reduction and sintering of ouxed iron ore pellets-a comprehensive review

Reduction and sintering of Ouxediron ore pellets-a comprehensive reviewby S. Prakash *

Synopsis

Iron ore pellets with or without additives are commonly used asalternate charge material in both conventional and newerironmaking processes. The performance characteristics of thesepellets are normally assessed on the basis of their reducibilityvalues, reduction strength and resistance to clustering. Theseproperties depend mainly on the type and amount of additives andon the sintering phenomena of the pellets during firing. In the caseof composite pellets (ore-coal pellets), reduction and sintering ofthe iron ore fines occur simultaneously. The thermodynamic andreaction mechanisms of the processes involved in the reduction ofiron ore pellets with and without reductants, additives, etc., maynot necessarily be identical.

The reduction and sintering characteristics of pellets greatlyinfluence the reaction mechanism and the surface morphology of thereduced pellets. The approaches in the literature to the analysis ofthese reduction and sintering phenomena are broadly categorized asfollows: (i) empirical models, (ii) models based on kinetics, and(ill) models based on morphological changes. The relative merits

and de-merits of these models are illustrated by their application tothe reduction of fluxed ore-coal composite pellets. A critical analysisof the investigations that have been conducted on the reduction andsintering behaviour of such pellets is given.

It is shown that, although elegant empirical models and kineticequations have theoretical significance and can be usefnlIyemployed in elucidating the reaction mechanisms of iron ore IUlllPS,they are sometimes not directly applicable to fluxed ore-coalcomposite pellets. In studies of the latter, the approach shouJd be tocorrelate the morphological changes with the reduction mechanism.The paper also gives a brief account of an investigation on thereduction and sintering aspects of fluxed composite reduced ironpellets.

Introduction

Although the reaction of iron ore has been asubject of research and discussion for manyyears, it is still beset with uncertainties, withinconsistent experimental data, and withconflicting theories. The practical applicationof many studies has been severely limitedbecause most of them have been confined tocontrolled experiments and idealizedconditions, which are very different from theconditions in industrial practice. Althoughseveral investigations have been carried outon the gaseous and solid reduction of iron

The Journal of The South African Institute of Mining and Metallurgy

oxides (ore), uncertainties arise because of thecomplex nature of these oxides. These studies,which generally simulate the conditions ofblast furnace or direct reduced ironmaking,may not necessarily be applicable to theconditions prevailing in the combinedprocesses of direct reduction and smeltingreduction (DR-SR).

The production of liquid iron from theDR-SR route conceptually avoids thecokemaking and ore-agglomeration processes.But these process routes which are at differentstages of their development, cannot directlyutilize ore fines and coal fines likeconventional processes. The feed is generallyin the form of lumps and/or pellets of selectedsize. However, the lumps/pellets of iron oreshould exhibit adequate reducibility, and theirgreen and reduction strength characteristicsshould be suitable under given processconditions. The complexities in the case ofpellets arise from uncertainties in theirchemical composition due to impurities, fluxadditions, other additives, and binders.

The presence of impurities or additives canalter the intra-crystaline structure, latticediffusivity, porosity, surface orientation, non-stoichiometry, and even the reductionsequence of the iron oxides during reduction.The nature and composition of the reducinggases, and the type of reductant, are asimportant as the temperature in determiningthe kinetic parameters. Iron oxides can formmixed crystals or compounds with otheroxides, which affect the thermodynamicdriving force for their reaction and can altertheir stability ranges. For instance,Von Bogdandy and Engel! have shown that,when SiOz-CaO was added to iron oxide, themagnetite stage disappeared from thedecomposition sequence of the iron oxide.

. Ferrous Process Division, National Metallurgical

Laboratory. /amshedpur, India 831007.@

The South Ajhcan Institute if Mining and

Metallurgy, 1996. SA ISSN 0038-223X/3.00 +

0.00. Paper received Sep. 1994; revised paper

received/ul. 1995.

JANUARY/FEBRUARY 1'196 3 <Ill

liquidF.

1600

IB I

I1400

11t- F. .

Uquid Oxide

I ,1200 . ¥-F..Wustite,

1000

Reduction and sintering of fluxed iron-ore pellets

Most of the studies in the literature on reduction andsintering phenomena in fluxed composite iron ore pelletsgenerally deal with one or more of the following:

~ experimental measurement of the progress of reactionby weight loss due to oxygen removal and chemicalanalysis in isothermal and non-isothermal conditions

~ microscopicinvestigationof the reductionand sinteringbehaviour of pellets

~ development of mathematical models to predict theeffect of process parameters on the reduction andsintering of pellets.

The present paper describes and analyses various aspectsof these studies, showing that the correlations established donot necessarily elucidate the reaction mechanism of fluxedcomposite iron ore pellets. A brief summary of an ongoinginvestigation on such pellets is also presented.

Thermodynamic aspects of reduction

Reduction of iron ore

Thermodynamically, the reduction of iron oxide occurs insteps. Of all the iron oxides, hematite is the richest naturally-occurring form. Accordingly, the following stable iron oxidescan occur during the reduction (above 570°C): hematite

(FezO3)' magnetite (Fep 4)' and non-stoichiometric wustite

(Fel- ye, Y= metal deficit).Magnetite and wustite exist within the temperature range

shown in Figure 1. During the stepwise reduction of hematiteto magnetite, wustite, and metallic iron, changes occur in thecrystal structure. In hematite, the oxygen atoms are arrangedin a closely packed hexagonal structure but, in magnetite andwustite, they form a face-centred cubic structure.

Hematite has one of two lattice structures: rhombohedralcorundum (Alz03 type of structure) or a spinal type ofstructure as magnetite. The latter is metastable and can beobtained3 by the oxidation of magnetite below 400°C.

Wustite is markedly non-stoichiometric, and it has a cubicface-centred, closely packed oxygen lattice with the ferrousions arranged in the octahedral sites between the largeoxygen ions4. At 1000°C, the oxygen-to-iron atomic ratiovaries from 1.05 to 1.14 in equilibrium with the Fe and Fe304respectively. Above 850°C, the composition of wustite inequilibrium with iron is almost constant (Figure 1). Thelattice parameter of wustite varies linearly3 with the oxygencontent from 4.301 to 4.2816 Aat 22.2 to 23.92 per centoxygen (by weight) respectively, with corresponding densitiesof 5.728 to 5.613 g/cm3 at 25°C. The increase in latticeparameter and density as the oxygen-to-iron atomic ratioapproaches unity indicates the formation of cation vacancies5.6.

Edstrom3 has shown an increase in volume of 25 per centin the transformation of hematite to magnetite, and a further7 to 13 per cent increase in its transformation to wustite.However, owing to shrinkage in the metal, the overallincrease in volume during reduction is about 25 to 27 per cent.Edstrom also found that, when natural magnetite crystals arereduced to metal, there is no corresponding volume increasebut there is a shrinkage of about 4 to 5 per cent in the finalproduct. This helps to explain why the reducibility is greatlyimproved when magnetite is oxidized to hematite before it isreduced.

~ 4 JANUARY IFEBRUARY 1996

LJ

°.QJ

2ro:vCO-EQJ

f--

If11j

II

Wu.tite .magnetite

I

Mag~etit.

Hematit. .

800

600

,,-Fe.magn.tit.

400'FeO, FezO,Fep,

O. 22 24 26 28 30

°1; wt.%

Figure 1-lron-oxygen equilibrium diagram2

Magnetite's lack of expansion during reduction results inthe formation of dense layers of metallic iron surroundingremnants of wustite, which cut off the access of the reducinggas to the oxide and prevent complete removal of theoxygen7.

Reduction of iron oxide composites

Owing to the expansion of the wustite phase field with anincrease in temperature, larger amounts of foreign ions canbe substituted. Wustite forms a continuous solid solution8with MgO and MnO. It can dissolve9 28 per cent of CaD at1100°C, and 5 per cent and 1 per cent of CrZ03at 1350°C and850°C respectively.

