Wetting of Solid Iron by Molten FeO-Containing Slag

Abstract

The study aims to contribute the approaches for optimizing the parameters of the cohesive zone in blast furnace, as well as for enhancing the efficiency of novel alternative ironmaking processes. Wetting conditions between primary slag and iron sponge determine carbonization of the latter, thus affecting the position in the blast furnace of the region where slag and metal separate into independent liquid phases. Wetting of solid iron by molten FeO-containing slags representing the primary blast furnace slag was studied experimentally using a sessile drop method. Effects of the FeO content and slag basicity on the contact angle of slag on the iron substrate have been revealed. The opportunities of controlling the wetting conditions between primary slag and solid iron by optimizing the basicity of iron ore materials are discussed.

1. Introduction

In the blast furnace ironmaking, the properties of primary FeO-rich slag determine the process of carbon transfer to the freshly reduced solid iron, followed by separation of slag and metal into the independent liquid phases [1]. In some alternative ironmaking technologies, aimed to substitute blast furnace, an interaction between primary slag and solid iron also plays an essential role [2].

The schematic picture of the blast furnace is shown in Figure 1. In the lower part of the shaft of the blast furnace, reduction and heating processes result in forming of a cohesive zone where the layers of partially reduced softened iron ore material (sinter or pellet), impermeable for the ascending gas flow, are pressed between the gas-permeable layers of coke. The form and the thickness of the cohesive zone significantly affect all aspects of the blast furnace operation [4].

Figure 1.

Schematic picture of blast furnace [3].

In the cohesive zone, chemical reactions take place mostly on the surface of softened material, notably on the lower part of the cohesive layer, where the primary slag, formed by interaction of iron monoxide (FeO) and the gangue compounds of iron ore material (predominantly, such as SiO₂, CaO, Al₂O₃ and MgO), contacts with the hot reducing gas and solid coke [1]. Due to pressure of the materials in the shaft of the blast furnace, FeO-rich primary slag squeezes from the iron sponge (formed due to reduction of iron oxides) to the surface of the cohesive layer.

Initially, high oxidizing potential of FeO-rich primary slag prevents carbonization of the iron sponge. However, while the materials descend to the area of blast furnace with elevated temperature and stronger reducing potential of gas, FeO is gradually reduced from the primary slag and oxidizing potential of the latter decreases. At certain point, change of physicochemical parameters of interaction among the solid iron, the primary slag, the gas atmosphere and the coke, results in very rapid carbonization and melting of iron at the lower part of the cohesive zone. After separation into the independent liquid phases, slag and metal start to drip through the active coke zone (a term applied to the area between the cohesive zone and coke combustion raceways) down to the hearth (cylindrical part at the bottom of blast furnace) where they settle into the layer of hot metal (pig iron) topped with the layer of slag. Pig iron and slag are periodically tapped from the blast furnace [1, 2, 4, 5].

Studies on the quenched blast furnaces, performed in the 1970s in Japan [6], revealed that dripping of metal in the lower part of the shaft of blast furnace starts after its carbonization to 0.8–1.0% C. In the belly (the widest region of the blast furnace below the shaft), the carbon content of metal droplets grows to around 2.0%; further, in the bosh (conic region of blast furnace between the belly and the hearth), it reaches around 4.0% (tapped pig iron contains 3.8–4.5% C). Same pattern was observed for the silicon transfer to metallic phase: In the softened cohesive layer, metal contains less than 0.03% Si, then, just after the onset of carbonization, Si content grows to 0.2%, reaching 2% or more at the top of the bosh-even if a conversion pig iron (a semi-product used to further produce steel) with less than 0.6% Si is produced (for the sake of brevity, the phenomena of partial oxidation of metal components near the raceways are not discussed in this chapter).

