Smelting Furnace of Iron Ore

Laboratory-scale experiments showed that pig irons and steels of acceptable grade can be made by arc-furnace methods from east Texas siliceous ores containing less than 25 percent iron. Under highly reducing conditions, 97 to 99 percent of the iron in the ore was recovered as pig iron. Small quantities of limestone, the minimum limited only by progressively higher slag viscosities, were used as flux. It was shown that satisfactory recoveries of high-silicon pig iron can be attained from very siliceous charges.

The technique for removing impurities from pig iron on a laboratory scale did not reach a high state of perfection, but refining results indicated that satisfactory steel can be consistently produced in an arc furnace from low-grade ores. Dolomitic limestone did not remove sulfur as effectively from pig iron as high-quality limestone.

Coke, bituminous coal, and lignite were found to be about equal on a fixed-carbon basis for reducing iron oxides. Laboratory-scale arc-furnace smelting experiments showed that power consumption was nearly inversely proportional to the grade of ore. The ability of the arc furnace to use low-grade iron ores and low-rank reductants effectively in coarse or fine sizes is noteworthy.

Furnace operation on low-grade ores was smoothest with a lignite reducing agent. Volatile matter in coal and lignite apparently affected overall reduction to only a small extent.

Estimated reserves of all classes of iron ore within the United States are over 76 billion tons, exclusive of the direct shipping ores of the Lake Superior region. A high percentage of these 76 billion tons require beneficiation by one or more of several well-known methods to produce a feed suitable for blast-furnace smelting. Beneficiation efficiency varies from poor to excellent, depending on the concentrating characteristics of the ore. Pyrometallurgical separation of iron from ore gangue results in high recovery of metal as pig iron. Pig irons from most types of iron ore are amenable to refining by standard methods.

The direct smelting of low-grade iron ores, in contrast with smelting high-grade ores, involves such considerations as relatively low production per furnace because of high slag volume, increased cost of power and furnace operation per unit of production necessitated by melting larger quantities of gangue and flux, and higher transportation and mining costs for each unit of iron produced, because more fuel, ore, and flux are required. Ore-dressing low-grade ores to obtain high-grade concentrate, rather than direct-smelting low-grade ores, involves less capital outlay and lower operational costs. However, recovery of iron by a concentrator is rarely as effective as by a furnace, and an additional loss of iron will be taken when concentrates are smelted. Electric furnaces can smelt soft or fine ores and utilize low-grade fuels, but the cost of operating power deters their widespread use. Proper evaluation of the relative merits of treating low-grade iron ores by beneficiation or by electric smelting requires the availability of a large amount of technical data. The purpose of this manuscript is to add facts to existing information.

Large deposits of iron ore occur in eastern Texas; reserves of measured and indicated ore, in terms of the washed product, have been estimated to be 160,000,000 tons. Shallow overburden, a layer of oxide ore underlain by carbonate ore, and a high silica content are characteristic features of many ore bodies. The ore was recognized in 1839, and the deposits were first worked during the Civil War. Production was intermittent until 1944; since that time yearly output has been considerable. The production recorded was 3,960,000 long tons in 1953, 2,240,000 long tons in 1954, and 3,110,000 long tons in 1955; all this material required beneficiation. The concentrates were smelted in Texas.

Present operations are carried on by the Lone Star Steel Co. and the Sheffield Steel Corp. Their beneficiation plants recover about 50 percent of the iron in the ores as concentrates, containing approximately 45 percent iron.

Ores from mines of the Lone Star Steel Co. in Morris County were selected as raw material for use in the metallurgical investigations. These represent some of the lowest grade iron ores employed commercially in steel production in the United States. Some of the developed ores of this company contain less than 25 percent iron; the oxide ores may be low in sulfur, but appreciable quantities of sulfur and phosphorus must be removed during the smelting of most such ores. Carbonate ores generally are relatively high in sulfur.

The samples of ore were obtained from the Rogers and Black Mountain mines and the concentrate, from the Lone Star Steel Co. mill. Later, drill cores obtained during exploration were composited and tested metallurgically.

