Effect of boron oxide additives on viscosity and melting point of the CaO – SiO₂ – Al₂O₃ – MgO system

Introduction

For blast furnace smelting to proceed efficiently, the slag formed during the process must possess stable physicochemical properties that are only slightly affected by fluctuations in chemical composition and temperature.

Slag melt viscosity is one of the most important physicochemical parameters that determine the stability and productivity of blast furnace operation. The final blast furnace slag should have good flowability during tapping. At tapping temperatures corresponding to those of steelmaking pig iron (1450 – 1550 °C), slag viscosity should be approximately 0.5 Pa∙s [1; 2].

In slags, the transition from solid to liquid occurs over a certain temperature range, so the melting point (Т_m) is a conditional value. The theoretical indicator of melting point is the liquidus temperature (Т_l) – the temperature at which the solid phase fully disappears upon heating. In practice, the melting point is considered the temperature at which slag begins to flow freely from the coke bed, which typically occurs when viscosity drops below 2.5 Pa∙s. This temperature is usually 200 – 300 °C lower than the tapping temperature and falls within the range of 1250 – 1350 ℃ [3].

Slag viscosity is greatly influenced by chemical composition, which is determined by the chemical and mineralogical makeup of the iron ore gangue, fluxes, and coke ash. It also depends on the nature of the smelting process, the thermal state of the furnace, and the grade of pig iron being produced. The content of major components in final slags at ironmaking plants across Russia, Ukraine, Europe, and the USA typically includes (wt. %: 30 – 40 SiO₂, 31 – 49 CaO, 3 – 18 MgO, 7 – 20 Al₂O₃. Minor components include MnO (0.1 – 3.0 %), FeO (0.2 – 0.8 %), and S (0.8 – 2.2 %) [4; 5]. Excluding trace oxides (MnO, FeO, S), it is reliably established that the base of any blast furnace slag is the four-component system CaO – SiO₂ – Al₂O₃ – MgO.

The relationship between slag viscosity and temperature in terms of these major components (CaO, SiO₂, Al₂O₃, MgO) has been studied extensively, with findings presented in numerous works.

In melts containing less than 15 % alumina, an increase in basicity (R) from 0.6 to 1.5 and in MgO content from 0 to 20 % raises the melting point to 1350 – 1400 °C and narrows the solidification interval. The slags become “short” (i.e., less fluid). At temperatures below 1400 °C, slags with more than 25 % MgO exhibit poor flowability [6 – 9].

Changing MgO content from 0 to 25 % in slags with basicity within 0.6 – 1.5 leads to a decrease in viscosity to a minimum value that depends on alumina content and temperature. In acidic slags, viscosity decreases more significantly than in basic ones [10].

For slags containing 5 % aluminum oxide, the minimum viscosity of 0.15 Pa·s at 1500 °C is reached in the composition range of R ≈ 0.9 – 1.1, with 17 – 20 % MgO and 36 – 38 % SiO₂. Reducing the temperature to 1400 °C raises the minimum viscosity to 0.35 Pa∙s and expands the MgO range for achieving this minimum to 13 – 20 %, while shifting toward more acidic slags with 39 – 41 % SiO₂.

Raising the alumina content to 10 % increases the minimum viscosity. As the temperature decreases from 1500 to 1400 °C, viscosity rises from 0.2 to 0.3 Pa∙s. The composition range for achieving this minimum narrows from R ≈ 0.8 – 1.2, 13 – 24 % MgO, 35 – 40 % SiO₂ (at 1500 °C) to R ≈ 1.05 – 1.2, 14 – 16 % MgO, and 39 –  41 % SiO₂ (at 1400 °C).

With 15 % Al₂O₃, the minimum viscosity increases further from 0.30 to 0.55 Pa∙s. The corresponding composition range narrows from R ≈ 0.9 – 1.2, 15 – 26 % MgO, and 30 – 33 % SiO₂ to R ≈ 0.8 – 1.05, 18 – 22 % MgO, and 33 – 35 % SiO₂ as the temperature drops from 1500 to 1400 °C. Increasing MgO content leads to a particularly sharp decrease in viscosity in acidic slags containing 25 – 35 % CaO. In slags with R ≈ 0.5 – 0.8, 13 – 18 % Al₂O₃, and 16 – 25 % MgO, sufficient flowability is maintained at 1350 – 1400 ℃.

In slags containing 20 % Al₂O₃, the melting point exceeds 1500 °C for any MgO content in the R ≈ 1.2 – 1.5 range. At R ≈ 1.1 – 1.2, crystallization occurs when MgO exceeds 16 %. As R decreases to 0.6, the critical MgO content increases to 20 %. When the MgO/Al₂O₃ ratio is approximately 0.5, Т_l is close to 1450 °C at R ≈ 1.1 – 1.2 and decreases to 1350 °C when R drops to 0.6. In such slags, the minimum viscosity varies from 0.4 Pa∙s (at 1500 °C) to 1.0 Pa∙s (at 1400 °C) at SiO₂ contents of 34 – 36 % [11 – 13].

