Volatilization Behavior of Manganese from Molten Steel with Di ﬀ erent Alloying Methods in Vacuum

: The volatilization loss of manganese during the vacuum smelting process is one of the key factors that determines the manufacturing cost and quality of manganese steel. In this study, the laboratory experiments and thermodynamic calculations were performed to investigate volatilization behavior of manganese from molten steels with di ﬀ erent alloying methods in vacuum process. Based on the thermodynamic analysis, with the increase of manganese content, the partial vapor pressure of the manganese component increased, resulting in manganese being easily volatilized from molten steel. The carbon content in the steel shows an evident inﬂuence on partial vapor pressure of manganese component, and a higher carbon content in steel leads to a lower partial vapor pressure of manganese, but it not inﬂuenced by the silicon content. Compared with the alloying method of high carbon ferromanganese, the volatilization loss of manganese in the alloying method of silicon manganese presents faster decay, agreeing well with the thermodynamic analysis. Besides, the volatile fraction generated in the alloying method of high-carbon ferromanganese is composed of a large amount of MnO nanorods with a lateral length approximately 500 nm and a small number of Mn 3 O 4 / Mn nanoparticles with a diameter less than 500 nm. Additionally, the volatile fraction generated in the alloying method of silicon manganese shows Mn 3 O 4 nanoparticles as the main phase. It can be inferred that the existence of the manganese oxide phase is attributed to the high chemical activity of nanoscale particles within air. of manganese was further characterized. This work provides an improved understanding of the volatilization behaviors of manganese and theoretical basis for the stable control of manganese yield during industrial vacuum secondary reﬁning of medium-manganese steels.


Introduction
Manganese is an indispensable alloying element in steel, owing to its ability to significantly improve the strength and toughness of steel through solid solution strengthening and dispersion strengthening mechanisms [1,2]. In particular, medium manganese steels (5-12 wt% Mn) as the third-generation advanced high-strength steels (AHSS) are in the limelight, and this is attributed to their exhibiting extraordinary combination of ultra-high strength, excellent plasticity, and low production costs, leading to great applications in automotive, construction, and high-speed trains industries [3][4][5][6]. In recent years, the microstructure and mechanical properties of medium manganese steels have been extensively studied [7,8]. However, compared with traditional steels, there are few publications of the smelting process depending on alloying composition like Mn, C, Al, Cr, Mo, Si, N, and V to produce these steel grades. Specifically, the control of high yield and narrow composition of manganese in the smelting process plays an important role in final steel quality and costs of raw materials.
The Ruhrstahl-Heraeus (RH) process was developed for degassing, decarbonization, efficient alloying, inclusion removal, and rapid homogenization of the molten steel, and now it is widely used to achieve high cleanliness for the industrial manufacture of high-quality manganese steels [9,10]. In this In the current study, the theoretical saturation vapor pressure of manganese in molten manganese steel was thermodynamically discussed. The effects of different time and alloying elements on the variation of manganese content in molten steel by the addition of different manganese alloys under vacuum were extensively investigated by the lab experiment. Meanwhile, the obtained volatile fraction of manganese was further characterized. This work provides an improved understanding of the volatilization behaviors of manganese and theoretical basis for the In the current study, the theoretical saturation vapor pressure of manganese in molten manganese steel was thermodynamically discussed. The effects of different time and alloying elements on the variation of manganese content in molten steel by the addition of different manganese alloys under vacuum were extensively investigated by the lab experiment. Meanwhile, the obtained volatile fraction of manganese was further characterized. This work provides an improved understanding of the volatilization behaviors of manganese and theoretical basis for the stable control of manganese yield during industrial vacuum secondary refining of medium-manganese steels.

Experimental Materials
The typical IF (interstitial free) steel and manganese alloy were supplied by Baogang Group (China) and Ningbo Iron and Steel Co., Ltd. (China), respectively. The corresponding chemical composition is shown in Table 1.

