Regular Article
Formation Mechanisms and Three-dimensional Characterization of Composite Inclusions in Low Aluminum Steel Deoxidized by Silicon
2023 Volume 63 Issue 2 Pages 338-345
Details
2023 Volume 63 Issue 2 Pages 338-345
Silicon deoxidized steel takes a large proportion in addition to aluminum deoxidized steel, since aluminum deoxidation may cause many problems, such as nozzle clogging and low toughness. In this paper, a low aluminum silicon deoxidized steel was prepared in the laboratory. The inclusions in steel mainly include MnS and composite inclusions made of MnS and oxides. The formation of MnS and composite inclusion was discussed based on the thermodynamic calculations. It was clarified that the MnS precipitate during the solidification process of steel. In addition, the formation mechanism of the composite inclusions was illustrated from the perspective of thermodynamics and lattice mismatch. The composite inclusions in steel were reconstructed by Micro-CT and the three-dimensional (3D) morphologies were analyzed in details. This method avoids artifacts and misleading of the two-dimensional observation on the cross sections of the steel samples. Meanwhile, the 3D observation by Micro-CT is an innovation way to characterize the different phases of the inclusion simultaneously, which paved a way for the analysis of the composite inclusions in the steel.
Deoxidation is an important process in steel making process due to its crucial influence on the type and distribution of inclusions in the steel. Aluminum is an economical and effective deoxidation agent in the steelmaking process. However, the Al2O3 inclusion clusters are frequently observed in Al-deoxidized steel, which caused many problems, such as the clogging of submerged entry nozzle during continuous casting,1,2,3,4,5) drawing fracture and fatigue failure during service course.6) The ferrosilicon and ferromanganese alloys with low aluminum content are widely used for the deoxidizing of some types of steel to reduce the number of non-deformable Al2O3-rich inclusions.7,8)
The types, quantities and distributions of inclusions are important to the quality of the steel.9,10,11,12) Many researchers have studied the types of inclusions and the formation mechanism in the steel after the deoxidation by silicon alloys. Thapliyal et al.13) observed the inclusions containing Mn–Si–Al–S–O in Si–Mn killed steels after Si–Mn deoxidation. Li et al.14) found that a large number of MnO-riched inclusions with 1–2 μm were generated by the oxidation of high [Mn] after reoxidation of Si-killed in stainless steel, which were transitional products. The inclusions in the silicon deoxidized steel were mostly SiO2, Al2O3–SiO2, SiO2–MnO and MnO–SiO2–Al2O3.15,16,17,18,19) The composite inclusions such as CaO–SiO2, MgO–Al2O3–SiO2 and other composite inclusions with more complicated compositions were found in calcium-treated or magnesium-treated steel.20) Besides, MnS wrapped Al–Si–O–Mn composite inclusions were found in some types of steel.21,22,23) The brittle inclusions such as Al2O3 and SiO2 which are coated with MnS can reduce the stress concentration of steel and avoid forming initial cracks.24,25) Thus, it is essential to study the formation of the inclusions of Al–Si–O–Mn wrapped by MnS.
At present, the most common method to analyze inclusions in steel is detecting the inclusions by scanning electron microscopy (SEM) or an automatic scanning electron microscope (ASPEX), which could obtain the morphology and composition of inclusions easily and quickly. The sizes of the inclusion are not be precious determined since the two-dimensional cross-sections could not always reveal the largest or central cross-section of the inclusions.26) Many researchers extracted the inclusions by electrolysis or chemistry dissolution and observed their morphology by SEM to reflect the three-dimensional morphology of inclusions.27,28) However, the electrolytic extraction (EE) or chemical extraction (CE) has many shortcomings. Different types of inclusions require various electrolytes or solutions. Besides, it is difficult to achieve nondestructive inclusions.
The 3D atomic probe is a powerful technique to obtain the 3D morphology in the precipitants, which can accurately analyze the morphology precipitates in steel.29,30,31) However, the area observed by this method is only a few hundred nanometers, which is appliable for the micron-level inclusions. Therefore, a nondestructive observation method with a larger detected dimension is required for the observation of the inclusions in the steel.
