Effect of Residual Element Tin on the Inclusion Characteristics and the Interaction between Lanthanum and Tin in Interstitial Free Steel

Abstract

Recycling of Scrap will lead to the continuous accumulation of residual element Tin (Sn) in steel production. The influence of Sn on interstitial-free (IF) steel and the formation of Sn compounds were investigated by experiments. The types of inclusions in IF steel with Sn is TiN, Al₂O₃, MnS and Ti₂S₃, which are consistent with those in IF steel. After adding Sn from 0% to 0.5%, the starting solidification temperature of molten steel decreases, resulting in the increase of inclusion number density. The average size of inclusions decreases from 2.39 µm to 1.89 µm. After adding 0.065%La, the inclusions containing Sn, such as La–Sn, La–Sn–O and La–Sn–Ti, generate in IF steel with 0.1%Sn or 0.5%Sn. With the increase of Sn content in IF steel containing 0.065%La, the number density of inclusions decreases and the number of large size inclusions containing Sn (≥10 µm) increases from 121 inclusions to 284 inclusions. Furthermore, the addition of Sn and La refines the macrostructure of ingots. Thermodynamic calculation shows that La₂Sn and La₅Sn₄ inclusions can form during the solidification process in sample with 0.1%Sn and 0.065%La. And in sample with 0.5%Sn and 0.065%La, La₂Sn and La₅Sn₄ inclusions can form directly in the molten steel. The generation of inclusions containing Sn can change the existing state of Sn and is beneficial in inhibiting the grain boundary segregation of Sn. The experiment lays a theoretical foundation for the use of La to minimize the harmful effects of Sn in IF steel in the future.

1. Introduction

Carbon emissions in iron and steel industry account for about 16% of countrywide carbon emissions in China.¹⁾ The proposal of the carbon peaking and the carbon neutrality goals has put tremendous pressure on the metallurgical industry to achieve energy saving and carbon reduction. Comparing the integrated iron and steel plant with iron ore as raw materials, electric arc furnace steelmaking (EAF) with scrap and direct-reduction ironmaking (DRI) as raw materials can reduce carbon emissions by 60%.¹⁾ EAF steelmaking using scrap will become an important way to achieve low-carbon metallurgy. What’s more, the automotive industry accounts for a large proportion of global steel consumption, with more than half of that being plates. To achieve the carbon neutrality goals, addressing how to use EAF steelmaking for high-grade sheet metal such as interstitial-free steel (IF) is a pressing issue for the metallurgical industry.

The continuous and stable growth of social scrap also provides the basic guarantee for EAF steelmaking.²⁾ However, the social scrap has a wide range of sources, various types and unstable composition. In conventional steelmaking processes, residual elements in the scrap, such as Cu, Sn, Sb and As, are difficult to remove due to their lower oxidation potential compared to Fe.³⁾ With the recycling of scrap, residual elements continue accumulation in steel production, causing dendritic segregation during the solidification of molten steel, grain boundary segregation and elements accumulation at the scale-steel interface.^4,5,6,7,8) These phenomena severely deteriorate the hot working properties, temper brittleness, and mechanical properties of steel.^9,10) And residual elements can affect the operation stability of continuous casting and rolling processes.^11,12,13) Moreover, residual elements also have some benefits, such as the improvement of corrosion resistance in steel by adding residual element Sn.^14,15,16) It is crucial to study how to reduce the content of harmful residual elements in steel or mitigate their detrimental effects.

Tinplate is widely used in food and beverage cans, among other applications. Therefore, Sn in scrap steel is an important and unavoidable residual element. Currently, there are several methods to reduce the content of Sn in steel, such as ore roasting, scrap pre-treatment, steam pressure method, calcium reaction method and so on.^17,18,19) However, due to the economic costs and environmental issues, these methods cannot be applied to industrial production. These methods do not fundamentally resolve the harmful effects of residual element Sn.

Adding rare earth elements (such as lanthanum, cerium and yttrium) to steel generally serves to purify the molten steel (such as deoxidation and desulfurization), micro-alloying, and modifying inclusions.^20,21,22) Moreover, researchers found that rare earth can form high melting point compounds with low melting point residual elements and fix them.²³⁾ Wang et al.²⁰⁾ had found that adding La to molten steel can form arsenic compounds like La–S–As, La–As, and LaAsO₄. The formation of As compounds changes the existing state of As, and reduces the As content in the matrix.²⁰⁾ What’s more, rare earth can also create competitive segregation with residual elements, and reduce residual element content at grain boundaries.²⁴⁾ Seah and Wang et al.^25,26) used first-principles calculations to explain the strengthening effect of rare earth segregation at grain boundaries and the embrittlement effect of residual elements on grain boundaries. The strengthening of grain boundaries can improve the high temperature ductility of steel billets, ensuring smooth continuous casting and hot working processes.^27,28,29) Xin et al.^30,31) also found the formation of Ce–As compounds by Ce treatment. The addition of Ce reduced the content of As at grain boundaries and improved the high-temperature ductility of steel. Sha et al.³²⁾ also observed the formation of La–Sn compounds in their research on the effect of La on fixing residual element Sn.

The harmful effects of residual elements are a bottleneck issue limiting the development of electric arc furnace smelting of high-grade sheet metals, such as IF steel. Previous research has shown that rare earth elements can mitigate the harmful effects of residual elements. However, most studies have focused on the effects of La and Ce on residual element As. There is a lack of systematic research on the stabilization and mitigation effects of La on residual element Sn. Therefore, this work initially investigated the influence of Sn on inclusions and the interaction between La and Sn in IF steel. The change in the chemical composition, morphology, size and number of inclusions were analyzed. The macrostructure of ingots was also observed. Furthermore, the formation of inclusions containing Sn was analyzed by combining thermodynamic calculations with experimental results. This work can provide a theoretical basis to mitigate the harmful effects of Sn in IF steel by La treatment.

