Melting Features and Viscosity of TiO₂-Containing Primary Slag in a Blast Furnace

Materials and preparation of the slag

The compositions of the synthetic slag (Table 1) in the present study were selected based on the primary slag composition of a Chinese commercial BF during Vanadium Titano-magnetite smelting in our previous study [13]. The primary-slag formation experiments was taken by the softening-melting equipment which can be found in some earlier literatures [13, 14]. The chemicals used in experiments for synthesizing the slag were all analytical reagents.

FeO was obtained by the pyrolysis of FeC₂O₄. Firstly, the mixture was heated to 1,473 K and held for approximately two hours in Mo crucible within Ar atmosphere to make the FeC₂O₄ fully decompose. And then the furnace was heated to 1,793 K and held for approximately one hour. Subsequently, the melted slags were quenched by water and then crushed. The quenched samples were analyzed for Fe²⁺ and the results are shown in Table 2. The results showed that FeO in the post experimental analysis were slightly but systematically lower than the original value. This suggests that the FeO in the slag was partially oxidized to Fe³⁺ during the experiments. Slightly change of the slag compositions by oxidizing Fe²⁺ to Fe³⁺ is not considered to have significant effect on the slag viscosity and structure. Then the slags were milled to particles of <0.074 mm and shaped by a cylindrical mould to obtain 3(diameter)×5(height) mm samples for melting features measurement.

Experimental apparatus and procedure

Melting features

Melting features measurement system includes three parts: sample hold, heating system and image system, as shown in Figure 1. The mechanism of the measure system is that the sample height will change with the slag melting since temperature rising. Samples were placed in the center of the heating furnace at a rate of 15 K/min below 1,173 K and then kept 5 K/min. The softening temperature (ST), melting temperature (MT), flowing temperature (FT) are defined as the temperatures at which the samples height decrease to 75 %, 50 % and 25 % of the original height, respectively [11, 15]. The initial softening temperature (IT) is considered as the temperature at which the samples height start to decrease. Figure 2 shows the melting features of one sample.

Figure 1:

Melting features measurement system. 1-Gas inlet, 2-Iron valve, 3-Alumina tube, 4-Thermolcouple, 5-Corumdum, 6-Spacer, 7-Sample, 8-C-Si tube, 9-Iron valve, 10-glass, 11-Video camera, 12-data record.

Figure 2:

Melting features of the slag.

Viscosity

The rotating cylinder method was employed for viscosities measurement. The experimental apparatus, measurement principle and the calibration method had been explained in detail in earlier literatures [16, 17]. The viscosities measurement were carried out at every 20 K interval during the cooling cycle with an equilibration time of 30 min at each temperature. The measurement stopped when the viscosities went up to 5 Pa · s and then the samples were reheated to 1,793 K. Finally, the spindle was taken out of the molten slag. The experimental apparatus for the slag viscosities measurement was shown in Figure 3.

Figure 3:

Experimental apparatus for the slag viscosities measurement.

Melting features

The effect of FeO and TiO₂ on the melting features of SiO₂-CaO-MgO-Al₂O₃-FeO-TiO₂ slag are shown in Figure 4. It is clear that all the slags have low initial softening temperature at about 1,423 K. It can be seen that the temperature of the melting features decrease sharply with increasing FeO from 5 wt% to 20 wt% and present a linear correlation between temperature and FeO content. The temperatures of the melting features increase about 15 K as FeO content increases every 5 wt%. The temperature difference between the lowest and the highest melting temperature is about 48 K. The temperature interval between initial softening temperature and softening temperature is higher than 95 K while the temperature difference between softening temperature and melting temperature or between melting temperature and flowing temperature is smaller than 5 K. In addition, the effect of the TiO₂ on the slag shows that the temperature of the melting features increase with increasing TiO₂ from 5 wt% to 10 wt%. As the content of TiO₂ is beyond 10 wt%, the temperatures of initial softening temperature and softening temperature increase slowly while melting temperature and flowing temperature rise sharply. The temperature interval between initial softening temperature and softening temperature is about 60 K. The temperature difference between softening temperature and melting temperature or between melting temperature and flowing temperature is smaller than 5 K when TiO₂ is lower than 10 wt%. The temperature difference between softening temperature and melting temperature is 12 K and the temperature between melting temperature and flowing temperature is 28 K when the TiO₂ content is 15 wt%, respectively. It indicates that the effect of TiO₂ on the slag is different to that of FeO. With the higher content of TiO₂, the melting temperature increases sharply.