The changes in free energy accompanying the reductionof various iron-oxide composites are discussed by Ward 10and by Richardson and Jeffesll. In composites, the oxygenpotential corresponding to the reduction to metallic iron isdecreased, and the thermodynamic driving force is lowered.Any elimination of oxygen in this case results in an enrichmentof the second phase in the oxide.

Richardson and Jeffesll demonstrated the effect of achange in the activity of an element or its oxide in dilutesolutions at the minimum reduction temperature and a givenoxygen potential. The presence of additional phases wouldintroduce thermodynamic activity gradients at fIXedtempera-tures and applied pressures, and would alter the availabledegree of freedom of the system. Fujii and MeussnerlZreported a decrease in the equilibrium oxygen pressure ofwustite with dilute chromium solutions. It is generally agreedthat the addition of only small amounts of foreign ionsproduces significant effects on the volume changes andreduction rates3. 13-Z0,Small additions of Na20, KzO, CaD,and LizO retard the FeZ03 to Fe304 reduction, but enhance the



FeO to Fe reduction21. While the addition of Si02 acceleratesthe initial stages of iron oxide reduction, AI2O3decelerates itat 550°c.

The addition of CaD or CaC03 increases the reducibility ofhematite by hydrogen. This is attributed to an increase inporosity, which promotes the wustite dissociationmechanism14.22. CaD and Si02 are major slag-forming oxidesduring induration of the pellets, and their ratio is important ingoverning the basicity of the gangue. Turkdogan and Vinters23found that an addition of CaD (up to 4 per cent by weight)and Si02 (up to 5 per cent by weight) affects the gaseousreduction of synthetic Fe203 and of ore pellets by carbonmonoxide differently: in the former reaction, it reduced theswelling of synthetic material but showed no pronouncedeffect on swelling in the case of the pellets.

Ca024-26and sodium oxide24.26-28cause abnormalswelling during reduction, but Si02 and AI2O3additionsreduce swelling28. 29mainly because of the improved slagbonds and the removal of basic oxides. The inability of MnOand MgO to promote swelling is a result of their highermobility and solubility in iron oxide.

EI-Geassy and his co-workers3o studied the effect of thedegree of crystallization of silica on the reduction kinetics of

Fe203 briquettes. They observed that the crystalline silicaacted as a surface promoter and accelerated the reduction ofiron oxide owing to the higher number of active sites forgas-solid reaction. However, they found that amorphoussilica retards the reduction because of the formation offayalite (Fe2Si04).

The effects of additions have been reported to vary withthe anionic nature of the compound. In the reduction of alkalichlorides and carbonates added to reagent-grade hematite,Bitsianes31 reported a catastrophic disintegration.Catastrophic swelling with Na2C03was observed by LU26.Fuwa and Banya28 report that NH4CIand FeCI3added to ironore formed whiskers of metallic iron, and Khalafalla andWeston15 observed that chloride and sulphide ions poisonedthe promoters of wustite reduction, although nitrate ions hadno effect. Na+was a better promoter when added as ahydroxide than as a chloride. When Fuwa and Banya28decreased the NaCIcontent of pellets that had showncatastrophic swelling from 0.3 to 0.08 per cent, there was nosignificant effect on the swelling. The absorption of sulphuron wustite surfaces inhibited the chemical reactions, andpromoted the growth of whiskers27.

Mechanism of reduction

The reduction of iron ore or any ferruginous raw materialwith a reducing material involves stepwise reactions (Figure 2).In the first stage of reduction, oxygen atoms undergo a severere-adjustment, which results in an increase in volume3 ofabout 25 per cent. This tends to open up the structure andfacilitates the subsequent reduction stages. In the transforma-tion of magnetite to wustite, the oxygen lattice remainsunchanged while iron atoms diffuse to fill the vacant sites inthe iron lattice. This can be accommodated with only a smallincrease in volume. Because wustite has a variable composition,there is also a small increase in volume as its compositionchanges from that in equilibrium with magnetite to that inequilibrium with metallic iron. On the other hand, the

The Journal of The South Afncan Institute of Mining and Metallurgy

Fe

\. \

\\\

b ~\\\\\IIIIIIII

I/

/

ibJp1b)

B

XoFexO

R(!'e)1

Superscript

b = Bulk gasFe = Ironh = Hematitem = Magnetite0 = Outer layers = Reactantt = Productw = Wustite

RW)

R(m)

)I R(w)

SA S,AR(rT,) Rlw)

S.B S.B

p(w)A

p(w)B

RS~R(Fi>J

S.B

RF.A

RF.B

R(O)Ap~)

Variablex = Interphase radiusR = ResistanceP = Pressure

Figure 2-Representation of a model with three interface concentricshells

nucleationgrowth of iron crystals results in shrinkage and alarge increase in the porosity of the metallic phase, whichgreatly facilitates the diffusion of the reducing gases from theouter surface of the particle to the wustite-iron interfaces.

Indirect reduction

For indirect reduction to take place, oxygen has to be removedby the contact of the reducing gas with the surfaces of ironoxide.

The fundamental steps involved in the reduction of ironoxide are as follows:

> diffusion of the reduci ng gas across the boundary layer> intraparticle diffusion of the reducing gas to the

reaction interface

> phase-boundary reaction: migration of Fe++andelectrons to the iron nucleus

> intraparticle diffusion of oxidized gas from the reactioninterface

> diffusion of oxidized gas across the boundary layer.

Diffusion control of the gas film

When the resistance of the gas film controls the reactions, theconcentration profile for the gas-phase reactant is as shownin Figure 3.

The mass transfer of the boundary-layer gas film becomesnegligible in the reaction of solids with a gas stream flowingabove critical gas velocities. Different kinetic models can beused for the various reduction mechanisms.

Control of the phase-boundary reaction

Control of the chemical reaction at the interface occurs whenthe reaction product is porous and offers no diffusionalresistances. At temperatures above 550°C, it was found

JANUARY/FEBRlI/':, !996 5 .....


i~Q)

'"os~0-rhosCl

'5c:0

~'E(I)(J

c:0

CJ

_l-,.- ......./' Gasfilm ""',

//

'\ SurfaceshrinkingI unreactedcore

I " SurfaceofI particle

l I:'

I I I~\ I I I

I \ I I I II ' I ,/ I

I" I /'. II I""',

I I.,,'" I II I -i---r I I

'-uL - _J - - _1- --~ :conc~~tration

I I I: 1 I mainbody

I I I I II of gas

I I I II I I

I 1 1

:I I I I I

I:

I I I11

For irreversibleI

1 I 1 I reactionI I I I CAe = 0I I I

CAg

CAs = CAe

re 0 re

Radial position

Figure 3-Representation of a reacting particle when gas-film diffusion isthe controlling resistance32

experimentally that, when the rate of reduction, ~, is given inmo1-cm-l,s-l, the reduction of iron oxide by H2 or COisshown by equation [1)33:

~ =k[po-pa, [1]where Po and Pi are the pressures of the reducing gas at theoxide surface and at the interface respectively.

The rate constant k for the phase-boundary reactiondepends on which of the iron oxides is reacting with the gas.McKewan34,in his study of the kinetics of iron oxide reduction,expressed phase-boundary reaction control as a shrinking-core model in terms of the fraction reacted, R:

poro [l-(l-R) 113]=kt, [2]where '0 is the initial radius of the pellet and Po is the density.The reaction rate d(i/dt in equation [2] is proportional to(l-R) 2/3. Spitzer et al.35 indicated that the summation of theremoval of oxygen at arbitrary points could possibly beproportional to (1-R)2/3. However, this rate equation does notdenote the mechanism unambiguously. Figure 4 illustratesthe concentration gradients within the particle when a chemicalreaction controls the rate.

