure present in the samples. The austempering treatment produced and increased hardness, due to the high volume fraction of un-tempered martensite, and the results are in accordance with the behavior reported in literature [46]. The Q&T sample was characterized by lower hardness, due to martensite tempering which led to martensite transformation and internal stress reduction.
Hardness and microstructural constituents can also explain tensile test results. The samples subjected to annealing treatment exhibited coarser microstructure and lower hardness and tensile strength compared to the other samples but were characterized by the highest ductility.
Considering the austempering treatment, the samples with the highest hardness values and the highest tensile strength, with a predominantly brittle behavior, were obtained. The brittle phenomenon was also confirmed by the fractographs, in which were not visible the features typical of a ductile fracture. This behavior is certainly determined by the microstructure, which mainly consisted of martensite, while the just mentioned plastic deformation was attributed to the ferrite volume fraction.

•
A novel high silicon steel with multiphase microstructure composed by ferrite, tetragonal martensite, and retained austenite was developed. Austempering treatment allowed us to stabilize a consistent volume fraction of retained austenite equal to 14.9%.

•
The complete XRD, EBSD, and TEM analysis revealed that RA is present in two morphologies: as film between martensite islands and in form of blocks.

•
The sample subjected to austempering exhibits the higher values of Vickers microhardness while Q&T and annealed samples are characterized by lower values.

•
Austempering treatment enhances material tensile strength (1650 MPa), in comparison with annealing and Q&T, which are characterized by lower tensile tests, 1130 and 1200 MPa respectively. However, the high amount of martensite, is responsible of low fracture strain (4.5%) and ductility, which is confirmed by the brittle fracture surface. Funding: The authors want to thank FONDAZIONE CARIPARO for the financial support with the visiting program TIP-STEP.

Acknowledgments:
The authors want to thank Marco Breda from Unilab Laboratori Industriali S.r.l., for the tensile tests.

Conflicts of Interest:
The authors declare no conflict of interest.