Optimum condition for smelting high carbon ferromanganese

Abstract

In the present study, pilot plant experimental heats were designed and carried out to determine the optimum condition for smelting high carbon ferromanganese through investigating some parameters affecting the smelting process including Mn/Fe ratio of the blend, coke ratio, slag basicity and dolomite/limestone ratio of the flux. The results of pilot plant experimental heats showed that using Mn blend with high Mn/Fe ratio decreases the consumption of charging materials: Mn ores, coke and fluxing materials (limestone and dolomite). This means that lower amounts of Mn ores, coke and flux materials can be used for attaining the same output alloy weight by increasing the Mn/Fe ratio of the blend. Furthermore, the produced slag decreases as Mn/Fe ratio of the blend increases. The higher Mn/Fe ratio of the blend does not only reduce the input materials and the produced slag, but also improves the produced alloy quality by increasing the Mn percent and Mn/Fe ratio of the produced high carbon ferromanganese. For obtaining standard HCFeMn alloy containing minimum 75%Mn, it is necessary to use Mn blend with the Mn/Fe ratio of higher than 6·2. The added coke must be adjusted according to the material balance and stoichiometric to prevent the over-coke and minimise the highly endothermic ‘Boudouard reaction’ to attain the highest Mn recovery and metallic yield. Furthermore, the slag basicity should be optimised with flux addition to attain the highest Mn recovery and metallic yield. The optimum slag basicity for attaining the highest manganese recovery and metallic yield depends on the used formula. The slag basicity can be determined from the chemical composition of raw material mixture with taking into consideration that about 2% of silicon in the raw material mixture will be reduced and goes into metal phase. Much higher slag basicity is not recommended as this practice produces viscous slag accompanied with lower Mn recovery and metallic yield. The results also revealed the negative effect of increasing MgO/CaO ratio of slag on manganese recovery and metallic yield. The MgO/CaO ratio in the slag decreases by decreasing the dolomite/limestone ratio in the flux.

Introduction

Manganese is an essential alloying element in nearly all types of steel and is used to increase strength, toughness, hardness and hardenability. It is also used for deoxidation and desulphurisation of ferrous metals and alloys. In iron and steel industry, manganese is not added as a pure element but as a ferromanganese alloy.

Ferromanganese can be produced either in a blast furnace or an electric arc furnace from ores containing both manganese and iron oxides. The production of ferromanganese in blast furnace consumes a large amount of coke (1·8–2·0 tons per ton of melt) and manganese raw materials (2·5–3·0 tons per ton of melt) due to the specific conditions of the process and the considerable losses of manganese to slag and gas (15–20%). Electric furnaces offer high capacity and productivity. In addition, they produce minimum environmental impact, as the furnace can be readily sealed, and only a small quantity of off-gas is generated. Electric furnaces present the most favourable choice of smelting technology where high process temperatures in excess of about 1400°C are required.

Currently, high carbon ferromanganese is produced by carbothermic reduction in submerged arc furnaces fed with a cold charge of Mn ores, carbonaceous agent and fluxing materials. Blending of ores from different sources is the common practice in the HCFeMn smelting process, and coke is used as a reducing agent. Limestone and/or dolomite are used as flux materials. These basic fluxes are added to give the slag suitable chemical properties, smelting temperature and viscosity in order to secure good furnace operation and a high manganese recovery. The amount of added limestone and dolomite depends on the required CaO and MgO to attain the specific slag basicity.

The composition of ferromanganese slag is very important as it determines the slag/metal equilibrium relations including the MnO content of final slag, melting relation and slag viscosity.1

Slag basicity is conveniently used to describe equilibrium and smelting relations, slag properties and final slag composition in the high carbon ferromanganese process.2, 3 Slag basicity is usually defined as the mass ratio between the basic and acid oxides and expressed by some basicity number.2

Different formulas are used to express the slag basicity in the high carbon ferromanganese process.1 ^– 6 The simplest expression is the (CaO)/(SiO₂) ratio. The flux materials commonly used in the high carbon ferromanganese smelting process are limestone and dolomite. Therefore, the produced slag contains a significant amount of MgO, and the slag basicity is commonly described as (CaO+MgO)/(SiO₂). When the smelting process produces slag containing a significant amount of Al₂O₃, it is preferable to define the basicity as (CaO+MgO)/(SiO₂+Al₂O₃). If all minor basic oxides like BaO and K₂O are included, extended basicity formula is recommended as given in equation (1) (1) where Σbas is the sum of all basic oxides, mainly CaO, MgO, BaO and K₂O. This formula is simplified as given in equation (2)2 (2) The determination of appropriate slag composition is very important for a high rate of recovery of manganese, high metallic yield and smooth operation of smelting high carbon ferromanganese. The main factor increasing the manganese recovery is a rise in the activity of manganese oxide in the slag melt, which is determined by the slag composition. The basic oxides CaO and MgO existing in the slag raise the activity of MnO.7 On the other hand, much higher CaO and MgO contents result in higher slag viscosity which impairs the reduction of manganese. In addition, as the dissociations of both limestone and dolomite are endothermic reactions, higher additions of these fluxes result in higher energy consumption. Thus, it is of prime importance to determine the optimum slag basicity for attaining the highest manganese recovery and metallic yield.