The various reductants used in smelting experiments were graphite, metallurgical coke, lignite, Oklahoma coking coals, and Illinois domestic heating coals. Fluxes were obtained from aggregate and agricultural dolomitic-limestone quarries near Rolla, Mo., and from high-purity limestone in eastern Missouri.

A single-phase 100-kv-a. Lectromelt arc furnace of about 200-pound charge capacity was used to smelt the ores and concentrates. Both an indirect-arc furnace and an induction furnace were used to refine pig iron. The indirect-arc furnace easily accommodated 40 to 60 pounds of pig iron; the induction furnace was used for crucible charges of about 1 pound. All furnacing was in batches.

Chemical Analyses

The partial chemical analyses of ores and mill concentrate (table 1) show that silica, alumina, phosphorus, and sometimes sulfur must be removed to make steel. The imbalance of slag-forming materials is also notable. The sulfur content of the concentrate appears higher than that of the ores, because the ore faces were unusually low in sulfur when the mines were sampled. Ores from this area normally contain quantities of sulfur similar to those in the drill-core sample.

Partial analyses of various reductants and limestone used in the experiments are presented in table 2.

Technical Studies

Fluxes and Reductants

To ascertain the smelting characteristics of the ores, fluxes, and reductants a number of small-scale tests were made in crucibles. Much of the iron oxide was reduced to metal shot in 1 hour at 1,400° C., but little metal collected on the crucible bottom; at 1,450° less than half of the iron was recovered in a metal button, and, at 1,500°, from 50 to 60 percent of the molten iron was recovered in 1 hour. Substantially all metallic iron was collected in 1-½ hours at 1,525° to 1,575° C.

Experimental gangue fluxing agents included fluorspar, soda ash, lime, and limestone. Adding restricted quantities of flux produced viscous, although liquid, slags. The fluidity of slags was increased slightly by adding fluorspar or soda ash up to 3 percent of the charge weight; but, after preliminary experiments, only limestone was used as flux. Limestones were obtained from local quarries producing agricultural limestone and from a high-purity eastern Missouri deposit. The local limestones were highly dolomitic; their magnesia contents were considered equivalent to lime in charge-balance calculations. Small charges of ore were smelted in a series of experiments in which the limestone-silica ratio was gradually decreased. The data from these experiments demonstrated that 1 unit of limestone to 2 units of silica was the minimum limestone addition that produced a workable slag.

To simplify making up charges, the ratio of limestone in the flux to silica in the ore was used as the flux-control variable. Table 3 gives the comparable basicity factors, (CaO + MgO/SiO2 + Al2O3) for each limestone-silica ratio. The ash content of reductants was not considered.

During the smelting of 0.5 limestone-silica factor charges, viscous slags formed; however, the iron content of these slags was low. Pig smelted from these low-limestone charges was relatively higher in silicon and sulfur than pig iron from higher limestone charges, because heavier additions of lime slagged more silica and shifted the calcium-sulfur equilibrium.

Variation in the character of slags and pig irons produced from smelting was more pronounced in limestone-silica-factor charges between 0.5 and 0.75 than between 0.75 and 1.0. When a constant quantity of reductant was added to 0.75- and 1.0-factor charges, the slag from the 0.75-factor charge contained the most iron. By adjustment of the quantity of reductant, the silicon content of pig iron from each type of charge could be made equal. Thus, the silicon content of pig iron depended primarily on the quantity of charge reductant and secondarily on the limestone-silica factor.

Using extra reductant produced low-iron slags and high-silicon pig iron from charges with a 0.5 limestone-silicon factor. Sulfur was removed from iron satisfactorily with any of the three quantities of limestone used.

Reductants were charged on the basis of a selected number of pounds per pound of iron in the feed. A few charges were made up with fine cuttings of graphite as the reductant, but the use of graphite was not investigated enough to permit making comprehensive conclusions. There was no sulfur or phosphorus pickup by the pig iron, however, when graphite was employed as the reducing agent.

