Effect of water vapour content in H 2 –H 2 O–CO–CO 2 mixtures on activity of iron oxide in slags relevant to novel flash ironmaking technology

Abstract

As an integral part of developing a novel flash ironmaking technology at the University of Utah, the activity of iron oxide in the slag was studied under three different gas atmospheres: H₂/H₂O (H₂), CO/CO₂/H₂/H₂O (reformed natural/coal gas), and CO/CO₂. The conditions of the slags investigated were MgO-saturated CaO–FeO–Al₂O₃–SiO₂–MnO (0·2–0·8 wt-%)–P₂O₅ (0·1–0·9 wt-%) in the temperature range 1550–1600°C with wt-% CaO/wt-% SiO₂ of 0·8 to 1·2, and under pO₂ = 2×10⁻¹⁰–2×10⁻⁹ atm. Water increased the activity coefficient of FeO in the slag and accordingly lowered the FeO content. The average FeO content was found to be 10, 11 and 16 wt-% under H₂/H₂O (H₂), CO/CO₂/H₂/H₂O (reformed natural/coal gas), and CO/CO₂, respectively. An empirical correlation for γ_FeO in slags under H₂/H₂O atmospheres was formulated to givelog γ_FeO = −3·0623 X _FeO−3·1421 X _CaO−2·5068 X _MgO+2·1957

Introduction

Iron oxide plays an important role in metal–slag reactions including those involved in sulphur and phosphorus distribution. In addition, FeO in the slag is considered as a loss in ironmaking and steelmaking processes. Moreover, FeO content in the slag affects its viscosity, corrosivity, and redox potential of the slag.¹ Thus, FeO activity (a _FeO) in slag and the related thermodynamics attracted the attention of many researchers.^2–6 The present work is an integral part of a research project that aims to develop a novel green ironmaking process based on the direct gaseous reduction of iron oxide fine concentrate in a flash process. The ultimate goal of this new process is to significantly reduce CO₂ emissions, energy consumption, and environmental pollution in the steel industry.⁷ Hydrogen, natural gas and coal gas are the proposed reducing agents in that new process. To date, only a few studies have been performed to measure FeO activities under gas atmospheres containing H₂O and H₂.^8–10 Thus, as part of the development of this new green ironmaking project, there was a need to investigate thoroughly the behaviour of FeO in H₂/H₂O and CO/CO₂/H₂/H₂O atmospheres, corresponding to an oxygen partial pressure (pO₂) range of 10⁻¹⁰–10⁻⁹ atm. In addition, the activity of FeO under CO/CO₂ was investigated for comparison. The major slag components were CaO, MgO, SiO₂, Al₂O₃ and FeO. The FeO activity coefficient was studied in the temperature range 1550–1650°C encompassing the expected operating temperatures in the proposed process.

There are two main methods to determine FeO activity. The first method is the thermodynamic relationships and EMF (electromotive force) measurements. In this method, FeO activity is measured using an oxygen sensor and thermodynamic relationship between Fe_(l)–FeO_(slag) expressed as follows¹¹ (1)

Using the EMF reading from an oxygen sensor, the equilibrium pO₂ can be calculated by the Nernst equation.¹² This technique was adopted by Liu et al.,^12,13 Ogura et al.,¹⁴ Iwase et al.,¹⁵ and Hamm et al. ¹⁶

The second method is the thermodynamic equilibrium technique with chemical analysis. This technique is the most common one used to measure a _FeO in which the slag sample is equilibrated with liquid iron or solid iron at a fixed temperature and under a stable atmosphere. In this technique, different principles could be employed to calculate the FeO or Fe_tO activity in the slags. These principles are listed as follows.

Slag–metal equilibrium technique

This technique is based on the following reactions (2) (3) (4) where a _FeO is FeO activity in the slag sample; (wt-%O) is analysed oxygen content in liquid iron in equilibrium with molten slag under consideration, equilibrium (2); is analysed oxygen content in liquid iron in equilibrium with pure liquid FeO, equilibrium (3).

This technique was adopted by Winkler and Chipman¹⁷ and Bishop et al.¹⁸ The technique was further modified to become more accurate by replacing the weight per cent oxygen in equation (4) with the activity (by multiplying both the numerator and denominator by the corresponding Henrian activity coefficients). Furthermore, the interaction between other impurities and alloying elements was considered.¹²

Alternatively, molten slags contained in an iron crucible would be brought into equilibrium under a CO/CO₂ or H₂/H₂O atmosphere. The FeO activity can be obtained from the expression (5) where pO₂ is the equilibrium oxygen partial pressure for a mixture of solid Fe+FeO in liquid slag; is the equilibrium oxygen partial pressure for a mixture of solid Fe+pure liquid FeO.

This technique was used by Schuhmann and Ensio.¹⁹ This technique requires a relatively complex experimental setup and long duration.¹⁵

Gas–slag–metal equilibrium technique

A different method to calculate the FeO activity considering the gas–slag–metal equilibria uses reaction (1), for which the equilibrium constant is given by (6) where x _FeO and γ_FeO are FeO mole fraction and Raultian activity coefficient in the slag. This technique was employed in this study. was calculated using HSC 5·11 (Outokumpu Oyj, Riihitontuntie 7, Finland).

Minor elements distribution techniques

Alternatively, FeO activity can be obtained from sulphide capacity or phosphate capacity.¹² Using the following equation from Bell’s work²⁰ (7)

Also, using a relation obtained by Lee and Fruehan,²¹ FeO activity can be obtained from the phosphorus distribution and phosphate capacity data.

FeO activity has been studied under different atmospheres. Basu et al.,³ Hamm et al.,¹⁶ Liu et al.,^12,13 Morales and Fruehan²² and Park and Lee²³ used argon atmosphere. More commonly, a larger number of researchers^4,24–26 determined FeO activity by controlling pO₂ through the CO/CO₂ equilibrium. Among the few studies performed to determine the free energy of formation of FeO using H₂/H₂O mixtures are that of Chipman⁸ who used the equilibrium between liquid electrolytic iron and H₂/H₂O gas mixtures. Also, Chipman and Marshall²⁷ studied the equilibria between solid iron and both liquid and solid wüstite (separately) under H₂/H₂O gas mixtures. To the best of the authors’ knowledge, there were no available data on the FeO activity coefficient under H₂/H₂O gas mixtures in ironmaking slags.

