A Review of Fluorine-free Mold Flux Development

Abstract

More than 95% of crude steel has been produced through the process of continuous casting technology, in which mold flux plays important roles inside the mold. As conventional mold flux system contains certain amount of fluorides that tend to be evaporated and cause the corrosion of casting facilities and environmental pollutions etc. It is urgent to optimize the mold flux composition without fluorides for the continuous casting process. In this review paper, the substitutes for fluorides in mold flux are summarized via two main strategies: TiO₂ and B₂O₃ based additives. Among them, CaO–SiO₂–Al₂O₃–Na₂O–B₂O₃ based fluoride-free (F-free) mold flux system shows great potential to replace conventional mold flux that can be widely used in the process of continuous casting, as the precipitated phase-Ca₁₁Si₄B₂O₂₂ has the most similar crystallization behavior to cuspidine in conventional mold flux, which can effectively control the horizontal heat transfer and reduce the occurrence of longitudinal cracks. This paper would provide technical guidance and research direction for the design and study of efficient environment-friendly F-free mold flux.

1. Introduction

In the mid-19 century, Sellers and Bessemer et al. first proposed the concept of continuous casting for molten steel, which was originally used for the casting of low melting point metals such as Lead. Until 1943, the first continuous casting machine for casting steel was invented. By the 1950s, the continuous casting technology was formally applied to the industrial process of steel production. Subsequently, the continuous casting technology has been rapid developed and wide applied, as it owns the advantages of short production process, high metal recovery rate and low energy consumption, et al. In modern steel industry, more than 95% of crude steel production has been produced through continuous casting technology.¹⁾

1.1. The Importance of Mold Flux

In the process of continuous casting, mold is called “heart” of the continuous casting machine, where the initial solidification of molten steel occurred inside the mold. When the molten steel enters inside the mold through the nozzle, the meniscus will be formed and begin to solidify around the meniscus area at the top of the mold, as shown in Fig. 1(a).²⁾ In order to ensure the smooth progress of continuous casting, mold flux is often added to the top of the mold. The slag power layer is first formed when the mold flux accumulates on the top of molten steel surface, which can (i) provide thermal insulation to prevent the steel from freezing.³⁾ Then, the mold flux is gradually melted and form molten slag layer, and the chemical reactions will occur and some volatile substances like fluorides, etc. will evaporate in this period. The molten slag layer can (ii) isolate the air to protect the steel from oxidation.⁴⁾ Besides, the molten slag layer can (iii) absorb the inclusions in the steel/slag interface that float up from molten steel. If the inclusions can not be absorbed effectively, they will gather at the steel/slag interface and be captured by the new shell that affect the surface quality of the slab. The significant absorption of the inclusions will change the chemical compositions of mold flux, such that its performances would be deteriorated. The molten steel that flows through submerged entry nozzle (SEN) will form an undertow on the upper part of the mold, which may catch the molten slag into the molten steel.⁵⁾ Therefore, the study on steel/slag interface behavior is important for the development of mold flux and the improvement on slab quality.

Fig. 1.

The schematic of mold inside; (b) The distribution of mold flux in the mold.²⁾

As the molten slag flows into the gap between copper mold and initial shell, it usually forms three typical different structure layers, i.e. a glassy flux layer adjacent to the copper mold, a liquid layer contact with the initial shell due to its high temperature, and a crystalline flux layer formed in between, as shown in Fig. 1(b). The liquid mold flux close to the high temperature initial shell can (iv) lubricate the strand during the oscillation period. Also, the liquid mold flux temperature tends to reduce in the vertical direction along the initial shell due to the further cooling of the mold, which in turn to affect the rheological behavior of mold flux and deteriorate its lubricating property. The glassy mold flux layer next to the copper mold is due to the high cooling rate inside the mold cooling channels. The solid crystalline layer precipitated in between the liquid and glassy layer can (v) inhibit the radiative heat transfer from the steel and generate air gaps between the mold and partially solidified shell during the initial solidification of molten steel, leading to the increase of the interfacial thermal resistance.^6,7) The distribution of slag film between the copper mold and initial shell is very important, as it can control the general heat transfer from the liquid pool to the copper mold and lubricate the strand during the mold oscillation.

