INTERNATIONAL JOURNAL OF ENERGY RESEARCH, VOL.

20,1103-1128 (1996)

DIRECT SMELTING AND ALTERNATIVE PROCESSES FOR

THE PRODUCTION OF IRON AND STEEL
T. ZERVAS, J. T. MCMULLAN AND B. C . WILLIAMS
Energy Research Centre, Uniuersiy of Ulster, Coleraine, Co. Londondeny, BT 52 1SA U.K.

SUMMARY
Processes for the direct smelting of iron and steel are presented together with a discussion of alternatives based on
plasmas and the direct production of iron carbide. The advantages and disadvantages of each technology are
presented and their energy efficiencies are discussed.

KEY WORDS: iron production; direct reduction; direct smelting; energy in steel production; iron carbide production; plasma
technology

1. INTRODUCTION
Over the past century, considerable research and development effort has been expended in developing
methods for producing iron that could substitute for pig iron from blast furnaces (Strassburger, 1969).
Much of this work was stimulated by a need to use lower-grade iron ores and non-coking coals that are
unsuitable for the coke oven unit in an integrated coke oven and blast furnace plant. This has led to
renewed interest in direct reduction and direct smelting processes. Direct reduction processes are those
which produce iron by the reduction of iron ore at temperatures below the melting point of the product;
the product is referred to as ‘direct reduced’ iron. Processes which produce a molten product similar to
blast-furnace hot metal directly from the ore are defined as ‘direct smelting’ processes. If the objective is
to produce liquid steel directly from the ore, the term ‘direct smelting’ processes. If the objective is to
produce liquid steel directly from the ore, the term ‘direct steelmaking’ is often used. These definitions
are clearly based on the characteristics of the product, though further treatment steps may be added to
produce special grades of steel.
As has been discussed by Zervas et al. (1996a, 1966b and 1966~1,the coke oven and blast furnace
integrated plant is expected to continue to be the world’s main source of raw material (pig iron) for
steelmaking production, as long as supplies last of suitable coke for the coke ovens, and are available at
a competitive price. Currently, direct reduction iron contributes a small portion of total world iron-mak-
ing capacity with about 30 million tonnes per year of direct reduction iron capacity having been installed
by June 1994 (Scott, 1994).
The direct reduction iron process has not been commercialized at the expected rate. In 1976 it was
estimated that direct reduction iron capacity would reach 45 million tonnes per year by 1985. Instead,
plant capacity had only reached 30 Mtonne per year by 1994, as mentioned above. The main reasons have
been the lower-than-expected demand for iron in the developing world and the consequent reductions in
plant capacity and utilization. The world-wide utilization of direct reduction iron plant has been about
60%.
Other research and development efforts in the reduction of iron ore have concentrated on the
development of the direct smelting of iron. The main effort here is to develop a molten iron-making
process that does not depend on metallurgical coal. Many processes are under development in the
U.S.A., Australia, Japan and Europe (Fruehan, 1993). In Austrialia, CRA Ltd and Midrex are cooperat-
ing in the development of HIsmelt process in which non-coking coal and pre-reduced iron ore are

CCC 0363-907X/96/121103-26 Received 11 December 1995

0 1996 by John Wiley & Sons, Ltd.
1104 T. ZERVAS, J. T. McMULLAN AND B. C. WILLIAMS

injected into a smelting reactor, and post-combustion of the off-gas with air is possible. In Japan, an
association of the steel industry and public organizations are collaborating in the direct iron ore smelting
(DIOS) process. In the U.S.A. the steel industry is responding to this challenge through a collaboration in
research and development for a process sponsored by the America1 Iron and Steel Institute (AISI) and
the U.S. Department of Energy (DOE) which involves a number of U.S.A. and Canadian steel firms and
academic research centres.
So far, only two technologies have come to full commerical application (see Table 1). These are the
Plasmadust variation of the Plasmasmelt process developed by SKF in Sweden, where it is used to process
70ktonne per year of allowed steelworks dust, and the Corex process which was developed in South
Africa by Korf Engineering and Voest Alpine with a capacity fo 300 ktonne of hot metal per year. The
ideal concept of a direct-smelting reduction process to convert iron ore and coal to steel has not yet been
realized.
The generalized iron and steel metallurgical complex is shown in Figure 1 and the underlying chemical
reactions are given in equations (1)-(12) (Davis et al., 1982; Lankford et al., 1985; Stephenson 1980).
3Fe,O3 + H, = 2Fe30 + H,O (1)
3Fe,03 + CO = 2Fe30, + CO, (2)
Fe,O, + H, = 3FeO + H,O (3)
Fe30, + CO = 3FeO + CO, (4)
FeO + H, = Fe + H,O (5)
FeO + CO = Fe + CO, (6)
3Fe + CO + H, + H,O
= Fe,C (7)
3Fe + 2CO = Fe,C + CO, (8)
CO, + C = 2CO (Boudouard reaction) (9)
H,O + C = CO + H , (10)
FeO + C = Fe + CO (11)
3Fe + C = Fe,C (12)
At temperatures below about 1000" C, the dominant reactions are given by equations (1)-(6). Metallic
iron will also absorb carbon according to equations (7) and (8). At temperatures above 1000"C, the
additional reaction of carbon, CO, and H,O begins to play a part as shown in equations (9) and (10). CO
and H, are produced, so increasing the reducing potential of the gas. The resultant net reaction of
equation (11) is important for processes that produce DRI directly from solid coal without prior
gasification of its fixed carbon. Above about 1200" C, metallic iron absorbs any carbon present, according
to equation (121, leading to fusing or melting of the solid even though the melting point of pure iron is
1530" C. Thus, 1200"C represents the watershed between direct reduction and direct smelting processes.
In practice, direct smeling processes operate above 1300"C to ensure that carbon is absorbed rapidly and
liquid hot metal is produced.
Regardless of the reduction mechanism as discussed above, a necessary precursor to any iron oxide
reduction process is the generation of a suitable reductant gas (H,,CO). This can be generated in a
one-step process within the reductant reactor or it can be produced externally in a two-step process
(Meyers, 1984). The main chemical reactions involved in gasification are as follows.
Cornbustion
C+O,= AH= -394kJmol-' (13)
 Table 1 . Direct reduction plants installed throughout the world (Steffen and Lungen, 1988).

Reduction Nominal Plant

agent capacity out of
Year No. of (G: gas; (ktonne service
Country Company and location Processes commissioned units K: coal) per year)

Egypt ANS, El Dikheila Midrex 1987 1 G 720

Argentina Acindar, Villa construction Midrex 1978/83 1 G 600
Dalmine Siderca, Campana Midrex 1976 1 G 330 5
rn
Acos Finos, Piratini n
z
Brazil Port0 Alegre SL/RN 1973 1 K 65 >
i Usiba, Salvador HYL 1974 1 G 325 =!
<
rn
Burma -
State owned company Kinglor- 1981 1 K 20 n
( Mandalay Metor 1984 1 K 20 P
\
Federal Republic Germany
( Hamburger Stahlwerke, Hamburg 1972
Midrex 1 G 400
Yrn
m
Great Britain r
W
British Steel, Hunterston Midrex 1979 2 G 800 X
8U
Sponge Iron India Ltd C
Paloncha SL/RN 1980/85 2 K 60
BSIL, Chandil SL/RN 1987 1 K 150 3
India Orissa Sponage Iron
Palasapunga Accar 1983 1 K 150
P
W
Ipitata, Joda TDR 1986 1 K 90 n
Sunflag, Bhandara Codir 1988 1 K 150
Essar, Hizira Midrex 1990 2 G 840 v1
F1rn
v1
Indonesia Krakatau Steel, Cilegon HyL/HyL I l l 1978/82 4 G 2240 c1
Iraq

Iran
I Sodco, Khor El Zubeir
Nisco, Ahwaz
Nisco, Ahwaz
HYL
Purofer
HYL
1979
1977
(?)
2
1
3
G
G
G
543
360
1100
X

