Effect of Na₂O and Rb₂O on Inclusion Removal in C96V Saw Wire Steels Using Low-Basicity LF (Ladle Furnace) Refining Slags

Abstract

Inclusion removal and modification of C96V saw wire steel using Na₂O- and Rb₂O- containing novel low-basicity LF (ladle furnace) Refining Slags have been researched. The results indicated that the addition of Na₂O deteriorates inclusion removal; by contrast, the addition of Rb₂O seems to significantly enhance inclusion removal. In detail, Rb₂O can improve the cleanliness in the as-quenched C96V saw wire steel melts compared to preexisting synthetic LF refining slag compositions: (i) The average inclusion diameter experienced a remarkable decrease after reaction between the liquid steel and the synthetic LF refining slag; (ii) In addition, the number of inclusions also suffered from a dramatic decrease, with the reaction time increasing from 900 to 2700 s (15 to 45 min); (iii) Furthermore, both of the MnO-SiO₂-Al₂O₃ and CaO-SiO₂-Al₂O₃ inclusion system mainly concentrated in the low melting zone when the composition of Rb₂O in synthetic refining slag was ≥5.0 wt%. This is mainly because Na₂O significantly reduces the viscosity of refining slag, while Rb₂O increases it. Then, there are two remarkable influences causing the increase of viscosity of refining slag with the addition of Rb₂O: the inclusions can be sufficiently entrained within the slag once absorbed due to the significant increase in the viscosity; and the slag entrapment during refining process weakened dramatically.

1. Introduction

In order to obtain a high productivity, higher tensile properties and smaller diameters (60 × 10^‒6 m (60 μm) to 80 × 10^‒6 m (80 μm)) than tire cord are very for saw wire [1]. Therefore, much a higher requirement for inclusion control should be put forward to prevent wire breakage caused by large size or hard inclusions which often initiate breakage during cord drawing and standing processes [2,3,4]. To this end, a lot of technology has been developed by metallurgical workers, such as Si-Mn deoxidation [5], boron addition to modify inclusions [6], and low basicity top slag refining [7] and Na₂CO₃ addition to modify inclusions [8].

On the other hand, there are still some problems that plagued metallurgical workers for many years, but which the subjects of new research and breakthroughs in recent years [9,10,11,12,13]. For example, in terms of the formation mechanism and evolution of SiO₂-type inclusions in Si-Mn killed steel wires, Wang et al. [9,10] found, in 2015, that the formation mechanism of SiO₂-type inclusions in wire rods involves two main steps: First, solid SiO₂ phase precipitates form MnO-SiO₂-Al₂O₃ system inclusions during the casting process to form dual-phase inclusions in the as-cast bloom, and second, dual-phase inclusions experience phase separation during the rolling process for heterogeneity and great compression ratios [9,10]. However, some of their results are not very persuasive. For example, they only analysed the inclusions in tundish and cast-bloom, but not in wire rod. As we know, dual-phase inclusions may experience phase separation during the drawing process for heterogeneity and great compression ratio. Thus, the formation mechanism and evolution of SiO₂-type inclusions in Si-Mn killed steel wires should be studied further. With regard to the formation mechanism of CaO-SiO₂-Al₂O₃-(MgO) inclusions, there were some controversies until Wang et al. [11,12] did some research in 2017 and came up with a very persuasive view. Wang et al. pointed out that it can be reasonably concluded that CaO-SiO₂- type inclusions in saw wire were exogenous particles form entrapped/emulsified top slag, but not products of slag-steel-inclusion chemical reactions [11,12]. However, this argument has not been confirmed by others or recognized by fellow researchers.

For the factors influencing oxide inclusion, deformability under high and low temperatures had not been noticed until Zhang et al. [13] did some research in 2018. They came up with a theory that the deformability of oxide inclusion under different temperatures could be scaled by Young’s moduli. However, some of the results are not very persuasive. For example, the theory of the influence of temperature on the deformability of oxide inclusions has not been confirmed by experiments.

Few researchers have looked at changing the chemical composition of LF refining (ladle furnace) slag with new components to improve inclusions absorption. The influence of alkali oxides enhancing inclusion removal had not been noticed until Sohn et al. [14,15] did some research at 2014. The effect of alkali oxides (Li₂O, Na₂O, K₂O, Rb₂O, Cs₂O) containing tundish fluxes on inclusion removal have been studied; the results indicated that K₂O, Rb₂O, and Cs₂O addition seems to significantly enhance inclusions removal, but the influence of Li₂O and Na₂O has the opposite effect. In detail, not only the number of inclusions, but also the average inclusions diameter, increased with minimal improvement with Li₂O and Na₂O addition. Sohn et. al came up with an explanation that K⁺, Rb⁺ and Cs⁺ have a large cationic radius compared to Li⁺ and Na⁺; therefore, the inclusions can be sufficiently entrained within the slag once absorbed due to the significant increase in the viscosity. However, they did not give any more detailed explanation about how the increase in viscosity of slag can improve the absorption ability of inclusions.

