Production of Ferroalloys - PDF Free Download (2023)

CHAPTER 1.10

Production of Ferroalloys Rauf Hurman Eric School of Chemical and Metallurgical Engineering, University of Witwatersrand, Johannesburg, South Africa

1.10.1. CLASSIFICATION, MANUFACTURE, AND USE OF FERROALLOYS The term Ferroalloy refers to various alloys of iron with a high proportion of one or more other elements such as chromium, manganese, and silicon. Ferroalloys are primarily used in the production of steels, stainless steels, and other grades of alloy steels as raw materials. They impart distinctive qualities to ferrous materials such as steels and cast irons or serve important functions during their manufacturing. Manganese is essential to the production of virtually all steels and is also important in the production of cast iron. Manganese neutralizes the harmful effect of residual sulfur, can act as a deoxidizer, and is also an important alloying element imparting hardenability to steel. Silicon is primarily added to steel as a deoxidizer, but it is also an alloying element in cast iron and some electrical grade steels. Chromium is the main element providing corrosion resistance in stainless steels and also added to special grade of alloy steels used in the manufacture of tools. It is quite clear that the ferroalloy industry is closely associated with the iron and steel industry, the leading consumer of its products. The leading ferroalloy-producing countries are China, South Africa, Russia, Kazakhstan, and Ukraine. The total annual world production of all ferroalloys combined together reached its peak amount of around 43 million tons in year 2010 very much in accordance with the annual peak reached in steel and stainless steel manufacturing during the same year. Since then due to the financial crisis seen in major world economies, the production of steel, stainless steel, and hence ferroalloys has decreased by about 20%. At the peak year of 2010, world production of bulk chromium, manganese, and silicon ferroalloys is estimated around 32 million tons made up of approximately 5.7 million tons of ferromanganese, 9.5 million tons of silicomanganese, 7.9 million tons of ferrosilicon, and 8.9 million tons of ferrochromium. The main ferroalloys (in alphabetical order) are: FeCr, ferrochromium; FeMn, ferromanganese; FeMo, ferromolybdenum; FeNi, ferronickel; FeSi, ferrosilicon; FeTi, ferrotitanium; FeV, ferrovanadium; and FeW, ferrotungsten.

Treatise on Process Metallurgy, Volume 3 http://dx.doi.org/10.1016/B978-0-08-096988-6.00005-5

© 2014 Elsevier Ltd. All rights reserved.

477

478

Rauf Hurman Eric

Other important ferroalloys are: FeAl, ferroaluminum; FeB, ferroboron; FeCe, ferrocerium; FeMg, ferromagnesium; FeMo, ferromolybdenum; FeNb, ferroniobium; FeP, ferrophosphorus; FeSiMg, ferrosilicon magnesium; and FeU, ferrouranium. In the following classification, use and manufacture of the main ferroalloys will be discussed briefly [1].

1.10.1.1. Ferrochromium Over 80% of the world’s ferrochrome output is utilized in the production of stainless steel. Stainless steel is dependent on chromium for its appearance and its resistance to corrosion. The average chromium content in stainless steel is 18%. FeCr is also used when it is desired to add chromium to carbon steel. FeCr from South Africa known as “charge chrome” and produced from a low-grade chrome ore is most commonly used in stainless steel production. High-carbon FeCr produced from high-grade ore found in Kazakhstan (among other places) is more commonly used in specialist applications such as engineering steels where a higher Cr to Fe ratio is important. Ferrochrome production is essentially a high-temperature carbothermic reduction operation. Chrome ore (an oxide of chromium and iron) is reduced by coke (and coal) to form the iron–chromium–carbon alloy. The heat for the process is provided typically from the electric arc formed between the tips of the electrodes in the bottom of the furnace and the furnace hearth in very large cylindrical furnaces known as “submerged arc furnaces.” As the name implies the three carbon electrodes of the furnace are submerged into a bed of mainly solid and some liquid mixture made up of the solid carbon (coke and/or coal), solid oxide raw materials (ore and fluxes) as well as the liquid FeCr alloy and molten slag droplets that are being formed. In the process of smelting, huge amounts of electricity are consumed. Tapping of the material from the furnace takes place intermittently. When enough smelted ferrochrome has accumulated in the hearth of the furnace, the tap hole is drilled open and a stream of molten metal and slag flows out down a trough into a chill or ladle. The ferrochrome solidifies in large castings, which are crushed for sale or further processed. Ferrochrome is classified by the amount of carbon and chromium it contains. The vast majority of FeCr produced is charge chrome from South Africa.

1.10.1.2. Ferromanganese Manganese ferroalloys consist of various grades of ferromanganese and silicomanganese. High-carbon ferromanganese, generally with 70–80% Mn and 6–7% C, is by far the largest tonnage ferroalloy used. It is a deoxidizing agent in steelmaking and an important alloying element. It has also the property of controlling the harmful effect of sulfur. Manganese combines with sulfur and forms manganese sulfide, which tends to float out of the liquid steel. Manganese has the effect of stabilizing the austenite phase; steels with 12–14% Mn are fully austenitic, which are used on a large scale for their wear and abrasion-resisting characteristics.

Production of Ferroalloys

Manganese ferroalloys are also manufactured by carbothermic reduction of manganese ores containing both iron and manganese oxides in submerged arc electric furnaces. The slags produced during the process contain significant amounts of manganese oxide; MnO and these slags can be reprocessed to ferrosilicomanganese (also called silicomanganese) and then to refined or low-carbon (LC) ferromanganese or even to manganese metal. Silicomanganese is generally made in two grades, both containing 65–75% Mn but with either 15–20% Si or 20–25% Si. These alloys are preferred by steelmakers who require to add manganese and silicon simultaneously for deoxidation purposes. The product of deoxidation is of course a fluid manganese silicate slag.

1.10.1.3. Ferromolybdenum The commercially important ore of molybdenum is the sulfide molybdenite, MoS2, produced as a concentrate, which has to be roasted to the trioxide to serve as the starting point for making the ferroalloy. MoO3 is readily reduced with carbon, reduction starting below 500 C. On the other hand, reduction of the oxide to the carbide proceeds simultaneously but the carbide decomposes in the presence of iron or molybdenum oxides. It is, therefore, entirely feasible to reduce the oxide with carbon provided an excess of MoO3 is allowed to pass into the slag. The alloys produced generally contain 55–70% Mo varying only in carbon content from 0.25% to 2.5%. In order to obtain an LC product, the smelting is done in several stages, the final alloy being produced by reduction with a deficiency of carbon, producing a molybdenum-rich slag. This is reduced in a second furnace with an excess of carbon (usually coke breeze), and an addition of iron turnings/scrap to produce an intermediate alloy, which together with raw concentrates and a small amount of carbon is smelted to give the final alloy. While the carbothermic reduction process is feasible as explained briefly above, the bulk of the tonnage of ferromolybdenum is made by the silicothermic reduction process with 75% Si ferrosilicon, using a small proportion of aluminum or calcium–silicon to make it completely self-sustaining and to ensure a good slag–metal separation. Some operators prefer the silicothermic reduction to take place in an arc furnace to ensure the best possible slag separation. More than 80% of the ferromolybdenum produced is consumed by the steel industry in the manufacture of stainless and full alloy steels.

1.10.1.4. Ferronickel Almost all the ferronickel produced is consumed by the steel industry in the manufacture of stainless and heat-resistant steels. The oxidized ores of nickel constitute by far the world’s largest reserves of this metal. These ores include the “laterites” in which the nickel oxide is intimately associated with limonitic iron oxide and the silicates that often contain the mineral garnierite. Many methods and techniques are used for producing ferronickel from the oxidized ores including selective carbothermic reduction

479

480

Rauf Hurman Eric

procedure. Usually, primary ferronickel is obtained by first calcining the iron- and nickel-bearing laterite ore in a kiln and then smelting the calcined ore in an electric furnace. The nickel content of marketable ferroalloy typically ranges from 19% to 38%.

1.10.1.5. Ferrosilicon This is an alloy of iron and silicon with an average Si content between 15% and 90%. The usual types on the market are ferrosilicon grades with 15%, 45%, 75%, and 90% silicon. The remainder is iron, with about 2% consisting of other elements like aluminum and calcium. Ferrosilicon is used as a source of silicon to reduce some metals from their oxides (metallothermic reductions) and to deoxidize steel and other ferrous alloys. It can be used to make other ferroalloys. Ferrosilicon is also used for manufacture of metallic silicon, corrosion-resistant and high temperature-resistant ferrous silicon alloys and silicon steel for electric motors and transformer cores. In the manufacture of cast iron, ferrosilicon is used for inoculation of the iron to accelerate graphitization. In arc welding, ferrosilicon can be found in some electrode coatings. The silicon metal is used as an alloying element mostly with aluminum and in the production chemicals, especially silicones. Silicon metal can also be refined highly to produce solar or even electronic grade (semiconductor) silicon. Ferrosilicon is a basis for manufacture of alloys like magnesium ferrosilicon (FeSiMg), used for modification of melted malleable iron. FeSiMg contains 3–42% magnesium and small amounts of rare earth metals. Ferrosilicon is also used as a reductant in the production of magnesium metal from dolomite. Ferrosilicon is produced by reduction of silica or sand with coke in the presence of scrap iron, mill scale, or other source of iron in submerged arc electric furnaces. An excess of silica over the stoichiometric requirement has to be used to prevent the formation of silicon carbide.

1.10.1.6. Ferrotitanium Titanium is used in steelmaking for deoxidation, grain-size control, carbon and nitrogen control, and stabilization. During steelmaking, titanium is introduced as ferrotitanium because of its lower melting temperature and higher density compared to those of titanium scrap. Producers of interstitial-free, stainless, and high-strength low-alloy steels are the major consumers of titanium within the steel industry. Ferrotitanium is usually produced by induction melting of titanium scrap with iron or steel; however, it is also produced directly from titanium mineral concentrates. The standard grades of ferrotitanium are 30% and 70% Ti. Ferrosilicon titanium is also manufactured to allow the simultaneous addition of silicon and titanium to steel.

1.10.1.7. Ferrovanadium South Africa, China, and Russia account for about 98% of world vanadium mine production. In these three countries, vanadium is primarily recovered from titanium-bearing

Production of Ferroalloys

magnetite ore processed to produce pig iron. The process produces a slag containing 20–24% vanadium pentoxide, which is further processed to ferrovanadium containing 40–50% vanadium. Most ferrovanadium is produced by aluminothermic reduction of such slags, which is self-sustaining, although in some operations raw materials are preheated to conserve aluminum. Sometimes this process is carried out in the electric furnace, the electrodes being used to initiate the reaction and then withdrawn till the reaction is complete. Electric heat is then applied to the molten slag to improve settling and to ensure the completion of the reaction. The silicothermic route, although also exothermic, produces insufficient heat to melt the slag and to give a good slag–metal separation. It is therefore carried out in the electric furnace. Some producers use little carbon to lower the oxidation state of vanadium, with ferrosilicon (generally 75% Si alloy) to finish of the reduction. Vanadium oxides are also recovered from petroleum residues, ashes, and poisoned refinery catalysts. These are used to produce catalysts, chemicals, and high up to 75–80% vanadium-containing ferrovanadium alloys. Ferrovanadium is consumed mainly by the steel industry in the manufacture of carbon steels, full alloy steels, and high-strength low-alloy and tool steels.

1.10.1.8. Ferrotungsten Tungsten is an important alloying element in high-speed and other tool steels and is used to a lesser extent in some stainless and structural steels. Tungsten is often added to steel melts as ferrotungsten, which can contain up to 80% tungsten. The ores of tungsten generally contain only 0.4–2.0% WO3, and they are principally tungstates of calcium (scheelite) or of iron and manganese (wolframite). These ores are upgraded by flotation or by magnetic or gravity separation to produce concentrates which run 60–75% WO3. The concentrates can either be reduced with carbon in the submerged arc furnace, or can be reduced by a combined silicothermic–aluminothermic reaction, which is self-sustaining. The choice of method is a question of economics as well as a question of alloy specification. The melting point of high tungsten containing (above 70%) ferrotungsten alloys is well over 2000 C and hence the tapping of the ferroalloy in the orthodox manner via tap hole from the furnace is not possible. In the so-called “solid block” process, the ferrotungsten is smelted in a removable-body furnace where it is never tapped but is allowed to accumulate as a solid button of alloy, which is eventually reclaimed when smelting is completed by dismantling the furnace and pulling out the button or block of ferrotungsten. There are two methods of carrying out this operation: one method is to smelt the tungsten concentrates plus fluxes and other additions, such as a slag, with an excess of carbon, such that the alloy will contain 1.5–2.0% C. This method produces a throwaway waste slag and an alloy, which is then crushed, smelted, and refined with

481

482

Rauf Hurman Eric

further concentrates and some iron ore. This refining operation produces a tungsten-rich slag that is returned to the first smelting operation. The other route is to smelt the tungsten concentrates with a deficiency of carbon in the first primary stage so that the alloy is relatively low in carbon; 0.5–0.6%, and to resmelt the resulting slag (generally contains about 25% WO3 in the second stage to an intermediate high-carbon alloy, often called “slag–metal,” which is recycled in the primary smelt. When the block of ferrotungsten is extracted, it is broken and then crushed, and sorted. It is usually found that only about 75–80% of the alloy is clean and free from adhering slag; the remaining portion is classed as reverts and returned to the primary smelt. It should be noted here that the final ferrotungsten alloy is obtained from the primary stage and the throwaway waste slag is obtained from the second stage.

1.10.2. THERMODYNAMICS IN THE PRODUCTION OF MAIN FERROALLOYS 1.10.2.1. Introduction The bulk ferroalloys, namely Fe–Mn, Fe–Cr, and Fe–Si are produced commercially by reduction smelting of their oxide minerals with coke or charcoal or another carbonaceous reductant in a submerged arc furnace and, to a lesser extent, by a metallic reductant such as aluminum. In general, the overall reduction reactions are highly endothermic, thus requiring large thermal energy inputs. The thermodynamic principles underlining reduction smelting of oxides are discussed with the help of free energy–temperature (Ellingham), diagram with a focus on the role of direct reduction, the smelting temperatures required for a given oxide mineral, the thermal energy requirements, the suitability of carbon as a reductant, the alloy grade and impurity levels, the slag–metal equilibria, and the formation of carbides.

1.10.2.2. Ellingham Diagram Figure 1.10.1 is an Ellingham diagram that shows the standard Gibbs energy of formation of various oxides as a function of temperature, with respect to 1 mol of oxygen gas [2,3]: (a) The more stable oxides appear on the lower part of the diagram in the order FeO, P2O5, Cr2O3, MnO, SiO2, Al2O3, MgO, and CaO. Thus, if conditions are made favorable for the reduction of MnO, then the oxides of iron, phosphorus, and chromium will also be reduced. (b) For a given oxide mineral, the oxides below it on the Ellingham diagram are more stable and therefore, the elements of the latter (such as Al, Si, and Ca) could readily act as reducing agents. This principle is utilized in the metallothermic reduction processes. (c) Whereas most lines in the Ellingham diagram slope upward, the CO line slopes downward because of entropy considerations. Thus, there is a temperature above

Production of Ferroalloys

Figure 1.10.1 Standard Gibbs energies of formation of metal oxides as a function of temperature [3].