Nisco, Ahwaz Midrex (?) 2 G 800

i Nisco, Mobarakeh Midrex (?) 5 G 3200
Sidbec Dosco,
Canada Contrecoeur 1 Midrex 1973 I G 400 X
Contrecoeur 2 Midrex 1977 1 G 600
 c
Table 1. Confiniced !e

Reduction Nominal Plant

agent capacity out of
Year No. of (G: gas; (ktonne service
Country Company and location Processes commissioned units K coal) per year)

Quatar Quatar Steel, Doha Midrex 1978 I G 400

Lybia State owned company, Misurata Midrex 1988 2 G 1 100
Malaysia Sabah Gas Ind. Labuan Midrex 1984 1 G 650
Trengganu Steel, Trengganu NSC 1985 1 G 600 X

Hylsa, Monterrey 1M 1957 1 G 100

9
HYL N
Hylsa, Monterrey 2M HyL 111 1960/79 1 G 250 rn
Hylsa, Monterrey 3M HyL 111 1974/82 1 G 550 icr
Mexico Hylsa, Puebla 1P 1969
<
HYL 1 G 335
Hylsa, Puebla 2P HYL 1977 1 G 700
-$
Sicartsa. Las Truchas HyL 111 1988 4 G 2000
Tamsa, Vera Cruz HYL 1967 1 G 280
New Zealand New Zealand Steel 1, Glenbrook SL/RN 1970 1 K 150
( New Zealand Steel 2, Glenbrook SL/RN 1986 4 K 700
Nigeria Delta Steel, Warri Midrex 1982/83 2 G 1000
Peru Siderperu, Chimbote SL/RN 1980 3 K 100
Saudi Arabia Saudi lron&Steel, Al Jubail Midrex 1982/83 2 G 800
Soviet Union OEMK, Kursk Midrex 1983/87 4 G 1650
Dunswart Steel, Benoni Codir 1973 1 K 150
Iscor, Vanderbijlpark SL/RN 1984/85 4 K 600
Iscor, Pretoria Corex 1988 1 K 300
South Africa Scaw Metals, Germiston DRC 1983 1 K 75
Union Steel Corp. Vereeniging Fluor/Huls 1985 1 G 250
Davsteel, Zondenvater Rotary Kiln 1985 1 K 35
Trinidad and Iscott, Point Lisas Midrex 1980/82 2 G 840
Tobago
USA. Georgetown Steel, Georgetown Midrex 1971 1 G 400
Minorca, Puerto Ordaz Midrex 1972/89 1 G 750
Fior de Venezuela, Matanzas Fior 1976 1 G 350
Sidor, Matanzas 1 HyL 1 1977 1 G 400
Venezuela Sidor, Matanzas 2 Midrex 1977 1 G 400
Sidor, Matanzas 3 HyL 1 1981/87 3 G 1920
Sidor, Matanzas 4 Midrex I979 3 G I275
Sivensa, Matanzas Midrex 1990 1 G 400
 ALTERNATIVE IRON / STEEL PRODUCTION PROCESSES 1107

I-

-- -- i " t
Fhul hot r

mBtdb&&C.
ow
coldrollhy,

C-hUO-
-
-g

Gasification
c + ;02= co A H = -lllkJmol-' (14)
C + H,O(g) = C O + H , AH = + 135.6 kJ mol- ' (15)
c +CO, = 2 c o AH = + 172.46 kJ mol- ' (9)
Shifr
C O + H,O(g) = CO, + H, A H = -41kJmol-' (16)
Methanation
C O + 3H, = C H , + H,O(g) A H = -206kJmol-' (17)
2CO + 2H2 = C H , + C O , A H = -247kJmol-' (18)
and the general process can be represented by
C,H, + ( n / 2 ) 0 , = nCO + ( m / 2 ) H 2 (19)
The combustion part of the process is exothermic, as shown by the heats of combustion, whereas most of
the gasification reactions are endothermic. The water gas shift reaction takes place in the gasifier and
helps to adjust the hydrogen-to-carbon monoxide ratio. It is also catalytically induced in a separate
methanation step to form SNG.

2. REACTORS
A reactor is a unit of equipment designed to carry out certain reaction processes. As the majority of the
reaction processes in metallurgy are heterogeneous, the main purpose of the reactor is to bring the
reacting phases together under favourable conditions of pressure and temperature, to supply the
necessary energy and to allow the separation of the product phases.
1108 T. ZERVAS, J. T. McMULLAN A N D B. C. WILLIAMS

Although the reactors are similar, there are differences and the reactors can be classified as
continuous or discontinuous, series or parallel, etc. (Rosenqvist, 1983).

2.1 Vertical shaft reactor

The vertical shaft reactor (Stephenson and Smaller, 1980) is a moving bed reactor in which iron ore
moves under gravity in counter current to a reductant gas stream. The shaft reactor has been always a
favoured reactor for the production of iron from iron ore and it has been commercially adopted as the
preferred reduction unit in a number of direct reduction processes.
As iron oxide requires higher concentrations of reductant to drive the reactions, a counter current
reactor is paticularly suited to the reduction of iron oxide. It also permits efficient heat transfer between
the gas and solids.
As the vertical shaft furnace is simple in construction, and existing and proven designs can be used
with a minimum of modification, it is attractive for the application of direct reduction processes.
The disadvantage of the shaft reactor is that it is not possible to use ore fines without agglomeration
because of the need to ensure adequate space within the bed to permit access and passage of the
reductant gas.
The development of mechanical devices to discharge the direct reduction iron from the shaft reactor is
essential for its successful application in the reduction processes.
The shaft is usually fitted with slowly oscillating charge feeders situated just below the reductant gas,
which helps to maintain the inflow of material into the reduction zone. The inflow of the charge and the
uniform distribution of reductant gas across the vertical shaft reactor are essential to assure uniform
metallization in the product. The vertical shaft reactor is designed to operate at ambient pressure so that
the feed and discharge legs can be sealed with a dynamic gas seal using neutral gas. This avoids the
capital costs associated with the mechanical sealing systems required for high pressure operation. The
high pressure process operates at 4-5atm and therefore requires sealing valve systems at the top and
bottom of the shaft, involving considerable complexity. Some shafts have no mechanical feeders but
depend on furnace geometry and good temperature control to avoid charge sticking and to maintain the
inflow.

2.2 Fluidized bed reactors

The main advantage of fluidized bed reactors is their ability to accept iron ore fines as the charge
material without agglomeration. Another advantage is that the rate of reaction is higher. These
advantages over the vertical shaft reactor are, however, offset by some disadvantages.

Pellets

""r1 Lump ore

.rc--------
Reducing

+I
gas

Sponge iron
Figure 2. Simplified diagram of the vertical shaft reactor (shaft furnace)
 ALTERNATIVE IRON /STEEL PRODUCTION PROCESSES 1109

As the main principle of operation of a fluidized bed reactor is good mixing of the charge, it is
necessary to employ two or more beds in series to obtain an approximation of counter current operation,
with associated problems of charge transfer between the beds (Stephenson and Smaller, 1980).
The reductant gas velocity must not exceed the sedimentation velocity of the smallest fines in the bed
so as to avoid carry over of charge to the exit gas stream. Competing with this, the heat and reduction
requirements of the process dictate that a minimum mass flow of gas must be maintained. In order to
satisfy both requirements, fluidized bed systems must operate at elevated pressure.
Another disadvantage is that operation of fluidized beds becomes difficult at high levels of metalliza-
tion as the sticking tendency of the material tends to interfere with fluidization. In practice, the particle
size distribution of iron ore fines may be far from ideal for good fluidization, and the bed will operate
best with fines having fairly narrow particle size range. Operation is more difficult with materials having
high levels of superfine particles.