2. Experimental Methodology

2.1. Experimental Procedure

The experimental materials include IF steel, pure tin granules (purity >99% Sn) and lanthanum (purity >99% La). IF steel was sourced from the IF steel slab produced by a commercial steelmaking plant. To investigate the effect of Sn on inclusion characteristics and the interaction between Sn and La in IF steel, five groups of experiments were set up (0%Sn, 0.1%Sn, 0.5%Sn, 0.1%Sn–0.05%La, 0.5%Sn–0.05%La). Experimental samples were smelted using a laboratory MoSi₂ resistance furnace, as shown in Fig. 1(a). IF steel raw materials (about 800 g), which had been polished to remove the oxide layer, were placed into a corundum crucible (60 mm outer diameter, 200 mm height) with a graphite protection crucible outside to prevent molten steel leakage. A corundum sleeve was placed on top of the graphite protection crucible to protect the furnace tube. Argon gas was continuously supplied throughout the smelting process as a protective atmosphere. The resistance furnace was heated to 1873 K and held for 20 minutes until the steel ingots melted. Different amounts of Sn and La were sequentially wrapped in iron sheets. The iron sheet containing La and Sn was attached to one end of an iron rod, then inserted into the molten steel and stirred to ensure uniform dispersion. After holding 10 minutes, the furnace entered the cooling program. The power was turned off when the temperature of furnace was decreased at a rate of 10 K/min to 1773 K. The samples were cooled to room temperature in the furnace.

Fig. 1. (a) MoSi₂ resistance furnace diagram; (b) schematic diagram of sampling locations for metallographic (Sample A) and macrostructure (Sample B) samples. (Online version in color.)

2.2. Analysis Methods

Using a drilling machine, chip-like samples were taken from the middle of the specimens for the detection and analysis of elements (C, S, Si, Mn, P, Al, Ti, Sn and La). The samples (Ф5 mm × 8 mm) were cut using wire cutting for the detection of O and N. The content of carbon and sulfur in the experimental samples was measured using a high-frequency infrared carbon-sulfur analyzer (HCS-801, Sichuan Science Instrument Co., Ltd). The content of oxygen and nitrogen was measured using an oxygen-nitrogen-hydrogen analyzer (ONH-3000, NCS Testing Technology Co., Ltd). Other elements (Si, Mn, P, Al, Ti, Sn and La) were measured using inductively coupled plasma atomic emission spectrometry (ICP-AES). The chemical composition of the samples is shown in Table 1. In the following text, Sn0, Sn1, Sn5, Sn1-La and Sn5-La represent the samples with about 0%Sn, 0.113%Sn, 0.526%Sn, 0.109%Sn–0.065% La and 0.537%Sn–0.063%La, respectively.

Table 1. Chemical composition of samples (Mass%).

Sample	C	Si	Mn	P	S	Al	Ti	N	O	Sn	La
Sn0	0.0016	0.0030	0.1040	0.0125	0.0060	0.0571	0.0568	0.0028	0.0019	–	–
Sn1	0.0017	0.0031	0.1038	0.0126	0.0058	0.0569	0.0566	0.0025	0.0018	0.113	–
Sn5	0.0017	0.0029	0.1037	0.0123	0.0059	0.0572	0.0565	0.0023	0.0019	0.526	–
Sn1-La	0.0016	0.0028	0.1039	0.0124	0.0057	0.0574	0.0569	0.0027	0.0019	0.109	0.065
Sn5-La	0.0018	0.0032	0.1042	0.0127	0.0060	0.0568	0.0567	0.0032	0.0022	0.537	0.063

As shown in Fig. 1(b), a metallographic sample (sample A) with dimensions of 10 mm × 10 mm × 5 mm was taken from 5 mm above the bottom center of the ingot. Metallographic samples were successively polished using sandpapers with grits of 400, 600, 800, 1200, 1500, and 2000, followed by polishing using abrasive pastes with grits of 3000, 8000, and 15000. A scanning electron microscope with energy spectrum analyzer (SEM-EDS, voltage 20 kV) was used to observe and preliminarily identify the types and compositions of inclusions. To observe the three-dimensional morphology of the inclusions, the metallographic sample was electrolyzed with a 2% TEA non-aqueous solution (2 v/v% triethanolamine - 1 mass/v% tetramethylammonium chloride - methanol solution) for 10 minutes (current 0.02 A) and observed using SEM-EDS. Moreover, in order to quantitatively characterize inclusions, an automatic inclusion analyzer (Aspex) (scanning area: 49 mm²) was used to analyze the number, size and chemical composition of inclusions. The minimum detection size of inclusions is 0.8 μm. Combining thermodynamic calculations, the formation of La–Sn inclusions was analyzed. Furthermore, a sample B was cut from 25 mm below the top of the ingot to observe the macrostructure of the horizontal cross-section at this location. Sample B was ground, polished with polishing paste and etched with 4% nitric acid in alcohol to observe macrostructure.

3. Experimental Results and Discussion

3.1. Types and Morphology of Inclusions

The morphology and composition of inclusions in the experimental samples were observed using SEM-EDS. The types of inclusions in different samples are shown in Table 2. Figure 2 shows the morphology of typical inclusions in Sn0, Sn1 and Sn5 samples. The inclusions are mainly TiN, Al₂O₃, MnS, Ti₂S₃ and their composite inclusions. The size of MnS and Ti₂S₃ inclusions are generally smaller compared to TiN and Al₂O₃–TiN inclusions. The addition of Sn does not change the types of inclusions. Inclusions containing Sn do not appear in the Sn1 and Sn5 samples.

Table 2. The main types of inclusions in samples.

Sample	Type
Sn0	TiN, Al2O3, MnS, Ti2S3
Sn1	TiN, Al2O3, MnS, Ti2S3
Sn5	TiN, Al2O3, MnS, Ti2S3
Sn1-La	TiN, LaS, La2O2S, La–P–S, La–Sn, La–Sn–Ti, La–Sn–O
Sn5-La	TiN, LaS, La2O2S, La–P–S, La–Sn, La–Sn–Ti, La–Sn–O

Fig. 2. SEM-EDS images of typical inclusions in Sn0, Sn1 and Sn5 samples: (a) TiN–MnS; (b) Al₂O₃–TiN; (c) MnS–Ti₂S₃; (d) Ti₂S₃. (Online version in color.)

After adding 0.065%La in IF steel with 0.1%Sn or 0.5% Sn, the types of inclusions are divided into inclusions with or without Sn. Figure 3 shows the typical inclusions without Sn in Sn1-La and Sn5-La samples, such as TiN, LaS, La₂O₂S, La–P–S and their composite inclusions. Figure 4 shows the typical inclusions containing Sn, such as La–Sn, La–Sn–O, La–Sn–Ti and their composite inclusions with La–P–S and TiN.