Coal samples from four Oklahoma mines of the Lone Star Steel Co. were used in many of these experiments. Metallurgical coke was employed in a few experiments, but no advantages from its use were apparent in the arc furnace.

Two samples of lump bituminous coal, obtained at different times from the same source in southern Illinois, were also used as a reductant. Table 2 shows that the sulfur content of these samples varied greatly. This coal also caused an unusual amount of gassing in the furnace; the gassing was particularly pronounced while the charge was partly molten. The semiwild furnace charge was attributed to large quantities of volatile matter escaping from the coal while the charge was semifused. In addition, volatilization of chemically combined water in the ore and carbon dioxide from the limestone added to the volume of gas.

Furnace operations were smoothest when Texas lignite served as the reductant. It is believed that the large volatile content of this material effectively escaped from the charge before fusion took place, thereby minimizing spewing of the furnace contents. It was also noted that arc control was most satisfactory with reduction by lignite.

Arc Furnace Smelting

A 100-kv.-a. Lectromelt, laboratory-type, size V, single-phase, arc furnace was used for most smelting experiments. In a few preliminary charges smelted in a Detroit, indirect-arc, rocking-type furnace dusting was excessive, and balling of the charges took place at the softening temperature.

The east Texas iron ores contain large quantities of gangue minerals; therefore, it was economically imperative to add the least possible flux to the furnace burden. As the gangue consisted chiefly of silica and alumina, a highly acid slag resulted when limestone additions were held to a minimum. Because the furnace was lined with magnesia brick, the first experiments were made with this lining, although these ore charges were expected to corrode magnesia. This lining was consumed in a few heats. A carbon lining that gave satisfactory service for many heats was subsequently installed.

Rogers-Mine Ore and Mill Concentrate

A charge consisting of 36 percent Rogers mine ore, 36 percent mill concentrate, 16 percent lime, 10 percent graphite, and 2 percent fluorspar was smelted in an indirect-arc rocking furnace to obtain iron contents of the slag at intervals as smelting progressed. This charge contained lime equivalent to a 0.5-limestone-silica factor charge. The charge was fused, and the surface portion sampled over a 2 hour period. The results are presented in table 4.

Charges of 50-50 Rogers ore and mill concentrate were prepared with approximately 0.5, 0.75, and 1.0 limestone-silica factors, plus coal in slightly varying portions. Ore and coal were crushed through ½-inch and limestone through 10-mesh. These charges were smelted in the carbon-lined 100-kv.-a. arc furnace, The time in the fused state was 2-½ hours. Table 5 gives data on several pig irons and slags produced from these smelting experiments.

Product analysis considerably varied from one test to another, but much general information and several trends are indicated from the results. As the percentage of reductant in the charge was decreased, the iron content of the pig rose, and silicon content decreased preferentially to carbon. The apparent trend of phosphorus in the metal was to rise somewhat with an increase of coal in the charges. Sulfur reduction was improved by adding more coal to augment reducing conditions. As limestone or reductant increased in the charges, the iron content of the slags decreased. The sulfur content of the slag rose as more lime was added to the charge.

Rogers-Mine Ore

A series of arc-furnacing tests was made on Rogers-mine ore by using a nearly constant iron-coal ratio and varying the limestone-silica factor from 0.5 to 1. The results of these experiments are presented in table 6. A column is included in this table for “basicity factor” or “V” ratio, terms used in the steel industry to indicate slag conditions

In contrast with erratic removal of sulfur from mixtures of Rogers-mine ore and mill concentrate, the data in table 6 indicate consistent removal of sulfur during smelting of straight Rogers-mine ore. Variation in limestone quantities was not reflected in the sulfur content of the pig irons, although the iron content of the pig iron rose as larger quantities of silica were slagged by increases of limestone. Carbon in the pig also increased as more silica was retained in the slag by the higher limestone additions. The slag removed only small quantities of phosphorus. The iron content of the slag increased as lime was charged in larger quantities but decreased when mixed ore and concentrate were smelted. These experiments show that pig irons of acceptable analyses can be produced by smelting low-grade siliceous iron ores with a small quantity of flux in an arc furnace.

