INTERNATIONALJOURNAL OF ENERGY RESEARCH, VOL.

20,69-91(1996)

DEVELOPMENTS IN IRON AND STEEL MAKING

T. ZERVAS, J.T. MCMULLAN AND B.C . WILLLAMS.
Energy Research Centre, Uniwrsi?yof Ulster, Coleraine, Co. Londondeny, BT52 ISA, UK

SUMMARY
A survey is presented of the direct reduction processes which are presently in commercial operation for the
production of iron and steel as an alternative to the classical blast furnace route. The relative advantages and
disadvantages of each technology are presented and their relative energy efficiencies discussed.

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

1. INTRODUCTION
Over the past thirty years there have been two separate lines of development in primary iron-making
technology. The mainstream has centred on the blast furnace, which still remains the principal process
unit for iron production throughout the world. The dominant alternative technologies are based on direct
reduction and smelting, which emerged during the late 1960s and early 1970s, and in which iron oxide
feedstocks are reduced to metallic iron by reducing gases, often at temperatures below the melting point
of iron itself, and avoiding the use of coke. The processes are summarized in Table 1 and Figures 1-4.
Processes which produce iron by reduction of iron ore at temperatures below the melting point of the
product are classified as direct-reduction processes, and the products referred to as direct-reduced iron
(DRI). 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-steel-making 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.
The continued success of the blast furnace as the major primary iron smelting process reflects the very
high levels of thermal and chemical efficiency which can be achieved, but the blast furnace has now
developed to a stage where it is only economically viable at a large scale of operation (typically of the
order of 3 Mty-'1. It also requires considerable associated infrastructure and auxiliary plant, and coke
ovens are needed to provide the high quality metallurgical coke which is the principal if not the sole fuel.
In general, recent developments in blast furnace technology have not led to any radical changes in the
furnace itself, but have involved gradual evolutions in furnace size, burden preparation, increased top
pressure, bell-less charging, improvements in refractory lining and cooling systems, and the application of
automated instrumentation and control technology.
The high capital intensity of a modem blast furnace means that most plants can only recover their
fixed costs when manufacturing at or near full capacity. In addition, the large scale of plant, the high
capital cost and the long construction times make it impossible to respond rapidly to market changes.
This inflexibility imposes a severe constraint on the ability of a blast-furnace-based iron and steel industry
to profit from market opportunities and exposes it to unwelcome commercial risks under depressed
market conditions.
As a result, there is considerable interest in the development of manufacturing routes involving
smaller and more flexible equipment (Smith, et af. 1987,1988). These objectives centre around a small
number of factors.

CCC 0363-907X/96/010069-23 Receiued 18th October 1995

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

Table 1. Commercial Technology Routes

Charge First Second Third Final

stage stage stage stage
process process

Blast Basic Ingot Hot

1 pellets furnace oxygen casting rolling
sinter furnace ingot mills mills

Blast Basic Continuous Hot

2 pellets furnace oxygen casting rolling
sinter furnace mills

Direct Electric Continuous Hot

3 pellets reduced arc casting rolling
iron (coal) furnace mills

Direct Electric Continuous Hot

4 pellets reduced arc casting rolling
iron (gas) furnace mills

Steel Electric Continuous Hot

5 Scrap arc casting rolling
furnace mills

Direct Basic Continuous Hot

6 smelting oxygen casting rolling
iron furnace mills

1.
-
I
BLAST
FURNACE I
I
STELMAKINa
MOLTEN
STEEL *

SCRAP

IRONORE DIRECT STEEL

ARC FURNACE
2' G A S I C O A L d REDUCTION &
-

-
SCRAP
MOLTEN
COAL DIRECT
3. + OXYOEN S
REDUCTTON
IRON SMELTING STEELMAKING
ORE
IRON
Figure 1. Block diagram illustrating various process routes to steel: 1. blast furnace: 2. direct reduction of iron; 3. direct smelting of
iron
 DEVELOPMENTS IN IRON AND STEEL MAKING 71

Figure 2. The classical iron and steel-makingmetallurgical complex

2Step process of Direct Reduction

Figure 3. Two-step process for direct reduction

Reduced cost. The primary opportunity in this area is to undercut the high fixed costs of a classical
blast furnace by reducing the amount of capital-intensive ancillary plant and equipment. One way of
achieving this is to use coal directly as process fuel and reductant, thereby removing the need for coke
ovens and by-product plants. Another possibility is to use fine iron ore directly, so avoiding the cost of
pelleting and sintering plants. The present wide availability of bulk oxygen also offers a way of achieving
high process temperatures, and, since bulk oxygen is normally delivered at high pressure, there is no need
for large compressors.
72 T.ZERVAS, J. T.McMULLAN AND B. C . WILLIAMS

IRON ORE

IRON SF'ONGE

lStep process of Direct Reduction

Figure 4. One-stepprocess for direct reduction

Flexibiliy. It is extremely attractive to develop iron and steel-making plant which is more flexible in
operation than the blast furnace and which can therefore react more quickly to market demands or to
problems of supply. Thus, small scale plant, with low fixed costs and a fairly self-contained process unit is
a beneficial option. Fuel flexibility would also be an advantage, with the attendant ability to use a wide
range of coals from different suppliers and avoid the need for metallurgical grade coking coals.

Small scale. If dependence on ancillary plant can be avoided, it should be possible to operate a small
iron producing plant economically at a capacity of about 500 kt y-'. This creates an opportunity for rapid
incremental expansion of capacity within an existing operation. It could also allow a new operator to
enter existing markets at a lower level of capital investment.

