News

Theoretical Aspects of Direct Reduction Process


Theoretical Aspects of Direct Reduction Process

In the direct reduction process of iron ore, the solid metallic iron (Fe) is obtained directly from solid iron ore without subjecting the ore or the metal to fusion. Direct reduction can be defined as reduction in the solid state at oxygen (O2) potentials which allow reduction of iron oxides, but not of other oxides (MnO, SiO2 etc.), to the corresponding elements. Since reduction is in the solid state, there is very little chance of these elements dissolving (at low thermodynamic activity) in the reduced iron, so the oxides which are more stable than iron remain essentially unreduced. Direct reduction of iron ore also takes place in the shaft of the blast furnace by the ascending gases.

The iron – oxygen system

The iron-oxygen (Fe-O) system is possibly one of the most extensively studied systems. The thermodynamics of the system is well understood and a lot of information is available about the kinetics of gaseous reduction involving iron oxides. The thermodynamically stable solid phases which occur between 400 deg C and 1400 deg C in the Fe-O system, at a total pressure of I kg/sq cm, are shown in the binary diagram (Fig 1). This diagram shows that Fe forms with O2 the three stable, solid compounds namely (i) hematite – Fe2O3, (ii) magnetite – Fe3O4, and wustite – FexO, where x is a little lower than 1. The non-stoichiometric FeO phase (wustite) is unstable below 570 deg C and decomposes into a mixture of metallic Fe and Fe3O4. Thus, reading from right to left across the phase diagram at constant temperature, below 570 deg C the phase sequence is Fe2O3 – Fe3O4 – Fe, whereas above 570 deg C the sequence is Fe2O3 – Fe3O4 – FeO – Fe.

Fig 1 Fe-O binary system diagram

The negligible solubility of O2 in solid alpha and gamma iron is less than 0.01 % of O2. Hence, the O2 content has no effect on the transition temperatures of the solid Fe modifications and is disregarded in the diagram.

Considering the reaction equilibrium, the reduction of Fe oxides involves one or more of these steps (i) hematite (Fe2O3) -> magnetite (Fe3O4), (ii) magnetite (Fe3O4) -> iron (Fe), (iii) magnetite (Fe3O4) -> wustite (FeO), and (iv) wustite (FeO) -> iron (Fe).

Wustite is stable only at a temperature which is higher than 570 deg C. The thermodynamic equilibria for the above reactions are well-known for the two major gaseous reducing agents used namely hydrogen (H2) and carbon monoxide (CO).



Iron – oxygen – carbon system

The equilibria of Fe and Fe oxides with a mixture of the CO and CO2 (carbon di-oxide) gases with solid carbon (C) are shown in Fig 2.

Fig 2 Fe-CO-O system diagram

From the Fig 2 it can be inferred that, at temperatures above 710 deg C and at a total pressure of 1 kg/sq cm all the Fe oxides can be reduced by CO/CO2 gas mixtures which are in equilibrium with C, and can, hence, be reduced by the C itself. At lower temperatures only those mixtures which are super-saturated with C and which, hence, according to the Boudouard equilibrium, react towards C deposition have a reducing action on wustite.

Iron – hydrogen – oxygen system

The equilibrium diagram for Fe and Fe oxides with a mixture of the gases H2 and H2O (steam) is shown in (Fig 3).

Fig 3 Fe-H-O system diagram

The major difference between this system and the Fe-O-C system is the absence of a ‘sooting line’ or corresponding phenomena. Thus it is theoretically possible to reduce hematite (and magnetite) to Fe with H2 at any temperature.

Comparison between reduction by CO and H2

From the studies of Fe-C-O, and Fe-H-O systems (Fig 4), it seems that above 815 deg C, H2 is a more efficient reducing agent than CO (i.e., the equilibrium H2/H2O ratios are lower than the corresponding CO/CO2 ratios), while it is opposite at lower temperatures. However, these equilibria are hardly achieved in industrial furnaces, since the rate of reduction becomes very slow as equilibrium is approached. When conditions deviate distinctly from equilibrium the respective reaction rates for reduction with H2 and CO are in the reverse order to those which is normally be expected from the equilibrium consideration. Thus, H2 is actually a more efficient reducing agent for a non-equilibrium process which is designed to operate at temperatures lower than 815 deg C and CO is more efficient at higher temperatures.