In this study, the following mass concentration ratio was used as slag basicity index (8)

The non-stoichiometry of ferrous oxide observed by Darken and Gurry²⁸ was found to be of negligible significance by Deo and Boom.²⁹ They suggested that the free energy of formation of ‘FeO’ and that of Fe_xO may be considered as the same. Thus, this convention was adopted in the present work.

Experimental details

Mixtures of dry powders of SiO₂, Al₂O₃ and MgO were pre-melted at 1600°C for homogeneity. The pre-melted slag was ground to less than 40 μm particle size and mixed for 36 h in a tumbler mixer (Bioengineering Inc., Cambridge, MA, USA). This was done to achieve homogeneity of the slag powder and reduce time to reach equilibrium in the subsequent experiments. Then the synthetic slag was stored in a desiccator to be used in the experiments. A horizontal electrical resistant furnace, shown in Fig. 1, heated by MoSi₂ heating elements with an alumina reaction tube (8 cm OD, 7 cm ID, 120 cm length) was used in the experimental set-up, which was equipped with a water vapour generator system. Temperature was controlled and monitored inside the tube by two B-type thermocouples (Pt6%Rh/Pt30%Rh): One was connected to a 708P temperature controller (MTI Corporation, Richmond, CA, USA) with an advanced PID adjustment to control the power supplied to the heating elements in the furnace to a temperature accuracy of ±1°C. The other thermocouple monitored the experimental temperature near the sample, indicating a constant temperature within 0·2–0·5°C. The gas flowrates were controlled by mass flow controllers (AALBORG, Orangeburg, NY, USA) with an accuracy of ±0·1, ±2, ±1·5, ±1·5 mL min⁻¹ for SO₂, H₂, CO and CO₂ gases, respectively. Water was injected as a liquid using a MASTERFLEX digital peristaltic pump drive (Cole-Parmer Instrument, Vernon Hills, IL, USA), which provided flowrates from 0·001 to 3400 mL min⁻¹ (using a set of 10 different diameter tubing) with the same brand pump head with an accuracy of ±0·1% of the flowrate.

1: fume hood; 2: valve; 3 water cooling jacket; 4 alumina shield; 5 NaOH scrubber; 6 MoSi₂ heating elements; 7 alumina sample holder; 8 alumina gutter; 9 heating tape; 10 B-type thermocouple

The initial slag composition other than FeO was chosen to be close to that in the blast furnace. The temperature and oxygen partial pressure ranges were chosen to be 1550–1650°C and 10⁻¹⁰–10⁻⁹ atm, respectively, which were wide enough to encompass the expected operating temperature of the new process based on the hydrogen or natural gas reduction of iron ore concentrate.

The samples were prepared as 2·5 and 1·0 g of slag mixed with 2·5 and 1·0 g of iron powder, respectively, for experiments under H₂/H₂O and the comparison experiments under the three gas atmospheres. They were mixed well to reduce the time to reach equilibrium in magnesia crucibles (1·8 cm OD, 4 cm height, 0·25 cm wall thickness) supplied by Ozark Technical Ceramics, Inc. (Webb City, MO, USA). The furnace was heated to the target temperature under a flow of N₂. At the target temperature, a four-sample alumina holder was placed in the even temperature zone of the furnace, which was measured accurately prior to the actual experiment by measuring the temperature profile inside the tube over its entire length. Then, N₂ was switched to the experimental gas mixture. From preliminary experiments,^30,31 it was confirmed that no measurable change in FeO content in the slag occurred after 4 h. To assure three-phase (gas–slag–metal) equilibria, 10 and 15 h were chosen in this work. After equilibrium, the system was purged with ultra-high purity N₂ for 5 min. The sample holder was quickly pulled out of the furnace, and the samples were quenched in ice bath or cold water. The crucible itself together with the sample was crushed, and iron was separated from the slag. Then, the slag was finely ground.

The composition of the slag phase was analysed by inductively coupled plasma optical emission spectroscopy. Prior to the analysis, the samples were digested in closed Savillex microwavable vessels. The chemical analysis results are listed in Tables 1 and 2, respectively. The equilibrium partial pressures of gases calculated by HSC 5·11 (Outokumpu Oyj, Riihitontuntie 7, Finland) are presented in Tables 3–5. Reproducibility of the experiments was confirmed by the consistency of the results of repeated experiments under the same conditions, as well as the reproducibility of the analysis method.^30,31 The experimental accuracy was within ±10%. Other details of the experimental procedure have been discussed elsewhere.^30–32

Table 1.