Integration within existing works. Because of existing investment, the large majority of iron and steel
producers are committed to the blast furnace/basic oxygen steel-making process route. Thus, the
attractions of any novel process lie in its suitability to provide small capacity increments and in the
possibility of reducing future dependence on coke ovens and burden preparation plants. Thus, the
product coming from the new plants must be capable of being absorbed into the existing process, must be
similar in temperature and chemical composition to blast furnace hot metal or crude steel. Of the two, an
iron product requires less stringent specification as the downstream processes afford greater scope to
correct and refine its chemical constituency.

Enuironmental factors. Industry is now faced with a continuously tightening circle of environmental
legislation and controls on emissions. In the iron and steel industry, coke ovens are the source of both
atmospheric and ground water pollution, and it is expected that significant investment will be required to
ensure that new coke oven plants comply with regulations. The new generation of coal-based smelting
processes, with reduced agglomeration and the reduction or elimination of coke as a process fuel, are
expected to play a major role in reducing the environmental impact of iron and steel making. Also, the
use of the gas discharged from the melting/reduction furnace as fuel for the generation of electricity
offers further opportunities for emissions reduction.

Alternatiw energy sources. Once the need for coke has been eliminated, it becomes attractive to
consider other solid reactants such as biomass, peat, lignite, wood and paper wastes, and municipal and
industrial wastes. These contain little inorganic matter, such as ash, and few of the elements, such as
sulphur and phosphorus, that are undesirable in steel. They also offer the possibility of improving the
greenhouse gas emissions performance of the iron and steel industry.
 DEVELOPMENTS IN IRON AND STEEL MAKING 73

Direct reduction and smelting technologies appear to offer attractive possibilities for such develop
ments. They are capable of economic operation at much lower outputs than a blast furnace (typically
300-600 kt y-'), are much less capital intensive and require little supporting infrastructure. Direct-re-
duced iron (DRI), the product from direct reduction processes, is physically similar to the ore feedstock
in form (usually pellets or lumps) and contains the gangue minerals present in the original ore. It is
suitable for use as a scrap substitute, for example in electric melting or as coolant in a basic oxygen
furnace, and has quality advantages over scrap because of its low residual content and its ease of
handling and feeding.
Currently, direct reduction plant makes up between three and four per cent of total world iron-making
capacity with about 24 Mt of installed capacity in 1993. Total DRI production in 1988 was about 14 Mt
of which 12.5 Mt were from natural gas based processes and the remainder mainly from coal based units.
A low rate of growth in direct reduction plants in the early 1980s was partly due to restructuring within
the steel industry, but other factors such as worldwide economic recession and the high energy prices of
the late seventies and early eighties played an important role. Today, direct reduction has increased its
share despite a general decline in steel consumption (Chatterjee, 1994).
Direct reduction processes use gas or coal to effect the reduction of iron ore to soiid sponge iron,
instead of the hot molten metal which derives from the blast furnace. Sponge iron normally has a metallic
iron content of approximately 85%, or a 90-95% degree of metallization, the rest being iron oxide and
gangue. The level of tramp and trace elements is determined by the ore charge and can therefore be kept
to a low level. A discussion of the terms used to characterize the products of iron refining is given in the
Appendix (Lankford et at., 1985).
Because of this, most DRI production is used as a substitute for high quality scrap in the electric-arc
steel-making furnace (EAF). DRI is a relatively pure material which dilutes contaminants in the scrap
and improves steel quality. Its use is limited in industrialized countries because of the high availability of
low cost scrap and the high cost of energy. Direct processes tend to be favoured in areas which have
access to suitable iron ores or agglomerates and where there are abundant reserves of natural gas,
non-coking coals or hydroelectric power.
More detailed discussion of direct reduction processes will be given in the next section, but, in
summary, the Midrex and HyL processes have become the standards for gas-reduction applications and
the rotary kiln process has gained acceptance in solid-state reduction. The other main development in the
reduction of iron ore concerns the direct production of liquid iron in the direct smelting process.
Once again, the driving force has been the need for coke in blast furnaces, the metallurgical grades of
coking coals required and the high cost of coke ovens at $300-400 per annual tonne. When the costs of
environmental pollution control measures are added, the break-even selling price of coke from a new
plant is about $140-150 per tonne, which is somewhat higher than the highest FOB prices from existing
plant. Thus, the attraction is evident of direct-smelt reduction technologies based on non-coking coals
(Nijhawan, 1991).
Again, as will be discussed later, a number of potential processes have been proposed, but so far there
are only two technologies which have come to full commercial application. These are the Plasmadust
process developed by SKF in Sweden (Steinmetz et at., 19861, where it is used to process 70000ty-' of
allowed steelworks dust, and the Corex process which was developed in South Africa by Korf Engineering
and Voest Alpine with a capacity of 300 000 t of hot metal per year. The ideal concept of a direct-smelting
reduction process to convert fine ore and coal to liquid product similar to steel has not yet realized.
The main factors influencing the economic analysis of DRI systems are the cost of iron ore, the cost of
energy, the cost of amortizing the investment, and the cost of operating the plant. Of these, energy costs,
which depend on the mass and energy balances for the system, and the investment and operating costs,
which depend on its design and operating characteristics, are the important aspects of current R&D.