Finally, the slag film will contact with the cooling water in the secondary cooling zone when it exits out of the mold outlet. At this time, some ions in the slag film are released into the cooling water. Finally, the used mold fluxes will be treated as wastes and dumped in the slag yard, in which the ions in slag films are also released into soil to pollute ground water.⁸⁾

1.2. The Function of Fluoride in Mold Flux

For a conventional mold flux system, CaO and SiO₂ are usually added as the base components. Besides, alkali oxides and fluorides are included as the fluxing agents. In order to control the melting speed of mold flux, carbon black, graphite and coke etc. are also required. A typical chemical compositions of a conventional mold flux is listed in Table 1.⁹⁾

Table 1. The part of chemical compositions of the conventional mold flux (Mass%).

CaO	SiO2	Al2O3	Na2O	K2O	MgO	MnO	Fe2O3	F	C
22–45	17–56	0–13	0–25	0–2	0–10	0–5	0–6	2–15	2–20

Among them, the addition of fluorides can significantly reduce the melting temperature and the viscosity of mold flux system,^10,11) and can effectively promote the solid-state slag crystallization, and the most important is that it will precipitate a vital phase-cuspidine (Ca₄Si₂O₇F₂), which can effectively control the horizontal heat transfer from the molten steel to the mold.^12,13,14) Therefore, it has become a common additive for mold flux. Generally, the content of fluoride in the mold flux is 6%–8%; sometimes, it can be up to 11%–13% for casting special steels.¹⁵⁾

However, fluorides will react with SiO₂ to generate the low melting point SiF₂ that tends to evaporate in the atmosphere during the melting process, also fluorides will generate HF once it contacts with cooling water,¹⁶⁾ the above reactions are shown as the following Eqs. (1) and (2). The generated fluorides will cause the corrosion problem of the casting facilities and environmental pollutions etc.^17,18,19,20) Therefore, it is urgent to develop environment-friendly fluorine-free (F-free) mold flux for the process of continuous casting.   

$$2{\text{CaF}}_{\text{2}}\left(\text{Slag}\right)+{\text{SiO}}_{\text{2}}\left(\text{Slag}\right)\to {\text{SiF}}_{\text{4}}\left(\text{g}\right)+\text{2CaO}\left(\text{Slag}\right)$$(1)

  

$${\text{CaF}}_{\text{2}}\left(\text{Slag}\right)+{\text{H}}_{\text{2}}\text{O}\to \text{2HF}\left(\text{g}\right)+\text{CaO}\left(\text{Slag}\right)$$(2)

However, the mold flux system without fluoride will result in many problems, in which the worst issues are the augment of mold flux viscosity, the reduction of crystallization performance and the deterioration of heat transfer, etc. In addition, it will change the slag interface properties, cause slag entrapment and affect the absorption of inclusions. The five functions of the mold flux without fluoride will be severely damaged, which cause the deterioration of slab quality and even lead to breakout of continuous casting production. Therefore, it is urgent to find alternative to modify the negative effects caused by the absence of fluoride. In the past decades, many oxides as alternatives have been tried to decrease the viscosity of F-free mold flux and form an effective slag film to control the heat transfer between the molten steel with the copper mold, primarily they are focused on TiO₂, B₂O₃ and Na₂O.

The objective of this review paper will systematically summarize the development of F-free mold flux, and the progress of the plant trials that mold flux with TiO₂, B₂O₃ and Na₂O to substitute fluorides will be introduced. Moreover, the shortcomings and the existing problems for the developed mold fluxes systems will also be discussed. The aim of this review paper is to provide technical guidance and research direction for the design and study of an efficient environment-friendly F-free mold flux.

2. Influence of Alternative on the Mold Flux Properties

2.1. Effect of TiO₂ on the Mold Flux Properties

TiO₂ is a typical amphoteric oxide, the addition of TiO₂ into the molten flux will lead to the appearance of [TiO₄]-tetrahedral and [TiO₆]-octahedral structure groups, which would destruct the major Si–O–Si network structure and lead to the formation of complex titanium-silicate structure groups, resulting in the improvement of the degree of polymerization (DOP) of silicate network.^21,22) The schematic evolution process of above network structure evolvement with the combination of titanium structure is shown in Fig. 2. Meanwhile, TiO₂ tends to lower the viscosity and the apparent activation energy of the slags, as the introduction of Ti⁴⁺ into the silicate network will weaken the strength of slag structure and a large proportion of Ti⁴⁺ would exist in the slag in the form of [TiO₄]-tetrahedral monomers, which results in the reduction of molten slag viscosity.²²⁾ Qi et al. suggested that the addition of TiO₂ would firstly increase the viscosity of F-free mold flux, and then lower the viscosity when its content is higher than 6.0%.²³⁾ The phenomena could be explained as that TiO₂ can firstly enhance the DOP of silicate network to increase the system viscosity. On the other hand, it would release more Ti⁴⁺ into the slag system to weaken the strength of slag structure and a large proportion of Ti⁴⁺ ions would form [TiO₄]-tetrahedral monomers, which results in a reduction of viscosity. So, the effect of TiO₂ on the viscosity of F-free mold flux depends on which effect plays a major role.