3. MELTING REACTOR PROCESSES

3.1 Fluidized bed melter gasifier

In this melter (Smith and Corbet, 19881, coal is charged into the top end of the melter where it is dried
and devolatilized by contact with gases produced during the melting process. The coal char creates a bed
which is fluidized by injected oxygen. Prereduced iron is charged by gravity from the prereducer reactor
into the fluidized bed, where it is reduced and melted, creating a pool of slag and metal under the
fluidized bed which is periodically discharged by tapping.
This melter operates in a counter-current mode with temperatures of about 1500°C in the lower
melting part and about 1O00-120O0C in the upper part where the gasification takes place. Thus, the
melter reactor operates as both gasifier and smelter. The metal and slag chemistry is the same as for the
blast furnace.
The engineering principles of this reactor are very similar to those of the blast furnace from which it
borrows its refractory technology and its tapping discharge process.
So far, there is only one commercial process, the Corex direct smelting process, that uses this melter
technology. The Corex process has the following characteristics.

(1) It can be used in conjuction with mini-mills, so producing pig iron at a lower cost and a lower
installed capacity than is possible with the blast furnace.
(2) It allows large integrated iron/steel making plant to increase pig iron capacity by up to 600 ktonne
per year and to supplement the large blast furnaces operating at capacities of up to 10 Mtonne per
year. This gives the integrated iron/steel making plant the flexibility to modify production
according to market demand.

3.2 Electric arc furnace

Today the most common commercial routes for the production of steel from iron are based either on
basic oxygen furnace (BOF) steel-making or on the electric arc furnace ( E M ) (Report, 1987; Open
University, 1979) (Table 2). Being an autogenous steel-making process, the basic oxygen furnace can only
take a limited quantity of iron scrap and direct reduction iron sponge. The electric arc furnace on the
other hand can operate with a charge of 100% iron scrap or 100% direct reduced iron sponge. For this
reason, it is an obvious candidate fo the basis of a mini-mill steel plant.
Steel-making is a refining process in which undesirable impurities such as sulphur and phosphorus are
removed by oxidation. The electric arc furnace was first used commercially in 1914 to produce stainless
and other forms of steel. The consumption of these alloy steels has increased since then and with it has
come the need for larger electrical arc furnaces. Initially, both the electric arc furnace and the open
1110 T. ZERVAS. J. T. McMULLAN AND B. C. WILLIAMS

Powder h n
Figure 3. Simplified diagram of the fluidized bed reactor

PREREDUCED DRI REDUCINGGAS

COAL

1 MELTER I

)SLAG
OXYGEN -+
Figure 4. Simplified diagram of the fluidized bed melter gasifier

hearth steel-making processes opeated on a cold charge which was composed entirely of either direct
reduction iron sponge or iron scrap. Today, the open hearth steel-making process is declining and the
electric arc furance is taking over, and is the prefered choice for mini-mill steel plant of capacity up to
about 500 ktonne per year.
The electric arc furnace is used for the smelting of iron and ferroalloys as well as for steel-making.
Electric energy is fed through carbon electrodes which are either dipped into the slag which acts as the
main resistor, or are terminated above the bath causing electric arcs to be produced between the
electrode and the bath. Another example is the submerged arc furnace, in which heat is generated partly
by arcs and partly by resistance heating in the solid and semimolten iron/steel.
The advantages are that heat is generated in the semimolten iron/steel or slightly above it and with
good heat transfer; there are no combustion gases and the effluent furnace gases are limited to those
produced by the process, with corespondingly smaller losses.
The electric arc furnace is the most common method of producing steel from direct reduction iron ore.
Like the basic oxygen furnace, it is a cylindrical refractory lined vessel, but its diameter-to-height ratio is
greater at between 4 and 6. The main components are the shell, where the hearth is lined with basic
 ALTERNATIVE IRON /STEEL PRODUCTION PROCESSES 1111

Table 2. Crude steel production by process, 1992 (IISI, 1993)

Production Oygen Electric Open Other Total

Country (Mtonneperyear) (%) (%) hearth (%) (%) (%)

Belgium 13.3 89.8 102 100

Denmark 06 100.0 100
France 18% 69.9 30.1 100
FR Germany? 39.7 77.1 25.6 1.3 100
Greece 09 100.0 100
Ire1and 0.3 100.0 100
Italy 24.9 408 59.2 100
Luxembourg 3.1 100.0 100
Netherlands 5.4 95.5 4.5 100
Portugal 0.7 55.4 44.6 100
Spain 12.3 40.6 59.4 100
United Kingdom 16.0 75.3 24.7 100
EC Total 132.2 66.8 32.8 0.4 100
Austria 4.0 90.9 91 100
Finland 3.1 83.1 16.9 100
Norway 04 100 100
Sweden 4.4 664 35.6 100
Turkey 10.3 34.1 59.6 6.3 100
Yugoslavia 0.7 74.6 20.9 4.5 100
Other W Europe 22.8 56.9 401 3.0 100
Total W Europe 155.0 65.3 33.9 08 100
Canada 13.9 67.3 32.7 100
United States 883 62.0 38.0 100
Japan 98.1 68.4 31.6 100
Australia 6.9 92.5 7.4 100
New Zealand 0.8 69.7 30.3 100
South Africa 9.1 63.4 35.6 1.0 100
Total Indust. Cts 368 1 65.9 33.7 03 00 100
Argentina 2.7 37.7 62.3 100
Brazil 23.9 78.9 19.3 1.9 100
Chile 1.o 95.1 4.9 100
Mexico 8.4 44.4 55.6 100
Peru 0.3 42.6 57.4 100
Venezuela 3.3 100.0 100
EDPt 25 35.4 58.8 5% 100
India 18.1 45.5 28.2 26.3 100
Rep. of Korea 28.1 69.8 30.2 100
Taiwan (ROC) 10.7 50.6 48.8 0.5 100
Total Market Econ 467.1 64-6 340 1.3 01 100
Bulgaria (E) 1.6 50.3 40.0 9.7 100
Czechoslovakia 11.1 53.9 11.0 35.0 100
Hungary 1.6 85.9 5.1 90 100
Poland 9.9 63.3 18.2 18.4 100
Romania 5.4 527 30.8 16.6 100
Russia 67.0 345 15.4 50.1 100
Ukraine 41.7 39.8 7.6 52.6 100
Total EITs $ 1382 41.2 13.6 45.2 100
Total § 605.3 592 294 11.3 0.1 100

tIncluding former GDR.

$Economy in transition.
§The countries included account for 84% of total world output.
1112 T. ZERVAS. J. T. McMULLAN AND B. C. WILLIAMS

Figure 5. Simplified diagram of electric arc furnace

refractories and holds the charge, the water cooled wall and roof panels to allow for greater power input,
an automatic positioning mechanism to hold the electrode and maintain the proper arc length, a
mechanism to remove the roof for feeding and another to tilt the furnace of tapping and deslagging.
Electric furnace shops are equipped with environmental equipment for water treatment and degreasing,
and noise and fume extraction.

3.3 Pneumatic reactors (basic oqgen furnace)

There are two kinds of pneumatic reactor (Lankford et al., 1985; Report, 1987). The basic oxygen furnace
is oxidizing, as in steel-making converters. Reducing reactors use a reducing gas in place of the oxygen, so
that reduction and deoxidation of the bath is obtained. Lime may also be introduced into the smelting
bath.
In the basic oxygen furnace, the oxygen reacts with a molten charge and is introduced through tuysres
directly into the smelting bath. Depending on the oxygen injection configuration, the furnace is
designated as top-blown, bottom-blown or side blown. The oxygen can also be introduced directly through
lances which dip into the bath.
In basic oxygen furnace steel-making, the reaction in the smelting bath is normally sufficiently
exothermic to supply the required heat. For reducing processes where a reductant gas is blown into the
reactor, the required heat may be obtained by introducing oxygen along with the reducing gas, or by
feeding a mixture of coal power and oxygen-enriched air.
The basic oxygen furnace is the most common process in the production of steel. It is one unit in the
integrated iron/steel making process, which also includes the coking unit, pelletizing plant, sintering
plant and blast furnace. In the basic oxygen furnace process, liquid metal, scrap and fluxes are fed into a
pressurized refractory lined vessel. First, the oxygen reacts with carbon and silicon to produce the
required heat to increase the temperature of the bath from about 1350" C to over 1650"C and melt the
scrap. The lime removes the sulphur and phosphorus from the liquid iron.

