Optimisation study and affecting mechanism of CaO/SiO 2 and MgO on viscous behaviours of titanium-bearing blast furnace slag

Abstract

To improve the metallurgical properties of titanium-bearing blast furnace slag, the orthogonal test was carried out and analysed by the multi-index synthetic weighted scoring method. The optimal slag compositions of titanium-bearing BF slag were CaO/SiO₂ 1.25, MgO 12.00 wt-%, Al₂O₃ 11.00 wt-% and TiO₂ 10.00 wt-%. In addition, the CaO/SiO₂ and MgO content had important effects on the break-point temperature, viscosity and activation energy of viscous flow. With an increasing CaO/SiO₂ from 1.10 to 1.30, the break-point temperature increased, the slag viscosity and the activation energy of viscous flow decreased. As the MgO content grew from 6.00 to 12.00 wt-%, the viscosity decreased, while the break-point temperature and the activation energy of viscous flow first reduced and then increased. Besides, the deeply relevant mechanism was researched by Factsage 7.0, Fourier transform infrared spectroscopy and X-ray diffraction.

Introduction

Vanadium-titanium magnetite (abbr. V–Ti magnetite) is a kind of multi-elements-coexistent mineral and has a very high comprehensive utilisation value. In China, V–Ti magnetite has abundant reserves in Panzhihua district and Chengde district [1–6]. Currently, blast furnace (BF) process is the predominant route to extract metal, such as Fe, V from V–Ti magnetite, while Ti is existed in the BF slag as an oxide [7,8]. BF slag is the main byproduct in iron-making process. The practical experience suggests that the proper metallurgical properties of BF slag benefit the slag/metal separation, improve desulphurisation capacity of slag and reduce the lining erosion. It facilitates the stability and productivity in BF operation [9,10]. The metallurgical properties of slag such as flowability and thermostability depend on the chemical composition and temperature [11,12]. Therefore, it is feasible to improve the metallurgical properties by optimising the slag composition for maximising the productivity and maintaining the iron-making operation.

Recently, the viscous behaviours of different BF slags have been widely studied by domestic and foreign scholars. Qiu et al. [12] investigated the high-alumina BF slag and obtained the optimal MgO/Al₂O₃ ratio (0.3–0.4) to save production costs, based on the multi-objective particle swarm optimisation algorithm. Liao et al. [7] investigated the effects of TiO₂ and CaO/SiO₂ on viscous behaviours of CaO–SiO₂–7.0 wt-%MgO–TiO₂–12.0 wt-% Al₂O₃ slags, and noted that the increasing TiO₂ contents (15.0–30.0 wt-%) and CaO/SiO₂ (0.5–0.9) all decreased the viscosity and activation energy. Sohn and Min elucidated the effects of various components such as Al₂O₃, MgO, alkali oxides and TiO₂ on the viscosity of the multi-component calcium–silicate-based slag system [13]. However, few researches perform an optimal slag composition by using the multi-index synthetic weighted scoring method.

The multi-index synthetic weighted scoring method is one of scientific evaluation method used to optimise parameters for multi-index system [14]. It contains three parts: (1) the weight of each index is determined by its significance for the whole test; (2) the multi-index results are converted to a final single result that is synthetic scoring value; (3) the synthetic scoring value is analysed by the normal single-index analysis method. Here, the synthetic weight of each index is a key point. To ensure a true and effective result, both the subjective cognition (experience) on the index significance by tester and the objective information from test results are considered in this method. Finally, the optimisation model of synthetic scoring value is established based on the optimisation theory, and its exact solution is given. So the analysis is reasonable and credible, and it has been widely applied in lots of fields such as agriculture, pharmacy and manufacturing industry [15–17].

In the present study, the base titanium-bearing BF slag was taken from the Chengde Steel in China. For further improving the viscous behaviours of BF slag, the orthogonal optimisation test was carried out and analysed by multi-index synthetic weighted scoring method. The optimal slag composition was obtained. In addition, the effects of CaO/SiO₂ and MgO content on the viscous behaviours of BF slag were further investigated.

Experiment methods

Materials

In this study, all experiments samples were synthesised by adding analytical reagent oxides of CaO, SiO₂, MgO, Al₂O₃ and TiO₂ into the base BF slag. The chemical compositions of the base BF slag, obtained from the Chengde Steel in China, are listed in Table 1. To improve the accuracy of experiments, the base BF slag and analytical reagent oxides were first roasted and then mixed adequately. The mixture samples were subsequently pre-melted with a molybdenum crucible in the induction furnace under the Ar atmosphere. After obtaining the homogeneous slag, the samples were crushed into powder and used for the viscosity measurements.

Table 1.

Chemical composition of base BF slag.

CaO (wt-%)	SiO₂(wt-%)	MgO (wt-%)	Al₂O₃(wt-%)	TiO₂(wt-%)	CaO/SiO₂(–)
36.02	28.51	9.09	13.16	8.22	1.26

Apparatus and procedure

In the present study, the continuous viscosity of CaO–SiO₂–Al₂O₃–MgO–TiO₂ BF slag was measured through the rotating spindle method by RTW-10 melt physical property comprehensive testing instrument with a digital viscometer developed by Northeastern University. The schematic diagram of the experiment device for viscosity measurements is shown in Figure 1, the device consists of a heating system, rotating system, a measuring system, a controlling system and a gas system. The heating system, rotating system and a measuring system are controlled by the controlling system which concludes a computer and controller. The gas cylinder, which is controlled by a flow meter, provides Ar to protect the sample, spindle and apparatus.

