Characterization of Metal Droplets in Slag after Desulfurization of Hot Metal

Abstract

Iron losses in slag during an intensive desulfurization of hot metal can reach 0.6–1.1% from the total amount of processed hot metal. The characteristics of different metal droplets (such as morphology, number, size, composition and solidification structure) in industrial slag samples after desulfurization were investigated using SEM. All metal droplets in the slag were classified into three groups according to morphology: Type A - spherical/oval; Type B - spherical/irregular; Type C - irregular. Thereafter, some mechanisms for involving of different metal droplets into the slag during desulfurization process were studied based on obtained characteristics of metal droplets. Moreover, a possibility to remove those metal droplets from the liquid slag was estimated based on Stokes law. In addition, the effect of some parameters (such as slag viscosity and size of different metal droplets in this slag) on the possibility to reduce the iron losses during desulfurization of hot metal was considered.

1. Introduction

It is well known that the intensive desulfurization of hot metal (HM) or steel in the ladle by injection of reagents with inert gas results in a dispersion of metal droplets into the slag.^1,2,3) This entrapment of the metal droplets by the slag can significantly increase the iron losses and, as a result, the cost of steel products at steel plants. Therefore, one important direction for increasing the steelmaking effectiveness and to decreasing of cost of steel products is a reduction of iron losses in slag during desulfurization of hot metal and steel.

The total iron losses during desulfurization of hot metal can be considered as a summary of iron losses from different sources:¹⁾ 1) iron droplets, which were entrapped into the slag during an intensive injection of reagents and stirring of hot metal directly during the desulfurization processes, 2) a part of the iron melt, which is entrapped into the slag during skimming of slag from a ladle after desulfurization and 3) deposition of some amount of hot metal on the ladle walls during tapping from the ladle. However, it is difficult to estimate the separate contribution of hot metal remaining on the ladle walls after tapping (Source 3) on the total iron losses during desulfurization. Therefore, reliable data of its evaluation are missing in the literature.

Shevchenko et al.⁴⁾ summarized the iron losses in the skimmed slag after desulfurization of hot metal by injection of Mg-granules in the absence of additives (named “Desmag–Ukraina” process in some metallurgical companies of China) and injection of Mg + CaO mixture (the ESM process in metallurgical company OAO “Severstal” in Russia). For instance, for a reduction of the sulfur content in the iron from 0.020% to 0.002%, the iron losses during desulfurization and slag skimming operations (Sources 1 and 2), were on average 0.76 and 1.02% from the total amount of hot metal processed by the above mentioned methods. Herewith, the iron loss from the slag skimming operations after desulfurization (Source 2) was determined as 0.15 and 0.19% for these processes. Thus, the iron losses in slag during desulfurization of hot metal (Source 1) were on average 0.61 and 0.83%, respectively.

Magnelöv et al.⁵⁾ estimated the total iron losses and quotas of magnetic fractions having different sizes in the skimmed slag after desulfurization of hot metal by injection of CaC₂ + 2 wt% cryolite (Campaign 1) and CaC₂ + 5 wt% nepheline syenite (Campaign 2). The magnetic fractions in the skimmed slags after desulfurizations (Sources 1 and 2) in Campaign 1 (average reduction of S content in the iron was from 0.026% to 0.0051%) and Campaign 2 (average reduction of S content in the iron was from 0.044% to 0.0037%) were 2.5 and 1.9% of the total processed hot metal, respectively. The magnetic fractions of large size pieces (from 5 to > 300 mm) in the slags after skimming in Campaigns 1 and 2 reached values of 1.8 and 0.8% of the desulfurized hot metal. It can be assumed that most of these large iron pieces (> 5 mm) correspond to additional iron losses from the slag skimming operations after desulfurization (Source 2). It depends on the slag characteristics (such as the amount of slag, fluidity or viscosity of slag etc.) and the techniques used for slag skimming (such as the method, equipment etc.). The amount of fine magnetic fraction (< 5 mm) in the slag after desulfurization (small size metal droplets dispersed within the slag - Source 1) can reach values of 0.7–1.1%. The amount of iron losses with those fine metal droplets in the slag depends on the parameters of desulfurization (intensity of stirring, injected materials, time of desulfurization etc.) and condition of slag. It should be pointed out that a recovering of small iron droplets (< 5 mm) from the slag by a magnetic separation is difficult because it needs a fine grinding of the solidified slag.