4. COMPARISON OF STEEL-MAKING PROCESSES

When we compare the basic oxygen furnace and the electric arc furnace we must remember that the
main function of the electric arc furnace is to supply heat for melting, while that of the basic oxygen
 ALTERNATIVE IRON/STEEL PRODUCTION PROCESSES 1113

Figure 6. Simplified diagram of the basic oxygen furnace

furnace is to refine by oxidizing the carbon and silicon and other elements and to increase the
temperature of the iron by using the heat of the reactions (Report, 1987). Consequently, they use
different charge materials, mainly direct reduced iron or scrap in the electric arc furnace, and molten
iron or iron carbide in the basic oxygen furance, to produce a similar product - the molten steel.
The major difference between the direct reduction electric arc furnace and the blast furnace basic
oxygen furnace steelmaking routes is the difference in energy consumption. For current efficient
integrated blast furnace/basic oxygen furnace plant, the energy requirement is typically 19.24 GJ tonne- '
of molten steel while that for scrap-based arc furnace mini-mill steelmaking is 8.49 GJ tonne-' molten
iron. The difference in energy consumption lies in the fact that a large part of the energy needed to
produce molten steel is associated with the reduction of iron oxide. With scrap, there is no need for
reduction because this stage is already completed.
With the basic oxygen furnace, the charge contains 25-30% scrap and the rest contains liquid iron and
flux, usually lime. Oxygen is introducd at a rate of 20 to 30 thousand cubic feet per minute. The oxidation
starts first with silicon, and is continued with carbon, then manganese. Through these oxidation reactions,
the scrap is melted and the iron temperature increases from about 1350" C. The lime flux merges with the
oxidized silicon, iron and manganese to form slag which removes about half of the sulphur and the
majority of the phosphorus. The process time is usually 15 to 20miutes. After oxidation, the liquid metal
is sampled and then discharged.
Electric arc furnaces for steelmaking have developed significantly over the past 20 years. Usually,
scrap is fed into an empty furnace and the arc melts a hole down through it using the scrap burden to
protect the walls from the arc flare. When all of the iron scrap is melted, during the refining and
superheating steps, the voltage is decreased to reduce the power, so protecting the walls and roof from
excessive radiant heat. Foamy slag practices have been used in which oxygen and carbon are introduced
into the bath to form a gas which foams the slag around the electrodes, protecting the walls and roof
from the arc and allowing higher power to be used. The carbon is removed by introducing oxygen.
Natural gas or oil can be used to supply extra heat to the electric arc furnace during melting, reducing
the melting time by about lominutes and the electrical consumption by 30 to 50KW h tonne-' of molten
steel.
It is usually not easy to evaluate in detail the advantages and disadvantages of each method of
1114 T. ZERVAS, J. T. McMULLAN AND B. C. WILLIAMS

producing molten steel because they depend on local conditions such as price and quality of the raw
materials and the condition of the market in which the steel products will be sold. Consequently, only
general conclusions of the advantages and disadvantages can be drawn.
The investment cost of an electric arc furnace steel-making plant is considerably less than that for an
integrated coke oven, blast furnace and basic oxygen plant. Also, producing steel from scrap costs less in
terms of cost, labour and environmental impact (see Table 3). Additionally, 60% less energy is needed to
produce molten steel in an electric arc furnace and relatively small plants of the order of 600 ktonne per
year capacity can be erected economically to produce a wide range of steel products competitively.
The disadvantages of electric arc furnace are as follows.

(a) The off-gas contains a dust classified as a hazardous waste.

(b) As the nitrogen content in the steel can be higher in the electric arc furnace than in the basic
oxygen furnace, it is difficult to produce certain steels.
(c) The main energy supply is electricity, which is expensive.

Despite these disadvantages, developments in the last 20 years have improved the efficiency of electric
furnace operation in terms of both its productivity and its ability to produce steel products competitively.

5. SMELTING PROCESSES

5.1 Corex process

The Corex process is the only commercial process that produces molten iron by direct reduction of iron
ore (Chatterjee, 1994; Delport, 1991; Feichtner et al., 1989; Scott, 1994; and Steffen, 1990). It was initially
developed as the K-R process by Korf engineering and Voest Alpine in West Germany in 1978. After a
successful pilot plant operation at Kehl/Rhine, Germany, in 1981, a plant of 300 ktonne per year was
constructed at ISCORs Pretoria Works and began successful operation in 1989.
The start-up was very successful and about 270 ktonne of hot metal were produced during the first year
of operation, representing 90% of the design capacity.
The Corex process combines a gasifier-smelter with a vertical-shaft reduction reactor. Hot reductant
gases are generated from a wide variety of untreated coals and oxygen in a smelter/gasifier and are
passed through a counter-current reduction shaft charged with lump ore and/or pellets.
The Corex process is designed to operate at a pressure of up to 5 bar. Charging of coal and iron ore is
done through a lock hopper feeder. The coal is stored in a feed bin and charged into the melter-basifier
by a feed screw conveyer. The coal falls by gravity into the gasifier where it comes into contact with a hot
reductant gas atmosphere at a temperature between 900°C and 1300°C and is instantly dried and
degasified. Reducing gas is generated by gasification of coal in a fluidized bed in the upper part of the
smelter.
The flow velocity is selected to maintain the fluidized bed at a temperature between 1500°C and
1800" C. The reductant gas has a high CO content of about 85% and after leaving the melter gasifier is
cooled to between 800" C and 900" C by mixing with cooling gas, and is prereduced in a hot dust cyclone
before being introduced to the vertical shaft reductant reactor. A small amount of the cleaned gas is
converted to cooling gas in a gas cooler. The fines captured in the hot cyclone are recirculated to the
gasifier.
The reductant gas is fed into the reduction furnace, flows counter-current through the descending iron
ore and escapes from the top end where it is exported for electricity generation or chemicals production,
or is recycled back to the process. Partial reduction takes place in the vertical shaft, the partially reduced
iron is continuously charged into the melter-gasifier, where further reduction and smelting takes place.
Molten iron and slag are discharged by conventional tapping, using a practice similar to that used in the
blast furnace. The tapping time is between 150 and 180 minutes.
 ALTERNATIVE IRON /STEEL PRODUCTION PROCESSES 1115

Table 3. Breakdown of steel-making costs (Report, 1987)

Item cost US$

Electric furnace Oxygen furnacet

Materials (ore, scrap fluxes etc) 10235 83-26

Energy (coal, oil, electricity,
oxygen, gas credits) 32.84 31.6
Maintenance, materials and
services 4 94
Labour 13.44 1913
Capital (interest and depreciation) 20 47.02
Total liquid steel cost 17263 190.41

?Oxygen furnace costs include components in coke and hot wet.

COAL

~~

GASIW

>El.AG

I & HOTMETfi
Figure 7. Simplified diagram of the Corex process

The degree of metallization of direct reduction iron in the vertical shaft furnace reactor is about 95%
and the carbon content varies from 2% to 5%, depending on the raw material and operating conditions.
The energy consumption is about 13.4 GJ tonne-' of molten iron.
The advantages of the Corex process are as follows.