Fig. 2.

The schematic of the combination process of silicate structure with titanium structure: (1) [TiO₄]-tetrahedral mainly exists in the form of [TiO₄]-tetrahedral monomers; (2) [TiO₆]-octahedral will destruct Si–O–Si bond in silicate structure to form titanium-silicate structure.

The combination of [TiO₄]-tetrahedral and [TiO₆]-octahedral structure into the silicate structure with addition of TiO₂, would promote the formation of CaO·TiO₂ (perovskite) and CaO·SiO₂·TiO₂ crystalline phase to affect the solidification and crystallization behavior of mold flux system.²⁴⁾ Research shows that TiO₂ will diminish the crystallization activation energy of mold flux and increase the crystallization ratio of mold flux, and the mineral phase of the slag film turns from akermanite into perovskite, which can effectively decrease the heat flux density of F-free mold flux.²⁵⁾ The precipitated crystalline phase-perovskite in the F-free powder can well control the heat transfer in the continuous casting mold, as the cuspidine does in the conventional F-containing mold slags.²⁶⁾ Besides, study suggests that another precipitated crystalline phase-CaO·SiO₂·TiO₂ can also be used to replace cuspidine, because the crystallization incubation time in CaO–SiO₂–TiO₂ based F-free mold flux system is similar to the conventional mold flux, but the thickness of the crystalline layer is thinner than that of cuspidine, and Na₂O is usually added to decrease the incubation time of CaO·SiO₂·TiO₂ at high temperature.²⁷⁾

2.2. Effect of Na₂O and B₂O₃ on Mold Flux Properties

Na₂O is often used as the fluxing agents to be added in mold flux, it works as network breaker and could provide free oxygen ions (O²⁻) to depolymerize the bridged oxygen (O⁰) in network structure.²⁸⁾ The addition of Na₂O can shorten the crystallization incubation time and enhance the crystallization performance of mold flux.^29,30) Besides, Na₂O can effectively reduce the melting temperature and viscosity of mold flux.³¹⁾ However, Na₂O is easy to evaporate with the rise of the mold temperature.³²⁾

B₂O₃ is a typical network former, its addition into the molten slag usually leads to the formation of [BO₃]-trihedral and [BO₄]-tetrahedral structure units in the mold flux, among which [BO₃]-trihedral is more abundant than [BO₄]-tetrahedral. However, with the further addition of B₂O₃ content, [BO₃]-trihedral would form [BO₄]-tetrahedral, as shown in Eq. (3):^33,34)   

$$\left[{\text{BO}}_{\text{3}}\right]+2{\text{Si-O}}^{-}\to \left[{\text{BO}}_{\text{4}}\right]+\text{Si}-\text{O}-\text{Si}$$(3)

Meantime, [BO₄]-tetrahedral structure units will introduce with silicate structure units to form complex borosilicate structure, through link the original Q⁰ (Si) (it means [SiO₄]-tetrahedral with no bridge oxygen) and Q² (Si) ([SiO₄]-tetrahedral with two bridge oxygen), in which the original Q⁰ (Si) and Q² (Si) transferred into Q¹ (Si) ([SiO₄]-tetrahedral with one bridge oxygen) and Q³ (Si) ([SiO₄]-tetrahedral with three bridge oxygen), resulting in the improvement of the DOP of silicate structure.^{33,34,35,36,37)} The evolution process of network structure with the addition of B₂O₃ is shown in Fig. 3. As an effective fluxing agent, B₂O₃ tends to significantly decrease the viscosity of mold flux.^37,38,39) It is because B₂O₃ can drastically decrease the break temperature of mold flux, which in turn to improve the superheat degree of the molten flux and lead to the reduction of the melts viscosity. Furthermore, once [BO₃]-trihedral (a simple two-dimensional structure) introduces into three-dimensional complex structure, it will greatly reduce the symmetry of the current network structure, and the present borate structure has preferentially connected to the near Si atoms and thus introduced into the silicate networks, which can lower the uniformity of the networks, resulting in the reduction of viscosity of mold flux.³⁷⁾ It could be found that the viscosity of mold flux is not only decided by molten structure, but also greatly affected by its solidification behavior.

Fig. 3.