Diffusion through the product layer

Figure 5 illustrates the situation in which the resistance to

- diffusion emanates from the product layer and controls the

reaction. Diffusion control occurs when the phase-boundaryreaction is faster than the rate of supply or than the removalof the reactants or products. When the geometry is sphericaland the texture porous, the rate-controlling model as developedby Crank, Ginstling and Brownshtein (CGB)32is as follows:

1-(2/3)R - (1-R)2/3 =De(Po - Pi)

t, [3]Po ToRT

~ 6 JANUARY IFEBRUARY 1996

'Eos

~~Q)

'"os~0-Il

~CAg = CAs = CAe

'5c:0

~'EQ)(Jc:0

CJ

""'.../' ....

",

/ ,/ '/ '

I \

I \1 \I \: I, I I n1\ I I /'I \ I I I

I

1 \ I I I II " 1/ 1I '1 /r II I ' I I ",/ I II I '1- -t'- I II 1 I --I 1 II I , I I I

I I

1 II II I

I:

CA.=Ofor Iirreversible I

reaction I

/: !

1IIIIIII

i

I

1IIIIIII I

0 rere

Radial position

Figure 4-Representation of a reacting particle when the chemicalreaction is rate controlllng32. 36

- - - °As = fluxofAthroughexteriorsurface

'"...

"" ofparticle(inward+.outward--)Gasfilm~

°A = fluxofAthroughsurfaceof any radius rd

°Ac =flux of A through

reaction surface

Typical position diffusion

region radius = rdconcentration A = CA

'Eos

~~

~CAg = CAs~0-~Cl

'5c:0~'Eg CAs= 00CJ

I II I

CA-_L_L-I I1 I1 I

re 0 re rd

Radial position

Figure 5-Representation of the situation in which product-layerdiffusion controls the reaction32. 36

where De is the effective gas diffusivity (Le. the coefficient ofporous diffusion of the gas).

Mechanism of mixed control

Mixed reaction control occurs when the rate of reaction isgoverned both by the phase boundary and by diffusion-control processes.

Seth and RoSS37developed an equation assuming ratecontrol by both the transportation of gas through the porous

The Joumal of The South African Institute of Mining and Metallurgy


product layer and the interfacial reaction. The equation is asfollows:

1/3 rokp [(1/2)-(R/3)-(I-R)2I3]I-(I-R) +

2kd

kp(CO - C)= t,

ropo

where [0 and [are the bulk-phase and equilibrium concen-tration of the reducing gas, Po is the density, and kp and kctare proportionality constants with the units of M/V t conc.and M/L t conc. respectively.

A similar model had been proposed earlier by LU38.

Model of nucleation and growth

When nucleation of the product at various preferential activesites in the crystal accompanied by the growth of these nucleiis the rate-limiting step, equation [5)39 applies:

In[~

]= kt.

I-R

This equation is derived from the Prout and Tompkinskinetic model, which assumes that the reaction starts at aseries of active centres. The }ohnson-Mehl or Avrami-Erofevequation has also been used4o to empirically describe thenucleation and growth type of reaction mechanism. The rateof reduction seems to be governed by the structure of the ironformed by reduction because of its influence on the processesof chemical reaction and diffusion. If the iron formed duringreduction is distributed over the surface of the FeO, and if itis difficult to disperse along the FeO and it grows vertically tothe FeO, the rate of chemical reaction becomes faster becausethe sum of the co-existing lines in the three phases per unitarea of the FeO surface becomes larger. The equation is

R = 1 - exp (-kt)n. [6]

[4]

Composition of the reducing gas32

The Fe-o-C (Figure 6) and Fe-O-H equilibrium diagramslshow that, thermodynamically, COis a more efficient reductantthan Hz below 820°C, and vice versa at higher temperatures.Reduction to lower oxides can be carried out by the use ofcontrolled oxygen potentia]s and gas mixtures of COor Hzwith their oxidized forms.

Investigations41-43 of reductions with Hz-CO mixtures ingas-based direct-reduction plants have shown that water-gasequilibrium is not attained at temperatures less than 1O00°C.Turkdogan and Vinters41 explain the internal reduction in

Hz-CO mixtures as consisting of two parallel reactions withHz and CO.Significant carbon deposition was obtained with25 per cent COor more. This can be attributed to the catalyticeffect of the reduced iron4Zand an inability to attain thewater-gas equilibrium in Hz-CO or Hz-COz mixtures41. Thereduction rates increase with an increase in Hz content.

The surface reaction rate is expected to decrease inproportion to the degree of dilution when inert gas diluentssuch as Ar or Nz are used with Hz or CO.Rao andAI-Kahtany44 reduced magnetite in Hz-He, HrAr, and Hz-Nzmixtures at 487°C and concluded that

~ Hz-He mixtures conform to first -order behaviour withrespect to the partial pressure of Hz

~ there are small retardations in reduction rate in Hz-Ar~ nitrogen has a strong poisoning effect on the reduction

kinetics.

Dam45 reported an enhancement of the kinetics bynitrogen in the reduction of hematite in Hz-Nz and Hz-Armixtures at 900°C.

[5]

Effect of temperature

The effect of temperature on the reaction rate is given by theArrhenius equation:

k =A exp (-E/Rl), [7]

600

Temperature of

ZZ 00 Z 600 30001.0

0.8 ( C)

(F~3CJN

8 0.6

8~~

0u 0.4t:C

0.2

0200 400

/" .

F/ \ LIq. e/ SFe \ ------

IIIII Liq. oxideIIIII,\

"- ....

1200 1400 1600


Figure 6-Equilibrium gas composition versus temperature for the Fe-O-C system, including the cementite equilibrium32

JANUARY/FEBRUi\f-(Y 199f 7 ....


whereE is the apparent activation energy and A a frequency

factor. Gas-solid non-homogeneous reactions include physicalprocesses in addition to the interfacial reaction.

Some remarkable retardations in reduction rates havebeen observed at various temperatures. Minimum rates havebeen reported at 600°C for the reduction of magnetite, at 650°Cfor the reduction of dense hematite in Hi, and at 975°C and1O25°Cfor the reduction of porous hematite with COandCO/H2respectively43. The lower reduction rates between 900and 950°C with H2 or H2-rich gas mixtures have beenattributed to the a-y-iron transformation45 at about 912°C.Moukassi et al.46attributed the minimum rate between 700and 900°C in the dry reduction of wustite by hydrogen to theformation of a dense layer of metal that acts as a barrieragainst the diffusion of gas through the pores.

Direct reduction

The thermodynamic aspects of the complex systems in iron-ore reduction are well known and easily understood. However,not much about the kinetics of the direct-reduction process isaccepted universally. Some kinetic aspects are brieflysummarized here. It is now accepted that there is neither anyuniversally accepted mechanism nor any particular value ofactivation energy. The rate is controlled by one or more ofseveral possible mechanisms, depending on the physicalnature of the system and the experimental conditions used totest it.

The direct reduction of iron oxides by solid carbon can becharacterized as reactions between a solid proceeding throughgaseous intermediates. Strictly speaking, such a reaction canbe truly visualized in an evacuated system, where the carbonmonoxide product is removed as fast as it is formed.

The reduction of iron oxide by solid carbon is generallyrepresented as follows, being based on two general hypothesesfor when a mixture of oxide and carbon (coke, graphite, coal,char, deposited carbon, etc.) is heated:

FexOy + C = FeXOY-l + CO

(solid) (solid) (solid) (gas) [8]

FexOy + CO = FeXOY-l + CO

(solid) (gas) (solid) (gas)

CO2 + C = 2 CO,(gas) (solid) (gas)

wherex= 1,2, or 3y = 1, 3 or 4.

The reaction between the two solid phases can begin onlyat points of contact between the carbon and the particles ofiron oxide. Contact is disrupted as soon as metallic iron appearsas an intermediate phase between the other two. From thenon, reduction can proceed only by the diffusion of carbonatoms through the layer of metallic iron to the metal oxideinterface. However, because carbon monoxide is one of thereaction products, it is generally accepted that the reductionof iron oxides by carbon proceeds indirectly by carbonmonoxide, and that the carbon dioxide formed then reactswith carbon to re-form carbon monoxide, as shown inequation [10].

