Dissolution Behavior of Lime into Steelmaking Slag

Abstract

Many studies have been conducted over the last 50 years or more to clarify the dissolution behavior of lime in slag. In this paper, the previous studies on enhancing the dissolution of lime are reviewed. The research subject is divided into industrial tests to verify the new technology, and laboratory tests to clarify the dissolution mechanism. To enhance the dissolution, the only feasible measure is to increase the interfacial area. For this purpose, various methods to increase surface area have been tried, including the decrease in lime size; powder blowing or injection; controlling the calcination conditions; and application of a pre-melted flux. Among them, powder blowing and the reuse of the refining slag are the most effective measures. For fundamental experiments, the immersion of a rotation rod in the slag, immersion of lime cube with a circular rotating rod, and the addition of lime to the slag under gas bubbling have been conducted. The mechanism and rate-controlling step of lime dissolution have become clear: the destruction of the dicalcium silicate layer is important to promote the dissolution of lime. Stirring-promoted fluid flow and the formation of CO₂ inside of the quick lime are proposed as mechanisms of the destruction.

1. Introduction

In steelmaking processes, lime is the most important flux used to make slag. Lime is made by the calcination of limestone by a rotary kiln or shaft furnace, and its size range is generally 7–35 mm in diameter. Its main component is CaO and, by the JIS (Japanese Industrial Standard), the CaO content has to be 93% or more and the residual CO₂ has to be lower than 2%. In English, lime for steelmaking is often called “quick lime”; the phrase comes from “living lime” in Old English.¹⁾ In Japanese, it is written as “生石灰,” in which the word of “living (生)” is added to the words of “lime (石灰).” Lime is not a stable material but is one that easily changes to its carbonate or hydroxide. This could be a reason why this material is named a “living” or “生” material.

In steelmaking slag, the CaO content is generally 45–65 mass%, but several percent of CaO called “free CaO” is also present. In the cross-section of industrial slag, many white or brown spots are observed, which correspond to the free CaO.^2,3) The definition of “free CaO” is the sum of the undissolved lime during the refining period and the precipitated CaO during its solidification. The undissolved lime deteriorates the efficiency of lime in the steelmaking reactions, and free CaO causes volume expansion and alkali elution during the reuse of slag in civil engineering applications.

The dissolution rate of lime in slag was frequently discussed after steelmaking changed from the open hearth to converter process for its short refining time. The dissolution of lime is sometimes called slag formation or “slagging,” and the ratio of dissolved CaO to the added CaO is termed the “slagging ratio.” In industry, the slagging ratio is determined by the ratio of actual basicity to designed basicity. The actual basicity is the ratio of CaO and SiO₂ contents in the slag, and designed basicity is the calculated ratio of CaO and SiO₂, obtained from the unit flux consumption and mass of oxidized silicon in the molten steel, respectively. However, contamination by tiny particles of free CaO in the sample taken for analysis and the dissolution of the slag adhered on the converter’s refractory lining mean that this value does not accurately describe the slagging ratio. It should be noted that the slagging ratio has no relation to the solid fraction of the steelmaking slag, as this is composed of coexisting solid and liquid states under saturation of the dicalcium silicate phase.

Despite the various studies on the dissolution behavior of lime, the slagging ratio is still far less than 100% and varies widely among the heats. The reasons for this technological stagnation can be considered as the following. (1) Compared with the scattering of temperature or FeO content, the scattering of the slagging ratio does not greatly affect the performance of the steelmaking reaction. (2) The economical merit of increasing the slagging ratio is limited to the decrease in the unit consumption of the undissolved lime.

Steelmaking is traditionally conducted using one vessel (the converter), but has recently been developed to include a hot metal dephosphorization process followed by a decarburization process. In most cases, these two processes are separately conducted using the converter as a reaction vessel. The treatment temperature of hot metal dephosphorization is much lower and its treatment time is shorter than those of traditional converter processes. These conditions are inadequate for the quick dissolution of lime. Previously, this process depended on the slagging effect of fluorspar, but the use of this flux was restricted by environmental regulations. Therefore, the improvement of slagging has again received much attention.

The research in this field is divided into industrial tests to verify the new technology, and laboratory tests to clarify the dissolution mechanism. Many studies have been conducted over the previous 50 years or more, especially in Japan. In this paper, the previous studies that sought to enhance the dissolution of lime are reviewed.