Fig 4 Equilibrium relations in the Fe-C-O, Fe-H-O, and C-O systems

Studies on the effect of gas composition mixture of CO and H2 at different temperatures have shown that as the H2 content of the reducing gas mixture increases the rate of reaction increases. This relationship has been found to be markedly nonlinear.

The reduction of Fe oxides to metallic Fe with H2 is endothermic and an external source of heat is necessary to maintain the required temperature. The corresponding reaction with CO is exothermic and under suitably controlled conditions, the reaction is thermally self-sustaining. In fact, it may be required to dilute the CO with H2 or other heat-absorbing gas to avoid overheating the charge. Some processes have been designed to take advantage of CO – H2 thermal balance and use mixtures of these gases for increasing the amount of reduction obtained during the heating of the ore, from the ambient temperature upto the maximum reaction temperature of around 1100 deg C.

Fig 4 shows that for all temperatures within the range where gaseous reduction is economically possible, the equilibrium gas mixtures contain at least 60 % of CO and/or H2. When equilibrium is not attained the unreacted concentrations of these gases are even higher and the great part passes unchanged through the reducing furnace. If the process is to be economical then it is necessary there is the utilization of the gas which remains after the reduction of wustite to metallic Fe, for reduction of the higher Fe oxides to wustite and/or for the regeneration of the gas mixture and the removal of the gaseous reaction products.

Gas-solid reaction and solid-solid reaction

Gas-solid reactions play a major role in technology, and encompass a very broad field including the extraction of metals from their ores (Fe oxide reduction, etc.). A common feature of all gas-solid reaction system is that the overall process can involve several intermediate steps. Typically, these intermediate steps involve (i) gaseous diffusion (mass transfer) of reactants and products from the bulk of the gas phase to the external surface of the reacting solid particle, (ii) diffusion of gaseous reactants or gaseous products through the pores of a solid reaction product or through the pores of a partially reacted solid, (iii) adsorption of the gaseous reactants on and desorption of reaction products from the solid surfaces, and (iv) the actual chemical reaction between the adsorbed gas and the solid.

In the area of the gas-solid reactions, there are several other phenomena which can affect the progress of reaction and the performance of furnace in which gas-solid reactions are carried out. These other phenomena include heat transfer, changes in the solid structure (such as sintering) which accompany the reaction, and the flow of gases and solids through the furnace in which gas-solid reactions are taking place. The rate of reduction is controlled by these factors depending on the process used.

The reactions between solids can be divided into two main groups namely (i) true solid – solid reactions which take place in the solid state between two particles in contact with each other, or through the migration of particles in the solid state, (e.g. formation of iron carbides through the reaction between Fe oxides and C), and (ii) reactions between solid reactants, which take place through gaseous intermediates (e.g. reduction of Fe oxides with carbon at 1 kg/sq cm pressure).

The reduction of Fe oxides with solid C can also be a true solid-solid reaction, provided it is carried out at very low absolute pressures. During one of the studies carried out by the reaction of mixtures of finely powdered graphite (C) and hematite ore under a vacuum of 0.0005 mm of Hg (mercury) it was found that at temperatures of upto 900 deg C, the reaction took place very slowly, and in 18 hours only Fe3O4 and FeO was formed, but no Fe. During the test an appreciable transformation was observed only at higher temperatures. It was concluded during the study that the rate of reaction is determined by the diffusion of the Fe ions within the oxide phase. A deduction made in another investigation was that the C diffuses in the Fe oxide, is perhaps only of historical interest. However, these studies have shown that the rate of reaction distinctly increases when the gas pressure above the powdered mixture increases. In tests of similar type, in which a stream of N2 (nitrogen) was passed through the mixture of C and Fe oxide, a marked decrease in the reaction rate was observed as the flow of N2 was increased. All these investigations, whether carried out in vacuum or under N2, with the rate of reduction of similar powdered Fe oxides in CO or H2, proved that the direct, solid-state reaction of C and ore (which is sometimes regarded as the actual mechanism of the true direct reduction) is of no importance for the progress of reduction process in an industrial furnace.