483

484

Rauf Hurman Eric

which CO is more stable than the oxide mineral to be reduced. However, the actual operation may have to be conducted at temperatures higher than the temperature of intersection between MO and CO lines in order to enhance the reaction rates and facilitate metal–slag separation. (d) All the lines plotted in the Ellingham diagram correspond to standard states (pure condensed substances). For impure oxides, the line would rotate clockwise (with respect to zero temperature) indicating greater ease of formation; conversely, for impure elements, the lines would rotate anticlockwise.

1.10.2.3. Direct and Indirect Reduction MO þ C ¼ M þ CO direct MO þ CO ¼ M þ CO2 indirect

ð1:10:1Þ ð1:10:2Þ

The indirect reduction is more exothermic (releases an extra 290 kJ/g mol) relative to the direct reduction. Consequently, the energy efficiency of a process could be significantly enhanced by utilizing the CO formed by direct reduction for subsequent indirect reduction as accomplished in the iron-making blast furnace. However, the latter calls for a long stack and elaborate burden preparation so that descending charge could be preheated and reduced indirectly by the upward moving hot gases before getting directly reduced. However, the relatively small scale of ferroalloy operations does not permit high-shaft furnaces and consequently, most of the CO generated by direct reduction leaves the furnace without being utilized in any indirect reduction reaction.

1.10.2.4. Smelting Temperature The final products are molten alloy and slag. The alloy composition should be within the specified/desired limits and their temperatures should be high enough (100 C above the liquidus) to impart adequate fluidity for good separation. In the case of ferromanganese, the equilibrium pressure of CO produced from the reaction MnOðsÞ þ CðgÞ ¼ MnðlÞ þ CO

DG3 ¼ 290,350 173:25T ð J=gmolÞ

ð1:10:3Þ ð1:10:4Þ

reaches 1 atm. at about 1400 C (corresponding to the intersection of MnO and CO lines on the Ellingham diagram). This temperature allows good slag–metal separation and hence is the normal smelting temperature. It should be noted, however, that even if an equilibrium pressure of 1 atm. for CO could be achieved at somewhat lower temperatures, the need for a clean metal–slag separation (if a slag is formed) may necessitate high-smelting temperatures.

485

Production of Ferroalloys

1.10.2.5. Thermal Energy Requirements As stated earlier, production of ferroalloys is generally an energy-intensive operation mainly due to the endothermic reduction reactions as shown below:

Enthalpy Change DH298

References [4,5]

kJ/g mol

kW h/kg-metal

MnO2 ðsÞ þ 2C ¼ MnðsÞ þ 2CO

300:0

1:52

ð1:10:5Þ

Cr2 O3 ðsÞ þ 3C ¼ 2CrðsÞ þ 3CO

803:1

2:15

ð1:10:6Þ

SiO2 ðsÞ þ 2C ¼ SiðsÞ þ 2CO

689:8

6:82

ð1:10:7Þ

Additional heat must be supplied for melting the metal and slag and raising their temperatures (for clean separation) as well as the gas temperature to the “requisite values.” Further, heat losses to surroundings and cooling water must be provided for. As a result, the heat requirement increases significantly; for manganese, it was estimated to be 2.45 kW h/kg. Because of such large thermal requirements, it is a standard practice to produce ferroalloys, in submerged arc furnaces, wherein the arc provides the necessary heat thus maintaining the reaction zone at suitably high temperatures. The final state of the metal–slag bath is governed by the hearth temperature, which should be reasonably high to effect metal–slag separation.

1.10.2.6. Alloy Grade and Impurity Levels The Ellingham diagram suggests that all the oxides of iron and phosphorus would get reduced during reduction smelting of MnO, Cr2O3, as well as SiO2. As a result, the overall grade of the molten alloy is determined by the extent of oxide reduction and the burden of iron oxides and other easy-to-reduce oxides in the charge. The slag composition is determined by the total gangue from ore, fuel, flux, etc., besides the unreacted mineral oxide. If the alloy specifications call for a maximum limit on phosphorous content, it must be adjusted in the burden itself (such as through mineral beneficiation) as there are no commercially established technologies available to remove phosphorous from the molten ferroalloys [6].

1.10.2.7. Slag–Metal Equilibria The reduction of oxide minerals of iron, manganese, and chromium takes place in stages as shown below: Fe2 O3 Fe3 O4 Fex O Fe MnO2 Mn2 O3 Mn3 O4 MnO Mn Cr2 O3 CrO Cr

ð1:10:8Þ ð1:10:9Þ ð1:10:10Þ

The reduction of Mn ores up to the stage of Mn3O4 occurs by thermal dissociation and indirect reduction (like that of iron ores to Fe3O4) and thereafter, by direct reduction.

486

Rauf Hurman Eric

In the last stage of direct reduction of oxides, the slag–metal phases are believed to approach thermodynamic equilibrium as follows: ðMOÞ þ ½CðalloyÞ ¼ ½M þ COðgÞ aM pCO DG ¼ RT ln K ¼ RT ln ðaMO ac Þ

ð1:10:11Þ ð1:10:12Þ

where parentheses for MO indicate the slag phase and brackets for C and M indicate the liquid alloy phase. Thus, knowledge of the equilibrium constant K, through standard Gibbs energy data, is useful in determining the ways to promote a reaction in the desired direction, thanks to the extensive thermodynamic studies done on liquid iron solutions and oxide slags [7–9]. The activities of various components can be estimated and used to assess the extent of equilibrium achieved in a given process. The multiphase equilibria such as those pertaining to ferroalloy processes have been analyzed by several investigators [10–13] with the help of the phase rule and Gibbs energy minimizing techniques for predicting the distribution of a metal between the slag and alloy phases for different charge compositions and operating conditions of temperature and pressure. In a recent study, Ding and Olsen [14] have analyzed the silicomanganese process by considering the main reactions as follows: ðSiO2 Þ þ 2C ¼ Si þ 2COðgÞ ðMnOÞ þ C ¼ Mn þ COðgÞ ðSiO2 Þ þ 2Mn ¼ Si þ 2ðMnOÞ

ð1:10:13Þ ð1:10:14Þ ð1:10:15Þ

The carbothermic reduction of silica in slag (reaction 1.10.13) being very sluggish, only reactions (1.10.14) and (1.10.15) were used for describing the equilibria as shown in Figure 1.10.2, which indicates that for a given alloy composition [%Si], the slag composition will shift to a lower (MnO) content in slag with increase in temperature. Another advantage of increasing the temperature will be to improve the slag/metal ratio significantly. The slag produced in ferroalloy process usually contains appreciable amounts of the valuable metal as shown below [15]: Alloy

Mn

Cr

C

Si

High-carbon ferrochrome

–

65.5

7.0

4.5

Medium-carbon ferromanganese

72

–

2.0

0.8

Slag Composition

Slag–Metal Ratio

MnO

Cr2O3

CaO

MgO

SiO2

Al2O3

High-carbon FeCr

–

6.2

–

37.2

30.6

26.0

Medium-carbon FeMn

14

–

38

5.0

28.0

3.0

0.83 –

487

Production of Ferroalloys

SiO2

Si

%

20

i %S Si 10% Si 5%

15

CaO+Al2O3+MgO (constant ratio)

MnO

Figure 1.10.2 Slag–metal equilibrium at various silicon contents in the Si–Mn alloy along the solid lines [14].

If the metal loss is considered excessive, such as in the case of medium-carbon ferromanganese shown above, and if the process economics warrants, then a proper control of slag composition should help lower the metal loss. In the case of ferromanganese, a slight increase in slag basicity will tend to increase the activity of MnO in slag as shown in Figure 1.10.3, thereby lowering the (MnO) content. Furthermore, a higher basicity should also help to lower the phosphorus content of the alloy.

1.10.2.8. Carbide Formation From an economic viewpoint, coke–carbon is considered a preferred reductant. However, the ferroalloy thus produced contains high levels of carbon-promoting formation (precipitation) of metal carbides such as: 7½MnðlÞ þ 3½C ¼ Mn7 C3 ðsÞ 7½CrðlÞ þ 3½C ¼ Cr7 C3 ðsÞ ½SiðlÞ þ ½C ¼ SiCðsÞ

ð1:10:16Þ ð1:10:17Þ ð1:10:18Þ

These carbides are refractory materials and the low temperature at the hearth bottom is likely to help precipitate Cr7C3 and SiC, which may cause furnace bottom build up. It is to be noted that in the case of manganese and chromium, the equilibrium carbide phase after solidification will be mostly M23C6; at the smelting temperatures, however, the first carbide to precipitate is likely to be Mn7C3 or Cr7C3 [16]. From a thermodynamic view point, it should be possible to reduce these carbides and recover the metal into the alloy by reacting the former with metal oxide at relatively high temperatures, as shown below: Mn7 C3 ðsÞ þ 3MnOðsÞ ¼ 10½MnðlÞ þ 3COðgÞ

ð1:10:19Þ

488

Rauf Hurman Eric SiO2 .90

.90

.80

.80

.70

.70

.60

.60 .50

.50

0.074 -0.094

0.12 0.18 0.34

.30

.30

0.50

.40

.40

0.15

.20

.20

.10

.10

CaO

0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 MnO

mole fraction

Figure 1.10.3 Isoactivity lines for MnO in the CaO–MnO–SiO2 system at 1650 C [4].

DG19 ¼ þ937,215 461:5T ð J=gmolÞ

ð1:10:19aÞ

For a molten bath saturated with Mn7C3 (aMn7 C3 ¼ 1) and in contact with solid MnO, with pCO ¼ 1 atm. and taking aMn ¼ 0.3 for a 75% Mn in ferromanganese alloys [18], one finds that the above reaction becomes feasible (negative Gibbs energy) at about 1400 C. Similarly, for 2SiCðsÞ þ SiO2 ðsÞ ¼ 3½SiðlÞ þ 2COðgÞ

DG20 ¼ 941,192 442:1T ð J=gmolÞ

ð1:10:20Þ ð1:10:20aÞ

Using aSiC ¼ 1, aSiO2 ¼ 1, pCO ¼ 1 atm., and aSi ¼ 0.82 for a 75% Si in ferrosilicon [4], it is found that DG equals zero at 1830 C. Thus, at temperatures above 1830 C, the above reaction will become feasible promoting dissolution of silicon carbide into the Fe–Si alloy.

1.10.2.9. Production of LC Ferroalloys LC alloys are usually produced by aluminothermic or silicothermic reduction smelting of the mineral and, to a smaller extent (for special purposes); by refining of a high-carbon liquid alloy through oxidation (of carbon) by fluxes and/or oxygen, as illustrated below for the case of ferrochrome:

489

Production of Ferroalloys

1.10.2.9.1 Silicothermic Reduction of Chrome Ore Using ferrochrome silicon (Si: 46–50%, C < 0.1%) as the reductant and lime as flux, LC ferrochrome can be and is produced by smelting of chrome ore as per the following reactions: 2Cr2 O3 þ 3½Si þ CaO ¼ 4½Cr þ ðCaO3SiO2 Þ

ð1:10:21Þ

2FeO þ ½Si þ CaO ¼ 2½Fe þ ðCaO SiO2 Þ

ð1:10:22Þ

For given compositions of the chrome ore, flux, and ferrochrome silicon [18], one can calculate charge requirements and product chemistries using mass balance equations along with simplifying assumptions as follows: Chrome ore

Cr2O3: 53, FeO: 14, SiO2: 7, MgO: 12, Al2O3: 11

Fe–Cr–Si

Si: 52, Cr: 31, Fe: 16.9, C: 0.02

Lime

CaO: 90

Alloy

0.06% C, 0.70% Si aim composition

Assume

Cr recovery: 82%, Fe recovery: 90% Si efficiency: 70%, Slag basicity (CaO þ MgO)/SiO2 ¼ 2.2

For a 100 kg of the chrome in the charge, mass balance for Si and CaO yields: Alloy

59.2 kg (Cr: 71.1, Fe: 28.0, Si: 0.7, C: 0.06)

Slag

184.2 kg (SiO2: 27.5, CaO: 54.04, MgO: 6.5, Al2O3: 5.98, Cr2O3: 5.18, FeO: 0.76)

1.10.2.9.2 Oxidation Refining by Oxide Fluxes/Oxygen 1 2 C þ Cr2 O3 ¼ ½CrðlÞ þ CO 3 3 DG23 ¼ 272,350 176:1T ð J=gmolÞ

ð1:10:23Þ ð1:10:24Þ

To attain very LC levels in the alloy melt, it is necessary to increase temperatures (to 1800 C) and/or reduce pCO using vacuum.

1.10.3. FERROCHROMIUM SMELTING TECHNOLOGY More than 80% of the world production of ferrochromium is used in stainless steel making. There are four grades of ferrochromium produced commercially, characterized broadly in terms of their carbon and chromium contents

490

Rauf Hurman Eric

High-carbon ferrochromium

(Cr: >60%, C: 6–9%)

Charge chrome

(Cr: 50–60%, C: 6–9%)

Medium-carbon ferrochromium

(Cr: 56–70%, C: 1–4%)

LC ferrochromium

(Cr 56–70%, C: 0.015–1.0%)

The demand for LC ferrochromium, produced by reacting Fe–Cr–Si alloy with a Cr2O3–CaO-based slag, has decreased dramatically during the last two decades mainly due to the commercial development of AOD (Argon-Oxygen-Decarburization) and VOD (Vacuum-Oxygen-Decarburization) processes, which allow removal of carbon from stainless steels with acceptable loss (oxidation) of chromium. These low and ultra-LC ferrochromium grades are used mainly for final adjustments of composition and for super alloys, which are melted in coreless induction furnaces. The ultra-low ferrochromium, produced by aluminothermic reduction of chromite, is relatively pure but very expensive and consequently not widely employed in the steel industry. As a result, the high-carbon ferrochromium has become the most widely produced and consumed grade of chromium-containing ferroalloys. The production of highcarbon ferrochromium is based on reduction smelting of chromite ore with coke in the presence of silica in a submerged arc furnace.