Baldwin47 studied the reduction of a heated bed of ironoxide and coke, and confirmed the mechanism given by

[9]

[10]

~ 8 JANUARY/FEBRUARY 1996

equations [9] and [10]. According to him, the rate of dis-solution of carbon is much slower than the reduction of ironoxides at higher temperature, and is the rate-controlling step.This view is largely supported by Ostuka and Kunii48, Szendreiand Van Berge49, Abraham and Ghosh50, and Srinivasan andLahiri51. Ghosh and Tewari52, however, report that thereduction of iron oxide is more likely to be rate-controlling.Their view is based on the estimation of activation energyvalues. They did not attempt to develop any kinetic modd,but reported kinetic data as a function of temperature, particlesize, ratio of reactants, etc.

Ra053, working with synthetic mixtures of ferric oxideand graphite powder, proposed a model based on theassumption that the rate-controlling step in the overallreduction process is the reaction between solid particles ofcarbon and carbon dioxide. He proposed the followingequation:

log (1.743 -J) = (-R/2.303) t+ log (1.743). [11]

In equation [11], the degree of reduction is expressed interms of pseudo-kinetic parameters defined as the ratio ofweight loss at time t to the maximum possible weight loss.This is not necessarily a true kinetic equation. However, theequation was found to represent the kinetics of reductionquite satisfactorily. It was seen that the addition of smallquantities of Li2Ogreatly enhanced the reaction rate, whilethe addition of ferrous sulphide retarded it. This was tojustify the assumption since the Boudouard reaction was therate-controlling step for the overall reduction process.

Equations for the reduction of hematite and magnetitepellets containing coal or char were proposed by Seatonet al.54as follows:

for hematite,1n(1 - 0.98J) = -kt

and, for magnetite,In (1 - 1.037f) =-kt. [13]

The J in these equations is also not clearly related to thedegree of reduction.

Equations [12] and [13] are similar to that given byTurkdogan and Vinters55 for the gaseous reduction of porouspellets. Seaton et al.54say that the rate of iron oxide reductionis controlled by the chemically controlled Boudouard reaction.They also postulate that the overall rate may be controlled byheat transfer, which affects the carbon-gasification reaction.The gasification of carbon by CO2is highly endothermic(tJ.H=41 220 callmol) and requires a high temperature toproceed at an appreciable rate. The overall reduction rate israte-controlled by the rate of gasification of carbon53, 56. Thisconclusion is based on the observation of a temperaturegradient in the system, the high endothermicity of thecarbon-gasification reaction, and the significant increase inthe overall rate with an increase in the temperature of thereaction.

Elguindy and Davenport57 report that, at temperaturesbelow 1O20°C,the reduction mechanism of an ilmenite-graphite system is solid-solid interaction that depends on thecontact points. However, at temperatures above 1O20°C,themechanism of reduction is represented by diffusion of theproduct layer, namely,

1 - (2/3)R - (I-RJ2/3", kt,

where COhas been suggested as the diffusing species.

[12]

[14]

The Joumal of The South Afncan Institute of Mining and Metallurgy


Maru58 has shown that the reaction between COandcarbon controls the overall rate for the reduction of chromiumoxide and chromium carbide. At Iow conversion levels, herepresented his experimental results in terms of chemical-reaction control. He explained the resistance, and indicatedthat the complete range of reduction could possibly beexpressed by mixed control involving interfacial reaction andproduct-layer diffusion, as proposed by Seth and ROSS37.

The rate of gasification of carbon by carbon dioxidedepends on the reactivity of the carbon, the temperature, andthe availability of heat to maintain the reaction at the reactiontemperature. Therefore, the rate of carbon gasification thatcontrols the rate of reduction in the reduction of solid carbonwill, in turn, be controlled by the rate of heat transfer fromthe heat source to the reacting materials.

The rate of heat transfer will be affected by the thermalconductivity of the material through which the heat must passand by the distance between the heat source and the reactionsites. Little can be done to improve the conductivity of thematerials involved in the reduction of iron ore but, in thedesign of the equipment, consideration should be given to thelocation of the point of heat production as close as possible towhere the heat is needed. In this respect, it is interesting tonote that, in the Wilberg process59, the gasification of carbonand the reduction of iron oxide take place in separate vessels,and this provides better heat transfer than in processes wherethe carbon and the iron oxide are mixed. Hogana's process60,which has been in use in Sweden for many years, uses largeamounts61 of iron-ore fines and coke fines packed in alternatelayers or concentric rings in ceramic saggers. Prakash et al.62.63ascertained the possibility of producing iron directly from oreand coal fines through the RESINESSroute (RE: REduction,SIN: SINtering, ESS: Electro Slag Smelting), and studied kineticaspects and heat transfer in this process, which was developedat the National Metallurgical Laboratory in Jamshedpur. Theyinferred from their studies that, like the reducibility of ironores, the reactivity of carbonaceous materials varies widelyfrom one material to another and from batch to batch of thesame material. The surface area of the carbon available toprovide reaction sites for the carbon dioxide is the importantfeature, and this is determined by the particle size and, moreimportantly, by the porosity of the material.

The choice of reductant is a factor in the determination ofthe operating cost and energy efficiency of direct-reductionprocesses. Processed materials such as charcoal, char, andcoke are more porous and reactive than natural carbon-bearingmaterials such as wood, coal. and graphite. Charcoal is generallymore reactive than coke, and both Iow-temperature coke andchar are more reactive than hard, burnt coke. Coke made fromdifferent types of coal may have an important influence. Someas components or additives have a definite catalytic effect.Ra053 found lithium oxide to be one of these. Nickel oxidehas a similar effect56. On the other hand, the gasification ofcarbon is inhibited by the presence of sulphur in the carbon53.

Fruehan56 studied the reduction of iron oxide by carbonin two stages, and his observations indicate that the carbonmonoxide formed by the gasification of carbon is consumedas fast as it is produced. This supports the claim that theBoudouard reaction is rate-controlling.

Similarly, in the second stage of reduction, the CO-to-C02ratio of the off gas was found to be close to and slightly


greater than the wustite-iron equilibrium, at least during thefirst 50 per cent of the reduction, but increased considerablyduring the last 20 per cent. This slowing down of thereduction rate is usually attributed to the sintering andrecrystallization of the newly formed metallic iron, whichoccurs readily at temperatures above about 650°C anddrastically decreases the porosity of the particles. Seth andROSS37showed that the porosity of hematite pellets sinteredat 1160°C and reduced in hydrogen at 800°C increased as thereduction proceeded up to 50 per cent, and then decreaseddrastically beyond that point. This also shows that thegasification of carbon is still rate-controlling, although thereduction rate of iron oxide decreases during the final stages.

In a similar manner, carbon can be gasified by watervapour, and iron oxides can be reduced by hydrogen producedby the gasification of carbon by water vapour. Again, becausethe gasification reaction is highly endothermic and requireshigh temperature to proceed at an appreciable rate, the factorsaffecting the rate of the reaction involving hydrogen are thesame as those for the gasification of carbon dioxide.

Of course, these laboratory experiments were conductedin isothermal reactors in which heat transfer was not a factor.Therefore, the translation of these results to an actual processwould have to be accompanied by proper accounting of theheat-transfer mechanism.

From the preceding discussions, it is seen that, in contrastto gas-solid reactions, not many kinetic studies have beenmade on solid-solid reduction. Elguindy and DavenportS7,Ra053, and Maru58 used equations analogous to thosedeveloped for non-porous gas-solid systems. Ra053 developeda model of his own, but his data fitted very well to the modelproposed by McKewan64 for rhe special case of Lip added asa catalyst. He considered the equilibrium relationship of theCO-to-C02 ratio in the reducing gas mixture with each of thethree oxides of iron: Fep3' Fe304' and FeO.

It can thus be seen that models developed for non-porousgas-solid systems can also be used for the interpretation ofexperimental data relating to solid-solid systems reactingthrough gaseous intermediates.