Furthermore, the suitable Mn/Fe ratio of the blend and coke ratio have to be optimised to reduce the charging materials, improve the proper production of ferromanganese performance, reduce the specific energy consumption and produce alloy with the standard specification.

This work is a part of complex study has been carried out through a project financed by the Science and Technological Development Fund (STDF), Egypt to enhance the economics of high carbon ferromanganese production process. In the first stage of the project, the real industrial data for high carbon ferromanganese production at Sinai Manganese Company have been collected and evaluated through statistical analysis of collected data and evaluation of the production technique. The results of this stage are under publication. In the second stage of the project, pilot plant smelting experiments for producing high carbon ferromanganese were designed and performed at the Pilot Unit of Steel and Ferroalloys, Central Metallurgical R&D Institute to determine the optimum conditions for producing HCFeMn with the highest yield and recovery through investigating some parameters affecting the smelting process including Mn/Fe ratio of the blend, coke ratio, slag basicity and dolomite/limestone ratio of the flux. In these experimental heats, the same charging materials used at Sinai Manganese Company for producing high carbon ferromanganese were used. This paper presents the results of pilot plant smelting experiments to optimise the condition for smelting high carbon ferromanganese. In the last stage of project, the results obtained have been applied in industrial scale at Sinai Manganese Company for producing standard HCFeMn containing minimum75%Mn. Adjusting the different production parameters in smelting high carbon ferromanganese has successfully reduced the consumption of the charged materials and electric energy consumption/ton high carbon ferromanganese. With using the optimum condition, the electric energy consumption/ton high carbon ferromanganese has been reduced from average 2993 to 2553 kWh with saving of 440 kWh/ton HCFeMn representing 14·7% and consequence reduction in the production cost. The industrial application results are under publication.

Experimental

Raw materials

The raw materials used in this investigation to perform the pilot plant experiments for smelting the high carbon ferromanganese are four manganese ores, Mn sinter, coke as reducing agents and limestone and dolomite as fluxing materials.

The charging materials were subjected to complete chemical analysis using both standard wet analysis methods8 and X-ray fluorescence (XRF) technique. X-ray fluorescence analysis was carried out on pressed powder of 10 g ore sample and 2 g wax (C₁₈H₃₆O₂N₂) using Panalytical Axios Advanced XRF and Omnian program of results database.

The chemical compositions of manganese ores and manganese sinter are given in Table 1. The chemical analyses of fluxing materials: limestone and dolomite, coke and coke ash are also given in Tables 2 4 respectively.

Table 1.

Chemical analysis of manganese ores and sinter/wt-%

	Mn	MnO₂	MnO	Fe	Fe₂O₃	CaO	MgO	SiO₂	Al₂O₃	P₂O₅	K₂O	BaO	LOI	H₂O	Mn/Fe
Mn Ore 1	49·1	46·8	25·13	7·94	11·35	2·6	0·71	7·35	2·52	0·186	0·35	1·45	8·9	0·86	6·18
Mn Ore 2	43·45	40·1	23·31	11·25	16·08	2·35	1·18	10·46	2·65	0·192	0·37	1·87	9·1	0·67	3·86
Mn Ore 3	44·5	47·5	18·6	9·2	13·15	2·45	1·43	10·3	2·65	0·192	0·38	2·1	8·9	0·86	4·84
Mn Ore 4	48·0	51·1	20·2	7·0	10·01	2·3	1·65	6·5	3·1	0·183	0·45	1·6	9·1	0·67	6·86
Mn Sinter	57·58	23·03	55·47	3·63	5·19	1·34	0·86	6·93	4·32	0·186				0·74	15·86

Table 2.

Chemical analysis of flux materials/wt-%

	CaO	MgO	SiO₂	Fe₂O₃	Fe	Al₂O₃	LOI	H₂O
Limestone	52	0·65	4·1	0·8	0·56	0·53	41·69	0·34
Dolomite	29·3	19·33	5	0·8		0·95	44·25	0·3

Table 3.

Chemical analysis of coke/wt-%

Fixed C	Ash	VM	H₂O
85·45	11·8	1·55	0·85

Table 4.

Chemical analysis of coke ash/wt-%

SiO₂	Al₂O₃	Fe₂O₃	CaO	MgO	MnO	TiO₂	S	P	H₂O
46·2	32·85	15·4	3	0·25	0·65	1	0·12	0·028	1·2

Furthermore, manganese ores and manganese sinter were subjected to X-ray diffraction (XRD) analysis. The XRD was carried out using a BRUKER, Axs D8 advance X-ray diffractometer with a Cu K_α X-ray source with a sec. monochromator. The scans were done from 10 to 80° at 1° min⁻¹ and a step size of 0·02°. The voltage used was 40 kV, with a current of 40 mA. The phase analyses of the four manganese ores and manganese sinter as detected by XRD analysis are illustrated in Fig. 1.

Figure 1.

X-ray diffraction pattern of Mn ores and Mn sinter

As it is clear from Table 1, manganese ores contain 43·45 to 49·1% Mn with a Mn/Fe ratio of 3·86 to 6·18, whereas manganese sinter has a higher Mn content and a Mn/Fe ratio of 57·58 and 15·86, respectively.