2. Measures to Improve the Dissolution Rate of Lime

The dissolution of lime in slag is categorized as a solid and liquid reaction, or the melting of a solid in liquid. To enhance this reaction, the only feasible measure is to increase the interfacial area. Strategies for this purpose have included decreasing the lime size; powder blowing or injection; controlling the calcination conditions; and application of a pre-melted flux.

2.1. Decrease of Lime Size

The most common measure is to use small-size lime. This is a traditional method and repeatedly studied by many researchers. Figure 1 shows the result presented in 1960 from Yawata Iron & Steel Co. Ltd.⁴⁾ In this research, the size range of the lime was changed from 30–50 mm to smaller than 25 mm. The size distribution was 34% 5–25 mm, 55% 5–0.25 mm, and 11% <0.25 mm. The decrease in phosphorus content was clearly observed, showing the improvement in dephosphorization. Recently, Nippon Steel & Sumitomo Metal showed the results of using small-size lime for converter-based hot metal dephosphorization.⁵⁾ In this case, a 7% decrease in the unit consumption of lime was observed. However, in general, the effect of a reduced lime size on slagging behavior is much smaller than that expected by the increase in the interfacial area.

Fig. 1.

Effect of lime size on the relation between carbon and phosphorus by ref. 4 used with permission.

2.2. Lime Powder Blowing

Lime powder blowing with oxygen via a top lance is known as the Linz-Donawitz-Arbed-Centre-National (LD-AC) method.⁶⁾ This old technology was firstly imported to Japan in 1964 by Nippon Kokan K.K. and their result is shown in Fig. 2.⁷⁾ This method was developed to treat the high-phosphorus content hot metal produced by the local iron ore in Europe. However, LD-AC was not a commonly used technology as the hot metal phosphorus content was decreased simply by changing the iron ore source. Recently, this method has received attention again for application to the hot metal dephosphorization process by converter (Fig. 3⁸⁾). After the restriction of fluorspar for slagging, many companies changed the reaction vessel for hot metal dephosphorization from a torpedo car to converter. The slag composition became a lower basicity with a higher T·Fe content, and this composition is not detrimental to the dissolution of lime. However, the added lime is not easily dissolved because of the lower temperature and shorter treatment time. Therefore, the old technology of LD-AC was revived and applied to hot metal dephosphorization by converter. Figure 4⁸⁾ shows the result presented by Sumitomo Metal in 2011. The application of LD-AC to hot metal dephosphorization is an effective measure to improve the slagging ratio. In addition, the blown lime powder directly reacts with FeO, formed at hot spots, and enhances the dephosphorization reaction.⁹⁾

Fig. 2.

Improvement of slagging ratio by using the LD-AC process by ref. 7 used with permission.

Fig. 3.

Schematic of LD-AC process by ref. 8 used with permission.

Fig. 4.

Improvement of slagging ratio by the application of LD-AC to the hot metal dephosphorization process using a converter by ref. 8 used with permission.

2.3. Calcination Conditions¹⁰⁾

The reactivity of lime is evaluated by various methods, i.e., the rates of neutralization or temperature increase when lime is immersed in water. The surface of the quick lime has many holes (Fig. 5¹⁰⁾) and the surface area is related to the pore density.¹¹⁾ The difference in reactivity is caused by the microscopic structure of porosity, which differs with calcination temperature, time, atmosphere, and the origin of the raw material. Figure 6¹²⁾ shows the influence of the lime surface area on desulfurization behavior, and reveals the optimum pore size for improving the reaction. However, the literature studying the influence of the calcination conditions on the steelmaking reaction is very limited.

Fig. 5.

Surface of quick lime with various calcining conditions. (a) 1173 K for 2 h, and (b) 1323 K for 2 h by ref. 10 used with permission.

Fig. 6.

Relation between the amount of sulfur removal from the steel (Δn_s) and the surface area of quick lime according to pore diameter range. (a) ≥5 mm, (b) ≥2 mm, and (c) ≥1 mm by ref. 10 used with permission.

2.4. Pre-melted Flux

Many patents have been published on using a pre-melted flux, and the production of calcium ferrite or calcium aluminate has been widely investigated. In Fig. 7, the result presented in 1971 by Nippon Steel is shown.¹³⁾ The composition of the flux was 64.4CaO-21.8Fe₂O₃-10.1Al₂O₃-2.3SiO₂ (by mass%), and the high dissolution rate and good dephosphorization performance was confirmed by laboratory experiments and industrial applications, respectively. In 2012, Kobe Steel used a pre-melted calcium ferrite-type flux with a manganese ore sinter to produce high-carbon steel with a low phosphorus content.¹⁴⁾ By this method, they clarified the improvement of dephosphorization by the increase in total Fe content in the slag. In general, the cost to produce these fluxes is higher than the economical merit obtained by the improvement of the steelmaking reaction. Therefore, the actual application of artificially made pre-melted fluxes is very limited.