Pore structure of the reduced iron

The reducibility tests on several natural ores have shown that the porosity of iron ore particles is one of the most important factors controlling reducibility. The reducibility expressed as the reciprocal of the time required for 90 % reduction, varied directly with the porosity. The relative reducibility increases with an increase in porosity as given by the equation “Relative reducibility = (porosity x 0.75) + 8.0”.

The reduction of Fe oxides always yields a porous reaction product. The nature of the oxide and the reduction conditions affect the structure of the pores in reduced iron. This is since the reduction proceeds from the surface of a particle towards inwards. The volume occupying space defined by the original wustite surface is diminished. This can only be accomplished by developing porosity. The scanning electron micrograph study of this porosity has shown that in general, H2 reduction gives a finer pore structure than that obtained by CO reduction. Also, from the scanning electron micrographs, it becomes apparent that the pore structure becomes coarser as the temperature of reduction in H2 is increased progressively from 600 deg C to 1200 deg C.

The initial pore surface area of the Fe oxide affects the pore surface area of the reduced iron formed by gaseous reduction. A decrease in the initial pore surface area of the Fe oxide decreases the pore surface area of the reduced iron. The pore surface areas of iron reduced from hematite ore in H2 measured by BET (Brunauer-Emmett-Teller) technique has indicated that it decreases with increasing reduction temperature.

A relationship between the reduction temperature and average critical pore size and the smallest pore radius has been obtained from the pore size distribution. The pore size has been found to increase slowly with reduction temperature upto 900 deg C, but has increased rapidly with further increases in temperature. These results are in accordance with the observation of fracture surfaces by scanning electron microscopy, which showed the distinct coarsening of the pore structure at reduction temperatures above 900 deg C.

The pore surface area of the reduced hematite ore is also affected by reduction temperature and gas composition. The pore surface area obtained from hematite reduced in CO/CO2 gas mixture is about two-third of that for reduction in H2/H2O gas mixtures. This is consistent with the coarser pore structure of CO/CO2 reduced iron observed under the microscope.

The gas diffusion in the pores of reduced iron has been measured. The diffusive flux in porous media occurs via two diffusion processes namely (i) Knudsen diffusion, independent of pressure and proportional to T (temperature) to the power 1/2, and (ii) molecular diffusion, inversely proportional to pressure and proportional to T the power 3/2.

The limiting ideal structure is assumed to have pores of uniform size which are all inter-connected and intersect each other with an angle of 45 degrees.

The effective diffusivity varies for a given porous medium with temperature and pressure and is different for different binary gas pairs. The pore structure becomes finer with decreasing reduction temperature.

Modes of reduction

The reduction of natural iron ore particles or sintered hematite pellets results in the formation of product layers. This well-known phenomenon was a subject of many studies. In one of the recent studies of the reduction of sintered hematite pellets by H2, it has been noticed that there is a typical example of layer formation in polished section of a partially reduced hematite pellet. Relatively smooth interfaces between the layers usually appear at low magnifications, though such an appearance can be misleading.

This indicates that gas diffusion is sufficient in the wustite layer to give some internal reduction ahead of the advancing Fe/FeO interface. The zone of internal reduction is extended as (i) the temperature is lowered, (ii) the porosity is raised, and (iii) the particle size becomes smaller.

The effect of particle size on the time needed to achieve a given percentage of reduction depends on the mode of reduction and hence on the type of rate-controlling process. Consideration of the modes of reduction of porous Fe oxides by gaseous reduction have shown three limiting rate controlling processes namely (i) uniform internal reduction, (ii) limiting mixed control, and (iii) diffusion in the porous Fe layer. If the reduction is controlled solely by any one of these, then the time of reduction is related to the particle (spheroidal) diameter in one of the three ways namely (i) uniform internal reduction that is time is independent of diameter, (ii) limiting mixed control, and (iii) diffusion in porous iron.