Pyrolusite (MnO₂) and haematite (Fe₂O₃) are the major phases identified by XRD analysis for manganese ores detected, whereas lower manganese oxides of Mn₃O₄ and MnO are the main phases of manganese sinter.

Manganese ores, manganese sinter, limestone and dolomite were crushed to particle sizes of <50 mm whereas coke was crushed to <10 mm.

Pilot plant submerged electric arc furnace

The pilot plant smelting experimental heats of high carbon ferromanganese were carried out in a 10 kg pilot plant submerged electric arc furnace (Fig. 2). The power is supplied to the furnace through an AC stepwise transformer with a primary electric power of 380 V capable of providing secondary current of 630 A with different voltages ranging between 35 and 44 V through two 40 mm diameter graphite electrodes. The electrodes can be moved up and down by a manual device. The outside dimensions of the overall furnace are 340 mm across the section and 300 mm height. The furnace wall and bottom were rammed with a thick magnesite layer and another layer of carbon paste.

Figure 2.

Schematic diagram of pilot plant submerged electric arc furnace

For selecting the voltage and current suitable for carrying out smooth melting of the charge, preliminary experiments were carried out. It was found that the best current for melting the charge has 480 A and 35 V.

Smelting technique

The material balance for the production of high carbon ferromanganese was computer programmed and the blends of Mn sinter and two Mn ores with the designed Mn/Fe ratios were determined. The amounts of coke required for reduction of different blends were calculated. The amounts of flux materials (limestone and dolomite) with different dolomite/limestone ratios for attaining different slag basicities for different blends were also calculated.

To carry out the experimental heats, the components of the charge, i.e. manganese ores, manganese sinter, coke and flux materials were manually well mixed together. The furnace was preheated and the graphite electrodes were moved down to place near to the bottom lining. A graphite bridge was placed between the tips of electrodes. The electric current was then switched on to raise the hearth temperature due to the produced arc. An initial portion of the mixed charge was added gradually in the furnace. As soon as the molten pool had been established, the furnace was then charged with another portion of the charge. This process was continued until all the charge was fed into the furnace. The furnace was operated at an average of 480 A and 35 V. After all of the main charge had been completely melted, the molten metal and slag was left for 30 min with the current switch on to ensure the maximum degree of reduction and complete setting of the formed high carbon ferromanganese through the slag.

The product was cast into metallic moulds of 100 mm inner diameter and 250 mm height where the metal covered by the slag was left to cool to room temperature.

The produced metal and slag of every experimental heat were weighed and representative samples were then taken for determination of their chemical compositions

Standard wet analysis methods were used to determine Mn, Fe, C and Si in the produced metal and MnO, FeO, CaO, MgO, SiO₂, Al₂O₃ in the produced slag for all heats. The chemical analysis was confirmed by carrying out XRF analysis on selected samples.

Different series of total 57 experimental heats for smelting high carbon ferromanganese were carried out on pilot plant scale. The different series of experimental heats were designed to investigate the effect of some parameters affecting the smelting process including Mn/Fe ratio of the blend, coke ratio, slag basicity and dolomite/limestone ratio of the flux.

The charging materials used in the experimental heats for producing high carbon ferromanganese using different variables are given in Table 5.

Table 5.

Charging materials in experimental heats to produce high carbon ferromanganese using different variables