On the other hand, the alkali oxides have not been applied in LF refining progress until now. The influence of alkali oxides on the number, size, morphology, and composition (especially for MnO-SiO₂-Al₂O₃ and CaO-SiO₂-Al₂O₃ inclusion systems) is still not completely understood.

Therefore, we have explored the influence of alkali oxides on enhancing inclusion removal by adding 2.5 wt%, 5.0 wt%, and 7.5 wt% Na₂O and Rb₂O into synthetic LF refining slag. The influence of alkali oxides (Na₂O, Rb₂O) on the number, average diameter, composition, and morphology of inclusions has been studied. Furthermore, we came up with a more detailed explanation about how the increase in viscosity of slag can improve the absorption ability of inclusions.

2. Experiment

2.1. Experimental Appratus and Procedure

With respect to simulating LF refining progress, the MoSi₂ furnace was utilized to fabricate the C96V saw wire steels in this experiment. A schematic diagram of the experimental equipment is shown in Figure 1.

In this experiment, the temperature of liquid metal is continuously measured by means of a B-type reference thermocouple. An argon atmosphere was kept in experiments all the time, blowing from the bottom to the top of the furnace tube. The experimental procedures were carried as follows. Firstly, 1.00 kg industrial pure iron is placed into a MgO crucible of 60 × 10^‒3 m (60 mm) inner diameter and 80 × 10^‒3 m (80 mm) depth. Secondly, the crucible is placed in a graphite crucible to prevent liquid metal from leaking. Finally, after the whole crucible is placed in the chamber, the power is switched on and the furnace is heated to the experimental temperature [1873 K (1600 °C)].

Alloys were added into the melts when the temperature reached 1873 K (1600 °C). Sampling No. 0 is the original chemical composition in steel without interaction of slag and metal. After that, 0.05 kg synthetic LF refining slag was put onto the surface of the molten metal by corundum funnel (Shandong Hao Yang wear resistant material Co., Ltd., Zibo, China). The height of solid slag and liquid slag are about 2.0 × 10^‒2 m (2.0 cm) and 1.5 × 10^‒2 m (1.5 cm); it completely covered the surface of the metal. Samples No.1 to No. 3 are taken from the molten metal after the slag melted for 15, 30, and 45 min. After each sampling, the steel liquid is stirred with a graphite rod for 2 min to make the molten steel and the refining slag uniform. All of the samples are taken by quartz tube sampler and quenched immediately in water.

In total, there are 7 heats made by treating with different synthetic LF refining slags, to which were added 2.5 wt%, 5.0 wt%, and 7.5 wt% Na₂O and Rb₂O, as shown in

Table 1. Chemical composition of synthesized refining slag.

Heat Number	CaO (wt%)	SiO2 (wt%)	Al2O3 (wt%)	Na2O (wt%)	Rb2O (wt%)	Time/min	CaO/SiO2
0#	42.22	52.78	5.00	-	-	0, 15, 30, 45	0.8
1#	40.11	50.14	4.75	2.50	-	0, 15, 30, 45	0.8
2#	38.00	47.50	4.50	5.00	-	0, 15, 30, 45	0.8
3#	35.89	44.86	4.25	7.50	-	0, 15, 30, 45	0.8
4#	40.11	50.14	4.75	-	2.50	0, 15, 30, 45	0.8
5#	38.00	47.50	4.50	-	5.00	0, 15, 30, 45	0.8
6#	35.89	44.86	4.25	-	7.50	0, 15, 30, 45	0.8

. Samples were taken at 0, 15, 30, and 45 min during the refining progress.

2.2. Analysis Methods

A direct reading spectrometer was utilized to detect the composition of Si, Mn, P, Cr, V, Ni, and Cu. Steel samples were also sent to Analysis and Testing Center (Chemical Laboratory, Shenyang, China) of Northeastern University to analyse Al content by the ICP (Inductively Coupled Plasma) method. For C and S, an infrared C/S analyzer was applied. Then, the LECO^® TC 500 O₂/N₂ analyzer (LECO, San Jose, MI, USA) was chosen to detect O and N. The chemical composition of C96V saw wire steel is shown in

Table 2. Chemical composition of C96V saw wire steels (mass %).

