3rd AMIREG International Conference (2009): Assessing the Footprint of 459

Resource Utilization and Hazardous Waste Management, Athens, Greece

Roasting reduction kinetics of an Indonesian nickeliferous laterite ore

E. Zevgolis, C. Zografidis, I. Halikia and M. Perraki

School of Mining and Metallurgical Engineering, National Technical University of Athens, Greece

ABSTRACT try even higher.

Investigation of the kinetics of a complex
In this work, the kinetics of roasting reduction
process like roasting reduction, is of noteworthy
of an Indonesian nickeliferous laterite ore, both
industrial importance, since understanding of
in the form of ore and pellets, with a gaseous
the rate controlling mechanisms is the funda-
reducing mixture -CO:N2-, is examined. It is
mental knowledge required for achieving dy-
deduced that reduction degree of the ore, in-
namic simulation or for identifying the inhibitor
creases within the first 20 and 60 minutes of re-
parameters of the same process. Moreover,
duction in the case of ore and pellets respec-
study of the process kinetics can contribute to
tively, and then it tends to equilibrium, for all
the optimization of the operation of a reactor
temperatures examined. Maximum reduction
and knowledge of kinetic parameters, such as
degree achieved was approximately 45% at
apparent activation energy or rate coefficients,
9000C for both ore and pellets. The diffusion
can be used as a useful tool in terms of a metal-
controlled mechanism clearly prevails for the
lurgical reactor design.
reduction of ore, as verified by the low activa-
There is an agreement between most of the
tion energy values, the study of selected reduced
researchers, that reduction of iron oxide in the
sample by SEM-EDS and the negligible effect
higher oxidation state, usually hematite (Fe2O3),
of temperature on the progress of the reduction.
to iron (Feo) is a multi-stage procedure:
On the contrary, the procedure is chemically
controlled up to the first 20 minutes regarding Fe2O3→Fe3O4(magnetite)→Fe(1-x)O(wuestite)→Feo
the reduction of pellets. After this time, diffu- The reported values of the activation energies
sion controlled mechanism prevails, till equilib- (Ea) for hematite reduction with CO, as well as
rium is attained. The predominant effect of the the suggested rate controlling mechanisms vary.
specific surface area and porosity of the laterite The range of the Ea values (Mondala et al.,
ore on its reducibility is verified. 2004/ 9.97-14.65 KJ/mol; Moon and Rhee,
1997-1998/ 14.6 - 19.8 KJ/mol; Pineau et al.,
1. INTRODUCTION 2006/ 72.3 KJ/mol) cannot provide clear evi-
dence about the controlling mechanism of iron
The objective of the current work is to investi- oxide. The same apply to the kinetics of iron ore
gate the kinetics of the roasting reduction of an reduction with CO (Swatantra et al., 1987/<
Indonesian nickeliferous laterite ore with carbon 80 KJ/mol; Hughes et al., 1982/< 40 KJ/mol).
monoxide, in order to determine the controlling Thus, it is deduced that the rate controlling
mechanism of the reductive procedure. Indone- mechanism of iron oxide reduction, either in the
sian laterite ore has been part of the metallurgi- form of pure hematite or as a constituent of
cal mixture fed into the rotary kilns of the Greek ores with various mineralogical and physical
ferronickel industry (Zevgolis et al., 2006a). Its properties, is not just mass transfer or surface
high content in nickel -about 2% on a dry basis-, chemical reaction, but it probably occurs in a
renders its importance for the ferronickel indus- succession of mechanisms along the different
3rd AMIREG International Conference (2009): Assessing the Footprint of 460
Resource Utilization and Hazardous Waste Management, Athens, Greece