Figure 1.

Schematic diagram of experimental apparatus (a) and dimensions of crucible and spindle (b).

First, 140 g sample was put into a molybdenum crucible. Then, the crucible was placed in the induction furnace and heated to 1500°C. After the slag sample was completely melted, its amount was about 40 mm in depth. The molten slag was maintained for 30 min to homogenise the chemical compositions and stabilise the temperature. Thereafter, the molybdenum rotating spindle was slowly immersed in the molten slag and kept at a distance of 10 mm above the crucible bottom to measure the viscosity at a speed of 200 rpm. The measurement results were recorded by the controlling system. When the viscosity reached about 3.5 Pa s, the viscosity measurement process was finished. Then, the slag sample was reheated to 1500°C and the molybdenum spindle was carefully taken out the molten slag. Meanwhile, the molten slag was taken out from the induction furnace and quenched under the Ar atmosphere. The viscosity measurements were performed during the cooling cycle and protected under an Ar atmosphere at 1.5 L min⁻¹.

Orthogonal optimisation of titanium-bearing BF slag

Orthogonal test scheme

To acquire a slag system with better flowability and thermostability, the orthogonal optimisation scheme is carried out to adjust the slag composition. As shown in Table 2, the test contains CaO/SiO₂ 1.10–1.30, MgO 6.00 wt-%–12.00 wt-%, Al₂O₃ 11.00 wt-%–16.00 wt-%, TiO₂ 5.00 wt-%–12.00 wt-% and there are four factors and five levels.

Table 2.

The slag composition for an orthogonal optimiSation test.

No.	CaO (wt-%)	SiO₂(wt-%)	A	B	C	D
No.	CaO (wt-%)	SiO₂(wt-%)	CaO/SiO₂(–)	MgO (wt-%)	Al₂O₃(wt-%)	TiO₂(wt-%)
1	39.36	35.78	1.10	6.00	11.00	5.00
2	37.79	34.35	1.10	7.00	12.00	6.00
3	35.17	31.97	1.10	9.00	13.00	8.00
4	32.55	29.59	1.10	11.00	14.00	10.00
5	29.93	27.21	1.10	12.00	16.00	12.00
6	38.05	33.09	1.15	6.00	12.00	8.00
7	35.91	31.23	1.15	7.00	13.00	10.00
8	33.24	28.90	1.15	9.00	14.00	12.00
9	34.84	30.30	1.15	11.00	16.00	5.00
10	36.45	31.69	1.15	12.00	11.00	6.00
11	36.08	30.06	1.20	6.00	13.00	12.00
12	38.81	32.34	1.20	7.00	14.00	5.00
13	36.08	30.06	1.20	9.00	16.00	6.00
14	36.62	30.52	1.20	11.00	11.00	8.00
15	34.44	28.70	1.20	12.00	12.00	10.00
16	39.52	31.62	1.25	6.00	14.00	6.00
17	36.75	29.40	1.25	7.00	16.00	8.00
18	37.30	29.84	1.25	9.00	11.00	10.00
19	34.52	27.62	1.25	11.00	12.00	12.00
20	37.30	29.84	1.25	12.00	13.00	5.00
21	36.82	28.32	1.30	6.00	16.00	10.00
22	37.95	29.19	1.30	7.00	11.00	12.00
23	40.21	30.93	1.30	9.00	12.00	5.00
24	37.95	29.19	1.30	11.00	13.00	6.00
25	35.69	27.45	1.30	12.00	14.00	8.00

Orthogonal range analysis of single index

Through the viscosity measurements in this work, the viscosity–temperature curves (η–t curves) of different slags are shown in Figure 2.

Figure 2.

η–t curves of orthogonal test.

Each η–t curve has an obvious turning point, which is usually corresponded to the tangency point of 45° line and η–t curve [18]. The turning point temperature was defined as the break-point temperature T _Br. At T _Br, a substantial variation in viscosity exists during the continuous cooling process of slag. In one sense, T _Br is the boundary between better flowability and worse flowability of slag, as a critical indicator of viscous behaviours.

The viscosity reflects the flowability of BF slag and higher viscosity can result in slower metal/slag separation and lower BF productivity. So it is expected to reasonably decrease. Specifically, the viscosity at T _Br (η _TBr) and the viscosity at 1500°C (η _1500°C) are as two indexes in the orthogonal optimizations of viscous behaviours of slag.

The activation energy of viscosity flow (E_η ) represents the frictional resistance for viscous flow, and variations in E_η can suggest a change in the structure of the molten slag or more directly a change in the cohesive flow units comprising the slag structure. E_η also can represents the temperature dependence of viscosity and reflect the thermostability of slags. Lower E_η indicates a better flowability and thermostability of slag and benefits the BF smooth operation [19]. The detailed calculation process of E_η was discussed in the following sections.

Therefore, T _Br, η_T _Br, η _1500°C and E_η were as four evaluation indexes in optimization of viscous behaviours of slag. Their values are listed in Table 3. Generally, to acquire appropriate fluidity, the optimal indexes should be reasonably lowered.

Table 3.

The values of indexes for the orthogonal test.