Along with the silicon, sulfur is another compound whose concentration in pig iron shall be controlled in blast furnace operation. Samplings performed by Volovik [7] on an operating blast furnace revealed that, although sulfur absorption by the iron ore sinter is observed yet while it descends from the blast furnace top to the middle of the shaft, the freshly reduced solid iron, being enveloped by the primary slag, does not significantly absorb sulfur. After the liquid metal is formed, its sulfur content drastically increases, reaching 0.3–0.4% S as a maximum (a value then decreases to final content of around 0.03% S in the tapped iron due to desulfurization of metal by slag in the hearth).

The industrial data referred above show that, in the blast furnace, freshly reduced solid iron, enveloped by the primary slag, remains virtually unaffected by chemical reactions within a certain period of time while the materials descend. However, at a certain point, the interplay of complex heating and reduction processes triggers very rapid change in composition and physical state of the metallic phase. Therefore, knowledge of the processes of interaction between the liquid FeO-rich slag and the solid iron is very important for the definition of iron ore materials’ composition-either for blast furnace cohesive zone optimization or for more energy-efficient operation of alternative ironmaking processes. However, our analysis reveals very few studies where interaction phenomena between solid iron and liquid primary slag are investigated (most studies focus on wetting phenomena at the interface of slag and metallic melts).

Iguchi et al. [8] studied wettability of solid iron by the slags in various CO₂/CO atmospheres, revealing that the liquid slags with more than 30 mol% FeO content perfectly wet the solid iron surface. It was found that the contact angle increases with the increase in the basicity of slags under the constant oxygen pressure, while addition of Al₂O₃ and MgO has no significant effect.

Hino et al. [9] studied some parameters of interaction between the solid iron and liquid slags representing FeO-2CaO⋅SiO₂⋅Al₂O₃(Gehlenite)-CaO⋅SiO₂ system with the ratio of Gehlenite/(CaO⋅SiO₂+ Gehlenite) = 0.3. Good wettability of iron by the studied slags was revealed, with the wetting angle decreasing in the range from 30° to 10° with the increased FeO content in the slag.

In both referred above studies, a sessile drop technique was applied with the temperature level fixed for all experiments-at 1350°C in [9] and at 1450°C in [8]. Noteworthy, these temperatures exceed the liquidus temperature for the slag studied systems and generally correspond to the conditions when the liquid slag phase is already separated from the iron phase-whether it is a lower boundary of cohesive zone in the blast furnace or, for example, an iron nugget, produced in the innovative ITmk3 ironmaking process [10]. In other words, by 1350°C a primary slag should be already long ago separated from the sponge iron, so the data available from the referred above studies [8, 9] are not very relevant to the conditions of sponge iron and primary slag interaction preceding their separation into the flowable phases.

Experimental approach with fixed temperatures, applied in [8, 9], allows for comparison of the wetting conditions in different slag systems. However, in reality, the temperature when primary slag and iron sponge separate into the independent liquid phases depends upon the reducing potential of gas, the reduction degree of the material and the gangue composition.

In our earlier study of the softening and melting properties of iron ore materials [11], it was shown that, in the viscous-plastic state under the load, an impermeable material is formed with the outer part coated by the slag relatively depleted in FeO due to the reduction, while its internal part contains iron sponge and FeO-rich slag. Under such conditions, the development of the physicochemical processes of iron oxides’ reduction and carbon transfer to the metallic phase, resulting in slag and metal separation, is to a great extent determined by the slag properties. As far as FeO content in slag determines both fluidity and oxidizing potential of slag, it should play a predominant role in these processes. Kim et al. [12] also found that after the iron oxide in the slag is reduced, the separation of the final slag and the Fe–C melt takes place since the wettability between them decreases.

In the current research, we studied wetting of solid iron by the FeO-rich slag under the temperature conditions close to the temperature of complete melting for the given slag-conditions typical for the blast furnace cohesive zone, where interaction between the heterogeneous phases (liquid slag and solid sponge iron) takes place.