Heat Number	C	Si	Cr	Mn	V	[Al]s	T.O	N	P	S	Ni	Cu
0#	0.96	0.16	0.21	0.34	0.10	≤0.0003	0.0012	0.0027	0.0067	0.0026	0.0027	0.0036
1#	0.96	0.15	0.21	0.36	0.10	≤0.0003	0.0013	0.0033	0.0070	0.0032	0.0028	0.0041
2#	0.97	0.15	0.20	0.35	0.11	≤0.0003	0.0012	0.0031	0.0071	0.0031	0.0031	0.0039
3#	0.95	0.17	0.23	0.35	0.11	≤0.0003	0.0014	0.0033	0.0072	0.0034	0.0029	0.0041
4#	0.96	0.15	0.20	0.34	0.12	≤0.0003	0.0015	0.0030	0.0070	0.0030	0.0022	0.0038
5#	0.96	0.15	0.21	0.36	0.11	≤0.0003	0.0012	0.0028	0.0068	0.0029	0.0035	0.0036
6#	0.97	0.16	0.20	0.37	0.10	≤0.0003	0.0013	0.0029	0.0061	0.0032	0.0033	0.0033

.

All of these steel samples were treated by 100–2000 mesh sand papers and polished. After that, photos (more than 50 pieces) around “S” route were taken. (On the surface of the sample, photos were taken in order from right to left, followed by from left to right). The Image J software (Image J 1.48, National Institutes of Health, Bethesda, MD, USA, 2014) was used for statistics of the size and count of inclusions; this software is a public domain, Java-based image processing program developed at the National Institutes of Health (NIH).

Finally, the metal samples were analysed by scanning electron microscope (SEM; Carl Zeiss AG, Niedersachsen, Germany) and Energy Dispersion Spectrum (EDS; Carl Zeiss AG, Niedersachsen, Germany) to determine the morphology and component of inclusions.

3. Results

3.1. Effect of Na₂O and Rb₂O on the Variation in Inclusion Diameter

Figure 2 shows the effect of reaction time on the average non-metallic inclusion diameter with the addition of Na₂O or Rb₂O. For normal heat, it is obvious that the count descends slowly with refining time being raised from 15 to 45 min.

The addition of Na₂O in Figure 2 shows that the inclusion diameter with the addition of Na₂O gets larger than that at normal heat at all sampling times; what is worse, increments in Na₂O seem to increase the inclusion diameter observed in steel samples, suggesting that Na₂O is not at all a suitable candidate for enhancing inclusion removal in C96V saw wire steels.

Contrary to Na₂O addition, Rb₂O addition seems to yield a dramatic trend of decreased average inclusion diameter with increasing reaction time. (i) Under the same refining time, the average diameter of inclusions was reduced sharply with the content of Rb₂O raised from 2.5% to 7.5%; (ii) On the other hand, the average diameter of inclusions decreased rapidly with the refining time ascending from 15 to 45 min when the weight percentage of Rb₂O fixed. (iii) Furthermore, the average diameter of inclusions in steel samples which were treated by Rb₂O (4^#, 5^#, 6^#) was smaller than that of normal heat (0^#).

3.2. Effect of Na₂O and Rb₂O on Inclusion Distribution Overlayed on the Phase Diagram

Figure 3 describes the distribution of inclusions in the MnO-SiO₂-Al₂O₃ and CaO-SiO₂-Al₂O₃ phase diagram in normal heat. It is obvious that the contents of SiO₂ and Al₂O₃ in inclusions increased gradually. In detail, MnO-SiO₂-Al₂O₃ inclusion mainly concentrate in the area of 70–90 wt% SiO₂, <18 wt% Al₂O₃, CaO-SiO₂-Al₂O₃ inclusion mainly concentrate in the area of 40–60 wt% SiO₂, 10–25 wt% Al₂O₃ after the steels are refined with 2700 s (45 min). In addition, some CaO-SiO₂-Al₂O₃ inclusions have a small amount of mass fraction of CaO and Al₂O₃ below 10 wt%.

The inclusion distribution overlaid in phase diagram with different refining time in 2^# (5.0 wt% Na₂O) steel samples is shown in Figure 4. Obviously, with the 15 min reactions shown in Figure 4a, the distribution of inclusions seems to be dominant toward the SiO₂-rich region. With longer reaction times, the distribution of inclusions are spread across the ternary phase diagram with the addition of Na₂O.

Contrary to the effect of Na₂O, the Rb₂O on the average inclusion diameter generally has a decreasing trend with longer reaction times. Figure 5 describes the distribution of inclusions with Rb₂O content of 5.0 wt%. As the reaction time increased, there is a strong tendency for inclusions to be distributed toward the less SiO₂ and Al₂O₃ region. In detail, the MnO-SiO₂-Al₂O₃ inclusion system mainly concentrate in the low melting area with 40–50 wt% SiO₂, <20 wt% Al₂O₃, and the CaO-SiO₂-Al₂O₃ inclusion system mainly concentrate in the low melting area with 40–50 wt% SiO₂, <20 wt% Al₂O₃ at the same time. For instance, the addition of Rb₂O seems to concentrate the distribution of inclusions toward the less SiO₂, Al₂O₃ phase, which is clearly shown in Figure 5.