The schematic of the combination process of silicate structure with borate structure: [BO₃]-trihedral transform to [BO₄]-tetrahedral and two non-bridging oxygen (O⁻) linked to form bridge oxygen; then [BO₄]-tetrahedral combine Q⁰ (Si) and Q² (Si) to form borosilicate structure with Q¹ (Si) and Q³ (Si).

Moreover, B₂O₃ tends to react with other substances to form lower melting point substance such as CaO·B₂O₃, 2CaO·B₂O₃ etc., which can greatly decrease the whole melting temperature range of the low fluorine mold flux.^19,38) The addition of B₂O₃ could lower the liquidus and the crystallization temperature of mold flux system, and it usually increases the incubation time of the slag crystallization.^40,41) Therefore, the crystallization behavior of mold flux is inhibited and the general heat transfer rate across the mold flux is improved. However, Na₂O plays an opposite role compared with that of B₂O₃.^30,42) Thus, the combined effects of B₂O₃ and Na₂O can be used to adjust the general crystallization and heat transfer properties of mold flux for the development of F-free mold flux.

3. Plants Trials

3.1. Substitution of Fluorides with TiO₂

According to previous laboratory experiments, Wen et al.²⁶⁾ designed two types of CaO–SiO₂–TiO₂ based F-free mold flux, and the major chemical compositions as well as the physicochemical properties are detailed in Table 2. They are used for plant trials for casting two types of peritectic steel grades, i.e. plain steel and low alloy steel (contain high Mn content). The results of the plant trials indicated that the F-free mold powders are uniformly melted in the mold, no stick phenomenon, also no formation of lumps and thick slag rims in the mold. The heat flux value of the F-free mold powder designed is close to that of the corresponding F-bearing mold powder used. The morphology and crystallization of the slag film are further analyzed by Scanning Electron Microscope (SEM), and shown in Fig. 4. It can be seen that the main precipitated phase is rod-like perovskite crystal in F-free mold flux, and the crystalline fraction of the slag film is about 42%. Study suggested that the surface quality of the continuous casting slab produced by using the two types of the F-free mold powders doesn’t show significant problems, and shown in Fig. 5.

Table 2. The major chemical compositions and physicochemical properties designed F-free mold fluxes.²⁶⁾

No.	Part of chemical compositions/mass%								Tm/°C	η1300°C/Pa·S	Tc/°C	η/%
	CaO	SiO2	Al2O3	Fe2O3	MgO	MnO	TiO2	C
1#	28.23	29.47	4.56	1.05	3.56	3.91	4.83	5.35	1030	0.187	1104	35
2#	30.75	27.65	5.38	1.47	3.49	3.45	6.82	8.45	1122	0.226	1163	60

T_m, η_1300°C, T_c and η represent hemisphere point temperature, viscosity at 1300°C, crystallization temperature, and crystallization fraction of the F-free mold powders, respectively.

Fig. 4.

Morphology of the slag film taken from plant mold: (a) F-free mold flux; (b) F-bearing mold flux.²⁶⁾

Fig. 5.

The surface quality of the cast slabs: (a) using 1# F-free mold flux; (b) using 2# F-free mold flux.²⁶⁾

However, TiO₂-containing F-free mold flux is subsequently proved to have some limitations, sometimes even cause sticker breakouts when it is used to cast crack-sensitive steels.⁴³⁾ One of the reasons for the breakouts might be the imbalance of the consumption of CaO and SiO₂ caused by the precipitation of perovskite.²⁷⁾ Another reason for the catastrophic breakouts is that TiO₂ in the slag pool would react with carbon and nitrogen to form the high melting point substances, such as titanium nitride (TiN), titanium carbide (TiC) and titanium carbonitride (Ti(C,N)) at the interface between carbon enriched layer and liquid slag layer, which would increase the viscosity and the break temperature of mold flux and lead to a lack of lubrication and severe friction during mold oscillation.⁴⁴⁾ Thus, TiO₂ is hard to replace CaF₂ in the F-free mold flux system, especially for casting crack-sensitive steels.