Fig. 7.

The improvement of dephosphorization behavior in a converter by the use of a pre-melted flux by ref. 13 used with permission.

In this sense, reuse of slag is a more profitable method to produce a pre-melted flux. For example, as ladle furnace slag contains high percentages of CaO and Al₂O₃, Sumitomo Metal reused this slag for the hot metal dephosphorization by converter, as shown in Fig. 8.¹⁵⁾ In this figure, K_CaO is calculated by Eq. (1); this value under normal operation is named K_CaO-Base, and is compared with the results using LD-AC.   

$$K_{\text{CaO}}=\text{ln}\left({\left[\text{P}\right]}_{\text{I}}/{\left[\text{P}\right]}_{\text{f}}\right)/W_{\text{CaO}}$$(1)

where [P]_I and [P]_f are the phosphorus contents (mass%) before and after the treatment, respectively, and W_CaO is the unit consumption of lime (kg/t).

Fig. 8.

The increase in K_CaO/K_CaO-Base with Al₂O₃ content in the slag caused by the addition of ladle slag, and its comparison with the results of LD-AC by ref. 15 used with permission.

In addition, by the application of hot metal dephosphorization, the concentration of P₂O₅ in converter slag for decarburization becomes low. As this slag contains high percentages of CaO and FeO, the recycling of this slag for hot metal dephosphorization is common. Nippon Steel recycles the decarburization slag for hot metal dephosphorization using the MURC (multi-refining converter) process.¹⁶⁾ In this process, after the tapping of decarburized molten steel, the hot metal of the next heat is directly charged on the slag (hot-recycling). The improvement of the distribution ratio of phosphorus by hot-recycling is shown in Fig. 9.¹⁷⁾ In this figure, the observed distribution ratio is divided by the calculated value by Eq. (2).¹⁶⁾   

$$\begin{array}{l}\text{log}\left(\%\text{P}\right)/\left[\%\text{P}\right]=\\ 2.5\text{ log}(\%\text{T}\cdot \text{Fe})+0.0715\left\{\left(\%\text{CaO}\right)+0.25(\%\text{MgO})\right\}\\ +7\mathrm{\hspace{0.17em} }710.2/T-8.55+\left(105.1/T+0.0723\right)\mathrm{\hspace{0.17em} }\left[\%\text{C}\right]\end{array}$$(2)

where (%M) indicates the content of M in the slag, [%N] indicates the content of element N in the molten steel (mass%), and T is temperature (K).

Fig. 9.

Effect of the hot-recycling of decarburization slag on the distribution ratio of phosphorus by ref. 17 used with permission.

Another example is shown in Fig. 10,⁵⁾ but in this case the decarburization slag was cooled and crushed before reuse. Reuse of the refining slag is a very effective method to decrease the undissolved lime and to decrease the total mass of slag generated from the steelmaking process.

Fig. 10.

Improvement of dephosphorization in a converter using decarburization slag by ref. 5 used with permission.

3. Fundamental Studies on the Dissolution of Lime

Many studies have conducted by dynamic and static experiments in the laboratory. In a dynamic study, the dissolution rate of lime is measured under the movement of the slag. In a static study, lime is immersed into a static slag and the interface between the lime and slag is observed.

In 1976, Y. Kawai and his group measured the dissolution rate of sintered CaO in slag using a rotating rod test (Fig. 11).^18,19) By this method, the dense sintered CaO rod was immersed into the slag and the diameter of the rod after rotation was measured with time. An example of the result is shown in Fig. 12. The following dimensionless equation was proposed, assuming that the rate-controlling step was mass transfer in the slag, using the Stanton number (St = k/u), Reynolds number (Re = d·u/ν), and Schmidt number (Sc = ν/D) (Fig. 13).   

$$J=St\times S{c}^{2/3}=0.384\times R{e}^{-0.31}$$(3)

where k is the mass transfer coefficient of CaO in the slag (m/s), u is relative velocity (m/s), d is the characteristic length (m), ν is kinematic viscosity (m²/s), and D is the diffusion coefficient of CaO in the slag (m²/s).