In summary, at longer reaction times, the distribution of inclusions is spread across the ternary phase diagram, and dominant more toward the SiO₂ and Al₂O₃ region, suggesting Na₂O addition and an increase in Na₂O would likely be detrimental to control inclusions systems concentrating in the low melting area. In contrast, Rb₂O addition would likely be helpful to control both MnO-SiO₂-Al₂O₃ and CaO-SiO₂-Al₂O₃ inclusion system concentrating in low melting area. This is also shown in Figure 6.

Figure 6 shows the average content of each component in inclusions during refining. Obviously, the content of SiO₂ and Al₂O₃ in inclusions decreases sharply when the steel samples are treated by Rb₂O, as shown in Figure 6c. Contrary to the effect of Rb₂O, the effect Na₂O on the content of SiO₂ and Al₂O₃ in inclusions yielded generally an increasing trend with longer reaction times, as shown in Figure 6b.

Figure 7 compares the change in the inclusion chemistry within the MnO-SiO₂-Al₂O₃, CaO-SiO₂-Al₂O₃ ternary phase diagrams after 2700 s (45 min) with varying amounts of Rb₂O additions from 2.5 wt% to 7.5 wt%. Additionally, he average content of each component in inclusions during refining process is shown in Figure 8.

It may be noted that the MnO-SiO₂-Al₂O₃ and CaO-SiO₂-Al₂O₃ inclusions system were distributed toward the less SiO₂ and Al₂O₃ phase with Rb₂O addition increased from 0 wt% to 7.5 wt%.

3.3. Influence of Na₂O and Rb₂O on the Number of Inclusions

Figure 9 shows the influence of reaction time on the number of inclusions with the addition of Na₂O or Rb₂O. For normal heat, it can be noted that the number of inclusions reduced gently when the reaction time increased from 900 to 2700 s (15 to 45 min). For heats with Na₂O additions from 2.5 wt% to 7.5 wt%, obviously, the number of inclusions exceeds normal level for all the concentrations of Na₂O. However, unlike Na₂O, reactions of the C96V saw wire steel samples with Rb₂O containing synthesized refining slag samples resulted in a significant decrease in the number of inclusions observed with both longer reaction times and higher concentrations of Rb₂O, as observed in Figure 9.

3.4. Morphology and Element Distribution of Typical Inclusions

Morphological observations of typical inclusions by SEM-EDS are shown in Figure 10, Figure 11 and Figure 12. The majority of inclusions in all of the experimental heats are SiO₂-MnO-Al₂O₃ and SiO₂-CaO-Al₂O₃ systems, the differences about inclusion components, sizes and numbers have been discussed above.

It is obvious that the size of inclusions in normal heat is around 5 μm. as shown in Figure 10. However, this increased sharply after the steel samples were treated by Na₂O; some are large than 20 μm, as shown in Figure 11. In contrast, inclusions became very small, with diameters of about only 1 μm after steel samples were treated with Rb₂O, as shown in Figure 12.

3.5. Element Distribution in Complex Inclusion

As mentioned, complex inclusions were discovered in all of the experimental steel samples. In order to describe inclusion structures exactly, SEM mappings were done, as shown in Figure 13, Figure 14 and Figure 15. Obviously, both MnO-SiO₂-Al₂O₃ and CaO-SiO₂-Al₂O₃ inclusions are multi-layered composite structures in all experimental steel samples. In detail, there was a SiO₂, MnO, CaO, and Al₂O₃ homogeneous composite in the center, and a periphery of MnS precipitate around it.

4. Discussion

The content of alkali oxides in metallurgical slag is usually low, but its influence is significant. This comprehensive change in the slag composition has a dramatic impact on the thermochemical and thermophysical properties of the slag, including its density, surface tension [16], and viscosity [17]. In particular, surface tension and viscosity have a direct impact on melt/slag separation efficiency, melt/slag reaction kinetics, and the permeability of the intermediate gases used for reduction and heat transfer.

4.1. Influence of Na₂O, Rb₂O on the Structure and Viscosity of Metallurgical Slags

Studies that discuss the structure and properties of metallurgical slags have been reviewed by Waseda and Toguri [18,19]. It is well known that the increase of Li₂O, Na₂O content in SiO₂-CaO-Al₂O₃-R₂O (R = Li, Na) slag system could decrease the viscosity. Sukenaga et al. [20] studied the effect of Na₂O on the viscosity for CaO-SiO₂-Al₂O₃-Na₂O slag. The results indicated that the viscosity of this quaternary melts decreased sharply with increasing the additive content of Na₂O. Similar results are also noted in the literature [21,22]. However, the influence of K₂O, Rb₂O, and Cs₂O is the opposite; this is due to the fact that the structure of SiO₂-CaO-Al₂O₃-R₂O (R = Li, Na, K, Rb, Cs) slag would like to change owing to the effects of alkali oxides (Li₂O, Na₂O, K₂O, Rb₂O, Cs₂O) [18,19,20,21,22,23].