Szekely et al.65, in their review of work done by variousworkers3. 7.66.67on the reduction of iron oxide by carbon,remark that, although the experimental results can beinterpreted by the use of Jander's or Ginstling Brownshtein'sequation, such interpretations serve only as indirect proofthat the overall rate is controlled by the reaction of CO2withcarbon. It should be noted that these models are somewhatinsensitive at a Iow degree of solid conversion and, in most ofthe studies, good fit is obtained only for Iow values of theconversion. More conclusive proofs arc required beforeinvestigators will be able to pin-point the exact mechanisminvolved. Hence, further theoretical and experimental work isrequired in this area for a proper understanding of the system.

Kinetically, a solid-solid reaction is expected to be muchslower than a gas-solid onc for the following two reasons:

(1) the solid-solid contact area is much smaller than thegas-solid contact area

(2) solid-state diffusion is much slower than mass transferin gases.

An attempt has been made by several investigators68-71to measure the rates of true direct reduction, viz thesolid-solid reduction between iron oxide and carbon, by

JANUARY /FEBF' . 9 ....


continuous evacuation of the chamber in the overalltemperature range 700 to 1150°C at pressures of 5 x 10-4 to10-2 torr. The underlying assumption in their investigationswas that the gas phase is substantially eliminated by theevacuation, and hence the rates obtained represent those ofsolid-solid reaction. It can be concluded from these resultsthat the rates under inert gas flushing are even smaller thansome of those obtained under evacuated conditions.

There is no experimental evidence to support the theoreticalexpectation that the rate of solid-solid reaction should bemuch slower than that of gas-solid reaction. However,although the chamber was evacuated in these experiments,the pressure inside the fine pores of the pellets was likely tohave been appreciable because the flow of COand CO2fromthe interior of the pellet to the chamber would require asubstantial pressure difference. It was found that, in order toallow the flushing of COand CO2out of the pellet, the pressureinside the pellet should be appreciable (10 mm Hg). If that isso, the gas phase cannot be ignored in such experiments68-71.

Reduction and sintering mechanism

The production of liquid iron outside the blastfurnace hasattracted considerable attention in the past ten to fifteen years.The development of smelting-reduction processes is aninnovative approach as an alternative route to steelmaking.However, these processes require very selective grades of rawmaterials: both the iron ore and the coal have to be fed in theform oflumps of selected size. The fines of ore, blue dust, orcoal produced during mining or crushing cannot generally beutilized in these processes. Some of the new smelting-reductionprocesses use a fluidized bed for the pre-reduction of ironore. Although fluidized beds can use iron-ore fines directly(without their agglomeration), sticking problems occur duringthe reduction. This unintentional agglomeration is the reasonfor numerous breakdowns in fluidized beds.

Fluxed composite reduced iron pellets

The concept of fluxed composite iron pellets requires thatwell-balanced amounts of fluxes are added to the mixture ofiron-ore fines and coal, and that, on being reduced by a solidor gaseous reductant, the mixture produces well-sinteredfluxed composite reduced iron pellets72. During melting orsmelting, such pellets will generate their own slag andcarbon. This is intended to reduce the residual oxygen of theiron oxide in the molten slag while the slag remains inchemical balance with the furnace lining, giving the necessaryassimilation of impurities like S, P, V,73.74as well.

However, one of the most imp~rtant properties of suchpellets is their strength during their transportation, reduction,and smelting to liquid iron. This depends largely on thecomposition and quality of the gangue in the ore and on theadditives. The development of pellets began with acid pelletsas early as 194074, and continued until pellets of neutral andbasic composition were gradually introduced in the 1970s74. 75.The basicity of these pellets has been reported in the range1.0 to 1.5. Edstrom3. 68, in his studies, found a notableimprovement in reduction properties when lime was added tomagnetite concentrate with a low content of gangue-D.5 percent (Si02 + Al2O3)' These improvements were attributedmainly to the formation of a binding phase consisting ofcalcium diferrite, Ca0-2Fe203'


Based on investigations69-77 into the gaseous reduction ofiron oxide pellets, it was concluded that the rate-determiningmechanism can be described as a combination of interparticlegaseous diffusion through the reduced layer and chemicalreaction at the interface between the reduced and the unreducedlayers. Several other investigators78-84 also support this view.The extent of reduction and its structural changes dependgreatly on various operating variables69. 77.79,80. The mostcritical of these is gaseous diffusion through the reducedlayer, which depends on the porosity of the pellets. Abrahamand Ghosh50 predicted a significant resistance to mass transferin the oxide. The mass-transfer limitations modify the ratebut do not influence the activation energy significantly. Thus,a high activation energy exists in spite of partial control bymass transfer. It further implies that, if composite pellets ofore and reductant can be used, the rate of the process can beenhanced by at least an order of magnitude. Ra053 found, inan investigation of the reduction kinetics of mixtures ofhematite and carbon powders in the temperature range 850 to1087°C, that the reduction process is characterized in terms ofthe availability of carbon monoxide. A solution-loss reactionoccurs between the carbon particles and the carbon dioxidewithin the mixture, and this governs the overall reductionprocess.

Although several investigators have attempted tocharacterize the reaction mechanism and rate-controllingsteps in terms of mathematical expressions, there is nowmuch experimental evidence based on microscopic studies tosuggest that temperature, carbon particle size, hematite-to-carbon ratio in the mixture, additions of catalysts, binders,fluxes, etc. all affect the reaction kinetics. It is also almostcertain that, at temperatures of 900 to 1000°C, the overallrate of reduction of iron oxide by carbon is influencedsignificantly by the rate of gasification of the carbon bycarbon dioxide.

With a more sophisticated approach, investigations on thereduction and sintering behaviour of fluxed ore and coal havebeen carried out to test the influence of CaO and Si02' thesintering temperature, and the sintering period on changes inthe iron morphology. Qi et al.82reported a chronologicalobservation by scanning electron microscope (SEM) of thereduction of iron ore. The reduction temperature was variedbetween 600 and 1000°C, and different gas compositionswere used. It was found that every precipitate could beassigned to a main area in the equilibrium diagram (Figure 7).The chemical composition had a strong influence on theprecipitation morphology of the iron82. Prior to the reductionof the material, the iron oxide was sintered at differenttemperatures and periods in an atmosphere of oxygen. Qiet al.82 observed that the precipitation of metallic iron tookless time than that of pure hematite. Consequently, the Caocan be referred to as an active compound. The results of theinvestigation are illustrated in Figure 8.

According to Philips and Muan's phase diagram83,calcium diferrite (CaO, 2Fe203 = CaFe407) is stable between1155 and 1126°C in air at 1 atm., as shown in Figure 9.Bergman84 recently found that calcium diferrite (CF2)wasformed at about 140°C at 1 atm. in air. This apparentdeviation can be attributed to the local oxygen potential atwhich it was formed, which differed from that of the ambientatmosphere.


100

90

CO 80CO + CO2

in "I. 70

60600

90

CO

CO + (02

in "I. 75

100

90

CO

CO + CO275

in%


-'-'-POROU

Fe

FeD

700 800 900Temperature °C

1000

Figure 7-Main regions of the different precipitation morphology of ironduring the reduction of iron ore82

100

60I

Fe FeD

700 800Temperature °C

950600

Figure 8-Changes in the precipitation of iron at different CaO contentsin the five zones of gas composition and temperature used in the study82

1600

Ho..ss

.'go.SS,

L~

L~

1400

2[10.FI,o. ,Iiq

1216'1200 [,0, [1011,0., L~

2[00.F.,o.[oOF.,o., [00.2Fo,o. HEM. . IIMA TITE

HAIIN. . HAIlNETITESS . SOLII SOLUTION1000

2[00.F.,o,'[101.,0. [O(!F.,o,,

"'"

0[00

602[_,0.

CURVE A-A' SHOWSCOMPOSITION OF

j Llo.llDUS AT LlD.IIIIJS

20TEHPW TUllE.