stages of the reductive process. This is a strong - continuous weighing of the test sample with
indication that the controlling mechanism is a weighing device having a resolution of 1 g,
strongly related to the mineralogical composi- - cooling to room temperature, at a low N2
tion of the starting raw material, the applied flow rate.
temperature range and the reducing gas compo- The degree of reduction after time t, Rt, rela-
sition. tive to the ferric iron was calculated according
to the requirements of the ASTM standard test
2. EXPERIMENTAL method by the following general equation:
Chemical analysis of the bulk ore sample is Rt = (mt/Oxh)*100 (1)
given in Table 1. In the same Table, the chemi- where:
cal analysis of iron and nickel of the -12.5+9.5 mt: is the mass loss, in grams of the test
mm and -9.5+6.3 mm fractions used in the ex- sample after reduction time t.
perimental procedure is also given.
As revealed by the XRD patterns (obtained Oxh is the hypothetical oxygen content of test
by a Siemens D-5000 diffractometer, Ni-filtered sample, assuming that the iron and nickel oxides
CuKa radiation /λ=1.5405 Å) the ore mainly present in the laterite sample are Fe2O3 and
consists of goethite, (a-FeOOH), nickeliferous NiO, respectively.
serpentine [(Mg,Fe,Ni)6Si4O12(OH)6,], which is The experimental data obtained concerning
the main nickel-bearing mineral in this particu- iron oxide reduction, were applied to the follow-
lar type of ore, maghemite (γ-Fe2O3) and quartz ing mathematical models:
(SiO2). It is noted that its mineralogical and che- i) Chemically controlled mechanism:
mical character mainly approaches the interme-
diate type of laterites (Zevgolis et al., 2009).
A thorough description of both the experi-
⎛ ri d i
⎜
⎜C −C
⎞
[ ]
⎟ 1−(1 − R )1 / 3 = k1t
⎟
(2)
mental set-up, as well as the preparation of the ⎝ o eq ⎠
test samples and the step-by-step conduction of where:
the experiments, has already been discussed ri, di: are the initial radius and density of the
(Zevgolis et al., 2009). The main steps of the grain respectively,
experimental procedure in brief, involved: R: the reduction degree,
- preheating - heating up to the desired tem- k1: the rate constant and t the time of reac-
perature and calcination till all the volatile tion,
matter (H2O and CO2) of the ore is expelled- Co, Ceq: are the fluid reactant concentration at
of the test sample (250 g) at a specified size the external surface of the grain and at
range (-12.5+9.5 mm) and (-9.5+6.3 mm), the surface of the core respectively.
- isothermal reduction with a gaseous reducing
mixture consisting of CO and N2 40/60 % by ii) Diffusion through the product layer:
volume,
⎛ ri 2 d i ⎞ ⎡ 1 R (1 − R )2 / 3 ⎤
Table 1: Chemical analysis of laterite Ore ⎜ ⎟⎢ − − ⎥ = k 2t (3)
⎜ Co − Ce ⎟ 2 3 2
Ore I
Ore I Ore I ⎝ q ⎠⎣ ⎦
Component (-12.5+9.5 (-9.5+6.3
(Bulk)
mm) mm) iii) Mixed control reaction, that is a combina-
Fe2O3 31.84 31.94 29.09 tion of equations (2) and (3), based on the
NiO 2.56 2.61 2.67 additivity of reaction times:
SiO2 36.24
⎡1 R ⎛1− R⎞2/3⎤
CaO traces
[ ]
2
rd rd
( i i ) 1−(1− R) +( i i ) ⋅⎢ − −⎜ ⎟ ⎥=t (4)
1/3
MgO 11.69 k1(Co −Ceq) k2(Co −Ceq) ⎣⎢2 3 ⎝ 2 ⎠ ⎦⎥
Al2O3 3.74
Cr2O3 1.38 It is noted that the flow rate of the gaseous
Mn3O4 0.39 reducing mixture CO/N2 -10 l/min- is in great
L.O.I. 11.37 excess relative to the stoichiometrically required
3rd AMIREG International Conference (2009): Assessing the Footprint of 461
Resource Utilization and Hazardous Waste Management, Athens, Greece

50 50
45
(a) (b)
45
40 40

Reduction Degree (R %)
Reduction Degree (R %)

35 35

30 30

25 750 25 750
800 20 800
20
900 900
15
15
10
10 Ore grain size: -9.5+6.3 mm
Ore grain size: -12.5+9.5 mm 5
5 CO/N2 : 40/60% by volume
CO/N2 : 40/60% by volume
0
0 0 20 40 60 80 100
0 20 40 60 80 100
Time (min)
Time (min)