Chemical analysis of samples under H₂/H₂O atmosphere

Sample	T/°C	pO₂	CaO/SiO₂	FeO/wt-%	MgO/wt-%	CaO/wt-%	Al₂O₃/wt-%	SiO₂/wt-%	X _FeO	γ_FeO	a _FeO
S5	1550	7·0×10⁻¹⁰	0·8	21·6	19·2	20·9	8·7	28·3	0·21	2·17	0·45
S6	1550	7·0×10⁻¹⁰	1	19·9	20·2	23·0	8·0	33·0	0·18	2·54	0·45
S7	1550	7·0×10⁻¹⁰	1·2	22·3	15·2	29·8	9·5	26·4	0·21	2·16	0·45
S8	1550	7·0×10⁻¹⁰	1·4	20·7	10·6	32·1	9·4	23·7	0·21	2·15	0·45
S9	1550	1·0×10⁻¹⁰	0·8	11·5	22·4	24·9	8·2	32·3	0·10	1·75	0·17
S10	1550	1·0×10⁻¹⁰	1	12·0	18·7	29·0	9·7	29·9	0·10	1·64	0·17
S11	1550	1·0×10⁻¹⁰	1·2	12·4	17·0	30·9	10·7	29·0	0·11	1·58	0·17
S12	1550	1·0×10⁻¹⁰	1·4	14·2	14·5	34·7	9·8	31·0	0·12	1·41	0·17
S13	1550	2·0×10⁻¹⁰	0·8	12·2	22·5	25·1	9·4	32·6	0·10	2·34	0·24
S14	1550	2·0×10⁻¹⁰	1	12·6	16·9	27·4	8·8	39·7	0·10	2·31	0·24
S15	1550	2·0×10⁻¹⁰	1·2	18·7	15·4	34·7	10·2	16·4	0·18	1·33	0·24
S16	1550	2·0×10⁻¹⁰	1·4	15·3	16·3	32·4	10·4	28·4	0·13	1·80	0·24
S17	1600	5·0×10⁻⁹	0·8	41·1	15·1	15·8	6·2	19·5	0·49	1·57	0·77
S18	1600	5·0×10⁻⁹	1	34·6	13·6	20·4	7·4	24·4	0·37	2·06	0·77
S19	1600	5·0×10⁻⁹	1·2	42·8	11·3	18·1	6·3	17·0	0·54	1·41	0·77
S20	1600	5·0×10⁻⁹	1·4	52·8	12·4	14·9	4·9	12·2	0·74	1·03	0·77
S21	1600	2·0×10⁻⁹	0·8	20·4	22·6	20·7	7·6	27·7	0·19	2·60	0·49
S21	1600	2·0×10⁻⁹	0·8	20·3	22·6	20·7	8·6	27·6	0·18	2·62	0·49
S22	1600	2·0×10⁻⁹	1	27·1	15·7	22·7	8·2	23·5	0·28	1·73	0·49
S22	1600	2·0×10⁻⁹	1	27·0	17·6	22·6	8·2	23·2	0·28	1·76	0·49
S23	1600	2·0×10⁻⁹	1·2	29·7	14·2	26·2	8·8	22·0	0·30	1·62	0·49
S23	1600	2·0×10⁻⁹	1·2	29·7	14·2	23·7	8·3	22·0	0·32	1·52	0·49
S24	1600	2·0×10⁻⁹	1·4	31·1	14·7	24·9	7·9	21·7	0·31	1·55	0·49
S25	1600	3·0×10⁻¹⁰	0·8	8·7	25·5	24·3	9·7	36·0	0·07	2·77	0·19
S26	1600	3·0×10⁻¹⁰	1	9·3	21·2	27·1	9·8	31·5	0·08	2·42	0·19
S27	1600	3·0×10⁻¹⁰	1·2	11·8	18·7	29·1	10·1	28·2	0·10	1·84	0·19
S28	1600	3·0×10⁻¹⁰	1·4	13·2	19·4	31·0	9·9	23·9	0·11	1·64	0·19
S29	1600	5·0×10⁻¹⁰	0·8	7·1	23·9	23·9	9·6	31·5	0·06	4·09	0·24
S30	1600	5·0×10⁻¹⁰	1	14·7	20·9	25·5	9·6	28·9	0·13	1·88	0·24
S31	1600	5·0×10⁻¹⁰	1·2	14·9	25·3	25·2	8·9	23·6	0·13	1·90	0·24
S37	1650	4·0×10⁻⁹	0·8	32·6	20·5	17·8	7·6	23·5	0·34	1·33	0·45
S38	1650	4·0×10⁻⁹	1	47·5	14·2	13·5	5·5	16·3	0·63	0·72	0·45

Table 2.

Calculated gas partial pressures at experimental temperatures and P _T = 0·85 atm* using CO/CO₂/H₂O/H₂ gas mixtures (1 atm = 101·3 kPa)

	Temperature/°C	Partial pressure/atm	Sample designation	Partial pressure/atm	Sample designation
pO₂/atm	1550	1·9×10⁻⁹	R10–R12	1·6×10⁻¹⁰	R19–R21
	1600	4·8×10⁻⁹	R13–R15	3·6×10⁻¹⁰	R22–R24
	1630	8·0×10⁻⁹	R16–R18	6·1×10⁻¹⁰	R25–R27
pCO/atm	1550	2·4×10⁻¹	R10–R12	2·7×10⁻¹	R19–R21
	1600	2·4×10⁻¹	R13–R15	2·7×10⁻¹	R22–R24
	1630	2·4×10⁻¹	R16–R18	2·7×10⁻¹	R25–R27
pCO₂/atm	1550	4·4×10⁻²	R10–R12	1·4×10⁻²	R19–R21
	1600	4·2×10⁻²	R13–R15	1·4×10⁻²	R22–R24
	1630	4·2×10⁻²	R16–R18	1·4×10⁻²	R25–R27
pH₂O/atm	1550	2·4×10⁻¹	R10–R12	1·0×10⁻¹	R19–R21
	1600	2·4×10⁻¹	R13–R15	1·0×10⁻¹	R22–R24
	1630	2·4×10⁻¹	R16–R18	1·0×10⁻¹	R25–R27
pH₂/atm	1550	3·3×10⁻¹	R10–R12	4·7×10⁻¹	R19–R21
	1600	3·3×10⁻¹	R13–R15	4·7×10⁻¹	R22–R24
	1630	3·2×10⁻¹	R16–R18	4·7×10⁻¹	R25–R27
Flowrates/mL min⁻¹†	H₂	90	R10–R18	128	R19–R27
	H₂O	0·054		0·022
	CO	69		77
	CO₂	13		5

*Atmospheric pressure at Salt Lake City.

†Flowrates are calculated at 25°C and 0·85 atm (atmospheric pressure at Salt Lake City).

Table 3.