(a) It is the only smelting process in which coke can be replaced by a wide variety of coals.
(b) In comparison with the blast furnace, it is more environmentally friendly as it avoids coke
production.
1116 T. ZERVAS, J. T. McMULLAN AND B. C. WILLIAMS

Table 4. Capital cost of a typical 300ktonne-per-year COREX plant (in M$US - 1987 estimates)
(Chatterjee, 1994)
~~

Item Corex unit Power

Engineering 12.6 2.1 6.3

Equipment (imported) 37.1 25,2 75.6
Other equipment 7.7 (cost included in power plant) 3.5
Civil and structural 6.3 0.7 4.2
Utilities 1.4 (cost included in power plant) 1.4
Erection supervision 9.8 0.7 2.1
Licence fee 2.1 - -
77.0 28.7 93.1
Total cost of Corex + oxygen 198.8
+power plant
Other costs @ 5% on Item 8 9.8
Subtotal of 8 + 9 208.6
Contingencies @ 15% 31.5
Total cost 240.1

Table 5. Operating cost of a 300 ktonne-per-year Corex plant

(in US$ - 1987 estimates) (Chatterjee, 1994)

Blast furnaces Corex

Raw materials 74.0 64.3

Fuel 0.1 -
Labour 3.O 3.0
Stores 1.o -

Transfers 17.0 18.0

Credit gas -( 5 ) -(19.0)
Total cost above 16.2 2.1
Total works cost 106.3 68.4
Credit for slag - (3.0) - (3.0)
Net works cost 103.3 65.4

(C) The process generates large volumes of process gases that have a high utility value and can be
exported.
(d) In the event that the off-gas cannot be exported economically, it can be recirculated - reducing
coal and oxygen consumption.
(e) Finally, Corex plant is relatively cheap to construct.

5.2 AISI process for direct smelting

Development of this process began in 1987 when AISI undertook a feasibility study of direct steel
making. In 1990, a pilot plant for preheating and prereduction of ore pellets was constructed with
HYLSA at Monterrey, Mexico, and a smelting pilot plant was build in Pittsburg (Fruehan, 1993;
Chatterjee, 1994; Scott, 1994; Aukrust, 1992).
 ALTERNATIVE IRON /STEEL PRODUCTION PROCESSES 1117

The main part of the project is the 5 tonne h-' pilot plant at Pittsburgh, Pennsylvania, where liquid
iron and semi-steel has been produced. This direct smelting process operates as follows (see Figure 8);
coal and prereduced iron are charged to a liquid metal and slag bath (smelting reactor) in which the
prereduced iron is reduced and combustion of the coal supplies the necessary heat for reduction and
melting of the iron. The important factor in this process is that the heat requirement is met by
post-combustion of the reductant off-gases. The off-gas contains a mixture of 40% carbon dioxide and
steam and is used for the prereduction of the iron ore pellets from haematite to wustite. This
prereduction is done in a conventional shaft furnace. The final reduction takes place in the smelter,
where gasification of the coal also occurs.
Although research and development on the AISI process began later than that on most other process
'
development programs, it has reached a more advanced state of development, and a 50 tonne h- of hot
metal demonstration plant is being designed with the following criteria.

(a) The prereduction of haematite to wustite will be about 255%.

(b) The post combustion will be at least 40%.
(c) The system with be pressurized.
(d) The process will involve continuous smelting.
(e) Batch tapping will be used.
(f) The production intensity will be 10 tonnes of hot iron per day per cubic metre.
(g) The refractory wear will be 0.25 mm h- I .

The AISI feasibility study reports that operation and investment cost will be lower and more favourable
than the classical route of coke oven and blast furnace.

5.3 DIOS process

The DIOS process has been developed through a collaboration of Japan Iron and Steel Federation, eight
major integrated steel producers and the Coal Mining Research Centre of Japan (Chattejee, 1994; Scott,
1994).

COAL

~TED&ppLpuIu(3D~m

HUI'BUZALISIAG

Figure 8. Simplified diagram of the AISI process

1118 T. ZERVAS, J . T. McMULLAN AND B. C. WILLIAMS

As shown in Figure 9, the DIOS process consists of three main units, a fluidized bed reactor where
prereduction takes place, a smelting reduction vessel where final reduction and smelting take place and a
gas reforming furnace. A pilot plant is under construction to examine the following aspects.

(a) When can liquid iron be produced in one step in a single unit in the smelter reduction vessel where
gasification, reduction and smelting take place in the same reactor?
(b) When can liquid iron ore be produced in two steps in two units, with prereduction in the
prereduction furnace and final reduction and smelting in the smelting reduction vessel?
(c) When can liquid iron ore be produced in three steps in three units, with prereduction in the
prereduction furnace, gasification in the gas reforming furnace and smelting in the smelting vessel?

5.4 Hismelt process

The Hismelt process was developed by the Australian company CRA and Klockner in the 1980s. The
process consists of two main units, a prereduction system in which prereduction takes place, and a smelt
reduction vessel in which final reduction and smelting take place (see Figure 10). The process is
somewhat similar to the AISI process (Chatterjee, 1994; Scott, 1994).
The Hismelt process is a two-step process. Iron ore is prereduced to about 22% in the prereduction
system and the prereduced iron is introduced at 850°C to the smelter reduction vessel where final
reduction and smelting take place. As with the DIOS and AISI processes, Hismelt involves a high
post-combustion of gas in the smelter. The Hismelt process differs from the AISI and DIOS processes
because, instead of using deep slag formation, it relies on violent agitation of the liquid metal to secure a
high level of heat exchange in the smelter. Some characteristics of the process are as follows.

(a) As the smelt reduction reactor is of the horizontal type, it is possible to achieve high injection
rates.
(b) Lower end pulverized coal injection maximizes the gasification of the coal and so increases the
reduction rate.
(c) Part of the off-gas is utilized by the post combustion step to preheat the air.

6. ALTERNATIVE ROUTES IN IRON- AND STEEL-MAKING

6.1 Plasma technology in iron-making

Steel refining and melting has used plasms technology for many years (Fey, 1985; Eriksson et al., 1987;
Santen et al., 1987; Barcza, 19861, but commercial application for direct reduction and smelting of iron
ore has not been applied (except for the Plasmadust process, which is a variation of the Plasmasmelt
system in which iron/steel dust is reduced to iron). Despite a considerable effort in research and
development over the last twenty years, no such process has passed the pilot stage.
The three main domains are being developed: the application of plasma in the blast furnace, in direct
reduction and in direct reduction/smelting.
In plasma processes, a plasma is generated either from an inert gas such as argon or from a reactive
gas such as hydrogen. Iron ore and carbon in the form of fines are then introduced to the system. Plasma
provides an intense concentrated source of energy for endothermic reactions and has a high energy
efficiency if primary energy is ignored and only net energy is considered for comparison with other
processes. Because of the energy density, the reactions are very fast, leading to smaller reactor sizes and
lower capital costs for the same plant capacity.
In spite of these advantages, the process has not been commercialized because of the cost of electric
energy for plasma generation. The only commercial application is the Plasmadust process of SKF steel. A
 ALTERNATIVE IRON / STEEL PRODUCTION PROCESSES 1119

LIMESTONE

Smelter Rduttion
VUrd

1"
ORE

LIMESTONE

k
Hot Mot81
Smoltor G8oMr

Figure 10. Simplified diagram of the Hismelt process

plant of 70ktonnes per year was erected between 1982-1984 at Landskrona, Sweden, to recover iron
from various waste oxide dusts (Steinmetz et al., 1986).
In conclusion, research and development on these processes is targeted towards replacing metallurgi-
cal coke with less expensive coal, to using lower quality characteristic iron minerals and building plants
with smaller capacities.