Fig. 3. SEM-EDS diagram of typical inclusions without Sn in Sn1-La and Sn5-La samples: (a) LaS–TiN; (b) La₂O₂S–TiN; (c) La₂O₂S–La–P–S. (Online version in color.)

Fig. 4. SEM-EDS diagram of typical inclusions containing Sn in Sn1-La and Sn5-La samples: (A1) La–Sn–TiN; (A2) La–P–S–Sn–Ti–O; (A3) La–P–S–Sn–Ti–O; (B1) La–Sn–O; (B2) La–P–S–Sn–O–TiN; (B3) La–P–S–Sn–Ti–O; (B4) La–P–S–Sn–O–TiN; (B5) La–P–S–Sn–TiN. (Online version in color.)

Comparing Sn1-La and Sn5-La samples, the Sn1-La sample contains fewer standalone inclusions containing Sn than the Sn5-La sample. According to the automatic inclusion analysis results, the highest Sn content in inclusions of Sn1-La and Sn5-La samples are 43.1% and 84.8%, respectively. Some inclusions containing Sn are mainly distributed outside La–P–S, liking the distribution of TiN inclusion. The standalone and composite inclusions containing Sn are shown in Figs. 4(B1) to 4(B5) in Sn5-La sample. There are many La–Sn compounds in La–Sn system. Sha et al.³²⁾ also observed the phenomenon of heterogeneous nucleation of La–Sn on La–P as the core. In addition, according to EDS results, the composition of La–Sn inclusion in Figs. 4(A1), 4(B5) and 5(a) is close to La₅Sn₄, La₂Sn and La₅Sn₄, respectively. It is indeed difficult to identify compounds with similar compositional ratios, such as La₅Sn₄ and La₁₁Sn₁₀. Although this is indeed not entirely accurate, La₅Sn₄ was identified form the ratio of the inclusions’ chemical element by EDS analysis. Line scanning of the inclusions reveals good consistency in the energy fluctuations of the La and Sn elements, as shown in Fig. 5(b). The formation of inclusions containing Sn may decrease Sn content in the steel matrix and suppress Sn segregation to grain boundaries.

Fig. 5. (a) The SEM images and elemental mapping of La–P–S–Sn–TiN inclusion in the Sn5-La sample; (b) the line scanning of La–P–S–Sn–TiN inclusion. (Online version in color.)

As shown in Fig. 6, the three-dimensional morphology of the typical inclusions was observed in the samples, such as Al₂O₃–TiN–Ti₂S₃, La–O–P–S–TiN, and La–O–P–S–Sn–TiN. Multiple TiN inclusions may heterogeneously nucleate on inclusions containing La as the core and aggregate into large size inclusions in samples containing rare earth La, as shown in Figs. 4(B4) and 6(d).

Fig. 6. The three-dimensional morphology of typical inclusions: (a) Al₂O₃–TiN–Ti₂S₃; (b) La–P–S–TiN; (c) La–O–P–S–Sn–TiN; (d) La–O–P–S–Sn–TiN. (Online version in color.)

3.2. Size, Number and Composition Analysis of Inclusions

3.2.1. Size Distribution and Number of Overall Inclusions

Using automatic inclusion analysis technology, inclusions within an analysis area of 49 mm² for each sample were scanned to determine the types, number, distribution, and size variation. The average diameter of inclusions is used as the size of the inclusions. Figure 7 shows the distribution of inclusions of various sizes in the samples (area: 5 mm×8 mm), which allows for a macroscopic observation of the changes in the number and size of inclusions. From Figs. 7(a) to 7(c), it can be seen that as the Sn content increases, the number of inclusions gradually increases. In particular, the number of inclusions around ≤5 μm significantly increases. Compared to the samples with only Sn added, the number of inclusions significantly increases in the samples with La and Sn elements added, as shown in Fig. 7. In samples containing 0.065%La, as the Sn content increases from 0.1% to 0.5%, the number of large inclusions (≥15 μm) in the steel significantly increases.

Fig. 7. The distribution of inclusions in samples: (a) Sn0; (b) Sn1; (c) Sn5; (d) Sn1-La; (e) Sn5-La. (Online version in color.)

To focus on the size distribution of inclusions, inclusion numbers in different sizes and average inclusion sizes were shown in Fig. 8. Number densities of inclusion were shown in Fig. 9. Based on the results of automatic inclusions analysis, the quantity of a particular type of inclusion is determined by screening the elemental content of the inclusions. From Fig. 9, it can be seen that as the Sn content increases, the number density of inclusions significantly increases from 24.51 mm⁻² to 87.25 mm⁻². From Fig. 8(a), it can be observed that as the Sn content increases, the number of inclusions smaller than around 5 μm increases, which is consistent with the macroscopic results in Fig. 7. The average size of the inclusions decreases from 2.39 μm to 1.89 μm, as shown in Fig. 8(b). As the Sn content increases from 0.1% to 0.5%, the number density of inclusions decreases from 185.47 mm⁻² to 151.99 mm⁻² under the same La content (0.065% La), as shown in Fig. 9. Moreover, the number of inclusions greater than 15 μm in the Sn5-La sample is significantly higher than in the Sn1-La sample, as shown in Figs. 7 and 8. This may be due to the higher Sn content in Sn5-La sample, resulting in the easier formation of inclusions containing Sn and promoting inclusions cluster, as shown in Figs. 4(B4) and 6(d). The heterogeneous nucleation of multiple TiN inclusions with rare earth inclusions (La–P–S–O–Sn–Ti) as the core resulted in the formation of large size inclusions with clusters.

Fig. 8. (a) The relationship between the number and size distribution of inclusions; (b) the average size of inclusions. (Online version in color.)

Fig. 9. Number density of all inclusions in each sample (mm⁻²). (Online version in color.)

Number densities and cumulative frequency distributions for inclusion sizes were shown in Fig. 10. From Figs. 10(a) to 10(c), it can be seen that as the Sn content increases, the size distribution of inclusions shifts to the left. When the Sn content changes from 0% to 0.5%, d₅₀ decreases from 1.37 μm to 1.07 μm and d₉₀ decreases from 4.68 μm to 2.81 μm. The cumulative frequency curve of inclusion sizes indicates that the maximum size of inclusions also decreases from 26.3 μm to 20.4 μm with the increasing of Sn content. Under the same La content, as the Sn content increases from 0.1% to 0.5%, d₅₀ decreases from 2.23 μm to 1.86 μm, indicating an increase in the number of smaller inclusions. However, d₉₀ increases from 5.32 μm to 5.83 μm, indicating that both the number of large size inclusions and the maximum inclusion size increase. In summary, adding Sn increases the number of inclusions in IF steel and decreases the average size of the inclusions. After adding La and Sn, the number of inclusions further increases. Under the same La content, as the Sn content increases, the number density of inclusions decreases and the average size of inclusions decreases. However, the number of large inclusions (≥15 μm) and the maximum inclusion size both increase.