Figure 4:

Effect of FeO and TiO₂ on the melting features of slag.

Viscosity

If a slag is melted at high temperature and then cooled slowly, the slag viscosity increases up to a specific temperature, and then increases abruptly below that temperature. Thermodynamically, the temperature is defined as the breakpoint temperature. If the slag is cooled further, solid phase will precipitate and increase the slag viscosity [18]. In this work, the viscosities of the slags varying with the contents of FeO and TiO₂ as a function of temperature are shown in Figures 5(a) and (b), respectively. From Figure 5(a), an increase of FeO content results in the decrease of the breakpoint temperature and slag viscosity. It can be expected that the coexisting regions of solid and liquid increase with FeO increasing. The breakpoint temperatures are 1,638 K, 1,648 K, 1,655 K and 1,687 K of sample 2, sample 3, sample 4 and sample 5, respectively. The viscosities in the high temperature indicate that the viscosities decrease with increasing FeO content. Figure 5(b) shows that the viscosities of the slags are affected by TiO₂ additions at various temperatures. Similar to previously published work [7], TiO₂ additions decrease the viscosity at the high temperature range of investigated. When the TiO₂ content is 5 wt%, the slag has a lower breakpoint temperature and a wider solid-liquid coexisting region than others. With the increase of TiO₂, the coexisting regions decrease obviously. As a result, the breakpoint temperatures increase with increasing TiO₂ from 5 wt% to 10 wt% and 15 wt%, which are 1,608 K, 1,653 K and 1,711 K, respectively.

Figure 5:

Viscosities of CaO–SiO₂–Al₂O₃–MgO–FeO-TiO₂ systems at C/S=1.32 by varying the FeO and TiO₂ contents as a function of temperature: (a) FeO=15 wt%; (b) TiO₂=10 wt%.

Analysis of smelting features

The melting temperature mainly depend on low melting temperature components in the slag, while the viscosity is largely related to the high melting temperature component. In this work, the effect of TiO₂ on melting temperature and viscosity is different while that of FeO is the same. That is, with increasing TiO₂, the melting temperatures increase while the viscosities decrease. However, with increasing FeO, both the melting temperatures and the viscosities decrease. The phenomena can be explained from the phase diagram (Figure 6) of SiO₂-CaO-MgO-Al₂O₃-FeO-TiO₂ slag system calculated by Factsage software. Figure 6(a) shows the phase diagram in the condition of 10 wt% TiO₂. The straight line labeled in the figure represents the slag basicity, which is 1.32. The points on the line are the liquidus of the samples 2, 3, 4 and 5 from right to left. The chemical compositions of the slag investigated in the present work demonstrate that the initial eutectic phase of the samples is CaTiO₃, indicating the slag samples have the similar melting temperature and the similar melting rate. Figure 6(b) shows the phase diagram in the condition of 15 wt% FeO varying with TiO₂. The liquidus of the samples 1, 4 and 6 are also marked on the labeled line. The liquidus increase with increasing TiO₂. It can also be seen that when TiO₂ is lower than 5 wt%, the initial eutectic phase is Ca₃MgSi₂O₈, while the eutectic phase is CaTiO₃ if TiO₂ content is larger than 5 wt%. Thus, the initial eutectic phase of all the samples are CaTiO₃ in the present work.

Figure 6:

Phase diagrams of SiO₂-CaO-MgO-Al₂O₃-FeO-TiO₂ slag system. (a) TiO₂=10 wt%; (b) FeO=15 wt%.