2. DIRECT REDUCTION PROCESSES

Direct reduction processes convert iron ore in the form of fines, pellets, sinter, etc. into sponge iron at
temperatures well below the melting point of iron itself. They differ from the conventional blast furnace
74 T. ZERVAS, J. T. McMUUAN A N D B.C.WILLIAMS

process in that solid metallized DRI is produced instead of molten iron, and a wide variety of reducing
agents may be used in place of metallurgical-grade coke. DRI is normally porous, but may be filamentary
or even consist of finely divided particles of iron. It must be further processed to convert it into a usable
end-product (Davis et al., 1982; Lankford et al., 1985; Stephenson et al., 1980).
The primary chemical reactions are the same for all iron ore reduction processes and are outlined in
equations 1-12.
3Fe,03 + H, = 2Fe,04 + H,O (1)
3Fe,0, + CO = 2Fe,04 + CO, (2)

Fe,04 + H, = 3Fe0 + H,O (3)

Fe304+ CO = 3FeO + CO, (4)

FeO + H, = Fe + H20
FeO + CO = Fe + CO,

3Fe + CO + H, = Fe,C+ H,O (7)

3Fe + 2CO = Fe3C + 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 lOOO"C, the dominant reactions are given by equations (1) to (6), the
resultant DRI will be porous and it will have much the same size and shape as the original iron-ore
particles. Metallic iron will also absorb carbon according to equations (7) and (8).
At temperatures above lOOo"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 ecpation (11) is important for processes that produce DRI directly from
solid coal without prior gasification of its fixed carbon.
At about 1200"C, a pasty, porous mass forms, and above this temperature, metallic iron absorbs any
carbon present, according to equation (12), leading to fusing or melting 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-smelting processes operate with product temperatures higher than 1300"C
to ensure that carbon is absorbed rapidly and liquid hot metal is produced.
Direct-steel-making processes must operate with product temperatures higher than 1530"C to produce
molten steel for subsequent treatment.
Regardless of the mechanism of reduction as discussed above, a necessary precursor to any iron oxide
reduction process is the generation of a suitable reductant gas (H,,CO). This reductant gas can be
generated internally in a one-step process within the reductant reactor or it can be produced externally in
a two-step process.
In one-step processes, the gasification-reduction takes place in one reactor. This approach is usually
adopted with liquid or solid fuels in a rotary kiln (Figure 4). Gasification occurs through partial oxidation
and is given by:
 DEVELOPMENTS IN IRON AND STEEL MAKING 75

C,H, + ( n / 2 ) 0, = nCO + ( m / 2 ) H, (13)

In two-step processes the reduction gas can be generated externally from the reduction reactor by one of
the following two methods (Figure 3).

(1) Catalytic steam reforming in which natural gas is reformed either with steam in a bed of nickel
catalyst supported on the refractory base (i.e. alumina)
CH, + H,O = CO + 3H2 (14)
or with the off-gas from the reduction reactor
CH, + CO, = 2CO + 2H2 (15)
CH, + H 2 0 = CO + 3H2 (16)
(2) Partial oxidation for the gasification of liquid hydrocarbons, heavy residuals and coal, which takes
place in an oxygen or air-fired gasifier, separate from the reduction reactor and following the
reaction of equation (13).

Partial oxidation produces less reductant gas than catalytic reforming per unit of fossil fuel (i.e.
according to reaction (14) one mole of CH, yields four moles of reduction gas, in contrast with reaction
(13), which yields three moles if n = 1 and m = 4).
Reactions (14) to (16) are endothermic and take place at temperatures between 850°C and 1000°C.
The heat needed to sustain the reformer reaction is obtained by combustion of the recycle off-gas or by
burning additional fuel. Reaction (13) is exothermic, and, to prevent soot formation, extra oxygen must be
introduced, which adds even more energy to the process.
The one-step processes are predominantly coal-based and, on the laboratory and pilot scales, many
variants exist using rotary kiln reactors, retort reactors and vertical shaft reactors. Commercially,
however, only the rotary kiln reactor is operated successfully in the SL/RN process which is discussed
later.
In principle, the two-step processes lend themselves to use with fossil fuels and their products in
vertical shaft reactors, retort reactors and fluidized bed reactors. A number of such processes exist on the
laboratory and pilot scales. Commercially, however, natural gas-based processes using the two-step
catalytic reforming route are more common because of the larger quantity of reductant gas which is
produced.
On a world-wide basis, 91% of all operating direct reduction plants in 1988 were natural-gas-based
(with 87% using the Midrex and HyL processes, mostly based on vertical shaft furnace reactors, as
discussed later). The remaining 9% are one-step coal-based processes (with 7% using the SL/RN process
and rotary kiln furnaces) (Steffen, 1989).

3. COMMERCIAL DIRECT REDUCTION PROCESSES

The processes adopted for commercial operation can be divided into two classes according to the fuel
used: natural gas or coal.

3.1 Natural-gas-based-processes

3.1.1 Muirex process: The Midrex process (Chatterjee, 1994; Davis et al., 1982; Faucher et al., 1979;
Lankford ef al., 1985; Rouillier et al., 1978; Rosenqvist, 1983; Scott, 1994; Stephenson et al., 1980) was
originally developed by Midland Ross Corporation of Cleveland, U.S.A., and involves three main
components, the shaft reactor where the iron ore is reduced, the gas reformer where the natural gas is
reformed to produce CO and H,, and the cooling gas system. The process is illustrated in Figure 5. It
operates at ambient pressure.
76 T.ZERVAS, J. T. McMULLAN A N D B. C. WILLIAMS

Natural gas is fed to the system and mixed with the off-gas. It is then preheated to 400°C and
introduced to the reformer where the gas is reformed according to equations (15) and (16). The reformed
gas at a temperature over 900°C is then introduced to the bottom of the reduction zone of the shaft
reactor. Iron ore is charged continually at the top of the shaft and falls in a counter-current flow relative
to the rising reductant gas. Initially, the prereduction of haematite to wustite takes place according to the
reactions (11, (21, (31, (41, after which the descending prereduced wustite is reduced to iron in the
reduction zone of the shaft reactor according to reactions (5) and (6). The iron sponge is cooled in the
lower part of the shaft reactor (cooling zone) and discharged at the bottom.
A recent development of the process is the production of hot briquetted iron that is more suitable for
storage, handling and transport, by elimination of the cooling zone of the shaft reactor and connecting
the reactor to a hot iron briquetting press.
The Midrex process has had great commercial success. Starting from a pilot plant of 20t capacity
which started operation in 1967, followed in 1969 by two plants of 400kty-I capacity (in Georgetown,
South Carolina, U.S.A., and in Hamburg, Germany), the number of plants has grown so that it has been
estimated that in 1986 54% of direct-reduction process plants for the production of sponge iron are used
in this process. Between 1977 and 1990, 25 new plants were built, increasing the.tota1 capacity to over
20 Mt.