No.	A	B	C	D	Results
No.	CaO/SiO₂	MgO	Al₂O₃	TiO₂	T _Br (°C)	η _TBr (Pa s)	η _1500°C (Pa s)	E_η (kJ)
1	1.10	6.00	11.00	5.00	1337	1.003	0.284	187.02
2	1.10	7.00	12.00	6.00	1332	1.169	0.299	198.91
3	1.10	9.00	13.00	8.00	1333	1.045	0.274	215.72
4	1.10	11.00	14.00	10.00	1318	1.022	0.214	239.00
5	1.10	12.00	16.00	12.00	1334	0.928	0.225	299.63
6	1.15	6.00	12.00	8.00	1310	1.367	0.271	190.65
7	1.15	7.00	13.00	10.00	1331	1.228	0.290	202.73
8	1.15	9.00	14.00	12.00	1333	1.009	0.228	223.76
9	1.15	11.00	16.00	5.00	1357	0.795	0.212	222.29
10	1.15	12.00	11.00	6.00	1336	0.794	0.188	178.93
11	1.20	6.00	13.00	12.00	1321	1.070	0.295	192.28
12	1.20	7.00	14.00	5.00	1360	0.982	0.331	178.96
13	1.20	9.00	16.00	6.00	1345	0.912	0.314	202.34
14	1.20	11.00	11.00	8.00	1347	1.014	0.331	176.35
15	1.20	12.00	12.00	10.00	1331	0.841	0.234	190.60
16	1.25	6.00	14.00	6.00	1363	0.820	0.303	180.74
17	1.25	7.00	16.00	8.00	1364	0.745	0.231	199.37
18	1.25	9.00	11.00	10.00	1329	0.706	0.182	174.99
19	1.25	11.00	12.00	12.00	1329	0.706	0.151	187.37
20	1.25	12.00	13.00	5.00	1354	0.598	0.201	172.63
21	1.30	6.00	16.00	10.00	1373	0.656	0.200	186.08
22	1.30	7.00	11.00	12.00	1354	0.725	0.212	171.61
23	1.30	9.00	12.00	5.00	1390	0.626	0.204	176.00
24	1.30	11.00	13.00	6.00	1372	0.598	0.216	174.11
25	1.30	12.00	14.00	8.00	1341	0.706	0.245	178.67

Orthogonal range analyses are listed in Table 4. Through above orthogonal range analyses of a single index, the affecting sequence of chemical compositions on each evaluation index and the respective structure of chemical compositions are obtained: (1) The order of slag compositions affecting the T _Br is CaO/SiO₂, TiO₂, Al₂O₃, MgO. To obtain lower T _Br, the optimal slag compositions are CaO/SiO₂ 1.10, MgO 12.00 wt-%, Al₂O₃ 12.00 wt-%, TiO₂ 12.00 wt-%. (2) The sequence of slag compositions affecting η _TBr is CaO/SiO₂, MgO, TiO₂, Al₂O₃. To obtain lower η_T _Br, the optimal slag slag compositions are CaO/SiO₂ 1.30, MgO 12.00 wt-%, Al₂O₃ 16.00 wt-%, TiO₂ 5.00 wt-%. (3) The order of slag compositions affecting the η _1500°C is CaO/SiO₂, MgO, TiO₂ and Al₂O₃. To obtain lower η _1500°C, the optimal slag compositions are CaO/SiO₂ 1.25, MgO 12.00 wt-%, Al₂O₃ 12.00 wt-%, TiO₂ 12.00 wt-%. (4) The sequence of slag compositions affecting E_η is CaO/SiO₂, Al₂O₃, TiO₂ and MgO. To obtain lower E_η , the optimised slag compositions are CaO/SiO₂ 1.30, MgO 6.00 wt-%, Al₂O₃ 11.00 wt-%, TiO₂ 6.00 wt-%.

Table 4.

Range difference analysis of orthogonal test.

Level		1	2	3	4	5	Range	Significance	Optimum
T _Br (°C)	A	1331	1333	1341	1348	1366	35	A > D > C > B	A₁
	B	1341	1348	1346	1345	1339	9		B₅
	C	1341	1338	1342	1343	1355	16		C₂
	D	1360	1350	1339	1336	1334	25		D₅
η _TBr (Pa s)	A	1.033	1.0390	0.964	0.715	0.662	0.376	A > B > D > C	A₅
	B	0.983	0.970	0.860	0.827	0.773	0.210		B₅
	C	0.848	0.942	0.908	0.908	0.807	0.135		C₅
	D	0.801	0.859	0.975	0.891	0.888	0.175		D₁
η _1500°C (Pa s)	A	0.259	0.238	0.301	0.214	0.215	0.087	A > B > D > C	A₄
	B	0.271	0.273	0.240	0.225	0.219	0.054		B₅
	C	0.239	0.232	0.255	0.264	0.236	0.032		C₂
	D	0.246	0.264	0.270	0.224	0.222	0.048		D₅
E_η (kJ)	A	228.056	203.672	188.106	183.020	177.294	50.762	A > C > D > B	A₅
	B	187.354	190.316	198.562	199.824	204.092	16.738		B₁
	C	177.780	188.706	191.494	200.226	221.942	44.162		C₁
	D	187.380	187.006	192.152	198.680	214.930	27.924		D₂

Multi-index synthetic weighted scoring method

To determine the final optimal slag composition, the multi-index synthetic weighted scoring method is carried out in the orthogonal test analyses [8,15]. Considering the importance of subjective cognizance and making the most use of objective test results, the synthetic weight value of indexes should accomplish the unification of the subject and object, then the scoring is scientific. The detailed processes are in the following section.