1.10.3.1. Raw Materials The natural ores of chromium are mainly composed of spinels, FeOFe2O3, FeOCr2O3, MgOCr2O3, and MgOAl2O3, all in solid solution in the chromite mineral, and accompanied by gangue frequently composed of serpentine (MgOSiO2nH2O). With suitable concentration treatment, it is possible to remove most of the siliceous gangue rendering a concentrate containing Cr2O3, Fe2O3, and Al2O3, with nearly stoichiometric amounts of FeO and MgO, i.e., moles of R2O3 equal to moles of RO. Indigenous chromite ores mostly fall into two broad categories: (a) High-silica hard lumps from igneous rock, containing serpentine (MgOSiO2nH2O) as the principal gangue with an overall Cr/Fe ratio of approximately 3. These siliceous ores are generally amenable to beneficiation by gravity methods. (b) Ferruginous friable ores from limonite deposits rich in secondary iron oxides. Since the chromite ore contains gangue, it is not possible to produce LC, low silicon ferrochromium in the conventional single-stage submerged arc furnace smelting process. In general, additional silica must be included in the charge as a flux to facilitate formation of a fluid slag during the smelting operation.

1.10.3.2. Submerged Arc Furnace The conventional smelting process for producing ferrochromium is carried out in electric reduction furnaces having Soderberg electrodes submerged in the burden material.

491

Production of Ferroalloys

To achieve steady process of the reduction reactions, the gas flow and partition through the burden should be uniform enough to avoid channeling of carbon monoxide gas produced by the reactions. This requires the charges to be primarily comprised of lumpy ore with a minimum of ore fines; also, the ore should not be friable so that excessive degradation of the ore does not occur in the furnace. However, due to increasing mechanization of the ore-mining operations and the necessity of using ferruginous friable ores, there is considerable generation of ore fines, with grain size smaller than 1 mm, which calls for some form of agglomeration (such as pelletizing, briquetting, and sintering) of the feed materials. Another important cost factor of the conventional technology is the high power consumption, which is in the range of 4000–4200 kW h/ton FeCr alloy.

1.10.3.3. Thermodynamic Considerations Most of the reduction reactions taking place in the submerged arc furnace are highly endothermic and are listed in Table 1.10.1 along with standard Gibbs free energies as a function of temperature. Table 1.10.1 also lists for various reactions the temperature at which the CO pressure is one atmosphere, under standard state conditions for the reactants and products. In actual smelting practice, the reduction temperature will be raised if the oxide is chemically combined in the gangue, and lowered if the reaction product can dissolve in the ferrochromium alloy. The data in Table 1.10.1 indicate that at these smelting temperatures, iron, chromium, and silicon form stable carbides. In the case of Cr2O3, it first reacts with carbon to form higher carbide, Cr7C3, which reacts back at higher temperatures with Cr2O3 to Table 1.10.1 Standard Gibbs Energies and Equilibrium Temperatures of Reactions During Carbothermic Reduction of Chromite [19]. Reaction

DH 298 K (kJ/mol)

DG (J/mol)

T in K for pCO ¼ 1 atm.

Fe3O4 þ C ¼ 3FeO þ CO

194.2

200,595 212.86T

942

FeO þ C ¼ Fe þ CO

161.5

153,251 152.08T

1001

Cr2O3 þ 3C ¼ 2Cr þ 3CO

803.1

825,029 486.0T

1698

Fe3O4 þ 5C ¼ Fe3C þ 4CO

703.8

670,083 682.28T

983

3FeO þ 4C ¼ Fe3C þ 3CO

509.6

470,114 469.42T

1001

5300.4

5426,008 3453.5T

1571

616.5

624,559 346.06T

1805

2,167,307 1025.48T

2113

937,000 450.0T

2082

7Cr2O3 þ 27C ¼ 2Cr7C3 þ 21CO SiO2 þ 3C ¼ SiC þ 2CO 3SiO2 þ 2SiC ¼ Si þ 4SiO þ 2CO Cr2O3 þ 3Cr7C3 ¼ Cr23C6 þ 3CO

2163 956.7

492

Rauf Hurman Eric

form lower carbide Cr23C6. This reaction occurs at approximately the same temperature as that required for reducing silica to SiC. 1.10.3.3.1 Carbon Content of Ferrochromium In submerged arc smelting, the carbon content of the ferrochromium is generally high and varies significantly because of the formation of various chromium carbides such as Cr3C2, Cr7C3, and Cr23C6. Table 1.10.2 shows the stability ranges for the reduction products and also the carbon contents of the carbides. The data indicate that temperatures over 1600 C are necessary to achieve carbon content under 9%. As mentioned earlier, chromite ores contain iron oxides and gangue that have significant effects on the reduction reactions in submerged arc smelting. Studies on solid-state reduction of chromite have shown that the iron oxides get reduced more readily than the chrome oxides. Under such conditions, iron forms only one carbide Fe3C, which can dissolve chromium to form a complex carbide (Cr,Fe)7C3. Thus, an ore rich in iron oxide will have high reducibility at relatively low temperatures. Accordingly, the reduction of chromic oxide in such ores will also occur at relatively lower temperatures with the likely formation of a carbon-rich chrome carbide (Cr3C2 or Cr7C3) known to be stable at lower temperatures. The gangue present in a chromium ore has a significant effect on the temperature of the smelting zone, thereby influencing the carbon content of the ferroalloy. An ore having a relatively high MgO content will require a higher smelting temperature. Further, if silica flux addition is reduced or lime is added, the liquidus temperature of the slag will increase thus promoting high-smelting temperatures. It has been argued that, because of their high surface-to-volume ratio, fine ores react readily at low temperatures forming a product high in carbon. The use of chromite–carbon agglomerates is reported to have produced high-carbon ferrochromium since they start reacting at lower temperatures. On the other hand, coarse-sized ores will not be as reactive and thus can survive to a lower depth in the burden and react with high carbon alloy, thereby promoting an LC ferrochromium. One method proposed for minimizing the carbon content of the ferrochromium is to create a region in the furnace that is oxidizing to the chromium carbide with the use of a high-density high-smelting ore. The high density (4 g/cm3) will allow it to pass Table 1.10.2 Stability Ranges for the Chromium Carbides [19]. Reduction Product

Stability Range ( C)

Weight Percentage Carbon in Product

Cr3C2

1150–1250

13.3

Cr7C3

1250–1600

9.0

Cr23C6

1600–1820

5.7

Cr

>1820

493

Production of Ferroalloys

through the low density slag (3 g/cm3) and form a chromic oxide-rich layer above the molten ferrochromium. Such an ore layer can oxidize the chromium carbides. Further, due to its viscosity it will remain in the furnace even after tapping. This explains how relatively LC levels can be obtained with the use of dense, lumpy, low reducibility ores. 1.10.3.3.2 Silicon Content of High-Carbon Ferrochromium The silicon content of the ferrochromium should be low as it is an undesirable element in stainless steels. Low silicon content is favored by a low operating temperature, high carbon content in the ferroalloy and a basic slag. Other impurities are not known to affect the silicon content significantly. In general, the specifications for ferrochromium call for a silicon level below 3%. This can be achieved by utilizing the idea, mentioned earlier, of forming a layer of chromic oxide ore in the lower portion of the slag, which oxidizes some silicon in addition to carbon. It has been amply demonstrated that, in the iron blast furnace, the silicon content of the hot metal depends on the silica content of the ash in coke in accordance with the following reactions: SiO2 ðcoke ashÞ þ CðcokeÞ ¼ SiOðgÞ þ COðgÞ SiOðgÞ þ CðcokeÞ ¼ ½Siðliquid ironÞ þ COðgÞ

ð1:10:25Þ ð1:10:26Þ

Thus, with increase in coke rate, the relative amount of SiO(g), and consequently, silicon in the metal will also increase. It is believed that under the highly reducing conditions in the submerged arc furnace, the silicon content of the ferrochromium would also depend on the silica content of the ash in the carbonaceous reductant used. High silicon ferrochromium, low in carbon, can be desiliconized by treatment with a chromite ore–lime slag, when it is necessary to produce LC ferrochromium. 1.10.3.3.3 Phosphorous Content of High-Carbon Ferrochromium Phosphorous is detrimental to both the mechanical properties and corrosion resistance of stainless steels. In the submerged arc smelting of chromite ore, a portion of the phosphorus contained in the charge is vaporized and removed with the off-gas; however, up to 60% can be retained in the alloy, For low phosphorous levels (<0.02%) in ferrochromium, the phosphorus content of the raw materials should be as low as possible. Also a relatively low operating temperature will promote removal of phosphorus into the slag phase, especially under oxidizing conditions. However, due to the highly reducing and hot conditions in the submerged arc furnace, there are no easy ways to produce a low-phosphorus ferrochromium from high phosphorus ore/coke. Even though several studies have been made on removal of phosphorus from liquid ferrochromium using fluxes such as CaF2–CaC2 and Ca–CaF2, industrial practices mostly rely upon the use of imported low-phosphorus metallurgical coke as a preventive measure.

494

Rauf Hurman Eric

1.10.3.4. Kinetic and Mechanism Considerations This section on the mechanisms and kinetics of chromite reduction is focused on providing insight into the reactions occurring during ferrochrome smelting by submerged arc furnace technologies [28]. The approach is to specifically follow and examine the reaction steps that result in the formation of metal. From investigations of excavated quenched submerged arc furnaces and previous overviews of the technology and the process steps [20–27], the following idealized reaction zones have been identified (see Figure 1.10.4). Note, however, that the exact positions of these zones appear to vary with furnace design and operating practice. The zones do not necessarily follow a simple layered structure as was envisaged in early models of the process [28]. 1.10.3.4.1 Upper Furnace Zone 1 Loose Charge The loose charge zone extends from the charge layer down to near to the tip of the electrode. The following reactions occur in this zone: • Preheating of the charge, • Decomposition of limestone fluxes, e.g., CaCO3, and other minerals, • Gasification of carbon; reaction with air and CO2, • Gaseous reduction of chromite ore and partial metallization of iron and chromium oxides. It is known that most of the volume in the submerged arc furnace is loosely sintered burden [27]. The average retention time in this zone is estimated to be 24 h, but only approximately 20% of reduction of the charge takes place in this loose charge zone; no liquid slag is formed. It was demonstrated [20], through the use of alumina tracers that the burden material descended in a V-shaped distribution, and rate of descent reaches a

Figure 1.10.4 Schematic diagram of the reaction zones in submerged arc furnace for ferrochromium production.

Production of Ferroalloys

maximum at positions between the furnace walls and the electrodes and between the electrodes themselves. Temperature profiles and excess gas pressures in this zone were measured and quoted [23–26]. These data showed that the 1873-K isotherm was achieved only close to the electrode tips, and that above 1673 K the gas pressure rises rapidly, the later temperature corresponding to the onset of slag formation. 1.10.3.4.2 Lower Furnace (Zones 2–6) The reactions occurring in the lower furnace involve: Slag formation, dissolution of chromite in slag, reduction of metal from the slag phase and metal alloy formation, and alloy/slag separation Sidewall slag/metal/ore/coke (Zone 2) Banks of rigid, partially fused/partly reduced material are formed in Zone 2 adjacent to the furnace walls. Sidewall/slag/metal (Zone 3) The material below Zone 2 contains mixtures of slag and metal. Beneath the electrodes (Zone 4) There is some uncertainty in the literature about the material present immediately under the electrode tips. These zones do not connect with the similar zones under the other electrodes (A, B, C). The presence of a void may have been due to contraction of the bed during cooling of the furnace [21]. Slag and coke under two electrodes but mainly slag under the third electrode were reported [24]. Another study [27] indicates the presence of a coke bed, containing a mixture of melted gangue minerals, fluxes, and MgO and Al2O3 liberated from the chromite during reduction. Due to the formation of partially solidified charge materials around the electrodes (Zones 2 and 3), the active slag reduction zone is restricted in size. The residence time in the high-temperature smelting zone, defined here as the coke bed (Zone 4), is relatively short, possibly of the order of 30–40 min. Slag (Zone 5) There is also significant uncertainty in the literature about the material existing in this zone. Some studies [21,24] found a distinctive slag layer; however, another one [27] reports that no distinctive slag layer was found. The presence of a large region of unmolten partly reacted lumpy ore between the slag and metal was also reported [21,26]. Metal (Zone 6) All previous studies show the formation of a distinct liquid metal alloy layer at the base of the furnace. As the chromite ore moves through the submerged arc furnace during processing it passes through the above-mentioned zones. In the upper furnace, Zone 1, the chromite particulates are present as loose charge and product gas from the lower furnace consisting of mainly CO and CO2 passes through this packed bed. Reactions between the gases and the solid chromite result in the partial reduction of the chromite and the formation of iron and chromium alloy, and, in some cases, metal carbides.

495

496

Rauf Hurman Eric

In the lower furnace, Zones 2–6, a series of reaction steps take place [28] involving the slag phase (Figure 1.10.5): • Dissolution of chromite spinel in slag • Reduction of the slag phase to form alloy • Alloy/slag separation. Although the overall reaction is described by the removal of oxygen by solid carbon, this reaction can take place through a number of pathways involving: carbon monoxide formation from solid carbon or carbon dissolved in alloy; reaction of CO gas with slag; metal ion diffusion in the slag phase. Each of these mechanisms contributes to the overall recovery of chromium as alloy. Optimization of the process to enhance the rates and extent to which these reactions proceed, and to maximize recovery, production rate, and energy efficiency is assisted by improved understanding of the various reactions and reaction mechanisms operative in the process, and by taking action through feed selection and preparation, and adjustment of process conditions. Temperature ~1973 K Slag Prereduced primary chromite spinel

Cr

C

dissolving

Cr–C

Alloy

Fe–Cr alloy released

C Cr–C Alloy Cr–C3

Fe–Cr Alloy agglomeration

Temperature ~1873 K

Relict primary spinel

Secondary spinel precipitation Cr–Si–C Alloy

Slag

Alloy Cr–Fe–Si–C

Figure 1.10.5 Alternative reaction pathways for reduction of ferrochromium slags with carbon [28].

497

Production of Ferroalloys

1.10.3.5. New Trends in Smelting Technology Increasing costs of electric energy and reduced availability of high-grade lumpy ores have prompted significant improvements in process technologies. Some of these developments, which have been commercially successful such as those involving, pelletizing of the ore fines, and preheating/prereducing the pellets, are briefly discussed as follows. 1.10.3.5.1 The Outokumpu Process This industrially proven process allows (a) utilization of a feed, comprised mainly/ exclusively of ore fines and (b) heat recovery from the arc furnace off-gas (CO) for preheating the burden is schematically shown in Figure 1.10.6. The process involves grinding and pelletizing of ore fines, followed by sintering of green pellets and preheating before smelting. The ore and coke fines are normally wet-ground to about 35% under 37 m (400 mesh) and then pelletized to approx. þ15 mm size. The bulk of the fuel for sintering and preheating the pellets is provided by the CO from the submerged arc furnace, which must be top-closed so as to recycle the CO gas generated inside. The preheating of the charge, usually in a rotary kiln, results in power saving of about 500–700 kW h/ton. STANDARD PROCESS Lump Ore Coke Fluxes

OUTOKUMPU PROCESS Chromite concentrate Coke

Grinding Mill

Smelting

Pelletizing Disc

Pellet Sintering Coke Fluxes Preheating Kiln Furnace Gas Power Requirement of Melting/Smelting 4000-4200 kWh/t

Smelting 3200-3600 kWh/t

Figure 1.10.6 Conventional smelting and Outokumpu process with sintering and preheating of pellets [19].