The rate controlling process becomes relatively simple only when (i) there is uniform internal reduction, hence a small particle size is needed, or (ii) the ultimate rate control by gas diffusion in the pores of the iron layer predominates, since the particle size is large. It is also to be realized that there can be a transition from one limiting rate-controlling process to another as the reduction progresses, depending on temperature, gas composition, particle size and type of oxide. The reduction of iron oxides can also show some unexplained and unusual behaviour.

Rate of reduction of porous Fe ore particles

The porosity and pore structure of the ore has a marked effect on the extent and uniformity of internal reduction. In one of the studies, the effect of particle size on the rate of reduction of hematite ore for a mixture of 90 % CO and 10 % C02, and for H2 at 1000 deg C has shown that with increasing particle size the internal reduction is confined to the outer regions of the particles, hence there is decrease in the overall rate of reduction with increasing particle size.

In the early stages of reduction of porous hematite particles, there is rapid conversion to FeO followed by internal reduction of FeO to Fe. In the limiting case of almost perfect gas diffusion in the pores of the Fe oxide, the internal reduction predominates and the rate is controlled primarily by a gas-solid reaction on the pore walls. A Fe layer, a few atoms thick is assumed to cover the pore walls of FeO. The reduction rate is presumed to be controlled jointly by rapid diffusion of O2 through the coating of the Fe layer on the pore walls and by the chemical reaction of H2 or CO with the O2 on the surface of this very thin Fe layer.

The effect of particle size shows that the rate of reduction increases with decreasing size of the particles. The typical micrograph indicates that the mode of reduction varies from one grain to another within the particle. This is because of the local differences in the porosity of the oxide grains. Because of variations in pore size and faster gas diffusion in larger pores, most of the reaction occurs on the walls of larger pores. That is, only a fraction of the total pore surface area is expected to be used for reaction. The rate of H2 reduction at 800 deg C achieved with various types of hematite ore particles increases non-linearly with increasing pore surface area of the Fe (or FeO) formed. These results substantiate the fact that the larger is the pore surface area, the smaller is the fraction of the total pore wall used in the reaction.

The rate of internal reduction in H2 – CO gas mixtures is usually the sum of the two individual rates of reduction with H2 and CO. Both the reduction data and the C deposition observed indicate that, below 1000 deg C, gas reactions leading to water-gas  equilibrium are slow.

Rate of reduction of iron ore (lump or pellet)

The rate of reduction of lump ore or ore pellets is of a complex nature in the stream of reducing gas in a packed bed, The complexity is because the overall rate of the reduction is controlled by several reaction processes in series, such as heat and mass transfer through the gas film boundary layer, gas-solid reactions and gas diffusion in porous product layers. Through mathematical analyses, facilitated by computer calculations, numerous equations have been derived to describe the rate of reduction of large oxide particles for various modes of reduction.

In several experiments with single pellets or iron ore particles, heat transfer is relatively fast and, with sufficiently high velocity gas streams, the gas film mass-transfer resistance is small enough to be neglected. Hence, there are primarily two major reaction steps in series which influence the rate of reduction namely (i) gas-oxide reactions, and (ii) gas diffusion in porous oxide and porous product layers. The relative effects of these rate processes depend on the particle size, gas composition, temperature and mode of reduction, and they change with the progress of reduction.

Gas diffusion in the porous Fe layer

In one of the studies, unidirectional reduction experiments have been carried out to demonstrate the effect of gas diffusion in the pores of Fe layers. Long cylinder samples were prepared from big pieces of the lump hematite ore and were packed inside a closely fitting nickel tube. After reduction in H2 for the needed time, the sample was partitioned axially and polished, and the thickness of the Fe layer was determined. The results of the experiments have shown that, when the thickness of the reduced Fe layer was around 1 mm, the further reduction proceeds in agreement with the parabolic rate law, which is similar to the outcome of the pore diffusion control. These tests have demonstrated that as the thickness of the porous Fe layer grows, the rate of reduction is eventually controlled by gas diffusion in the pores of the Fe layer.

Partial internal reduction preceding the main advancing front of the Fe layer can lead to the entrapment of some FeO in the reduced layer. This situation can lead to sluggish removal of O2 in the final stages of reduction.