Determination of standardise evaluation matrix

The orthogonal test scheme including 25 groups experiments is marked as and T _Br, η _TBr, η _1500°C and E_η are evaluation indexes: . The makes up the standardise evaluation matrix and the exact data can be obtained from Table 3. According the Equation (1), is obtained. To unify the tendency requirement of each index and remove its incommensurability, the evaluation matrix should be standardised. (1) (2) For unifying, the index’s magnitude order and eliminate the dimension, the z_ij is conducted as the Equation (2). is the final standardise evaluation matrix. The results are as follows:

Synthetic weight determination of each index

First, based on the previous experience, the expert investigating method is used to obtain the subjective weight for each index [15,20]. The subjective weight of evaluation indexes is assigned α ₁= 0.25, α ₂ = 0.25, α ₃ = 0.25, α ₄ = 0.25, respectively. Therefore, the matrix of the subjective weight is , .

The objective weight is written as .

Combining the formulas below, the objective weight of each index could be calculated. (3) (4) (5)

Notes, , when . Matrixes about p_ij , h_j and β_j are calculated in the following. To achieve a unity of subjective favouritism and objective information properly, the optimisation decision model is established as Equation (6). Where, μ (0 < μ < 1) is the inclination coefficient which indicates the preference between subject weights and object weights and is assigned with μ = 0.5. Here, the values of subject weights, object weights and inclination coefficient are substituted into Equation (7). The final synthetic weights of each index are gained as ω = (0.202, 0.253, 0.194, 0.351) ^T . (6) (7)

Calculation of synthetic weighted score

The final synthetic weights of indexes are substituted into Equation (8). The synthetic weighted score for 25 group experiments is calculated as f_i = (38.699,47.787,45.871, 41.259,60.008,43.489,49.558,41.937,38.805,19.002,39.504,46.652,45.149,43.716,27.446,39.542,34.680,12.612,12.670,16.750,27.030,21.840,27.990,23.306). (8)

Single-index test analysis evaluation

Subsequently, the optimisation on viscous behaviours of BF slag is investigated by the single-index test analysis evaluation and the results are listed in Table 5. Through the whole multi-index synthetic weighted scoring method, the optimal results of BF slags are obtained as follows: the sequence of chemical components affecting viscous behaviours of BF slags is CaO/SiO₂, Al₂O₃, MgO, TiO₂; the structure of optimal slag composition is: CaO/SiO₂ 1.25, MgO 12.00 wt-%, Al₂O₃ 11.00 wt-%, TiO₂ 10.00 wt-%.

Table 5.

Test results and conclusions of orthogonal test.

Level	A	B	C	D	Result
k ₁	46.73	37.65	27.17	33.78
k ₂	38.56	40.10	31.88	34.96	ω ₁ = 0.202, ω ₂ = 0.253, ω ₃ = 0.194, ω ₄ = 0.351
k ₃	40.49	34.71	35.00	38.24	Inclination coefficient μ = 0.500
k ₄	23.25	31.95	38.56	31.58	β = (0.25, 0.25, 0.25, 0.25)
k ₅	24.72	29.33	41.13	35.19	Significance A > C > B > D
R	23.47	10.78	13.96	6.66	Optimum A₄B₅C₁D₄

In addition, the effects of inclination coefficient μ and subjective weight α on the optimisation are discussed, respectively, in Tables 6 and 7. Under different inclination coefficient μ or subjective weight α, the sequence of slag components on viscous behaviours of BF slag remains barely changed: CaO/SiO₂, Al₂O₃, MgO, TiO₂, and the structure of slag components also is almost same: CaO/SiO₂ 1.25, MgO 12.00 wt-%, Al₂O₃ 11.00 wt-%, TiO₂ 10.00 wt-%.

Table 6.

The results of single-index evaluation under different inclination coefficient μ.

μ	Significance	Factors
μ	Significance	CaO/SiO₂ (–)	MgO (wt-%)	Al₂O₃ (wt-%)	TiO₂ (wt-%)
0	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.30	12.00	11.00	5.00
0.2	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.25	12.00	11.00	10.00
0.5	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.25	12.00	11.00	10.00
0.8	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.25	12.00	11.00	10.00
1	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.25	12.00	11.00	10.00

Table 7.

The results of single-index evaluation under different subjective weight α.

α				Significance
α ₁	α ₂	α ₃	α ₄	Significance	CaO/SiO₂ (–)	MgO (wt-%)	Al₂O₃ (wt-%)	TiO₂ (wt-%)
0.30	0.10	0.30	0.30	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.25	12.00	11.00	10.00
0.25	0.25	0.25	0.25	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.25	12.00	11.00	10.00
0.23	0.31	0.23	0.23	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.25	12.00	11.00	10.00
0.20	0.40	0.20	0.20	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.25	12.00	11.00	10.00
0.10	0.70	0.10	0.10	CaO/SiO₂ > Al₂O₃ > MgO > TiO₂	1.30	12.00	11.00	5.00

Comparison of the optimal slag and base BF slag

According the above analyses, the optimal slag composition is CaO/SiO₂ 1.25, MgO 12.00 wt-%, Al₂O₃ 11.00 wt-%, TiO₂ 10.00 wt-%. The metallurgical properties of optimal slag is analysed and verified through the viscosity experiment and theoretical analysis.

As shown in Figure 3 and Table 8, compared with the base BF slag, the evaluation indexes T _Br, η ₀, η _h and E_η of optimal slag are decreased and better. As a result, the flowability and thermostability are improved. Besides, the experimental result is verified in thermodynamics using Factsage 7.0 package, which is widely applied for metallurgical field. As shown in Figure 4, the optimised slag composition is located in the crystalline region of melilite and the liquidus temperature of optimised slag is lower than that of the base BF slag, so the optimised slag can be melted easily and the flowability is improved. In addition, the distributions of liquidus in the optimised slag crystalline region are sparser. Thus, the slag thermostability gets better.