3.2. Substitution of Fluorides with B₂O₃ and Na₂O

Fox et al.⁸⁾ designed a CaO–SiO₂–8.3%Al₂O₃–12.35%Na₂O–1.49%B₂O₃ based F-free mold flux without TiO₂ for billets casting and carried out comparative plant trials with the traditional F-containing flux. The billet is a typical low-alloyed hardening and tempering steel 42CrMo4E, and the results of the billets surface quality are shown in Fig. 6. It can be seen there are no visible cracks in the cast product from F-free mold flux, but there exist some minor cracks in the case of F-containing mold flux, and the oscillation marks are more pronounced on the F-free flux cast products. Recently, Klug et al.⁴⁵⁾ also designed a B₂O₃-containing F-free mold flux as a reference and conducted laboratory scale experiments to compare the its physicochemical properties with conventional F-bearing mold flux, its major composition and physicochemical properties are listed in Table 3. Industrial trials are also performed for casting low carbon steel slabs, in which the slabs section is 1300×250 mm, the average casting speed is about 1.15 m/min. It is observed from the results that the melting performance and the alumina absorption ability of the F-free mold flux is similar to the F-bearing one; the slag pool thickness for F-free and F-bearing mold flux is in the range of 15–20 mm; SEN erosion rate with F-free mold flux is significantly lower than F-bearing mold flux; and mold powder consumption of casting every ton of steel for F-free mold flux is about 0.37 kg and for F-bearing is about 0.4 kg. The slabs produced during the industrial trials with the F-free mold flux presented a good surface quality.

Fig. 6.

The billet surface of 42CrMo4E carbon steel casted by: (a) using F-free mold flux; (b) using F-containing mold flux.⁸⁾

Table 3. Elemental compositions and physicochemical properties of F-bearing and F-free mold flux.²⁶⁾

Flux	Main chemical compositions/wt%								Tm/°C	η1300°C/Pa·S	Tf/°C	Tbk/°C
	CaO	SiO2	Al2O3	MgO	Na2O	F	B2O3	Cfree
F-bearing	35.1	37.8	2.9	5.2	4.6	7	0	1.9	1191	0.34	1202	1083
F-free	31.1	33.6	1.8	5.8	7.9	0	5.6	2.5	1069	0.33	1084	1078

T_m, η_1300°C, T_f and T_bk represent hemisphere point temperature, viscosity at 1300°C, flowing temperature, and the temperature of the rapid change of viscosity of the mold fluxes, respectively.

4. Outlook

Although many industrial trials on F-free mold flux are carried out and some claimed to successfully replace the conventional F-bearing mold flux; however, the majority study is only limited to crack-insensitive steels. Very few studies have been conducted for the casting of medium carbon (MC) steels, which a typical representative of crack-sensitive steels, due to the volumetric shrinkage during the δ-γ phase transformation that gives rise to the thermal stresses within the solidified shell and results in the frequent occurrence of longitudinal cracking.⁴⁶⁾ In order to minimize the possibility of the longitudinal cracking formation, the in-mold heat transfer for casting MC steels needs to be reduced with the aim to keep a thin and uniform solidified shell to minimize the effect of thermal stresses. The reduction of heat transfer is usually achieved by maintaining a thick solid crystalline layer of slag film, as the crystallization of mold flux tends to block radiative heat transfer and increase the interfacial thermal resistance by the formation of mold/slag gap.^6,7) Besides, the thick solid crystalline layer can resist in-mold ferrostatic pressure. Conventional F-bearing mold flux can precipitate cuspidine crystalline phase, which can effectively control the thickness of the solid crystalline layer of slag film to reduce the heat transfer and achieve the reduction of the longitudinal cracking formation. However, cuspidine (Ca₄Si₂O₇F₂) cannot be precipitated in the mold flux without fluoride, thus the horizontal heat transfer cannot be effectively controlled, and the development of F-free mold flux for casting MC steels becomes very challenge.

Wang et al.^30,47,48) conducted research on the development of low fluoride mold flux system for casting MC steels, which is based on the CaO–SiO₂–4%Al₂O₃–Na₂O–B₂O₃–3%CaF₂ slag system, and the research is mainly on the crystallization behavior and heat transfer performance. The results indicated that the crystallization behavior of the low fluoride mold flux is very similar to the benchmark conventional F-containing mold flux, and a new crystalline phase-calcium borosilicate (Ca₁₁Si₄B₂O₂₂) together with cuspidine are precipitated in in low temperature region, which exhibits the most similar crystallization behavior to that of cuspidine, as shown in Fig. 7. It can effectively increase the thickness of the solid crystalline layer of slag film to reduce the heat transfer and make the reduction of the longitudinal cracking occurs.

Fig. 7.