Recently, the X-ray Micro-CT has been developed rapidly gradually applied in industrial manufacturing and scientific research. The X-ray Micro-CT was employed to analyze and detect the morphology of micron-scale particles and defects in solid materials. Liu et al.32) studied the 3D spatial distribution of steel fibers in fiber-reinforced cementitious composites through the micro-CT technique. Skarżyński33) analyzed the size and distribution of pores and cracks with an X-ray micro-computed tomography system in reinforced concrete beams in the 3-point bending experiment. Shang et al.34) characterized the three-dimensional (3D) inclusions of TiN and MnS in steel by X-ray micro-CT. Li et al.35) developed a method of observing 3D inclusion clusters in metal and then measured the Al–TiB2 and Al–SiC system by the X-ray micro-CT in beamline BL20XU at Spring-8. In our previous work, MnS and other single-phase inclusions were 3D characterized by this beamline, and the 3D morphology and distribution of inclusions were obtained.36) However, all of these inclusions are single-phase in steel. The composite inclusions are seldom characterized by micro-CT.
Although micro-CT is able to detect the internal morphology of inclusions, it cannot analyze the chemical composition of inclusions. Thus, Micro-CT technology could be used to analyze the composite inclusions assisted by SEM-EDS which can determine the composition and 2D morphology. In this paper, the samples containing composite inclusions of (Al2O3–MnO–SiO2)–MnS were prepared by experiments. The mechanism of the formation of the inclusions was explained by the thermodynamic calculation. After the type of inclusions in the experimental steel was determined by SEM-EDS, the micro-CT available in the beamline of BL13W1 at the Shanghai Synchrotron Radiation Facility (SSRF) was utilized to obtain a 3D morphology for the typical inclusion of (Al2O3–MnO–SiO2)–MnS in steel with stereological analysis. It would be the first 3D characterization of the (Al2O3–MnO–SiO2)–MnS.
The chemical composition of the experimental steel is listed in Table 1. The high pure iron was smelted by an electromagnetic induction furnace in the shielding gas of argon. The powder of ferric oxide was then added to the molten iron. The ferrochrome, ferromanganese and pyrite were used to adjust the composition. Pure silicon and a small pieces of aluminum pellets were added for deoxidation. The melt was kept at 1530°C for 10 minutes and cooled to room temperature. A sample with a size of 10×10×5 mm was cut by wire electrical discharge machining (WEDM). The sample was polished and observed by scanning electron microscope with an energy dispersive spectrometer (SEM-EDS) to obtain the morphology and composition of the inclusions in the steel.
| C | S | Mn | Cr | Al | Si | Fe |
|---|---|---|---|---|---|---|
| 0.046 | 0.011 | 0.75 | 0.21 | 0.050 | 0.30 | Remainder |
Some cylindrical samples with a diameter of 0.3 mm were cut from the experimental steel for Micro-CT scanning detection. The cylindrical samples containing composite inclusions were observed by using X-ray Micro-CT in beamline BL13W1 at the SSRF, which is an advanced third-generation light source.
The difference of the X-ray linear attenuation coefficients (LAC) between Fe and inclusions should be greater than 10% to distinguish the composite inclusions from the steel matrix. The photon energy of the beamline in SSRF ranges from 8 to 72.5 keV. Figure 1 shows the comparison of X-ray linear attenuation among the Fe, AMS, and MnS inclusion. The photon energy was set to 38 keV after testing the quality of X-ray images.
The comparison of X-ray linear attenuation coefficient of Fe and different inclusions. (Online version in color.)
Figure 2 shows the typical inclusions detected in the sample. The sizes of the inclusions are about 5 μm. Some composite inclusions even contain three phases, which can be seen intensity of the images of the inclusions. The normalized composition according to the EDS analysis of all regions are shown in Table 2 for comparison. The inclusions in the deoxidized silicon steel were mainly MnS and composite inclusions. All the composite inclusions were oxides wrapped by the light-colored MnS inclusions as shown in Figs. 2(b) to 2(f). The individual Al2O3 and SiO2 inclusions were not found in the steel. The composition of the composite inclusion can be expressed by (Al2O3)x-(MnO)y-(SiO2)z-MnS.