Calculated gas partial pressures at experimental temperatures and P _T = 0·85 atm* using H₂/H₂O/SO₂ gas mixtures

	Temperature/°C	Partial pressure/atm	Sample designation	Partial pressure/atm	Sample designation
pO₂	1550	2·2×10⁻⁹	S1–S4	7·0×10⁻¹⁰	S5–S8
	1600	4·8×10⁻⁹	S17–S20	1·8×10⁻⁹	S21–S24
	1650	…	…	4·2×10⁻⁹	S37–S38
pH₂	1550	4·7×10⁻¹⁰	S1–S4	1·2×10⁻³	S5–S8
	1600	4·1×10⁻⁴	S17–S20	1·3×10⁻³	S21–S24
	1650	4·3×10⁻⁴	…	1·3×10⁻³	S37–S38
pS₂	1550	1·4×10⁻³	S1–S4	1·0×10⁻³	S5–S8
	1600	1·7×10⁻³	S17–S20	1·3×10⁻³	S21–S24
	1650	2·0×10⁻³	…	1·6×10⁻³	S37–S38
pH₂O	1550	3·6×10⁻¹	S1–S4	2·6×10⁻¹	S5–S8
	1600	3·6×10⁻¹	S17–S20	2·6×10⁻¹	S21–S24
	1650	3·6×10⁻¹	…	2·6×10⁻¹	S37–S38
Flowrates/mL min⁻¹†	H₂	125		149
	H₂O (liq.)	0·047		0·032
pO₂	1550	2·4×10⁻¹⁰	S13–S16	1·1×10⁻¹⁰	S9–S12
	1600	5·3×10⁻¹⁰	S29–S32	2·7×10⁻¹⁰	S25–S28
	1650	…	…	…	…
pH₂	1550	1·3×10⁻³	S13–S16	1·3×10⁻³	S9–S12
	1600	1·3×10⁻³	S29–S32	1·4×10⁻³	S25–S28
	1650	1·4×10⁻³	…	1·4×10⁻³	…
pS₂	1550	8·3×10⁻⁴	S13–S16	7·3×10⁻⁴	S9–S12
	1600	1·0×10⁻³	S29–S32	9·5×10⁻⁴	S25–S28
	1650	1·3×10⁻³	…	1·2×10⁻³	…
pH₂O	1550	1·7×10⁻¹	S13–S16	1·3×10⁻¹	S9–S12
	1600	1·7×10⁻¹	S29–S32	1·3×10⁻¹	S25–S28
	1650	1·7×10⁻¹	…	1·3×10⁻¹	…
Flowrates/mL min⁻¹†	H₂	170		181
	H₂O (liq.)	0·018		0·011

*Atmospheric pressure at Salt Lake City.

†Flowrates are calculated at 25°C and 0·85 atm (atmospheric pressure at Salt Lake City).

Table 4.

Calculated gas partial pressures at experimental temperatures and P _T = 0·85 atm* using CO/CO₂ gas mixtures

	Temperature/°C	Partial pressure/atm	Sample designation	Partial pressure/atm	Sample designation
pO₂	1550	3·49×10⁻⁸	R37–R39	1·37×10⁻¹⁰	R46–R48
	1600	9·31×10⁻⁸	R40–R42	3·67×10⁻¹⁰	R49–R51
	1630	1·63×10⁻⁷	R43–R45	6·44×10⁻¹⁰	R52–R54
pCO	1550	4·78×10⁻¹	R37–R39	8·10×10⁻¹	R46–R48
	1600	4·78×10⁻¹	R40–R42	8·10×10⁻¹	R49–R51
	1630	4·78×10⁻¹	R43–R45	8·10×10⁻¹	R52–R54
pCO₂	1550	3·72×10⁻¹	R37–R39	3·96×10⁻²	R46–R48
	1600	3·72×10⁻¹	R40–R42	3·96×10⁻²	R49–R51
	1630	3·72×10⁻¹	R43–R45	3·96×10⁻²	R52–R54
Flowrates/mL min⁻¹†	CO	103	R37–R45	66	R46–R54
Flowrates/mL min⁻¹†	CO₂	30	R37–R45	7

*Atmospheric pressure at Salt Lake City.

†Flowrates are calculated at 25°C and 0·85 atm (atmospheric pressure at Salt Lake City).

Table 5.

Calculated gas partial pressures at the experimental temperatures and P_T = 0·85 atm* using H₂/H₂O gas mixtures

	Temperature/°C	Partial pressure/atm	Sample designation	Partial pressure/atm	Sample designation
pO₂	1550	3·49×10⁻⁸	R64–R66	1·39×10⁻¹⁰	R73–R75
	1600	8·46×10⁻⁸	R67–R69	3·35×10⁻¹⁰	R76–R78
pH₂O	1550	6·43×10⁻¹	R64–R66	1·39×10⁻¹	R73–R75
	1600	6·43×10⁻¹	R67–R69	1·39×10⁻¹	R76–R78
pH₂	1550	2·06×10⁻¹	R64–R66	7·10×10⁻¹	R73–R75
	1600	2·06×10⁻¹	R67–R69	7·10×10⁻¹	R76–R78
Flowrates/mL min⁻¹†	H₂	68	R64–R72	97	R73–R81
Flowrates/mL min⁻¹†	H₂O	0·04		0·017

*Atmospheric pressure at Salt Lake City.

†Flow rates are calculated at 25°C and 0·85 atm (atmospheric pressure at Salt Lake City).

Results and discussion

The first part of this paper is devoted to the discussion of the effects of different slag parameters such as basicity, FeO concentration and MgO on γ_FeO under H₂/H₂O, which were calculated using equation (6). In addition, various models to estimate γ_FeO were evaluated to predict γ_FeO under an H₂/H₂O atmosphere. An empirical formula was developed to correlate γ_FeO under the experimental conditions adopted in this work. The second part of this paper compares the effect of gas composition on γ_FeO. A separate set of experiments was carefully carried out under the same conditions but different gas atmospheres (CO/CO₂/H₂/H₂O, CO/CO₂ and H₂/H₂). It is noted that although CO/CO₂ mixtures in the similar pO₂ range were tested for comparison, this does not closely represent the blast furnace conditions in which the 4–5 wt-%C³³ in iron in the hearth makes pO₂ at the interface between molten iron and the slag much lower (10⁻¹⁵–10⁻¹⁴ atm)^11,34–36 than the 10⁻¹⁰ to 10⁻⁹ atm range expected in the new flash ironmaking process. Thus, the FeO content in the blast furnace slag is expected to be accordingly lower at around 1%.

Effect of X _FeO on its activity a _FeO

Figure 2 presents the variation of a _FeO with X _FeO obtained in this work. It is apparent that FeO exhibits a positive deviation from ideality for all slag compositions investigated. Similar behaviour was reported by several authors,^13,37,38 whereas Basu et al. ³ observed the positive deviation only up to X _FeO = 0·4 as shown in Fig. 3.

