kinetics of fluidized bed iron ore reduction

1. Introduction

In the near future many so-called mini mills will replacethe conventional route of steelmaking in integrated mills.Mini mills use scrap or scrap substitutes (e.g. direct re-duced iron) as feed material in electric arc furnaces.1) Tosatisfy the increasing demand on scrap substitutes newprocesses for the direct reduction of iron ores have to be de-veloped. Direct reduction processes operate in shaft fur-naces, rotary kilns and fluidized beds. The reduction in flu-idized bed reactors offers the advantages of no feed ag-glomeration, uniform temperature in the reactor and excel-lent heat and mass transport. Some of these fluidized bedprocesses (e.g. FIOR process, FINMET process2) andCircored process) produce hot briquetted iron, whereas oth-ers (e.g. Iron Carbide process) produce Fe3C. Both productscan be melted in electric arc furnaces. The second applica-tion of fluidized bed reactors in iron ore reduction is theprereduction stage in a smelting reduction process (e.g.FINEX process,3,4) DIOS process. For efficient use of rawmaterials fundamental knowledge on reduction kinetics isrequired.

In the 1960s intensive investigations on the fluidized bedreduction of iron ore fines by pure hydrogen or hydrogen/nitrogen-mixtures were started.513) The reduction processesuse a reducing gas produced by steam reforming of naturalgas. Therefor the reducing gas is a mixture of H2, CO, CO2,CH4, N2 and H2O. Thus a strong interest in the kinetics ofreduction by a reducing gas containing CO and CH4 exists.Some authors tested the kinetics of reduction by mixturesof H2, CO, CO2 and H2O,

1416) others the carburization ofiron ore fines by adding CH4 to the reducing gas.

17,18)

Industrial scale fluidized bed reduction units operate at ele-vated pressure, but only a few kinetic studies for the reduc-tion of iron ore in fluidized beds were done at elevated pres-sures.1922)

The scope of this work was to investigate the reductionkinetics of iron ore fines under conditions similar to indus-trial scale units. That means that a laboratory fluidized bedreactor for reduction tests had to be operating at elevatedpressure, in wide temperature range and with a reducinggas containing all species of reformed natural gas.

ISIJ International, Vol. 40 (2000), No. 10, pp. 935942

935 2000 ISIJ

An Experimental Study on the Kinetics of Fluidized Bed Iron OreReduction

Arno HABERMANN, Franz WINTER, Hermann HOFBAUER, Johann ZIRNGAST1) and Johannes Leopold SCHENK1)

Christian Doppler Laboratory, Institute of Chemical Engineering, Fuel and Environmental Technology, Vienna University ofTechnology, Getreidemarkt 9/159, A-1060 Vienna Austria. E-mail: [email protected]) VOEST-ALPINE Industrieanlagenbau GmbH (VAI), Turmstrasse 44, P. O. Box 4, A-4031 Linz Austria.

(Received on November 22, 1999; accepted in final form on June 5, 2000)

To optimize existing iron ore reduction processes or to develop new ones, it is necessary to know the re-duction kinetics of the iron ore of interest under the relevant operating conditions. In this work the reductionkinetics of hematite fine iron ore was studied for industrial-scale processes using the fluidized bed technolo-gy. Especially designed batch tests were performed in a laboratory-scale fluidized bed reactor fluidized withH2, H2O, CO, CO2, N2 at atmospheric and elevated pressures to simulate the relevant process conditions. Toobtain the reduction rates and the degree of reduction, the concentrations of H2O, CO, and CO2 in the out-let gas were analyzed by FT-IR spectroscopy.

Preliminary reduction tests showed a strong effect of the sample weight on the reduction rates, especial-ly in the early stages of reduction. The optimum sample weight was determined by partly replacing thehematite with silica sand. Additionally, the silica sand provided a constant and stable flow pattern through-out the reduction tests. The effects of temperature, gas composition, particle size and pressure on the ratesof reduction were tested and discussed.

Rate analysis showed the existence of two phases with different rates during the reduction tests.Additional investigations (microscope analysis, SEM) demonstrated that in the first phase the rates werecontrolled by mass transport in the gas phase and in the second phase by the reduction process within thesmall grains of the iron ore particles.