Vacuum Experiment Procedure
The manganese volatilization experiment was conducted on a 10 kg vacuum induction furnace ( Figure 2), which consists of a vacuum system, smelting system, sampling/alloying system, pulling system, and cooling system. During experiment, an IF-steel sample was put into a high-purity magnesia crucible (inner diameter, 110 mm; outer diameter, 130 mm; and height, 195 mm), which was assembly positioned crucible-coil in the center of a water-cooled vacuum chamber. The chamber was then vacuumed completely by a mechanical pump, followed by backfill argon to 60 KPa. To ensure there was no residual air in the vacuum chamber, the above process was repeated several times. Subsequently, the IF-steel block was heated by induction heating via adjusting the power. When the temperature reached 1873 K, manganese alloy packed in a tie foil was added into the melt. After homogenization of molten steel for 4 min, the first sample was taken by a cylinder sampler and quenched in water. At this time, the chamber continuously was vacuumed to about 530 Pa. The other two samples were taken at 20 and 40 min, respectively. Two sets of experiments were performed at the same experimental procedure, except for the amount of addition of IF steel and manganese alloy (Table 2). Finally, the manganese steel melt was poured in a mold with the protection of argon. After completion of each experiment for 10 min until the completely solidified, the vacuum furnace was cooled to room temperature, in the air, by opening the furnace door, and the volatile fraction was collected in the vacuum chamber wall.

Sample Characterization
The chemical compositions of the as-obtained manganese steel ingots and steel samples were evaluated by the inductive coupled plasma (ICP, PerkinElmer Analyst ICP-OES 8300, Waltham, Massachusetts, USA,) and direct reading spectrometer. The manganese element distribution of steel ingots was also characterized by electron probe microstructure analysis (EPMA, JXA-Isp100, Tokyo, Japan). Phase compositions of the volatile fraction were analyzed by powder X-ray diffraction (XRD, Smart-lab, Tokyo, Japan), employing a scan rate of 10 • min −1 from 10 • to 90 • with Cu Kα radiation. Meanwhile, morphologies of manganese alloy and volatile fraction were characterized by field-emission scanning electron microscopy (FESEM, Hitachi SU8010, Tokyo, Japan) equipped with X-ray energy-dispersive spectrometry (EDS).

Thermodynamic Analysis
The fundamental physical volatilization of metal elements from the steel can occur in a specific temperature range, which is based on the principle that the saturation vapor pressure significantly contributes to the evaporation characteristics. In general, the higher the saturated vapor pressure of the element, the easier it is to volatilize it. According to the Clausius-Clapeyron relation, the saturation vapor pressure of a pure substance can be calculated by Equation (1) [16]: where p * is the saturated vapor pressure of the pure component (Pa); A, B, C, and D are the evaporation constants for each element; and T is the absolute temperature (K). The relationship between the saturated vapor pressure and temperature of Mg, Ca, Mn, Al, Fe, Si, and Ti element is plotted in Figure 3a [17]. Clearly, the saturated vapor pressure of each pure element significantly increases with the increasing temperature, and the saturated vapor pressure of pure manganese is second only to Ca and Mg at 1600 • C, reaching up to 5370 Pa, indicating that manganese is extremely volatilize into vapor phase under vacuum.
where is the molar fraction of manganese in the molten steel; is the Raoultian activity coefficient of Mn; and and are the activity coefficient and mass percentage based on 1% concentration in Henry's Law. According to Wagner's model [19], the interaction coefficients (Table  4) at the same temperature (1873 K) are used to calculate the activity coefficients in Equation (3).
The partial vapor pressure ( ) of the manganese component is a function of manganese, carbon, and silicon contents, as shown in Figure 3b. It can be clearly seen that the partial vapor pressure of the manganese component increases with the increase of manganese content in molten steel, proving that the high manganese content can facilitates volatilization. In addition, the carbon content in the molten steel has evident effects on the partial vapor pressure of manganese in the same conditions of manganese and silicon content, and a higher carbon content in steel leads to a lower partial vapor pressure of manganese. When the content of silicon increases to 2.0%, the partial vapor pressure of manganese still remains, revealing the partial vapor pressure of manganese is not affected by the silicon content, which is attributed to the interaction coefficient of silicon to manganese that is equal to 0.   Furthermore, the partial vapor pressure (p Mn ) of different content manganese components (Table 3) in molten steel, based on the thermodynamic equilibrium between liquid and vapor phases, can be expressed by the following Equation (2) [18]: where x Mn is the molar fraction of manganese in the molten steel; r Mn is the Raoultian activity coefficient of Mn; and f Mn and w Mn are the activity coefficient and mass percentage based on 1% concentration in Henry's Law. According to Wagner's model [19], the interaction coefficients (   The partial vapor pressure (p Mn ) of the manganese component is a function of manganese, carbon, and silicon contents, as shown in Figure 3b. It can be clearly seen that the partial vapor pressure of the manganese component increases with the increase of manganese content in molten steel, proving that the high manganese content can facilitates volatilization. In addition, the carbon content in the molten steel has evident effects on the partial vapor pressure of manganese in the same conditions of manganese and silicon content, and a higher carbon content in steel leads to a lower partial vapor pressure of manganese. When the content of silicon increases to 2.0%, the partial vapor pressure of manganese still remains, revealing the partial vapor pressure of manganese is not affected by the silicon content, which is attributed to the interaction coefficient of silicon to manganese that is equal to 0.