Variation of activity of FeO with mole fraction under H₂/H₂O atmospheres in temperature range 1550 to 1650°C for slags with CaO/SiO₂ of 0·8 to 1·4

Variation of activity of FeO with mole fraction in complex slags reported by others^1–4

The results of the present work have been presented in Fig. 3 along with those of the earlier work^3,13,37,38 for comparison. Positive deviation of the activity of FeO was observed by these workers. Turkdogan and Pearson³⁸ had reported positive deviation of a _FeO in CaO–MgO–MnO–FeO–SiO₂ slags at all FeO concentrations with the exception of negative deviation only when X _SiO2 was less than 0·008. Ichise and Iwase³⁹ observed negative deviation in a few cases at low iron oxide concentrations (X _FeO<0·1), whereas Wrampelmeyer et al. ⁴⁰ observed some incidents of negative deviation of FeO at intermediate concentrations (X _FeO = 0·25 to 0·4). In contrast, Chipman⁸ reported no positive deviation at all, but a small negative deviation of a _FeO at X _FeO exceeding 0·4.

Effect of basicity on γ_FeO

Figures 4–7 show the variation of γ_FeO with basicity, defined as (wt-% CaO/wt-% SiO₂), equation (8), for different pO₂, corresponding to different FeO concentrations, and temperatures. Other more comprehensive basicity indices were attempted but the results were too scattered to draw a conclusion. The current work used slag composition resembling ironmaking conditions and thus the basicity was in the range of 0·8–1·4. It is notable that γ_FeO showed a small decrease with an increase in basicity over the studied range at 1550°C, as shown in Fig. 4, whereas at 1600 and 1650°C γ_FeO decreased significantly with an increase in basicity, as shown in Figs. 5 and 6.

Effect of basicity on γ_FeO at 1550°C and different gas compositions under H₂/H₂O

Effect of basicity on γ_FeO at 1600°C and different gas compositions under H₂/H₂O

Effect of basicity on γ_FeO at 1650°C and pO₂ = 4×10⁻⁹ atm under H₂/H₂O

Effect of basicity on γ_FeO at 1550–1600 and 1630°C under CO/CO₂/H₂/H₂O and CO/CO₂, respectively

Basu et al.,³ in their experiments under an Ar atmosphere in the temperature range 1600 to 1650°C, reported that the change in basicity as well as temperature has only a negligible effect on the activity coefficient of FeO for any given range of FeO concentrations. In their work, basicity was varied in the range of 1·2 to 3·5 resembling the conditions in a basic oxygen furnace. Over this wider range of basicity, only a moderate decrease in γ_FeO with increasing basicity was observed. In the present study, smaller increments in basicity were used, 0·2, over a narrower range where a noticeable decrease in γ_FeO with increasing basicity was observed, especially at 1600 and 1650°C, as shown in Figs. 5 and 6.

Kishimoto et al. ³⁷ also reported a similar trend to that of Basu et al. ³ under an Ar atmosphere. They concluded that the increase in basicity over the range X _CaO/X _SiO2 = 2·0 to 6·0, at any level of FeO concentration, caused only a marginal decrease in the activity coefficient of FeO. Bishop et al. ¹⁸ reported a small decrease in a_FeO when the basicity, defined as the mole % ratio (CaO+MgO+MnO)/(SiO₂+PO_2·5+Al₂O₃), increased beyond 2, for Fe_tO concentrations within the range of 10 to 60 mol.-%. It is seen that all the above-mentioned authors reported smaller effects of basicity on γ_FeO than that in the current work over a wider basicity range.

Effect of MgO concentrations on γ_FeO

Figure 8 shows the effect of X _MgO on log γ_FeO over different ranges of FeO concentrations and at different temperatures under H₂/H₂O. It can be seen that γ_FeO increases with X _MgO in the investigated slag compositions, regardless of the temperature, for a slag with small X _FeO (≤0·1); whereas MgO has a negligible effect on γ_FeO in slags with higher X _FeO. Park and Lee²³ concluded that MgO has a negligible influence on γ_FeO for X _FeO>0·04 whereas an increase in X _MgO decreases γ_FeO for slags with lower X _FeO in the CaO–Al₂O₃–MgO_sat.–SiO₂–Fe_tO–MnO–P₂O₅ system. Basu et al. ³ showed that γ_FeO was unaffected by the change in MgO content regardless of X _FeO in the CaO–MgO_sat.–SiO₂–FeO–MnO–P₂O₅ slag.

Effect of X _MgO on log γ_FeO at different FeO concentrations and temperatures under H₂/H₂O

Thermodynamic and mathematical models of γ_FeO

In this section, an assessment of five models, available in the literature, for predicting γ_FeO in multicomponent slags was made with the experimentally obtained γ_FeO. These models will be referred to as Ohta and Suito’s model,⁴¹ Tao’s model,⁴² Yan’s model,⁴³ Park and Lee’s model²³ and Yang’s model.⁶

Ohta and Suito’s model⁴¹ is an empirical model developed based on experimental data at 1600°C under deoxidised Ar atmosphere. They used MgO, Al₂O₃, SiO₂ and CaO as the only variables that affect γ_FeO. Ohta and Suito⁴¹ used multiple regression analysis to obtain this empirical model without using any thermodynamic basis. Thus, the validity of this model is confined to the ranges of slag compositions and conditions used in their work. The model predicts values that are quite different from the measured values in this work. Thus, it was excluded from the evaluation of the experimental data. Tao’s model⁴² is only applicable to CaO–SiO₂–FeO ternary slag system, which is quite different from the slag of the current study. Moreover, Tao’s model is too complicated, many parameters of which need to be measured from binary slag systems. As a result, Tao’s model was not evaluated further.

Yan’s model⁴³ was derived by quadratic formalism based on the regular solution model, which is expressed as follows (9)

Park and Lee’s model²³ is also based on the regular solution model, as follows (10)

In equations (9) and (10), x _i is the mole fraction of slag component i, T is absolute temperature, and R is the universal gas constant (J mol K⁻¹).