Fig. 10. The number density and cumulative frequency of inclusions are compared with the size distribution: (a) Sn0; (b) Sn1; (c) Sn5; (d) Sn1-La; (e) Sn5-La. (Online version in color.)

3.2.2. Variation of TiN Inclusions

The types of inclusions obtained through automatic inclusion analysis technology are shown in Fig. 9. And the results are consistent with the SEM-EDS results. The main inclusions are TiN in samples with the increase of the Sn content from 0 to 0.5%. Moreover, as the Sn content increases, the quantity of TiN inclusions in the samples significantly increases. Figure 11(a) shows the changes in the mass fractions of the liquid phase and BCC phase with temperature in the Sn0, Sn1, and Sn5 samples, calculated using FactSage8.2. It can be seen that as the Sn content increases, starting temperature for formation of BCC decreases. This indicates that the addition of Sn can lower the initial temperature of solidification. The main inclusions in the Sn0, Sn1, and Sn5 samples are TiN, which form during the solidification of the molten steel, as shown in Fig. 11(b). The solubility product required of TiN formation decreases with the decrease of initial solidification temperature. The lower the initial solidification temperature, the greater the amount of TiN nucleation cores formed. Therefore, the addition of Sn increased the number of inclusions in the samples.

Fig. 11. (a) The mass fraction of liquid phase and BCC phase of Sn0, Sn1 and Sn5 samples changed with temperature during the solidification process; (b) the relationship between Ti and N formed by TiN in liquid steel and the relationship between a_[Ti]‧a_[N] at solidification front and solid phase ratio. (Online version in color.)

3.2.3. Inclusions Containing Sn

In the Sn1-La and Sn5-La samples, the number of large size inclusions increases with the increase of the Sn content. Inclusions larger than 10 μm were classified in the Sn1-La and Sn5-La samples, as shown in Fig. 12. It was found that as the Sn content increases from 0.1% to 0.5%, the number of large size inclusions (≥10 μm) in the steel within the scanning area (49 mm²) increases from 121 to 284. In the Sn5-La sample, 84.5% of the large size inclusions (≥10 μm) contain Sn element, while only 21.5% of the large size inclusions (≥10 μm) in the Sn1-La sample contain Sn element.

Fig. 12. The number of large-sized inclusions (≥10 μm) in the Sn1-La and Sn5-La samples. (Online version in color.)

Figure 13 shows the size distribution of inclusions containing Sn in the Sn1-La and Sn5-La samples. It can be seen that there are more large-sized inclusions containing Sn in the Sn5-La sample compared to the Sn1-La sample. The maximum size of inclusions is 26 μm in the Sn1-La sample. However, there are inclusions containing Sn with sizes around 34 μm in the Sn5-La sample. It is considered that the formation of inclusions containing Sn reduces the Sn content in the steel matrix, thereby suppressing Sn segregation to the grain boundaries. Moreover, the addition of the rare earth element La can form competitive segregation with Sn and reduce the grain boundary segregation of Sn at grain boundaries, which decreases its impact on the high-temperature thermos-plasticity of the steel. Furthermore, the presence of large-sized inclusions containing Sn provides a potential for reducing the Sn content in the steel through the flotation and removal of these large inclusions.

Fig. 13. (a) Size distribution of Sn-containing inclusions in Sn1-La and Sn5-La samples; (b) local magnification of size distribution of inclusions larger than 7 μm. (Online version in color.)

3.3. Macrostructure Characterization

Using a 4% nitric acid alcohol solution to etch the ingots, the macrostructure of the samples was obtained, as shown in Fig. 14. And the average grain size of macrostructure (with an area of approximately 490 mm²) was measured by the intercept method, as shown in Fig. 15. Figures 14(a) to 14(c) reveal that as the content of Sn increases from 0% to 0.5%, the macrostructure of the samples becomes refined and the average grain size of macrostructure decreases from 1.049 mm to 0.654 mm. Combined with the results of automatic inclusion analysis, the refinement of macrostructure may be due to the increase of inclusions number. The main inclusions in the Sn0, Sn1, and Sn5 samples are TiN and TiN–Al₂O₃–(MnS/Ti₂S₃) composite inclusions. TiN can act as a heterogeneous nucleation core for δ-Fe and promote the formation of δ-Fe.^33,34) The number density of TiN and composite inclusions in the samples increases with the increase of Sn content, resulting in the increase of heterogeneous nucleation cores for δ-Fe. Zhou et al. found that the increase in the number density of TiN inclusions can cause the higher upper limit of the I_heter. (the heterogeneous nucleation rate). And the higher the value of the I_heter., the more refinement of the macrostructure. Additionally, the addition of rare earth La leads to the formation of rare earth inclusions in the samples, which can also act as a heterogeneous nucleation core for TiN and promote the formation of TiN. TiN inclusions can also serve as heterogeneous nucleation cores for δ-Fe, further refining the macrostructure of the ingots, as shown in Figs. 14(d) to 14(e) and 15. Many previous studies have found that rare earth elements have the effect of refining grain.³⁵⁾

Fig. 14. Macrostructure of ignots: (a) Sn0; (b) Sn1; (c) Sn5; (d) Sn1-La; (e) Sn5-La. (Online version in color.)

Fig. 15. Average grain size of macrostructure: (a) Sn0; (b) Sn1; (c) Sn5; (d) Sn1-La; (e) Sn5-La. (Online version in color.)

Furthermore, grain boundary energy is the driving force for grain growth. The segregation of residual Sn at the grain boundaries reduces the grain boundary energy, weakening the driving force for grain growth and affecting grain growth. As the Sn content increases, the segregation amount of Sn at the grain boundaries gradually increases, effectively inhibiting grain boundary migration, and leading to a significant reduction in grain size.³⁶⁾ Additionally, rare earth elements also tend to segregate at grain boundaries and hinder grain growth. Therefore, the refinement of macrostructure in the samples is likely the result of the combined effects of residual element Sn and rare earth element La grain boundary segregation and the increase in the number density of inclusions.