Overall, the amount of iron losses in a slag only during desulfurization operations (Source 1) can reach from 30–60%⁵⁾ to 80%⁴⁾ (6–11 kg/t HM) from the summary iron losses during desulfurization of hot metal depending on the used technology. Therefore, the purposes of this study were: a) to characterize the metal droplets in the size range smaller than 5 mm in the slag just after desulfurization of hot metal, b) to consider the different mechanisms of involving of different metal droplets into the slag during the desulfurization process, c) to discuss the effects of some technological parameters of desulfurization on a possibility to decrease the amounts of different metal droplets in the slag. Therefore, some characteristics of different metal droplets in slags (such as morphology, number, size and composition) were determined and compared in industrial trials just after desulfurization of hot metal.

2. Experimental

2.1. Sample Preparation

In this study, desulfurization of hot metal from the blast furnace was carried out according to the normal praxis at the steel plant, but in two different ways: in a torpedo-car (trials TC1 and TC2) and in a transport ladle (trial TL1), as schematically shown in Fig. 1. The injection of reagents for desulfurization is made through a ceramic lance by using nitrogen as a carrier gas. The main technological parameters of desulfurization are listed in Table 1. Injection time and consumption of desulfurization reagents were varied depending on the initial sulfur content, the mass and temperature of hot metal in a ladle, and on the required final sulfur content. After desulphurization in trials TC1 and TC2, the hot metal was poured from a torpedo-car into a transport ladle. Then, the desulfurized hot metal in all trials after a skimming of formed slag in a transport ladle was transported to the BOF.

Fig. 1.

Schematic illustration of the desulfurization processes of hot metal in the industrial trials.

Table 1. Main parameters for desulfurization of hot metal during the industrial trials.

Trial No*	Weight of hot metal (ton)	Reagents	Consumption of reagent (kg/ton HM)	Temperature (°C)	Injection time (min)	Sinit/Sfinal (%)
TC1	296	95% CaC2 + 5%C	5.6	1424	34	0.074/0.005
TC2	261	95% CaC2 + 5%C	7.0	1381	37	0.077/0.003
TL1	140	CaC2	7.0	1347	17	0.040/0.001

*:  TC – samples taken from torpedo car, TL – samples taken from transport ladle.

Samples of slag and hot metal were taken from the torpedo-car or ladle immediately after desulphurization of hot metal and lifting of an injection lance. The compositions of slag and hot metal were analyzed according to the common practice in a steelmaking plant.

2.2. Investigations of Metal Droplets and Non-metallic Inclusions

Characteristics (such as morphology, size, number and composition) of metal droplets in slag samples were investigated on a polished cross section of slag samples by using a Scanning Electron Microscopy (SEM) equipped with energy-dispersive spectroscopy (EDS). The SEM investigations were done at a magnification of ×100. All metal droplets having size larger than 3 μm were measured at this magnification. Thereafter the obtained SEM images were analyzed by using an image analyzer for determination of the morphology, size and number of different metal droplets. The total observed area for each sample was on average about 9 mm². The number of investigated metal droplets in each slag sample was varied from 80 to 1420 depending on the trial and type of metal droplets.

All metal droplets in slag samples were classified depending on their morphology based on a circularity factor, CF. The value of CF for each metal droplet on SEM images was determined by the image analyzer as follows:   

$$CF=4·\pi ·A/P^2$$(1)

where A and P are the area and perimeter of investigated metal droplet, respectively.

