The raw materials include 49MnVS3 steel casting billet as well as Ni-Mg alloy (75%:25%), quartz tubes, molybdenum rods, and high purity Ar gas (99.999%). The chemical composition of the sample is shown in Table 1.

A total of five groups of magnesium treatment modification experiments were conducted, which were divided into one contrast group and four experimental groups with different quality magnesium alloys. The experimental procedures were as follows: 600 g of original billet steel was placed into a magnesia crucible (ID: 53 mm, OD: 61 mm, H: 120 mm), and transferred into a pit-type resistance furnace. Then, the sample was heated to 1873 K under high-purity argon flow (15 L/min), and the molten steel was held for 20 min to achieve sufficient homogenization.

The final content of magnesium of the melts is listed in Table 2.

After the experiment, the metallographic samples of 10 mm × 10 mm × 12 mm were obtained at the 1/4 diameter of the center of the ingot. The samples were polished and observed under optical microscope (Zeiss) and scanning electron microscope (SEM, Model: Phenom Pro) to observe the morphology of the inclusions.

Define abbreviations and acronyms the first time they are used in the text, even after they have been defined in the abstract. Abbreviations such as IEEE, SI, MKS, CGS, sc, dc, and rms do not have to be defined. Do not use abbreviations in the title or heads unless they are unavoidable.

The morphology and distribution of inclusions in the sample were observed by metallographic microscope, and the typical distribution of the five groups of inclusions was shown in Figure 1. As shown in Figure 1(a), when only Ca is added to the non-quenched and tempered steel, inclusions are irregularly distributed, and inclusions are aggregated and distributed in the grain boundary in the form of long chains. In addition, there are many inclusions in large size and few in small size. After adding Mg and Ca for composite treatment, it can be seen from Figures 1(b)-(e) that the distribution of inclusions becomes more uniform and the morphology of inclusions transforms from chain shape to spherical or spindle shape, while the large particle inclusions are significantly reduced.

In order to explore the detailed changes of the inclusions in the steel during the composite treatment of different mass fractions of magnesium, the morphology and composition of the inclusions in the sample were analyzed by SEM with energy spectrometer, as shown in Figure 2.

According to Figure 2, magnesium can simultaneously modify oxides and sulfides to form oxide-sulfide composite inclusions. The mechanism of Mg modification is analyzed: Magnesium bubbles have a strong binding ability with dissolved oxygen in molten steel. When Mg is added into molten steel, it will preferentially react with Al and O to form MgO∙Al₂O₃. If the mass fraction of magnesium continues to increase, MgO∙Al₂O₃ will further react with Mg to form MgO. In addition, a very small amount of magnesium reacts with sulfur to form MgS. At the end of solidification, MgO∙Al₂O₃, MgO and a tiny amount of MgS induce the encapsulation of sulfide to form composite inclusions.

The change of the oxide core in the composite sulfides with the increase of magnesium content is: Al₂O₃ → Al₂O₃ + MgO∙Al₂O₃ → MgO∙Al₂O₃ + MgO → MgO → MgO + MgS. The Mg in the composite sulfide (Mg, Mn)S wrapping around the Mg-Al-O oxide core is mainly derived from the diffusion of Mg in the oxide core into the outer MnS.

The ratio of composite sulfide and the ratio of individual MnS with different magnesium content are obtained, as shown in Table 3. The content of composite sulfide with core corresponding to magnesium addition of 0 ppm, 7 ppm, 10 ppm, 19 ppm, 22 ppm are 6.7%, 13.7%, 40.2%, 44.6%, 50.5%, and the percentages of individual MnS are 93%, 83%, 57.8%, 53.4%, 47.5%, respectively.

In order to explore the deformation mechanism of the inclusions after magnesium treatment in the actual rolling process, this experiment used the Gleeble-3500 thermal simulation tester to simulate the hot rolling process. The simulated rolling material is the half of the ingot in the high-temperature furnace modification experiment. The deformation temperature, deformation speed and degree of deformation of the sample can be accurately controlled on the thermal simulation testing machine. It can also deform the sample in multiple passes. After deformation, it can control the cooling of the sample. Some steel samples (diameter

of 10 mm and height of 15 mm) were used to conduct single-direction compression rolled experiments.

The process of this experiment is as follows. First, the rolled sample was heated to a rolling temperature of 1200˚C at 5˚C∙s⁻¹, and then the temperature was maintained for 5 minutes. Next, as shown in Figure 3, the sample was rolled by gleeble-3500, the total deformation along the horizontal axis was 60%, and the rolling time was 0.1 s. After rolling, the rolled sample is cooled by air, and then polished and imaged. The metallographic image is shown in Figure 4.

It can be seen from Figure 4 that the sulfide in the original 49MnVS3 steel is still in a long strip shape after the simulated compression. After modification by adding Mg, the morphology of the sulfide varied from a long strip to a short rod or spindle shape. The result shows that as the magnesium content increases, the degree of deformation of MnS decreases. This is due to the effect of Mg modification treatment, the proportion of single MnS in the steel decreases and the proportion of composite inclusions with core increases. Some studies [12] [13] [14] have shown that composite inclusions have the characteristics of soft outside and hard inside. Compared with MnS with good plasticity, such inclusions are not easy to deform along the rolling direction during hot working, which can significantly improve the mechanical properties of steel and eliminate anisotropy.

A statistical software is used to perform statistics on metallographic photos to explore the degree of deformation of MnS. For each sample, count the inclusions in 100 metallographic photos. According to the observation results of the metallurgical microscope, the shape of inclusions is defined as length/width (L/W), which can be divided into three types: spherical, spindle, long strip and irregular shape. In order to quantitatively describe the deformability of inclusions, the spindle ratio h of the morphology of the inclusions is defined. The spindle ratio h of the inclusions is calculated according to Equation (1), as follows:

h = Spherical number + Spindle number Total inclusion number (1)

After the rolling, the L/W values are recorded as shown in Figure 5.

Figure 5 shows the h-value (1 ≤ L/W ≥ 3) of sulfide inclusion in non-quenched and tempered steel before and after Mg-Ca composite modification. After Gleeble extrusion of the Mg-added sample in the laboratory, the inclusion spindle rate h

increases compared to the original steel with only calcium. When the Mg content in steel changed from 0 to 22 ppm, the spindle rate h of inclusions in steel increased from 31% to 53%. Combined with Table 3, it can be seen that this is because the inclusion core rate of sample e is 50.5%, which is significantly higher than that of sample (a) of 6.7%. The greater the h-value, the number of nucleation cores of composite sulfides. However, numerous composite sulfides with oxide cores could contribute to the small deformation of inclusions during hot-processing and further improve the machinability of steel. Therefore, the Mg addition of 10 to 22 ppm in this non-quenched and tempered steel is beneficial to obtaining more spindle inclusions. Simultaneously, the ratio of composite inclusions with cores can be reaching about 40% - 50%.