As the reduction temperature decreases the pore structure becomes much finer, presumably with many narrow channels and bottlenecks on connected capillaries, when Knudsen diffusion predominates, hence low values of the ratio of the effective molecular diffusivity/effective average Knudsen diffusivity. As the pore structure becomes coarser with increasing reduction temperature, bringing about easy passage of gas through the pores, the ratio becomes higher.

The effect of gas composition on the time to achieve 50 %, 75 %, 90 %, and 95 % reduction for sintered hematite ore pellets and magnetite ore pellets reduced at 900 deg C by H2-CO-CO2 mixtures (with CO/CO2 ratio equal to 9 to suppress soot deposition), is that as H2 is replaced by CO the time of isothermal reduction to achieve a given percentage of O2 removal increases gradually up to about 50 % CO and with further addition of CO there is a marked increase in the time of reduction. The time of reduction for 100 % (CO/CO2 ratio equal to 9), is about 10 times more than in H2 at the same temperature. The molecular gas diffusivity in a binary system, such as H2-H2O or CO-CO2, as derived from the kinetic theory of gases, is an invariant for the system and essentially independent of the gas composition. However in ternary and multi component systems, each species has a different diffusivity and varies with the gas composition. Furthermore, the rate equation for diffusive flux is complicated.

The reduction behaviour of hematite ore pellets in the H2-CO mixtures has shown a pattern similar to that observed in the H2 and CO, which is the rate of reduction beyond about 50 % O2 removal is controlled by gas diffusion in the pores of the Fe layer.

Limiting mixed control in initial rate

In the early stages of the reduction, the rate of reduction is controlled jointly by (i) gas diffusion in the pores of the FeO (solid state diffusion in FeO can be ignored), and (ii) reaction on the pore walls of the FeO. This implies a thin porous Fe layer and rapid gas diffusion therein. Depending on the porosity of FeO and the gas diffusivity therein, there is partial internal reduction ahead of the nominal Fe/FeO interface. The reaction of H2 with porous FeO is usually confined to the pore mouths close to the nominal Fe/FeO interface.

Partial internal reduction

Depending on gas composition, temperature, pellet size and total gas pressure, there is mixed rate control during some period of the reduction within the framework of the limiting rate laws. The rate equation is normally based on the assumption that the gaseous reduction of the pellet is controlled jointly by slow counter-current diffusion of gas through the inter particle pores of the pellet, and by slow chemical reaction of the gas with the Fe oxide at the Fe oxide-/Fe interface of the particles.

Water-gas shift reaction

The water gas shift reaction plays a significant role in the direct reduction processes which use reformed hydro-carbons as reductant in the reduction of iron oxides. It is generally agreed, from the different rates of reduction of iron ore by CO and by H2 and the marked effect which even slight amounts of H2 contained in a CO/CO2 mixture have on the reduction rate, that the H2 is the actual reducing component in such gas mixtures. The CO is considered to serve mainly to reduce the resultant steam (H2O) back to H2. The reactions are (i) H2 + FeO = H2O + Fe, and (ii) H2O + CO = H2 + CO2.

The second sub-process of this reaction is known as the water-gas shift reaction. It is well known that this process needs a catalyst. In iron ore reduction all of the products (Fe3O4, FeO, and Fe) come under consideration as possible catalysts. Out of these particularly active is solid Fe. The process of reduction of iron ore in CO/CO2 mixtures containing H2 is, hence, to be understood, when metallic Fe is present, as a reaction sequence. Sub-reaction (i) the reduction proper, takes place at the Fe oxide surface while the sub-reaction (ii), the regeneration of the H2 by the water-gas reaction, takes place at the Fe surface.

The spatial separation of the two sub-reactions needs their connection by a transport process, which is to take place as a gaseous diffusion or surface diffusion by one of the participants in the reaction. The optimal conditions occur at the 3 phase boundary Fe/Fe oxide/gas.