The morphology of typical inclusion in experimental steel.
| Inclusion | Region | Mn | S | Si | Al | AMS-MnS |
|---|---|---|---|---|---|---|
| (a) | A | 65.66 | 34.44 | – | – | Mn0.52S0.47 |
| (b) | A | 54.24 | 24.44 | – | – | Mn0.56S0.44 |
| B | 39.87 | 5.91 | 7.33 | 1.15 | (Al2O3)0.03-(MnO)0.65-(SiO2)0.32-MnS | |
| C | 37.37 | 7.75 | 6.98 | 1.14 | (Al2O3)0.04-(MnO)0.61-(SiO2)0.35-MnS | |
| D | 46.58 | 3.65 | 9.04 | 1.19 | (Al2O3)0.02-(MnO)0.68-(SiO2)0.30-MnS | |
| E | 41.94 | 10.70 | 11.28 | 0.90 | (Al2O3)0.02-(MnO)0.50-(SiO2)0.47-MnS | |
| (c) | A | 54.16 | 27.43 | – | – | Mn0.53S0.47 |
| B | 38.71 | 15.22 | 6.13 | 2.97 | (Al2O3)0.12-(MnO)0.45-(SiO2)0.43-MnS | |
| C | 25.91 | 9.96 | 20.04 | 0.81 | (Al2O3)0.02-(MnO)0.18-(SiO2)0.80-MnS | |
| (d) | A | 37.82 | 10.68 | 6 | 3.27 | (Al2O3)0.11-(MnO)0.55-(SiO2)0.34-MnS |
| B | 30.94 | 10.71 | 9.43 | 2.02 | (Al2O3)0.07-(MnO)0.38-(SiO2)0.55-MnS | |
| C | 32.81 | 16.80 | – | – | Mn0.53S0.47 | |
| (e) | A | 40.81 | 1.55 | 12.74 | 4.62 | (Al2O3)0.08-(MnO)0.56-(SiO2)0.36-MnS |
| B | 30.66 | 3.93 | 21.80 | 2.37 | (Al2O3)0.04-(MnO)0.34-(SiO2)0.62-MnS | |
| C | 37.62 | 4.26 | 11.06 | 3.75 | (Al2O3)0.08-(MnO)0.54-(SiO2)0.38-MnS | |
| D | 55.29 | 25.10 | – | – | Mn0.56S0.44 | |
| (f) | A | 31.37 | 14.21 | 1.31 | – | (MnO)0.73-(SiO2)0.27-MnS |
| B | 41.06 | 19.68 | – | – | Mn0.55S0.45 |
The value of x in (Al2O3)x-(MnO)y-(SiO2)z-MnS could not be negligible, despite it was small. Since the pure MnS detection results of Mn and S ratio are greater than 1 in Table 2. It is probably because those Mn elements dissolve in steel and inclusion. The proportion of Al2O3 was higher than the calculated value. The mapping scanning of the composite inclusions was carried out in order to better reveal this issue. The mapping distributions of elements of one composite inclusion is shown in Fig. 3. It can be seen that the Al, Si, and O elements were distributed in the same region.
Mapping distributions of elements of one composite inclusion. (Online version in color.)
The precipitation of inclusion is related to the solubility product of its constituent elements. The solidus and liquidus temperatures of the experimental steel were calculated by Eqs. (1)37) and (2)38) to verify whether MnS precipitated in liquid steel, during solidification, or after solidification.
| (1) |
| (2) |
The liquidus and solidus temperature of low-sulfur steels was respectively 1526°C (1799 K) and 1492°C (1765 K), which is consistent with the values calculated by Thermo-calc software.
The solubility product for MnS was computed based on the equilibrium analysis in Eq. (3). The solubility constant K for MnS was described as Eq. (4).27)
| (3) |
| (4) |
| (5) |
The generation of MnS was computed according to the composition in Table 1 and the element interaction coefficient39) in liquid iron, as shown in Fig. 4. The composition of the test steel is far from the equilibrium curve of the solidus temperature, which indicates that the MnS precipitated after the steel completely solidified.
Solid solution curve of MnS in steel. (Online version in color.)