Figure 7 gives the liquidus of the samples compared with the melting temperatures and the breakpoint temperatures, which indicates that the measurement and the calculation are consistent. The temperatures decrease with increasing FeO content and increase with increasing TiO₂. It can also be obtained that the breakpoint temperatures are higher than the liquidus temperatures and the melting temperatures. The breakpoint temperatures are 10–20 K higher than the liquidus with increasing FeO from 5 wt% to 20 wt% and 15–50 K higher than the liquidus with increasing TiO₂ from 5 wt% to 15 wt%.

Figure 7:

Smelting features of the slag. (a) TiO₂=10 wt%; (b) FeO=15 wt%.

Apparent activation energy

The temperature dependence of the viscosity is usually expressed by Arrhenius equation [7].

where A is a proportionality constant, E_w is the apparent activation energy for viscous flow, R is the gas constant and T is the absolute temperature.

The apparent activation energy for slags E_w represents the frictional resistance for viscous flow. E_w represents the energy barrier that the cohesive flow units in slags have to overcome when the units move between different equilibrium states [19]. The variation of E_w suggests the structure changes in the slags and its value is expected to be constant for a certain slag system before the solid precipitation occurs. Based on the experimental data, the apparent activation energies for viscous flow of TiO₂ containing primary slags are evaluated according to Weymann–Frenkel’s equation. The results present in Figure 8. The curve of lnη changing with FeO and TiO₂ can be divided into four periods according to the increase rate. In this study, the viscosity curves derivative of the temperature are conducted and the turning points are regarded as the dividing location. Those four periods are named slow period (I), rapid period (II), slow down period (III) and dramatically rise period (IV).

Figure 8:

Natural logarithm of the viscosity for various TiO₂ slags vs reciprocal temperature. (a) TiO₂=10 wt%; (b) FeO=15 wt%.

The E_w values are evaluated from the slope of the lines according to the Arrhenius relationship and present in Figure 9 in respect of the slowly period (before solid precipitation occurs). It shows that the apparent activation energies (E_w) decrease with increasing TiO₂ and FeO content, indicating that the reduce of energy barrier for viscous flow and the formation of some simpler structural units in the slags [20, 21]. It suggests that both TiO₂ and FeO may primarily behave as a basic oxide and act as network modifier in the I period. Besides, the decrease of E_w value is significant when TiO₂ content increases from 10 to 15 wt%, which is different from the results of H. Park et al. [22] 5 wt% TiO₂ decreases the viscosity significantly compared to the TiO₂ free slags. In this work, free oxygen provided by the FeO in the slag may make much more pronounce than the influence of the TiO₂. In the rapidly period (II), the solid phase CaTiO₃ has not participated firstly because the breakpoint temperatures are lower than that of liquidus. The increase of the viscosities of the slags indicates that the introduction of TiO₂ into silicate network may act as a weak acidic oxide in the basic slag system at a relative lower temperature [23, 24]. TiO₂ can capture oxygen ions and the existence of Ti₂O₆^4– chain units gives rise to a higher degree of polymerization for silicate network [25, 26]. Then the titanium ion is advantage to generate CaTiO₃ when the temperature is lower than the liquidus temperature. In the slow down period (III), the effect of Ti₂O₆^4– may decrease and the effect of CaTiO₃ is gradually enhanced. In the IV period, an amount of the solid phase participated. As a result, the viscosities of the slags rise dramatically. It can be seen from Figure 8(a) that the regions of the slow period (I) extend and that of periods of II, III and IV postpone with increasing FeO, which indicates that FeO in the slag can reduce the effect of Ti₂O₆^4– chain units and decrease the precipitation temperature of the solid phase. It can be seen from Figure 8(b) that the four periods are not obvious when the TiO₂ content is 5 wt% while it becomes more and more obvious with increasing TiO₂.

Figure 9:

Effect of FeO and TiO₂ concentration on the apparent activation energy of viscous flow of slags in I period.