Analysis of the data shows that in slags with basicity below 1.0, MgO content may reach 15 – 20 % without significantly complicating smelting, since the slags remain sufficient fluid and melt below 1350 °C. According to calculations [14], such slags form from blast furnace charges containing approximately 20 – 30 % siderite ore with 10 – 15 % MgO. A further increase in MgO content raises the slag melting point and leads to poor fluidity and process instability. Smelting with such charges becomes technically challenging or even unfeasible.

The gangue of siderite ores contains about 50 % MgO [15 – 17]. In this regard, in blast furnace smelting, siderite ore is used as an additive either directly in the initial charge or during sinter production. Smelting with a monocomponent charge of siderite ore from the Bakalskoye deposit is not feasible, since the resulting slags would have excessively high melting points. However, it is known [18 – 21] that the addition of boron oxide to blast furnace slags reduces their viscosity across the entire temperature range and makes them more fluid.

Методы и материалы исследования

Using a mathematical balance logical-statistical model, we assessed the feasibility of conducting blast furnace smelting with increased magnesium oxide content in the slag (from 15 to 30 %) due to the addition of 50 % roasting–magnetic separation concentrate (RMC). Within this framework, the effect of adding 1 – 3 % B₂O₃ on smelting parameters was evaluated [14].

According to the calculations, the changes in process performance parameters do not exceed 3 %. Productivity decreases, while coke consumption and total ore consumption increase. The addition of B₂O₃ reduces the content of all oxide components in the slag, including MgO, and introduces boron oxide into the system.

To evaluate the influence of boron oxide additions on the viscosity and melting point of the CaO – SiO – Al₂O₃ – MgO system, a simplex-lattice design of experiments was employed. This method provides a process model that closely reflects real-world behavior by accounting for the simultaneous influence of all varying factors.

Composition–property diagrams are geometric representations of multicomponent equilibrium systems. They include a compositional base, which reflects the chemical composition of the system, and a geometric property component (response surface) represented by a set of points, lines, or surfaces located above it. These diagrams provide quantitative information about specific properties corresponding to particular compositions within a multicomponent system.

However, constructing such diagrams requires a large number of experimental runs to determine how a property depends on the chemical composition of a multicomponent system.

The labor intensity of experimental studies necessitates optimizing the number of experiments. To achieve this, mathematical modeling is used in experimental design to analyze the influence of different components on the property under investigation and to obtain sufficiently complete information with minimal experimental effort. This approach significantly reduces the number of required experiments while ensuring that the results are obtained with an adequate level of reliability. One such method is the simplex-lattice approach, which analytically expresses the property–composition relationship as a continuous function [22].

In the simplex-lattice design of experiments, it is assumed that the properties of any mixture depend solely on the relative proportions of its components rather than on the total amount of the mixture. Examples of such properties include density, surface characteristics, viscosity, specific electrical conductivity of homogeneous metal and slag melts, gas solubility in solvent mixtures, and the hydrogen ion concentration (pH) of aqueous solutions, provided that their phase composition remains unchanged.

The preparation of the experimental design matrix was aimed at studying the methods of experimental planning, execution, and statistical processing of the results.

The matrix was developed based on the assumption that the target dependency could be approximated by a third-degree polynomial model, with the following initial conditions:

\(\frac{{{\rm{CaO}}}}{{{\rm{Si}}{{\rm{O}}_2}}}\) = R = 0.9 ÷ 1.2;

MgO = 15 – 36 %;

B₂O₃ = 0 – 15 %; Al₂O₃ = 5 – 20 %.

These starting conditions were based on the following: according to the calculations, increasing the share of siderite ore in the charge from 0 to 50 % results in SiO₂ and MgO contents rising from 35 to 38 % and from 9 to 36 %, respectively, while CaO and Al₂O₃ contents decrease from 40 to 17 % and from 14 to 8 %, respectively. The addition of up to 3 % boric anhydride to the charge leads to up to 15 % B₂O₃ in the final slag.

Considering that MgO has significantly lower desulfurization capacity than CaO, the basicity of the slag was maintained in the range of 0.9 – 1.2. The MgO content in blast furnace slags was varied between 15 and 36 %, as concentrations below 15 % have little effect on the smelting process. In most final blast furnace slags produced during pig iron smelting in Russia and Western countries, the Al₂O₃ content varies from 5 to 20 %.

Thus, the studied composition field within the full five-component CaO – SiO₂ – Al₂O₃ – MgO – B₂O₃ system is represented by a tetrahedron, the vertices of which correspond to pseudocomponents Y1 , Y2 , Y3 , and Y4. For each experimental slag composition in the design matrix, the precise content of each component was calculated (Table 1).

Synthetic slags were prepared for experimentation with CaO and SiO₂ in a ratio corresponding to R = 0.9 – 1.2. The calcium oxide used (analytical grade) was pre-calcined in a muffle furnace at 910 °C for 6 h. The base samples were produced by heating and melting a mixture of oxides (CaO–SiO₂) in a graphite crucible at 1500 – 1550 ℃ (30 min hold). The melt was poured into a mold and cooled.