4. Thermodynamic Analysis on Formation of Inclusions

Due to the absence of La–P and La–Sn databases in FactSage8.2 software, the predominance area diagram of La–Al–Ti–O–S–Fe system at 1873 K was selected for calculation, as shown in Fig. 16. It can be seen from Fig. 16 that Al₂O₃ inclusions are present in IF steel without La or a small amount of La at 1873 K. It could be seen from Fig. 11(b) that TiN inclusion forms during solidification of molten steel. When 0.065%La was added to the liquid steel, LaS and La₂O₂S can directly generated in the liquid steel. Due to the lack of reliable thermodynamic data for La–P–S, La–Sn–Ti, and La–Sn–O, we are currently unable to calculate their relevant thermodynamics.

Fig. 16. La–Al–Ti–O–S–Fe system advantage region diagram at 1873 K. (Online version in color.)

4.1. The Formation of La–Sn Inclusions under Equilibrium Condition

La and low melting point residual element Sn can form high-melting point La–Sn compounds. According to the SEM results, the distribution of La–Sn inclusions is mostly around the periphery of La–P–S inclusions, similar to the distribution of TiN. In order to explore the formation of La–Sn compounds, thermodynamic calculations were performed on some La–Sn compounds. Many thermodynamic data for La–Sn compounds are incomplete. Table 3 shows the standard Gibbs free energy expressions for some La–Sn compounds, such as La₂Sn and La₅Sn₄.^32,37)

Table 3. Standard Gibbs free energy of La and Sn reactions in molten steel.^32,37)

Reactions	ΔGθ=A+BT (J/mol)
2[La]+[Sn]=La2Sn(s)	−580366+250.2 T
5[La]+4[Sn]= La5Sn4(s)	−1224390+452.95 T

Taking La₂Sn as an example, the relationship between the concentrations of La and Sn when La₂Sn is formed was calculated at the liquidus and solidus temperatures. Additionally, it is necessary to calculate the relationship between the actual activity product ${\text{(}{a}_{\text{[La]}}^{\text{2}}\cdot a_{\text{[Sn]}}\text{)}}_{a\text{c}}$ and the equilibrium activity product ${\text{(}{a}_{\text{[La]}}^{\text{2}}\cdot a_{\text{[Sn]}}\text{)}}_{\text{eq}}$ of La₂Sn with temperature. The formation reaction of La₂Sn in the molten steel is shown as follows:

  

$$2[\mathrm{L}\mathrm{a}]+[\mathrm{S}\mathrm{n}]=\mathrm{L}{\mathrm{a}}_2\mathrm{S}\mathrm{n}(\mathrm{s})$$(1)

  

$$\mathrm{\Delta }{\mathrm{G}}^{\theta }=-580\hspace{0.17em}366+250.2\hspace{0.17em}\mathrm{T}=-\mathrm{R}\mathrm{T}\mathrm{l}\mathrm{n}\mathrm{K}$$(2)³²⁾

where [La] and [Sn] refer to the La and Sn elements within the molten steel; R is the molar gas constant, 8.314 J/(mol·K). K is the equilibrium constant and can be expressed as:

  

$$\mathrm{K}=\frac{a_{{\text{La}}_{\text{2}}\text{Sn(s)}}}{a_{\text{[La]}}^{\text{2}}\cdot a_{\text{[Sn]}}}=\frac{\text{1}}{\text{(}{f}_{\text{[La]}}^{\text{2}}{w}_{\text{[La]}}^{\text{2}}\text{)}\cdot \text{(}{f}_{\text{[Sn]}}{w}_{\text{[Sn]}}\text{)}}$$(3)

  

$$\begin{array}{l}\mathrm{L}\mathrm{o}\mathrm{g}{a}_{\text{[La]}}^{\text{2}}\cdot a_{\text{[Sn]}}=\mathrm{L}\mathrm{o}\mathrm{g}{w}_{\text{[La]}}^{\text{2}}{w}_{\text{[Sn]}}+\text{2Log}{f}_{\text{[La]}}+\text{Log}{f}_{\text{[Sn]}}\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}=-\frac{\text{30}\mathrm{\hspace{0.17em} }\text{310}\text{.84}}{\text{T}}+\text{13}\text{.07}\end{array}$$(4)

where $a_{{\text{La}}_{\text{2}}\text{Sn(s)}}$, a_[La] and a_[Sn] are the activities of La₂Sn, La and Sn in the liquid steel, respectively; w_[La] and w_[Sn] are mass% of La and Sn in the molten steel, mass%. The calculation uses Hernian 1 mass% standard state. f_[La] and f_[Sn] are the activity coefficients of La and Sn in the molten steel and can be calculated from the activity interaction coefficients $e_i^j$. Since the activity of pure substances was considered to be 1 during calculations, the value of $a_{{\text{La}}_{\text{2}}\text{Sn(s)}}$ was regarded as 1.

Due to the low content of each component in IF steel, it was assumed that the effect of temperature on activity is small enough to be negligible. The temperature dependence of activity coefficients was not considered in the following calculation. The activity coefficient f_i of elements in steel can be obtained by Eq. (5) and Table 4.^38,39,40) Table 4 is the activity interaction coefficients of the components in the molten steel at 1873 K. The activity of elements in liquid steel can be calculated by Eq. (6), so as to calculate the actual activity product.

  

$$\mathrm{L}\mathrm{o}\mathrm{g}{f}_i={\sum }_{\text{j}=\text{1}}^{\text{n}}{e}_i^j[\text{mass\%}j]$$(5)

  

$$a_i=f_i[\text{mass\%}i]$$(6)

Table 4. First-order interaction coefficients ($e_i^j$) in molten steel at 1873 K.^38,39,40)

Element	C	Si	Mn	P	S	Al	Ti	N	O	Sn	La
${\text{e}}_{\text{La}}^j$	0.03	−0.351	0.28	–	−82	–	–	–	−4.98	–	0.0078
${\text{e}}_{\text{Sn}}^j$	0.37	0.057	–	0.036	−0.028	–	–	0.027	−0.11	0.0016	–

The liquidus and solidus temperatures of the Sn1-La and Sn5-La samples were calculated using the Equilib module in FactSage8.2 software. The liquidus and solidus temperatures of the Sn1-La sample are 1808.3 K and 1799.7 K, respectively. For the Sn5-La sample, the liquidus and solidus temperatures are 1805.1 K and 1784.8 K, respectively.