The crystallization behavior: (a) TTT diagram of F-containing mold flux; (b) TTT diagram of low fluoride mold flux; (c) XRD result of F-containing mold flux; (d) XRD result of low fluoride mold flux.³⁰⁾

Subsequently, Wang et al.^42,49,50) designed a group of CaO–SiO₂–Na₂O–B₂O₃ based F-free mold flux for casting MC steel based on a commercial F-bearing mold flux for casting MC steel, through replacing fluorides with B₂O₃ and adjusting basicity, Na₂O, Li₂O contents to compensate the negative effects of the mold flux properties caused by the absence of fluorides. The research on the performance of the F-free mold fluxes was studied and compared with the commercial F-bearing mold flux. The results suggested that the performances of the F-free mold flux with basicity of 1.15, B₂O₃ contents of 6%, Na₂O contents of 8% and Li₂O contents of 2% are most close to the F-bearing mold flux, the more components of the F-free mold flux and F-bearing mold flux are listed in Table 4 and their physicochemical properties are showed in Table 5. The results suggested that the major properties of the F-free mold flux are very similar with the F-bearing mold flux, except for the viscosity of the F-free mold flux that is slightly higher than the F-free mold flux. Besides, the crystallization behavior of the F-free mold flux and the F-bearing mold flux also were studied by using Double Hot Thermocouple Technology (DHTT), and their results were showed in Fig. 8. It could be observed from the crystallization behavior of the F-bearing mold flux that fine crystals begin to precipitate from the upper right corner of the cold side during the cooling process, then the they are growing up and moving toward the hot side; meantime the new crystals are also formed, till the crystallization process is finished and remains a steady state. It takes 41 s to complete the whole crystallization process. The liquid slag layer in the final steady state is only about 4.63% and the crystalline layer is 95.37%. As for the F-free mold flux, the whole crystallization process takes only 27 s, and the liquid slag layer in the final steady state in the case of F-free mold flux system is about 6.47% and the crystalline layer is 93.53%. The steady state mold fluxes were quickly cooled and the removed for Scanning Electron Microscope (SEM) and XRD analysis, the results were shown in Figs. 9 and 10, respectively. It can be seen from the combination of SEM images and XRD results, the precipitated crystalline phase of the F-bearing is cuspidine (Ca₄Si₂O₇F₂) and a small number of gehlenite (Ca₂Al₂SiO₇) and they appear a larger grain size. While, the precipitated crystalline phase in the F-free mold flux is calcium borosilicate (Ca₁₁Si₄B₂O₂₂) and a small part of calcium magnesium silicate (Ca₁₄Mg₂(SiO₄)₈), and they are smaller and more compacted. Those results further confirm that the calcium borosilicate can be formed in the F-free mold fluxes system and it shows good potential to replace cuspidine in conventional F-bearing mold fluxes system.

Table 4. The components of mold fluxes (Mass%).⁴²⁾

Flux	CaO/SiO2	CaO	SiO2	Al2O3	Na2O	B2O3	CaF2	Li2O
F-bearing	1.25	43.06	34.44	4	7.5	0	8	1
F-free	1.15	41.72	36.28	4	8	6	0	2

Table 5. The physicochemical properties of mold fluxes.⁴⁹⁾

Flux	Melting temperature range/°C	η1300°C/Pa·S	Tbk/°C	Slag film structure (total thickness of 4.4 mm)			Heat transfer/ KW/m2
				Liquid	Crystalline	Glassy
F-bearing	1064–1216	0.136	1253	0.23	3.51	0.66	499
F-free	1071–1224	0.256	1220	0.28	3.43	0.69	505

Fig. 8.

(a)–(d): The crystallization evolution process of the commercial F-bearing mold flux; (e) the final steady state of the commercial F-bearing mold flux; (f) the final steady state of the F-free mold flux.⁴⁸⁾

Fig. 9.

The SEM images of mold flux: (a) the commercial F-bearing mold flux; (b) the F-free mold flux.

Fig. 10.

The XRD results of mold flux: (a) the commercial F-bearing mold flux; (b) the F-free mold flux.

Aiming at the problem of high viscosity of the F-free mold flux and poor thermal stabilization of the calcium borosilicate, Zhang et al.³⁷⁾ suggested that further increasing B₂O₃ content on the basis of 6% can effectively reduce the viscosity of the F-free mold flux, also further addition of basicity on the basis of 1.15 has no obvious effect on the viscosity,³⁷⁾ but it will improve the crystallization temperature and promote the crystallization of mold flux.⁵¹⁾ Study suggested that ZrO₂ acting as heterogeneous nucleating agent could be added into the F-free mold flux, which can strengthen the crystallization ability of F-free mold flux and promote the precipitation of calcium borosilicate phase.⁵²⁾