KEY WORDS: iron ore reduction; high temperature fluidized bed; reduction kinetics; elevated pressure; H2CO gas mixture.

2. Experimental

2.1. Experimental Setup

The operating conditions of the newly developed flu-idized bed reactor can be summerized as follows:

absolute pressure up to 10 bar temperature range from 773 to 1 173 K the reducing gas contains H2, CO, CO2, CH4, H2O and

N2The concept of the experimental setup is shown in Fig. 1:

The fluidized bed reactor and the electrical heating shellswere set into a pressure vessel to avoid pressure diffferenceat the hot walls of the reactor (up to 1 173 K). This conceptguarantees the fulfillment of requirements and safe opera-tion under the conditions listed above. The welding exami-nation and pressure test of the apparatus were performed bythe Technical Inspection Association (TUEV). The reduc-ing gas was stored in gas bottles and the demanded mixturewas provided by mass flow controllers. If the reducing gascontained H2O the gas was sent through a heated water bathbefore entering the reactor, i.e. the reducing gas was en-riched with H2O to the demanded content. To heat up thereducing gas before entering the fluidized bed where the re-duction of the iron ore takes place, the gas was preheated ina fluidized bed filled with pure silica sand. Both fluidizedbeds were heated by electrical heating shells. The tempera-tures were controlled by PID-controllers and the tempera-tures of the fluidized bed for reduction, the preheating bed,the freeboard (cyclone height), the exhaust gas and insidethe pressure vessel were recorded. The pressure within thefluidized bed reactor was controlled by the exit valve. Thepressure in the pressure vessel was regulated to minimize

differential pressure between the fluidized bed reactor andthe pressure vessel. At the exit valve the pressure was re-duced to atmospheric pressure and the gas was analyzed,burned and sent to the chimney.

To minimize iron ore losses during the reduction tests,the upper part of the reactor was designed of a conicalshape to reduce the superficial velocity. The cyclone locat-ed within the reactor separated the iron ore particles fromthe exhaust gas. One important aspect of the reactor designwas to guarantee the same operating conditions along theheight of the reactor. Therefore, the heating shells had to befixed up to the top flange of the reactor.

The iron ore samples were placed into the reactor and thereactor purged with pure nitrogen as inert gas while heatingup to the desired temperature. As soon as the desired oper-ating temperature and pressure had been reached the flu-idizing gas was switched to the reducing gas composition.While the reduction test gas analysis was obtained by usingFT-IR spectroscopy in combination with a long-path length,low-volume heated gas cell. The following species weremeasured: CO2, CO, CH4 and H2O. Since H2O was ana-lyzed, the reactor exit tubes and the total sample line(pump, filter, gas cell) had to be heated to prevent conden-sation. At the end of the reduction test the fluidizing gaswas switched back to pure nitrogen again and the reactorwas cooled down to ambient temperature. The reduced ironore samples were removed and analyzed by a titriometricmethod (iron chloride method) for total Fe, Fe and Fe12.

To obtain knowledge about the structural changes andspatial element (Fe, O) distribution within the iron ore par-ticles during reduction, partly and totally reduced particleswere analyzed by optical and scanning electron microscopy

ISIJ International, Vol. 40 (2000), No. 10

2000 ISIJ 936

Fig. 1. The experimental setup.

techniques.The experiments were carried out with Mt. Newman

hematite ore from Western Australia. Chemical analysisand physical properties of the ore are listed in Table 1 andTable 2. The particle size was 0.1250.5 mm, unless stateddifferently.

2.2. Determination of the Rate of Reduction

FT-IR spectra were recorded every 10 s (Figure 2 showsa typical spectrum obtained during the first stage of reduc-tion.). Using predefined calibration functions the progressof the height or area of a selected peak led to the concentra-tion plot shown in Fig. 3. All oxygen containing species(H2O, CO, and CO2) were detected by FT-IR spectroscopy.The number of oxygen atoms reduced from the iron ore byReactions 1 and 2 can be calculated by the following bal-ance of the oxygen atoms (Eq. (1)):

FexOy1H2fi FexO(y21)1H2O Reaction 1

FexOy1COfi FexO(y21)1CO2 Reaction 2

ORed5Oout2Oin....................................................(1)