The polythermal projection of the Al2O3–MnO–SiO2 ternary system was calculated to illustrate the formation of the core of the composite inclusions, as shown in Fig. 5. The calculated temperature is 900–1700°C. While the minimum temperature of the isothermal projection line is 1200°C, which indicates that the minimum liquidus of the inclusion is 1200°C. The chemical components of the oxide cores of the composite inclusions in Table 2 are marked in Fig. 5. All the oxide cores are precipitated above 1200°C. The chemical composition of the experimental steel is located under the curves of 1200°C in Fig. 4, which indicates the MnS inclusions in the experimental steel are precipitated below 1200°C. Therefore, the oxides could act as the nucleus of the MnS inclusions.
The isotherm projection of the Al2O3–SiO2–MnO. (Online version in color.)
The heterogeneous nucleation is affected by the interface energy of materials. However, the interface free energy is difficult to be measured. It is composed of many factors, such as chemical composition, electronic potential, interface lattice parameters, etc. The lattice mismatch provides a method to predict the relative effectiveness of nucleating agents and substrates.40) The phase diagram of Al2O3–SiO2–MnO at 1100°C was calculated to determine the possible chemical formula of the substance at the interface, as shown in Fig. 6. The composition of the nucleus was possible located in partition 1 to 3 due to the low aluminum content in steel. The phase containing aluminum were Mn3Al2Si3O12. The Bramfitt mismatches were calculated between MnS and the possible substrates (Mn3Al2Si3O12, MnSiO3, SiO2 and Mn2SiO4). The results are shown in Table 3. It is generally considered that the substrates precipitate on the core surface when the mismatch is less than 12%. Hence, all the SiO2, Mn2SiO4 and Mn3Al2Si3O12 could act as effective nucleus for MnS inclusions.
Ternary phase diagram of Al2O3–MnO–SiO2 at 1100°C. (Online version in color.)
| Nucleus | (h k l)nucleus | Mismatch (%) | (h k l)MnS |
|---|---|---|---|
| Mn3Al2Si3O12 | (1 0 0) | 10.0 | (1 0 0) |
| MnSiO3 | (1 0 0) | 14.9 | (1 0 0) |
| SiO2 | (1 1 1) | 8.9 | (1 1 0) |
| Mn2SiO4 | (1 0 1) | 11.3 | (1 0 0) |
Schematic diagram of the Micro-CT in beamline BL13W1, SSRF. (Online version in color.)
The SEM images show two-dimensional sections of inclusion, which couldn’t reflect the real sizes and the volumetric ratios of the different phases. Hence, the 3D of the composite inclusions was carried out by using Micro-CT at beamline BL13W1 in the SSRF. Figure 10 shows the schematic diagram of the Micro-CT. The projection images were captured with an angular step of 0.15°. The samples were rotated from 0° to 180°. In total, 1200 projection images were obtained for each sample. Another two types of images were acquired during the experiment to eliminate the background features introduced by the monochromator. The flatfield images were acquired by temporarily removing the sample out of the viewing field during the experiment. While the dark filed images were obtained without X-ray penetration.
Areas of different phases in sequential slices with different directions. (Online version in color.)
The phase retrieval method was applied to increase the contrast of the different phases in the original image obtained by CT before the reconstruction of the slices. Then the 2D slices were generated, which were based on the phase recovery information. These processes were accomplished by using PITRE software, which was developed by SSRF.41) The X-ray propagation-based phase contrast imaging (PPCI) handles phase retrieval of X-ray propagation-based phase contrast computed tomography (PPCT) projections. Then the sinograms were created, for the sinogram pre-processing and slice reconstruction. The 2D images of the slice was generated following the process shown in Fig. 8 for giving an example. The composite inclusion was marked in Fig. 8(d).
The Images outputted in slice reconstruction processing: (a) The raw projections; (b) X-ray propagation-based phase contrast imaging (PPCI); (c) Sinograms of PPCI; (d) Slices of phase retrieval. (Online version in color.)
Figure 9 shows the 3D morphology and 2D slices in different directions by using the softwares of Avizo and ImageJ. The blue layer represents the outer MnS, while the red core represents the inner Al2O3–MnO–SiO2 phase in Fig. 9(a). Slices with regular intervals along different directions are demonstrated in sequence, as shown in Figs. 9(b), 9(c) and 9(d). The light gray parts represent MnS and the dark gray parts represent the core of oxide. The morphology of slices varies constantly in each direction. The two or three oxide cores in Fig. 9(d) are actually connected to each other and forms an abnormal shape rather than individuals, as seen in Figs. 9(b) and 9(c). It is clearly that the number of cores observed in the 2D cross sections is not necessarily the real number, which would mislead the research works on the composite inclusions.