Equipment. The Midrex process uses a counter-current shaft reactor through which the iron ore
descends continuously under gravity and is heated and reduced by the rising reforming gas. The shaft
furnace reactor comprises two zones, an upper reduction zone which is insulated by refractory lining
bricks, and a lower, conically shaped, uninsulated cooling zone. The raw materials are introduced at the
top of the reactor using a continuous-feed hopper. The preheated reforming natural gas is fed into the
shaft furnace reactor at a point one-third up from the bottom of the upper reduction zone. In the lower

F&03
wante g u I
I
1 I
i I /

PROCESS
Figure 5. Simplified diagram of the Midrex process
 DEVELOPMENTSIN IRON AND STEEL MAKING 77

cooling section, the descending hot sponge iron is cooled with the help of recirculated gas from the gas
cooling system.

Operation. In the Midrex process the reductant gas is generated in the gas reformer unit by reforming
natural gas and partial recycled furnace off-gas. The reducing gas is a 95% mixture of hydrogen and
carbon monoxide containing 55% H, and 33% CO. The reformed gas is introduced to the shaft furnace
reactor at a temperature of about 900 "C.
The reductant gas is taken off at the top of the furnace. The off-gas passes through a scrubber where it
is cooled and scrubbed to remove dust particles. About 70% of it consists of a mixture H, and CO. The
largest portion of the cooled and cleaned off-gas is compressed, enriched with natural gas, preheated to
400°C and piped into the reformer catalyst tubes (nickel catalyst supported on an alumina refractory
base) where it is converted to a mixture of hydrogen and carbon monoxide according to equations (15)
and (161, and is then sent to the vertical shaft furnace reactor.
The iron ore, which may be either totally pelletized or a mixture of pellets and lumps, is fed to the top
of the furnace through a conical hopper. The ore descends under gravity and is successively preheated,
prereduced and reduced in the upper part of the furnace reduction zone, following equations (1) to (6).
The iron sponge is cooled to below 50°C in the lower cooling zone by recycled cooling gas and, because
the hot DRI leaving the reducing zone is a good reforming catalyst, a controlled amount of natural gas
may be added as the reforming reactions (14) to (16) are endothermic and help to cool the hot
direct-reduction iron. The cooling gas is introduced at the bottom of the cooling zone and leaves at the
top, where it is passed through a scrubber for cooling and cleaning. After the gas is cooled, it is
compressed and passed through a demister. The complete reduction of iron ore to iron sponge with 92%
to 93% metallization takes about six hours.
Combustion of the remaining off-gas provides the heat needed for the reformer. Waste gas from the
reformer provides supplementary heat to preheat the combustion air, the natural gas and the recycle
off-gas before introduction to the reformer unit.

Aduantages. The advantages of the Midrex process may be summarized as follows.

(a) Initially, the Midrex used iron ore pellets as raw material. However, it then developed to
accommodate lumps, and it is now possible to use a mixture of 30% to 50% lumps and 70% to 50%
pellets. The addition of lumps reduces the operating cost of the process because of lower feedstock
costs. It also increases the process efficiency by reducing the sticking problems and allowing higher
operation temperatures.
(b) Energy intensity has improved by about 25% over the last 25 years from 12.5-14GJt-' of DRI to
today's average of 9.5-10.5 GJ t-' of DRI.
(c) As air is used instead of oxygen, no oxygen production unit is needed and both investment and
operating costs are reduced.
(d) Recent developments in the Midrex process have allowed the cooling zone of the shaft furnace
reactor to be eliminated, with direct connection of the reactor to a hot iron briquetting press. This
produces hot iron briquettes (HIB) which are more suitable for storage, handling and transport,
and are much less prone to reoxidation than iron sponge, particularly when water is present.
(e) Midrex process plants can be built for a range of capacities from 100-1200kty-', but are now
-'
usually constructed in 440 ky year modules.
(f) Lastly, Midrex process plants can be constructed with comparative low capital investment costs and
operation costs.

3.1.2 HyL process: The HyL I process was developed in Maunder, Mexico in the early part of 1950s by
Hojalata y Lamina SA (HyLSA) in an effort to find an alternative method of producing iron (Chatterjee,
1994; Davis et al., 1982; Lankford et al., 1985; Scott, 1994; Stephenson et al., 1980). Starting from a pilot
78 T.ZERVAS, J. T.McMULLAN AND B. C. WILLIAMS