Yang et al. ⁶ developed their model based on the ion and molecule coexistence theory (IMCT) to quantify γ_FeO. This model is composed of a set of non-linear equations to be solved using relevant thermodynamic data to predict the equilibrium X _FeO and a _FeO accordingly. The predicted γ_FeO using these three models was compared to the observed γ_FeO from this work, it was found that the coefficient of determination (r ²)<0·5 that implies a poor ability of these models in calculating γ_FeO for the studied slags under H₂/H₂O. This is attributed to the fact that none of the aforementioned models considered the effect of gas atmosphere on γ_FeO. It is worth noting that the range of the obtained γ_FeO values under H₂/H₂O was comparable to those ranges reported in the literature under Ar or CO/CO₂ atmospheres,^{3,4,13,23,37,44} as summarised in Table 6.

Table 6.

Comparison of measured values of FeO activity coefficients in complex slags under different gas atmospheres

Reference	X _FeO	γ_FeO	T/°C	Atmosphere
Nassaralla et al. ⁴⁴	0·006–0·014	4·2–9·3	1500–1550	Ar
Park and Lee²³	0·009–0·052	1·8–2·5	1600	Ar
Liu et al. ¹³	0·005–0·037	2·2–4·7	1550	Ar
Basu et al. ³	0·051–0·444	0·803–4·698	1600–1650	Ar
Kishimoto et al. ³⁷	0·013–0·719	1·318–2·967	1450	Ar
Henao and Itagaki⁴	0·13–0·51	0·3–1·92	1400	CO/CO₂
Current study	0·056–0·483	0·483–4·086	1550–1650	H₂/H₂O

An empirical correlation was formulated using linear regression to adequately predict γ_FeO under H₂/H₂O in the studied slags. As mentioned earlier the temperature was found to have a negligible effect on γ_FeO under the investigated conditions. Thus, it was excluded from the empirical correlation (11)

Figure 9 used equation (11) to compare the measured values of γ_FeO to the predicted results. It is concluded that γ_FeO can be adequately estimated using this correlation. Incorporating additional compositional variables did not yield any significant improvement.

Comparison of measured γ_FeO versus calculated using equation (6)–(11) under H₂/H₂O atmosphere in temperature range 1550 to 1650°C and wt-% CaO/wt-% SiO₂ range of 0·8 to 1·4 under pO₂ range of 10⁻¹⁰ to 10⁻⁹ atm

Effect of gas atmosphere on γ_FeO

The major reactions involving FeO in the investigated slags under the experimental conditions are as follows, according to IMCT⁶ (12) (13)

It is worth noting that according to IMCT, SiO₂ and Al₂O₃ exist as simple oxides, SiO₂ and Al₂O₃, equilibrated with complex molecules such as Fe₂SiO₄ and FeAl₂O₄. Iron oxide exists in the form of ion couples Fe²⁺ and O²⁻, equilibrated with complex compounds such as Fe₂SiO₄ and FeAl₂O₄. We will focus on reaction (12) for the following reasons:

(i)

no previous research reported the interaction of H₂O with aluminate

(ii)

the higher silica content, at least 2–4 times higher, compared with alumina

(iii)

no change was observed in spectroscopic analyses that could be related to alumina species as a result of changing the gas composition. Therefore, FeO would be stabilised in the slag mostly in the form of Fe₂SiO₄, i.e. Fe²⁺ forms stable silicates with the least polymerised silicate anion, .

Previous research^45–51 has shown that water has a significant solubility in slag in the temperature range of 1500–1650°C. In weakly to moderately basic slags, water dissolution may involve the following reaction⁵¹ (14) where the dotted and the solid lines represent coordination and covalent (stronger than coordination bond) bonds, respectively. It is noted that reaction (14) competes with reaction (12). Most FeO in the slag is combined with silicate ions according to reaction (12) and, therefore, any reaction that might perturb equilibrium (12) will affect γ_FeO. FeO is a network modifier, which breaks the bridging oxygen bonds; hence, its dissolution implies the depolymerisation of the silicate anions. Based on the spectroscopic results, the H₂O-containing atmospheres tend to stabilise the polymerised silicate anions, i.e. the higher Qⁿ (n: the number of bridging oxygen; the higher the n, the more the polymerisation). Therefore, it is expected that H₂O increases γ_FeO, resulting in lower FeO wt-% in the slag, whereas CO₂ decreases it. Figures 10–12 show the variation of X _FeO with pO₂ and temperature for various CaO/SiO₂ ratios. It is seen that the slags under CO/CO₂ have higher FeO contents following this order: CO/CO₂>CO/CO₂/H₂/H₂O>H₂/H₂O. In support of that observation, Figs. 13–15 show a reverse order of the atmospheres with respect to γ_FeO. Moreover, Fig. 16 indicates that H₂O-containg atmospheres have higher γ_FeO than CO/CO₂ atmospheres based on the average of all experimental points. Thus, the hitherto unknown role H₂O plays in affecting γ_FeO in the slag compared with other gases was evidenced. That has also been substantiated by the effect of pH₂O on γ_FeO at constant pO₂ and CaO/SiO₂, as shown in Figs. 17 and 18. It is noted that γ_FeO increases with pH₂O within the range tested. All the slags under the three gas atmospheres exhibited positive deviation of a _FeO from ideality, as Fig. 19 shows.

10.

Effect of gas atmosphere on X _FeO at various pO₂ and temperature values for slags with wt-% CaO/wt-% SiO₂ of 0·8 and pH₂O of 0 and 0·1 atm in CO/CO₂ and H₂/H₂O/CO/CO₂, respectively

11.

Effect of gas atmosphere on X _FeO at various pO₂ and temperature values for slags with wt-% CaO/wt-% SiO₂ of 1·0 and pH₂O of 0, 0·1 and 1·0 atm in CO/CO₂, H₂/H₂O/CO/CO₂, and H₂/H₂O/CO, respectively

12.

Effect of gas atmosphere on X _FeO at various pO₂ and temperature values for slags with wt-% CaO/wt-% SiO₂ of 1·2 and pH₂O of 0, 0·1 and 1·0 atm in CO/CO₂, H₂/H₂O/CO/CO₂, and H₂/H₂O/CO, respectively

13.

Effect of gas atmosphere on γ_FeO at various pO₂ and temperature values for slags with wt-% CaO/wt-% SiO₂ of 0·8 and pH₂O of 0 and 0·1 atm in CO/CO₂ and H₂/H₂O/CO/CO₂, respectively

14.