According to Eqs. (4), (5), (6), the concentration relationship of La and Sn for the formation of La₂Sn in molten steel at liquidus and solidus temperatures was calculated, as shown in Fig. 17(a). The relationship between the equilibrium activity product and the actual activity product with temperature was also calculated, as shown in Fig. 17(c). Similarly, the formation of La₅Sn₄ was calculated, as shown in Figs. 17(b) and 17(d). From Fig. 17, it can be observed that the actual activity products of La₂Sn and La₅Sn₄ are lower than their equilibrium activity products above the solidus temperature in sample with 0.1% Sn and 0.065% La, indicating they cannot form directly in the molten steel. However, with the addition of 0.5% Sn and 0.065% La to IF steel, La₂Sn and La₅Sn₄ can form directly in the molten steel.

Fig. 17. (a)–(b) The concentration relationship between [La] and [Sn] formed by La₂Sn and La₅Sn₄ in liquid steel; (c)–(d) the relationship between the equilibrium activity product and the actual activity product with temperature in Sn1-La sample. (Online version in color.)

4.2. The Formation of La–Sn Inclusions during Solidification Process

During the solidification process of molten steel, due to the redistribution of solutes between the solid and liquid phases, La and Sn elements in the liquid phase at the solidification front will continuously enrich. When they accumulate to a certain extent, the actual activity product of La₂Sn ${\text{(}{a}_{\text{[La]}}^{\text{2}}\cdot a_{\text{[Sn]}}\text{)}}_{a\text{c}}$ in the molten steel may exceed the equilibrium activity product required for the formation of La₂Sn ${\text{(}{a}_{\text{[La]}}^{\text{2}}\cdot a_{\text{[Sn]}}\text{)}}_{\text{eq}}$. Once ${\text{(}{a}_{\text{[La]}}^{\text{2}}\cdot a_{\text{[Sn]}}\text{)}}_{a\text{c}}$ exceeds the equilibrium activity product, La₂Sn inclusions will form at the solidification front. Qu et al.⁴¹⁾ found that the Ohnaka(4α) model was suitable for describing segregation phenomena during continuous casting. Therefore, to further study the formation of La–Sn compounds, the Ohnaka(4α) model was used to calculate the formation of La–Sn compounds in Sn1-La sample (0.1% Sn-0.065% La) during the solidification process. The micro-segregation equation of the Ohnaka(4α) model is shown as follows.⁴¹⁾

  

$${\text{C}}_{\text{L,}i}={\text{C}}_{\text{0,}i}{[\text{1}+f_s({\beta }_i{k}_i-\text{1})]}^{(\text{1}-k_i)/({\beta }_i{k}_i-\text{1})}$$(7)

  

$$\begin{array}{cc}{\beta }_i={\displaystyle \frac{\text{4}{\alpha }_i}{\text{1}+\text{4}{\alpha }_i}}& {\alpha }_i={\displaystyle \frac{{\text{2D}}_{\text{s,}i}{\text{t}}_f}{{\left({\displaystyle \frac{\lambda }{\text{2}}}\right)}^{\text{2}}}}\end{array}$$(8)

  

$$\begin{array}{cc}{\text{t}}_f=\frac{{\text{T}}_{\text{L}}-{\text{T}}_{\text{s}}}{{\text{R}}_{\text{c}}}& \lambda =\text{146}\times {\text{10}}^{-\text{6}}\times {\text{R}}_{\text{c}}^{-\text{0}\text{.39}}\end{array}$$(9)

where C_L,i is the concentration of solute element i in the liquid phase at the solidification front (wt%); C_0,i is the initial concentration of solute element i in the molten steel (wt%); f_s is the solidification fraction; k_i is the partition coefficients of solute element i, where k_La = 0.02 and k_Sn = 0.5;⁴²⁾ β_i is the inverse diffusion coefficient; α_i is the Fourier series of solute element i; D_s,i is the diffusion coefficient of solute element i in the solid phase(m²/s), where D_s,La = 1×10⁻²⁰,⁴²⁾ D_s,Sn = 2.24 × 10^−7.4;⁴³⁾ t_f is the local solidification time (s); λ is the secondary dendrite spacing (m), related to cooling rate R_c (K/s). The cooling rate was set as 0.1 K/s.

Using Eqs. (5), (6), (7), (8), (9), the actual activity product for the formation of La₂Sn ${\text{(}{a}_{\text{[La]}}^{\text{2}}\cdot a_{\text{[Sn]}}\text{)}}_{a\text{c}}$ during solidification was calculated. The equilibrium activity product for the formation of La₂Sn during solidification was calculated using Eq. (4). And the temperature T_L–S of the liquid phase at the solidification front was calculated using Eq. (10).

  

$${\text{T}}_{\text{L}-\text{S}}={\text{T}}_{\text{m}}-\frac{{\text{T}}_{\text{m}}-{\text{T}}_{\text{L}}}{\text{1}-f_{\text{s}}\frac{{\text{T}}_{\text{L}}-{\text{T}}_{\text{s}}}{{\text{T}}_{\text{m}}-{\text{T}}_{\text{s}}}}$$(10)

By substituting Eq. (10) into Eq. (4), the equilibrium activity product of La and Sn can be obtained. The calculation formula is shown as follows:

  

$$\mathrm{L}\mathrm{o}\mathrm{g}{\text{(}{a}_{\text{[La]}}^{\text{2}}\cdot a_{\text{[Sn]}}\text{)}}_{\text{eq}}=\frac{\text{30}\mathrm{\hspace{0.17em} }\text{310}\text{.84}}{{\text{T}}_{\text{m}}-\frac{{\text{T}}_{\text{m}}-{\text{T}}_{\text{L}}}{\text{1}-f_{\text{s}}\frac{{\text{T}}_{\text{L}}-{\text{T}}_{\text{s}}}{{\text{T}}_{\text{m}}-{\text{T}}_{\text{s}}}}}+\text{13}\text{.07}$$(11)

T_m, T_L and T_S are the melting point of pure iron (1809 K), liquidus temperature and solidus temperature (K), respectively. In the same way, the equilibrium activity product and actual activity product of La₅Sn₄ during solidification were calculated with the change of solidification fraction. The calculation results are present in Fig. 18. As can be seen from Fig. 18, the actual activity product in the liquid phase at the solidification front exceeds the equilibrium activity product when the solidification fraction is above 0.1147 for La₂Sn and 0.409 for La₅Sn₄, allowing their formation at the solidification front. However, due to the lack of relevant thermodynamic data, the formation mechanisms of La–Sn–O and La–Sn–Ti inclusions cannot be calculated at present.