An equivalent size of metal droplets, D_(MD), was determined as a diameter of circle having the same area with an investigated metal droplet on a SEM image by using an image analyze software (= $\sqrt{4A/\pi }$ ). The number per unit area, N_A(MD), and area fraction, f_A(MD), of different metal droplets on a polished surface of slag sample were calculated as follows:   

$$N_{\mathrm{A}\left(\mathrm{M}\mathrm{D}\right)}=n_{(\mathrm{M}\mathrm{D})}/A_{\mathrm{o}\mathrm{b}\mathrm{s}}$$(2)

  

$$f_{\mathrm{A}\left(\mathrm{M}\mathrm{D}\right)}=N_{\mathrm{A}(\mathrm{M}\mathrm{D})}·A_{(\mathrm{M}\mathrm{D})}$$(3)

where the n_(MD) is the number of metal droplets for corresponding type. A_obs is the total area of sample observed by SEM. A_(MD) is the average area for corresponding type of the metal droplets.

In this study, solidification structure of different metal droplets and non-metallic inclusions in the metal droplets and in the slag were investigated by a SEM on surfaces of the slag samples after electrolytic etching. Specimens from different slag samples after polishing were used for electrolytic etching with a 10%AA electrolyte (10 v/v% acetylacetone - 1 w/v% tetramethylammonium chloride-methanol). In this process, the metal matrix of droplets in the slag specimens can be dissolved in an electrolyte. However, the non-metallic inclusions (such as oxides, nitrides, carbides and complex multicomponent inclusions), which are more stable, do not dissolve during the electrolytic process. Therefore, after electrolytic etching the size, morphology and composition of non-metallic inclusions can be determined more accurately.

Compositions of different types of metal droplets (15–20 measurements for each type) were determined by point analyses using SEM in combination with EDS.

3. Results and Discussion

3.1. Metal Droplets in the Slag

Typical SEM images of different metal droplets in the slags from experimental trials are shown in Fig. 2. It can be seen that both slag samples contained metal droplets with various morphologies and sizes. In this study it was assumed that the morphology of different metal droplets is determined by the conditions of formation and solidification of respective metal droplets in the slags. Therefore, all metal droplets in the slag samples were classified into three groups according to the morphology, as shown in Table 2: spherical/oval (Type A), spherical/irregular (Type B) and irregular (Type C).

Fig. 2.

Typical SEM images of different metal droplets in the slags after desulfurization of hot metal.

Table 2. Classification of the different metal droplets in the slag samples.

This classification was done based on a value of the circularity factor, CF. A typical relationship between the values of CF and equivalent size, D_(MD), of different metal droplets in slag samples is shown in Fig. 3. Despite the considerable scatter of the data, it is obvious that the values of CF tend to decrease with an increased equivalent size of metal droplets in the slag samples. According to obtained results, it was concluded that most of metal droplets for Type A, B and C correspond to CF values in the ranges 0.8≤ CF ≤1.0, 0.6≤ CF < 0.8 and CF < 0.6, respectively. The average values and error bars of the standard deviations of CF and D_(MD) for metal droplets having different morphology are shown in Fig. 3(b).

Fig. 3.

Typical relationship between the circularity factor, CF, and equivalent size, D_(MD), of different metal droplets in slag samples.

Figure 4 shows the number of different metal droplets per unit area (a) and their frequencies (b) in slag samples depending on the droplet morphology and time of reagent injection during desulfurization in various industrial trials. It is apparent that the number of Type A droplets is significantly larger compared to those of Type B and C droplets. The frequency of Type A droplets reached 66–84% of the total number of metal droplets in the slags while those values for Types B and C droplets were only 11–21% and 5–13%, respectively. Moreover, it should be pointed out that the number of metal droplets in the slag tends to increase with an increased time of reagent injection during desulfurization of hot metal. This is particularly evident for the Type A metal droplets.