Heat no.	Variable	Mn Ore 1 wt/g	Mn Ore 2 wt/g		Mn Ore 3 wt/g		Mn Ore 4 wt/g	Mn Sinter wt/g	Mn Ores blend wt/g		Coke wt/g	Dolomite wt/g	Limestone wt/g
1	Mn/Fe	1750	2340		…		…	1750	5840		1100	360	190
2	Mn/Fe	2310	1730		…		…	1730	5770		1100	310	162
3	Mn/Fe	2940	1760		…		…	1175	5875		1112	320	166
4	B	2175	1630		…		…	1630	5435		1070	0	0
5	B	2175	1630		…		…	1630	5435		1070	73	38
6	B	2340	1755		…		…	1755	5850		1100	317	164
7	B	2175	1630		…		…	1630	5435		1070	517	270
8	B	2030	1522		…		…	1522	5074		1050	695	365
9	B	2175	1630		…		…	1630	5435		1070	370	193
10	B	2175	1630		…		…	1630	5435		1070	591	297
11	B	2050	1555		…		…	1555	5160		1070	625	310
12	B	2030	1522		…		…	1522	5074		1050	695	365
13	Coke	2175	1630		…		…	1630	5435		750	517	270
14	Coke	2175	1630		…		…	1630	5435		900	517	270
15	Coke	2175	1630		…		…	1630	5435		1070	517	270
16	Coke	2175	1630		…		…	1630	5435		1230	517	270
17	Coke	2000	1500		…		…	1500	5000		762	495	244
18	Coke	2000	1500		…		…	1500	5000		1016	495	244
19	Coke	2000	1500		…		…	1500	5000		1400	495	244
20	Coke	2000	1500		…		…	1500	5000		1525	495	244
21	Coke	2000	1500		…		…	1500	5000		1270	495	244
22	Coke	2000	1500		…		…	1500	5000		1525	495	244
23	Lime/dol	2000	1500		…		…	1500	5000		1270	0	591
24	Lime/dol	2000	1500		…		…	1500	5000		1270	336	294
25	Lime/dol	2000	1500		…		…	1500	5000		1270	365	266
26	Lime/dol	2000	1500		…		…	1500	5000		1270	390	243
27	Lime/dol	2000	1500		…		…	1500	5000		1270	410	224
28	Lime/dol	2000	1500		…		…	1500	5000		1270	436	200
29	Lime/dol	2000	1500		…		…	1500	5000		1270	665	0
30	Lime/dol	2000	1500		…		…	1500	5000		1270	332	290
31	Lime/dol	2000	1500		…		…	1500	5000		1270	301	316
32	Lime/dol	2000	1500		…		…	1500	5000		1270	276	356
33	Lime/dol	2000	1500		…		…	1500	5000		1270	254	338
34	Lime/dol	2000	1500		…		…	1500	5000		1270	228	379
35	Lime/dol	2000	1500		…		…	1500	5000		1270	193	510
36	Lime/dol	2000	1500		…		…	1500	5000		1270	580	169
37	Lime/dol	2000	1500		…		…	1500	5000		1270	0	100
38	Lime/dol	2000	1500		…		…	1500	5000		1270	0	389
39	Lime/dol	2000	1500		…		…	1500	5000		1270	0	581
40	Limestone	2000	1500		…		…	1500	5000		1270	0	676
41	Limestone	2000	1500		…		…	1500	5000		1270	0	772
42	Limestone	2000	1500		…		…	1500	5000		1270	0	591
43	Limestone	2000	1500		…		…	1500	5000		1270	0	591
44	Limestone	2000	1500		…		…	1500	5000		1270	0	591
45	Mn/Fe	…	…		1500		2500	1000	5000		1267	445	221
46	Mn/Fe	…	…		1500		2000	1500	5000		1264	427	210
47	Mn/Fe	…		…		2000	2000	1000		5000	1263	450	222
48	Mn/Fe	…		…		2500	1500	1000		5000	1259	451	223
49	Mn/Fe	…		…		2500	1000	1500		5000	1260	431	213
50	Mn/Fe	…		…		2500	2000	500		5000	1262	472	233
51	Mn/Fe	250		4500		…	…	250		5000	938	507	250
52	Mn/Fe	500		4000		…	…	500		5000	936	467	230
53	Mn/Fe	500		3500		…	…	1000		5000	931	425	210
54	Mn/Fe	1000		3000		…	…	1000		5000	934	387	191
55	Mn/Fe	1500		2500		…	…	1000		5000	937	348	172
56	Mn/Fe	1500		1500		…	…	2000		5000	925	265	130
57	Mn/Fe	3000		500		…	…	1500		5000	941	191	95

Results and discussion

The weight and chemical composition of the produced high carbon ferromanganese and slag are also presented in Tables 6 and 7, respectively.

Table 6.

Weight and chemical composition of the produced high carbon ferromanganese

Heat no.	Alloy wt/g	Alloy chemical composition/wt-%
Heat no.	Alloy wt/g	Mn	Fe	C	Si
1	2000	75	17	7·1	0·17
2	2250	76	16·7	6·8	0·27
3	1875	74	19·05	6·5	0·29
4	1950	73·8	19·12	6·5	0·19
5	1930	74·1	19·01	6·29	0·203
6	2400	76·1	15·81	6·8	0·15
7	2450	75·9	15·2	6·59	0·204
8	1700	73	20	6·8	0·15
9	2180	76·01	17·1	6·68	0·2
10	2000	74·1	18·47	6·93	0·18
11	1950	74·5	19·82	6·5	0·18
12	1870	74·5	17·98	6·97	0·21
13	1900	73·13	18·1	6·6	0·19
14	1950	74·1	17·5	6·7	0·2
15	1870	74·8	18·1	6·8	0·19
16	2050	75	17	6·98	0·22
17	1550	70·7	21·9	6·74	0·22
18	1700	72	20·3	6·9	0·28
19	1900	72·8	19·6	6·9	0·21
20	1550	70·57	22·3	6·71	0·2
21	2180	75·8	16·1	6·8	0·29
22	1650	71·5	21·5	6·71	0·21
23	2430	76	15	6·87	0·24
24	1700	73·1	19·8	6·2	0·21
25	2040	75	17·5	7·01	0·28
26	1750	72·8	19·9	6·97	0·24
27	1830	73·9	18·89	7	0·24
28	1900	74·01	19	6·93	0·21
29	1750	72·5	20·1	6·91	0·28
30	2120	75·8	17	6·9	0·22
31	2200	76	16·7	6·93	0·25
32	1940	74·3	18·3	7·02	0·25
33	1850	74	18·6	7	0·21
34	2000	75·8	17	6·96	0·22
35	1850	73·1	19·61	6·97	0·23
36	2000	74·9	17·7	6·98	0·25
37	1800	74·1	18·6	6·98	0·23
38	1950	74·5	18	7·03	0·21
39	2100	75·3	17·49	6·96	0·23
40	1900	74·9	17·9	6·95	0·25
41	1800	75·2	17·5	7	0·28
42	2050	76	16·4	6·99	0·27
43	2000	75·3	17·4	6·95	0·32
44	1940	75·1	17·47	7·03	0·3
45	2000	75·5	17·01	7·01	0·32
46	2100	76·1	16·8	6·9	0·25
47	2125	76·5	16·5	6·82	0·26
48	2150	77·1	14·83	6·9	0·31
49	2300	73	15·15	6·85	0·22
50	2050	75	17·2	6·8	0·29
51	1550	72·5	20·4	6·85	0·37
52	1450	71·5	21·1	6·27	0·23
53	1500	73·2	19·8	6·76	0·29
54	1750	74·2	18·6	6·82	0·3
55	1600	74·4	18·6	6·87	0·3
56	1850	75·3	17·8	6·71	0·31
57	1950	76	16·8	6·9	0·26

Table 7.