plant in 1955, followed in 1957 by a prototype plant with four fixed bed reactors, a commercial HyL plant
was built in 1960. Continued research and development led to the improved HyL I1 process which has
greater flexibility in handling a range of oxide pellets and lumps. It is illustrated in Figure 6.
In principle, the HyL process is not very different from Midrex. The main differences lie in the
reformer, which uses catalytic steam reforming instead of off-gas from the reduction reactor, and in the
reactor, in which the feedstock is held stationary in a number of fixed beds and the reductant gas feed is
circulated between the beds according to their requirements. A mixture of CO and H, is still needed as
in all direct reduction processes.
The HyL I1 process retains the static fixed bed design of the prototype HyL I process, but makes more
efficient use of the reductant gas by introducing high temperature tubes into the reductant gas heating
furnace. These allow the reductant gas to be heated to high temperatures and reduce the number of
heating furnaces from four to two.
The heart of the HyL process is a module consisting of four static fixed bed reactors which are cycled
through the various treatment steps (phases) for the reduction of iron ore. Three of the beds are in
operation while the fourth is engaged in charging and discharging. The other major component is the gas
reformer in which natural gas is catalytically reformed using steam to produce CO and H, according to
the equation (14), in contrast with the Midrex process which uses off-gas for the catalytic reforming
according to equations (15) and (16).
The production of iron sponge is performed in three steps, each of three hours duration. In the first
step, the iron ore charge is heated and prereduced using reductant gas coming from another reactor in
which the reduction itself takes place. When prereduction is complete, the fixed bed reactor changes its
role to the second step of reduction by being fed with gas with a richer reductant composition. Once most
of the reduction has been completed, the fixed bed reactor changes its role again and passes to the stage
in which cooling, carburization (Fe,C) and final reduction takes place and the iron sponge is ready to be
discharged.
A system of automatic valves permits each reactor to be connected in any desired order, so the
reactors can be switched from one process step to another when required. The gas used in the third step
comes directly from the reformer, then to the second step and then to the first step.
Pre-reduction and reduction take place according to reactions (1) to (6) in the first and second steps.
The reductant gas is first heated in a gas fired furnace to about 815 "C and is then further heated to about
1050 "C by direct injection and combustion of hydrocarbons before being introduced to the fixed bed
reactor.

, I

HyLPRocrss
Figure 6. Simplifieddiagram of the HyL process
 DEVELOPMENTS IN IRON AND STEEL MAKING 79

The advantage of the HyL process lies in its high temperatures in the pre-reduction and reduction
steps, which improve the reduction efficiency of the hydrogen rich reductant gas. In the third cooling
step, the temperature is about 550°C and carbon is added as Fe,C to the reduced product according to
reactions (7) and (8).
The reductant gas from the reformer ( 75% H,, 14% CO, 7% CO,, 4% CH,) is introduced to the
fixed bed reactors in their third cooling step mode. The operation takes place at a pressure of 4.5 bar,
uses lumps or pellets and has an energy intensity of 11-12.5GJ/t of direct reduced iron.
In spite of their early success as the first direct-reduction processes, HyL and HyL I1 are no longer
competitive either in terms of energy consumption, or in terms of operation and investment costs. The
process has now been replaced by the more successful HyL I11 process, which was developed in 1979.

3.1.3 HyL ZZI Process: In 1979, after many years of research, HyL 111, a new direct reduction process, was
developed from the original HyL process of HyLsA of Maunder, Mexico (Chatterjee, 1994; Dancy, 1990;
Lankford et al., 1985; Quintero et al., 1989). In the HyL I11 process, as shown in Figure 7, the four static
fixed bed reactors are replaced by a single vertical shaft furnace moving bed reactor. The same catalytic
gas reformer, gas heater and auxiliary equipment are retained. A HyL I11 plant began production in 1980
with a capacity 275 kty-'of iron sponge.
The process involves three main components, the shaft reactor where the iron ore is reduced, the gas
reformer where the natural gas is reformed to produce CO and H,, and the reductant gas heater. The
vertical shaft reactor operates on the same principles as the Midrex vertical shaft reactor. The differences
from the Midrex process lie in its high pressure operation and its use of catalytic steam reforming instead
of off-gas from the reducing reactor. Iron ore in the form of lumps/pellets is charged through a charge
sealing mechanism at the top of the shaft and falls in a counter-current flow relative to the rising
reductant gas. Initially, pre-reduction of haematite to wustite takes place according to reactions ( 0 , (21,
(3), (4), after which the descending pre-reduced wustite is reduced to iron in the reduction zone of the
shaft reactor according to reactions ( 5 ) and (6). The iron sponge is cooled to below 50 "C in the lower part
of the shaft reactor (cooling zone) by recycled cooling gas and is discharged at the bottom through a
discharge sealing mechanism. The cooling gas is introduced at the bottom of the cooling zone and leaves
at the top, where it is passed through a scrubber for cooling and cleaning.

H Y L I I I P R ~
Figure 7. Simplified diagram of the HyL 111 process
80 T.ZERVAS. J. T. McMULLAN AND B. C. WILLIAMS

Reductant gas is generated in the reformer in a manner similar to the HyL process as described
previously. It is heated at 950°C in the gas heater section, is then introduced to the vertical shaft reactor
in the lower part of the reduction zone and is taken off at the top of the furnace. The off-gas is passed
through a scrubber where it is cooled and scrubbed to remove dust particles. The HyL I11 process
operates at between 4 to 5 atm and uses iron ore lumps or pellets or a mixture of both. Its energy
intensity is 10-11GJt of DRI.

Aduantuges. The advantages of the HyL I11 process are as follows

(a) The investment cost is decreased because the high pressure operation increases the rate of reaction
and reduces the size of the reactor.
(b) The process can operate with 100% pellets or 100% lumps or any mixture of both; this increases
the flexibility of the process and reduces the operating costs.
(c) HyL I11 is more energy efficient than the original HyL process,
(d) As the reductant gas is richer in hydrogen than in the Midrex process, it can operate at higher
temperatures, above 950 "C. As a result, reduction efficiency is increased.
(el As air is used instead of oxygen, no oxygen production unit is needed, with consequent savings in
both investment and operating costs.
(f) HyL I11 process plant usually can be built for a range of capacities from 200 to 50ky-'. A plant was
constructed by HyLsA and MAN GmbH for SICARTSA, on the Pacific cost of Mexico, with a
capacity of 2Mty-'. It began operation in August 1988. The plant consists of two independent
modules of two shaft reactors each.
(g> Lastly, HyL I11 process plants can be constructed with comparatively low capital investment costs
and operating costs.