The morphology of a typical composite inclusion of (Al2O3–SiO2–MnO)–MnS: (a) 3D morphology; (b) 2D morphology in length direction (X direction); (c) 2D morphology in width direction (Y direction); (d) 2D morphology in height direction (Z direction). (Online version in color.)
The areas of the core oxidation, MnS, and total composite inclusion in various slices in all the three directions were calculated by image processing software respectively, as shown in Fig. 10. The area of the core in the X direction began to increase from the 10th slices disappears at the last 10 slices. It indicates that the coating layer of MnS was only 0.15 um and the oxide cores in the X direction was almost exposed to the surface. The core in the Y and Z directions appears from the 50th slices and end up around the last 50 slices. The layers of MnS in the Y and Z directions are relatively thick, which are about 0.76 um. Thus, for the current observed inclusions, the stress in the X direction is more liked to induce crack due to the brittleness of the oxide core of the composite inclusions.
The proportions of the core in the 2D cross sections of the composite inclusions reflect the encapsulation rate of the MnS in the composite inclusions. It indicates the MnS completely encapsulates the oxide core in a certain direction when the proportion of the core is 0. The distribution of the area fraction of oxide core in the composite inclusions in the three directions was statistically analyzed in Fig. 11. It can be seen that 26% of the slices along the Y direction are occupied by pure MnS inclusions. The maximum area fraction of the oxides in the 2D cross sections is 50%–60%, which indicates MnS occupies at least 40% in the series of slices. However, MnS is not well wrapped in 2D slices as shown in the series of the cross sections in Fig. 9. In the X and Z directions, the proportion of oxides area is 30%–40% in maximum. The morphology of the composite inclusion is assumed into a spherical oxide core with the MnS uniformly wrapped in the outer layer, as shown in Fig. 12. The volume fraction of oxide core in composite inclusion can be expressed by the formula (6).
| (6) |
The distribution of the core area ratio in composite inclusion. (Online version in color.)
Schematic diagram of a hypothetical composite inclusion. (Online version in color.)
In this study, the silicon deoxidized steel with low aluminum content was prepared. The inclusions type was determined by SEM-EDS. The mechanism of the composite inclusion is illustrated by thermodynamic analysis. The morphology of the composite inclusion was characterized by CT and quantitatively analyzed that revealed the limitation of the 2D observation of the cross sections.
(1) Inclusions in silicon-deoxidized steel with low Al are mainly MnS and composite inclusions containing MnS. The cores of the composite inclusion are oxides, whose composition can be expressed as (Al2O3)x-(SiO2)y-MnO.
(2) MnS inclusions are mainly formed during the solidification process. The oxides inclusions of SiO2, Mn2SiO4 and Mn3Al2Si3O12 could act as effective nucleus for MnS inclusions.
(3) Micro-CT is applied to characterize the composite inclusions with different phases, which is helpful to avoid the misleading of the 2D morphology of inclusions.
(4) The volume fraction of the core in the composite inclusion was 0.32, and the maximum area fraction of the core in the cross sections of the composite inclusion is less than 60%.
The authors greatly acknowledge the support of the projects from the Science Fund for Distinguished Young Scholars of Hebei Province, China (E2021209039), Hebei Financial Support Project for the Introduced Overseas Student (C20210309), the National Natural Science Foundation of China (No. 52074056), the Tangshan Science and Technology Bureau for Fundamental Innovation Team of High Quality Clean Steel in Tangshan (21130209D) and the Natural Science Foundation of Chongqing, China (No. cstc2020jcyj-msxmX0449). The experimental work was supported by the Sharing Service Platform of CAS (Chinese Academy of Sciences) Large Research Infrastructures (No. 2018-SSRF-PT-005960, No. 2020-SSRF-ZD-000490 and No. 2019-SSRF-PT-008456). We also thank the staffs at SSRF for their help during our experiments there.