Fig. 4.

Numbers of different metal droplets per unit area, N_A(MD), and their frequencies in slag samples taken from different trials.

It is believed that an area fraction of metal droplets on a surface of slag sample, f_A(MD), correlates to a volume fraction of hot metal in this slag. In this study, the f_A(MD) values were used to estimate the contribution of each type of metal droplets to the total iron losses during desulfurization of hot metal (Source 1). According to the obtained results shown in Fig. 5, it can be concluded that the area fractions of irregular Type C droplets are considerably larger than those for metal droplets of Types B and A in all industrial trials. For instance, the frequency of f_A(MD) for Type C droplets in the slag samples of all industrial trials reached 53–65% of the total area fraction of metal droplets. However, the values for Types B and A droplets were only 27–33% and 5–20%, respectively. Overall, a clear dependence of the f_A(MD) values for different metal droplets on the injection time was not found in this study. It may be explained by the larger effect of other technological parameters in the different industrial trials (such as the intensity of reagent injection into the hot metal, condition of surface slag, etc.). Moreover, the mechanisms of entrapment of different metal drops into the slag and their behaviours during the intensive desulfurization processes can be different.

Fig. 5.

Area fractions of different metal droplets, f_A(MD), and their frequencies in slag samples taken from different trials.

It was found that the size of different metal droplets varies in the wide range. Furthermore, that they increased drastically from Type A (3–250 μm) to Type B (3–550 μm) and to Type C (3–890 μm). The typical size distributions and cumulative area fractions for different metal droplets in slag samples from the TC1 trial are shown in Fig. 6. It can be seen that most metal droplets having sizes smaller than 100 μm are the droplets of Types A (about 86%) and B (about 10%). Though the amount of all small droplets in the slag, N_{A(MD<100 μm)}, reached a value of 98% of the total N_A(MD) value, the area fraction of these droplets was less than 25% of the total f_A(MD) value. However, it can be seen (Fig. 6(b)) that the cumulative area fraction of metal droplets drastically increased in the size range of droplets larger than 100 μm. Although the amount of these metal droplets, N_{A(MD>100 μm)}, is relatively small (only 2% of the total amount of droplets in the slag), the f_{A(MD>100 μm)} value reached almost 75%. Moreover, most of these large sized metal droplets corresponded to the Type C (42–49%) and Type B (40–43%) droplets. Therefore, an investigation of the formation and behavior of these large sized metal droplets have a large importance with respect to the reduction of iron losses during desulfurization of hot metal.

Fig. 6.

Typical size distributions (a) and area fractions (b) of different metal droplets in slag samples of the TC1 trial.

It may be assumed that the metal droplets that have various morphologies also have different compositions. Therefore, the compositions of the different types of metal droplets in the slag samples were evaluated in this study by using a SEM equipped with an EDS. The contents of the main elements (such as C, Si, Mn, Ti and V) in different metal droplets in slag (MD) and hot metal (HM) from various trials are listed in Table 3. Despite the considerable scatter of the results for composition analysis, it was found that the carbon content in the metal droplets of different droplet types is significantly lower compared to that in the hot metal. This fact can be explained by the precipitations of graphite and/or Fe₃C phases during solidification of the metal droplets. Furthermore, carbon that dissolves in the hot metal droplets can reduce some amount of Mn and Si from the MnO and SiO₂ of the liquid slag. These concentrations in the slag after desulfurization were 0.3–0.4% and 2.1–5.5%, respectively. The concentrations of the main components in slag samples of industrial trials before and after desulfurization are given in Table 4. This fact is confirmed by the higher contents of Mn and Si in most metal droplets in comparison to the hot metal values. It is obvious that the longer the hot metal droplets will be in contact with the molten slag, the larger the amount of Mn and Si can be reduced from the oxides of slag. So, the Mn and Si concentrations in most metal droplets considerably increase from Type C to Types B and A droplets. The contents of Ti and V also tend to increase on average from Type C to Types B and A in most of the metal droplets. Although, to a much lesser extent than those for Mn and Si. It may also be explained by the possibility for a partial reduction of Ti and V from the liquid slag (0.6–2.5% TiO₂ and 0.3–0.4% V₂O₃) by carbon from the liquid metal droplets. Here, it should be noted that the precipitation of different non-metallic inclusions (such as nitrides, carbides and carbo-nitrides of Ti and V) and their characteristics in various metal droplets will be considered and discussed in a separate article. Thus, based on the obtained results from the composition analysis of metal droplets, it is believed that the Type A metal droplets had remained longer in the liquid state in the slag melt before their solidification in comparison to the Type C droplets.