ORed [mol] number of oxygen atoms reduced from theiron ore

Oout [mol] total number of oxygen atoms leaving thereactor in the form of H2O, CO, and CO2

Oin [mol] total number of oxygen atoms entering thereactor

ORed in each time step i divided by the total number of oxy-gen atoms initially present in the iron ore and the time stepgives the actual rate of reduction:

...............................(2)

O0 [mol] total number of oxygen atoms initiallypresent in the iron ore

[1/min] rate of reduction

D t [min] time step

The fractional reduction can be obtained by:

...................................(3)

R [2] fractional reductionDO [mol] loss of oxygen

DO is calculated by summarizing the number of oxygenatoms reduced from the iron ore (ORed in [mol]) in eachtime step i:

..............................(4)

Using Eq. (4) the rate of reduction and the fractional reduc-tion can be determined (Fig. 4). For a better comparison ofthe reduction tests carried out at different conditions therate of reduction is plotted versus fractional reduction (R)(Fig. 5).

The fractional reduction of the iron ore samples at theend of a reduction test was calculated as shown above and

O O ii

t

55

Red,

0

RO

O5

0

dR

dt

dR

dt

O

O ti5 Red,

0


937 2000 ISIJ

Fig. 2. Typical FT-IR spectrum obtained during the first stage ofiron ore reduction (Temperature: 1 053 K, pressure: 1.2bar, particle size: 0.1250.5 mm, inlet gas composition:55 vol% H2, 9 vol% CO, 5 vol% CO2, 6 vol% CH4, 25vol% N2)

Fig. 3. Concentration of exhaust gas measured by FT-IR spec-troscopy for conditions stated in Fig. 2.

Table 1. Chemical analysis of the hematite ore in mass%.

Table 2. Physical properties of the hematite ore.

the result was compared with the titriometric analysis per-formed with the sample. The difference between these twodifferent methods was always less than 5%.

2.3. Variation of Sample Weight

Preliminary reduction tests were performed with a sam-ple weight of 300 g of iron ore. This sample weight (at a su-perficial velocity of 0.25 m/s) led to a well mixed bubblingfluidized bed with a bed height of 70 mm. The quantitativeanalysis of the exhaust gas showed that in the first stages ofreduction the H2O content exceeded 40 vol%, i.e. most ofthe H2 at the inlet (55 vol%) was consumed by the reductionof the iron ore and the rate of reduction was controlled bythe supply of H2 in the reducing gas. Under those condi-tions the determination of the reduction rate will give mis-leading results. To prevent this effect, the mass of iron orein the reactor had to be decreased. The iron ore was partlyreplaced by the inert material silica sand to maintain con-stant hydrodynamic conditions in the fluidized bed alongthe whole reduction process. The bed of silica sand provid-ed a constant matrix for the iron ore particles in terms ofhydrodynamic properties. However, the size of silica sandparticles had to be adapted to the higher density of the iron

ore so that the minimum fluidization velocity for both ma-terials was equal and the iron ore particles mixed well withthe silica sand. This was proved by tests performed in acold model plexiglass unit (Fig. 6). To obtain hydrodynam-ic similarity between cold model tests and reduction teststhe operating conditions for the cold model tests were de-fined using scaling criteria.23) Optical observations showedperfect mixing of the iron ore and the silica sand.

The mass of the iron ore sample was reduced stepwisefrom 300 g to 7 g to find the optimum sample weight: Asshown above a high mass of iron ore led to gas supply con-trolled reactions whereas a small mass of iron ore producedinaccurate results especially at the end of the reduction testwhen the concentration changes in the exhaust gas werevery low. In Fig. 7 the effect of the iron ore sample weightis shown: The rate of reduction increased strongly with de-creasing sample weight. At a sample weight less than 30 ginaccurate results (caused by very low changes in the ex-haust gas) arose. The maximum H2O concentration in theexhaust gas for a sample weight of 30 g was 8 vol%. Thismeans that the reducing potential of the exhaust gas was


2000 ISIJ 938

Fig. 4. Rate of reduction obtained from gas concentration mea-surements (Fig. 3) and fractional reduction plotted versustime for conditions stated in Fig. 2.

Fig. 5. Rate of reduction versus fractional reduction for condi-tions stated in Fig. 2.