3.1.4 FIORprocess: The FIOR process was developed by Esso Research & Engineering Company, U.S.A.,
with cooperation from Arthur D. Little Inc. (Chatterjee, 1994; Dancy, 1990; Davis et al., 1982; Lankford
et uL, 1985; Stephenson et al., 1980). In 1962 Esso built a pilot plant at Baton Rouge, Louisiana, U.S.A.,
and in 1965 a demonstration plant was built at Dartmouth, Nova Scotia, which led to the erection in 1976
of a commercial plant with a capacity of 400kty-' of DRI briquettes at Puerto Ordaz, Venezuela.
The FIOR process (see Figure 8) involves three main components, a module of four fluidized bed
reactors erected on a tower structure to take advantage of gravity flow of solids in which the iron ore is
reduced; the gas reformer where the natural gas is reformed to produce CO and H,, and a briquetting
section in which the reduced iron ore fines are compressed into a dense, compact, pillow-shaped product.
The ore is charged continuously into the preheated reactor where it is raised to the reduction
temperature of 800°C by combustion of natural gas. As the iron ore is fluidized and preheated by the
combustion products it overflows into the highest of the three fluidized bed reduction reactors whence, in
a fluidized state, it flows downwards through each of the three fluidized bed reductant reactors being
reduced by the circulating reductant gas which is flowing upwards, counter-current to the descending iron
ore.
The reduction takes place in three stages. In the first stage, in the first reactor at a temperature of
about 700"C, the ore is reduced by about 10% before it overflows into the second reduction reactor
(second stage of reduction) and hence to the third reactor where final reduction take place (third stage of
reduction). From the third reactor the 91% to 93% metallized iron fines are discharged through a
depressurizing device and enter a briquettes feed drum where they are compacted using double roll
presses at 650 "C.
The natural gas is passed through a reformer and shift reactor to produce a reductant gas with a
composition of 90% H,, 10% CO. The natural gas is reformed with steam according to equation (14) and
the reformed gas is reacted with H,O in the shift reactor according to the equation (CO + H,O = H, +
 DEVELOPMENTSIN IRON AND STEEL MAKING

IUCYCLE GAS SCRUBBER

-
BRlQWWIWtFlELDDRUM
81

+;om ORE PPEEAT

REDUCED SOLIDSTo
BRIQIJETIlNG

AIR

REDUCING REACTORS

(
1 -
FIOR PROCESS REDUCING GAS FIJRNAtX

Figure 8. Simplified diagram of the FIOR process

CO,). The reductant gas is compressed and introduced to the bottom of the third (lowest) reductant
fluidized bed reactor.
The energy consumption is about 16.7GJ t" of direct reduced iron. The major advantage of the
process that it uses iron ore fines instead of pellets, thereby avoiding the cost of pelletizing.

3.2 Coal-basedprocesses

3.2.1 SL/RNprocess: The original process was developed by the Republic Steel Company and National
Lead Corporation in Norway between 1920 to 1930 for the recovery of titanium oxide from titanium ore,
and was labelled the RN process (Bold et at., 1977; Chatterjee, 1994; Dwis et al., 1982; Lankford et al.,
1985; Scott, 1994; Stephenson et al., 1980). In 1964 Lurgi obtained the patent rights for the process and
with collaboration from the Steel Company of Canada has developed it into a process for the direct
reduction of iron that today is called the SL/RN process (see Figure 9). A 36Okty-' plant began
operation in 1975 in Ontario, Canada, following another more specialized plant which was established
early in 1970 at Auckland, New Zealand.
The main process equipment is a cylindrical refractory lined rotary kiln 4 to 6m in diameter and 60 to
150 m in length. The kiln slopes downward from the charge to the discharge end. Haematite and coal are
introduced at the upper end and iron sponge is discharged from the lower end. Air and coal are
introduced through a burner at the lower end to provide the necessary heat. Air for combustion is also
introduced from secondary ports.
The SL/RN process is a one-step process. Gasification and reduction take place in the same rotary
kiln reactor. To preheat the charged materials to the reaction temperature of about 980"C, any
convenient fuel can be used for the start up of the operation. There are two phases in the process, a
preheat one that takes place in the preheat zone, usually 40% to 50% of the kiln length, and a reduction
phase that take place in the reduction zone where the temperature is between 1050 and 1100°C.
82 T.ZERVAS, J.T. MCMULLAN AND B.C. WILLIAMS

BupNPRocIsr3
Figure 9. Simplified diagram of the SL/RN process

Reduction begins after carbon gasification according to the Boudouard reaction (equation (9)). Any
excess fuel is burnt at the discharge end. A charge of iron lumps or pellets, limestone or dolomite, and
coal or lignite is introduced to the upper part of the rotary kiln and discharged from the lower part to a
rotary cooler via a transfer chute.
The operation takes place at atmospheric pressure and consumes about 14.6 GJ/t of DRI. The
advantages of the SL/RN process may be summarized as follows.

(a) The high flexibility of the process with regard to the type of raw materials, the use of pellets and
lumps in any proportions.
(b) The possibility of using minerals that cannot be used in the blast furnace, e.g. titaniferous
ironsands at the Auckland, New Zealand, plant.
(c) Flexibility of the process with regard to the type of fuels and reductants, e.g. lignite, char, low
temperature coke, coke breeze, and anthracite coal have all been used satisfactorily.
(d) As air is used instead of oxygen, no oxygen production unit is needed and both investment and
operating costs are reduced.
(e) Lastly, SL/RN process plants can be constructed with comparatively low capital investment and
operating costs.