Effects of Alloying Methods on the Manganese Volatilization
To evaluate the effect different alloying methods on volatilization behavior of manganese in molten steel, the high-carbon ferromanganese and silicon manganese alloy were added during the vacuum process and investigated. The steel samples and ingots obtained are shown in Figure 4a. Figure 4b shows acicular ferrite structure in the ingot, and the corresponding EPMA mapping for manganese element, which indicates manganese element presented severe dendrite segregation, also proving that the existence forms of manganese in the steel.
To evaluate the effect different alloying methods on volatilization behavior of manganese in molten steel, the high-carbon ferromanganese and silicon manganese alloy were added during the vacuum process and investigated. The steel samples and ingots obtained are shown in Figure 4a. Figure 4b shows acicular ferrite structure in the ingot, and the corresponding EPMA mapping for manganese element, which indicates manganese element presented severe dendrite segregation, also proving that the existence forms of manganese in the steel. Furthermore, the analysis results of obtained steel samples are listed in Table 5. It is found that the content of manganese in molten steel with the addition of high-carbon ferromanganese and silicon manganese alloy present the same decreasing tendency with the extension of vacuum time. The content of manganese in molten steel with the alloying method of silicon manganese decreases rapidly in the first 20 min, from 11.05 to 8.95 wt%, which is attributed to the fast pressure drop result in the acceleration of the evaporation of manganese. The slow decrease of manganese content in molten steel in the latter stage and until 60 min reached 7.92 wt%, owing to the gradually decreasing partial vapor pressure caused by the reduced manganese content. Similarly, the content of manganese in molten steel with the addition of high-carbon ferromanganese had a slight attenuation from the initial 11.34 to 8.41 wt% for 60 min. It is confirmed that the increased carbon content in molten steel will lower the partial vapor pressure of component manganese, agreeing with the thermodynamic calculation results in Figure 3b. In addition, the carbon monoxide gas generated in the decarburization reaction of the molten steel in vacuum can also increase the pressure of the vacuum chamber, thus hindering the volatilization of manganese. Furthermore, the effect mechanism of silicon and carbon element on the volatilization of manganese requires more in-depth research.  Furthermore, the analysis results of obtained steel samples are listed in Table 5. It is found that the content of manganese in molten steel with the addition of high-carbon ferromanganese and silicon manganese alloy present the same decreasing tendency with the extension of vacuum time. The content of manganese in molten steel with the alloying method of silicon manganese decreases rapidly in the first 20 min, from 11.05 to 8.95 wt%, which is attributed to the fast pressure drop result in the acceleration of the evaporation of manganese. The slow decrease of manganese content in molten steel in the latter stage and until 60 min reached 7.92 wt%, owing to the gradually decreasing partial vapor pressure caused by the reduced manganese content. Similarly, the content of manganese in molten steel with the addition of high-carbon ferromanganese had a slight attenuation from the initial 11.34 to 8.41 wt% for 60 min. It is confirmed that the increased carbon content in molten steel will lower the partial vapor pressure of component manganese, agreeing with the thermodynamic calculation results in Figure 3b. In addition, the carbon monoxide gas generated in the decarburization reaction of the molten steel in vacuum can also increase the pressure of the vacuum chamber, thus hindering the volatilization of manganese. Furthermore, the effect mechanism of silicon and carbon element on the volatilization of manganese requires more in-depth research.