Fig. 11.

Experimental setup to measure the dissolution rate by ref. 18 used with permission.

Fig. 12.

Typical experimental result obtained by the rotating rod method for 40CaO-40SiO₂-20Al₂O₃ (mass%) slag (Slag A) by ref. 18 used with permission.

Fig. 13.

Relation between J-factor and Reynolds number for 40CaO-40SiO₂-20Al₂O₃ (mass%) slag (Slag A) and 40CaO-40SiO₂-20FeO (by mass%) slag (Slag D) by ref. 18 used with permission.

The formation of dicalcium silicate layer was reported by many researchers according to experiments under the static condition.^20,21,22) In the study of Kawai et al., the formation of a dicalcium silicate layer on the surface of the CaO rod was also observed, but they concluded the formation of this layer was not the rate-controlling step.

This literature first investigated the dissolution of lime by chemical engineering procedures, and Eq. (3) was useful to estimate the dissolution rate. However, as the decrease in diameter was less than 1.5 mm after rotation for 20 min, the measured dissolution rate was considered to be different to that of actual quick lime. By the same technique, Hamano et al.²³⁾ studied the influence of CaF₂, CaCl₂, and Al₂O₃ on the dissolution rate of lime, and Sasaki et al.²⁴⁾ measured the dissolution rate of decarburization slag.

On the other hand, D. Sichen and his group pointed out that the forced convection did not occur around the rod in the rotating rod method²⁵⁾ and proposed a different technique (Fig. 14).²⁶⁾ In this method, molten slag with a lime cube was stirred by a bar moving circularly. After stirring for a given time, the crucible containing the slag and lime cube was quenched and cut to measure the size of the residual cube. By this method, about 50% of the lime was dissolved in 6 min, and a strong dependence on the density (porosity) was observed, as shown in Fig. 15.²⁶⁾ In this experiment, the dissolution rate of the lime cube was evaluated by Eq. (4). (L₁ + L₂)_initial and L₁ + L₂ indicate the sum of the lengths of the cube (measured in two perpendicular directions of the cube cross section) before and after the experiment, respectively.   

$$L_N=\frac{L_1+L_2}{{(L_1+L_2)}_{initial}}$$(4)

Fig. 14.

Experimental method using a lime cube with a circularly moving bar by ref. 26 used with permission.

Fig. 15.

Dissolution behavior of cube lime with various densities. Type 1 = 1658 kg/m³, Type 2 = 2144 kg/m³, and Type 3 = 1774 kg/m³ by ref. 26 used with permission.

They also showed the detachment of dicalcium silicate from the lime surface and concluded that the mechanical separation of this layer had an important role in the dissolution. These results are more realistic to simulate the industrial operation.

Recently, we have measured the dissolution rate under the gas bubbling condition (Fig. 16).^27,28,29) In this method, lumpy lime was added to the slag which was stirred by gas bubbling; the dissolution rate was evaluated by the concentration change of CaO in the slag. This technique is the most realistic method to simulate the actual situation, and the industrial quick lime can be evaluated directly. By this method, the dissolution behavior of sintered CaO and industrial quick lime with and without preheating was studied. The residual CO₂ in the quick lime was completely eliminated by preheating. In addition, to clarify the effect of the interfacial layer, the slag of the FeO–CaO–B₂O₃ system was compared with the result of the FeO–CaO–SiO₂ system. In the case of sintered CaO and preheated quick lime, the interfacial layer was observed only in the case of the FeO–CaO–SiO₂ system. Also, the dissolution rate was much slower in this slag than that in the FeO–CaO–B₂O₃ system where no interfacial layer was formed. The dissolution rate of sintered CaO and preheated quick lime was almost the same, although the porosity density was largely different. On the other hand, the dissolution rate of quick lime without preheating showed the greatest value even in the FeO–CaO–SiO₂ system (Fig. 17). These results indicate that the formation of CO₂ gas from quick lime has an important role in promoting the dissolution, as it destructs the interfacial layer. Furthermore, porosity does not play a part when the interfacial layer covers the lime surface.

Fig. 16.

Experimental method using slag under the gas bubbling condition by ref. 27 used with permission.

Fig. 17.

Dissolution rate of various samples for three types of slag systems. sCaO is sintered CaO powder, pCaO is preheated quick lime, and mCaO is non-preheated quick lime by ref. 27 used with permission.