2. Experimental methods

Synthetic CaO-SiO₂-FeO slags were studied. FeO was produced using a mix of iron (purity of 99.99%, fineness <10 μm) and Fe₂O₃ (purity of 99.95%, fineness <5 μm) in a stoichiometric ratio corresponding to the reaction Fe₂O₃ + Fe = 3FeO. Quartz tube with the mix was closed with an iron wire sponge ball to prevent oxidation and purged with an Ar gas, then heated to 1050°C, held for 5 hours and quenched in water. “Pure for analysis” grade SiO₂ and CaO powders were used to prepare ternary CaO-SiO₂-FeO mixtures by weighing, mixing and grinding in a mortar.

Slag compositions (shown in Table 1) with relatively low liquidus temperature were chosen from different crystallization fields of a ternary CaO-SiO₂-FeO system (Figure 2) representing Wollastonite, Olivine and Wüstite. Noteworthy, in 3D phase diagrams of a quaternary slag systems (with MgO or Al₂O₃ as the fourth slag component), liquidus surfaces in these crystallization fields are plateau-like, in contrast to surrounding steep surfaces of Tridymite and Dicalcium Silicate crystallization fields. Composition of low-melting slags from Wollastonite, Olivine and Wüstite crystallization fields is most relevant to the primary slags, formed in partly reduced iron ore materials.

#	Chemical composition, mass %				Contact angle, degrees	Temperature of measurement (complete melting temperature + 20°C), °C	Surface tension at temperature of measurement, calculated using method of Mills [13], mN/m
	SiO2, %	CaO, %	CaO/SiO2	FeO, %
1	50.0	22.0	0.44	28.0	60.0	1200	482
2	43.0	29.0	0.67	28.0	50.0	1200	509
3	38.0	34.0	0.89	28.0	44.0	1200	528
4	43.0	11.0	0.26	46.0	38.5	1150	512
5	38.0	16.0	0.42	46.0	19.0	1150	533
6	26.0	28.0	1.08	46.0	9.0	1150	580
7	38.0	0.0	0.00	62.0	5.5	1250	515
8	32.0	6.0	0.19	62.0	8.5	1150	555
9	25.0	13.0	0.52	62.0	22.2	1150	583

Table 1.

Composition and properties of the studied slags.

Figure 2.

Ternary phase diagram for the CaO-SiO₂-FeO system [14]. Points denote slag compositions for the experiments listed in Table 1.

Wettability of solid iron was studied using a sessile drop method. Experimental setup is demonstrated in Figure 3. Slag cone (1) with 5 mm diameter and 5 mm height and average mass of approximately 200 mg was placed on the ARMCO® pure iron (99.98% Fe) substrate (2) established on the chamotte plate in the quartz glass tube (6) of the horizontal furnace with SiC heating elements (7). Two Pt-Pt/Rh type B thermocouples (9) were used: The first one, used for data recording, was placed near the studied sample and the second one-just under the quartz tube-to control isothermality of the reaction zone.

Figure 3.

Experimental setup: 1—slag sample, 2—iron substrate, 3—thermocouples, 4—gas outlet, 5—photo camera, 6—quartz tube, 7—SiC heating elements, 8—gas inlet, and 9—quartz glass.

Technically pure Ar (99.99% Ar, Linde Gas) was constantly supplied to the reactor. Gas was exiting the reactor through the line (4) down to the water lock. Heating rate of 10° C per minute was applied. A digital photo camera was used to record the processes of melting and wetting. Obtained images were digitally processed using ImageJ [15] freeware. When the temperature of complete melting for a given slag was reached (i.e., when the entire sample formed a hemisphere), the temperature in the furnace was allowed to rise by another 20°C and stabilize within 5 minutes, and then, the contact angle was measured.

To ensure that the results are representative, three slag mix samples were studied. Two results with the smallest and the highest contact angle values were excluded, while the remaining result was retained for the analysis.