6.1.1 Plasmasmelt process

The plasmasmelt process (Figure 11) (Smith and Corbett, 1987,1988; Eriksson et al., 1987) was developed
by SKF in Sweden. The process consists of three main units, two are fluidized beds, one of which is for
 Y

Table 6. Details of large-scale applications of plasma systems (Barcza, 1986)

Process name Production

Commodity Company Process type Device or contractor capacity

1. DRI SKF Sweden Indirect non-transferred Plasmatech arc heater Plasmared process 70 ktonne
furnace gas reformer-shaft lx6MW per annum
LPG or coal slurry
2. DRI USCO, South Indirect non-transferred gas Huls arc heaters Fluor/Davy McKee 320 ktonne
Africa reformer-shaft furnace 3X8MW per annum
syngas
3. Pig iron Cockerill Indirect non-transferred Westinghouse arc Pirogas process Industrial blast
Steel-CRM superheating of gas blast to 1x3MW furnace
and a U.S. steel blast furnace
producer
Iron and steel 4. Pig iron (alloyed SKF Smelting in a coke-filled Plasmatech arc heaters Plasmadust 70 ktonne per annum
with Cr, Mn and shaft furnace-Zn, Sn fumes 3X6MW (Scandust) (250 ktonne-per-annum
others) produced 7X6MW (Plasmasmelt) study only
5. Pig iron ASEA d.c. transferred-arc open Hollow graphite Elred process (600 ktonne pe annum
bath electrode 40 MW study only
6. Steel M.A.N. d.c. transferred-arc Solid graphite Unarc process 15 tonne furnace
GmbH open-bath scrap melting electrode 6 MW
7. Steel Freital-Voest d.c. transferred-arc Water-cooled
Alpine, open-bath scrap melting plasmatrons
3X7MW 35 tonne
4X7MW 45 tonne
8. Stainless steel ASEA-Krupp d.c. transferred-arc Solid graphite SE process (single 55 tonne
open-bath electrode 18 MVA electrode)
1. Ferro-chromium Mintek- ASEA d.c. transferred-arc Hollow graphite d.c. chrome 50 or 20 ktonne per annumt
open-bath co-melting and electrode 16 MVA process
smelting
Ferro-alloys 2. Ferro-chromium SFK Direct non-transferred shaft Plasmatech arc heaters plasmachrome 78 ktonne per annum
furnace (coke-filled) - 4X6MW
preheated ore feed
3. Ferro-manganese Samancor d.c. transferred-arc Freital-Voest Alpine Freital-Voest 50 ktonne per annum
open-bath melting plasmatron Alpine
1 0 8 MVA (8 MW)

t50 Kt/a Melting and smelting 2.1

20 Kt/a Smelting
 ALTERNATIVE IRON/ STEEL PRODUCTION PROCESSES 1121

drying and preheating and the other is for prereduction of the iron ore. The third is a shaft furnace
plasma reactor in which final reduction and smelting take place. Iron oxide fines are mixed with lime and
charged to the preheat fluidized bed reactor, which uses the off-gas from the shaft smelter reactor or the
fluidized bed prereduction reactor. After this, the dried iron ore is passed to the second fluidized
prereducer reactor, where prereduction takes place by contact with the off-gas from the shaft smelter
reactor. Prereduced iron oxide with 70% reduction is pneumatically injected with coal fines to the shaft
smelter reactor. The vertical shaft smelter reactor is fed at the top with coke. Plasma generators are
mounted in the lower part of the shaft reactor. Molten iron is collected from the lower end of the shaft
smelter by tapping. The majority of the off-gas from the smelter reactor consists of CO and H,, and after
cleaning it is recirculated back to the two fluidized bed reactors. A small part of the cleaned off-gas is
compressed and used as plasmagas and for pneumatic injection of coal fines and the prereduced
haematite. Energy consumption is estimated to be about 10GJ tonne-' of direct smelted liquid iron.
Plasmadust is a variation of this process in which various iron oxide dust wastes are reduced to iron, and
Plasmachrome is another in which chromite concentrate is reduced to ferrochrome.

6.I.2 Elred process

The Elred process (Smith and Corbett, 1987,1988; Davis et al., 1982; Chatterjee, 1994) was developed by
ASEA in Sweden and consists of two main units, a fluidized bed reactor prereducer and a d.c. plasma arc
furnace reducer-smelter. Combustion air and pulverized coal is introduced into the fluidized bed reactor
to heat and maintain the bed at 980" C and also to produce the CO and H, for the prereduction. Iron ore
is charged and heated to about 200°C in a venturi preheater by off-gas from the fluidized bed. The
burden is reduced about 75% in the prereducer fluidized reactor and is fed through a hollow electrode at
about 700°C to the arc plasma furnace where final reduction takes place. The excess gas is used to
produce electricity for the process. The energy consumption is about 15.5 GJ thm-'.

slag Hot-
Figure 1 1 . Simplified diagram of the Plasmasmelt process
1122 T. ZERVAS, J. T. McMULLAN AND B. C. WILLIAMS

:r-
Hot air Prereduction

Slag
'I
Prereduced
carbonaceous
material

+( ,gee )4-l(
f
Combu8tlon (I.8

/
Electricity

Substation )
1
Liquid Iron I
To the grid

Figure 12. Outline of the Elred process

6.1.3 Pirogas process

The Pirogas process (Smith et al., 1988, 1987; Eriksson et al., 1987) is an application of plasma technology
in the blast furnace. It was originally developed at the Centre de Research Metallurgiques (CRM), and
the name stands for 'plasma injection of reducing overheated gas'. The intention of the process is to
decrease coke consumption by replacing part of the hot blast blown through the blast furnace tuysres by
reductant gas heated to 2000" C in a plasma heater. The advantage of the process lies in the increased
production of the blast furnace and in the possibility of using off-peak electicity. Depending on the
operating conditions, the ratio of electricity to coke can be varied over a wide range. The optimum
economic conditions occur with operation at a temperature of about 1300"C. Operation is normally
between 1300 and 2200" C, and electrical energy consumption is around 500 KW h tonne-' of liquid
metal.

6.2 Iron carbide process

The iron carbide process (Dancy, 1990; Munson, 1995; Scott, 1994; Chatterjee, 1994) was invented and
developed by F. Stephens with his company Iron Carbide Holding Ltd. U.S.A. A 25 tonne per day pilot
plant was erected and began operation in 1989. The first commercial iron carbide process plant has been
completed, with start-up in Trinidad during 1995. In this process, iron ore fines are reduced and iron
carbide formed in a fluidized bed reactor using a mixture of hydrogen and carbon monoxide. The
reduction of the iron ore fines (haematite) takes place at 600" C and 1.8atm in a fluidized bed reactor.
The iron carbide is discharged with a typical composition of 93% Fe,C, 4% Fe,O, and 3% gangue and is
transferred to a basic oxygen furnace at a temperature of 500°C. The iron carbide can be a 40%
substitute for molten iron produced in a blast furnace and a 100% substitute for scrap in a basic oxygen
furnace operation. Energy consumption is about 12.9GJ tonne-'.

7. DIRECT STEEL-MAKING
The convesion of iron minerals to steel takes place in four different steps: gasification, reduction,
smelting and refining. Commercial processes today use at least two reactors. The classical route of
 ALTERNATIVE IRON / STEEL PRODUCTION PROCESSES 1123

PLASMAARC HEATER

‘I I I

Figure 13. Plasma arc blow-pipe and tuyere assembly

IRON ORE YO
FLUIDIZED
CONSENTRATE BED UNIT PREHEATER

Fe,C
A
co : 440
SCRAP HOT COOLAND
CLEAN
, & MAKEUP

*
STEEL
Figure 14. Outline of the iron carbide process

coke-making, blast-furnace and oxygen steel-making (refining) uses three reactors; the alternative route
of direct reduction (gas based) and electric arc furnace (refining) uses two reactors. Finally, the new
alternative routes try to convert iron minerals to steel in only one reactor in which all four reaction steps
take place (OTA, 1980; Fruehan, 1990; Nabi, 1993; Tanaka, 1988). The advantages of direct steel-mak-
ing are as follows.
1124 T. ZERVAS. J. T. McMULLAN AND B. C. WILLIAMS

(a) Steel is produced in one reactor only by replacing the classical route of cokemaking, blast-furnace
and oxygen steelmaking.
(b) Steel is produced in one reactor only by replacing the alternative route of direct reduction/smelt-
ing and electric arc furnace.
(c) The direct steel-making process can generate large quantities of process off-gases which have a
high value for electricity generation or chemicals production (e.g. Methanol).
(d) As air is used instead of oxygen, no oxygen production unit is needed and both investment and
operating costs are reduced.
(e) Environmental problems are decreased.
(f) Any grade of coal or any alternative reductant/fuels and any type of iron minerals with different
characteristics may be used, leading to increased flexibility.
(g) Capital costs are low.