Several smelting experiments were conducted similar to those of which the results are shown in table 6, except that duplexing slag treatments were made for removing phosphorus. At the end of 1-¾ hours of smelting the slag was skimmed and largely replaced with an oxidizing basic cover made up of limestone and either roll scale or ferric oxide. After the cover became fluid, it was skimmed into a slag mold, and the metal layer was slowly poured through it. Some reduction in silicon content of the pig was attained by this simulated Perrin treatment, carbon content was slightly reduced, and phosphorus content was scarcely affected. Too little oxidation was attained under these conditions to refine the pig iron.

Black Mountain-Mine Ore

By using data gained from smelting Rogers mine ore, Black Mountain ore was smelted without preliminary runs. In a series of smelting experiments, ore was blended with varying portions of limestone, while the iron to coal ratio was maintained nearly constant. Coal made up either 7 or 8 percent of the total charge weight, while limestone in the charges varied from 18 to 30 percent. Surface temperatures of the molten charges averaged between 1,570° and 1,590° C. (2,855° to 2,890° F.) . The molten charges were in the arc furnace 2-¾ hours. A mixture of four Oklahoma coking coals crushed through ½ inch was used as reductant. Table 7 shows smelting results from varying lime in the charges.

Simulated Perrin treatment was tried but was not effective enough to produce steel under laboratory conditions.

The data in table 7 indicate an overly reduced condition during smelting. Although the iron-coal ratio approximated that used in the experiments on the Rogers ore, larger quantities of silicon were reported in the pig irons produced from Black Mountain ore. Smelting concentrated the sulfur in the slag layer and the phosphorus in the metallic layer. Slag from the low-lime charge was highly viscous and difficult to handle; however, very little iron shot remained in the slag, indicating that fluidity was adequate for metal-slag separation.

Electric-Furnace Refining of Pig Iron

High-silicon pig irons smelted from Rogers and Black Mountain ores were refined in indirect-arc and induction furnaces. The objectives were to remove major portions of the silicon, carbon, and phosphorus through oxidation and slagging or, in the case of carbon, oxidation and volatilization. Refining involved maintaining oxidizing conditions over the molten iron by using oxidizing slags or lancing directly with air, oxygen, or air-oxygen mixtures.

In preliminary experiments, impurities were removed in small crucibles heated by induction. Oxidizing slags were placed over the molten samples of pig iron, and oxygen lancing was also employed to increase the oxidation of impurities. These experiments in crucibles did not produce steel but indicated the conditions necessary to achieve the desired degree of purification.

An indirect-arc rocking-type furnace, operated with both an alumina and a magnesia lining, was used for larger scale experiments to purify pig irons. Removal of impurities was more effective with the latter lining.

Rogers-Mine Ore Plus Mill Concentrate

A composite of pig irons from two smelting heats analyzed, in percent: Fe 94.3, Si 1.4, C 4.1, and S 0.10. This material was melted in a carbon-lined arc furnace with a closed top. When molten, a reducing flux cover, weighing 50 percent of the metal and containing 61 percent limestone, 20 percent graphite, 10 percent soda ash, and 9 percent silica, was placed on the metallic bath. The charge was held between 1,500° and 1,600° C. for 2-½ hours to remove sulfur. At the end of the period, the slag was skimmed and the metal sampled. The sulfur was effectively removed, as shown by analysis of the treated metal: Fe 93.6, Si 1.9, C 3.7, P 0.21, and S 0.01. The molten pig iron was transferred to a hot magnesia-lined indirect-arc furnace where it was covered with an oxidizing flux weighing about 18 percent of the metal. The flux material contained 69 percent limestone, 28 percent Rogers-mine ore plus mill concentrate, and 3 percent soda ash. The resulting slag was skimmed at the end of 1-¼ hours. A second fluxing treatment was made in the same manner. Finished metal analyzed, in percent, as follows: Fe 97.4, Si 0.09, C 2.3, P 0.20, and S 0.01. Silicon and sulfur were virtually removed by the treatments, but only half of the carbon and very little phosphorus were affected. Conditions for oxidation in the indirect-arc furnace did not permit making steel from this pig iron without excessive treatment time and large volumes of flux.