Table 3. Composition of different metal droplets (MD) in slag and hot metal (HM) from samples various trials.

Trial	Type of MD	Content of main elements (wt%)*
		C	Mn	V	Si	Ti
TC1	A	2.87 ± 1.53	1.17 ± 0.28	0.58 ± 0.50	0.47 ± 0.33	0.12 ± 0.10
	B	2.62 ± 2.52	0.99 ± 0.48	0.31 ± 0.28	0.56 ± 0.51	0.12 ± 0.10
	C	1.77 ± 0.46	0.66 ± 0.34	0.14 ± 0.08	0.50 ± 0.29	0.06 ± 0.06
	HM	4.40	0.32	0.31	0.66	0.12
TC2	A	2.60 ± 1.49	1.74 ± 0.68	0.19 ± 0.17	1.39 ± 0.72	0.15 ± 015
	B	2.74 ± 1.52	1.72 ± 0.73	0.30 ± 0.41	1.15 ± 0.51	0.27 ± 0.54
	C	2.30 ± 1.33	1.05 ± 0.80	0.23 ± 0.14	1.17 ± 0.75	0.12 ± 015
	HM	4.50	0.32	0.31	0.68	0.16
TL1	A	3.78 ± 0.78	3.36 ± 2.04	0.57 ± 0.57	2.43 ± 1.67	0.14 ± 0.23
	B	3.73 ± 0.46	3.11 ± 1.34	0.80 ± 0.58	2.36 ± 1.14	0.15 ± 0.19
	C	4.07 ± 1.09	1.74 ± 1.50	0.29 ± 0.16	1.11 ± 0.98	0.06 ± 0.07
	HM	4.73	0.37	0.31	0.40	0.12

*:  average value ± standard deviation for metal droplets.

Table 4. Contents of main components in slag samples of industrial trials (wt%).

Trial	Slag	CaO	SiO2	Al2O3	TiO2	MnO	S	C	FeO + Fe*
	before	36.1	35.8	2.9	10.4	2.2	0.6	0.3	11.7
TC1	after	22.7	5.5	0.5	2.0	0.3	2.4	7.1	59.5
TC2	after	19.9	5.2	1.4	2.5	0.4	1.4	6.5	62.7
TL1	after	25.3	2.1	0.6	0.6	0.4	1.4	8.7	60.9

*:  contents of FeO and metallic Fe in iron droplets in the slag samples are determined together.

3.2. Formation and Behavior of Different Metal Droplets in Slag during Desulfurization

Different forms of metal droplets in the slag were caused by various conditions of formation and solidification of these droplets in the slag.

It may safely be suggested that the spherical or oval forms of Type A droplets indicates testified that these metal droplets were liquid and dispersed into the liquid slag. Therefore, they had a possibility to keep the spherical (or oval) form with a minimum surface area during solidification. The liquid Type A droplets were dispersed into the liquid part of slag together with nitrogen gas bubbles during the intensive desulfurization, as shown in Fig. 7 (Mechanism 1).³⁾ According to this mechanism, small size metal droplets were formed when the gas bubbles passes through the interface between the liquid slag and the hot metal and the metal film around the gas bubbles was ruptured (Figs. 7(a)–7(c)). Large size metal droplets were formed by the jet formation after a detachment of the gas bubble from the surface of the hot metal (Fig. 7(d)). All these metal droplets formed by this mechanism had a round shape. A liquid state of the Type A metal droplets in the slag was also confirmed by the directed solidification structures of these droplets, as will be discussed below.