Figure 3.

η–t curves and fitting results between lnη/T and 1/T of the base slag and optimal slag.

Figure 4.

The phase diagram of CaO–SiO₂–MgO–13.16 wt-%Al₂O₃–9.22 wt-%TiO₂ slag.

Table 8.

The metallurgical properties of the base slag and optimal slag.

slag	T _Br (°C)	η _TBr (Pa s)	η _h (Pa s)	E_η (kJ)
The base BF slag	1358	0.723	0.235	206.09
The optimal slag	1304	0.698	0.171	111.93

Good flowability benefits the slag/metal separation, improves desulphurization capacity of slag gas permeability and the vanadium yield. It also can reduce the lining erosion and prolong working life of BF. Good thermostability will benefit the smooth BF operation. According to the optimal slag composition, Chengde Steel adjusts the scheme and technical parameters for rational matching in BF process, and it could reduce the energy consumption and maximise the productivity of ironmaking.

Effects of CaO/SiO₂ and MgO on viscous behaviours of titanium-bearing BF slag

According to the industrial BF operation, CaO/SiO₂ and MgO strongly affect the viscous behaviours of slag. Therefore, the effects of CaO/SiO₂ and MgO on viscous behaviours of BF slags are deeply researched through viscosity measurements. The experiment scheme is shown as Table 9.

Table 9.

Slag compositions for viscosity measurements.

No.	Factors	CaO (wt-%)	SiO₂ (wt-%)	MgO (wt-%)	Al₂O₃ (wt-%)	TiO₂ (wt-%)	CaO/SiO₂ (–)
1	CaO/SiO₂	34.92	31.75	9.09	13.16	8.22	1.10
2		35.66	31.01	9.09	13.16	8.22	1.15
3		36.37	30.31	9.09	13.16	8.22	1.20
4		37.04	29.63	9.09	13.16	8.22	1.25
5		37.68	28.99	9.09	13.16	8.22	1.30
6	MgO	38.89	30.87	6.00	13.16	8.22	1.26
7		38.34	30.43	7.00	13.16	8.22	1.26
8		37.22	29.54	9.00	13.16	8.22	1.26
9		36.11	28.66	11.00	13.16	8.22	1.26
10		35.55	28.21	12.00	13.16	8.22	1.26

Effects of CaO/SiO₂ and MgO on the break-point temperature

Effect of CaO/SiO₂ on the break-point temperature

According to Figures 5(a) and 6(a), the T _Br increases obviously with an increase of CaO/SiO₂ from 1.10 to 1.30. As the related reports, the T _Br and liquidus temperature of slags have similar variation. Therefore, it is feasible to use the variation of liquidus temperature to investigate the change of T _Br [21,22].

Figure 5.

η–t curves of experimental slags with different CaO/SiO₂ (a) and MgO content (b).

Figure 6.

T _Br of experimental slags with different CaO/SiO₂ (a) and MgO content (b).

The X-ray diffraction (XRD) analysis results for different CaO/SiO₂ slags are shown in Figure 7. Melilite, CaTiO₃, CaO·Al₂O₃ and pyroxene are discovered mainly in slags, and melilite is the basic phase in different CaO/SiO₂ slags. The diffraction peak intensity of melilite increases slightly with an increasing CaO/SiO₂ from 1.10 to 1.30. The diffraction peak intensity of phases, containing CaTiO₃ and CaO·Al₂O₃, also increases. However, the diffraction peak intensity of pyroxene becomes weak. According to the previous literatures, the melting points of melilite, CaTiO₃, CaO·Al₂O₃ are higher than that of pyroxene [23,24]. As a result, the amount of low melting-point phase is relatively decreased. Finally, the liquidus temperature of BF slags increases and the T _Br also increases.

Figure 7.

XRD analysis for different CaO/SiO₂ slags.

The Factsage 7.0 is applied to calculate the phase diagram. As shown in Figure 8, the liquidus temperature increases with an increasing CaO/SiO₂. The liquidus temperatures of different CaO/SiO₂ slags are 1387, 1391, 1393, 1394 and 1401°C. Its change matches well with the variety of T _Br.

Figure 8.

Phase diagram of CaO–SiO₂–MgO–13.16 wt-% Al₂O₃–8.22 wt-% TiO₂ slag.

Effect of MgO on the break-point temperature

As shown in Figure 6(b), when the MgO content increases, the T _Br initially decreases and subsequently increases. To explain the change of T _Br, the XRD analyses results for different MgO content slags are shown in Figure 9. Melilite, CaTiO₃, MgO·Al₂O₃, CaO·MgO·SiO₂ and pyroxene are mainly discovered in the slags, and melilite are the basic phase. When the MgO content raises from 6.00 to 11.00 wt-%, the diffraction peak intensity of high melting-point phases (melilite, CaTiO₃ and MgO·Al₂O₃) decreases, but the diffraction peak intensity of low melting-point phases (pyroxene) increases. As a result, the sum of low melting-point phases increased relatively. Finally, the liquidus temperature decreases and the T _Br decreases. However, as the MgO content further increases to 12.00 wt-%, a high melting-point phase (CaO·MgO·SiO₂ 1498°C) suddenly appears and the diffraction peak intensity of low melting-point phases (pyroxene) decreases. The sum of high melting-point phases increases, so the liquidus temperature increases and the T _Br increases [25,26].

Figure 9.

XRD analysis for different MgO content slags.