Fig. 6. Mixing of iron ore particles (dark) in a fluidized bed ofsilica sand observed in the cold model. The photo showsthe state of mixing after 60 min. of fluidization.

Fig. 7. Effect of sample weight on reduction rate of hematite re-duced at 1 bar and 1 053 K (reducing gas: 55 vol% H2, 9vol% CO, 5 vol% CO2, 31 vol% N2).

still high and a differential operation of the reactor was pos-sible, i.e. the conditions (e.g. gas composition, temperature)in the reactor did not significantly change along the heightof the reactor. The first stage of the reduction was con-trolled by gas-phase mass transport (identified by a relative-ly constant rate of reduction, see also Sec. 3). For these rea-sons 30 g iron ore (combined with 270 g silica sand) wasconsidered to be the optimum sample weight. All furtherexperiments were carried out with this sample weight.

3. Results and Discussion

3.1. Variation of Temperature

In Fig. 8 the effect of temperature on the reduction rate ispresented: Increasing the reaction temperature led to higherrates of reduction. The strongest effect was observed be-tween 823 K and 973 K, while this effect was less strong athigher temperatures. It was noticed that the rate of reduc-tion dropped to almost zero for the test performed at 823 Kwhen the fractional reduction reached 0.5. Formation ofcarbon was observed at the tests carried out at 823 K and898 K. These results are in agreement with those of experi-ments performed by several researchers.5,911,22,24)

For all tests at temperatures higher than 898 K two phas-es with different rates of reduction can be seen (Fig. 8): Thefirst very fast peak at the beginning of the experiment is aresult of a start effect and the very fast reduction ofhematite to magnetite and is not considered as a specialphase of the reduction. The following first phase of reduc-tion is characterized by constant rates of reduction and endswith a strong decrease in the rate of reduction. During thelast phase the fractional reduction increases from 0.6 to0.85 at very small rates of reduction.

3.2. Variation of Pressure

The effect of absolute pressure is shown in Fig. 9. Thetests were carried out with constant superficial gas velocity,i.e. the overall molar flow for a test at 3 bar was three timeshigher than for the tests at atmospheric pressure. For test Bthe molar flow (or partial pressure) of H2 was kept constant(at 0.55 bar), for test C the molar flow was increased tokeep the H2 concentration of the reducing gas constant (at55 vol%). The increase of the absolute pressure with con-

stant H2 partial pressure led to no change in the rate of re-duction during the early and medium stages of reduction,whereas the increase of the H2 partial pressure led to signif-icantly higher rates of reduction.

An increase of the rate of reduction with increasing pres-sure was also observed by Ahner et al.,19) Ashie et al.21) andSato et al.22) Most of these tests were carried out with an in-creased molar flow at increased pressure and gave the sameresults as shown above. Reduction tests carried out at in-creased pressure and constant total molar flow (i.e. at con-stant molar flow H2) led to higher rates of reduction andseem to be in disagreement to test B in Fig. 9. But these ex-periments at constant total molar flow led to a decrease inthe superficial velocity (i.e. increase in residence time ofthe gas in the reactor) at increased pressure19) and can notbe compared with test B for that reason. In agreement tothe results presented in this paper Kawasaki et al.25) foundthat the variation of absolute pressure had no effect on therate of reduction.

3.3. Variation of Gas Composition

In Figs. 10 and 11 the influence of the gas compositionon the rate of reduction is given. Higher concentrations ofH2 in the reducing gas gave higher rates of reduction, addi-tion of H2O to the reducing gas decreased the rates of re-duction. If the reducing gas contained 55 vol% H2 the addi-tion of 9 vol% CO and 5 vol% CO2 did not affect the rate ofreduction. The reason for this effect is the slow rate of re-duction by only COCO2 containing mixtures (Fig. 11).The addition of small amounts (e.g. 6 vol%) of CH4 did notchange the rate of reduction.

The same efffect of changing H2 and H2O concentrationsin the reducing gas was observed by Hutchings et al.15)

Reduction by COCO2 gas mixtures was also studied byHutchings et al., Jess et al.17) and Kawasaki et al.25) leadingto similar result as shown above.