Analysis of Manganese Volatile Phase
The morphologies of high-carbon ferromanganese and silicon manganese alloy conducted by field-emission scanning electron microscopy (FESEM) are shown in Figure 5, and irregular lump shape was observed. As for volatile fraction generated in the vacuum process, which was condensed on the inner wall of the vacuum chamber and deposit to a certain thickness, it can be observed in Figure 6a, b. To further explore the effects of different alloying methods on manganese volatile phase structure, XRD analysis of volatile fraction, in addition to high-carbon ferromanganese and silicon manganese, was carried out. As shown in Figure 6c, the volatile fraction generated in the alloying method of high-carbon ferromanganese was mainly composed of MnO (JCPDS No. 89-2804) and a small amount of Mn 3 O 4 (JCPDS No. 75-1560) and Mn (JCPDS No. 17-910). The appearance of manganese oxide phase is attributed to continuous evaporation of manganese at first condensation on the inner wall of the vacuum chamber and then oxidized by air, to form Mn x O, after opening the furnace door. After the silicon manganese was added, the main phase of volatile fraction was transformed into Mn 3 O 4 , and a small number of Mn and MnO phases were generated (Figure 6d). This may be correlated to the particle size of volatile manganese, which affects the thermodynamic stability of manganese oxide phase [20].
The morphologies of high-carbon ferromanganese and silicon manganese alloy conducted by field-emission scanning electron microscopy (FESEM) are shown in Figure 5, and irregular lump shape was observed. As for volatile fraction generated in the vacuum process, which was condensed on the inner wall of the vacuum chamber and deposit to a certain thickness, it can be observed in Figure 6a, b. To further explore the effects of different alloying methods on manganese volatile phase structure, XRD analysis of volatile fraction, in addition to high-carbon ferromanganese and silicon manganese, was carried out. As shown in Figure 6c, the volatile fraction generated in the alloying method of high-carbon ferromanganese was mainly composed of MnO (JCPDS No. 89-2804) and a small amount of Mn3O4 (JCPDS No. 75-1560) and Mn (JCPDS No. 17-910). The appearance of manganese oxide phase is attributed to continuous evaporation of manganese at first condensation o  field-emission scanning electron microscopy (FESEM) are shown in Figure 5, and irregular lump shape was observed. As for volatile fraction generated in the vacuum process, which was condensed on the inner wall of the vacuum chamber and deposit to a certain thickness, it can be observed in Figure 6a, b. To further explore the effects of different alloying methods on manganese volatile phase structure, XRD analysis of volatile fraction, in addition to high-carbon ferromanganese and silicon manganese, was carried out. As shown in Figure 6c, the volatile fraction generated in the alloying method of high-carbon ferromanganese was mainly composed of MnO (JCPDS No. 89-2804) and a small amount of Mn3O4 (JCPDS No. 75-1560) and Mn . The appearance of manganese oxide phase is attributed to continuous evaporation of manganese at first condensation on the inner wall of the vacuum chamber and then oxidized by air, to form MnxO, after opening the furnace door. After the silicon manganese was added, the main phase of volatile fraction was transformed into Mn3O4, and a small number of Mn and MnO phases were generated (Figure 6d). This may be correlated to the particle size of volatile manganese, which affects the thermodynamic stability of manganese oxide phase. [20]   Simultaneously, the morphology of volatile fraction was characterized by FESEM, which is completely different from the morphology of the manganese alloy, further confirming that the volatile fraction only comes from the volatilization of manganese element in the molten steel. As shown in the FESEM images in Figure 7a,b, volatile fraction generated in alloying method of high-carbon ferromanganese exhibited a nanoscale morphology, including a large amount of nanorods with a lateral length approximately 500 nm and a small number of particles with less than 500 nm. Based on XRD analysis results, the nanorod structure can be ascribed to MnO phase. In contrast, all uneven nanoparticles with a diameter less than 500 nm can be discovered in Figure 7c,d, which corresponds to volatile fraction generated in alloying method of silicon manganese. It is worth mentioning that the volatile manganese was oxidized owing to the high chemical activity of nanoscale particles [21]. Besides this, the EDS mapping images (Figure 7e) further verify element Mn and O uniform distributed in entire nanoparticles, matching well with XRD analysis results. In addition, the corresponding element content analysis (inset of Figure 7e) shows that the mass fraction of Mn and O is 82.67% and 17.33%, respectively, confirming the existence of the manganese oxide phase.
vacuum treatment. XRD patterns of volatile fraction with different alloying methods of (c) highcarbon ferromanganese and (d) silicon manganese.
Simultaneously, the morphology of volatile fraction was characterized by FESEM, which is completely different from the morphology of the manganese alloy, further confirming that the volatile fraction only comes from the volatilization of manganese element in the molten steel. As shown in the FESEM images in Figure 7a,b, volatile fraction generated in alloying method of highcarbon ferromanganese exhibited a nanoscale morphology, including a large amount of nanorods with a lateral length approximately 500 nm and a small number of particles with less than 500 nm. Based on XRD analysis results, the nanorod structure can be ascribed to MnO phase. In contrast, all uneven nanoparticles with a diameter less than 500 nm can be discovered in Figure 7c,d, which corresponds to volatile fraction generated in alloying method of silicon manganese. It is worth mentioning that the volatile manganese was oxidized owing to the high chemical activity of nanoscale particles [21]. Besides this, the EDS mapping images (Figure 7e) further verify element Mn and O uniform distributed in entire nanoparticles, matching well with XRD analysis results. In addition, the corresponding element content analysis (inset of Figure 7e) shows that the mass fraction of Mn and O is 82.67% and 17.33%, respectively, confirming the existence of the manganese oxide phase.