Weight and chemical composition of produced slag

Heat no.	Slag wt/g	Chemical composition of the produced slag/wt-%
Heat no.	Slag wt/g	MnO	FeO	CaO	MgO	SiO₂	Al₂O₃
1	1550	24·24	0·33	19·45	5·9	31·16	10·86
2	1431	27·29	1·14	17·96	5·39	28·88	10·88
3	1478	25·84	0·34	18·64	5·44	29·81	10·53
4	1026	28·49	0·13	7·47	2·99	33·93	13·26
5	1107	24·31	0·16	11·13	3·9	34·42	13·33
6	1451	27·06	0·69	18·13	5·44	29·16	10·99
7	1590	26·71	0·17	23·9	6·77	26·42	9·69
8	1832	28·45	0·12	27·43	7·53	23·39	8·36
9	1429	27·69	0·3	20·08	5·88	27·97	10·45
10	1682	23·65	0·27	26·82	7·35	26·38	9·6
11	1682	26·01	0·3	27·59	7·4	24·62	8·89
12	1710	17·11	0·21	31·76	8·72	27·07	9·68
13	1534	29·5	0·15	22·9	6·52	25·36	9·04
14	1567	26·03	0·22	24·07	6·83	26·62	9·66
15	1606	28·78	0·15	23·26	6·57	25·69	9·5
16	1642	17·15	0·1	27·15	7·65	29·96	11·26
17	1393	15·05	0·07	28·13	7·81	30·27	10·6
18	1435	17·16	0·1	27·53	7·6	29·56	10·61
19	1498	16·55	0·1	27·87	7·64	29·86	11·08
20	1519	25·87	0·13	24·78	6·78	26·53	9·95
21	1477	24·37	0·12	25·2	6·92	27·03	9·93
22	1519	26·3	0·15	24·62	6·74	26·37	9·89
23	1343	21·74	0·14	26·39	2·7	29·93	11·13
24	1448	17·14	0·08	27·7	6·63	30·06	11·07
25	1457	25·68	0·1	24·82	6·23	26·82	9·87
26	1465	19·83	0·11	26·75	6·97	28·83	10·6
27	1472	21·63	0·1	26·14	7·03	28·09	10·32
28	1480	25·9	0·12	24·69	6·88	26·44	9·71
29	1553	22·95	0·12	25·51	9·24	26·55	9·68
30	1448	20·71	0·11	26·5	6·34	28·76	10·59
31	1438	19·51	0·1	26·93	6·12	29·33	10·81
32	1431	26·18	0·11	24·71	5·37	27	9·96
33	1424	30·59	0·15	23·23	4·85	25·46	9·4
34	1415	32·42	0·16	22·63	4·49	24·89	9·19
35	1396	21·83	0·1	26·25	4·53	29·11	10·77
36	1500	23·59	0·11	25·42	7·69	27	9·9
37	1000	21·66	0·12	11·47	3·14	36·01	13·97
38	1172	25·92	0·13	18·77	2·75	30·97	11·76
39	1286	30·82	0·18	21·53	2·44	27·22	10·19
40	1343	27·18	0·13	24·56	2·52	27·86	10·36
41	1400	23·58	0·23	27·56	2·58	28·39	10·48
42	1343	22·07	0·11	26·29	2·69	29·82	11·09
43	1343	21·32	0·14	26·54	2·72	30·1	11·19
44	1343	20·07	0·11	26·97	2·76	30·59	11·37
45	1463	16·88	0·08	25·57	8·51	28·97	11·18
46	1407	24·31	0·12	23·09	7·65	26·14	10·87
47	1466	19·39	0·09	24·84	8·19	28·07	10·68
48	1470	29·04	0·13	21·89	7·15	24·67	9·25
49	1413	19·31	0·1	24·71	8·02	27·81	11·25
50	1526	21·56	0·1	24·4	8·01	27·54	9·59
51	1690	21·01	0·1	23·04	6·72	35·17	7·74
52	1610	19·86	0·09	23	6·72	35·22	8·42
53	1500	20·91	0·11	22·22	6·57	34·2	9·2
54	1470	21·49	0·1	21·67	6·36	33·45	9·57
55	1328	21·25	0·1	21·31	6·2	32·98	10·07
56	1147	21·77	0·11	19·9	5·98	31·33	12·58
57	975	19·81	0·09	19·1	5·45	30·27	13·63

Effect of Mn/Fe ratio of blend

In the first series of the pilot plant experimental heats, different blends of two Mn ores and Mn sinter with different Mn/Fe ratios were used to investigate the effect of this parameter. All the experiments of this series were carried out using the same calculated basicity and coke ratio. From the obtained results, the metallic yield, expressed as the weight of produced alloy ×100/total weight of the input manganese and iron, was calculated and plotted against Mn/Fe ratio of the blend (Fig. 3a ). As it is clear from Fig. 3a , the metallic yield increases with increasing manganese/iron ratio of the blend. Increasing the metallic yield, i.e. increasing the output alloy, is accompanied with decreasing consumption of Mn ores blend, coke and fluxing materials. Thus, as the Mn/Fe ratio of the blend increases, the blend weight/alloy weight, coke weight/alloy weight and flux weight/alloy weight all decrease, as shown in Fig. 3a–c . This means that lower amounts of Mn ores, coke and flux materials can be used for attaining the same output alloy weight by increasing the Mn/Fe ratio of the blend.