...0

/

Figure 9- The Cao-Fe203 system (in air)83


60600 700 800 950

Temperature °C

Figure 1O-Changes in the precipitation of iron at different SiO2 contentsin the five zones of gas composition and temperature used in the study82

As in the case of CaD, SiOz cannot be reduced in thetemperature range 600 to 1000°C. In contrast to CaD, SiOzcan form only the chemical compound fayalite (2FeOSiOz),which also cannot be reduced by CO. Consequently, SiOz canbe referred to as a sluggish compound that hinders thereduction of iron oxide. The influences of SiO2and CaD onthe iron morphology are shown in Figure 10.

Under identical changes of temperature and CO-CO2reducing-gas composition, Qi et al.8z assigned differentzones, (i) to (v), to each precipitate, as shown in Figures 8and 10. These five zones comprise various combinations ofthe gas composition and temperature under which the iron-precipitation morphology was studied.

All the investigations on the influence of CaD and SiOzhave shown that different concentrations of foreign compoundsresult in a change in iron morphology under different reductionconditions. These changes are characterized by fibrous, porous,and dense morphology. Fibrous and dense morphologies arethe two extreme cases, and porous morphology is an inter-mediate stage between these two.

A schematic representation of the effects of differentfactors on surface morphology is shown in Figure 11.

One of the main problems during reduction in a shaftretort under unfavourable conditions is clustering. This canresult in uneven gas distribution and problems duringdischarging. The pellets therefore need to be tested for theirproperties such as porosity, green and compression strengthafter reduction (CSAR) , and stickingcharacteristics.However,the standard methods of test are rather rough simplificationsof the process steps and treatments that they are intended tosimulate. Thus, there is a danger that one may concentrate onoptimizing the behaviour of an ore product under test conditionsrather than its properties under actual process conditions.

In order to minimize smelting costs, it is essential that thesponge iron should be highly reduced, and to achieve thiswithout sacrificing productivity in the direct-reduction furnaceneeds an easily reducible burden. The limiting factor is usuallythe shape of the reduction curve towards the end of reduction.The time required for a reduction of 90 per cent under iso-thermal conditions can be minimized53 by the use of areductant in the fluxed pellets.

JANUARY/FEBRUNN 199(j 11 ...


PRECIPIT A TIOH MORPHOLOGY

DENSE POROUS FIBROUS

caO - CONTENT

[>

<]SlOz - CONTENT

CIO + Stoz - CONTENT

<]SINTERING TEMPERATURE

<]SINTERlNG PERIOD

<]Figure 11-Effects of different factors on surface morphology82

Results and discussions

Prakashet al.8s-87 recentlyconductedlaboratory studies onthe reduction and sintering behaviour of fluxed compositereduced iron pellets. The starting material was chemical gradeiron oxide (85 per cent FezO3' the remainder being made upof non-FeO at a particle size of -0.1 +0.03 mm), commercial-grade graphite powder (99.8 per cent Cat a particle size of-0.01 mm), lime powder (99.8 per cent Cao at a particle sizeof -300 mesh) and a binder. In all the experiments, variousproportions of these ingredients were used to make fluxedcomposite (iron oxide and solid reductant) reduced ironpellets. The COgenerated in situ served to reduce the pellets.

To maintain a reducing atmosphere inside the furnace,excess COwas passed through the bed (Figure 12). The pelletsamples were heated to pre-determined reduction temperaturesof 1073, 1173, 1273, and 1323 K for specific reduction times.

Experiments were also carried out on the effects of theFe-to-C ratio and the quantity of flux on the reduction charac-teristics. The plots of the fraction reacted (calculated fromweight-loss data) against time are shown in Figure 13 andwere subjected to isothermal kinetic analysis.

The fluxed composite reduced iron pellets were fracturedfor a study under the SEM of the changes in microstructureduring simultaneous reduction and sintering after knownreaction intervals. These results, illustrated in Figures 14 to16, show that, with an increase in time and temperature(Figure 14), the ferrite phase tends to grow on a substrate ofwustite, progressively producing bridges that consolidate andform aggregated particles. However, apparently in the initialstages, the pores opened up due to gas pressure inside themand were further increased by the reduction from wustite tometallic iron. As a result of this, there was a significantvariation in the pore-size distribution. The SEM photographsat the higher temperatures indicate the enhanced meltformation89-91, presumably CaO-FeO-SiOz' However, theliquid did not close the pores, and reduction continued. Thesephotographs show significant changes in morphology due tothe presence of CaO, and these must account for the changesin reducibility. These results corroborate the findings ofGudenau and Schiller9z.


@

Nz

16

1 Balance2 Ore sample3 Electric furnace4 Thermocouple5 Reducing-gas inlet6 Flow meter7 Drying unit8 CO2-absorption unit9 °2-absorption unit

10 Gas meter11 Mixing vessel12 Reaction tube13 Electric furnace14 Coke bed15 CO2 cylinder16 N2cylinder17 Stainless-steel cage18 Alumina balls

Figure 12-5chematic diagram of the isothermal setup86-68

60c~ 50u.~~ loO

~i301--u~ 20~

10

00 10 20 30 40

TIME, min

(al ISOTHERMAL fr - time plot

50

50 60

0

t 40.."",.,I .. 30::0I ..."" 20.....,.,,~ 10

0 10 20 30 40 50TIME.min

(bl CGB MODEL - g (f) - t PLOT (isothermal I

"10

....ii:'~:;;

- 0~:;;...,C'=I;

-:-'c~.. a -10:I~C

60

A:E = 173.0; 8:E = 184."; C:E. 180.5D D:E = 188.7 kJ/mol

~~A~

-27 8 9 10

TEMP. RECIPROCAL;1010 IT. K-1(C) : ARRHENIUS PLOT - DIFFERENTIAL AND INTEGRAL METHODS

Figure 13-Evaluation of isothermal kinetic parameters(size: -0.1 + 0.04 mm; FetC =2.9; %CaO = 2.0)



(a) Time: 5 min (c) Time: 20 min

(b) Time: 15 min (d) Time: 30 min

Figure 14-Changes in microstructure of the core part: variation in pore-size distribution, coercion of the structure with an increase in residence time(FetC: 2.9; CaO: 2%; temp: 1273 K)

Summary

On the basis of the studies reviewed here, it can be concludedthat, of the many possible parameters, the temperature andpotential of the reduction gas are generally considered to beof great importance to the kind of iron precipitation that occurs.Every kind of iron precipitation is found in certain areas ofthe Baur-Glassner diagram (Figure 7) under correspondingpotentials of temperature and reduction gas. The chemicalcomposition and presence of foreign substances (CaO, Si02,ete.) have a strong influence on the surface morphology ofthe reduced iron.

In the quantitative analysis of the reduction of iron ore,ore-coal pellets, fluxed ore pellets, and fluxed compositepellets, the following three types of approaches were used:empirical models, models based on kinetics, and modelsbased on morphological changes. The relative merits andlimitations of these models were discussed, and the role ofsome of these approaches in the simulation of the reductionof fluxed (ore-coal) composite pellets was illustrated. It was,however, noted that gaseous diffusion through the productlayer cannot be used alone as rate-limiting in the determina-


tion of the kinetic parameters: the mass-transfer processbetween the reducing gas and the iron oxide must also beconsidered. The reduction process can be characterized interms of the availability of carbon monoxide. Althoughelegant empirical models and kinetic equations havetheoretical significance and are indeed necessary forelucidating the reaction mechanism, they are sometimes notdirectly applicable in the kinetic study of fluxed ore-coalcomposite pellets. It is prudent to explore the possibility ofcorrelating the change in iron-precipation morphology withthe reactionmechanism.Such correlationsestablished forvariations of parameters within specific ranges maysignificantly ameliorate the reduction and strength charac-teristics of fluxed pellets. However, fluxed ore-coal peIletshave not been studied as widely as this.