5. Further Works

The CaO–SiO₂–Na₂O–B₂O₃ based F-free mold flux shows a great potential to be applied to the industry for the casting process of various steels, where the discovery of calcium borosilicate (Ca₁₁Si₄B₂O₂₂) appears very similar crystallization behavior as cuspidine (Ca₄Si₂O₇F₂) in CaO–SiO₂ based traditional F-containing mold flux system. However, there are still many works to be done for the development of CaO–SiO₂–Na₂O–B₂O₃ based F-free mold flux system. First, it is necessary to explore the influence of the major additive components and their interactions on the comprehensive properties of the F-free mold flux. Then, the study on the molten structure of the mold flux and the relationship between structure and properties need to be carried out, due to the fact that B₂O₃ as a network former in mold flux system, the addition of B₂O₃ would corporate into the original single silicate structure and form complex borosilicate structure, leading to the variation of the F-free mold flux properties. Besides, the most important issue is that the research focused on the crystallization and heat transfer behavior of F-free mold flux and the effective control of the precipitation and growth of calcium borosilicate (Ca₁₁Si₄B₂O₂₂), becomes the key issues of the development of F-free mold flux system.

In summary, the research of CaO–SiO₂–Na₂O–B₂O₃ based F-free mold flux is essential to be conducted in the future. More efforts should be made to provide the fundamental understanding of the molten structure, viscosity, crystallization, heat transfer, etc. of the F-free mold flux system.

Acknowledgments

The financial support from National Science Foundation of China (U1760202, 51661130154) and the Newton Advanced fellowship (NA150320) is great acknowledged.