3.4. Variation of Particle Size

Most industrial applications use iron ores with a wideparticle size distribution. Figure 12 shows the rate of re-duction for particles larger than 0.5 mm compared with the


939 2000 ISIJ

Fig. 8. Effect of temperature on reduction rate of hematite re-duced by H2CO gas mixtures (55 vol% H2, 9 vol% CO, 5vol% CO2, 31 vol% N2).

Fig. 9. Effect of absolute pressure on reduction rate of hematitereduced at 1 053 K.A: pH250.55 bar, pCO50.09 bar, pCO250.05 barB: pH250.55 bar, pCO50.09 bar, pCO250.05 barC: pH251.65 bar, pCO50.27 bar, pCO250.15 barBalance is N2.

usually used particle size of 0.1250.5 mm. It can be recog-nized that the larger particles reach the same fractional re-duction at the end of the reduction, but the rates of reduc-tion decreased for increasing particle size. This is in agree-ment with the results obtained by McKewan26) and differsfrom the results obtained by Meissner et al.14) This apparentdisagreement may be caused by diffferences in structureand porosity of the analyzed ores.

3.5. Rate Controlling Step

Most of the diagrams rate of reduction versus fractionalreduction showed a peak at the beginning of the experiment(as a result of a start effect and the very fast reduction ofhematite to magnetite) followed by two phases with signifi-cant different rates of reduction. These two phases are sepa-rated by a strong decrease of the rate at a fractional reduc-tion of 0.5. One might interpret that this decrease was

caused by the transition from magnetite reduction towustite reduction. Samples taken during the second (slow)phase of reduction (80 min of reduction by 55 vol% H2 at1053 K) were analyzed and consisted of 91% total Fe,75.2% Fe, 11.5% Fe12 and 4.3% Fe31. Fe31 was found insamples taken during the slow phase of reduction of alltests and indicates the existence of magnetite. Different ki-netics of magnetite reduction and wustite reduction cantherefor not cause the strong decrease of the rate and a dif-ferent mechanism has to be found. Hence the possible ratecontrolling steps for iron ore reduction will be analyzed inthe next section.

The reduction of iron oxide particles in a fluidized bedreactor must proceed through the following steps:

1. Transport of the gaseous reactants (H2 or CO) fromthe bubble phase of the fluidized bed into the emul-sion phase.

2. Transport of the gaseous reactants from the emulsionphase to the external surface of the iron ore particle.

3. Diffusion of the gaseous reactants through the poresof the particle to the internal surface.

4. Reduction of the iron oreH21O

22fi H2O22e

2

Fe2112e2fi Fe5. Diffusion of products (H2O, CO2) from the pores of


2000 ISIJ 940

Fig. 10. Effect of reducing gas composition on the reduction rateof hematite at 1 bar and 1 053 K.A: 55% H2, 9% CO, 5% CO2B: 75% H2C: 55% H2D: 55% H2, 9% CO, 5% CO2, 6% CH4E: 55% H2, 9% CO, 5% CO2, 10% H2OBalance is N2.

Fig. 11. Effect of reducing gas composition on the reduction rateof hematite ore reduced at 1 bar and 1 053 K.A: 55% H2, 9% CO, 5% CO2B: 20% CO, 5% CO2C: 30% CO, 2.5% CO2Balance is N2.

Fig. 12. Effect of particle size on reduction rate of hematite re-duced at 1 053 K, 1 bar and by H2CO gas mixtures (55vol% H2, 9 vol% CO, 5 vol% CO2, 31 vol% N2).

Fig. 13. Calculated rate of reduction for different rate controllingsteps: A: Gas transport control; B: Interface reactioncontrol; C: Pore diffusion control.

the particles to the external surface.6. Transport of products from the external surface into

the emulsion phase7. Transport of products from the emulsion phase into

the bubble phase of the fluidized bed.Steps 1, 2, 6 and 7 can be identified as gas transport re-

sistances, steps 3 and 5 are referred to as shell layer resis-tance, step 4 as interface resistance.