Oxidizing conditions were intensified by using oxygen and additional air. A charge of pig iron containing, in percent, Fe 89.3, Si 5.1, C 3.8, P 0.19, and S 0.01 was melted in a zirconium silicate crucible by induction heating. Limestone, to the extent of 20 percent of the charge weight, was placed on the molten metal at 1,350° C. The bath was lanced with 50-50 air-oxygen mixture for 5-½ minutes when the slag became too pasty to permit further blowing. This treatment lowered the carbon and silicon content of the metal to 3.1 and 4.4 percent, respectively.

Pig iron of the above analysis was melted in the magnesia-lined, indirect-arc, rocking furnace under a flux weighing 30 percent of the charged metal. The flux was made up of 62 percent limestone, 35 percent Rogers-mine ore plus mill concentrate, and 3 percent soda ash. One-half of the flux was added and the charge held at 1,500° C. for 1-¼ hours; then the slag was skimmed, the remaining flux added, and the charge held molten an additional 1-¼ hours. Surface temperature of the melt during the latter period averaged 1,540° C. Throughout the entire refining period a jet of air was blown into the furnace over the molten charge. This procedure increased oxidation of the charge impurities to such an extent that the analysis of the poured metal was: Fe 99.3, Si 0.05, C 0.29, P 0.004, and S 0.01.

In general, oxygen lancing in a crucible under laboratory conditions did not provide proper oxidation of the impurities. Most of these treatments were terminated in less than 10 minutes. Holding molten pig iron in an indirect-arc furnace under an oxidizing flux and constantly forcing air into the furnace converted the pig iron to mild steel.

Black Mountain-Mine Ore

Pig iron was produced by arc-furnace smelting a charge consisting of 69 percent Black Mountain mine ore, 23 percent limestone, and 8 percent Illinois coal No. 2. The pig iron contained, in percent: Fe 91.9, C 3.6, Si 3.0, P 0.28, and S 0.012. It was charged to the 50-kw indirect-arc rocking furnace and then covered by 12 ½ percent of its weight of an oxidizing flux consisting of 70 parts of limestone and 30 parts of ferric oxide. A 50-50 mixture of air and oxygen was injected into the furnace above the charge during the entire treatment. The first refining period was 45 minutes at an average surface temperature of 1,566° C. At the end of this time the slag was skimmed and the bare bath held molten under the air-oxygen jet 15 minutes before another portion of oxidizing flux was added. The second flux remained on the metal for 45 minutes at an average surface temperature of 1,570° C. The charge was poured at 1 hour and 45 minutes. The finished metal contained, in percent: C 0.076, Si 0.037, P 0.063, and S 0.024. The indicated removal of impurities, in percent was: C 98, Si 99, and P 86.

Figure 1 shows the furnace operator charging cold pig iron to the indirect-arc furnace. Figure 2, photographed with the furnace in operation, indicates the mode of air injection over the molten charge. An air hose attached to a ceramic tube protruding from the furnace door permitted introduction of air or oxygen while the furnace was in motion.

Power Utilization

Power-consumption data are essential to evaluate the feasibility of electrically smelting low-grade iron ores. Data for this purpose were derived through smelting charges of low-grade iron ore and a high-grade iron concentrate in the laboratory. Such data must be considered in relation to laboratory or small-scale work, which inherently requires more heat per unit of production. Thus, the power required to produce 1 pound of pig iron from iron ore on a laboratory scale may be several times that required in a large industrial furnace. The relationship between the power required to produce a unit of iron from low-grade ore and that necessary to produce a unit of iron from high-grade ore is constant, regardless of the scale of operations; therefore, the following figures on power for low-grade ore may be readily projected for large-scale production, when the power consumption of the large furnace is known for an ore of a definite iron content. Some adjustment may be necessary for flux requirements of various ores; however, this adjustment will be small if a constant basicity factor is adhered to.