498

Rauf Hurman Eric

1.10.3.5.2 The Showa–Denko Process The logical step, after the Outokumpu Process, in further reducing electrical energy consumption was implemented by the Showa–Denko process, which includes a prereduction treatment of the ore in a rotary kiln downstream of the pelletizing plant, is illustrated in Figure 1.10.7. The hot prereduced burden with about 60% reduction degree is directly charged into the submerged arc furnace. As a result of the prereduction, the power consumption is reported to be in the range of 2000–2300 kW h/ton alloy. Because of a strong affinity during the prereduction treatment in a rotary kiln of chromium for oxygen, it is necessary, therefore, to keep the carbonaceous reductant in microscopic contact with the chromite ore particles and allow them to react at temperatures higher than 1280 C. To produce pellets of good strength and reactivity, ore and coke are pulverized to a fine homogenous mixture of size generally less than 100 m. The pellets are subjected to a roasting treatment in a rotary kiln at temperature 1340–1450 C under air. Reaction between the coke and ore within the pellets proceeds unhampered by the roasting atmosphere because of the protection provided by the formation of a hard insulating film of the flux (mostly forsterite 2MgOSiO2 and spinel MgOAI2O3) outside the pellets. The pellets must be strong enough to withstand the tumbling action in the rotary kiln to ensure proper temperature distribution. Furthermore, since the difference between the reaction temperature and the softening temperature of the pellets is rather small, special attention is required for controlling the flame and the furnace temperature. By treating low-medium-grade ore fines, which are more abundantly available, and using

SHOWA DENKO/CMS Process Chromite Concentrate Coke Grinding Mill Pelletizing Disc

Preheating Reduction Kiln

Reduction degree ~ 60%

Coal Burners Ladle Transfer Coke Fluxes Smelting

Power Requirement of Melting/Smelting 2000/2100 kWh/t FeCr

Figure 1.10.7 Prereduction of pellets and smelting/melting in Showa–Denko/CMI process [19].

499

Production of Ferroalloys

a closed-top electric furnace, the process achieves substantial reduction not only in unit power consumption but also in pollution–abatement costs, that would be otherwise incurred in handling the dusty ore in the open. The Showa–Denko process is credited with two industrial plants, having a total production capacity of 200,000 ton/year, operating in Japan and South Africa. 1.10.3.5.3 The CODIR Process for Chromite Fines The CODIR process as shown schematically in Figure 1.10.8 was conceived and tested by Krupp in the 1980s and later commercialized by Mannesmann Demag. It aims at (1) Direct charging of ore fines into a rotary kiln. (2) Achievement of almost complete reduction (90%) of chromium and iron oxides in the rotary kiln with the use of cheap noncoking coal as exclusive reductant and fuel. (3) Hot charging of prereduced calcine into an electric furnace for open-bath smelting resulting in a low power consumption of 1050–1200 kW h/ton alloys only. The process is reported to have been in full industrial application at Samancor’s plant in South Africa. The plant employs a rotary kiln, 4.8 m in diameter and 80 m long; rate at 120,000 ton/year along with a submerged furnace with a transformer capacity of 33 MVA. In the rotary kiln, chromite ore fines and coal react at temperatures approaching 1450 C to produce a semisolid product consisting of highly metalized ferrochromium, slag, gangue, and char. High temperature and intimate contact between the chromite CODIR FOR CHROMITE FINES Chromite Concentrate Coal Fluxes

Reduction Kiln Coal Burners Reduction Fluxes degree ~ 90% Conditioning Drum

Ladle Transfer

Smelting Power Requirement of Melting/Smelting 1050/1200 kWh/t FeCr

Figure 1.10.8 Prereduction and smelting/melting in CODIR process [19].

500

Rauf Hurman Eric

grains and carbon are necessary over the entire residence time in the kiln to achieve a high degree of metallization (>90%). For this purpose, the burden is transformed into a semisolid (pasty) state with some phases molten and other solid, so that only a small portion of the surface of the metalized phase is exposed to the furnace atmosphere, and yet macroscopic phase separation does not take place. Phase separation is carried out subsequently by melting the kiln discharge in the submerged arc furnace. It has been reported recently that the smelter is being upgraded to a 56-MVA furnace for transferring energy to the process/charge via a DC plasma arc produced by a single hollow graphite cathode.

1.10.4. REDUCTION OF MANGANESE OXIDES AND PRODUCTION OF MANGANESE ALLOYS In this section, production of manganese-containing ferroalloys will be discussed in a systematic way based on an excellent previous review [29]. Manganese is only produced commercially from blends of ore in which manganese exists as oxide. Knowledge of the thermodynamics of the reduction of these oxides is therefore essential to the understanding of the production of alloys from manganese ores.

1.10.4.1. Manganese–Oxygen System Of the known oxides of manganese (Mn2O7, MnO2, Mn5O8, Mn2O3, Mn3O4, and MnO), the oxides found naturally in Manganese ores are MnO2, Mn2O3, Mn3O4, and MnO. Under normal conditions, manganese dioxide occurs as stable b-MnO2. Mn2O3 is cubic above 20 C. Trimanganese tetra oxide (Mn3O4) exists at atmospheric pressure in the tetragonal modification and changes to the cubic form at 1170 C. Manganese oxide (MnO) in normal conditions exists in only one modification. This oxide is, however, nonstoichiometric and is more accurately described as Mn1xO, with x ¼ 0–0.25. The complete phase diagram of the manganese–oxygen system above 1000 C is given in Figure 1.10.9. The equations for the thermal dissociation of the important manganese oxides are as follows: 4MnO2 ! 2Mn2 O3 þ O2 DH298 ¼ 166:0kJ 6Mn2 O3 ! 4Mn3 O4 þ O2 DH298 ¼ 64:8kJ 2Mn3 O4 ! 6MnO þ O2 2MnO ! 2Mn þ O2

ð1:10:27Þ ð1:10:28Þ ð1:10:29Þ ð1:10:30Þ ð1:10:31Þ ð1:10:32Þ

The oxygen partial pressures exerted by these reactions are given as a function of the temperature in Figure 1.10.10. According to the thermodynamics data, the decomposition temperature at 101.3 kPa for reaction (1.10.27) is 500 C and for reaction (1.10.29),

501

Production of Ferroalloys

2500 –6

–2 -10

1 Log(pO ), bar 2

–7

L+MnO2 Liquid

–8

Temperature, K

L1+L2

2000

(2083K)

2057K 2124K

(1954K) (1928K)

L+Mn1−xO

1500

1518 K d–Mn+Mn1−xO 1411 K 1360 K g–Mn+Mn1−xO

Mn2O3+MnO2

1835K –1 –2 Mn1−xO

–3

+b–Mn3O4

–4

b–Mn3O4+Mn2O3 1443K

–5

1

–6 Mn1−xO+a–Mn3O4 –7

0 a–Mn3O4+Mn2O3 –1

b–Mn+Mn1−xO

–8 –2

1000 0.45

0.50

0.60 0.55 Mole fraction O

0.65

0.70

Figure 1.10.9 Mn–O phase diagrams above 1000 K [30].

Figure 1.10.10 Predominance diagram of oxides in Mn-O system as a function of oxygen partial pressure and reciprocal temperature [29].

502

Rauf Hurman Eric

980 C. For reaction (1.10.31) to occur at atmospheric pressure, a reductant is necessary. Reaction (1.10.32) requires a high reducing potential. It is important, therefore, to consider the reduction of the oxides of manganese in the presence of carbon monoxide, hydrogen, and carbon. These are the reactions that usually take place in the blast furnace and electric furnaces used to produce manganese alloys.

1.10.4.2. Reduction of Manganese Oxides with Carbon Monoxide, Hydrogen, and Carbon The reduction of MnO2 to Mn occurs in four main steps (see Equations 1.10.27, 1.10.29, 1.10.31, and 1.10.32).The reduction is strongly exothermic in the presence of carbon, methane, and carbon monoxide; the following equations apply: 1 1 1 MnO2 þ C ! Mn2 O3 þ CO2 4 2 4 DH298 ¼ 56:8kJ 1 1 1 1 MnO2 þ CH4 ! Mn2 O3 þ H2 O þ CO2 8 2 4 8 DH298 ¼ 56:8kJ 1 1 1 MnO2 þ H2 ! Mn2 O3 þ H2 O 2 2 2 DH298 ¼ 79:5kJ 1 1 1 MnO2 þ CO ! Mn2 O3 þ CO2 2 2 2 DH298 ¼ 99:6kJ

ð1:10:33Þ ð1:10:34Þ ð1:10:35Þ ð1:10:36Þ ð1:10:37Þ ð1:10:38Þ ð1:10:39Þ ð1:10:40Þ

The reduction reactions of Mn2O3 in the presence of the same reducing agents are also exothermic. The following equations apply: 1 1 1 1 Mn2 O3 þ C ! Mn3 O4 þ CO2 2 12 3 12 DH298 ¼ 16:3kJ 1 1 1 1 1 Mn2 O3 þ CH4 ! Mn3 O4 þ CO2 þ H2 O 2 24 3 24 12 DH298 ¼ 17:2kJ 1 1 1 1 Mn2 O3 þ H2 ! Mn3 O4 þ H2 O 2 6 3 6 DH298 ¼ 24:3kJ 1 1 1 1 Mn2 O3 þ CO ! Mn3 O4 þ CO2 2 6 3 6 DH298 ¼ 31:0kJ

ð1:10:41Þ ð1:10:42Þ ð1:10:43Þ ð1:10:44Þ ð1:10:45Þ ð1:10:46Þ ð1:10:47Þ ð1:10:48Þ

503

Production of Ferroalloys

The third step is the reduction of Mn3O4, which is endothermic with carbon or methane and exothermic with hydrogen or carbon monoxide. 1 1 1 Mn3 O4 þ C ! MnO þ CO2 3 6 6 DH298 ¼ þ11:8kJ 1 1 1 1 Mn3 O4 þ CH4 ! MnO þ CO2 þ H2 O 3 12 12 6

ð1:10:49Þ ð1:10:50Þ ð1:10:51Þ

DH298 ¼ þ 10:5kJ

ð1:10:52Þ

1 1 1 Mn3 O4 þ H2 ! MnO þ H2 O 3 3 3

ð1:10:53Þ

DH298 ¼ 3:3kJ

ð1:10:54Þ

1 1 1 Mn3 O4 þ CO ! MnO þ CO2 3 3 3

ð1:10:55Þ

DH298 ¼ 17:2kJ

ð1:10:56Þ

The last step, the reduction of MnO, is more difficult and requires a high reducing potential. The presence of carbon is therefore necessary, and at a carbon monoxide partial pressure of 101.3 kPa, a temperature of 1420 C is necessary for the reaction to proceed. However, the formation of manganese carbides in the presence of carbon lowers the reaction temperature to 1300 C. The following, highly endothermic overall reaction applies: MnO þ C ! Mn þ CO DH298 ¼ þ274:5kJ

ð1:10:57Þ ð1:10:58Þ

This reaction can be broken down into MnO þ CO ! Mn þ CO2

ð1:10:59Þ

and the reaction of CO2 with C to produce CO CO2 þ C ! 2CO

ð1:10:60Þ

which ensures that a high reducing potential is maintained. Due to the importance of both carbon monoxide and carbon as reducing agents in the commercial production of manganese containing alloys, these are discussed in more detail. It is possible to predict the ratio CO/CO2 that will reduce the individual manganese oxides by considering the relationship between the oxygen partial pressure and the equilibrium constants for reactions (1.10.27), (1.10.29), (1.10.31), and (1.10.32). Kp ¼ pO2

ð1:10:61Þ

504

Rauf Hurman Eric

Assuming that the activities of the solid phase are unity and the equilibrium constant of the reaction which is

2CO þ O2 ! 2CO2

ð1:10:62Þ

1 p2CO :pO2 ¼ 2 pCO2 Kp

ð1:10:63Þ

DG ¼ RT ln Kp

ð1:10:64Þ

DG ¼ RT ln pO2

ð1:10:65Þ

and the relationships

and

The oxygen potentials of the individual reactions at various temperatures can be calculated. This is shown graphically in the Ellingham diagram (see Figure 1.10.11), which is illustrated again to show greater detail and the shift of the CO line as a function of carbon monoxide partial pressure. The same applies to the use of hydrogen as a reductant. Although the reduction of MnO2, Mn2O3, and Mn3O4 takes place at low partial pressure of CO, the presence of carbon is necessary for the reduction of MnO. In this case, the following reaction is important: 2C þ O2 ! 2CO

ð1:10:66Þ

log pO2 ¼ 11,672=T 9:16

ð1:10:67Þ

for which

This reaction is pressure dependent due to the change in the number of moles of gas and again the Ellingham diagram can be used to predict under which conditions the reduction of MnO will occur (see Equations 1.10.57, 1.10.59, and 1.10.60). However, manganese carbides are always formed during the reduction of MnO with carbon, and this lowers the temperatures at which the reaction occurs. The products produced in this process always contain carbon. The phase diagram for the manganese–carbon system is shown in Figure 1.10.12. The melting point increases with increasing carbon content. Commercial ferromanganese (76% Mn, 16% Fe) contains 7% carbon. The solubility of carbon decreases with increasing silicon content.

1.10.4.3. Reduction of Mixed Oxides and Minerals Containing Manganese Oxides In the discussion of the reduction of manganese oxides, it has been assumed that the activity of the solid phase is unity. However, this is usually not true in practice because in most

Production of Ferroalloys

Figure 1.10.11 Temperature dependence and oxygen potentials of selected oxide systems [3]. The equilibrium for reaction 2CþO2 =2CO is presented in four different pressures: 1) 3.0 bar, 2) 1.2 bar, 3) 0.5 bar, 4) 0.3 bar.

manganese ores the manganese oxides are combined with iron oxides, silicates, aluminum oxides, calcium oxides, and phosphorous oxides. The oxides FeO and P2O5 are reduced at lower temperature than MnO, whereas SiO2 needs slightly higher temperature (see Figure 1.10.11). Any iron oxide and oxides of phosphorous occurring in the ore, fluxes, and reductants used in commercial processes are reduced together with the manganese oxides and report to the product. A judicious

505

506

Rauf Hurman Eric Atomic Percent Carbon 10

20

30

1400 1300

L 1248⬚C

1200

Temperature ⬚C

1100

(δMn)

13

ε

1143⬚C

~1090⬚C

1079⬚C (γMn)

~1050⬚C 1020⬚C

1000

Mn3C

990⬚C

~970⬚C

(βMn)

900

Mn15C4

890⬚C

~850⬚C

Mn5C2

(αMn)

600

Mn7C3 + graphite

710⬚C

Mn23C6

700

Mn7C3

800

500 0

Mn

1

2

3

4

5

6

7

8

9

Weight Percent Carbon

Figure 1.10.12 Phase diagram for the Mn–C system [16,29].

choice of ores and raw material is therefore necessary to limit the amount of these elements, which report to the metal. The silica contained in the ores can also be reduced, and in the production of silicomanganese this is desirable. Quartz or quartzite is therefore an important raw material for the production of silicomanganese. The presence of other oxides in the ores reduces the activity of the manganese oxide. This must be compensated for by increasing the reducing potential by increasing the CO/CO2 ratio of the gas mixture. Alternatively, the reduction temperature must be increased if the reaction is carried out with carbon.