Swelling during reduction

The apparent volume of iron ore or pellet usually increases during reduction. This is called swelling. There are broadly three kinds of swelling behaviour which can be seen. These are known as (i) normal swelling, (ii) catastrophic swelling in which there is a sudden volume expansion with the conversion of FeO to Fe, the Fe appearing in the form of filamentary growths, known as whiskers wires of fibrous Fe, and (iii) bursting expansion, a typical behaviour of Fe-rich materials containing small quantities of alkalis. This latter kind of behaviour is different to catastrophic swelling (although it is no less serious) in that a major part of the expansion takes place before the appearance of Fe as a reaction product.

It can be said that neither lump ore nor sinter are known to swell abnormally or catastrophically, whereas certain types of pellets do, and give rise to operational problems by reducing the permeability of the burden as abnormally swollen pellets are soft, spongy and tend to disintegrate.

The specific volumes of different Fe oxides and Fe as reported in literature is 0.272 cc of Fe2O3 per gram of Fe (at room temperature), 0.270 cc of Fe3O4 per gram of Fe, 0.231 cc of FeO per gram of Fe (23.5 % O2), and 0.128 cc of Fe per gram of Fe. Hence, the volume is expected to decrease during each stage of reduction. However, the main cause of swelling of Fe ores is caused by the transformation of the hexagonal hematite ore into cubic magnetite ore and the resulting lattice disturbances. Lattice disturbances cause pore formation, whereby there is a considerable increase in the apparent volume of the Fe ores during the transformation from hematite to magnetite.

In general, during reduction in CO rich gas the swelling is much greater than in H2-rich gas. The reason for this behaviour is that metal dusting occurs during C deposition in CO containing gas mixtures. However, it is difficult to explain swelling which can occur during reduction in CO-CO2 gas mixtures when there is no C deposition. The cause and effect of swelling or shrinkage accompanying reduction have not yet been solved.

Usually there are two types of impurities in ore pellets. These are (i) impurities with a impeding effect on swelling, and (ii) impurities with an improvement effect on swelling.  The example of first is silica (SiO2) while for the second is alkalis (K2O, Na2O). It has been noticed that reagent grade Fe2O3 pellets containing SiO2 upto 5 % do not swell when reduced in CO – CO2 gas mixtures and also certain amount of SiO2 is necessary in acid pellets to maintain strength and to prevent catastrophic swelling. In second case, it is seen that small additions of alkalis Na2CO3 or K2CO3 in the range of 0.1 % to 1 % can result in the catastrophic swelling in H2 or CO of otherwise normal ore pellets. The effect of alkalis becomes more pronounced with increasing basicity (CaO/SiO2) ratio in the pellet. The adverse effect can be prevented by the addition of a fine-grained acidic gangue to form stable alkali silicates.

There are some contradictory observations of the effect of impurities in ore pellets (e.g. lime content). A small amount of CaO addition (less than 0.1 %) to hematite ore pellets causes considerable swelling during reduction and this suggested that CaO is a cause of catastrophic swelling. On the other hand, it has been noticed that around 1 % CaO addition to hematite ore pellets suppresses swelling during reduction. These variations in the observed effect of CaO on swelling can be due to the presence or absence of other impurities in the iron ore, such as alkalis.

Reduction of hematite ore by C

The reaction between hematite ore and C is of fundamental importance in the preparation of metallized ore pellets. Much of the new interest has been stimulated by the development of the rotary kiln process which uses solid C as the reductant in the production of direct reduced iron (DRI). It is generally accepted that reduction of Fe oxide by C occurs through gaseous intermediates CO and CO2, except under a very high vacuum where the true solid-solid reaction is the predominant mechanism.

The reaction mechanism through gaseous intermediates which takes place during the reduction of hematite ore by C is through reactions (i) C(s) + 0.5 O2 = CO(g), (ii) FexOy(s) + CO(g) = FexO(y-1) (s) + CO2(g), and (iii) CO2(g) + C(s) = 2CO(g).