Based on the laws studied by past researchers, we calculated the viscosities for the slag compositions in our experiment by Factsage 7.2 (Bale, C.W.; Pelton, A.D.; Thompson, W.T.; Eriksson, G.; Hack, K.; Chartand, P.; Decterov, S.; Jung, I.H.; Melanson, J.; Petersen, S. Thermfact/CRCT (Montreal, QC, Canada), GTT-Technologies (Aachen, Germany), 2017) at 1873 K (1600 °C). In detail, the viscosities are 0.351 Pa·s (0^#, normal heat), 0.343 Pa·s (1^#, 2.5 wt% Na₂O), 0.318 Pa·s (2^#, 5.0 wt% Na₂O), 0.273 Pa·s (3^#, 7.5 wt% Na₂O). The calculated results are in good agreement with the expected law.

According to the literature [24,25], the structure of slag is very complex, including chains, rings, and three-dimensional network structures, as shown in Figure 16.

(i) The influence of Na₂O on the viscosity of SiO₂-CaO-Al₂O₃ slag can be understood when considering the change in the degree of ploymerization (DOP) of slag structure [26].

$${\mathrm{R}}_2\mathrm{O}=2{\mathrm{R}}^{+}+{\mathrm{O}}^{2-}$$

(1)

$${\left[\mathrm{S}{\mathrm{i}}_3{\mathrm{O}}_9\right]}^{6-}(\mathrm{r}\mathrm{i}\mathrm{n}\mathrm{g})+{\mathrm{O}}^{2-}={\left[\mathrm{S}{\mathrm{i}}_3{\mathrm{O}}_{10}\right]}^{8-}\left(\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{i}\mathrm{n}\right)$$

(2)

$${\left[\mathrm{S}{\mathrm{i}}_3{\mathrm{O}}_{10}\right]}^{8-}\left(\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{i}\mathrm{n}\right)+{\mathrm{O}}^{2-}={\left[\mathrm{S}{\mathrm{i}}_2{\mathrm{O}}_7\right]}^{6-}(\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r})+{\left[\mathrm{S}\mathrm{i}{\mathrm{O}}_4\right]}^{4-}\left(\mathrm{m}\mathrm{o}\mathrm{n}\mathrm{o}\mathrm{m}\mathrm{e}\mathrm{r}\right)$$

(3)

where R⁺ is the basic oxide cation such as Na⁺. Addition of Na₂O would provide O²⁻, which can modify the complex silicate structure into simple [Si₂O₇]⁶⁻ and [SiO₄]⁴⁻. In addition, according to Sohn [14], after treatment with Na₂O, the [AlO₄]⁵⁻-tetrahedral structure will be depolymerized. We have provided the schematic diagram for Equations (2) and (3) in Figure 17. A schematic diagram for how Na₂O can depolymerize the [AlO₄]⁵⁻-tetrahedral structure is shown in Figure 18.

Therefore, the viscosity of slag decreases sharply.

(ii) The influence of Rb₂O and Cs₂O on the structure and properties of metallurgical slag have not been studied enough. Even so, some researchers [15,18,19] speculated that Rb₂O and Cs₂O have the same effect as K₂O. Furthermore, Sohn et al. [14,15] did some experiments to research the effect of K₂O, Rb₂O and Cs₂O on inclusion removal based on the above inference. The results indicated that Rb₂O, Cs₂O have the same effect as K₂O. In detail, with the introduction of the large cationic radius of the K⁺, Rb⁺ and Cs⁺ compared to Na⁺ and Li⁺, the inclusions can be sufficiently entrained within the alumino-silicate-based slag once absorbed due to the significant increase in the viscosity.

It is well known that K₂O addition can catalyze the formation of [AlO₄]⁵⁻ tetrahedral units, which act as network formers in the slag structure [27,28,29], as expressed by Equation (4) [27]. A schematic diagram for Equation (4) is presented in Figure 19.

$$2\mathrm{A}{\mathrm{l}}^{3+}+7{\mathrm{O}}^{2-}+{\mathrm{R}}_2\mathrm{O}=2{\left[\mathrm{A}\mathrm{l}{\mathrm{O}}_4\right]}^{5-}+2{\mathrm{R}}^{+}$$

(4)

Furthermore, according to Sohn et al. [14,15], K₂O addition could polymerize the slag to form complex aluminates and alumino-silicate structures. In order to help us understand this process, we provided the schematic diagram of how K₂O, Rb₂O and Cs₂O polymerize the slag network structure (as shown in Figure 20). Thus, the viscosity of slag increases significantly with an increase in the content of Rb₂O.