In spite of these advantages, no commerical plant has yet been built. An experimental pilop plant (AISI)
claims to have succeeded in producing semi-steel directly from iron ore.

8. ADVANTAGES OF IRON DIRECT SMELTING PROCESSES

The advantages of direct smelting processes may be summarized as follows (Tanaka, 1988).

(a) Flexibility in capacity is improved because, as the process is faster than the balst furnace, the
reactor size can be reduced. Also, as iron ore lumps or fines can be used directly, there is no need
for pretreatment such as pelletizing or sintering.
(b) Iron direct smelting process use coal, so there is no need to construct a coke oven.
(c) There is high flexibility in raw materials and reductant fuels. This means that iron ore fines, iron
sand, scrap or any mixture of these materials can all be used. As a result, there is a wide choice of
cheap raw materials for the smelting reduction process. Concerning reductants/fuels, non-metal-
lurgical coke, low-grade coal, anthracite and lignite can all be used.
(d) Operation is highly flexible, which means that the direct iron smelting process can be stopped and
started with relative freedom.
(e) There is wide flexibility in the use of the off-gas which are generated in large quantities. These
process gases have a high utility value as energy or chemical raw materials. Thus, the smelting
reduction furnace can also be used as an energy generator. In fact, it can be run in a variety of
modes determined by the rate of prereduction in the production furnace and the rate of secondary
combustion of the gases produced in the smelting reduction furnace. The lowest mode generates
only the minimum energy required for the production of molten pig iron. In the highest operating
mode, the furnace operates like a coal gas generator to produce more energy than is required by
the steel-making system. The surplus is supplied to outside needs.

9. COMPARISON OF DIRECT REDUCTION WITH DIRECT SMELTING PROCESSES

Direct reduction and direct smelting processes have many similar advantages (Gudenau et al., 19891, most
of which have already been discussed. In addition, direct smelting processes allow the use of a wide range
of raw materials and reductants/fuels, and the smelting furnace can also act as a gasifier.
Direct smelting processors usually employ a vertical shaft furnace for prereduction of the iron ore
(Corex process). The advantage here is that, as the iron ore charge does not need to be agglomerated,
investment and operating costs are reduced.
Direct reduction has a major disadvantage (Eketorp, 1987) in the nature of the product itself the solid
sponge iron. After reduction and discharge, it has to be protected against reoxidation. Also, melting of
 ALTERNATIVE IRON /STEEL PRODUCTION PROCESSES 1125

the solid sponge requires electric power, which can be uneconomical in many countries. Sponge iron also
has to complete with scrap, which is often cheaper and more easily available. Sticking is also a problem
during the production of sponge iron. By contrast, the direct smelting process has resolved the sticking
problem as the iron ore is prereduced in the reductant shaft furnace, avoiding metallization as final
reduction, metallization and melting all take place in the smelting furnace (Corex process).
Commercial direct reduction process plants use air instead of oxygen, while the Corex direct smelting
process uses oxygen.

10. RAW MATERIALS FOR DIRECT REDUCTION AND SMELTING

The designer of a direct reduction and smelting plant must take into consideration the characteristics of
the raw materials (size distribution, strength and chemical properties) that it is intended to use because
this determines the type of process (vertical shaft reactor, fluidized bed reactor) that can be adopted. The
characteristics of the iron ore determine the type of reactor and vice versa (Stephenson and Smaller,
1980).

(a) Fines iron ore is used in fluidized bed reactor processes, but the iron content in the mineral should
be at least 64% (e.g. the commercial application of fluidized bed reactor for prereduction in the
Plasmasmelt process).
(b) A mixture of pellets and lumps can be used in shaft reactors, but there is a maximum allowable
percentage of lumps for efficient operation. Any reduction in the concentration of lumps increases
the efficiency of the process, and optimum operation is actually obtained when the charge is 100%
pellets (e.g. in the commerical application of a shaft reactor for prereduction in the Corex process).
The iron content in the mineral should be at least 65% (Delport, 1991).

The important role played by the propeties of the raw materials in the direct smelting process is obvious
and it is important to examine how this is affected by factors such as factory siting, the type of process
and the raw materials used (Steffen, 1989).
The iron and steel industries normally pay a price that includes transport (the CIF price) (Hoddinott,
1992, 1993, 1994). If, on the other hand, industry is connected with an iron mining operation and has a
local direct reduced iron plant, they will be charged the FOB price (or less, as they include an element of
the mining costs). It is very difficult to be precise about the correct variation between FOB price and CIF
price as the cost of transportation is different for every location. As an indication, an agreement was
made in 1991 between Australia’s Hamersley and Nippon Steel Corporation, Japan, for an FOB price on
a long-term basis of 33.49 cents per Fe unit for Hamersley fines and 41-48 cents per Fe unit for
Hamersley lump.
Sea freight is an important element in the cost of iron ore (Duisemberg, 1993, 1994; Metals and
Minerals Annual Review, 1992). The rates are variable and 1991 the spot freight rates varied from
between $US 8 and $US9.4 per tonne from Brazil to Europe to between $9.5 and $ 11.2 per tonne from
Australia to Europe, and between $11.7 per tonne and $13.5 per tonne from Brazil to Japan.
Energy cost is the single most important factor in the total cost of direct smelting of iron. In all
processes the energy requirement is met by consuming coal, and in some cases, such as Plasmasmelt and
Elred, supplementary energy to meet the heat requirement is also provided by using electricity. In 1991,
steam coal prices in Europe were between 37 and $39 per tonne from South Africa and between $45 and
$46 per tonne from many U.S. mines. The average price of coking coal was $61 per tonne.

11. CRITERIA FOR EFFICIENT OPERATION O F DIRECT SMELTING PROCESSES

In direct smelting processes, energy efficiency can be increased by the following measures.
1126 T. ZERVAS, J. T. McMULLAN AND B. C. WILLIAMS

(a) The gas should be operated in counter-current flow for high gas utilization (Oeters and Saatci,
1987).
(b) The experience of years of blast furnace operation must be applied to direct smelting processes,
particularly during reduction.
(c) The process should operate with minimum off-gas for high productivity.
(d) A high degree of post-combustion degree is needed for optimal prereduction, reduction and
smelting, and for high energy efficiency (Fruehan et al., 1989; Wright et al., 1991).
(e) Waste heat should be used for electricity generation.

The necessary heat requirement for gasification of carbon and smelting can be provided by:

(a) combustion of fuels;

(b) electrical energy where cheap electricity is available (Mitsubishi, 1991).

12. ENERGY CONSUMPTION

When comparing direct smelting processes and their alternatives, it is apparent, as shown in Table 7, that
the Plasmasmelt process has the lowest energy consumption at about 10GJ tonne-' of liquid melt.
Nonetheless, the process is economically feasibly only where cheap electricity is available.

13. CONCLUSIONS AND RECOMMENDATIONS

Coal represents about 28% of total world energy consumption (Hoddinott, 1992, 1993, 19941, and with
estimated reserves of 1040Btonne has a reserve life of 230 years at a consumption of 4.5 Btonne per year.
By contrast, oil has an estimated reserve life of 42 years and natural gas has an estimated 59 years.
Additionally, 70% of oil and gas reserves are concentrated in the Middle East and the former Soviet
Union while coal has a more even distribution. Nonetheless, the increasing cost and scarcity of
metallurgical coal for producing coke for blast furnaces ensure that direct processes for producing iron
will continue to be developed. Today, the most widely used direct processes for producing iron are based
on natural gas, but this is neither available nor cheap on a world-wide basis. Thus, coal must have a high
R& D priority.
The main research and development thrusts needed for the new iron-making direct smelting processes
are (Fruehan, 1993) as follows.

(a) Higherproduction intensity: the target is further to reduce the working volume of the prereducer
and the smelter furnace relative to the blast furnace.
(b) Increased post-combustion and heat tmnsfer: the majority of the energy for melting must be met by
post combustion, and this energy must be efficiently transferred to the smelting system.
(c) New rejructories must be developed because of the large amount of energy released.