The initial formation of CO is an important step in the overall reaction rate. O2 of the entrapped air together with O2 gas released by the dissociation of Fe oxides reacts with C to yield CO (first reaction). In addition, some CO can also be formed by true direct reduction occurring at the points of contact between the C and Fe oxide particles. CO gas thus produced readily reacts with hematite ore particles (second reaction). The Boudouard or the solution-loss reaction between CO2 gas and C particles regenerates CO gas (third reaction) and thereby tends to restore the reducing potential of the gas-phase contained within the pores of the sample. The oxidation of certain types of C in CO2 is catalyzed in the presence of certain metals and metal compounds. The rate enhancement of the process has been observed with the addition of Li2O (lithium oxide) and the inhibiting effect has been reported with addition of FeS (ferrous sulphide). Metallic Fe has been found to be a good catalyst for the gasification of graphite (C). Because of this unpredictable catalytic reaction in the mixture, equations derived through mathematical modeling to describe the overall rate of the reaction are of limited value and can be applicable only to those systems where reactions are not catalyzed.

At moderately high temperatures (e.g. 1000 deg C) the rates of the Fe oxide reactions (at temperature greater than 570 deg C and the sequence is Fe2O3, Fe3O4, FeO, Fe) are much greater than that of the Boudouard reaction. In other words the overall process becomes limited by the availability of the CO gas according to the Boudouard reaction. Thus at steady-state the composition of this gas-phase closely corresponds to the equilibrium gas-phase composition for FexOy/FexO(y-1).

Fe oxides reduction with hydrocarbons

Hydrocarbons can be used in two ways as a reducing agent for the production of DRI. These are (i) direct use of hydro-carbons or a mixture of gas containing hydro-carbons, and (ii) use of the reformed hydrocarbon products (CO, H2), by reforming within the reduction reactor (it has been found that auto-catalytic reforming of some hydro-carbons within the reducing furnace provided an access of macro and micro porosity which leads to more extensive reduction and also which leads to the deletion of the capital cost of gas reformer and processing.

There are a few studies using directly hydrocarbons or a mixture of gas containing hydrocarbons as reductant for direct reduction of iron ores. Two important points emerge from these studies. The first is that the rate of reduction with hydrocarbons is slow and the production of a high quality of DRI is troublesome and uneconomical. The second point is that these studies have been done under isothermal conditions in a thermo-gravimeter with single particle or powder compact, thus the results are of only theoretical value.

Theoretical importance of investigations with hydrocarbons – The kinetics of ferric oxide reduction by pure methane (CH4) has been studied in the three temperature ranges of (i) low temperature (500 deg C to 600 deg C), (ii) medium temperature (650 deg C to 750 deg C) and (iii) high temperature (800 deg C to 950 deg C). At the low temperature, the reduction proceeds only from Fe2O3 to Fe3O4. A prolonged holding of the sample in a stream of CH4 has not led to any process extension beyond this stage. The rate became appreciable at 650 deg C. In special experiments after the Fe3O4 composition has been reached, the sample has been reduced further by H2 and CH4. It has been shown that CH4 reduction in the low temperature range beyond the Fe3O4 stage occurs only if a sufficient quantity of metallic Fe has been built up. In this case the reducing agent has not been CH4, but its decomposition product, H2. C formed by CH4 decomposition takes almost no part in the reduction and gets accumulated in the sample.

In the medium temperature range the conversion of Fe3O4 to FeO takes place but at low rates. A sharp rise in reduction rate is observed on going from 750 deg C to 800 deg C. The process becomes very sensitive to temperature changes beyond 800 deg C, and accelerated considerably in the high temperature range, when metallic Fe appeared in the sample. The appearance of metallic Fe at the FeO to Fe stage, at comparatively high temperatures indicates a decisive role of metallic Fe as a catalyst for reforming CH4 by the reduction products (CO2, and H2O). In the absence of a catalyst, the decomposition of CH4 and its reforming by the reduction products (CO2, H2O) do not occur to any substantial extent and no C accumulation in the sample has been observed. When the Fe catalyst is present, CH4 dissociation into the elements takes place only at very late stages of reduction, when there is insufficient CO2 and water vapour to convert all the CH4 diffused into the sample. C build-up in the sample starts from that stage.

In the 2-stage production of DRI with CH4, it has been found that the complete decomposition of CH4 in the presence of the Fe bearing material occurs at temperatures of 850 deg C to 900 deg C, which is 400 deg C to 450 deg C lower than on an inert surface (e.g. fire clay), while the reaction rate, conversely, has been 10 times higher. The products of the first stage are a sooty Fe containing 30 % to 50 % C and technically pure H2.