Effect of gas atmosphere on γ_FeO at various pO₂ and temperature values for slags with wt-% CaO/wt-% SiO₂ of 1·0 and pH₂O of 0, 0·1 and 1·0 atm in CO/CO₂, H₂/H₂O/CO/CO₂ and H₂/H₂O/CO, respectively

15.

Effect of gas atmosphere on γ_FeO at various pO₂ and temperature values for slags with wt-% CaO/wt-% SiO₂ of 1·2 and pH₂O of 0, 0·1 and 1·0 atm in CO/CO₂, H₂/H₂O/CO/CO₂ and H₂/H₂O/CO, respectively

16.

Comparison of average values of γ_FeO of three gas atmospheres at experimental temperatures for slags with wt-% CaO/wt-% SiO₂ of 1·2 and pH₂O of 0, 0·1 and 1·0 atm in CO/CO₂, H₂/H₂O/CO/CO₂ and H₂/H₂O/CO, respectively, and pO₂ range of 10⁻¹⁰ to 10⁻⁹ atm. Error bars are experimental standard deviation and labels are average X _FeO

17.

Effect of pH₂O on γ_FeO wt-% CaO/wt-% SiO₂ of 0·8 at different temperatures and pO₂

18.

Effect of pH₂O on γ_FeO wt-% CaO/wt-% SiO₂ of 1·2 at different temperatures and pO₂

19.

Variation of activity of FeO with FeO mole fraction under different gas mixtures in temperature range of 1550–1630°C, wt-% CaO/wt-% SiO₂ range of 0·8 to 1·2, and pH₂O ranges of 0, 0·1–0·2, and 0·1–0·2 atm in CO/CO₂, H₂/H₂O/CO/CO₂ and H₂/H₂O, respectively, and pO₂ range of 10⁻¹⁰–10⁻⁹ atm

For the slag compositions investigated, MgO-saturated CaO–FeO–Al₂O₃–SiO₂–MnO(0·2–0·8 wt-%)–P₂O₅(0·1–0·9 wt-%), in the temperature range 1550–1600°C, with CaO/SiO₂ range 0·8 to 1·2, and under pO₂ = 2×10⁻¹⁰–2×10⁻⁹ atm, the average FeO wt-% was found to be 10 (9·5–10·5), 11 (9·8–13) and 16 (14·7–16·9) under H₂/H₂O, CO/CO₂/H₂/H₂O, and CO/CO₂, respectively. These gas mixtures resemble H₂ reductant, reformed natural or coal gas (NG/CG), and CO/CO₂ conditions, respectively. The average content of FeO under H₂ and NG/CG was presented relative to that under CO/CO₂ conditions in Fig. 20. The effect of gas atmosphere on slag chemistry was further explored in this laboratory via spectroscopic techniques.⁵² In addition, it was found that water in the gas atmosphere affected the distribution of the elements between slag and molten iron such as Mn.⁵³ Moreover, water was determined to affect the interaction between the furnace ceramic lining and the slag.⁵⁴

20.

Influence of type of reductant gas on FeO content in slag. Labels show average FeO wt-%

It is noted that although CO/CO₂ mixtures in the similar pO₂ range were tested for comparison, these conditions are not close to the blast furnace in which the FeO content in the slag is of the order of 1% because of the presence of carbon in molten iron. The 4–5 wt-% C in iron in the hearth of the blast furnace makes pO₂ at the interface between molten iron and the slag equal to 10⁻¹⁵–10⁻¹⁴ atm.^11,34–36

Conclusions

Water was found to decrease the content of FeO by increasing its activity coefficient in the slags investigated, MgO-saturated CaO–FeO–Al₂O₃–SiO₂–MnO(0·2–0·8 wt-%)–P₂O₅(0·1–0·9 wt-%), in the temperature range 1550–1600°C, with wt-% CaO/wt-% SiO₂ of 0·8 to 1·2, and under pO₂ = 2×10⁻¹⁰–2×10⁻⁹ atm. The average FeO content was found to be 10, 11 and 16 wt-% under H₂/H₂O (H₂), CO/CO₂/H₂/H₂O (reformed natural/coal gas), and CO/CO₂, respectively. In other words, the slags under H₂ and reformed natural/coal gas contain, respectively, 37 and 31% less FeO wt-% than under CO/CO₂.

The following correlation for γ_FeO in slags under H₂/H₂O was formulated using linear regression

Footnotes

Acknowledgements

The authors thank Adirek Janwong for help with the analytical work using inductively coupled plasma. In addition, the authors would like thank the staff of Micron Microscopy Core at the University of Utah, especially Dr Brian Van Devener, for the valuable help with characterisation work. The authors acknowledge the financial support from American Iron and Steel Institute (AISI) through a Research Service Agreement with the University of Utah under AISI’s CO₂ Breakthrough Program. This material also contains results of work supported by the US Department of Energy under Award Number DE-EE0005751.

References

Lee

Min

Jung

and Yi

: ISIJ Int., 2004, 44, (8), 1283–1290.

Nikolic

Henao

Hayes

and Jak

: Metall. Trans. B, 2008, 39B, (2), 189–199.

Basu

Lahiri

and Seetharaman

: Metall. Trans. B, 2008, 39B, (3), 447–456.

Henao

and Itagaki

: Metall. Trans. B, 2007, 38B, (5), 769–780.

Henao

and Itagaki

: ‘Sohn International Symposium; advanced processing of metals and materials volume 1: thermo and physicochemical principles: non-ferrous high-temperature processing’, 755–766; 2006, Somerset, NJ, Wiley.

Yang

Shi

Zhang

and Zhang

: Steel Res. Int., 2012, 83, 244–258.

Sohn

Choi

Zhang

and Ramos

: Iron Steel Technol. (AIST Trans.), 2009, 6, (8), 158–165.

Chipman

: J. Am. Chem. Soc., 1933, 55, 3131–3139.

Chipman

and Fetters

: Trans. Am. Soc. Met., 1941, 29, 953–966.

10.

Bodsworth

: J. Iron Steel Inst., 1959, 193, 13–24.

11.

Suito

and Inoue

: Trans. Iron Steel Inst. Jpn, 1984, 24, 301–307.