Low-Grade Ore

Selected drill cores from an east Texas iron-ore-exploration project were used as low-grade ore in power-consumption studies. A composite of this material contained, in percent: Fe 28.6, SiO2 32.0, Al2O3 10.7, CaO 0.14, P .087, and S 24. Arc-furnace smelting experiments were made on this ore, with variations in the reductant and basicity factor. Residence of charge in the furnace was held constant at 2½ hours. This time was near minimum for reduction of the iron in both the low- and high-grade ores. Reductants used were coke, coal, and lignite. Basicity factors of the charges were 0.30, 0.35, and 0.40. The fixed-carbon (in the reductant)-iron ratio in the charges was either 0.45 or 0.60. Within limits of the experimental variations, the quantity or type of reductant did not affect power consumption enough to delineate a trend. There was no definite advantage for coke, coal, or lignite but furnace operation with lignite was less variable. In contrast with the spewing of “wild” charges when coal was used, charges containing lignitic reductant smelted quickly and with a steady arc. This difference is attributed to the release of volatile matter from lignite before incipient fusion of the charge.

As slag has a higher specific heat than pig iron, the large slag volume of the high-basicity charges were expected to require more power. Experiments did not bear this out, as more kilowatt-hours per pound of pig iron were required to smelt 0.30-basicity-factor charges than for 0.40-basicity-factor charges. Further analysis of smelting conditions showed that, as the basicity factor of the charges was lowered, the slags became increasingly viscous and required higher finishing and pouring temperatures. The added wattage used to raise the charge temperature 15° or 20° C. apparently was greater than that necessary to heat and flux the additional limestone in the charges having higher basicity factors.

High-Grade Concentrates

A truckload of specular hematite gravity concentrate was obtained from the Ozark Ore Co., Iron Mountain, Mo. This particular lot of ore contained, in percent: Fe 55.3, SiO2 9.3, Al2O3 3.5, CaO 4.8, MgO 0.60, P 0.007, and S 0.009. Portions of this material were arc-furnace-smelted in experiments that paralleled those on low-grade ore. Conditions were maintained to produce pig irons of comparable iron content from both the low- and high-grade feeds.

Melts of concentrate were quiet in the furnace with coke, coal, or lignite reductants. Because of low slag volume and high metal fall, total power consumption for smelting a charge of concentrate was expected to be less than for smelting the same weight of charge made up of low-grade ore. Smelting data, however, did not confirm this assumption. The insulating effect of slag is believed to have reduced the heat loss from low-grade charges; total power consumption per charge slightly favored the low-grade iron ore.

Consumption of electric power per pound of pig iron is shown for corresponding basicity factor charges in table 8. A comparison of the energy consumed in making 1 pound of pig iron from the 2 grades of ore shows that the ratio of power consumed to pig iron produced is nearly inversely proportional to the iron tenor of the ores. This relationship was extended and confirmed by laboratory smelting of 22-percent iron ore.

Furnace Capacity

To arrive at relative furnace capacity for smelting low- and high-grade ores, a series of furnacing experiments was made, in which choke feeding was practiced. Low- and high-grade ores were blended with fluxes and lignite to produce balanced charges. The usual quantity of charge material was added to the furnace and an arc struck. After a pool of molten material had formed, more cold charge was added to the furnace in increments until molten slag began running from the charging hole. Weights of charges at these points were noted and compared for low- and high-grade ores. The relative furnace capacity was 89 parts by weight of low-grade charge to 100 parts of high-grade charge. Relative parts of ore were 98 and 100, respectively, representing 51.5 and 100 parts of iron. Thus the capacity of the furnace to produce pig iron was in direct proportion to the iron tenor of the ore, except for a small discrepancy due to a difference in iron lost in the slag.