The following options look promising for the future development of direct smelting processes.

Table 7. Comparison in energy consumptions

Iron
Process Corex Plasmasmelt Elred Carbide

Energy (GJ tonne-' of product) 13.4 10 15.5 129

Reductant coal coal coal gas
 ALTERNATIVE IRON /STEEL PRODUCTION PROCESSES 1127

(a) Combination of a smelter reactor with a gas-based shaft reactor. In this way, the waste gas from the
smelter reactor can be used in the direct reduction stage.
(b) Use of the waste gas from the smelter reactor for producing electricity or marketable chemicals
such as methanol.
(c) If the smelter reactor is used as a gasifier, any gas-based process can be incorporated.
(d) Use of a combination of coal or peat with biomass to charge the smelter reactor. In this way, the
sulphur content of the reductant gas and the liquid iron product can be reduced.
(e) Use of waste material that can be produced in a form and quality similar to peat (in terms of
humidity, density, etc.) to charge the smelter.
(f) Combination of a large capacity gasifier with a plasma heater for direct smelting of iron or for
injection to a blast furnace (Pirogas) with electrictric power generation or a chemical plant.

REFERENCES
Aukrust, E. (1992). ‘Results of the AISI/DOE direct steelmaking program’, Int. Symp. on Smelting, Quebec, Canada, pp. 591-610.
Barcza, N. (1 986). ‘The development of large-scale thermal-plasma systems’, J. South Afncan Imt. Mining & Metallurgy, August,
317-333.
Chatterjee, A. (Ed.) (1994). Beyond the Blast Furnace, CRC Press, U.S.A.
Dancy, T. (1990). ‘Solid state reduction or iron ores’, Elliot Symp. Proc., Cambridge, Massachusetts, U.S.A., pp. 11-31.
Davis, C., McFarlin, J. and Pratt, H. (1982). ‘Direct-reduction technology and economics’, Ironmaking & Steelmaking, 9, 93-129.
Delport, H. (1991). ‘The COREX process’, 2nd European Ironmaking Congress, Glasgow, U.K., pp. 289-302.
Duisemberg, E. dos S. (1993). ‘Steel industry metals’, Metals & Minemls Ann. Reu., 58-62.
Duisemberg, E. dcs S. (1994). ‘Steel industry metals’, Metals & Minemls Ann. Reu., 51-55.
Eketorp, S. (1987). ‘Energy considerations of classical and new iron and steel-making technology’, Energy, 12, 1153-1168.
Eriksson, S., Johansson, B. and Santen, S. (1987). ‘Applications of plasma technology to steelplant waste treatment’, Plasma
Technology in Metallurgical Processing, Iron & Steel Society, U.S.A.
Feichtner, H.,Maschlanka, W. and Helten, F. (1989). ‘The COREX process’, Skillings’ Mining Reu., 14, 20-27.
Fey, M. (1985). ‘Plasma technology for the steel industry’, Iron & Steel Engr, 38-45.
Fmehan, R. (1990). ‘Iron bath smelting-current status and understanding’, Elliot Symp. Proc., Cambridge, Massachusetts, U.S.A.,
pp. 1-10,
Fruehan, R. (1993). ‘Challenges and opportunities in the steel industry’, Iron & Steelmaker, March, 59-64.
Fruehan, R., Ito, K. and Ozturk, B. (1989). ‘Analysis of bath smelting processes for producing iron’, Steel Res., 60,129-137.
Gudenau, H.,Fang, J., Hirata, T. and Gebel, U.(1989). ‘Fluidized bed reduction as the prestep of smelting reduction’, Steel Res., 60,
138-144.
Hoddinott, P. (1992). ‘Energy minerals (coal)’, Metals & Minerals Ann. Reu., 118.
Hoddinott, P. (1993). ‘Energy minerals (coal)’, Metals & Minemls Ann. Reu., 121.
Hoddinott, P. (1994). ‘Energy minerals (coal)’, Metals & Minerals Ann. Reu. 105.
IlSl (1993). World Steel in Figures, International Iron and Steel Institute.
Lankford, W., Samways, N., Craven, R. and McGannon, H. (Eds) (1985). The Making, Shaping and Treating of Steel, United States
Steel, U.S.A.
Meyers, R. (1984). Handbook of Synfuels Technology, (Eds), McGraw Hill Book Company, U.S.A., Metals and Minerals Annual
Review (1992). ‘Steel industry metals’, 57-64.
Mitsubishi (1991). ‘Apparatus for manufacturing iron by melting and reduction’, European Patent Publication Number: 0 488 429
A2.
Munson, M. (Ed.) (1995). ‘Events and trends in metal and mineral commodities’, JOM, 47, April, 54-58.
Nabi, G. (1993). ‘Methods and apparatus for steelmaking’, Iron & Steel Maker, 71-77.
Oeters, F. and Saatci, A. (Eds) (1987). Mass and Heat Balances During the Reduction oflron Process, Stahleisen, Germany.
Open University (Eds) (1979). Iron and Steel Production, The Open University Press, U.K.
OTA Report (1980). Technology and Steel in dust^^ Competitiueness, OTA, U.S.A.
Report (1987). ‘Technoeconomic Assessment of Electric Steelmaking Through the Year 2000’. Center for Metals Production,
Pittsburgh, Pennsylvania, U.S.A.
Rosenqvist, T. (Ed.) (1983). Principles of Ertmctiw Metallurgy, McGraw-Hill, U.S.A.
Santen, S., Bentell, L., Johansson, B. and Westerlund, P. (1987). ‘Applications of plasma technology in ironmaking’, Plasma
Technology in Metallurgical Processing, Iron & Steel Society, U.S.A.
Scott, D. (Ed.) (1994). Developments Affecting Metallurgical Uses of Coal. IEA Coal Research, England, U.K .
Smith, R. and Corbett, M. (1987). ‘Coal-based ironmaking’, Ironmaking & Steelmaking, 14, 49-76.
Smith, R. and Corbett, M. (1988). ‘Coal-based ironmaking’, Process Technology Conf. Proc., Toronto, Canada, pp. 147-178.
Steffen, R. (1989). ‘Direct reduction and smelting reduction-an overview’, Steel Res.,60, 96-103.
Steffen, R. (1990). ‘The COREX process’, Metallurgical Plant Technol., 20-23.
Steffen, R. and Lungen, H.(1988). ‘Current state of the direct reduction of iron ores to sponge iron’, MetaNurgical Plant & Technol.
18-25.
Steinmetz. E., Steffen, R. and Thielmann, R. (1986). ‘Present state and development potential of processes for the direct reduction
and smelting reduction’, Metallurgical Plant & Technol., 24-39.
1128 T. ZERVAS, J. T. McMULLAN AND B. C. WILLIAMS

Stephenson, R. and Smaller, R. (Ed9 (1980). Direct Reduced Iron, The Iron & Steel Society of AIME. U.S.A.
Strassburger, J. (Ed.) (1969). Blast Furnace Theory and Practice. Gordon & Breach, U.S.A.
Tanaka. K. (1988). ‘Future trends in the technological evolution of the Japanese steel industry’, Iron & SteelMaker, 11-15.
Wright. J. Taylor, 1. and Philip, D. (1991). ‘A review of progress of the development of new ironmaking technologies’, Minerals
Engng, 4, 983-1001.
Zervas,T., McMullan, J. T. and Williams, B. (1996a). ‘Developments in iron and steel making’, Int. J. Energy Research, 20, 69-91
( 1996).
Zervas, T., McMullan, J. T. and Williams, B. (1996b). ‘Gas-based direct reduction processes for iron and steel production’, Int. J .
Energy Research, 20, 157-186 (1996).
Zervas, T., McMullan, J. T. and Williams, B. (1996~).‘Solid-based direct reduction processes for iron and steel production’, Int. J.
Energy Research, 20, 255-278 (1996).