Figure 3.

Mn/Fe ratio of Mn ores blend versus

Furthermore, the amount of produced slag decreases as the Mn/Fe ratio of the blend increases, as shown in Fig. 4. The decrease in the amount of produced slag could be attributed to the decrease in flux materials consumption (Fig. 5), as well as the decrease in the blend weight.

Figure 4.

Mn/Fe ratio of Mn ores blend versus slag wt/alloy wt

Figure 5.

Flux wt/alloy wt versus slag wt/alloy wt

The higher Mn/Fe ratio of the blend does not only reduce the input materials but also improves the produced alloy quality by increasing the Mn per cent and Mn/Fe ratio of the produced high carbon ferromanganese as illustrated in Fig. 6.

Figure 6.

Mn/Fe ratio of Mn ores blend versus

Effect of coke ratio

The stoichiometric amount of carbon required to reduce the manganese and iron oxides to metal and produce HCFeMn (7%C) alloy was calculated and consequently the stoichiometric amount of coke (coke ratio = 1) was also calculated.

In the second series of the pilot plant experimental heats, different coke ratios were used with keeping constant blend of two Mn ores and Mn sinter with constant Mn/Fe ratio and constant basicity to investigate the effect of coke ratio. From the obtained results, manganese recovery and iron recovery were calculated and plotted against the coke ratio, Fig. 7a and b respectively. Fig. 7b illustrates slight increase in iron recovery as the coke ratio increases up to 1·2. On the other hand, manganese recovery gradually increases as the coke ratio increases up to an optimal value of 1. Further coke addition higher than the stoichiometric amount decreases the manganese recovery (Fig. 7a ). The same trend is also revealed in the relation between the coke ratio and metallic yield (Fig. 7c ).

Figure 7.

Coke ratio versus

The lower manganese and iron recoveries and metallic yield at lower coke addition than the stoichiometric could be attributed to the lack of reducing agent required to the reduction process. On the other hand, over-coke encourages the highly endothermic ‘Boudouard reaction’ to occur,9 resulting in lower temperature and consequently hinders the direct reduction of MnO by carbon. This effect may be the reason for the lower manganese recovery and consequently lower metallic yield at higher coke addition than the stoichiometric.

Effect of slag basicity

In the smelting process of high carbon ferromanganese, the higher manganese oxides that predominate in manganese ores (MnO₂, Mn₂O₃ and Mn₃O₄) are relatively unstable and easily reduced in solid state in the presence of CO gas (3) (4) (5) These are exothermic reactions producing a considerable amount of heat, and thereby preheating the charge materials in the furnace.

Iron is always present in manganese ores, and the reduction of iron oxides run parallel to reduction of the manganese higher oxides. Complete reduction in the solid state is possible (6) However, the most possible reaction is the indirect reduction of higher iron oxides in the solid state by CO to FeO (7) When the temperature has reached 800–1000°C, the reaction at the surface of the coke is sufficiently rapid to make the ore reduction and the ‘Boudouard reaction’ to run simultaneously. As a result, the CO₂ gas formed by reduction of ore may in turn react with carbon to give the reaction (8) The ‘Boudouard reaction’ is strongly endothermic. The relative extent of the gas reduction and ‘Boudouard reaction’ is reflected in the furnace off-gas CO/CO₂ ratio.

Although the possibility of complete reduction of iron oxides in the solid state, FeO is often reduced with solid carbon (9) Unlike the reduction of iron oxides, further gas reduction of solid MnO to manganese metal is not possible. MnO is far more stable than FeO.

Considerable smelting of the remaining oxide mixture starts at about 1250°C, and the final reduction of MnO to Mn metal will take place with solid carbon in the coke bed (10) The reduction of silica and P₂O₅ takes place also with solid carbon (11) (12) The direct reduction reactions by solid carbon are highly endothermic and will consume high amount of electric energy.

Slag is formed by the oxides of silicon, calcium, magnesium, aluminium, manganese and iron, which have failed to be reduced in smelting process. Silicon almost entirely enters slag. Only a small portion of silicon is reduced from slag and appears in the metal. The reduced manganese passes to the metal, while the unreduced oxides remain in the slag. Almost all of the iron is reduced and enters the metal, so the slag is low in ferrous oxide.

In production of high carbon ferromanganese, slag plays an important role in the distribution of elements between slag and metal phases and consequently greatly affects the recovery of different elements and metallic yield.

The third series of the pilot plant experimental heats were designed to create different slag compositions, by adding different amounts of fluxing materials (limestone and dolomite), to attain different slag basicities. In all experimental heats of this series, constant blend of Mn sinter and two Mn ores with the same Mn/Fe ratio was used and maintaining the same coke ratio.