Fig. 18. The relationship between the activity product of La₂Sn and La₅Sn₄ and the solid phase ratio during solidification in Sn1-La sample: (a) La₂Sn; (b) La₅Sn₄. (Online version in color.)

5. Conclusions

Experimental samples with different La and Sn content were smelted by a laboratory resistance furnace. The effect of Sn on IF steel and the formation of inclusions containing Sn were investigated by experiments and thermodynamic calculation. The following conclusions can be drawn:

(1) The main inclusions are TiN, Al₂O₃, MnS and Ti₂S₃ in IF steel. The addition of Sn does not change the types of inclusions. After adding 0.065%La in IF steel with 0.1% or 0.5% Sn, inclusions containing Sn appear, including La–Sn, La–Sn–O and La–Sn–Ti.

(2) With the increase of Sn content from 0% to 0.5% in IF steel, the initial temperature for the formation of the BCC phase decreases. The number density of inclusions increases from 24.51 mm⁻² to 87.25 mm⁻² and the average size of inclusions decreases from 2.39 μm to 1.89 μm.

(3) The number of inclusions sharply increases by adding La and Sn in IF steel. When Sn content increases from 0.1% to 0.5% in IF steel with 0.065% La, the number density of inclusions decreases from 185.47 mm⁻² to 151.99 mm⁻². And the proportion of inclusions containing Sn in large inclusions (≥10 μm) increases from 24.5% to 84.5%. The maximum size of inclusions containing Sn increases from 26 μm to 34 μm.

(4) Fine macrostructures were obtained by adding Sn and La. This may be due to an increase in the number of inclusions that could act as heterogeneous nucleation cores for δ-Fe.

(5) Through thermodynamic analysis, La₂Sn and La₅Sn₄ inclusions can form during the solidification process in Sn1-La sample with 0.1%Sn–0.065%La. In Sn5-La sample with 0.5%Sn–0.065%La, La₂Sn and La₅Sn₄ inclusions can form directly in the molten steel.

(6) The generation of inclusions containing Sn may decrease Sn content in the matrix, thereby suppressing Sn segregation to the grain boundaries. This study suggests that residual element Sn could be reduced by consuming Sn as inclusions and by floating removal of large inclusions containing Sn.

Statement for Conflict of Interest

The authors declare that they have no conflict of interest.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant Nos. 52104337, 52274339), the China Baowu Low Carbon Metallurgy Innovation Foundation (Grant No. BWLCF202317), and the Natural Science Foundation of Gansu Province (Grant No. 24JRRB001).