Figure 1: Reduction Degree as a function of time of the: (a) (-12.5 + 9.5) mm, (b) (-9.5 + 6.3) mm fraction of the ore.
amount of CO for the total transformation of the The reduction rate of Indonesian laterite is
ferric iron to metallic iron. Thus, we manage in significantly higher compared with the limonitic
this way to eliminate the effect of external mass type Greek nickeliferous laterites, which consti-
transfer on the process kinetics. Moreover, al- tute 80-90% of the laterite feed of the Rotary
though the effect of heat transfer was not inves- kilns in the Greek ferronickel industry (Zevgolis
tigated within the framework of the current et al., 2009). This can be attributed to the con-
work, it cannot be excluded as rate controlling siderably higher specific surface area of the Ore,
mechanism, since as reported, either thermal ra- determined by using the single- point Brunauer -
diation for higher temperatures (>1000oC) (Ray, Emmett - Teller (BET) method to be 162.6
1993), or conduction of heat in porous solids is m2/g, compared to the respective values fluctu-
a matter for investigation (Sun and Lu, 1999; ating among 15.5-26.5 m2/g for the same frac-
Huang, and Lu, 1993), can have a critical effect tion of the limonitic type Greek laterites. The
on reduction kinetics. initial reduction rate of the examined Ore is
comparable with that of the intermediate type of
Greek nickeliferous laterite of Kastoria origin,
3. RESULTS AND DISCUSSION in which goethite is also the main iron carrier
The results of the roasting reduction of the -12.5 mineral. Moreover, calcination prior to the re-
+ 9.5 and -9.5 + 6.3 mm fractions of the ore, ex- duction step, causes a decrease in the specific
pressed as reduction degree versus time, are surface area of the sample after goethite dehy-
presented in Figure 1. The experiments were droxylation at approximately 400 oC and proba-
carried out at the temperatures of 750, 800 and bly a coalescence of the narrow pores of the raw
900 oC. ore, due to high temperature treatment. This re-
It is apparent from Figure 1 that reduction sults in an appreciable increase in total porosity
degree increases within the first 20 minutes of and the production of larger pores, which corre-
the process, and then reaction tends to equilib- spond to easier diffusitivity of the gaseous re-
rium, for all temperatures examined. From the ducing agent.
experimental results it was observed that maxi- The progress of iron metallization regarding
mum reduction degree achieved was 44% for the -9.5+6.3 mm reduced fractions after 90 min-
the -9.5+6.3 mm fraction. It is also noted that utes reduction, was determined by wet chemical
under the examined experimental conditions, analysis (ASTM E277, 1988; Xu et al., 2003),
temperature does not play an important role in as the ratio Femetal : Fetotal. The degree of metal-
the progress of the reduction, since the final re- lization was: 10.11%, 10.44%, and 15.45% at
duction degree obtained was fluctuating be- 750, 800 and 900 oC respectively. The above
tween 32.5 and 44%, in the temperature range values verify the reduction degree values calcu-
examined. Moreover, it is deduced by Figure 1 lated by equation (1) according to the weight
that decrease of the ore grain size favors the loss technique, taking into consideration that a
progress of the reduction. reduction degree of 33.3% corresponds to the
3rd AMIREG International Conference (2009): Assessing the Footprint of 462
Resource Utilization and Hazardous Waste Management, Athens, Greece

total transformation of hematite to magnetite min, since after 20 minutes the reduction proc-
and the start point for metallic iron formation. ess practically tends to equilibrium.
The mineralogical characterization of the - It is deduced that either the diffusion or the
12.5+9.5 mm reduced laterite fractions at the mixed control kinetic mechanism -i.e. equations
temperature of 900 ºC, verified the presence of (3) or (4)- seem to prevail. It should be noted
metallic iron (a-Fe) and magnetite (Fe3O4), as though, that since equation (4) comes from the
the main iron-bearing mineral phases. Quartz, combination of equations (2) and (3) and it is
olivine [(MgFe)2SiO4] and pyroxene not a mathematical model based on the intro-
[(Mg,Fe)SiO3] were also indentified. duction of new physicochemical parameters of
For the kinetic analysis of the experimental
the process, cannot give a clear picture whether
results, the kinetic model equations (2) - (4)
have been applied to the experimental data for diffusion or surface chemical reaction phenom-
the temperatures of 750, 800 and 900 oC. The ena constitute the controlling mechanism. The
kinetic model equations that best fit the experi- conclusions, are in agreement with those drawn
mental data for the two examined fractions are by former reducibility kinetic studies of Greek
presented in Figures 2 and 3, based on the ap- nickeliferous laterites with solid reducing agents
plication of the least square method for the line- (solid fuels) (Zevgolis et al., 2006b; Halikia
arity assessment of the kinetic equation dia- et al., 1998) and gaseous reducing agent - (mix-
grams. The kinetic analysis of the experimental ture of CO - N2).
data was conducted for the time period 2-20 Within the framework of further assessment
of the reduction kinetic data, a methodology of
0.14 work was additionally used, for approaching the
0.12 2
rate controlling step of the process within the
R = 0.9493
2
examined temperature range, based on the ap-
1-(1-R) +1/2-R/3-[(1-R) ]/2