In addition, according to Figure 8, the liquidus temperature calculated are 1428, 1418, 1396, 1391 and 1413°C, respectively. The change of calculated liquidus temperature with different MgO content roughly resembled with that of T _Br.

Effect of CaO/SiO₂ and MgO on viscosity

Effect of CaO/SiO₂ on viscosity

The effect of CaO/SiO₂ on the slag viscosity is observed in Figure 10(a). When the CaO/SiO₂ increases from 1.10 to 1.30, the viscosity decreases, similar to the Park’s and Zhang’s works [27,28].

Figure 10.

Effect of CaO/SiO₂ on the viscosity of slags at various temperatures (a) experiment results (b) calculated viscosity by Factsage 7.0.

The viscous behaviour of molten slag is mainly influenced by chemical composition, structure and temperature [29,30]. It is well known that the depolymerisation of slag structure results in a decrease of the viscosity. The molten slag has a large silicate network, which are combined through bridging oxygen (O⁰) [27]. Since the increasing CaO/SiO₂ is dissociated to provide more free oxygen icons (O²⁻), the (O²⁻) reacts with the bridged oxygen (O⁰) to form the non-bridging oxygen (O⁻). It can break the Si–O bonds in [SiO₄]⁴⁻ tetrahedral units. As a result, the complex silicate networks were depolymerised to small units, leading to a reduction in viscosity.

Figure 10(b) shows the effect of CaO/SiO₂ on the slag viscosity calculated by Factsage7.0. It is noticed that the calculated viscosity values also decreases with increasing the CaO/SiO₂. The change of calculated results is consistent with that of experimental results.

To gain the effects of CaO/SiO₂ on the slag structure and reveal the mechanism of viscosity changes, the experimental samples are researched by Fourier transform infrared spectroscopy analysis. The vibration region between approximately 750 and 1200 cm⁻¹ is identified as tetrahedral, and the vibration region from 590 to 750 cm⁻¹ is on behalf of tetrahedral. Besides, the vibration region from 400 to 590 cm⁻¹ approximately represents the Si–O–Al bonds [14,17].

The effect of CaO/SiO₂ on the slag structure is shown in Figure 11. The characteristic symmetric stretching vibration band for tetrahedral becomes weak, and the width of band is broadened and the depth of convoluted band becomes shallower, as increasing the CaO/SiO₂ from 1.10 to 1.30. It indicates that the distance between the tetrahedral units is decreased. Furthermore, the width of tetrahedral asymmetric stretching vibration band is also broadened, and the trough of tetrahedral asymmetric stretching vibration band dampened obviously. This suggests that the tetrahedral units in the slag are also depolymerised. Meanwhile, the depth of Si–O–Al bending vibration trough is decreased, and it indicates the linkages between Si/Al and O in the slag are decreased. Overall, the above-mentioned changes indicate that the complex anion groups are simplified to simpler structural units and the slag structure is depolymerised. This leads in the decrease of viscosity with increasing CaO/SiO₂ [29,30].

Figure 11.

FTIR for the experimental slags with different CaO/SiO₂.

Effect of MgO content on viscosity

As shown in Figure 12(a), when MgO content increases from 6.00 to 12.00 wt-%, the slag viscosity decreases apparently. The viscosity variation of slags is similar with that of slag studied by Kim et al. [31] and Feng et al. [9].

Figure 12.

Effect of MgO content on the viscosity of slags at various temperatures (a) experiment results (b) calculated viscosity by Factsage 7.0.

Similar to the study about CaO/SiO₂ on the slag viscosity, MgO behaves as a network-modifying oxide and causes that the , and Al–O–Si bending decreases. Generally speaking, MgO additions depolymerise the silicate and the alumino-silicate structures within the slag, which results in a lower viscosity [17].

Meanwhile, as shown in Figure 12(b), the slag viscosity with different MgO contents at various temperatures is calculated by Factsage 7.0. It can be seen that the calculated viscosity and the experimental results are roughly same with the trends.

Effects of CaO/SiO₂ and MgO on the activation energy of viscosity flow

There are many methods for calculating the activation energy. It can be expressed by the Weymann–Frenkel equation modified by Urbain, as shown in Equation (9) [32,33]. The present paper used the formula and the viscosity data of slags obtained to calculate E_η through the linear regression method. (9) where, η is the viscosity in Pa s, T is the temperature in K, A is the pre-exponential, R is gas constant, 8.314 J/(mol K) and E_η is activation energy of viscous flow in J mol⁻¹. Besides, E_η can not only reflect the flowability of slags, but also represent the temperature dependence of viscosity and reflect the thermostability of slags [21].

Finally, Equation (9) is given by taking the logarithm for Equation (10). It can be seen that ln(η/T) has a linear relationship with η/T. By the linear regression method, the E_η is obtained from the slop, and A is got from the intercept, based on the experimental data. (10) Figure 13 shows the fitting results between ln(η/T) and 1/T for all slags. Obviously, lnη has a good linear relationship with 1/T for all the investigated slags.

Figure 13.

Fitting results between lnη/T and 1/T for the experimental slags with different CaO/SiO₂ (a), MgO content (b).

As shown in the Figure 14, E_η decreases with an increasing CaO/SiO₂ from 1.10 to 1.30. The variation of E_η is roughly consistent with the change of viscosity. When the MgO content grows from 6.00 to 11.00 wt-%, the E_η decreases from 177.72 to 174.15 kJ mol⁻¹. When MgO content further increases from 11.00 to 12.00 wt-%, E_η increases to 178.49 kJ mol⁻¹.