The schematic diagram of how Na₂O and Rb₂O change the structure of refining slag is shown in Figure 21. For the sake of understanding, we simplified the structure of refining, showing only the [AlO₄]⁵⁻-tetrahedral structure, the Si-O-Al structure, and the Si-O- structure.

4.2. Influence of Na₂O, Rb₂O on the Ability of Slags to Absorb Inclusions

Most authors report that inclusion absorption by slag occurs in three stages [30,31,32,33]:

(i) Flotation in the bath—transport of the inclusion to the steel/slag interface.

(ii) Separation of liquid steel—movement of the inclusion to the interface, breaking the surface tension of steel.

(iii) Dissolution in slag—removal of the inclusion from the steel/slag interface for full incorporation into the slag.

An inclusion can only be considered that is eliminated from steel when it is completely dissolved in the slag. According to Valdez et al. [31] and Reis et al. [34,35], the viscosity of slag significantly influences the ability of slag to absorb inclusions. Stage (i) takes place in molten steel; thus, the effect of Na₂O and Rb₂O occurs mostly in stages (ii) and (iii).

4.2.1. The Slag Entrapment during Refining Process

The influence of alkali oxides (Li₂O, Na₂O, K₂O, Rb₂O, Cs₂O) on the slag entrapment is an important factor that shouldn’t be ignored. According to J. Strandh et al. [36], if the slag viscosity is too low, the risk of slag entrainment to the steel is promoted, resulting in more inclusions in the steel.

When Na₂O was added to the slag, the reaction products ([Si₃O₉]⁶⁻, [Si₃O₁₀]⁸⁻, [Si₂O₇]⁶⁻, [SiO₄]⁴⁻, [AlO₄]⁵⁻ and so on) from Equations (1)–(3) might be incorporated into the molten steel. Therefore, the number and average diameter of inclusion increased rapidly, and the content of SiO₂ and Al₂O₃ contained in inclusions increased at the same time.

In contrast, the effect of Rb₂O is the opposite. When Rb₂O was added into slag, the slag entrapment during refining process was significantly weakened. Thus, the number and average diameter of inclusions, and the content of SiO₂ and Al₂O₃ contained in inclusions, decrease sharply compared to normal heat; this is described in Figure 22.

4.2.2. Influence of Na₂O, Rb₂O on the Second Stage: Separation of Liquid Steel

In terms of the second stage, Bouris et al. [37] and Nakajima et al. [38] have already done some correlative research, and described models for predicting the separation times for spherical, rigid, and chemically inert particles. During this process, the inclusion’s motion is decided by a force balance between a capillary force ($F_{\sigma ,\mathrm{Z}}$), a buoyancy force ($F_{\mathrm{b}}$), a drag force ($F_{\mathrm{d}}$) and a fluid force ($F_{\mathrm{m}}$), which is visualized in Figure 23. The force, when the inclusion is in contact with both steel and slag, are:

$$F_{\sigma ,\mathrm{Z}}=\left(-2\pi R_{\mathrm{i}}+2\pi \mathrm{Z}\right){\sigma }_{\mathrm{M}\mathrm{S}}+2\pi R_{\mathrm{i}}{\sigma }_{\mathrm{S}\mathrm{i}}-2\pi R_{\mathrm{i}}{\sigma }_{\mathrm{M}\mathrm{i}}$$

(5)

$$F_{\mathrm{b}}=\frac{4}{3}\pi R_{\mathrm{i}}^3({\rho }_{\mathrm{s}}{\tilde{\Delta }}_{\mathrm{b}}-{\rho }_{\mathrm{i}})g$$

(6)

$$F_{\mathrm{d}}=6\pi R_{\mathrm{i}}{\eta }_{\mathrm{s}}{\tilde{B}}_2\frac{\mathrm{d}Z}{\mathrm{d}t}$$

(7)

$$F_{\mathrm{m}}=\frac{1}{2}\frac{4}{3}\pi R_{\mathrm{i}}^3{\rho }_{\mathrm{s}}{\tilde{\Delta }}_{\mathrm{b}}\frac{{\mathrm{d}}^{2}Z}{\mathrm{d}{t}^2}$$