The resulting slags were then blended in appropriate proportions with magnesium oxide (pre-calcined at 400 °C), aluminum oxide, and boric anhydride (pre-melted in a resistance furnace at 900 °C for 4 h). These mixtures were melted in a graphite crucible at 1500 – 1550 °C (30 min hold), poured into molds, cooled, and ground.

The powders were pressed into pellets, placed in a molybdenum crucible, heated to 1550 °C, and subjected to viscosity measurements. A vibrational viscometer operating in forced oscillation mode was used for this purpose [23; 24], with the melt temperature recorded using a tungsten–rhenium thermocouple. The measuring probe was made of molybdenum. Viscosity was measured during cooling at a rate of 5 – 7 °C/min.

Results and discussion

The results of the experiments are presented in Table 2 and Figs. 1 – 4.

Fig. 1 shows the isolines of the response function for the temperature at which the given viscosity is achieved, with increasing B₂O₃ content from 0 to 10 %. According to the data in Fig. 1, an increase in B₂O₃ content in the slag leads to a decrease in the temperature required to achieve the specified viscosity, particularly in the simplex region with high MgO content.

Fig. 2 presents the isolines of the response function for the temperature corresponding to the given viscosity, as the basicity of the slag increases from 0.9 to 1.1. As shown in the figure, higher basicity results in higher temperatures needed to reach the given viscosity in simplex regions with elevated MgO and Al₂O₃ content. At the same time, in regions with high B₂O₃ content, the required viscosity is achieved at relatively low temperatures – between 1100 and 1300 °C – regardless of the slag basicity.

Fig. 3 shows isolines of the response function for the temperature at which the given viscosity is reached as MgO content increases from 15 to 29 %.

The results presented in Fig. 3 indicate that, despite nearly doubling the MgO content in the slag, in the simplex region with maximum B₂O₃ content, the increase in the temperature needed to reach the given viscosity is minor. The temperature remains below 1380 °C at minimum viscosity and below 1340 °C at maximum viscosity.

Fig. 4 presents isolines of the response function for the temperature at which the given viscosity is reached, as Al₂O₃ content increases from 5 to 15 %.

The results in Fig. 4 demonstrate that increasing the Al₂O₃ content in the slag lowers the temperature required to reach the desired viscosity in the region with high MgO content. This temperature is further reduced when B₂O₃ content is also increased.

Across the entire composition range of the CaO – SiO₂ – Al₂O₃ – MgO –B₂O₃ system, in the absence of boron oxide, the temperature corresponding to the slag melting point (defined as the temperature at which viscosity reaches 2.5 Pa·s) exceeds 1390 °C. At basicity below 1.1, MgO content below 20 %, and Al₂O₃ content above 10 %, the melting point lies in the range of 1300 – 1400 °C. The viscosity of 0.5 Pa·s, which corresponds to the tapping viscosity, is reached at 1300 – 1440 °C for any basicity level, provided that MgO content is below 20 % and the Al₂O₃/MgO ratio exceeds 0.5. Moreover, the higher this ratio, the lower the temperature required to reach that viscosity. This indicates that blast furnace smelting with slags of this composition can proceed without operational difficulties.

A further increase in magnesium oxide content causes a sharp rise in slag melting point – up to 1500 °C – alongside a narrowing of the crystallization interval. The tapping viscosity temperature rises to 1540 °C when MgO content reaches 36 %, making smelting under such conditions difficult or even infeasible.

The addition of B₂O₃ lowers the temperature at which the viscosity of the melt reaches 2.5 Pa·s, and at 15 % B₂O₃, this temperature drops to below 1150 °C. the slags exhibit improved fluidity and greater thermal stability.

Conclusions

At present, the content of major components in most blast furnace slags is as follows (wt. %): 30 – 40 SiO₂, 31 – 49 CaO, 3 – 18 MgO, and 7 – 20 Al₂O₃. Minor components include MnO (0.1 – 3.0 %), FeO (0.2 – 0.8 %), and S (0.8 – 2.2 %). Such slags exhibit good fluidity (viscosity below 0.5 Pa·s) at temperatures above 1450 °C. An increase in magnesium oxide content above 25 % significantly reduces slag fluidity and raises the melting point, making the slags viscous and difficult to process, which complicates blast furnace operation. For this reason, materials produced from siderite ore using various preparation technologies (raw ore, roasting–magnetic separation, agglomeration) are added to the blast furnace charge only as supplementary components. Their share in the initial charge is kept below 20 % to ensure that the resulting slag does not contain more than 15 – 20 % MgO.

The addition of boron oxide–containing materials to the initial charge lowers the slag melting point. This enables blast furnace smelting with slags containing approximately 40 % MgO, which corresponds to a siderite ore share of 40 – 50 % in the initial charge.