Conclusions
In the current study, thermodynamic calculation was performed to determine the partial vapor pressure of manganese component in molten steel. The volatilization behavior of manganese was investigated by laboratory experiments for manganese steels, with different alloying methods, in vacuum process. The following conclusions can be obtained:

Conclusions
In the current study, thermodynamic calculation was performed to determine the partial vapor pressure of manganese component in molten steel. The volatilization behavior of manganese was investigated by laboratory experiments for manganese steels, with different alloying methods, in vacuum process. The following conclusions can be obtained:

1.
With the increase of manganese content, the partial vapor pressure of manganese component increased, resulting in manganese components being easily volatilized from molten steel. Moreover, the greater the carbon content in the steel, the lower the partial vapor pressure of manganese component. However, the partial vapor pressure of manganese component is not influenced by the silicon content in molten steel, which is depends on the value of the interaction coefficient.
As a result, an obvious decrease trend of manganese content in the steel under vacuum process with time can be found. Compared with the alloying method of high-carbon ferromanganese, the volatilization loss of manganese in the alloying method of silicon manganese presents faster decay. The observed results were in good agreement with thermodynamic analysis. 3.
The volatile fraction generated in alloying method of high-carbon ferromanganese is composed of a large amount of MnO nanorods with a lateral length approximately 500 nm and a small number of Mn 3 O 4 /Mn nanoparticles with diameter less than 500 nm. Meanwhile, the volatile fraction generated in alloying method of silicon manganese shows Mn 3 O 4 nanoparticles as the main phase. It can be inferred that the existence of the manganese oxide phase is attributed to the high chemical activity of nanoscale particles in air.