Figure 14.

The variations of E_η for the experimental slags with different CaO/SiO₂ (a), MgO content (b).

The E_η is not only affected by the slag structure, but also affected by the liquidus temperature. When MgO content increases from 6.00 to 11.00 wt-%, the viscosity decreases. It indicates the simplification of slag structure [9]. Meanwhile, the low melting-point phase increases and the liquidus temperature reduces, leading the decrease of E_η . With MgO content increasing from 11.00 to 12.00 wt-%, the slag structure is further simplified, but the sum of high melting-point phases increases, which leads to the increase of liquidus temperature of BF slag. Therefore, the combination of two factors increases the E_η of slag with MgO content increased from 11.00 to 12.00 wt-%.

Conclusions

In the present study, an orthogonal optimisation test was carried out and analysed using the multi-index synthetic weighted scoring method. The optimal slag compositions were CaO/SiO₂ 1.25, MgO 12.00 wt-%, Al₂O₃ 11.00 wt-%, TiO₂ 10.00 wt-%, and the affecting sequence of components on the viscous behaviours of slag was CaO/SiO₂, Al₂O₃, MgO, TiO₂.

With an increase of CaO/SiO₂ from 1.10 to 1.30, the polymerisation degree of complex viscous units decreased. The slag viscosity and the activation energy of viscous flow E_η decreased. However, the liquidus temperature raised, and the break-point temperature (T _Br) increased.

When MgO content grew from 6.00 to 12.00 wt-%, the complex viscous units within slags were depolymerised. The slag viscosity decreased while the T _Br and E_η first reduced and then increased.

Footnotes

Disclosure statement

No potential conflict of interest was reported by the authors.

References

Feng

Chu

Tang

et al. Effect of CaO/SiO₂ and Al₂O₃ on viscous behaviors of the titanium-bearing blast furnace slag. Steel Res Int. 2016;87:1274–1283. doi: 10.1002/srin.201500355

Zhou

Yang

Jiang

et al. Influence of basicity on high-chromium vanadium-titanium magnetite sinter properties, productivity, and mineralogy. JOM. 2015;67:1203–1213. doi: 10.1007/s11837-015-1326-7

Sun

Zheng

Liao

et al. Effect of P₂O₅ addition on the viscosity and structure of titanium bearing blast furnace slags. ISIJ Int. 2014;54:1491–1497. doi: 10.2355/isijinternational.54.1491

Wang

Zhao

Chu

et al. Effect and function mechanism of sinter basicity on softening-melting behaviors of mixed burden made from chromium-bearing vanadium-titanium magnetite. J Cent South Univ. 2017;24:39–47. doi: 10.1007/s11771-017-3406-z

Kim

Matsuura

Tsukihashi

Effect of Al₂O₃ and CaO/SiO₂ on the viscosity of calcium-silicate–based slags containing 10 mass Pct MgO. Metall Mater Trans B. 2013;44:5. doi: 10.1007/s11663-012-9759-7

Sara

Vladimir

Sergy

et al. Investigation of thermal properties of blast furnace slag to improve process energy efficiency. J Clean Prod. 2017;149:137–145. doi: 10.1016/j.jclepro.2017.02.052

Liao

Wang

et al. Influence of TiO₂ and basicity on viscosity of Ti bearing slag. Ironmaking Steelmaking. 2012;39:133–139. doi: 10.1179/1743281211Y.0000000064

Zhou

Jiang

Yang

et al. Vanadium–titanium magnetite ore blend optimization for sinter strength based on iron ore basic sintering characteristics. Int J Miner Process. 2015;142:125–133. doi: 10.1016/j.minpro.2015.04.019

Feng

Chu

Tang

Effects of MgO and TiO₂ on the viscous behaviors and phase compositions of titanium-bearing slag. Int J Min Met Mater. 2016;23:868–880. doi: 10.1007/s12613-016-1302-4

10.

Wang

Sun

Sridhar

Effect of Al2O3 on the viscosity and structure of CaO-SiO₂-MgO-Al₂O₃-FetO slags. Metall Mater Trans B. 2015;46:537. doi: 10.1007/s11663-015-0303-4

11.

Kim

Sohn

Effect of CaF₂ and Li₂O additives on the viscosity of CaO–SiO₂–Na₂O slags. ISIJ Int. 2011;51:2.

12.

Qiu

Zhan

Jiang

et al. Optimisation of the MgO/Al₂O₃ ratio of high-alumina BF slag based on MOPSO algorithm. Ironmaking Steelmaking. 2017;149:137.

13.

Sohn

Min

DJ.

A review of the relationship between viscosity and the structure of calcium-silicate-based slags in ironmaking. Steel Res Int. 2012;83:611–630. doi: 10.1002/srin.201200040

14.

Tang

Chu

Feng

et al. Melting separation process of high chromium vanadium-bearing titanomagnetite metallized pellet and its optimization by multi-index synthetic weighted scoring method. ISIJ Int. 2017;57:1156–1165. doi: 10.2355/isijinternational.ISIJINT-2016-714

15.

Tao

JM.

New study on determining the weight of index in synthetic weighted mark method. Systems Eng Theor Pract. 2001;46:45.

16.

Wang

Jiao

Optimization of ethanol extraction technique for Schisandrae Fructus and Salviae Miltiorrhizae Radix by multi index comprehensive weighted score evaluation. Chinese Tradit Herbal Drug. 2015;7:998.

17.

Yao

Yue

CY.