References

• 1)   R.  Zhu,  G. S.  Wei and  H. J.  Zhang: Iron and Steel, 57 (2022), 1 (in Chinese). https://doi.org/10.13228/j.boyuan.issn0449-749x.20220381
• 2)   Z. F.  Cui,  F. Q.  Shangguan,  F. J.  Wang,  J. C.  Zhou and  A. J.  Xu: Iron and Steel, 58 (2023), 126 (in Chinese). https://doi.org/10.13228/j.boyuan.issn0449-749x.20230006
• 3)   E. T.  Stephenson: Metall. Trans. A, 14 (1983), 343. https://doi.org/10.1007/bf02644212
• 4)   R. Y.  Chen and  W. Y. D.  Yuen: ISIJ Int., 45 (2005), 807. https://doi.org/10.2355/isijinternational.45.807
• 5)   Y.  Lan and  S.  Sridhar: Metall. Mater. Trans. B, 42 (2011), 1031. https://doi.org/10.1007/s11663-011-9528-z
• 6)   H. B.  Peng,  W. Q.  Chen,  L.  Chen and  D.  Guo: High Temp. Mater. Process., 34 (2015), 19. https://doi.org/10.1515/htmp-2014-0010
• 7)   I.  Kapoor,  C.  Davis and  Z. S.  Li: Ironmak. Steelmak., 48 (2021), 712. https://doi.org/10.1080/03019233.2021.1898869
• 8)   A.  Ramadan,  A. Y.  Shash,  I. S.  El-Mahallawi,  D.  Senk and  T.  Mattard: J. Iron Steel Res. Int., 22 (2015), 582. https://doi.org/10.1016/s1006-706x(15)30043-1
• 9)   W. T.  Nachtrab and  Y. T.  Chou: J. Mater. Sci., 19 (1984), 2136. https://doi.org/10.1007/BF01058089
• 10)   H.  Matsuoka,  K.  Osawa,  M.  Ono and  M.  Ohmura: ISIJ Int., 37 (1997), 255. https://doi.org/10.2355/isijinternational.37.255
• 11)   X.  Su,  R. J.  Brown and  R. W.  Tyson: Steel Res. Int., 77 (2006), 825, https://doi.org/10.1002/srin.200606468
• 12)   S. C.  Chen,  R.  Zhu,  L. Q.  Xue,  T. C.  Lin,  J. S.  Li and  Y.  Lin: Int. J. Miner. Metall. Mater., 22 (2015), 141. https://doi.org/10.1007/s12613-015-1054-6
• 13)   M.  Godec,  M.  Jenko,  H. J.  Grabke and  R.  Mast: ISIJ Int., 39 (1999), 742. https://doi.org/10.2355/isijinternational.39.742
• 14)   W. L.  Lin,  X. T.  Deng,  Q.  Wang,  Y. J.  Zhou and  Z. D.  Wang: Iron and Steel, 58 (2023), 113 (in Chinese). https://doi.org/10.13228/j.boyuan.issn0449-749x.20230024
• 15)   X. J.  Zhang,  F.  Gao and  Z. Y.  Liu: J. Iron Steel Res. Int., 23 (2016), 1044. https://doi.org/10.1016/s1006-706x(16)30156-x
• 16)   Z.  Zhong,  W. Y.  Lv,  T. Y.  Liu,  Y.  Xie,  Y.  Yang,  Z. Z.  Wang,  D. K.  Li,  X. B.  Li and  T. Q.  Wu: J. Iron Steel Res. Int., 31 (2024), 998. https://doi.org/10.1007/s42243-023-01122-0
• 17)   X.  Zhang,  G. J.  Ma,  M. K.  Liu and  Z.  Li: Metals, 9 (2019), 1. https://doi.org/10.3390/met9080834
• 18)   A. I.  Zaitsev,  N. E.  Zaitseva,  E. K.  Shakhpazov and  B. M.  Mogutnov: ISIJ Int., 44 (2004), 957. https://doi.org/10.2355/isijinternational.44.957
• 19)   T.  Kékesi,  T. I.  Török and  G.  Kabelik: Hydrometallurgy, 55 (2000), 213. https://doi.org/10.1016/S0304-386X(99)00091-2
• 20)   H. P.  Wang,  J.  Kang and  Y.  Wang: J. Mater. Res. Technol., 18 (2022), 2922. https://doi.org/10.1016/j.jmrt.2022.03.171
• 21)   C.  Liu,  J.  Zhang,  Q.  Ren and  L.F.  Zhang: J. Iron Steel Res. Int., 30 (2023), 2471. https://doi.org/10.1007/s42243-023-01022-3
• 22)   Q. C.  Zhao,  H.  Luo,  Z. M.  Pan,  X. F.  Wang and  H. X.  Cheng: Mater. Today Commun., 34 (2023), 9. https://doi.org/10.1016/j.mtcomm.2023.105329
• 23)   S. H.  Song and  S. F.  Sun: Steel Res. Int., 87 (2016), 1435. https://doi.org/10.1002/srin.201500425
• 24)   H. B.  Peng,  W. Q.  Chen,  L.  Chen and  D.  Guo: High Temp. Mater. Process., 33 (2014), 179. https://doi.org/10.1515/htmp-2013-0028
• 25)   M. P.  Seah: Acta Metall., 28 (1980), 955. https://doi.org/10.1016/0001-6160(80)90112-1
• 26)   H. Y.  Wang,  X. Y.  Gao,  H. P.  Ren,  H. W.  Zhang and  H. J.  Tan: Acta Phys. Sin., 63 (2014), 369. https://doi.org/10.7498/aps.63.148101
• 27)   C. I.  Garcia,  G. A.  Ratz,  M. G.  Burke and  A. J.  Deardo: J. Met., 37 (1985), 22. https://doi.org/10.1007/bf03258635
• 28)   X.  Jiang and  S. H.  Song: Mater. Sci. Eng. A, 613 (2014), 171. https://doi.org/10.1016/j.msea.2014.06.094
• 29)   S. H.  Song,  Y. W.  Xu,  X. M.  Chen and  X.  Jiang: J. Rare Earths, 34 (2016), 1062. https://doi.org/10.1016/s1002-0721(16)60135-7
• 30)   W. B.  Xin,  B.  Song,  M. M.  Song and  G. Y.  Song: Steel Res. Int., 86 (2015), 1430. https://doi.org/10.1002/srin.201400332
• 31)   W. B.  Xin,  J.  Zhang,  G. P.  Luo,  R. F.  Wang,  Q. Y.  Meng and  B.  Song: Metall. Res. Technol., 115 (2018), 8. https://doi.org/10.1051/metal/2018030
• 32)   A. X.  Sha,  F. M.  Wang,  J. R.  Wu,  L. R.  Dong,  S. Y.  Li,  Y. X.  Chen and  L. X.  Zhu: Rare Met., 24 (2000), 287 (in Chinese). https://doi.org/10.13373/j.cnki.cjrm.2000.04.012
• 33)   X. Z.  Zhou,  D. Y.  Wang,  X. L.  Li,  T. P.  Qu,  D.  Hou,  S.  Hu,  Z. X.  Zhang and  H. H.  Wang: Mater. Charact., 205 (2023), 14. https://doi.org/10.1016/j.matchar.2023.113250
• 34)   X. Z.  Zhou,  D. Y.  Wang,  J.  Tian,  T. P.  Qu,  Z. X.  Zhang,  Y. P.  Wang,  X. L.  Li and  D.  Hou: Mater. Charact., 208 (2024), 23. https://doi.org/10.1016/j.matchar.2024.113659
• 35)   J. B.  Hu,  W. B.  Jiang,  C. C.  Huang,  L.  Qi,  Z. G.  Wang and  Z.  Ge: Mater. Today Commun., 38 (2024), 10. https://doi.org/10.1016/j.mtcomm.2024.108077
• 36)   S. Y.  Chu,  Y.  Gan,  S. T.  Qiu,  L.  Xiang,  Y. S.  Tian and  B.  Liu: China Metall., 32 (2022), 1 (in Chinese). https://doi.org/10.13228/j.boyuan.issn1006-9356.20210820
• 37)   B.  Fu: MA thesis, University of Science and Technology Wuhan, (2010), https://doi.org/10.7666/d.y1739629, (accessed 2010-04-25), (in Chinese).
• 38)   Q.  Ren,  Z. Y.  Hu,  Y. X.  Liu,  W. C.  Zhang,  Z. Q.  Gao and  L. F.  Zhang: J. Iron Steel Res. Int., (2024), 12. https://doi.org/10.1007/s42243-023-01135-9
• 39)   J. X.  Chen: Handbook of Common Charts and Data for Steelmaking, Beijing, Metall. Industry Press, (2010), 1. ISBN978-7-5024-4774-8.
• 40)   L. M.  Wang,  T.  Du,  X. L.  Lu,  Z. B.  Li and  Y. C.  Gai: J. Chinese Society of Rare Earths, 03 (2003), 251 (in Chinese). https://doi.org/10.3321/j.issn:1000-4343.2003.03.001
• 41)   T. P.  Qu,  J.  Tian,  K. L.  Chen,  Z.  Xu and  D. Y.  Wang: Ironmak. Steelmak., 46 (2019), 353. https://doi.org/10.1080/03019233.2017.1396722
• 42)   W. B.  Xin: Ph.D. thesis, University of Science and Technology Beijing, (2016), https://xueshu.baidu.com/usercenter/paper/show?paperid=d40f7de9befdd66c6ca7d281ab9fd00c, (accessed 2015-12-31), (in Chinese).
• 43)   X.  Zhang: Ph.D. thesis, Wuhan University of Science and Technology, (2021), https://doi.org/10.27380/d.cnki.gwkju.2021.000028, (accessed 2021-03-01), (in Chinese).