From the results obtained, different slag basicity indices were determined as illustrated in equations (13)–(16) (13) (14) (15) (16) The determined different slag basicity formulas were correlated with the different parameters of alloy output. From the results obtained, Figs. 8–11 are plotted illustrating the effect of the different slag basicity formulas on the recovery of Mn, Fe and Si as well as the metallic yield. These figures reveal the same trend for all different slag basicity formulas. Iron recovery slightly increases whereas silicon recovery slightly decreases as slag basicity increases in the range studied. On the other hand, both manganese recovery and metallic yield increase by increasing the slag basicity up to optimum figure. Further increase in slag basicity higher than the optimum results in decreases in both manganese recovery and metallic yield. The optimum slag basicity for attaining the highest manganese recovery and metallic yield is about 0·9 for B1, 1·1 for B2, 0·8 for B3 and 0·9 for B4.

Figure 8.

Slag bascity, expressed as (CaO)/(SiO₂), versus

Figure 9.

Slag bascity, expressed as (CaO+MgO)/(SiO₂), versus

Figure 10.

Slag bascity, expressed as (CaO+MgO)/(SiO₂+Al₂O₃), versus

Figure 11.

Slag bascity, expressed as (100–SiO₂–Al₂O₃–MnO)/(SiO₂+Al₂O₃), versus

The slight increase in iron recovery by increasing slag basicity could be attributed to the increase in the activity of FeO in the basic slags. However, this only slight increase in iron recovery by increasing slag basicity indicates that appreciable amount of iron oxides may be indirect reduced in the solid state by CO. Furthermore, the higher slag viscosity of the basic slags retards the direct reduction of FeO by C.

The lower manganese recovery and consequently lower metallic yield in the heats with lower slag basicity can be explained by the low activity of MnO in these acidic slags and formation of manganese silicates. At higher slag basicity, the activity of MnO increases as more MnO is free.

According to the regular solution model, the activity coefficient of the component in the slag can by expressed by the following equations10 (17) (18) where, X _j is cation fraction, and α _ij is the interaction energy between cations, i.e. (i cation)–O–(j cation). The reference state of the activity in this case is taken to hypothetical pure liquid. Therefore, the activity coefficient of MnO (γ _MnO) can be calculated as the function of slag composition and temperature by equation (2) and the values of interaction energy.10 Then, the activity of the MnO (a_MnO) can be deduced from the following relation (19) where X _MnO is the molar fraction of MnO.

From the results of calculated γ _MnO and a _MnO, Figs. 12 and 13 were plotted illustrating the effect of slag basicity, expressed as (CaO+MgO)/(SiO₂), on the activity coefficient and activity of MnO. Figure 12 illustrates increasing the activity coefficient of MnO as the slag basicity increases. At a given MnO mole fraction, Fig. 12 reveals higher MnO activity at higher slag basicity.

Figure 12.

Slag bascity, expressed as (CaO+MgO)/(SiO₂), versus activity coefficient of MnO

Figure 13.

MnO mole fraction versus activity of MnO at different slag bascities

The appearance of optimum slag basicity for attaining the highest manganese recovery and metallic yield was examined by determining the approximated viscosity of the different slags formed at different slag basicities. Due to the complication of the slag system, it seems possible to consider the slag is composed of CaO, MgO, MnO, SiO₂ and Al₂O₃ as the summation of these oxides represent mass contents of more than 95% of the slag composition. Furthermore, MgO has a character similar to CaO, therefore, to simplify the slag system the percentage of CaO can be assumed to be the sum of CaO+MgO in the slag. Using this assumption, it seems possible to compare the slag viscosity values at different basicities through the Al₂O₃–CaO–MnO–SiO₂ system.11 The deduced viscosity values at different slag basicities, expressed as (CaO)/(SiO₂) and (CaO+MgO)/(SiO₂) are shown in Figs. 14 and 15, respectively. These figures illustrate that the lowest viscosity is attained in slag with (CaO)/(SiO₂) and (CaO+MgO)/(SiO₂) ratios of 0·9 and 1·1, respectively.

Figure 14.

Slag bascity, expressed as (CaO)/(SiO₂), versus slag viscosity

Figure 15.

Slag bascity, expressed as (CaO+MgO)/(SiO2), versus slag viscosity

This is in good agreement with the obtained results, in which the highest manganese recovery and metallic yield are obtained at optimum slag basicity: B1 = 0·9 and B2 = 1·1. The retardation of the reduction process at slag basicity lower than the optimum is the result of two negative effects, i.e. low MnO activity and high slag viscosity. For slag basicity higher than the optimum, the negative high slag viscosity effect seems to overcome the positive effect of increasing the MnO activity with the net result of retardation of the reduction process.

The decrease in silicon recovery by increasing slag basicity is a result of two negative effects: decrease in the silica activity due to the formation of stable silicate compounds, and increase in slag viscosity. However, the very low silicon recovery even at the lowest slag basicity (∼3%) is a result of the lower working temperature of high carbon ferromanganese smelting process comparing with that required for silicon reduction.

The slag basicity also affects the consumption of fluxing materials and the produced slag. As the slag basicity increases, both the fluxing materials consumption and the produced slag increase (Fig. 16).