These steps offer resistance in series of the overall reduc-tion reaction. If one of these steps is considerably slowerthan the others, it may be called the rate controlling step.The rate of reduction is given by dR/dt (12R)2/3 for inter-face reaction control, by dR/dt (12R)1/3/{12(12R)1/3}for pore diffusion control and by dR/dt const. for gastransport control.27) In Fig. 13 the calculated rate of reduc-tion versus fractional reduction is shown for either gastransport control, interface reaction control and pore diffu-sion control. If one compares the calculated rates of reduc-tion (Fig. 13) with the ones obtained by experiment (Figs.812) the following conclusions can be drawn: The firstphase of reduction is characterized by constant rates of re-duction (after an initial peak from a fractional reduction of0 to approx. 0.15), i.e. for most of the reduction tests, dur-ing the first phase, the rate of reduction is controlled by thetransport of the reduction gas from the bubble phase of thefluidized bed to the iron ore particle and the transport ofproduct gas from the particle to the bubble phase. Duringthe second phase of reduction the rate of reduction seemedto be controlled by interface reaction control or by pore dif-fusion, as the curves obtained by the reduction tests aresimilar to curves B and C in Fig. 13.

3.6. Microscope Analysis

To get detailed knowledge about the structural changesduring reduction of iron ore particles and to verify the as-sumptions stated above scanning electron micrographs(SEM) were made of cuts through partly reduced particles.Cuts through an iron ore particle reduced by H2 for 10 min.at 1 053 K showed that a high amount of pores is spreadthrough the whole particle (Fig. 14). The particle is subdi-vided into small grains (210 mm in diameter) surroundedby the pores. All grains in the particles consist of a brightshell on the outer surface (very likely metallic iron) and adark core (iron oxide). As this structure of the grains is thesame throughout the particle it is evident that the removalof oxygen takes place throughout the particle and the re-duction is not influenced by the pores.

A detailed view at the structure of the grains is shown inFig. 15: The thickness of the shell is 0.21.0 mm, that leadsto a total removal of oxygen from the smallest grains. Toverify the assumption that the shell consists of metallic ironand the core of iron oxide the distribution of selected ele-ments (Fe, O) was determined. It can be seen in Fig. 16 thatthe concentration of iron is the same throughout the grainsand that the concentration of oxygen is considerable higherin the core of the grains. This shows that the shell consistsof metallic iron while the iron oxide is present in the core.Since no micro-pores through this shell can be detected, adirect access of the reducing gas to the iron oxide via thepores is not possible. Hence the removal of the oxygenfrom the core of the grains is carried out by solid state dif-fusion and the rate of reduction in the second phase is con-trolled by this solid state diffusion.


941 2000 ISIJ

Fig. 14. Scanning electron micrograph of a cut through a singleiron ore particle (0.1250.5 mm) after 10 min of reduc-tion by 55 vol% H2 at 1 053 K.

Fig. 15. Scanning electron micrograph of a cut through an ironore particle partly reduced by H2 at 1 053 K. Brightareas at the outer surface of the single grains indicatemetallic iron.

Fig. 16. Distribution of oxygen (middle) and iron (right) in single grains (Fig. 15) surrounded by black pores. Dark-greyareas mark higher concentration of selected elements.

4. Conclusions

Fine iron ore was reduced in a wide range of fluidizedbed conditions. The rate of reduction was measured bymeans of FT-IR spectroscopy and the following conclusionscould be drawn:

(1) Exhaust gas analysis by FT-IR spectroscopy is avery useful technique for rate measurements of fluidizedbed iron ore reduction because all important oxygen con-taining gaseous species (H2O, CO, CO2) can be analyzed.Another advantage of FT-IR spectroscopy is the high timeresolution, that allows sampling every 8 seconds (importantat the first, very fast phase of reduction).

(2) It was shown by the analysis of the exhaust gas thatthe dilution of the iron ore sample with the silica sand(90% silica sand and 10% iron ore) in the reactor offers thepossibility to measure the rate of reduction not affected bythe supply of H2 in the reducing gas.

(3) The rate of reduction increased with higher temper-atures and higher molar flow of reducing gases (H2 andCO), but was not affected by the absolute pressure in the re-actor and small quantities of CH4 in the reducing gas (6vol%). Particles with a size of 0.54.0 mm showed signifi-cantly lower rates of reduction than particles with a size be-tween 0.125 and 0.5 mm, but reached the same fractionalreduction at the end of the reduction.

(4) Highly reduced particles are subdivided into smallgrains (210 mm in diameter) surrounded by pores. Allthese grains consist of a shell of metallic iron on the outersurface and a core of iron oxide.