1.10.4.4. Reduction of Manganese Oxides by Silicon Figure 1.10.11 shows that the reduction of manganese(II) oxide by silicon is also possible: 2MnO þ Si ! SiO2 þ 2Mn

ð1:10:68Þ

for which aSiO2 a2Mn ð1:10:69Þ a2MnO aSi This reaction is the basis of the Frisch process for the production of LC ferromanganese and manganese metal from silicomanganese. The silicon content of the metal is restricted by using a basic slag that reduces the activity of SiO2 by reaction with CaO, displacing reaction (1.10.68) as far to the right as possible. K¼

507

Production of Ferroalloys

1.10.4.5. Reduction of Manganese Oxide by Aluminum It is easier to reduce manganese(II) oxide with aluminum than with silicon because the difference in the oxygen potential of the dissociation equilibrium of MnO and Al2O3 are greater than for MnO and SiO2 (see Figure 1.10.11). As in the case of carbon and silicon, the higher oxides of manganese are also reduced if present in the ore. The aluminothermic process is used in the production of manganese metal.

1.10.4.6. Production of Manganese-containing Ferroalloys A number of manganese-containing ferroalloys are manufactured and used largely in the mild steel, foundry, and stainless steel industries. The names and typical compositions of these alloys are given in Table 1.10.3, and the international standards for the most commonly used alloys, namely high-carbon ferromanganese HC–FeMn and silicomanganese FeSiMn, are given in Table 1.10.4. Generally, high-carbon ferromanganese and silicomanganese are produced from a blend of manganese-containing ores, and in the case of silicomanganese, slags and silica are added. Ferromanganese can be produced in either electric submerged furnaces or blast furnaces, although only few blast producers exist in the world, whereas silicomanganese is largely produced in submerged arc furnaces. High-carbon ferromanganese can be converted to medium-carbon manganese by an oxygen-blowing process, and silicon manganese can be further refined into medium or LC ferromanganese as well as manganese metal (see Figure 1.10.13). 1.10.4.6.1 Production of Ferromanganese in Electric Arc Furnaces The majority of producers of ferromanganese in the world use submerged arc electric furnaces. Although electric furnaces have lower capacities than blast furnaces they have the advantage that the heat requirement is provided by electricity, and coke and coal are

Table 1.10.3 Types of Ferromanganese and Their General Compositions [29] Alloy

Composition % Manganese

Carbon

Silicon

High-carbon ferromanganese (carbure´)

72–80

7.5

<1.25

Medium-carbon ferromanganese (affine´)

75–85

<2.0

Low-carbon ferromanganese (surraffine´)

76–92

0.5–0.75

Silicomanganese

65–75

<2.5

15–25

Ferromanganese silicon

58–72

0.08

23–35

Spiegeleisen

16–28

<6.5

11–45

508

Rauf Hurman Eric

Table 1.10.4 Ferromanganese Standards [29] Alloy

Country or Organization

Standard

Reference

FeMn

International Standards Organization

DIS

5446

France

AFNOR NF

A-15-020

Japan

JIS

G2301

United States

ASTM

A99-66

Former Soviet Union

GOST

4755-70

Federal Republic of Germany

DIN

17564

International Standards Organization

DIS

5447

France

AFNOR NF

A-13-030

Japan

JIS

G2304

United States

ASTM

A701

Former Soviet Union

GOST

4756-70

Federal Republic of Germany

DIN

17564

FeSiMn

added to the feed only as reductants. Consequently, the coke consumption is lower in electric furnaces than in blast furnaces, which is a considerable advantage in the light of dramatically increasing coke prices. An additional advantage is that the process is not entirely dependent on high-strength coke unlike blast furnaces, and a portion of the carbon requirement can be supplied in the form of coal. In South Africa, where coking coal is in short supply, up to 70% of the carbon for the production of ferro- and silicomanganese is supplied in the form of bituminous coal. Originally electric furnaces were small (3–8 MVA); however, furnaces have grown progressively larger with time. More recently built electric furnaces have capacities of 75–90 MVA. Smaller furnaces are still popular with producers because they offer flexibility in that they can be switched more easily between different products than their larger counterparts. The larger size and more stable operation of modern electric furnaces, due largely to computer control, have resulted in lower electricity consumption. Electric furnaces used in the production of manganese alloys are generally circular and have three electrodes, each coupled to a separate electric phase (see Figure 1.10.14). The diameters of these furnaces range from 2 to 20 m. In modern electric arc furnaces, the raw material is usually fed by gravity from bunkers above the furnace. Fresh burden therefore automatically enters the furnace as the raw materials are melted and slag and metal are removed from the system. In older furnaces, charging cars are still used to introduce raw materials to the top of the units.

509

Production of Ferroalloys INTERMEDIATE PRODUCT

CONSUMPTION

END USE

Direct use

Pig iron

1%

0.5 kg/t of steel Mainly low grade ores; 1%

Ferromanganese

Steelmaking

92% high-carbon/ standard 8% medium- and low carbon FeMn

Direct use Ferromanganese Silicomanganese Manganese metal

Silicomanganese

Average Mn-content of steel: 0.7% Mn High-strength low-alloy steels: 1.5-2% Mn

90% Negligible 3.5 kg/t steel 0.5 kg/t 0.1 kg/t

84 %

13 %

2%

Manganese metal

2%

Chemical industry

3%

Uranium extraction Brick and ceramic coloring Chemical oxidizer and catalyst

Synthetic manganese dioxide

(a)

Copper and aluminum alloys

Batteries (dry cell)

80% electrolytic 20% chemical

4%

Metallurgical Grade Ores Power 2400 kWh/t Limestone

Coke

Electric Arc Furnace

Oxygen

Oxygen Furnace

Medium-carbon FeMn

High-carbon FeMn

Slag

Fines

Ferrosilicomanganese

Electric Arc Furnace

Nodulizing, Sintering

Mn Ore

Power 4000 kWh/t

Pelletizing Manganese Ore

Beneficiation

High-Silica Ores

Coke

Quartz Power

Arc or Ladle Furnace

Electric Arc Furnace

Medium-carbon FeMn Low-carbon FeMn Industrial Mn metal

Ferrosilicomanganese Power Sulfuric Acid

Dissolution Dissolution

(b)

Battery or Chemical Grades Ores

Manganese metal 99.5% pure (cathode)

Electrolysis Chemical precipitation

Electrolytic Manganese dioxide (EMD)(anode)

Chemical Manganese dioxide

Figure 1.10.13 a) Manganese intermediate products and end uses; b) Summary of process routes [29].

510

Rauf Hurman Eric

Figure 1.10.14 Layout of an electric arc furnace. (a) charging bins, (b) charging tubes, (c) electrodes, (d) electrodes slipping device, (e) electrode-positioning devices, (f) current transmission to electrodes, (g) electrode scaling, (h) furnace transformer, (i) current bus bar system, (j) furnace cover, (k) furnace shell, (l) top hole, (m) furnace bottom cooling, and (n) refractory material [29].

As the raw materials move down the furnace, the higher oxides of manganese are reduced to MnO by the gas leaving the furnace. The reduction of manganese(II) oxide occurs by the contact of carbon with the molten oxide in the slag phase. The overall reaction is: 10 1 C ! Mn7 C3 þ CO 7 7

ð1:10:70Þ

DG ¼ 265:7 0:18T ðkJ=molÞ

ð1:10:71Þ

MnO þ for which

The heat required for this endothermic reaction, for heating the burden, and to compensate for heat losses is supplied by the electrical input to the furnace. Heating takes

511

Production of Ferroalloys

place by the flow of electricity from the tips of the electrode, which are submerged in the burden, through the burden and slag to the metal, as well as through the flow of electricity between the electrodes. 1.10.4.6.2 Design and Operation of Electric furnaces The degree of heating depends on the electrical current flow as well as the resistance provided by the burden and the slag to the flow of electricity. In the production of ferromanganese, the resistivity of the burden is low; hence low voltages between the electrodes are necessary to maintain satisfactory penetration of the electrode in the charge. The vapor pressure of manganese is high; therefore, overheating of the charge must be avoided. The current densities on the electrodes should accordingly be lower for ferromanganese production than for other alloys. The diameters of the electrode are therefore larger in ferromanganese furnaces than in other ferroalloys furnaces to facilitate the high currents required for low-voltage operation. The distance between electrode centers is usually larger than in other furnaces; hence the furnace diameters tend to be greater. The values of these design parameters for a number of operating furnaces are given in Table 1.10.5. Most electric furnaces have two tap holes offset by 60 , which are used alternatively to tap both slag and metal. The slag and metal are then separated either in the ladle or by the means of a skimmer plate in the runner between the furnace and the ladle. In larger furnaces, separate tap holes are included for metal and slag. Tap holes are usually opened by tap-hole drills and closed with automatic mud guns. Table 1.10.5 Furnace Design Parameters [29] Elkem Elkem Beauharnois Sauda

Temco Samancor Bell Meyerton Bay M 10

Samancor Meyerton M4

Former Former Soviet Union Soviet Union PRO 2.5 PKZ 33

Inside shell 15.1 diameter (m)

12.5

10.0

16.0

9.8

2.7

8.7

Shell 8.8 height (m)

6.0

5.2

8.0

5.7

1.2

3.0

1.4

1.90

1.20

0.30

1.5

2.5

33

Electrode diameter (m)

1.90

1.90

Tapholes

2 metal, 2 slag

2 2 metal, metal, 1 slag 1 slag

2 metal, 1 slag

2 metal, 1 slag

Megawatt rating

4.5

30

46

20

13

512

Rauf Hurman Eric

An important feature in the design of submerged arc furnaces is the Soderberg electrodes. These are used because the larger electrode diameter required for the production of manganese alloys make the use of prebaked electrodes uneconomic. The Soderberg electrode consists of mild steel or stainless steel casing that is stiffened with internal fins and is filled with a carbonaceous paste, consisting of solid aggregate, usually calcined anthracite, and a binder of coal–tar pitch. The paste becomes plastic when hot and fills the entire volume of the casing. On further heating of the electrode by the electric current and furnace heat, the paste is baked and becomes solid. As the electrode is consumed, additional casings are welded onto the top. The current-carrying capacity and strength of an electrode is a function of the quality of the paste, the electrode baking rate, and the cross-sectional area. Breakages of Soderberg electrodes are a major cause of downtime in electric furnaces, and proper management of the electrodes is therefore essential for efficient production. A number of devices are commercially available to control the electrode movement and slipping rate (rate at which the electrode is allowed to move through the rings to compensate for its consumption in the furnace). One example case is designed and manufactured by Elkem. The electrode is clamped by hydraulically operated rings and is moved up and down on hydraulic cylinders. Current is fed to the electrode through brass contact shoes clamped around its diameter. Larger electric furnaces are usually completely closed at the top, and the CO-rich gas leaves the furnace at approximately 290 C and is cleaned in cyclones and venture scrubbers. The gas is then either flamed off to the atmosphere or, more recently, is used to generate electricity. At some plants, the off-gases are used to fuel auxiliary equipment in the plant. Smaller furnaces are either open or closed. In the case of open furnace, the gas is usually withdrawn by fans and cleaned in a bag filter plant. In this case, the gases are completely burnt in the furnace and have no commercial value. Raw materials are usually batch weighed and blended accordingly to a predetermined recipe and are then fed to bunkers above the furnace. The mix then gravitates into the furnace through feed chutes. To ensure an even distribution of material over the furnace, up to 10 feed chutes are radially distributed around each electrodes and one is positioned in the center of the furnace. After the metal is tapped from the furnace it is cast into molds formed from ferromanganese fines or cast irons and allowed to solidify. The alloy is then removed, crushed, and screened into various size fractions, depending on the requirement of the user. An alternative to this practice is the use of a casting machine. In this case, the metal is tapped directly onto a moving train consisting of small molds. The metal then solidifies and is ejected from the mold at the end of the strand. The advantage of this process is that the product is more even in size and cubical in shape. The generation of fines (6 mm), which are generally unsalable, is also minimized.

513

Production of Ferroalloys

1.10.4.6.3 Raw Materials Required for the Manufacture of HighCarbon Ferromanganese Manganese ores from different sources vary widely in their contents of manganese, iron, silica, alumina, lime, magnesia, and phosphorous. To produce standard ferromanganese (78% Mn) and a slag containing 30% MnO, the manganese-to-iron ratio in the charge must be 7.5:1. Since a single manganese ore of this ratio is seldom available, blending of ores from different sources is common practice to reach the desired manganese-to-iron ratio and to control the level of deleterious elements, particularly phosphorous. The use of sinter as a source of manganese is becoming increasingly popular. In the sintering process, a degree of prereduction is achieved reducing the energy requirement in the furnace. The additional advantage of sinter is that fine ores, which are otherwise unusable, are agglomerated in the sintering process. Bag-house dust and sludge from gas-cleaning plants can also be recycled to the furnace in the form of sinter. The maximum amount of sinter that can be fed to the furnace is a moot point and depends on its mineralogy and state of prereduction. When sinter replaces ore of high MnO2 content, the energy required in the furnace increases because the highly exothermic reduction of MnO2, Mn2O3, and Mn3O4 no longer takes place in the furnace. The bottom size of ore is also important because close packing of the ore in the furnace must be avoided. This can result in the formation of calcined bridges in the furnace, which disrupt the distribution of gas and can cause eruptions when the gas entrained under a bridge is suddenly released on its collapse. Generally, ores larger than 6 mm are used in large furnaces. The addition of small amounts of 6-mm material to small furnaces is possible. Generally, ore is screened prior to batch weighing of the furnace mix. The carbon required in the furnace is generally added in the form of coke. The size of the reductant is important. Coke and coal of too small a size can also cause close packing as well as affect the resistivity of the burden. Limestone, dolomite, or silica is added to the process as fluxes to adjust the basicity of slag. The amount and type of flux added depends on the blend of ores and whether a discard or high-slag practice is used. 1.10.4.6.4 Chemistry of the Process A dig-out of a 75-MVA ferromanganese furnace in the Republic of South Africa showed that nine distinct zones exist around each electrode (see Figure 1.10.15). This study showed that material descends rapidly down the side of the electrode (a) into the semi-active zone (b) where prereduction of higher manganese oxides to MnO takes place. Thereafter, the material moves into the active zone of the furnace (e, f), where reactions take place between the manganese(II) oxide in the melt and the coke particles in the coke bed: 7MnO þ 10C ! Mn7 C3 þ 7CO

ð1:10:72Þ

514

Rauf Hurman Eric

Figure 1.10.15 Zones in a ferromanganese furnace. The zones are (a) loosely sintered burden; (b) loosely sintered material enriched in carbonaceous reducing agents; (c) coke and slag region showing the active zone away from electrode; (d) coke bed; (e) coke enriched layer with CaO–MnO–SiO2 slag; (f) MnO melt layer with some slag, coke, and additional carbonaceous reducing agent; (g) ferromanganese alloy layer intermixed with MnO melt; (h) graphite and carbon-rich material; (i) carbon lining; (j) brick lining; (k) slag tap hole level; (l) metal tap hole level; and (m) pieces of electrode [29].