R = 0.9795 2
2/3

R = 0.9941
0.10
plication of the diagnostic equation to the ex-
0.08
750 perimental data:
800
0.06 900 ln[− ln(1 − R )] = n ln t + ln b (5)
1/3

0.04
where:
0.02
Ore grain size: -12.5+9.5 mm
CO/N2 : 40/60% by volume
R: reduction degree (%) of iron oxides
calculated by equation (1)
0.00
0 5 10 15 20 25 t: time (sec)
Time (min) b: constant and
Figure 2: Kinetic models application to the experimental n: constant depending on the rate control-
data of the reduction of the (-12.5 + 9.5) mm fraction of ling mechanism and the geometrical
the ore at the temperature of 750, 800 and 900 ºC. characteristics of the ore.
0.0120
The obtained ‘n’ values from application of
2
R = 0.9985
the experimental data R - t concerning the (-
0.0100
9.5+6.3) mm fraction of the ore, which repre-
sent the slopes of the linear graphic representa-
1/2-R/3-[(1-R) ]/2

0.0080
tion of equation (5), are compared with the theo-
2
2/3

R = 0.9966

0.0060 800 retical values of the widespread used kinetic

900 equations, presented in Table 2. Equations D1 -
0.0040 750
D3 correspond to the diffusion rate controlling
step, and F1, R1 and R2 correspond to chemical
0.0020
2
R = 0.975 Ore grain size: -9.5+6.3 mm reaction mechanism
0.0000
CO/N2 : 40/60% by volume
The determination of the slope values, as
0 5 10 15 20 25 seen in Table 2, enhances the conclusion that
Time (min)
the diffusion mechanism prevails within the ex-
Figure 3: Kinetic models application to the experimental
data of the reduction of the (-9.5 + 6.3) mm fraction of the
amined temperature range, since they are very
ore at the temperatures of 750, 800 and 900 ºC. low and approach those applying to the diffu-
3rd AMIREG International Conference (2009): Assessing the Footprint of 463
Resource Utilization and Hazardous Waste Management, Athens, Greece

Table 2: Theoretical and experimentally obtained ‘n’ Val- increases with time. Moreover, it can be said
ues of the Kinetic Equations for Gas - Solid Reductive that the diffusional resistance after 20 minutes is
Reaction.
such that the reaction practically tends to equi-
Experimental
Theo- librium.
‘n’ Values
retical A metallographical section of grains of the
Kinetic Equation Grain size:
‘n’ reduced fraction -12.5+9.5 mm of the ore at
(-9.5+6.3)
Values
mm
800oC was studied by SEM/EDS (JEOL® JSM-
D1: (1-R) ln (1-R) + r = Kt 0.57
6380LV). A back-scattered electron image of a
D2: [1- (1-R)1/3]2 = Kt 0.54
partly reduced iron oxide grain, is presented in
750oC: 0.66 Figure 4. This is a typical feature of a reduced
D3: 1- (2/3)R – (1-R)2/3 = Kt 0.57
800oC: 0.49 ore grain when diffusion of the gaseous reduc-
F1: -ln(1-R) = Kt 1.0
900oC: 0.41 ing agent through the pores as well as through
R2: 1- (1-R)1/2 = Kt 1.11
R3: 1- (1-R)1/3 = Kt 1.07 the iron product layer controls the process. As
can been seen, an unreacted core of the iron ox-
ide -magnetite and probably wuestite- is pre-
sion controlling step (Equations D1-D3).The Ar- served; in the mantle area a metallic iron layer
rhenius activation energy values of the rate- formed, whilst at the outermost rim of the grain
determining step at various conditions have the iron oxide remained unreduced . This indi-
been evaluated by considering the initial rate cates that there is not just a certain reaction zone
values from the first two pairs of points in R - t but the reducing agent penetrated through the
diagrams, i.e reduction degree values obtained pores of the ore, as well as the cracks inside the
for t = 0 and t = 2 sec and calculating ”mean“ grain formed probably by the dehydroxylation
rate rmean as: of goethite and its transformation to porous
R − R0 hematite. The presence of pyroxene
rmean = (6) [(Mg,Fe)SiO3] detected by X-Ray analysis, was
t − t0
also verified by SEM/EDS; it probably acted as
where R0 and t0 correspond to zero recovery and an additional kinetic inhibitor for the diffusion
time respectively. Thus, the activation energy of the gaseous reducing agent.
for the reduction of the two fractions (- The same reducibility tests were applied to
12.5+9.5) and (-9.5+6.3 mm), was determined pellets of the ore (diameter of -9.5 + 6.3 mm),
to be 0.88 and 6.73 kcal/mole respectively, by for two temperature values (750 and 800 ºC).
Bentonite 1% by weight was used as a binding
the use of ”mean“ rate values, which are typical
agent. The kinetic curves of the process are pre-
values corresponding to the diffusion control-
sented in Figure 5. The kinetic model equations
ling step.
(2) - (4) that best fit the experimental data for
The conclusion deduced by the kinetic as-
sessment of the experimental data, regarding the
prevalence of diffusion as the rate controlling
mechanism, also justifies the negligible effect of
temperature on the progress of the reduction
within the temperature range examined. On the
contrary, in case that the effect of the surface
chemical reaction was predominant, increase of
temperature should enhance the reduction rate.
Moreover, the reduction rate is favoured by the
decrease of the ore grain size, since decreasing
of the ore particle size, results in an increase of
the free surface and concequently the total rate
of reduction. The kinetic curves (Fig. 1), show
clearly that the reaction rate declines with time,
something which indicates that the diffusion Figure 4: Back–scattered electron image of grains of the
path lengths and therefore diffusional resistance (-12.5+9.5) mm fraction of the ore reduced at 8000C.
3rd AMIREG International Conference (2009): Assessing the Footprint of 464
Resource Utilization and Hazardous Waste Management, Athens, Greece