Basically, smelting-reduction processes should completelyavoid ore agglomeration and cokemaking. The concepts ofthese processes are impaired93 broadly because of thearrangements for the pre-reduction of iron-ore fines and thehighly efficient post-combustion that occurs in the smeltingvessel. Conceptually, fluxed composite iron-ore pellets

JANUARY /FEBRI 13 <IllI gm


(a) Temperature: 1273 K

(b) Temperature: 1323 K

(Core) (c) Temperature: 1373 K

Figure 15--Microstructure of the fractured pellets, showing the effects of temperature (FetC: 2.9; CaD: 2%; time: 45 minI

(Periphery) (d) Temperature: 1373 K

utilizing ore and coal fines seem to be a better alternativecharge material for DR-SR processes. However, these pelletsshould have the requisite properties relating to reduction andcompressive strength after reduction.

Prakash et al.72,85obtained some very encouragingresults on the reduction and strength characteristics of fluxedcomposite reduced-iron pellets (FCRIP).A laboratory study isbeing undertaken on the sintering and reduction properties ofsuch pellets, and work is under way on the development of aprocess94 for the production of FCRIP.

Acknowledgements

The author thanks Professor P. Ramachandra Rao, Director ofthe National Metallurgical Laboratory, )amshedpur, for hisadvice and encouragement in the preparation of this review,for providing facilities for the experimental work, and foraccording permission to publish the paper.

Thanks are also due to Professor ).0. Edstrom, of theproduction Technology Mining and Steel Industry of RoyalInstitute of Technology, Stockholm, Sweden, and to ProfessorH.W. Gudenau of the Institute of Ferrous Metallurgy, AachenUniversity, Germany, for providing scarce and unique literatureon the subject.

~ 14 JANUARY (FEBRUARY 1 996

~-~--~-

~

References

1. VaN BOGDANDY,lo, and ENGELL,H.). The reduction qfiron ores. 3rd edn.

Berlin, Springer Verlag, 1971. Chap. 1, p. 44.

2. DARKEN,loS., and GURRY,R.W.I. Am. Chem. Soc., vo\. 67.1945. p. 1398.

3. EDSTROM,).0. Trans. lIS!, no. 170. 1953. p. 289.

4. RoTH, W.lo Acta CIYst., vo\. 13. 1960. p. 140.

5. KOFSTAD, P. High temperature oxidation qfmetals. New York, John WHey,

1966. Chap. 3, p. 74.

6. GILLOT, B., and FERRIOT, ).G.! Phys. Chem. Solids, vo\. 37.1976. p. 857.

7. Ross, H.D. Direct reduction iron technology and economics qfproduction

and uses. ElIiot, ).F. et al. (eds.). Iron and Steel Society. AIME, 1980. p. 29.

8. VaN BOGDANDY, lo, and ENGELL, H.). The reduction qfiron ores. 3rd edn.

Berlin, Springer Verlag, 1971. Chap. I, p. 41.

9. LEVIN, R.lo, and WAGNER, ).B. Jr. Trans. Metal!. Soc. AlME, no. 233.1965.

p.159.10. WARD, R.G.Introduction to physical chemistry qfiron and steelmaking.

London, Edward Arnold, 1962. Chap. I, p. 7.

11. R!CHARDSON,F.D., and )EFFES,).H.E.ISlf, no. 163. 1949. p. 397.

12. FuJu, C.T., and MEUSSNER,R.A. (1968). Trans. Metal/. Soc. AlME, no. 242.

1968. p. 1259.

13. GEIGER,G.H., and WAGNER,).B. )r.Ibid., no. 233.1965. p. 2092.

14. STRANGWAY,P.K., and Ross, H.D. Can. Metall. Qutly, vo\. 4.1965. p. 97.

15. KHALAFALLA,S.E., and WESTON,P.L. )r. Trans. Metal!. Soc. AlME, no. 239.

1967.p.1494.

16. EL KAsABGY,T., and Lu, W.K. Metall. Trans., vo\. liB. 1980. p. 409.

17. NISHIKAWA,Y., SAYAMA,S., DEDA,Y., and SUZUKl,Y. Trans. pS!, vo\. 23,

no. 8.1983. p. 639.

18. KUBASCHEWSKI,0., and HOPKlNS,B.E. Oxidation qfmetals and alloys.

London, Butterworths Scientific Publications, 1953. Chap.1, p. 46.



(a) Fe/C: 3, % Lime: 2

(b) Fe/C: 6, % Lime: 2

Figure 16-Effect of lime addition on Fe/C ratios (temp: 1273 K; time: 45 min.)

19. WAGNER,c., and ZIMENS,K.ActaChem.Scand., no. 547.1947. 1'.1.20. OHMI, M., USUI, T., and NAKAJIMA, K. Trans. ISIj, vo!. 22. 1982. pp. 30-38.21. KHALAFALLA, S.E., REIMERS,G.w., and BAIRD, M.J. Metal/. Trans., vo!. 5.

1974. 1'1'.1013-1018.22. SETH, B.B.lo, and Ross,H.U. Can. Metal/. Qut(y, vo!. 2. 1963. p. 15.23. TURKDOGAN,E.T., and VINTERS,J.V.lbid., vo!. 12, no.!. 1973. p. 9.24. FREDRIKSSON,H., and STEVENSSON,1. Scand. J. Metal/., vo!. 3. 1974.

1'1'.185-192.25. BUEFUSS,R.lo Trans. Metal/. Soc. AIM£, no. 247. 1970. p. 225.26. Lu, W.K. Scand. J. Metal/. vo!. 2, 1973. p. 169.27. NICOLLE,R., and RIST, A. Metal/. Trans., vo!. lOB. 1979. p. 429.28. FUWA, T., and BANYA, S. Trans. ISI/, vo!. 9. 1969. p. 137.29. Lu, W.K. Scand. J. Metal/., vo!. 2. 1974. p. 169.30. EL-GEASSY,AA, SHEHATA,K.A., and Ezz, S.Y.lron and Stee/lntemational

no. 55.1977. p. 329.31. BITSIANES,G., and )OSEPH,T.lo J. Metals, vo!. 7. 1955. p. 639.32. PRAKASH,S. Nonisothermal kinetics of iron ore reduction. Ph.D. thesis.

Chap. 2, pp. 9-123.33. VON BOGDANDY,loV., and ENGELL,H.). Loc. cit., chap. 3, p. 137.34. McKEwAN, W.M. Trans. Metal/. Soc. AIME, no. 218.1960. p. 2.35. SPITZER,R.H., MANNING,F.S., and PHILBROOK,W.O. Ibid., no. 236. p. 726.36. LEVENSPIEL,O. Chemical reaction engineen'ng. 2nd edn. New Delhi,

John WHey Eastern, 1972. Chap. 12, pp. 343-350.37. SETH, B.B.lo, and Ross, H.U. Trans. Metall. Soc. AIM£, no. 233. 1965.

1'.180.38. Lu, W.K. Ibid., no. 227. 1963. p. 203.39. EvANs, ).W" and Koo. c.H. Rate process qf extraction metal/urgy.

Sohn, H.Y., and Wadsworth, M.E. (eds.). New York, Plenum Press, 1979.Chap. 45, p. 259.


(c) % Lime: 2, Fe/C: 2.9

(d) % Lime: 8, Fe/C: 2.9

40. USUI, T., OHMI, M., HIRASHIM.I.S.. and 0,,",11 \, Y. Proceedings Qllht' 6thProcess Technology CorJferen(('. Tile IroIl and Steel Society of :\IME, 1986.pp. 545-553.

41. TURKDOGAN, E.T., and VINTERS. I.V JIetal! lians., vo!. 3. Jun. 1972.p. 1561.

42. TOWHIDI,N., and SZEKELY,J./il1flll/ilking ,md Steelmaking,vo!. 6. 1981.1'.237.

43. EL-GEASSY,AA, SHEHATA,KA. dIld El!, S.Y.TransjlSL VD!.17. 1977.1'.629.

44. RAo,Y.K., and AL-KAHTANY,\\.M. Ironmilking and Steelmaking vo!. 11.

1984. p. 34.45. HAYASHI,S., IGUCHI,Y., and Hili le'. I. Trans. fSIj, VD!. 24. 1984.