The knowledge of the actual dissolution behavior of lime in an industrial furnace with strong stirring is poor. Recently, the slagging behavior of lime in 6 ton converter of IMPHOS project³¹⁾ and 320 ton industrial converter was published.³⁰⁾ In the 6 ton converter, slag was sampled several times after the start of blowing using a special sampler to obtain from various depths in the slag bath. With the 320 ton converter, slag was sampled during slag tapping using a steel chain. For both cases, the interface between the undissolved lime particles and molten slag was observed. The results showed that no interfacial layer of dicalcium silicate was found in any sample. They concluded that under the strong stirring condition, the interfacial layer was actually mechanically removed and the formation rate of this layer, or the mass transfer in this layer, do not have important roles in lime dissolution.

4. Conclusion and Remarks

In this paper, the studies on the dissolution of lime were reviewed. Over these ten years, a lot of research has been conducted with dynamic tests,^{26,27,28,29,32)} stagnant tests,^33,34,35) theoretical modelling,³⁶⁾ and process simulation to consider the slagging behavior.^37,38) This indicates that the dissolution of lime is still one of the unsolved subjects in steelmaking reactions.

By the fundamental experiments, the mechanism and rate-controlling step of lime dissolution has become clear. The destruction of the dicalcium silicate layer is most important to promote the dissolution of lime. The proposed mechanisms of the destruction are fluid flow by stirring and the formation of CO₂ inside the quick lime; however, it should be remembered that the interfacial layer was observed by the dynamic experiments using sintered CaO. In addition, it should be discussed that the slagging process is finished when the interfacial layer detaches, or further dissolution of the detached dicalcium silicate is necessary. By the industrial trials, most of the ideas to enhance the slagging rate have already been conducted. Although, they showed promising results, the slagging ratio was not improved much. As lime is not an expensive material, the economical merit of enhancing slagging is not enough if only the reduction of lime unit consumption is counted. However, the scattering of the slag formation behavior causes scattering of the refining performance and the instability of operation. The improvement of slagging should be evaluated by the economical merit for various viewpoints in steel refining processes.

Based on the existing data, it is already obvious that the static experiment is meaningless. Even in laboratory experiments, the dissolution rate has to be evaluated under the stirring condition using actual quick lime. Moreover, the research should not be finished in the laboratory, but further industrial trials combined with fundamental knowledge are imperative. In this sense, the study has to be conducted under the collaboration of industry and universities.