References

• 1)  Steel Statistical Yearbook 2017, World Steel Association, Brussels, (2017), 3.
• 2)   Y.  Meng and  B.  Thomas: Metall. Mater. Trans. B, 34 (2003), 707.
• 3)   K.  Mills: ISIJ Int., 56 (2016), 1.
• 4)   K.  Mills and  A.  Fox: ISIJ Int., 43 (2003), 1479.
• 5)   R.  Mcdavid and  B.  Thomas: Metall. Mater. Trans. B, 27 (1996), 672.
• 6)   K.  Gu,  W.  Wang,  L.  Zhou,  F.  Ma and  D.  Huang: Metall. Mater. Trans. B, 43 (2012), 937.
• 7)   W.  Wang and  A.  Cramb: Steel Res. Int., 81 (2010), 446.
• 8)   A.  Fox,  K.  Mills,  D.  Lever,  C.  Bezerra,  C.  Valadares,  I.  Unamuno,  J.  Laraudogoitia and  J.  Gisby: ISIJ Int., 45 (2005), 1051.
• 9)   M.  Srinivasan: Science and Technology of Casting Processes, InTech, New York, (2012), 205.
• 10)   C.  Li and  B.  Thomas: Metall. Mater. Trans. B, 35 (2004), 1151.
• 11)   S.  Sridhar,  K.  Mills and  S.  Mallaband: Ironmaking Steelmaking, 29 (2002), 194.
• 12)   P.  Grieveson,  S.  Bagha,  N.  Machingawauta,  K.  Liddel and  K. C.  Mills: Ironmaking Steelmaking, 15 (1988), 181.
• 13)   T.  Watanabe,  H.  Hashimoto,  M.  Hayashi and  K.  Nagata: ISIJ Int., 48 (2008), 925.
• 14)   M.  Hanao,  M.  Kawamoto and  T.  Watanabe: ISIJ Int., 44 (2004), 827.
• 15)   P.  Riboud,  Y.  Roux,  L.  Lucas and  H.  Gaye: Fachber. Huttenprax. Metallweiterverarb., 19 (1981), 859.
• 16)   M.  Persson,  S.  Seetharaman and  S.  Seetharaman: ISIJ Int., 47 (2007), 1711.
• 17)   Y.  Kashiwaya and  A.  Cramb: Metall. Mater. Trans. B, 32 (2001), 401.
• 18)   S.  Choi,  D.  Lee,  D.  Shin,  S.  Choi,  J.  Cho and  J.  Park: J. Non-Cryst. Solids, 345–346 (2004), 157.
• 19)   G.  Li,  H.  Wang,  Q.  Dai,  Y.  Zhao and  J.  Li: J. Iron Steel Res. Int., 14 (2007), 25.
• 20)   A.  Zaitsev,  A.  Leites,  A.  Lrtvina and  B.  Mogutnov: Steel Res. Int., 65 (1994), 368.
• 21)   Z.  Wang,  Q.  Shu and  K.  Chou: High Temp. Mater. Process., 32 (2013), 265.
• 22)   K.  Zheng,  Z.  Zhang,  L.  Liu and  X.  Wang: Metall. Mater. Trans. B, 45 (2014), 1389.
• 23)   X.  Qi,  G.  Wen and  P.  Tang: J. Iron Steel Res. Int., 17 (2010), 6.
• 24)   S.  He,  Q.  Huang,  G.  Zhang,  Y.  Lu and  Q.  Wang: J. Iron Steel Res. Int., 18 (2011), 15.
• 25)   X.  Qi,  G.  Wen and  P.  Tang: J. Non-Cryst. Solids, 354 (2008), 5444.
• 26)   G.  Wen,  S.  Sridhar,  P.  Tang,  X.  Qi and  Y.  Liu: ISIJ Int., 47 (2007), 1117.
• 27)   H.  Nakada and  K.  Nagata: ISIJ Int., 46 (2006), 441.
• 28)   J.  Gao,  G.  Wen,  Q.  Sun,  P.  Tang and  Q.  Liu: Metall. Mater. Trans. B, 46 (2015), 1850.
• 29)   J.  Klug,  D.  Silva,  S.  Freitas,  M.  Pereira,  N.  Heck,  A.  Vilela and  D.  Jung: Steel Res. Int., 83 (2012), 791.
• 30)   J.  Wei,  W.  Wang,  L.  Zhou,  D.  Huang,  H.  Zhao and  F.  Ma: Metall. Mater. Trans. B, 45 (2014), 643.
• 31)   H.  Kim,  W.  Kim,  J.  Park and  D.  Min: Steel Res. Int., 81 (2010), 17.
• 32)   Z.  Tong,  J.  Qiao and  X.  Jiang: Ironmaking Steelmaking, 44 (2017), 237.
• 33)   Y.  Sun and  Z.  Zhang: Metall. Mater. Trans. B, 46 (2015), 1549.
• 34)   J.  Li,  Y.  Sun,  Z.  Li and  Z.  Zhang: ISIJ Int., 56 (2016), 752.
• 35)   Z.  Li,  Y.  Sun,  L.  Liu and  Z.  Zhang: Metall. Mater. Trans. E, 3 (2016), 28.
• 36)   Z.  Wang,  Q.  Shu and  K.  Chou: ISIJ Int., 51 (2011), 1021.
• 37)   L.  Zhang,  W.  Wang,  S.  Xie,  K.  Zhang and  I.  Sohn: J. Non-Cryst. Solids, 460 (2017), 113.
• 38)   L.  Zhou,  W.  Wang,  B.  Lu,  G.  Wen and  J.  Yang: Met. Mater. Int., 21 (2015), 126.
• 39)   Z.  Wang,  Q.  Shu and  K.  Chou: Steel Res. Int., 84 (2013), 766.
• 40)   Z.  Wang,  Q.  Shu and  K.  Chou: Can. Metall. Q., 52 (2013), 405.
• 41)   Q.  Shu,  Z.  Wang,  J.  Klug,  K.  Chou and  P.  Scheller: Steel Res. Int., 84 (2013), 1138.
• 42)   L.  Zhou,  W.  Wang,  J.  Wei and  K.  Zhou: ISIJ Int., 55 (2015), 821.
• 43)   Q.  Wang,  S.  He and  K.  Mills: Proc. 4th Int. Conf. on Continuous Casting of Steel in Developing Countries, The Chinese Society for Metals, Beijing, (2008), 651.
• 44)   Q.  Wang,  Y.  Lu,  S.  He,  K.  Mills and  Z.  Li: Ironmaking Steelmaking, 38 (2011), 297.
• 45)   J.  Klug,  M.  Pereira,  E.  Nohara,  S.  Freitas,  G.  Ferreira and  D.  Jung: Ironmaking Steelmaking, 43 (2016), 559.
• 46)   K.  Mills,  A.  Fox,  Z.  Li and  R.  Thackray: Ironmaking Steelmaking, 32 (2005), 26.
• 47)   B.  Lu,  W.  Wang,  J.  Li,  H.  Zhao and  D.  Huang: Metall. Mater. Trans. B, 44 (2013), 365.
• 48)   L.  Zhou,  W.  Wang,  J.  Wei and  B.  Lu: ISIJ Int., 53 (2013), 665.
• 49)   L.  Zhou,  W.  Wang and  K.  Zhou: ISIJ Int., 55 (2015), 1916.
• 50)   L.  Zhou and  W.  Wang: Metall. Mater. Trans. E, 3 (2016), 139.
• 51)   W.  Wang,  X.  Yan,  L.  Zhou,  S.  Xie and  D.  Huang: Metall. Mater. Trans. B, 47 (2016), 963.
• 52)   L.  Zhou,  H.  Li,  W.  Wang and  I.  Sohn: Steel Res. Int., 88 (2017), 1.