3. Results and discussion

The results of contact angle measurement are represented in Table 1 and plotted against FeO content of slag in Figure 4. Data from the sources [8, 9] were approximated by the curves and also represented in Figure 4 (for the data from [8] molar concentration was converted to mass concentration). As can be observed from Figure 4, slags with higher FeO content exhibit substantially lower contact angles with iron: Increase of FeO content from 28 to 62% is followed by fivefold drop in contact angle. The obtained results are generally consistent with the trends reported in literature [8, 9]. At the same time, it is interesting to mention that, as seen from Figure 5, surface tension of the slag slightly grows with the increasing of FeO content (data were calculated using the spreadsheet available from [13], representing modified partial molar method developed by Mills [16]). Such discrepancy might be explained by better wetting of the solid iron surface (due to chemical interaction of FeO and iron) at higher FeO content, which offsets the slight increase of surface tension.

Figure 4.

Measured contact angle versus FeO content in slag in comparison with data of Iguchi et al. [8] and Hino et al. [9].

Figure 5.

Surface tension calculated using method of Mills [13] versus FeO content in slag.

Although, both in our experiments and in studies of the other authors, contact angle decreases with the increased FeO content in slag, our data show substantially higher values of contact angle for the comparable slag compositions (e.g., around 28% FeO). This difference can be explained by higher temperatures of measurement, applied in the referred studies [8, 9].

At the first sight, the experimental data for dependency of the contact angle upon the slag basicity are substantially scattered; however, as demonstrated in Figure 6, after the slags are grouped by the FeO content, it is possible to observe that, for the slags with FeO content of 28% and 46%, the increased basicity enhances wettability of the iron (contact angle decreases). On the contrary, for the slags with 62% FeO, contact angle increases with the increased basicity. Such ambivalent influence of CaO might be explained by the surface tension increase. A similar effect was previously reported by Kozakevitch [17] who studied surface tension of various slag mixes-addition of CaO to pure FeO is initially followed by surface tension’s decrease; however, after CaO content in binary FeO-CaO melt reaches approximately 15%, surface tension starts increasing. Our slags with 62% FeO content may follow the behavior of the FeO-CaO binary system. Grouping of the slags by the crystallization fields in the ternary phase diagram (also demonstrated in Figure 6) shows that the wetting conditions in the studied system can be efficiently controlled by tuning the slag composition and adjustment of the reduction degree of iron ore material (resulting in certain FeO content in slag).

Figure 6.

Contact angle versus slag basicity. Slags compositions presented in Table 1 are grouped here according to crystallization fields as follows: 1, 2 and 3—wollastonite; 4 and 5—eutectic valley between olivine and wollastonite; 7 and 8—olivine; 6 and 9—Wüstite.

In the study of Iguchi et al. [8], increase of the slag basicity was followed by increased contact angle; however, analysis of their data shows that the contact angles for the CaO-SiO₂-FeO slags with the basicity ratios (CaO/SiO₂) of 1.0 and 1.1 (both from the Wollastonite crystallization field) are higher when compared to the slag with the basicity of 0.5 from the Tridymite crystallization field. Therefore, it is possible to presume that the effect of slag basicity on the wettability of solid iron is rather complex and depends upon FeO content in the slag and other parameters of slag composition.

As discussed above in the introduction, wetting conditions between FeO-rich primary slag and the freshly reduced iron sponge determine iron carbonization and its intake of sulfur and silicon. Therefore, the obtained results might be used to develop a method for controlling the composition of metallic phase by tuning the composition of iron ore material and the regime of reduction. In our further studies, we aim to reveal the primary slag compositions (FeO content and basicity) favorable for limiting the transfer to the iron sponge of such elements as silicon, sulfur and probably even phosphorus (the latter under the blast furnace conditions is by 100% reduced to the hot metal), which might be applied for developing of a novel ironmaking method.