50
of the reduction. This is enhanced by the kinetic
45
assessment of the experimental data, conducted
40 by the application of equations (2) - (4) (Fig. 6).
Thus, it is seen that for both temperatures, the
Reduction Degree (R %)

35
30 procedure is chemically controlled at the initial
25 stage up to 20 minutes, justifying in such a way
20
750
800
the positive effect of temperature. This is
15 probably attributed to the fact that the reductive
10 gas penetrates easier through the spherical po-
5
Pellet diameter (from the ore): -9.5+6.3 mm rous pellets contrary to the irregularly shaped
CO/N2 : 40/60% by volume
0 grains of the examined fraction of the bulk ore.
0 20 40 60 80 100 On the contrary, for the time range 20-60 min-
Time (min)
utes, till equilibrium is attained, the formation of
Figure 5: Reduction Degree as a function of time of the the iron layer probably renders the diffusion of
pellets of the Ore. CO as the controlling mechanism.

the reduction of pellets, based on the application

of the least square method, are presented in Fig- 4. CONCLUSIONS
ure 6. It is noted that the kinetic analysis for The current work constitutes an attempt to as-
both temperatures was conducted for two time sess the experimental results of the reduction
periods, 2-20 and 20-60 minutes, due to the fact roasting of an Indonesian nickeliferous laterite
that a change of the slope in the kinetic curves ore with a gaseous reducing agent -CO-N2-,
indicates a probable change of the rate control- from a kinetic point of view. The kinetic analy-
ling mechanism. sis of the isothermal reduction within the tem-
As deduced from Figure 5, the equilibrium is perature range 750-900°C, revealed that reduc-
attained after 60 minutes for the pellets, con- tion degree increases within the first 20 minutes
trary to the reduction of bulk ore, where equilib- of the process, and then reaction tends to equi-
rium was attained earlier, after 20 minutes for librium. The diffusion controlled mechanism
the same fraction (-9.5+6.3 mm). The maximum prevails, as verified by the negligible effect of
reduction degree achieved was approximately temperature on the progress of the reduction
45%, which is indicative of the fact that the fi- within the examined range, the application of
nal result after 90 minutes of reductive proce- the diffusion kinetic equations, the low activa-
dure is almost the same with that regarding the tion energy values for the two fractions -0.88
bulk ore. and 6.73 kcal/mole-, as well as the examination
Moreover, it is noted that increase of tem- of reduced laterite sample by SEM-EDS.
perature from 750 to 800oC favors the progress Application of the same reducibility test to
0.14 0.02
2
(a)/2-20 min 0.01 (b)/20-60 min R = 0.9371
0.12 2
R = 0.9951
0.01
0.10
1/2-R/3-[(1-R) /2

0.01
2/3

0.08
1/3

2
R = 0.9949
1-(1-R)

2
R = 0.9609 750 0.01
0.06 800 750
0.01 800

0.04
0.00

0.02 Pellet diameter (from the ore): -9.5+6.3 mm

Pellet diameter (from the ore): -9.5+6.3 mm 0.00
CO/N2 : 40/60% by volume
CO/N2 : 40/60% by volume
0.00 0.00
0 5 10 15 20 25 0 20 40 60 80
Time (min) Time (min)