(5) The reduction of iron ore particles proceedsthrough two phases with highly different rates of reduction:In the first, fast phase of reduction the rate is controlled bythe transport of the reducing gas from the bubble phase ofthe fluidized bed to the iron ore particle and the transport ofproduct gas from the particle to the bubble phase. Duringthe second, slow phase the rate of reduction is controlled bysolid state diffusion in the small grains of the iron ore parti-cles.

Acknowledgements

The authors wish to express their thanks to Mr. W.Steinbach, Institute of Chemical Engineering, for experi-mental assistance and to the Christian Doppler Forschungs-gesellschaft for the financial support.

REFERENCES1) K.-H. Schubert, H. B. Lngen and R. Steffen: Stahl Eisen, 116

(1996), 71.2) S. Zeller, G. Deimek, K. Milionis and L. Gould: La Revue de

Metallurgie-CIT, 95 (1998), 426.3) J. L. Schenk, W. L. Kepplinger, F. Wallner, H. G. Kim, S. Joo and I.

O. Lee: Proc. of 2nd ICSTI / 57th Ironmaking Conf., ISS,Warrendale, PA, (1998), 1556.

4) S. Joo, H. G. Kim, I. O. Lee, J. L. Schenk, U. R. Gennari and F.Hauzenberger: Proc. of SCANMET I, 1st Int. Conf. on ProcessDevelopment in Iron and Steelmaking, MEFOS, Lulea, Sweden(1999), 293.

5) S. Y. M. Ezz: Trans. Metall. Soc. AIME, 218 (1960), 709.6) J. Feinman and T. D. Drexler: AIChE J, 7 (1961), 584.7) W. Wenzel, F. R. Block and E. Wortberg: Arch. Eisenhttenwes., 43

(1972), 805.8) J. S. Sheasby and J. F. Gransden: Can. Metall. Q., 13 (1974), 479.9) L. Krol and W. Zymla: Arch. Eisenhttenwes., 49 (1978), 463.

10) C.-H. Koo and J. W. Evans: Trans. Iron Steel Inst. Jpn., 19 (1979),95.

11) N. Inoue, Y. Nakano, M. Ishida and T. Shirai: Int. Chem. Eng, 19(1979), 484.

12) K. Sakuraya, K. Kamiya and M. Tanaka: Trans. Iron Steel lnst. Jpn.,20 (1980), 92.

13) M. V. Srinivasan and J. S. Sheasby: Metall. Trans. B, 12 (1981), 177.14) H. P. Meissner and F. C. Schora: Trans. Metall. Soc. AIME, 221

(1961), 1221.15) K. M. Hutchings, J. D. Smith, S. Yrk and R. J. Hawkins:

Ironmaking Steelmaking, 14 (1987), 103.16) D. Neuschtz and T. Hoster: Steel Res., 60 (1989), 113.17) A. Jess and H. Depner: Steel Res., 69 (1998), 77.18) S. Hayashi and Y. Iguchi: ISIJ Int., 38 (1998), 1053.19) W. D. Ahner and J. Feinman: AIChE J, 10 (1964), 653.20) C. S. Kim: Diss. Abstr. Int. B, 31 (1970), 1248.21) T. Ashie, Y. Shinohara, R. Watanabe, M. Onoda and K. Mori: Trans.

Iron Steel lnst. Jpn., 25 (1985), B169.22) K. Sato, Y. Ueda, Y. Nishikawa and T. Goto: Trans. Iron Steel Inst.

Jpn., 26 (1986), 697.23) L. R. Glicksman: Chem. Eng. Sci., 39 (1984), 1373.24) J. S. Sheasby and J. F. Gransden: Can. Metall. Q., 13 (1974), 479.25) E. Kawasaki, J. Sanscrainte and T. J. Walsh: AIChE J, 8 (1962), 48.26) W. M. McKewan: Trans. Metall. Soc. AIME., 218 (1960), 2.27) L. von Bogdandy and H.-J. Engell: Die Reduktion der Eisenerze,

Springer Verlag, Berlin/Verlag Stahleisen, Dsseldorf, (1967), 156.


2000 ISIJ 942

kinetics of fluidized bed iron ore reduction

Documents