Equilibrium between the slag and metal was thought to exist under each electrode, and further from the electrode, layers of unreacted ore and coke were found to be present (h). This suggests that heat is concentrated under each electrode. The path of electrical transfer was deduced to be from the electrode tip through the coke bed and into the alloy layer (g). The efficient production of high-carbon ferromanganese therefore depends on the degree of reduction of MnO by carbon as well as the prereduction that occurs in the upper region of the furnace. The ratio of CO to CO2 in the off-gas is important and can be used to monitor the condition of the furnace. The higher the CO2 content of the off-gas, the higher is the energy efficiency of the process because the reducing potential of the gas is being more fully utilized (see Figure 1.10.16). Good operation of the furnace is indicated by a CO2/(CO2 þ CO) ratio of 0.55. This ratio, as well as the MnO content of the slag, can be used to control the coke rate of the furnace. Undercoking of the furnace is indicated by high MnO content of the slag and a low CO2 content in the off-gas. A further influence on the MnO content of the slag is the basicity ratio (CaO þ MgO)/SiO2. Generally the addition of basic oxides increases the melting point

515

Production of Ferroalloys

3000

Power consumption, kWh/t

2800

2600

2400

2200

2000

1800

10

30

50

70

90

CO2/(CO + CO2) in Furnace off-gas, %

Figure 1.10.16 Relationship between ferromanganese production [29].

off-gas

composition

and

power

consumption

in

of the slag. The hotter slag improves the reaction between the slag and the coke and, consequently, more MnO is reduced. Increasing the basicity of the slag thus decreases the residual MnO content. The MnO content of the slag is also reduced by increasing the penetration of the electrodes, which also increases the slag temperature. The target MnO content of the slag depends on whether a discard or high-slag process is used. 1.10.4.6.4.1 Discard Slag Practice

Discard slags generally have MnO contents of 8–12%. Slag is produced by the silica and other basic oxides entrained in the ore. By increasing the basicity of the slag, the recovery of manganese as metal is increased; however, consumption of carbon and electricity also increases. The discard slag process is therefore only used where power is relatively cheap and delivered cost of manganese ore is high due to high transport costs. The recovery of manganese to the metal is between 70% and 75% when this practice is used. The practice usually involves the addition of limestone or dolomite to the furnace. 1.10.4.6.4.2 High-Slag Practice

In the high-slag practice, less coke is required for reduction and little or no basic fluxes are used because the MnO content of the slag satisfies the requirement for basic oxides. Slags of this nature tend to contain more than 25% MnO. In the high MnO practice, the power

516

Rauf Hurman Eric

consumption is reduced because a higher proportion of the reduction occurs by the gases and less MnO is reduced by solid carbon. Maximum use is therefore made of the exothermic nature of the prereduction reactions. The recovery of manganese is low in the ferromanganese furnace, but the overall recovery of manganese is high because the slag is usually used for the production of silicomanganese. An additional attraction of the high-slag process is that an artificial ore, with manganese-to-iron ratio of up to 100:1, can be made from relatively low-grades ores. This artificial ore can then be used to produce the highest grade of ferromanganese without the need to purchase costly high-grade ores. An additional benefit is the extremely low phosphorous content of the slag, which hence lowers the phosphorous content of any mix in which it is used. In countries having only ores with high phosphorus contents, the first stage of the process is the production of a slag high in MnO and a low manganese alloy. Since the energy input in the production of ferromanganese by the high-slag process is lower than that of the discard slag process, it is used by most producers. 1.10.4.6.5 Production of Silicomanganese Unlike ferromanganese, silicomanganese is only produced in electric arc furnaces, most of which can be used interchangeably to produce either of the manganese-containing alloys. Silicomanganese is used either as a substitute for ferromanganese and ferrosilicon in steel making or as a raw material for the production of medium and LC ferromanganese and industrial manganese metal. Although silicomanganese generally contains 14–19% Si, grades containing up to 35% are produced for the production of extremely LC alloys. The solubility of carbon in the alloy decreases with increasing silicon content (Figure 1.10.17). On cooling, sparingly soluble SiC comes out of solution. 7 6 5 Equilibrium with C Carbon (%)

4 3 2 Equilibrium with SiC

1 0

5

10 15 Silicon (%)

20

25

30

Figure 1.10.17 Carbon solubility in the Fe–Mn–Si–C system (50–80% Mn) at 1420 C [29].

517

Production of Ferroalloys

There are three general routes for the production of silicomanganese: (1) Reduction of manganese ores and silica with coke and coal. (2) Reduction of MnO-rich slags from ferromanganese production and quartzite with coke and coal. (3) Reaction of standard ferromanganese and quartzite with coke. The first two processes are used for the production of alloys containing 15–25% Si and can be carried out in the same furnaces used for ferromanganese production. The third method is used to produce alloys containing 30–35% Si and is generally performed in smaller furnaces. 1.10.4.6.5.1 Raw Materials

The raw materials used in silicomanganese production are similar to those used in making ferromanganese. Silica is added to the furnace as quartz or quartzite, and ferromanganese slag can be used as an alternative additional source of manganese instead of manganese ore or sinter. As in the case of ferromanganese, production advantages can be gained by using sintered ore. 1.10.4.6.5.2 Chemistry of the Process

In a dig-out of a small electric arc furnace, four zones were distinguished: the burden zone, the zone of the coke bed, the melting zone, and the metal layer. In the burden zone, the higher oxides of manganese are reduced to MnO while the higher oxides of iron are reduced to FeO and partially to metallic iron. Manganese(II) oxide is converted to complex silicates, which begin to melt at the bottom of the burden zone. Fine metallic particles exist in the coke bed zone, which are possibly caused by condensation of silicon and manganese in the hot areas under electrodes. In the upper and lower parts of the melt zone, the reduction of manganese and silicon oxides occur. The equilibrium is determined by the following reaction: SiO2 þ 2Mn ! Si þ 2MnO

ð1:10:73Þ

This equation is important in determining the silicon and manganese contents of the metal that is collected in the lowest part of the furnace and is influenced by the slag chemistry and temperature of the process. Increasing the CaO content of the slag reduces the silicon content of the metal. The basicity requirement of the slag is therefore better supplied in the form of MgO. Higher temperature tends to drive Si into the metal at the expense of Mn. Higher temperatures are therefore required in the production of silicomanganese than in the production of ferromanganese. To produce manganese metal of 97% Mn by the silicothermic method, a silicomanganese containing 28% Si is required that is particularly low in phosphorous and iron. This is made from a manganese slag produced from the partial reduction of manganese ores.

518

Rauf Hurman Eric

1.10.4.6.5.3 Operation of the Furnace

The operation of the furnace for silicomanganese production is similar to that of ferromanganese production. However, deeper penetration of the electrode is necessary to provide the higher temperature required to drive the reduction of silicon. The resistivity of the burden is therefore important, and the size and the activity of the reductant are critical for stable operation of the furnace. 1.10.4.6.6 Production of Medium-Carbon Ferromanganese Medium-carbon ferromanganese contains 1–1.5% carbon and has a manganese content of 75–85%. Medium-carbon ferromanganese can be produced either by refining highcarbon ferromanganese with oxygen or by the silicothermic route, whereby the silicon in silicomanganese is used to reduce additional MnO added as ore or slag. The former process has considerable advantage and is used by most producers. The silicothermic and oxygen-based processes are shown schematically in Figure 1.10.18.

Silicothermic reduction

MOR Process Electrodes Mn ore MOR fume

FeMn slag Electrodes

Flux

Mn ore

Coke

Lime

OPEN ARC FURNACE

SUBMERGED ARC FURNACE 8% Mn slag

Ore-Lime Melt

MOR slag Power

Coal

Power

SUBMERGED ARC FURNACE 15% Mn slag To HC FeMn HC FeMn To SAF Oxygen

TRANSFER LADLE Fume

High-Mn Slag

Slag

SiMn REACTION LADLE 2

REACTION LADLE 1

MC FeMn

High-Silicon Alloy

Mn Slag

LADLE (BLOWING VESSEL)

MC FeMn

To SiMn FURNACE

Figure 1.10.18 Comparison of process flow sheets for silicothermic reduction and the MOR process [29].

519

Production of Ferroalloys

1.10.4.6.6.1 Production of Medium-Carbon Ferromanganese by Oxygen refining of HighCarbon Ferromanganese

In the oxygen-based process (patented by Union Carbide as the “MOR”—manganese– oxygen refining—process), high-carbon ferromanganese is decarburized in a similar manner to the steelmaking process in the basic oxygen furnace [29]. However, several distinctive differences are encountered in the case of ferromanganese: (1) a final temperature of 1750 C compared to 1550 C, (2) refractory attack is more severe, (3) difficult casting of the final alloy, (4) the higher vapor pressure of manganese, (5) the higher volume and temperature of the off-gas. In the MOR process, oxygen is blown into the molten high-carbon ferromanganese and the temperature is increased from its tapping value of 1300–1750 C. The heat required is supplied by the oxidation of manganese to manganese(II) oxides and carbon to carbon monoxide. The need to increase the temperature is shown by the carbon-temperature relationship in Figure 1.10.19. In the early part of the blowing process, most of the oxygen is consumed by oxidation of manganese and the temperature of the melt increases from 1300 to 1550 C. Thereafter, carbon is rapidly oxidized and the temperature rises to 1650 C. Above this temperature, the rate of carbon removal decreases and manganese is once again oxidized. The process is stopped at 1750 C, which corresponds to a carbon content of 1.3%. Further reductions in carbon content result in unacceptably high losses of manganese. In the MOR process, the recovery of manganese is about 80% and the distribution of manganese can be broken down as follows: Alloy MC FeMn 80% Fume formed by vaporization 13% Slag formed by oxidation of Mn 5% Other losses, splashing 2% 1800

Temperature (°C)

1700 1600 1500 1400 1300 7

6

5

4

3

2

1

Carbon (%)

Figure 1.10.19 Dependence of carbon content on temperature for a ferromanganese alloy (Mn: Fe ¼ 6:1) at 101.3 kPa [29].

520

Rauf Hurman Eric

The manganese lost in the fume is recovered in the gas-cleaning plant and is then pelletized and returned to the high-carbon ferromanganese furnace. The slag, which contains about 65% MnO, is also returned to the high-carbon ferromanganese furnaces. The successful operation of this process depends on the design of the blowing vessel and oxygen lance as well as giving careful attention to operational procedures. The MOR process has many advantages over the silicothermic process: lower energy consumption, lower capital investment, lower production cost, and greater flexibility. The main disadvantage of the process is that its use is limited to production of medium-carbon ferromanganese because the carbon content cannot be reduced to below 1.3%. 1.10.4.6.6.2 Silicothermic Production of Medium-Carbon Ferromanganese

In the silicothermic production of medium-carbon ferromanganese, a high-grade slag or a melt containing manganese ore and lime is contacted with silicomanganese containing 16–30% silicon. The silicon in the alloy acts as the reducing agent in the process, which reduces the manganese(II) oxide in the melt. Similar to silicomanganese production, the equilibrium is determined by the reaction: Si þ 2MnO ¼ SiO2 þ 2Mn

ð1:10:74Þ

The purpose of the lime addition is to reduce the activity of the SiO2 in the melt, thus forcing the above reaction as far to the right as possible. The ratio of CaO to SiO2 in the slag should be greater than 1.4 to ensure a sufficient reduction in the activity of SiO2. The carbon entering the process in the silicomanganese remains entirely in the metal phase and is therefore found in the product. Thus to produce a medium-carbon ferromanganese containing 1% C, a silicomanganese containing 20% Si is necessary (see Figure 1.10.17). The heat produced by the silicothermic reduction is not sufficient to sustain the process; hence it is usually carried out in an electric arc furnace. These furnaces are usually small and, unlike ferromanganese furnaces, are lined with magnesia bricks, which are fairly resistant to highly basic slag. The power consumption is between 1000 and 3000 kW h/ton. These furnaces can be tilted so that the slag can be separated from the metal. Although the silicothermic reduction process is more energy intensive than the decarburization of high-carbon ferromanganese, it has the advantage that the final carbon content is limited only by the carbon content of initial silicomanganese. The silicothermic process can therefore be used to produce LC ferromanganese and industrial manganese metal. 1.10.4.6.7 Production of LC Ferromanganese LC ferromanganese contains 76–92% Mn and 0.5–0.75% C. The production of LC silicomanganese is not possible by the decarburization of high-carbon ferromanganese

521

Production of Ferroalloys

without incurring extremely high losses of manganese. Therefore silicothermic reduction process is used. The process is similar to that used in the silicothermic production of medium-carbon ferromanganese. High purity ores are used and in particular ores containing iron and phosphorus should be avoided. An artificial manganese ore produced as a high-grade slag is particularly suitable because of its low impurity level and because all the manganese is present as MnO. The reduction of the higher oxides of manganese is therefore unnecessary. The operating figures for 1 ton of ferromanganese containing 85–92% Mn, 0.1% C, and 1% Si with a manganese recovery of 75% are: Calcined manganese ore

1250–1350 kg

Silicomanganese (32–33%Si)