Figure 16.

Slag bascity, expressed as (CaO+MgO)/(SiO₂), versus

Effect of MgO/CaO of slag

The fourth series of the pilot plant experimental heats were designed to investigate the effect of (MgO)/(CaO) ratio of slag by using different dolomite/limestone ratios with all other parameters are constant.

From the results obtained, Fig. 17 illustrates the effect of MgO/CaO ratio of slag on manganese recovery and metallic yield, respectively. As it is clear from this figure, increasing the MgO/CaO ratio in the slag is accompanied by decreasing manganese recovery and metallic yield. This behaviour can be attributed to the increase in slag viscosity by increasing MgO content in slag.12 The high viscous slag would retard the diffusion process and increase the residence time of metal drops in the slag bath and thus deteriorates the smelting process and slag/metal separation.

Figure 17.

MgO/CaO ratio of slag versus

The MgO/CaO ratio in the slag decreases by decreasing the dolomite/limestone ratio in the flux (Fig. 18).

Figure 18.

Dolomite/limestone ratio vesus (MgO/CaO) ratio in slag

Conclusions

From the results of this study, the following conclusions can be deduced.

Using Mn blend with high Mn/Fe ratio decreases the consumption of charging materials: Mn ores, coke and fluxing materials (limestone and dolomite). This means that lower amounts of Mn ores, coke and flux materials can be used for attaining the same output alloy weight by increasing the Mn/Fe ratio of the blend.

Furthermore, the produced slag decreases as the Mn/Fe ratio of the blend increases.

The higher Mn/Fe ratio of the blend does not only reduce the input materials and the produced slag but also improves the produced alloy quality by increasing the Mn per cent and Mn/Fe ratio of the produced high carbon ferromanganese.

For obtaining standard HCFeMn alloy containing minimum 75%Mn, it is necessary to use the Mn blend with the Mn/Fe ratio of higher than 6·2.

The added coke must be adjusted according to the material balance and stoichiometric to prevent the over-coke and minimise the highly endothermic ‘Boudouard reaction’ to attain the highest Mn recovery and metallic yield.

Slag basicity should be optimised with flux addition to attain the highest Mn recovery and metallic yield.

The optimum slag basicity for attaining the highest manganese recovery and metallic yield is about 0·9 for B1, 1·1 for B2, 0·8 for B3 and 0·9 for B4.

The basicity is approximately defined by the raw material mixture composition. It slightly increases during reduction process due to the slight amount of silica reduced in the smelting of high carbon ferromanganese. Thus, the slag basicity can be determined from the chemical composition of the raw material mixture with taking into consideration that about 2% of silicon in the raw material mixture will be reduced and goes into metal phase.

Much higher slag basicity is not recommended as this practice produces viscous slag accompanied with lower Mn recovery and metallic yield.

10.

The results revealed the negative effect of increasing MgO/CaO ratio of slag on manganese recovery and metallic yield. The MgO/CaO ratio in the slag decreases by decreasing the dolomite/limestone ratio in the flux.

Footnotes

Acknowledgements

This work is a part of complex study has been carried out through a project financed by the Science and Technological Development Fund (STDF), Egypt. The authors would like to acknowledge the STDF due to their financial support and all facilities they offered to perform this work. Cordial thanks and deep appreciation are due to Eng. Mohamed Abdel Samie Chairman & Managing Director of Sinai Manganese Company (SMC), for his encouragement, sound support and providing the raw materials for performing this study. Special thanks and gratitude are due to all members in Steel and Ferroalloys Department, CMRDI and technical staff of SMC Company.

References

Tang

Olsen

: Proc. Conf. INFACON 11, New Delhi, India, February 2007, IFAPA, 335–343.

Olsen

Ligstad

: Electric Furnace Conference Proceedings, San Antonio, TX, USA, November 2002, ISS, 205–216.

Olsen

: Proc. Conf. INFACON 9, The Ferroalloys Association, Quebec City, Canada, June 2001, 280–285.

Olsen

Tangstad

Lindstad

: ‘Production of manganese ferroalloys’, 247; 2007, Trondheim, SINTEF and Tapir Academic Press.

Tangstad

Calvert

Brun

Lindseth

: Proc. Conf. INFACON 10, Johanesbrg: Saimm, Cape Town, South Africa, February 2004, 213–222.

Vorob’ev

Godunov

Ignat’v

: Steel Translation, 2009, 39, (3), 243–247.

Nikoleav

: Steel USSR, 1974, 299–302.

Vogel

: ‘A textbook of quantitative inorganic analysis including elementary instrumental analysis’, 925; 1992, London, Longman Group Ltd.

Eissa

: ‘Energy saving in high carbon ferromanganese alloy production’, Second Progress Report of Project ID: 97, Submitted to Science and Technological Development Fund (STDF), Cairo, 2010.

10.

Ban-Ya

: ISIJ Int., 1993, 33, (1), 2–11.

11.

Verein Deutscher Eisenhüttenleute , ed (ed.): ‘Slag Atlas’, 282; 1981, Düsseldorf, Verlag Stahleisen GmbH.

12.

El-Fawakhry

Mishreky

Ebaid

Eissa

Matter

: Proc. Middle East Sulphur Symp., Cairo, Egypt, February 1990, NRC, 79–91.