Figure 6: Kinetic models application to the experimental data of the reduction of pellets from the ore (diameter: -9.5 +
6.3 mm), for the time range: (a) 2-20 min, (b): 20-60 min.
3rd AMIREG International Conference (2009): Assessing the Footprint of 465
Resource Utilization and Hazardous Waste Management, Athens, Greece

pellets of the same origin and the same diame- tuating Temperature Conditions, Thermochimica
ter, showed that the final reduction degree ob- Acta, 111, pp. 143-166.
Zevgolis, E.N., C. Zografidis, J. Gaitanos, I.-P. Kostika
tained after 90 minutes reduction, was almost and I. Halikia 2006a. Energy requirements in nicke-
the same with that of the bulk ore. liferous laterite treatmen. Paper presented at the EPD
The reductive procedure tends to equilibrium Congress, San Antonio, Texas, 12-16 March 2006, pp.
after 60 minutes and increase of temperature 487-496.
among 750-800°C favours the progress of the Zevgolis, E.N., Ch. Zografidis, I. Halikia and E. Devlin,
reduction. This is indicative of the fact that as 2009. Roasting reduction study of Greek nickeliferous
laterites, Paper presented at the 138th TMS Congress,
verified by the kinetic analysis of the experi- San Francisco, California.
mental data, the procedure is chemically con- Zevgolis, E.N., I. Halikia and I.-P. Kostika, 2006b. Re-
trolled up to the first 20 minutes. Later on, dif- ductive behavior of the recycled dust during nicke-
fusion becomes the rate controlling step, till liferous laterite treatment, Erzmetall - The World of
equilibrium is attained. Metallurgy, 59 (6), pp. 350-359.
Xu, Z., J. Hwang, R. Greenlund, X. Huang, J. Luo and S.
The significantly higher reducibility of po- Anschuetz, 2003. Quantitative Determination of Me-
rous Indonesian laterite, where iron is mainly tallic Iron Content in Steel-Making Slag, Journal of
present in the form of goethite, than that of the Minerals & Materials Characterization & Engineering,
Greek nickeliferous limonitic type laterites, 2(1), pp. 65-70.
verifies the predominant effect that the physical
parameters such as specific surface area and po-
rosity have on the progress of the reduction.

REFERENCES
ASTM Standard E277, 1988. Test Method for Total Iron
in Iron Ores by Stannous Chloride Reduction and Di-
chromate Titration.
Halikia, I., P. Neou-Syngouna and M. Katapotis, 1998.
Reductive Roasting of Iron-Nickel Ore Using Greek
Lignite Thermodynamic and Kinetic Approach, ed.
NTUA, In honor of Professor Emeritus of NTUA A.Z.
Fragiskos, Athens, (Greek Text).
Huang, B-H. and W.-K. Lu, 1993. Kinetics and Mecha-
nisms of Reactions in Iron Ore/Coal Composites, ISIJ
International, 33 (10), pp. 1055-1061.
Hughes, R., et al., 1982. The Reduction of Iron Ores by
Hydrogen and Carbon Monoxide and their Mixtures,
Thermochimica Acta, 59, pp. 361-377.
Mondala, K., H. Lorethovab, E. Hippoa, T. Wiltowskia,
and S.B. Lalvani, 2004. Reduction of iron oxide in
carbon monoxide atmosphere-reaction controlled ki-
netics, Fuel Processing Technology, 86, pp.33-47.
Moon, I.J. and C.H. Rhee, 1998. Reduction of hematite
compacts by H2-CO gas mixtures, Paper presented at
the TMS Annual Meeting, Orlando, 9-13 February,
pp. 302-306.
Pineaua, A., N. Kanaria and I. Gaballah, 2006. Kinetics of
reduction of iron oxides by H2. Part I: Low tempera-
ture reduction of hematite, Thermochimica acta, 447,
pp. 89-100.
Ray, H.S., 1993. Kinetics of metallurgical reactions, Ox-
ford & IBH Publishing CO.PVT.LTD.
Sun, S. and W.-K. Lu, 1999. A Theoritical Investigation
of Kinetics and Mechanisms of Iron Ore Reduction in
an Ore/Coal Composite, ISIJ International, 39 (2), pp.
123-129.
Swatantra Prakash and Hem Shanker Tay, 1987. Reduc-
tion of Iron Ore Under Rising Temperature and Fluc-