4. DIRECT SMELTING PROCESSES

4.1 COREX process

The COREX process (Figure 10) is the only commercial process that produces molten iron by direct
reduction of iron ore (Chattejee, 1994; Delport, 1991; Feichtner et al. 1989; Scott, 1994; Steffen, 1990).
The process was initially developed as the K-R process by Korf Engineering and Voest Alpine in 1978.
 DEVELOPMENTSIN IRON AND STEEL MAKING 83

After successful operation of a pilot plant at Kehl/Rhine, Germany, in 1981, a plant of 300kty-' year
capacity was constructed at ISCORs Pretoria works and started successful operation in 1989. The
start-up was very successful and about 270 thousand tomes 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 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-gasifier
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 from about 900°C to 1300°C when it is instantly dried and
degasified. Generation of the reducing gas is done in a fluidized bed, by partial oxidation of coal
according to equation (13).
The flow velocity is selected to maintain a fluidized bed at temperature between 1500"C and 1800"C.
The reductant gas has a high CO content of about 85% and, after leaving the melter-gasifier, it is cooled
to a temperature of 800 to 900°C by mixing with cooling gas and is pre-reduced 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 back to the gasifier.
The reductant gas at a temperature of 800 to 900°C is fed to the reduction furnace and flows
counter-current to the descending iron ore, escaping from the top end from where it is recycled back to
the process or is exported to be used for electricity generation or the production of chemicals.
Partial reduction takes place in the reduction vertical shaft, and the partially reduced iron is
continuously charged into the melter gasifier at 800 to 900"C, where further reduction and smelting takes

COAL

Figure 10. Simplified diagram of the COREX process

84 T.ZERVAS. J. T. McMULLAN AND B. C. WILLIAMS

place. Molten iron and slag are discharged by conventional tapping from the gasifier melter. The hot
metal tapping temperature is about 1400 to 1500°C and uses similar techniques as for the blast furnace.
The tapping time is between 150 and 180min.
The degree of metallization of direct-reduced iron from the vertical shaft furnace reactor is about 95%
on average. The carbon content varies from 2 to 5%, depending on the raw material and operation
conditions. The energy consumption is about 13.4GJ t-' molten iron.

Aduantages. The advantages of the COREX process may be summarized 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 environment-friendly as it avoids the production of
coke.
(c) It generates a large quantity of process gases that have a high utility value.
(d) In cases where the export gas cannot be utilized economically, the off-gas can be recirculated, so
reducing coal and oxygen consumption.
(e) Lastly, the COREX process can be constructed with a comparatively low capital investment.

5. ALTERNATIVE ROUTES OF IRON AND STEEL MAKING

5.1 Plasma technology in iron making

Plasma technology has been used for many years in steel refining and melting (Fey, 1985; Eriksson et al.,
1988; Santen et al., 1988) but, with the exception of the PLASMADUST process, it has not been applied
commercially to the direct reduction and smelting of iron ore. Despite a considerable research and
development effort over the last twenty years other projects have never passed the pilot stage.
Three main areas have been developed. The application of plasma in blast furnaces, in direct
reduction and in direct reduction/smelting.
In processes based on plasma technology, the 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. The plasma provides an intense, concentrated, source of energy for en-
dothermic reactions and has a high energy efficiency as long as primary energy is igfored and only net
energy is considered for comparison with other processes.
Because plasma technology provides an intense and concentrated energy source, the reactions are very
fast and the reactor size is decreased for the same capacity of product. As a result, the investment cost
decreases.
In spite of the advantages of reduced energy requirement and reduced investment, plasma-based
processes have not yet been commercialized because of the use of electrical energy for generation of the
plasma generation. This is only economic in countries which have access to cheap electricity from
hydroelectric power stations. The only commercial application is the PLASMADUST process from SKF
-'
steel. A plant of 70 kt y was erected between 1982 and 1984 at Landskrona, Sweden. The process allows
recovery of iron from various waste oxide dusts.
Thus, in conclusion, research and development into plasma processes have concentrated on replacing
metallurgical coke with less expensive coal, on using lower quality iron minerals, and on building plants
with smaller capacities.

5.2 Direct steel making

For many years there has been a continuous effort within the steel industry aimed at replacing the
several operations of steel making in order to produce steel directly from iron minerals (Aukrust, 1992;
Fruehan, 1990; Nabi, 1993; Tanaka, 1988).
The conversion 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 coke
 DEVELOPMENTS IN IRON AND STEEL MAKING 85

making, blast-furnace and oxygen steel making (refining) uses three reactors and the alternative route of
direct reduction (coal-based) and electric-arc furnace (refining) uses two reactors. Finally, the new
alternative route, of which there are still no commercial examples, tries to convert iron minerals to steel
in one reactor (divided or not) in which the four steps of gasification, reduction, smelting and refining all
take place.

The advantages of direct steelmaking are the following.

(a) Steel is produced in one reactor by replacing the classical coke making, blast-furnace and oxygen
steel making route.
(b) Steel is produced in one reactor by replacing the alternative direct-reduction/smelting and
electric-arc furnace route.
(c) The direct steel making process can generate a large quantity of process off-gases that have a high
utility value for electrical energy generation or the production of chemicals such as methanol.
(d) As air is used instead of oxygen, no oxygen production unit is needed and both investment and
operating costs are reduced.
(el Environmental problems are decreased.
(f) Any grade of coal, any alternative reductant/fuels and any type of iron mineral can be used. This
increases flexibility.
<g, Lastly, a direct steelmaking process can be constructed with comparatively low capital investment.
In spite of all these advantages, no commercial plant has yet been constructed. An experimental pilot
plant (A.I.S.I. process in the U.S.A.) claims to have succeeded in producing semi-steel directly from iron
ore by using only one reactor.