References

• 1)  Farlex: The free dictionary, http://www.thefreedictionary.com/quicklime, (accessed 2017-02-17).
• 2)   J. S.  Chen,  H. C.  Ho,  T. K.  Hsu,  B. L.  Hsu,  N. P.  Lee and  I. Y.  Chen: Proc. 8th European Slag (EUROSLAG) Conf., EUROSLAG, Duisburg, (2015).
• 3)   I. Z.  Yildirim and  M.  Prezzi: Adv. Civ. Eng., 2011 (2011), 13, http://dx.doi.org/10.1155/2011/463638.
• 4)   S.  Maehara,  K.  Wakabayashi,  T.  Kotani,  K.  Tagiri and  S.  Narita: Tetsu-to-Hagané, 46 (1960), 283.
• 5)   T.  Suzuki,  A.  Sato,  M.  Miyata and  A.  Ota: CAMP-ISIJ, 28 (2015), 661, CD-ROM.
• 6)   P.  Metz,  A.  Decker and  J.  Nepper: Stahl Eisen, 80 (1960), 20.
• 7)   M.  Kishida,  Y.  Nishio,  S.  Cho,  S.  Wakabayashi,  K.  Saito,  G.  Monma and  N.  Kimura: Tetsu-to-Hagané, 51 (1965), 1937.
• 8)   T.  Tanigaki,  T.  Kanbayashi,  Y.  Tajiri,  J.  Tani,  M.  Miyata and  T.  Tamura: CAMP-ISIJ, 24 (2011), 157, CD-ROM.
• 9)   M.  Miyata,  T.  Tamura,  T.  Matsuo,  A.  Matsumoto and  T.  Tanigaki: Materia Jpn., 51 (2012), 16.
• 10)   N.  Ishiwata and  H.  Ito: Tetsu-to-Hagané, 95 (2009), 188.
• 11)   T.  Yasue and  Y.  Arai: J. Soc. Inorg. Mater. Jpn., 258 (1995), 356.
• 12)   M.  Ueda,  Y.  Fujita,  K.  Nakai,  T.  Toshimori and  Z.  Morita: Tetsu-to-Hagané, 75 (1989), 58.
• 13)   S.  Yamamoto,  M.  Yoshii and  M.  Ichinohe: Tetsu-to-Hagané, 57 (1971), S102.
• 14)   A.  Tazuki,  T.  Fujita,  T.  Abiko and  T.  Sugitani: CAMP-ISIJ, 25 (2012), 309, CD-ROM.
• 15)   A.  Matsumoto,  T.  Endo,  H.  Yamazoe,  K.  Fujiwara,  M.  Miyata and  T.  Tamura: CAMP-ISIJ, 24 (2011), 158, CD-ROM.
• 16)   Y.  Ogawa,  M.  Yano,  S.  Kitamura and  H.  Hirata: Tetsu-to-Hagané, 87 (2001), 21.
• 17)   K.  Miyamoto,  N.  Sasaki,  M.  Matsuo and  N.  Sano: CAMP-ISIJ, 17 (2004), 642.
• 18)   M.  Matsushima,  N.  Yadoomaru,  K.  Mori and  Y.  Kawai: Tetsu-to-Hagané, 62 (1976), 182.
• 19)   M.  Matsushima,  S.  Yadoomaru,  K.  Mori and  Y.  Kawai: Trans. Iron Steel Inst. Jpn., 17 (1977), 442.
• 20)   L.  Hachtel,  W.  Fix and  G.  Trömel: Arch. Eisenhüttenwes., 43 (1972), 3617.
• 21)   F.  Oeters and  R.  Sheel: Arch. Eisenhüttenwes., 45 (1974), 575.
• 22)   J.  Yamg,  M.  Kuwabara,  T.  Asano,  A.  Chuma and  J.  Du: ISIJ Int., 47 (2007), 1401.
• 23)   T.  Hamano,  M.  Horibe and  K.  Ito: ISIJ Int., 44 (2004), 263.
• 24)   N.  Sasaki and  Y.  Ogawa: CAMP-ISIJ, 23 (2010), 949, CD-ROM.
• 25)   T.  Deng,  B.  Glaser and  D.  Sichen: Steel Res. Int., 83 (2012), 259.
• 26)   T.  Deng and  D.  Sichen: Metall. Mater. Trans. B, 43B (2012), 578.
• 27)   N.  Maruoka,  A.  Ishikawa,  H.  Shibata and  S.  Kitamura: CAMP-ISIJ, 25 (2012), 217, CD-ROM.
• 28)   A.  Ishikawa,  N.  Maruoka,  H.  Shibata and  S.  Kitamura: CAMP-ISIJ, 24 (2011), 216, CD-ROM.
• 29)   N.  Maruoka,  A.  Ishikawa,  H.  Shibata and  S.  Kitamura: High Temp. Mater. Process., 32 (2013), 15.
• 30)   Z. S. M.  Li,  M.  Whitwood,  S.  Millman and  J.  van Boggelen: Ironmaking Steelmaking, 41 (2014), 112.
• 31)   M. S.  Millman,  A.  Overbosch,  A.  Kapilashrami,  D.  Malmberg and  M.  Bramming: Ironmaking Steelmaking, 38 (2011), 499.
• 32)   W.  Yan,  W.  Chen,  Y.  Yang,  X.  Zhao and  A.  McLean: Proc. AISTech 2015, AIST, Warrendale, PA, (2015), 2164.
• 33)   S.  Amini,  M.  Brungs and  O.  Ostrovski: ISIJ Int., 47 (2007), 32.
• 34)   B.  Deo,  P. K.  Gupta,  M.  Malathi,  P.  Koopmans,  A.  Overbosch and  R.  Boom: Proc. 5th Europe Oxygen Steelmaking Congress, Steel Institute VDEh, Düsseldolf, (2006), 202.
• 35)   X.  Guo,  Z. H. I.  Sun,  J.  van Dyck,  M.  Guo and  B.  Blanpain: Ind. Eng. Chem. Res., 53 (2014), 6325.
• 36)   R.  Sarkar,  U.  Roy and  D.  Ghosh: Metall. Mater. Trans. B, 47B (2016), 2651.
• 37)   N.  Dogan,  G.  Brooks and  M. A.  Rhamdhani: ISIJ Int., 49 (2009), 1474.
• 38)   S.  Kitamura,  K.  Miyamoto,  H.  Shibata,  N.  Maruoka and  M.  Matsuo: Tetsu-to-Hagané, 95 (2009), 313.