Table 2 outlines the expected aspects of such novel technology in comparison with blast furnace (major ironmaking technology within an integrated steelmaking route), Midrex® (major technology in relevant segment of integrated steelmaking) producing Direct Reduced Iron/Hot Briquetted Iron (DRI/HBI) and ITmk3® (one of the alternative ironmaking technologies, commercialized in 2010, so far at a single plant). As shown in Table 2, we expect that thermodynamic conditions of iron ore reduction can be adjusted to prevent carbonization of metallic phase as well adsorption of silicon, sulfur and phosphorus. Certain losses of iron with FeO-containing slag are an unavoidable aspect of such approach to some extent, and this can be considered as a rebirth of an ancient bloomery process but on a current technology control level. However, we believe that novel technology should have competitive advantages to offset this drawback. Such advantages include

• superior product composition;

• lower (compared with blast furnace and ITmk3®) temperature level, resulting in lower energy requirement;

• absence of gangue material in the product should allow usage of medium grade iron ore in contrast to Midrex® where gangue material stays in the product, so its amount shall be kept low in order to minimize slag yield in the electric arc furnaces producing steel from DRI/HBI, which requires usage of scarce high-grade ores with >67% Fe [21].

Product parameters		Blast furnace [18]	DRI/HBI (Midrex®) [19]	ITmk3® [20]	Novel technology (expected)
Product composition	Fe metallic, %	≈94	83–89	97.2	≈99
	C, %	4.0–4.5	1.0–3.5	2.0–2.4	≈0.8
	Si, %	0.3–1.0		≈0.8	≈0.03
	S, %	0.03	0.001–0.03	0.07–0.12	≈0.03
	P, %	<0.2	0.005–0.09	0.01–0.025	≈0.01
	Gangue, %	Removed to slag	2.8–6.0	Removed to slag	Removed to slag
Product temperature, °C		≈1450 (determined by melting points of a product and of low-FeO slag)	≈950 (determined by thermodynamics of the reduction processes)	≈1350 (determined by melting point of a product)	1150–1250 (determined by melting point of Fe-rich slag)
State of product		Liquid (pig iron)	Solid pellets (DRI) or briquettes (HBI)	Liquid, then solidified into “nugget”	Solid (sponge)

Table 2.

Parameters of product in ironmaking technologies.

Certainly, the precise aspects and, needless to say, technological layout of the proposed concept are yet to be considered in our further research where study of the wettability phenomena should play substantial role.

4. Conclusions

Interaction between the freshly reduced solid iron and the primary FeO-containing slag considerably determine carbonization of iron and subsequent separation of metal and slag into the independent liquid phases in the blast furnace operation conditions. Transfer of sulfur and silicon among the solid, gaseous and liquid phases is also affected by FeO content in slag which is determined by the wettability phenomena on the slag-metal interface. Wettability of solid iron by molten ternary CaO-SiO₂-FeO slags in an inert atmosphere under the temperature conditions close to the complete melting for the given slags was studied experimentally, and the main results are as follows:

1. The contact angle of a molten FeO-containing slags on the solid iron substrate decreases from 60.0° to 5.5° when the FeO content in slag increases from 28 to 62%. At the same time, surface tension of the slag slightly grows with the increasing of FeO content. Such discrepancy is explained by better wetting of the solid iron surface at higher FeO content due to chemical interaction of FeO and iron, which offsets the slight increase of surface tension.

2. Effect of the slag basicity on the contact angle depends upon the FeO content in slag: For the slag with 28% and 46% FeO, increased basicity ratio (CaO/SiO₂) of slag is followed by enhanced wetting of the solid iron. However, for the slags with 62% FeO, contact angle increases with the increased basicity. Such ambivalent effect of FeO in the ternary slag systems is in the agreement with previous studies of binary FeO⋅CaO melts.

3. A possibility of controlling the wetting conditions between the primary slag and the solid iron by tuning the composition of slag-forming compounds of iron ore material can be used to optimize the parameters of cohesive zone in blast furnace and to control the composition of hot metal or to design innovative energy-efficient ironmaking processes with limited transfer of sulfur and, possibly, phosphorus to the metallic phase.

Conflict of interest

The authors declare no conflict of interest.