Method for multiple attribute decision-making under risk based on synthetic weighting. Eng Electron. 2005;12:2047.

18.

Shu

Chou

et al. Effect of Al₂O₃/SiO₂ mass ratio on viscosity of CaO–Al₂O₃–SiO₂–CaF₂ slag. Ironmaking Steelmaking. 2015;42:155.

19.

Hsieh

Tong

LI.

Optimization of multiple quality responses involving qualitative and quantitative characteristics in IC manufacturing using neural networks. Comput Ind. 2001;46:1–12. doi: 10.1016/S0166-3615(01)00091-4

20.

Park

Min

Song

HS.

Amphoteric behavior of alumina in viscous flow and structure of CaO-SiO₂(-MgO)-Al₂O₃ slags. Metall Mater Trans B. 2004;35:269–275. doi: 10.1007/s11663-004-0028-2

21.

Gao

Wang

Hong

et al. Effects of basicity and MgO content on the viscosity of the SiO₂–CaO–MgO–9wt% Al₂O₃ slag system. Int J Min Met Mater. 2014;21:353–362. doi: 10.1007/s12613-014-0916-7

22.

Jin

Yang

HY.

Effect of the CaO/SiO₂ mass ratio and FeO content on the viscosity of CaO–SiO₂–“FeO”–12wt%ZnO–3wt%Al₂O₃ slags. Int J Min Met Mater. 2017;24:758.

23.

et al. Metallurgical properties of BF slag and slagging regime. Ironmaking Steelmaking. 2006;41:19. doi: 10.1179/1743281212Y.0000000095

24.

Zhang

Jiang

Xue

et al. Influence of MgO/Al₂O₃ ratio on viscosity of blast furnace slag with high Al₂O₃ content. Steel Res Int. 2016;87:88.

25.

Qin

Liu

Yang

YH.

Structure evolution of blast furnace slag with high Al₂O₃ content and 5 mass% TiO₂ via molecular dynamics simulation and Fourier transform infrared spectroscopy. Met Res Tec. 2018;115:1.

26.

Kim

Lee

Min

et al. Influence of MgO and Al₂O₃ contents on viscosity of blast furnace type slags containing FeO. ISIJ Int. 2004;44:1291–1297. doi: 10.2355/isijinternational.44.1291

27.

Zhang

Jahanshahi

Review and modeling of viscosity of silicate melts: Part I. Viscosity of binary and ternary silicates containing CaO, MgO, and MnO. Metall Mater Trans B. 1998;29:178.

28.

Park

J-Y

Kim

et al. Effect of TiO₂ on the viscosity and slag structure in blast furnace type slags. Steel Res Int. 2012;83:150–156. doi: 10.1002/srin.201100249

29.

Kim

Sohn

Min

DJ.

A study on the viscous behavior with K₂O additions in the CaO-SiO₂-Al₂O₃-MgO-K₂O quinary slag system. Steel Res Int. 2010;81:735–741. doi: 10.1002/srin.201000067

30.

Zeng

Sheng

MQ.

et al. Viscosity of low silica CaO–5MgO–Al₂O₃–SiO₂ slags. Ironmaking Steelmaking. 2015;41:489.

31.

Kim

Sohn

et al. The effect of MgO on the viscosity of the CaO-SiO₂-20wt%Al₂O₃–MgO slag system. Steel Res Int. 2010;81:262.

32.

Ren

Zhang

et al. A model for estimating the viscosity of blast furnace slags with optical basicity. Int J Min Met Mater. 2012;19:1089.

33.

Shanker

Gornerup

Lahiri

et al. Estimation of viscosity for blast furnace type slags. Ironmaking Steelmaking. 2007;34:479.

Optimisation study and affecting mechanism of CaO/SiO 2 and MgO on viscous behaviours of titanium-bearing blast furnace slag

Abstract

Introduction

Experiment methods

Materials

Chemical composition of base BF slag.

Apparatus and procedure

Orthogonal optimisation of titanium-bearing BF slag

Orthogonal test scheme

The slag composition for an orthogonal optimiSation test.

Orthogonal range analysis of single index

The values of indexes for the orthogonal test.

Range difference analysis of orthogonal test.

Multi-index synthetic weighted scoring method

Determination of standardise evaluation matrix

Synthetic weight determination of each index

Calculation of synthetic weighted score

Single-index test analysis evaluation

Test results and conclusions of orthogonal test.

The results of single-index evaluation under different inclination coefficient μ.

The results of single-index evaluation under different subjective weight α.

Comparison of the optimal slag and base BF slag

The metallurgical properties of the base slag and optimal slag.

Effects of CaO/SiO2 and MgO on viscous behaviours of titanium-bearing BF slag

Slag compositions for viscosity measurements.

Effects of CaO/SiO2 and MgO on the break-point temperature

Effect of CaO/SiO2 on the break-point temperature

Effect of MgO on the break-point temperature

Effect of CaO/SiO2 and MgO on viscosity

Effect of CaO/SiO2 on viscosity

Effect of MgO content on viscosity

Effects of CaO/SiO2 and MgO on the activation energy of viscosity flow

Conclusions

Footnotes

Disclosure statement

References

Effects of CaO/SiO₂ and MgO on viscous behaviours of titanium-bearing BF slag

Effects of CaO/SiO₂ and MgO on the break-point temperature

Effect of CaO/SiO₂ on the break-point temperature

Effect of CaO/SiO₂ and MgO on viscosity

Effect of CaO/SiO₂ on viscosity

Effects of CaO/SiO₂ and MgO on the activation energy of viscosity flow