800–850 kg

Burnt lime

1000–1100 kg

Electrodes

10–12 kg

Electricity

1800–2500 kW h

Since the required silicon content of the metal is low, a slag high in MnO is necessary. The MnO content of the slag can, however, be reduced by use of a two-stage refining operation. In the first stage, an excess of silicomanganese is maintained and a slag containing 6–8% Mn is teemed and discarded. The second refining stage with manganese ore and lime results in a slag containing 10–14% Mn, which is returned to the silicomanganese furnace. 1.10.4.6.8 Gas Cleaning In all the processes described in this chapter, large volumes of gas are generated. These gases consist generally of CO, CO2, and N2, and contain large quantities of dust from the raw materials and condensed manganese droplets. These gases require cleaning prior to their venting to the atmosphere. In open furnaces, in which the gas is totally combusted, hoods are incorporated in the design through which the gas is removed. This gas is cooled in trombone coolers to around 200 C and the dust is removed in bag filters. Ideally, after recovery, the dust should be agglomerated and returned to the furnace. In closed furnaces, the gas is usually cleaned in cyclones and venturi scrubbers prior to combustion to the atmosphere or used in downstream processes. The emissions from the gas plants are less than 50 mg/m3, which is statutory maximum. It is expected that dust limits will be reduced to 25 mg/m3. In addition to this, the exposure limits allowed in the working environment in manganese-producing facilities have been reduced due to the toxicity of manganese dust; the following limits apply:

522

Rauf Hurman Eric

United States, United Kingdom, Australia, Belgium, Brazil, Federal Republic of Germany

5 mg/m3

Former Yugoslavia

2 mg/m3

Former Soviet Union, Poland

0.3 mg/m3

Bulgaria

0.02 mg/m3

1.10.4.6.9 Developments and Future Trends 1.10.4.6.9.1 Energy Saving Measures

In regions with high electricity prices, developments have concentrated on the saving of energy. This is of particular importance to Japanese producers of manganese alloys that use electric furnaces. Developments include the use of the off-gas from the furnace to preheat and mildly prereduce the ore, either in a rotary kiln or in a shaft above the furnace. In the process in which a shaft kiln is positioned above the furnace, the gas from the furnace is burnt under the furnace roof and the hot burnt gas leaves the furnace through the vertical shaft. The raw ore mix is introduced to the top of the shaft and is heated in a countercurrent fashion by the exhaust gas. The ore is then fed to the furnace, where further heating takes place as the ore is exposed to the radiant heat produced by the burning gas in the roof. 1.10.4.6.9.2 Use of Plasma Furnaces

Nontransferred arc plasma furnaces have been used successfully in the production of charge chrome. Attempts to produce ferromanganese in plasma furnaces have not been economically successful due to the high losses of manganese caused by volatilization in the arc zone of the submerged arc furnace. In an effort to solve these problems and to utilize the higher specific throughputs that are obtainable in a plasma unit, a combined plasma/shaft furnace is under consideration.

1.10.4.7. Production of Silicon Alloys This section briefly deals with the manufacture of silicon-rich alloys, including all the ferrosilicon alloys that are made in the submerged arc furnace [31]. These range from metallic silicon of þ98%Si content down to low-grade ferrosilicon of 16–18%Si content. 1.10.4.7.1 Raw Materials The principal raw material is the oxide SiO2 normally referred to as silica, and this occurs widely in nature in many forms. The minerals that are principally of interest in this connection are crystalline quartz, quartzites, and certain sandstones. When making ferrosilicon, the minimum requirement is generally þ96% SiO2 with certain limitations on the alumina, calcium, and phosphorus contents. For making silicon metal, however, the SiO2 content must be in excess of 99% with control of the Al, Ca, and P contents depending upon the grade of silicon metal to be made. One important requirement is that the silica

523

Production of Ferroalloys

must not unduly decrepitate on heating and this appears to be related with the content of combined water in the rock. Another important consideration is the amount of fines produced on crushing as this will affect the amount of rock that has to be mined to produce the furnace feed, which should be approximately 75 25mm in size with all the fines (9 mm) screened out. Fines cannot be tolerated in an electric furnace mix because they segregate and more importantly they reduce porosity of the charge causing premature fusion and crusting, which results in a buildup of gas pressure, and the production of “blows.” These take the form of sudden release of high-temperature gas consisting of CO and SiO, which means a loss of silicon and reduction of overall efficiency. As to impurities, iron is not important when making ferrosilicon but for silicon metal, the iron content of the silica must be such that the final product is within specification, bearing in mind that 100% of the iron will report in the final product metal. Phosphorus is also mostly reduced and about two thirds pass into the metal or alloy. Although most forms of silica are generally very low in phosphorus, it is important to ensure that the ash in the carbonaceous reductant does not contain sufficient P to make the alloy unacceptable with respect to this element. Both calcium oxide and alumina are partially reduced when making silicon alloys, the former to the extent of 40–50% depending upon the grade of the alloy, and the latter to the extent of 55–60%. In both cases these elements can be removed to a considerable extent by treating the molten metal. In the making of ferrosilicon alloys, the iron is mostly added as scrap, although occasionally a little mill scale may be used. Iron ore is rarely used because it tends to form slag with the silica before the iron is completely reduced and it can introduce unwanted elements such as phosphorus, calcium, and aluminum. The iron scrap must also be free from elements such as copper, tin, chromium, nickel, and the like. Ahostofcarbonaceousreductantsareinuse.Fromachemicalpointofview,thefixedcarbon contentisofprimaryimportanceasistheanalysisoftheashcomponent.Fromfurnaceoperation point of view, one of the most important factors is the electrical resistivity of the reductant, which affects the overall charge resistance. This governs the penetration of the electrodes into the charge at a given operating voltage and is important in obtaining optimum efficiency. 1.10.4.7.2 Fundamental Aspects The overall reaction for the reduction of silica with carbon is simple, but it involves the absorption of considerable quantity of heat as well as attainment of very high temperature in order that the reaction: SiO2 þ 2C ¼ Si þ 2CO

ð1:10:75Þ

shall proceed to the right. The standard Gibbs free energy change is zero at 1937 K so that a temperature well in excess of this is needed to drive the reaction in a forward direction. In reality, however, the above reaction does not represent the actual mechanism of the reduction process that occurs through a number of intermediate ones, the principal ones being:

524

Rauf Hurman Eric

SiO2 þ 3C ¼ SiC þ 2CO

ð1:10:76Þ

SiO2 þ C ¼ SiOðgÞ þ CO

ð1:10:77Þ

and

It is noteworthy to see that reaction (1.10.77) produces silicon monoxide gas at high temperatures which due to its gaseous nature may result in silicon losses if not properly handled and engineered during the process in the submerged arc furnace. Reaction (1.10.76)-producing silicon carbide is well-known to occur and accretions of silicon carbide are found in the cooler parts of a furnace when it is shut down and dug out. The reaction is favored by an excess of carbon in the charge. The reason why silicon carbide is not found in the hotter parts of the furnace is probably because it reacts at a high temperature with silicon monoxide as well as with silica itself: SiC þ SiOðgÞ ¼ 2Si þ CO 2SiC þ SiO2 ¼ 3Si þ 2CO

ð1:10:78Þ ð1:10:79Þ

The silicon monoxide reaction (1.10.77) is favored by a deficiency of carbon, and in a furnace operated with a cool top much of this is condensed on the carbon particles to be reduced to metal on their descent in the furnace. It can thus be postulated that the overall reaction probably takes place in two stages, namely the formation of silicon carbide in the upper relatively cooler parts of the charge followed by reaction with silicon monoxide as well as with silica in the hotter regions in the vicinity of electrodes, eventually producing liquid silicon. In the making of ferrosilicon, the reduction process is facilitated because the solution of silicon in liquid iron is a process with a favorable free energy change as well as with an exothermic enthalpy change, so the reduction can take place at a lower temperature; for example, the change in free energy where pure liquid silicon dissolves to give a 1% solution: SiðlÞ ¼ Sið1%Þ DG ¼ 119,240 25:48T ð J=molÞ

ð1:10:80Þ

at smelting temperatures. Obviously for concentrated solutions, the activity of silicon dissolved in iron needs to be taken in account in calculating the free energy change in any particular set of conditions. In fact, when making low silicon alloys (high Fe content-dilute solutions), the presence of silicon carbide is not detected in contrast to the making of silicon metal. The occurrence of silicon carbide, especially on the hearth, will require the charging of an excess of silica for a time in an attempt to clear and eventually convert it to silicon metal or the charging of iron oxide (generally mill scale) and a reversion of the process to manufacturing ferrosilicon until the SiC accretions have been eliminated. 1.10.4.7.3 Physical Factors The melting points of ferrosilicon alloys are lower than that of iron, mostly below 1673 K [28]. Because the reduction temperature is well above the liquidus, there is

Production of Ferroalloys

generally no difficulty in pouring these alloys, and it is normal to tap the metal into a ladle and to pour it into molds. Ferrosilicon alloys absorb gases, especially hydrogen during manufacture, and when cast it is usual to find that ingot has blow holes due to the alloy having given up some of its gas content on solidification. This is particularly noticeable in the 75% Si-containing alloy. Because the pouring temperature is usually well above the solidification temperature there is a tendency of gravitational segregation in ferrosilicon ingots. To avoid this, ferrosilicon is normally cast into large flat slabs. Significant progress has been made in continuously casting a ferrosilicon ingot. All these alloys, as well as silicon metal, are brittle and the large slabs can be easily broken prior to crushing to size. 1.10.4.7.4 Operating Practice The resistance of the charge in smelting ferrosilicon is considerably more than that obtained in ferromanganese operations. This means that higher voltages can be used, permitting the operation of larger furnaces. As is well known, high silicon-containing alloys do not absorb carbon to any appreciable extent, so that carbon is the natural refractory material for the hearth and the lower side walls of silicon alloy smelting furnaces. The rest of the furnace is generally lined with a high-grade firebrick. The carbon lining can be of prebaked carbon blocks or can be of carbon paste carefully rammed and baked in. At smelting temperatures of the order of 1973–2073 K, these alloys are extremely fluid and it is important to ensure a “liquid tight” bottom. Furnaces used for the smelting of high-silicon alloys, i.e., over 70% Si, are mostly of the open-top type, and the carbon monoxide generated is burnt on the top of the furnace. Recently attempts to recover this gas and to close the top of the furnace have been intensified both from emissions and energy recovery points of view. The main problem with respect to closing the top of the furnace was the fact that direct accessibility to the charge from the top was necessary from time to time for poking the charge material to keep it porous and to break up the crust formations. The same operating principles apply to the smelting of silicon metal as to ferrosilicon, but it is a much more difficult product to make because there is no iron to collect the reduced silicon. The carbon control must be more stringently applied because an error of over- or under-coking is more difficult to correct. It should also be mentioned that the overall recovery of silicon is lower around a maximum of 80%, compared with say 75% ferrosilicon practice where it is commonly around 90%. 1.10.4.7.5 Ferrosilicon Grades and Compositions There is a considerable range of ferrosilicon alloys manufactured from as low as 10% Si to 95% and to pure silicon metal. Table 1.10.6 gives a summary of the grades and their respective compositions [32].

525

526

Rauf Hurman Eric

Table 1.10.6 Ferrosilicon Types and Respective Compositions (Maximum Contents for All Elements Except Si and Al—Balance Fe) [32] FeSi Type

wt.% Si

Al

P

S

C

Mn

Cr

Ti

0.15

0.06

2.0

3.0

0.8

0.30

0.30

FeSi10

8.0–13.0

<0.2

FeSi15

14.0–20.0

<1.0

1.5

1.5

FeSi25

20.0–30.0

<1.5

1.0

1.0

FeSi45

41.0–47.0

<2.0

0.20

1.0

0.5

FeSi50

47.0–51.0

<1.5

0.05

0.8

0.5

FeSi65

63.0–68.0

<2.0

0.04

0.4

0.4

FeSi75Al1

72.0–80.0

<1.0

0.5

0.3

FeSi75Al1.5

1.0–1.5

FeSi75Al2

1.5–2.0

FeSi75Al3

2.0–3.0

FeSi90Al1

87.0–95.0

<1.5

0.05

0.05

0.05

0.04

0.15

0.20

0.3 0.20

0.3

0.30

0.5 0.04

0.04

0.15

0.5

0.2

0.30

1.10.5. GENERAL PROCESS DESCRIPTION In this section, the more general aspects of the ferroalloy manufacturing processes will be briefly discussed independent of the specific alloys produced, which will include descriptions of the submerged arc furnace, exothermic–metallothermic reduction processes as well as emissions and controls [33]. A variety of furnace types, including submerged electrical arc furnaces, exothermic (metallothermic) reaction furnaces, and electrolytic cells, can be used to produce ferroalloys. Furnace descriptions and their ferroalloy products are given in Table 1.10.7.

1.10.5.1. Submerged Electric Arc Process In most cases, the submerged electric arc furnace produces the desired product directly. It may produce an intermediate product that is subsequently used in additional processing methods. The submerged arc process is a reduction smelting operation. The reactants consist of metallic ores (ferrous oxides, silicon oxides, manganese oxides, chrome oxide, etc.) and a carbon-source reducing agent, usually in the form of coke, charcoal, high- and low-volatility coal, or wood chips. Limestone may be also be added as a flux material. Raw materials are crushed, sized, and, in some cases, dried, and then conveyed to a mix house for weighing and blending. Conveyors, buckets, skip hoists, or cars transport

527

Production of Ferroalloys

Table 1.10.7 Ferroalloy Processes and Respective Groups [33] Process

Product

Submerged arc furnace

a

Silvery iron (15–22% Si) Ferrosilicon (50% Si) Ferrosilicon (65–75% Si) Silicon metal Silicon/manganese/zirconium (SMZ) High-carbon (HC) ferromanganese Silicon manganese HC ferrochrome Ferrochrome/silicon FeSi (90% Si)

Exothermic

b

Silicon reduction

Low-carbon (LC) ferrochrome, LC ferromanganese, mediumcarbon (MC) ferromanganese

Aluminum reduction

Chromium metal, ferrotitanium, ferrocolumbium, ferrovanadium

Mixed aluminothermal/ silicothermal

Ferromolybdenum, ferrotungsten

Electrolyticc

Chromium metal, manganese metal

Vacuum furnaced

LC ferrochrome e

Induction furnace

Ferrotitanium

a

Process by which metal is smelted in a refractory-lined cup-shaped steel shell by submerged graphite electrodes. Process by which molten charge material is reduced, in exothermic reaction, by addition of silicon, aluminum, or a combination of the two. c Process by which simple ions of a metal, usually chromium or manganese in an electrolyte, are plated on cathodes by direct low-voltage current. d Process by which carbon is removed from solid-state high-carbon ferrochrome within vacuum furnaces maintained at temperatures near melting point of alloy. e Process that converts electrical energy into heat, without electrodes, to melt metal charges in a cup or drum-shaped vessel. b

the processed material to hoppers above the furnace. The mix is then gravity-fed through a feed chute either continuously or intermittently, as needed. At high temperatures in the reaction zone, the carbon source reacts with metal oxides to form carbon monoxide and to reduce the ores to base metal. Smelting in an electric arc furnace is accomplished by conversion of electrical energy to heat. An alternating current applied to the electrodes causes current to flow through

528

Rauf Hurman Eric Carbon electrodes

Figure 1.10.20 Typical submerged arc furnace design [33].