Fig. 7.

Schematic illustration of entrainment of hot metal droplets into slag due to rising gas bubbles according to mechanism considered in [3].

The Type B droplets having a spherical/irregular form represent an intermediate group between the spherical (Type A) and irregular (Type C) metal droplets. In this study, it was assumed that some amount of the Type B droplets was involved into the slag in the same manner as the Type A droplets (the Mechanism 1). Another part of the Type B droplets in the slag was formed similarly to the irregularly shaped metal droplets.

The irregular shape of Type C metal droplets can be explained by the formation of these droplets during the splashing in an open eye zone of hot metal due to an intensive injection of reagents during desulfurization (Mechanism 2). Dispersions of hot metal dropped on an upper surface of the formed slag, as shown in Fig. 8. These droplets cooled rapidly during their flight. Moreover, when they dropped on a surface of a solid slag, these metal droplets solidified very fast. In this case, these metal droplets could not to maintain a spherical shape. The fact of the fast solidification of these irregular metal droplets was confirmed by an undirected solidification structure of these droplets.

Fig. 8.

Schematic illustration of splashes in an open eye zone of hot metal.

Different solidification rates of metal droplets in the slag had various morphology, which was confirmed by investigations of droplet solidification structures. Figure 9 shows the revealed microstructures of metal droplets of Type A and Type C on the surface of the slags after electrolytic etching. It is apparent that the metal droplets of Type A have a dendritic structure, which directed from the surface to the center. This can be explained by the direction of solidification of the liquid metal droplets in the slag. However, the metal droplets of Type C have an undirected solidification structure. This may be explained by a faster solidification (“quenching”) of these liquid metal droplets. It is also confirmed by the irregular form of the Type C droplets. The metal droplets of Type B have a microstructure which is a mixture of the ones found in the droplets of Type A and C. Similar microstructures of metal droplets were observed in slag samples from all trials.

Fig. 9.

Typical microstructure of metal droplets found after electrolytic extraction: (a) Type A and (b) Type C.

It should be pointed out that the metal droplets (such as Type A and B) in a liquid slag have a possibility to return to the hot metal bath with time due to the higher density of metal droplets in comparison to the slag. Therefore, some conditions of the formed slag during desulfurization of hot metal (such as low solidification temperature and viscosity, high fluidity, etc.) are very important to minimize the iron losses with metal droplets of Types A and B in the slag. For instance, an increased fluidity of slag can significantly increase the possibility for metal droplets to return to the hot metal bath. As a result, the number of Type A and B droplets in the slags during and after the desulfurization process can be decreased, as will be discussed in another paper. Moreover, an increased amount of liquid phase in the slag should also decrease the possibility for a formation of irregular Type C droplets. On the other hand, the increased fluidity of slag can significantly increase the area of an open eye zone of the hot metal during the intensive desulfurization. In this case, the number and size of Type C droplets can significantly be increased.