(8)

$${\tilde{\Delta }}_{\mathrm{b}}=\frac{1}{4}\left(\frac{{\rho }_{\mathrm{M}}}{{\rho }_{\mathrm{s}}}-1\right){\left(\frac{Z}{R_{\mathrm{i}}}\right)}^3-\frac{4}{3}\left(\frac{{\rho }_{\mathrm{M}}}{{\rho }_{\mathrm{s}}}-1\right){\left(\frac{Z}{R_{\mathrm{i}}}\right)}^2+\frac{{\rho }_{\mathrm{M}}}{{\rho }_{\mathrm{S}}}$$

(9)

$${\tilde{B}}_2=1\mathrm{for}\frac{Z}{R_{\mathrm{i}}}\ge 1$$

$$\left(\frac{{\eta }_{\mathrm{M}}}{{\eta }_{\mathrm{S}}}-1\right){\left(\frac{Z}{R_{\mathrm{i}}}\right)}^2-2\left(\frac{{\eta }_{\mathrm{M}}}{{\eta }_{\mathrm{S}}}-1\right)\left(\frac{Z}{R_{\mathrm{i}}}\right)+\frac{{\eta }_{\mathrm{M}}}{{\eta }_{\mathrm{S}}},\mathrm{for}0\le \frac{Z}{R_{\mathrm{i}}}\le 1$$

(10)

where $R_{\mathrm{i}}$ is the inclusion’s radius, $g$ is acceleration due to gravity, ${\rho }_{\mathrm{s}}$, ${\rho }_{\mathrm{i}}$ and ${\rho }_{\mathrm{M}}$ are the density of slag, inclusion and the metal, respectively. Z, $\mathrm{d}Z/\mathrm{d}t$ and ${\mathrm{d}}^{2}Z/\mathrm{d}{t}^2$ is the inclusion’s position, speed and acceleration, ${\eta }_{\mathrm{M}}$, ${\eta }_{\mathrm{s}}$ are the viscosity of metal and slag, respectively. ${\sigma }_{\mathrm{M}\mathrm{S}}$, ${\sigma }_{\mathrm{S}\mathrm{i}}$ and ${\sigma }_{\mathrm{M}\mathrm{i}}$ are the interfacial tensions between metal and slag, and slag and inclusion, and metal and inclusion, respectively.

The equation of motion is given by:

$$F_{\mathrm{a}}+F_{\mathrm{m}}=F_{\mathrm{b}}-F_{\mathrm{d}}-F_{\sigma ,\mathrm{Z}}$$

(11)

where $F_{\mathrm{a}}$ is the force from the acceleration of the inclusions, given as:

$$F_{\mathrm{a}}=\frac{4}{3}\pi R_{\mathrm{i}}{\rho }_{\mathrm{i}}\frac{{\mathrm{d}}^{2}Z}{\mathrm{d}{t}^2}$$

(12)

Obviously, viscosity is an important factor on the stress of inclusions during the separation process, as shows in Equation (11). According to Valdez et al. [31], decreasing viscosity causes the inclusions easier enter into the slag. By contrast, that increasing viscosity causes the inclusions to require much more time to separate. Furthermore, in 2014, Yang [39] researched the influence of viscosity of slag on the separation time in depth, and observed the similar conclusions.

Therefore, $F_{\mathrm{a}}=F_{\mathrm{b}}-F_{\mathrm{d}}-F_{\sigma ,\mathrm{Z}}-F_{\mathrm{m}}$ decreased gradually with the viscosity of slag increasing caused by Rb₂O addition. In other words, Rb₂O addition will exacerbate the kinetic conditions of inclusion removal during the separation stage due to Rb₂O addition decrease the viscosity of slag. By contrast, the effect of Na₂O addition is opposite. Namely, $F_{\mathrm{a}}=F_{\mathrm{b}}-F_{\mathrm{d}}-F_{\sigma ,\mathrm{Z}}-F_{\mathrm{m}}$ increased gradually with the viscosity of slag decreasing caused by Rb₂O addition.

4.2.3. Influence of Na₂O, Rb₂O on the Third Stage: Dissolution in Slag

Strandh et al. [36] developed a mathematical model to study inclusion behavior at the interface, and came up three types (remain, oscillating and pass) of inclusion behavior at the interface depending on the inclusion size, the velocity of the inclusion, and the interfacial properties of the system. Furthermore, Yang [39] put forward a supplement that a forth type of inclusion behavior at the interface is “pass, and then return back”. We provided the schematic diagram of the four inclusion behaviors in Figure 24 in order to help the reader to understand these theories.

In Na₂O addition experiment heats, inclusions (especially for SiO₂- containing and Al₂O₃- containing inclusion) are difficult to dissolute into slag due to the depolymerization by Na₂O. In detail, inclusions (especially for SiO₂- containing and Al₂O₃- containing inclusion) will be depolymerized into smaller unites with more simple structure, but not combine with slag units. As a result, inclusions may return back into molten steel, which is visualized in Figure 24d. Thus, the number and SiO₂ and Al₂O₃ content of inclusions increased remarkably with increasing not only reaction time, but also the content of Na₂O.