12.

Liu

: ‘Measurement of FeO activity in bath smelting slags’, PhD thesis, Carnegie Mellon University, Pittsburgh, PA, USA, 1996.

13.

Liu

Fruehan

Morales

and Ozturk

: Metall. Trans. B, 2001, 32B, (1), 31–36.

14.

Ogura

Fujiwara

Mochizuki

Kawamoto

Oishi

and Iwase

: Metall. Trans. B, 1992, 23B, 459–466.

15.

Iwase

Uchida

Tomii

McLean

and Katogi

: High Temp. Mater. P, 2001, 20P, (3–4), 269–277.

16.

Hamm

Cundiff

McCauley

and Frigm

: ‘The use of slag iron oxide activity measurements for steel making process control’, Steelmak. Conf. Proc., 1994, 77, 156.

17.

Winkler

and Chipman

: Trans. AIME, 1946, 167, 111–133.

18.

Bishop

Jr Grant

and Chipman

: Trans. Metall. Soc. AIME, 1958, 212, 890–892.

19.

Schuhmann

Jr and Ensio

: J. Met., 1951, 3, 401–411.

20.

Venkatradi

and Bell

: J. Iron Steel Inst., 1969, 207, 1110–1113.

21.

Lee

and Fruehan

: Ironmak. Steelmak., 2005, 32, 503–508.

22.

Morales

and Fruehan

: Metall. Trans. B, 1997, 28B, 1111–1118.

23.

Park

and Lee

: Steelmak. Conf. Proc., 1996, 79, 165–171.

24.

Fujiwara

Ichise

Kitou

and Matsui

: Metall. Trans. B, 1999, 30B, (3), 419–427.

25.

Yang

and Belton

: Metall. Trans. B, 1998, 29B, (4), 837–845.

26.

Fujiwara

Ichise

Kitou

and Matsui

: Metall. Trans. B, 1997, 28B, (2), 243–250.

27.

Chipman

and Marshall

: J. Am. Chem. Soc., 1940, 62, (2), 299–305.

28.

Darken

and Gurry

: J. Am. Chem. Soc., 1945, 67, 1398–1412.

29.

Deo

and Boom

: ‘Fundamentals of steelmaking metallurgy’; 1993, New York, Prentice Hall International.

30.

Mohassab-Ahmed

Sohn

and Kim

: Ind. Eng. Chem. Res., 2012, 51, (20), 7028–7034.

31.

Mohassab-Ahmed

Sohn

and Kim

: Ind. Eng. Chem. Res., 2012, 51, (9), 3639–3645.

32.

Mohassab Ahmed

: ‘Phase equilibria between iron and slag in CO/CO₂/H₂/H₂O atmospheres relevant to a novel flash ironmaking technology’, PhD thesis, The University of Utah, Salt Lake City, UT, USA, 2013.

33.

Habashi

: ‘Handbook of extractive metallurgy’, 53–55; 1997, Weinheim, Wiley-VCH.

34.

The Japan Society for the Promotion of Science : ‘Steelmaking data sourcebook’, 1988, New York, Gordon and Breach Science Publishers.

35.

Shankar

: Ironmak. Steelmak., 2006, 33, (5), 413–418.

36.

Peacey

and Davenport

: ‘The iron blast furnace theory and practice’; 1979, New York, Pergamon Press.

37.

Kishimoto

Hasegawa

Ohnuki

Sawai

and Iwase

: Steel Res. Int., 2005, 76, 341–347.

38.

Turkdogan

and Pearson

: J. Iron Steel Inst., 1953, 173, 217–223.

39.

Ichise

and Iwase

: Trans. Metall. Soc. AIME, 1984, 517–531.

40.

Wrampelmeyer

Dimitrov

and Janke

: Steel Res., 1989, 60, 539–549.

41.

Ohta

and Suito

: Metall. Trans. B, 1998, 29B, (1), 119–129.

42.

Ping Tao

: Metall. Trans. B, 2006, 37B, (6), 1091–1097.

43.

Ban-Ya

: ISIJ Int., 1993, 33, (1), 2–11.

44.

Nassaralla

Sarma

Morales

and Myers

: Proc. E.T. Turkdogan Symp., Pittsburgh, PA, USA, May 1994, TMS, 61–71.

45.

Brandberg

and Sichen

: Metall. Trans. B, 2006, 37B, (3), 389–393.

46.

Rosenqvist

: ‘Principles of extractive metallurgy’; 2004, Trondheim, Tapir Forlag.

47.

Orrling

and Cramb

: Metall. Trans. B, 2000, 31B, (2), 403–406.

48.

and Kim

: Steel Res., 2000, 71, 15–21.

49.

Leekes

Nowack

and Schlegelmilch

: Steel Res., 1988, 59, 406–416.

50.

Sosinsky

Maeda

and McLean

: Metall. Trans. B, 1985, 16B, (1), 61–66.

51.

Iguchi

Ban-Ya

and Fuwa

: Trans. Iron Steel Inst. Jpn, 1969, 9, 189–195.

52.

Mohassab-Ahmed

and Sohn

: JOM, 2013, 65, (11), 1559–1565.

53.

Mohassab-Ahmed

and Sohn

: Steel Res. Int., 2013, 85, 875–884.

54.

Mohassab-Ahmed

Sohn

and Zhu

: Ironmak. Steelmak., 2014, 41, (8), 575–582.

Effect of water vapour content in H 2 –H 2 O–CO–CO 2 mixtures on activity of iron oxide in slags relevant to novel flash ironmaking technology

Abstract

Introduction

Slag–metal equilibrium technique

Gas–slag–metal equilibrium technique

Minor elements distribution techniques

Experimental details

Results and discussion

Effect of X FeO on its activity a FeO

Effect of basicity on γFeO

Effect of MgO concentrations on γFeO

Thermodynamic and mathematical models of γFeO

Effect of gas atmosphere on γFeO

Conclusions

Footnotes

Acknowledgements

References

Effect of X _FeO on its activity a _FeO

Effect of basicity on γ_FeO

Effect of MgO concentrations on γ_FeO

Thermodynamic and mathematical models of γ_FeO

Effect of gas atmosphere on γ_FeO