An effect of different parameters on possible iron losses in the slag during desulfurization can be evaluated by considering the balance between droplets being involved inside, f_V(MD)^IN, and droplets being removed outside, f_V(MD)^OUT. Removal of different metal droplets from the unstirred slag melt can be estimated based on Stokes law. The lowering velocity of liquid metal droplets in a calm slag melt (having Reynolds number lower than unit), v_(MD), can be evaluated by using the following equation:   

$$v_{\left(\mathrm{M}\mathrm{D}\right)}=\frac{2g}{9}\cdot \frac{p_{\left(\mathrm{M}\mathrm{D}\right)}-p_{\left(\mathrm{s}\mathrm{l}\mathrm{a}\mathrm{g}\right)}}{{\eta }_{\left(\mathrm{s}\mathrm{l}\mathrm{a}\mathrm{g}\right)}}\cdot {\left(\frac{D_{\left(\mathrm{M}\mathrm{D}\right)}}{2}\right)}^2$$(4)

where g is the gravitational acceleration (= 9.81 m/s²). p_(MD) and p_(slag) are the densities of the hot metal droplets (~7150 kg/m³) and slag melt (~3200 kg/m³). η_(slag) is the dynamic viscosity of slag melt.

The value of η_(slag) for ladle slag during desulfurization of hot metal can be varied in the wide ranges depending on the slag composition, temperature and injected reagents. For instance, the viscosity of industrial slags from a blast furnace production and similar laboratory slags in the temperature range from 1350 to 1450°C varies from 0.3 to more than 15.0 depending on the slag basicity and contents of such components as FeO, CaF₂ and other.⁶⁾ Moreover, reliable η_(slag) values for similar slags, which were obtained in given trials during desulfurization of hot metal, are missing in the literature. Therefore, the values of v_(MD) for different types of metal droplets in this study, were estimated at the various values of slag viscosity, η_(slag), as a function of the equivalent size of metal droplets, as shown in Fig. 10. The time for removal of different metal droplets, τ_(MD), from the liquid slag (Fig. 10(b)) was calculated for a thickness of slag layer in the ladle about 100 mm. This value is close to the data from industrial practice of desulfurization of hot metal.

Fig. 10.

Calculated velocity, v_(MD), and time for removal of different metal droplets from the liquid slag depending on the equivalent size of metal droplets, D_(MD), and slag viscosity.

In this study, a possibility for removal of different metal droplets from liquid slags was considered based on Stokes law for two groups of slags having different viscosity: η_(slag) ≤ 1 Pa·s for Slag I and 1 < η_(slag) ≤ 15 Pa·s for Slag II. Though the Slag I and Slag II are hypothetical slags, it was assumed in this study that some part of industrial slags is liquid during desulfurization. Furthermore, that they have the viscosity values in the ranges given for Slag I or Slag II. The presence of spherical droplets (Type A) in the slag samples taken after desulfurization confirms that some part of slag is liquid during desulfurization. Moreover, the presence of TiN inclusions in the slag samples, which can form in the hot metal and which only can be introduced into the liquid slag by pop-up nitrogen gas bubbles during desulfurization, also confirms the presence of some part of liquid slag. It can be seen that the v_(MD) values for droplets smaller than 100 μm is too low (≤ 0.07 mm/s) in both slag groups. As a result, the removal time of these metal droplets from both slags will be too long (> 23 min for Slag I and > 77 min for Slag II). Therefore, it can be expected that most of metal droplets smaller than 100 μm (containing most number of Type A and Type B droplets) cannot be removed from liquid slag during desulfurization according to Stokes law. In this case, these metal droplets will be accumulated in the slag and increase the iron losses during desulfurization of hot metal (f_V(MD)^IN >> f_V(MD)^OUT).

The v_(MD) values for droplets, having the size from 100 to 300 μm, varies from 0.03 to 0.65 mm/s in Slag I. In this case, some part of these droplets can come back from the slag to the hot metal bath because the removal time for most of these droplets is 3–30 min. However, these droplets cannot be removed practically from the slag with a higher viscosity (Slag II) due to the significantly larger removal time (> 30 min). The metal droplets larger than 300 μm can be removed almost completely from the liquid Slag I (v_(MD)~0.2–7.2 mm/s and τ_(MD)~0.2–9.0 min) and partially from the Slag II (v_(MD)~0.01–2.2 mm/s and τ_(MD)~0.8–130 min). Thus, the metal droplets of Types A and B having the size larger than 100 μm, have a relatively high possibility to be removed from the liquid slag during desulfurization. The same sized irregular droplets of Type C cannot practically be removed from the slag due to their fast solidification, as was shown above. As a result, the Type C droplets will also be accumulated in the slag during desulfurization (f_V(C)^IN >> f_V(C)^OUT). This will increase the total iron losses.