In contrast, in Rb₂O addition experiment heats, it is easy for inclusions (especially for SiO₂- containing and Al₂O₃- containing inclusion) to dissolute into slag due to the polymerization by Rb₂O. In detail, inclusions (especially for SiO₂- containing and Al₂O₃- containing inclusion) will be polymerized with slag units into larger unites of more complex structure, and they will stay in the molten slag, which is visualized in Figure 24c. Thus, the number, average diameter, and SiO₂ and Al₂O₃ content of inclusions decreased significantly, not only with the reaction time but also the content of Rb₂O.

In summary, in order to understand the influence of Na₂O and Rb₂O on the ability of slag to absorb inclusions, its comprehensive effect on all of stages should be considered.

The schematic diagram of the dissolution model of inclusions in Na₂O- containing or Rb₂O- containing refining slag is shown in Figure 25. For the sake of understanding, we simplified the structure of refining and inclusions: only the [AlO₄]⁵⁻-tetrahedral structure, the Si-O-Al structure and Si-O- structure are shown.

(i) For Na₂O addition experiment heats:

The inclusions (especially for SiO₂- containing and Al₂O₃- containing inclusion) may return back molten steel due to the fact that it is difficult to dissolve in slag, although it could dissolve through the molten-slag interface more easily. Additionally, the reaction products ([Si₃O₉]⁶⁻, [Si₃O₁₀]⁸⁻, [Si₂O₇]⁶⁻, [SiO₄]^4‒, [AlO₄]⁵⁻ and so on) from Equations (1)–(3) might be dissolved into molten steel, as shown in Figure 25a.

Therefore, the number and average diameter of inclusion increased rapidly, and the content of SiO₂ and Al₂O₃ contained in inclusions increased at the same time.

(ii) For Rb₂O addition experiment heats:

When the quantity of Rb₂O addition was low, most of the inclusions could pass through the molten-slag interface. It is easy for inclusions (especially for SiO₂- containing and Al₂O₃- containing inclusion) to be polymerized with slag units into larger unites with more complex structure, and then remain in the molten slag. Additionally, slag entrapment during refining process will be significantly weakened. Therefore, the number and average diameter of inclusion decreased rapidly, and the content of SiO₂ and Al₂O₃ contained in inclusions decreased at the same time, as shown in Figure 25b.

5. Conclusions

The effect of Na₂O and Rb₂O containing synthetic LF refining slag on the absorption ability of inclusions for C96V saw wire steels has been studied. Kinetic studies using a MoSi₂ furnace at 1873 K (1600 °C) on the reaction of synthetic LF refining slag containing alkali oxides (Na₂O, Rb₂O) with C96V saw wire steels suggested minimal improvement with Na₂O additions would deteriorate inclusion removal. In contrast, Rb₂O additions seems to remarkably enhance inclusion removal for to two reasons: after Rb₂O additions (i) the inclusions can be sufficiently absorbed into the refining slag due to the significant increase in the viscosity; (ii) the slag entrapment during refining process will be weakened dramatically. In detail, with Rb₂O additions:

(1) Not only did the number of inclusions, but also the average inclusion diameter, decrease in the steel samples;

(2) On the other hand, both of the MnO-SiO₂-Al₂O₃ and CaO-SiO₂-Al₂O₃ inclusion systems mainly concentrated in the low melting zone when the composition of Rb₂O in synthetic refining slag was ≥5.0 wt%.

In summary, Rb₂O additions seem to significantly improve the cleanliness in the as-quenched C96V saw wire steel melts, compared with preexisting synthetic refining slag compositions.

6. Future Work

It should be noted that much work needs to be done in future:

(i) the influence of stirring on slag entrainment that supported by rod. Obviously, there are many factors that should be considered in order to explore the influence of stirring on slag entrainment, such as the time of stirring, the velocity of stirring, the shape of the crucible, the diameter of the crucible, the diameter of the rod, the temperature of molten slag. (Other factors may also have important effects on the process, we can not describe all of them at present.)

(ii) The viscosities of CaO-SiO₂-Al₂O₃-R₂O (R = Na or Rb) should be measured. We got the viscosities of slag after Na₂O addition by calculating with Factsage 7.2, but it must be measured in experiments, although the calculated results are in good agreement with the expected outcomes. Furthermore, we don’t have any value about the viscosities of slag after Rb₂O addition due to Factsage 7.2 lack of enough data base.

(iii) The influence of Na₂O, Rb₂O addition on the structure of slag (CaO-SiO₂-Al₂O₃) should be researched further, especially for Rb₂O addition. This is mainly because there is hardly any literature to have noticed this topic.