the charge between the electrode tips. This provides a reaction zone at temperature up to 2000 C. The tip of each electrode changes polarity continuously as the alternating current flows between the tips. To maintain a uniform electric load, electrode depth is continuously varied automatically by mechanical or hydraulic means. A typical submerged electric arc furnace design is depicted in Figure 1.10.20. The lower part of the submerged electric arc furnace is composed of a cylindrical steel shell with a flat bottom or hearth. The interior of the shell is lined with two or more layers of carbon blocks. The furnace shell may be water cooled to protect it from the heat of the process. A water-cooled cove and fume collection hood are mounted over the furnace shell. Normally, three carbon electrodes arranged in a triangular formation extend through the cover and into the furnace shell opening. Prebaked or self-baking (Soderberg) electrodes ranging from 76 to over 100 cm (30–40 in.) in diameter are typically used. Raw materials are sometimes charged to the furnace through feed chute from above the furnace. The surface of the furnace charge, which contains both molten material and unconverted charge during operation, is typically maintained near the top of the furnace shell. The lower ends of the electrodes are maintained at about 0.9–1.5 m (3–5 ft.) below the charge surface. Three-phase electric current arcs from electrode to electrode, passing through the charge material. The charge material melts and reacts to form the desired product as the electric energy is converted into heat. The carbonaceous material in the furnace charge reacts with oxygen in the metal oxides of the charge and reduces them to base metals. The reactions produce large quantities of carbon monoxide (CO) that pass upward through the furnace charge. The molten metal and slag are removed (tapped) through one or more tap holes extending through the furnace shell at

Production of Ferroalloys

the hearth level. Feed materials may be charged continuously or intermittently. Power is applied continuously. Tapping can be intermittent or continuous based on production rate of the furnace. Submerged electric arc furnaces are of two basic types, open and closed. Open furnaces have a fume collection hood at least 1 m (3.3 ft.) above the top of the furnace shell. Moveable panels or screens are sometimes used to reduce the open area between the furnace and hood and to improve emissions-capture efficiency. Carbon monoxide rising through the furnace charge burns in the area between the charge surface and the capture hood. This substantially increases the volume of gas the containment system must handle. Additionally, the vigorous open combustion process entrains finer material in the charge. Fabric filters are typically used to control emissions from open furnaces. Closed furnaces may have water cooled steel cover that fits closely to the furnace shell. The objective of covered furnaces is to reduce air infiltration into the furnace gases, which reduces combustion of that gas. This reduces the volume of gas requiring collection and treatment. The cover has holes for the charge and electrode to pass through. Covered furnaces that partially close these hood openings with charge material are referred to as “mix-sealed” or “semi-closed furnaces.” Although these covered furnaces significantly reduce air infiltration, some combustion still occurs under the furnace cover. Covered furnaces that have mechanical seals around the electrode and sealing compounds around the outer edges are referred to as “sealed” or “totally closed.” These furnaces have little, if any, air filtration and undercover combustion. Water leaks from the cover into the furnace must be minimized as this leads to excessive gas production and unstable furnace operation. Products prone to highly variable releases of process gases are typically not made in covered furnaces for safety reasons. As the degree of enclosure increases, less gas is produced for capture by the hood system and the concentration of carbon monoxide in the furnace increases. Wet scrubbers are used to control emissions from covered furnaces. The scrubbed high carbon monoxide content gas may be used within the plant or for electricity generation. The molten alloy and slag that accumulate on the furnace hearth are removed at 1–5 h intervals through the tap hole. Tapping typically lasts 10–15 min. Tap holes are opened with pellet shot from a gun, by drilling, or by oxygen lancing. The molten metal and slag flow from the tap hole into a carbon-lined trough, then into a carbon-lined launder that directs the metal and slag into a reaction ladle, ingot molds, or chills (chills are low, flat iron, or steel pans that provide rapid cooling of the molten metal). After tapping is completed the furnace is resealed by inserting a carbon paste plug into the tap hole. Chemistry adjustments may be necessary after furnace tapping in the ladle to achieve a specified product. Ladle treatment reactions are batch processes and may include metal and alloy additions. During tapping and/or in the reaction ladle, slag is skimmed from the surface of the molten metal. It can be disposed of in landfills, sold as road ballast, or used as raw material in a furnace or reaction ladle to produce a chemically related ferroalloy product.

529

530

Rauf Hurman Eric

After cooling and solidifying, the large ferroalloy castings may be broken with drop weights or hammers. The broken ferroalloy pieces are then crushed, screened (sized), and stored in bins until shipment. In some instances, the alloys are stored in lump form in inventories prior to sizing for shipping.

1.10.5.2. Exothermic (Metallothermic) Process The exothermic process is generally used to produce high-grade alloys with low-carbon content. The intermediate molten alloys used in the process may come directly from submerged electric arc furnace or from another type of heating device. Silicon or aluminum combines with oxygen in the molten alloy, resulting in a sharp temperature rise and strong agitation of the molten bath. Low- and medium-carbon content ferrochromium (FeCr) and ferromanganese (FeMn) are produced by silicon reduction. Aluminum reduction is used to produce chromium metal, ferrotitanium, ferrovanadium, and ferrocolumbium (ferroniobium). Mixed alumino/silicothermal processing is used for producing ferromolybdenum and ferrotungsten. Although aluminum is more expensive than carbon or silicon, the products are purer. LC ferrochromium is typically produced by fusing chromium ore and lime in a furnace. A specified amount is then placed in a ladle (ladle no. 1). A known amount of intermediate-grade ferrochrome silicon is then added to the ladle. The reaction is extremely exothermic and liberates chromium from its ore, producing LC ferrochromium and a calcium silicate slag. This slag, which still contains recoverable chromium oxide, is reacted in second ladle (ladle no. 2) with molten high-carbon ferrochrome silicon to produce the intermediate-grade ferrochrome silicon. Exothermic processes are generally carried out in open vessels and may have emissions similar to the submerged arc process for short periods while the reduction is occurring.

1.10.5.3. Emissions and Controls Particulate is generated from several activities during ferroalloy production, including raw material handling, smelting, tapping, and product handling. Organic materials are generated almost exclusively from the smelting operation. The furnaces are the largest potential sources of particulate and organic emissions. Particulate emissions from electric arc furnaces in the form of fumes account for an estimated 94% of the total particulate emissions in the ferroalloy industry. Large amounts of carbon monoxide and organic materials also are emitted by submerged electric arc furnaces. Carbon monoxide is formed as a byproduct of the chemical reaction between oxygen in the metal oxides of the charge and carbon contained in the reducing agent (coke, coal, etc.). Reduction gases containing organic compounds and carbon monoxide continuously rise from the high-temperature reaction zone, entraining fine particles and fume precursors. The mass weight of carbon monoxide produced sometimes exceeds that of the metallic product. The heat-induced fume consists of oxides of the products being produced and carbon from the reducing agent. The fume is enriched by silicon dioxide, calcium oxide, and magnesium oxide, if present in the charge.

Production of Ferroalloys

In an open electric furnace, virtually all carbon monoxide and much of the organic matter burns with induced air at the furnace top. The remaining fume, captured by hooding about 1 m above the furnace, is directed to a gas-cleaning device. Fabric filters are used to control emissions from 85% of the open furnaces, scrubbers are used on 13% of the furnaces, and electrostatic precipitators on 2%. Two emissions capture systems, not usually connected to the same gas-cleaning device, are necessary for covered furnaces. A primary capture system withdraws gases from beneath the furnace cover. A secondary system captures fumes released around the electrode seals and during tapping. Scrubbers are used almost exclusively to control exhaust gases from sealed furnaces. The scrubbers capture a substantial percentage of the organic emissions which are much greater for covered furnaces than open furnaces. The gas from sealed and mix-sealed furnaces is usually flared at the exhaust of the scrubber. The carbon monoxide-rich gas is sometimes used as a fuel in the kilns and sintering machines. The efficiency of flares for the control of carbon monoxide and reduction of volatile organic compounds has been estimated to be greater than 98%. A gas heating reduction of organic and carbon monoxide emissions is 98% efficient. Tapping operations also generate fumes. Tapping is intermittent and is usually conducted during 10–20% of the furnace operating time. Some fumes originate from the carbon lining but most are a result of induced heat transfer from the molten metal or slag as it contacts the runners, ladles, casting beds, and ambient air. Some plants capture these emissions to varying degrees with a main canopy hood. Other plants employ separate tapping hoods ducted to either the furnace emission control device or a separate control device. After furnace tapping is completed, a reaction ladle may be used to adjust the metallurgy by chlorination, oxidation, gas mixing, and slag–metal reactions. Ladle reactions are an intermittent process, and emissions have not been quantified. Reaction ladle emissions are captured by the tapping emissions control system.

ACKNOWLEDGMENTS Sincere appreciation is extended to PhD student Mr Amit Bhalla for his help especially during the very busy and hectic periods of the academic year which seemed to be much more frequent during this exercise. Grateful acknowledgement and thanks is also extended to Professor Lauri Holappa for giving the oppurtunity to contribute to this book.

REFERENCES [1] L.A. Corathers, J. Gambogi, P.H. Kuck, J.F. Papp, D.E. Polyak, K.B. Shedd, U.S. Geological Survey Minerals Yearbook-2008, U.S. Department of the Interior, Washington, 2010, pp. 25.1–25.15. [2] F.D. Richardson, J.H.E. Jeffes, J. Iron Steel Inst. 160 (1984) 261–270, Ibid-150, vol. 166, pp. 123–235; ibid-1952, vol. 171, pp. 167–175. [3] H.-G. Lee, Materials Thermodynamics with Emphasis on Chemical Approach, World Scientific Publishing Co, Singapore, 2012, p. 221. [4] J.F. Elliot, M. Gleiser, V. Ramakrishna, Thermochemistry for Steel Making, vol. 12, Addison-Wesley Publishing CO, Reading, MA, 1963.

531

532

Rauf Hurman Eric

[5] M.W. Chase, Jr.,C.A. Davies, J.R. Downey, Jr.,D.J. Frurip, R.A. Mcdonald, and A.N. Syverud. JANAF Thermochemical Tables, third ed., ACC-NBS, The Dow Chemical Co., Midland, Michigan, 1985. [6] P.N. Chaudhary, R.P. Goel, G.G. Roy, Ironmaking Steelmaking (2001) 396–402. [7] H. Gaye, J. Welfringer, in: H.A. Fine, D.R. Gaskell (Eds.), Proceedings of the Second International Symposium on Metallurgical Slags and Fluxes, TMS-AIME, Warrendale, Pennsylvania, 1984, pp. 357–375. [8] R.P. Goel, H.H. Kellogg, Mathematical description of the thermochemical properties of iron-silicates slags containing lime, Ibid., pp. 347–355. [9] Y. Iguchi, P.J. Koros, G.R. St.Pierre, in: Proceedings of the Elliott Symposium on Chemical Process Metallurgy, ISS-AMIE, Warrendale, Pennsylvania, 1990, pp. 129–147. [10] E.T. Turkdogan, R.A. Hancock, Trans. Inst. Min. Metall. 67 (1957–1958) 573–600. [11] W.J. Rankin, Trans. Inst. Min. Metall. C 88 (1979) 167–174. [12] G.J.W. Kor, Met. Trans. 10B (1979) 367–374. [13] A. Tanaka, Tetsu-to-Hagane 66 (1980) 1474–1483. [14] W. Ding, S.E. Olsen, in: Proceedings of the Electric Furnace Conference, ISS-AIME, Warrendale, Pennsylvania, 1991, pp. 259–264. [15] M.V. Zode, private communication, MEL, December 1991. [16] T.B. Massalski et al., (Ed.), Binary Alloy Phase Diagrams, vol. I, American Society of Metals, Ohio, 1986, pp. 558–575. [17] G.W. Healy, in: R.A. Bergman (Ed.), Proceedings of the International Symposium on Ferrous and Nonferrous Alloy Processes, CIM, Hamilton, Ontario, August, Pergamon Press, New York, 1990, pp. 97–108. [18] M. Riss, Y. Khodorovsky, Production of Ferroalloys, Mir Publishers, Moscow, 1967, pp. 187–191. [19] R.P. Goel, Smelting technologies for ferrochromium product-recent trends, in: A.K. Vaish, S.D. Singh, N.G. Gaswami, P. Ramachandrarao (Eds.), Ferroalloy Industries in the liberalized Economy, NML, Jameshpur, 831007, 1997, pp. 37–50. [20] R.C. Urquhart, A study of the production of high carbon ferrochromium in the submerged arc furnace, Ph.D. thesis, University of Witwatersrand, Gauteng, South Africa, 1972. [21] K. Yamagishi, K. Endo, J. Saga, in: INFACON 1, Johannesburg, South Africa, 1974, pp. 143–178. [22] I. Ostrovski Ya, V.P. Vorob’ev, G.N. Kozhevnikov, Steel Transl. (7) (1976) 618–620. [23] I. Ostrovski Ya, Investigation of some particularties of technology of melting carbon ferrochromium, Ph.D. thesis, Institute for Metal Physics, Sverdlovsk, USSR, 1976. [24] A. Wedepohl, N.A. Barcza, Observation made during the dig out of a 48 MVA ferrochromium furnace, NIM Report No. 2009, Randburg, South Africa, 1981. [25] W. Pei, Study of some phenomena during the smelting reduction of chromite ore, Licentiate thesis, Royal Institute of Technology, Stockholm, 1991, ISSN 0821 8604. [26] K.L. Kossyrev, S.E. Olsen, H C FeCr smelting experiments, reduction mechanisms and slag properties, SINTEF Report STF24F947111, 1994. [27] E. Ringdalen, The high carbon ferrochromium process, reduction mechanisms, Ph.D. thesis, Norwegian University of Science and Technology, Trondheim, 1999. [28] P. Hayes, in: Proceedings of 10 th International Ferroalloy Congress: INFACON X; Transformation through Technology, Feburary 1–4, 2004, CapeTown, SouthAfrica, SAIMM, Johannesburg, 2004, pp. 1–14. [29] D.B. Wellbeloved, P.M. Craven, J.W. Waudby, Manganese and Manganese Alloys, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 1990, pp. 1–48. [30] A.N. Grundy, B. Hallstead, L.J. Gauckler, Assesment of the Mn-O system, J. Phase Equil. 24 (1) (2003) 21–38. [31] A.G.E. Robiette, Electric Smelting Processes, Charles Griffin & Company Limited, London, 1973, set and printed by Chorley & Pickersgill Limited, Leeds. [32] M.I. Gasik, M.M. Gasik, Electrothermal Silicon Technology, National Metallurgical Academy of Ukraine, Dnipropetrovsk, 2011, p. 487. [33] Background Report AP-42 Section 12.4 Ferroalloy Production, U.S. Environmental Protection Agency QAQPS/ISD/EIB, Research Triangle Park, NC, 27711, U.S.A., 1992.