Overall, it should be noted that the metal droplets can easily be removed from slag layers having a thickness significantly smaller than 100 mm. Moreover, an intensive stirring during desulfurization can significantly change the considered relationship between metal droplets involved into the slag and removed from this slag.

4. Conclusions

Some mechanisms for involving of different metal droplets into the slag are discussed based on characterization of metal droplets (morphology, number, size and composition) in industrial slag samples after desulfurization of hot metal. Moreover, a possibility to remove those metal droplets from the liquid slag is considered based on Stokes law. The most important results can be summarized as follows:

(1) All metal droplets in the slag samples can be classified into three groups according to the morphology: Type A – a spherical/oval shape (size range 3–250 μm), Type B – a spherical/irregular shape (3–550 μm) and Type C – an irregular shape (3–890 μm).

(2) Though the number of metal droplets increases from Type A (66–84% from total number of metal droplets) to Type B (11–21%) and Type C (5–13%), the area fraction, f_A(MD), of these metal droplets corresponds to 5–20%, 27–33% and 53–65% from the total area fraction of all metal droplets, respectively.

(3) Small sized droplets (≤ 100 μm) consisted mostly of Type A (84–86%) and Type B (9–11%) droplets. These droplets corresponded to 98% of the total number and less than 25% of the total area fraction of all metal droplets in the slag samples. Large sized droplets (> 100 μm) consisted mostly of Type C (42–49%) and Type B (40–43%) droplets. These droplets corresponded to only 2% of the total number and almost 75% of the total area fraction of all metal droplets in the slag samples.

(4) Based on the consideration of morphologies, compositions and solidification structures of different metal droplets, it was assumed that most of metal droplets of Type A and B were involved into the slag melt during passing of gas bubbles through the interface between the hot metal and liquid slag and rupture of the metal film around the gas bubbles (Mechanism 1). Most of the irregular Type C metal droplets were formed by the splashing of hot metal during an intensive injection of reagents for desulfurization (Mechanism 2).

(5) A decreased slag viscosity and an increased size of metal droplets can significantly accelerate a removal of Type A and B metal droplets from a liquid slag. However, large sized irregular Type C droplets cannot practically be removed from the slag due to their fast solidification. Therefore, they will be accumulated in the slag during the whole desulfurization period.

Acknowledgments

The authors are grateful for financial support from the Swedish Energy Agency (Energimyndigheten), Swedish Steel Producers association (Jernkontoret) and Hugo Carlssons Foundation.

References

• 1)   C.  Brodrick: Proc. 7th Int. Heavy Minerals Conf. on ‘What next’, The Southern African Institute of Mining and Metallurgy, Johannesburg, South Africa, (2009), 141.
• 2)   V. F.  Chevrier and  A. W.  Cramb: Metall. Mater. Trans., 31B (2000), 537.
• 3)   Z.  Han and  L.  Holappa: ISIJ Int., 43 (2003), No. 3, 292.
• 4)   A. F.  Shevchenko,  L. P.  Kurilova,  V. G.  Kislyakov,  A. M.  Bashmakov and  A. L.  Rudenko: Fundamental and Application Problems of Iron and Steel Industry, Issue 17 (2008), 116.
• 5)   M.  Magnelöv,  J.  Eriksson,  J.  Drugge,  J.  Björkvall and  B.  Björkman: Ironmaking Steelmaking, 40 (2013), No. 6, 436.
• 6)  Slag Atlas, 2nd Ed., VDEh, Germany, (1995), 387, ISBN 3-514-00457-9.