1'1'.143-146.46. MOUKASSI,M., STEINMETZ,P., DII'IU . B., alld GLEITZER,C.Metal!. hans.,

vo!.14B.Mar. 1983.1'1'. 125 131.47. BALDWIN, B.G.joumaIISL no. 179. 1955. p. 30.48. OSTUKA,K.J., and KUNII,D. J. Cllt'lI/. Engng lapan, vo!. 2, no. 1. 1969.

1'.46.49. SZENDREI, T., and VANBERGE. F.e. rhenn,'cllim. Acta, 110.44, 1981 p. 11.50. ABRAHAM,M.c., and GHOSH,A. fl(1nmaki/(g and Stee/making, va!. I. 1979.

1'.14.51. SRINIVASAN,N.5., and LAHIRI,,\K. Metall hans. B, 110.83. 1977. P 175,52, GHOSH,P.c., and TEWARI,s.r\. If Sf. no. 208. 1970. p. 255.

53, RAo, Y.K. Metal/. Trans., vo!. 2. I CJ71.p. 1439.

54, SEATON, C.E., FOSTER, ).S., and VII. \SCO,It]. frans. jlSL vo!. 23. 1983.

1'.490.55, TURKDOGAN, E,T., and VINTERS. V.I. Metall hans., VO!.2. 1971. p. 1375.

56, FRUEHAN, R.). Metall. Trans. 8. !W. 83. 1977. p. 279.

57. ELGUINDY, M.I., and DAvENPoln, W.G. Me/ill! Trans., vo!. 1. 1970. p. 1729.

58. MARu, Y.lbid., vo!. 4. 1973. p. 2591.

JANUARY /FEBI'

15 ...19[',


59. EKETROP, S. Altemative routes to steel. London, Iron and Steel Institute,

1971.p.17.60. NARlTA,K., and KANEKO,D. SEAlSI QuarterlY, 1979. pp. 19-31.

61. I'RAKASH,S. NML Report No. 07488083,1989. Production ifhigh quality

steel using RESINESS process. Prakash, S., and Banerjee, S. (eds.). Chap. 3,

p.38.62. PRAKASH,S., and BANERJEE,S. (1988). Improved process for making high

quality steel directly from fine particles of iron rich materials and non-

coking coal fines. Indian Patent, no. 908/DeI/88. App!. Dec. 1987, ace.

lun.1988.63. DAS, S.K., I'RAKASH,S., and Banerjee, S. Proceedings International

CoTJference on Advances in Chemical Metallurgy. (ICCM-91) Bombay,

Bhabha Atomic Research Centre, 1993. pp. 681-695.

64. McKEwAN, M.M. Trans. Metall. Soc. AIM£, no. 222. 1958. p. 791.

65. SZEKELY,I., EVANS,I.W., and SoHN, H.Y. Gas-solid reaction. New York,

Academic Press, 1976. Chap. 4, p. 176.

66. ABRAHAM,M.C. Investigations on kinetics of reduction of ferric oxides with

graphite. Indian Institute of Technology (Kanpur), Ph.D. thesis, 1976.

chap.2.67. EOSTROM,1.0.lemkontorets Annaler, no. 140. 1956. pp. 101-115.

68. EOSTROM,1.0. Ibid., no. 142. 1958. p. 401.

69. MCKEWAN,W.M. Trans. Metall. Soc. AIM£, no. 224.1962. p. 2.

70. HARA, Y., and SHIN-ICHI-KoNDO.Trans. lIS!, no. 8. 1968. p. 300.

71. I'RocrOR, M.I., HAWKINS,RI., and SMITH,I.D.Ironmaking and Steelmaking.

vo!. 19, no. 3.1992. p. 194.

72. I'RAKASH, S., GOSWAMI, M.C., and MAHAPATRA, A.K.S. Proceedings if theFirst International Congress on Science and Technolcgy if Ironmaking.

ICSTI-94. Tokuda, M. (ed.). ISII Publication, 1994. pp. 110-115.

73. I'RAKASH,S. On the improvement of test methods for the assessment of raw

materials for ironmaking. Malayasia, South East Asia Iron and Steel

Institute, 1995. p. 62.

74. EOSTROM,1.0.lemkontorets Annaler, no. 141. 1957. p. 809.

75. WRlGHT,I.K. Trans. lIS!, vo!. 17. 1977. p. 726.

76. TANIGUCHI,S., OiIMJ,M., and HUKUHARA,H. Ibid., vo!. 20, no. 11. 1980. p. 753.

77. PANIGRAHY,S.C., IENA,B.C., and RIGAURO,M. Metal/. Trans. B, vo!. 21.

1990. p. 463.

78. ICHIRO,S., SHOII,S., KuNIHIKO,T., and NOBUHIRO,H.JlSIInternationa£

vo!. 30, no. 3.1990. p. 199.

79. SHARMA,T.Ibid., vo!. 34, no. 12. 1994. pp. 960-963.

80. SRlVASTAVA,G.C., and SHARMA,T. Trans. Indian Institute if Metals, vo!. 41,

no. 6.1988. p. 567.

81. SHARMA,T.Ironmaking and Steelmaking. vo!. 19, no. 5. 1992. p. 372.

82. QI, Y., GUOENAU,H.W., and LI, V. Proceedings if the First InternationalCongress on Science and Technology if Ironmaking. ICSTl-94. Tokuda, M.

(ed.). ISII Publication, 1994. pp. 110-115.

83. Philips, B., and Muan, A.J. Am. Ceram. Soc., vo!. 41, no. 11. 1958. p.

445.84. BERGMAN,B.Ibid., vo!. 69, no. 8. 1986. p. 608.

85. I'RAKASH,S., BASU, P., and RAY, H.S. Proceedings if the 3rd InternationalSymposium on Bendiciation and Agglomeration. ISBA-91. Indian Institute

of Metals Publication, 1991. p. 476.

86. I'RAKASH,S., and RAY, H.S. IISIInternationa£ vo!. 30, no. 3. 1990. p. 185.

87. PRAKASH,S. On the dynamic and non-isothermal characterisation of

ironmaking raw materials. Scand. f Metall., in press, 1995.

88. I'RAKASH,S. Ironmaking and Steelmaking. vo!. 21, no. 3. 1994.pp. 237-243.

89. HUFFMAN,G.P., HUGGINS,F.E., and DUNMYRE,G.R Fuel, vo!. 60, no. 7. 1981.

p.585.90. HAMILTON,I.D.G. Trans. Intstn Min. Metall. C. vo!. 85, no. 3. 1976.

p. C-30.91. NEKRASOV,Z.\., DROZOOV,G.M., SHMELEV,Y.5., BOLOENKO,M.G., and

DANKO,Y.G. Steel in USSR, vo!. 8, no. 8. 1978. p. 429.

92. GUOENAU,H.W., and SCHILLER,M. Proceedings European Ironmaking

Congress, Dusseldorf3. PII/4. 1986.

93. EOSTROM,1.0., and VONSCHEELE,I. The balanced oxygen blastftmace(BOBF) compared with other process altemativesjor hot metal produc-

tion. Royal Institute of Technology, Department of Production Technology

Mining and Steel Industry, Stockholm, Report no. 1992: 14. p. 1.

94. PRAKASH,S., and GOSWAMl,M.C. A process for making fluxed composite

reduced iron pellets from fines of iron rich materials and carbonaceous

materials. Indian Patent, no. 1235/DeU93. App!. Nov. 1993, ace. lan. 1994,app!.lun. 1993. .

~ 16 JANUARY IFEBRUARY 1996 The Joumal of The South African Institute of Mining and Metallurgy

reduction and sintering of ouxed iron ore pellets-a comprehensive review

Documents