In the second stage, the product of the first stage (sooty Fe with highly dispersed C in the pores of DRI and on the surface of the Fe particles) has been used as an active reducing agent and mixed with mill scale or concentrate. The mixture has been reduced in the temperature range 1050 deg C to 1100 deg C with a make-up reducing agent of H2 reformed natural gas. The results of industrial trials has shown that the use of sooty Fe instead of soot, petroleum coke and the other known carbonaceous reducing agents considerably intensified the Fe-oxide reduction process. As is well known, the direct reduction of Fe oxides with C is directly related to the rate of reaction between the C and CO2. The sooty Fe can have intensified the rate of Boudouard reaction.

The isothermal reduction of hematite ore pellets (with 10 % to 15 % porosity) in a thermo-balance with a mixture of CH4-H2 (containing 4.5 % CH4) within the temperature range 700 deg C to 1000 deg C has shown that the reduction is chemical – controlled initially and diffusion – controlled in the later stages. It has been shown that reduction in pure H2 is faster than in the CH4- H2 mixture. This difference is attributed to C deposition in the outer reduced layers of the pellet, causing resistance to gas diffusion when the reducing gas contained CH4. It has been shown that the excess residual C can be removed from the reduced iron at lower temperature by its hydrogenation.

In another study, it has also been demonstrated that it is possible to hydrogenate residual C in direct reduced products to CH4. The C formed as a result of the reduction of Fe oxide in a mixture of CH4 and H2 (containing 20 % CH4) reacted with steam (H2O) according to the water gas reaction to regenerate H2 and produce CO.

Pure ferric oxide briquettes were reduced at temperatures ranging from 800 deg C to 1050 deg C, in gas mixtures containing H2, CO, CH4, N2 and CO2, which has been obtained by partial oxidation of natural gas with air. The CH4 content of the reformed gas mixture was between 13 % and 16 %. The overall reduction rate again has been controlled initially by chemical reaction and the gaseous diffusion has been applicable during the latter stages. It has been shown that the hematite ore briquettes have swelled and considerable porosity has been was developed during reduction. The solid-state diffusion rates increased more rapidly with temperature than it did by interfacial or gaseous diffusion reaction rates. The reduction of porous (30 % porosity) Fe ore in CH4 has indicated that the reaction proceeded stepwise from Fe2O3 to Fe3O4, FeO and Fe. The Fe catalyzed the CH4 cracking reaction. Optimum conditions for CH4 utilization occurred at around 1000 deg C.

The above findings are not consistent with the earlier studies on the understanding of high-grade porous (around 30 %) or dense hematite ore reduction kinetics, which had shown that the rate of reduction can be considered to fall between 3 limiting cases, namely (i) uniform internal reduction, (ii) limiting mixed control, and (iii) diffusion in porous iron layer, respectively with the rate of reduction corresponding to, (i) chemical control, (ii) the overall chemical control and diffusion control, and (iii) diffusion control. The overall rate of reduction is not controlled by only one of these rate controlling mechanisms and can be changed from one limiting case to another during the course of reduction.

In one of the studies it has been found that the most important factors controlling the extent of reduction are (i) the temperature, (ii) the composition of gas, presence of unreacted hydrocarbons in the reducing gas, the ratio of H2/C in it, and reducing capacity, (iii) the ore particle size, and (iv) the residence time for reduction.

Reduction of Fe oxides with the products of CH4 reformed with H, O within the reduction furnace – In early 1981 a commercial process has been introduced, using gaseous mixtures containing upto around 30 % by volume of CH4 (e.g. coke oven gas), for the direct gaseous reduction of Fe ore in a counter current moving bed shaft furnace. The furnace contained a reduction zone, a cooling zone, and an intermediate reforming zone. A hot mixture of coke oven gas and steam has been fed to the intermediate zone and reduced Fe ore therein catalyzed the reforming of the CH4 to CO and H2. The reformed gas flows upward into the reduction zone for the reduction of Fe ore.


Leave a Comment