5.2.1 Iron carbide process: The iron carbide process was developed by F. Stephens in company Iron
Carbide Holding Ltd. (Chattejee, 1994; Munson, 1995; Scott, 1994). A 25 tonne per day pilot plant began
operation in 1989, after which a commercial plant started in Trinidad during 1995. In this process iron
ore fines are reduced and iron carbide is produced in a fluidized bed reactor using a mixture of hydrogen
and carbon monoxide. The reduction of iron ore fines (haematite) takes place in a fluidized bed reactor
at a temperature of 600°C and a pressure 1.8atm. The iron carbide is discharged with a typical
composition of Fe,C 93%, Fe,O, 4% and 3% gangue, from Canadian concentrate, and is transferred to a
basic oxygen furnace (BOF) at a temperature of 500 "C. The iron carbide can be used as a 40% substitute
for molten iron produced in a blast furnace and as a 100% substitute for scrap in a basic oxygen furnace
operation.

6. RAW MATERIALS AND REDUCTANTS/ FUELS FOR DIRECT REDUCTION

The characteristics of raw materials and reductants/fuels dictate the type of reactor used in the process
and impose some practical constraints on the operation of the reactor.

6.1 Reductants /fuels for direct reduction

The type of reductants/fuels, their availability and their price usually determine the character of the
process. There are two categories of reductant fuels:

(1) Gas;
(2) Solid.

These two categories create alternative technology routes to the direct reduction of iron and
determine the type of reactor used in the process. The gas route normally uses the shaft reactor, the
86 T.ZERVAS, J. T.McMULLAN AND B.C. WILLIAMS

fluidized bed and the retort reactor. The solid route normally uses the rotary kiln reactor. Through
gasification, solid reductants can also be used and follow the gas route for iron reduction. The excess heat
requirement of the process can be provided by the combustion of fuel or by electrical energy (e.g. in a
plasma). Energy recovery can be achieved by combusting a portion of the off-gas in the reactor by
injecting oxygen/air into the reactor. The degree of post- combustion (Fruehan et al., 1989; Wright et al.,
1991) is defined by the equation
100(CO,% + H,%)
Post combustion =
(CO,% + H,O% + CO% + H,%)

6.2 Raw materials for direct reduction

The designer of a direct-reduction process plant must take into consideration the characteristics of the
raw materials (size distribution, strength property and chemical properties) that it is intended to use
because this is related to the type of process (rotary kiln reactor, vertical shaft reactor, fluidized bed
reactor and retort reactor). The characteristics of the iron ore determine the type of the reactor and vice
versa as follows.

(a) If fines iron ore is used in a fluidized bed reactor process, the iron content in the mineral should be
at least 64% (e.g. the FIOR process).
(b) Shaft reactors can accommodate a mixture of pellets and lumps, but there is a maximum allowable
percentage of lumps for efficient operation. Optimum efficiency is obtained when the charge to the
furnace consists of 100% pellets, but in practice sticking problems are reduced if a small
proportion of lumps is included. The iron content in the mineral should be at least 65%.
(c) Rotary reactor-based processes are more flexible and is possible to use lumps and pellets in any
proportions with an iron content 54-69%. The basic features of the process are that any grade of
coal is mixed with any type of iron ore, including low iron content minerals (e.g. the SGRN process,
one example of which is in operation in New Zealand using titaniferous iron sands).
(d) Retort reactors use a mixture of lumps and pellets with 66% minimum iron content (e.g. the HYL
process) (Stephenson, et al., 1980).

7. CRITERIA FOR EFFICIENT OPERATION OF D I R E f f REDUCTION AND DIRECT

SMELTING PROCESSES

7.1 Gas based processes

In the gas processes the energy efficiency is increased by certain actions (Oeters et al., 1987):

(a) the gas should flow in a counter-current manner (shaft reactors, etc.) for high gas utilization;
(b) the gas should be operated with complete gas recycling and CO, scrubbing;
(c) the gas mixture(H,,CO) should have a high HJCO ratio;
(d) the use of waste heat for electricity generation (Mitsubishi, 1991).

The necessary heat requirement can be met by:

(i) burning fuels or, more often, by combustion of excess reactor gas with oxygen/air in the
combustion chamber of the reformer (Midrex process);
(ii) electrical energy where cheap electricity is available usually from hydroelectric power stations.
(iii) thermal energy from a nuclear power station.
 DEVELOPMENTS IN IRON AND STEEL MAKING 87

7.2 Coal-Based Processes

In coal-based processes, energy efficiency is increased by following certain guidelines (Oeters er al., 1987).

(a) The gas should flow in a counter-current manner for high gas utilization.
(b) The experience of years of optimal operation of blast furnaces must be applied to direct smelting
processes, particularly during reduction.
(c) The process must operate with minimum off-gas for high productivity.
(d) A high degree of post combustion is necessary for optimal pre-reduction, reduction and smelting,
and for high energy efficiency.
(e) The use of waste heat for electricity generation.

In the coal based processes the necessary heat requirement for gasification of carbon and smelting is
provided by the following:

(i) By combustion of fuels;

(ii) By electrical energy where cheap electricity is available usually from hydroelectric power stations.

8. ADVANTAGES OF DIRECT REDUCTION PROCESSES

Direct reduction processes are an alternative to the blast furnace for the production of iron from iron ore
and allow substitution of iron scrap in the manufacture of steel products in the electric furnace. Not all
the advantages of using direct reduction would apply to every plant because there is a wide variety of
direct reduction processes from which to choose, and each has its particular advantages. The advantages
of the direct reduction of iron processes may be summarized as follows.

(a) Direct reduction plants can be constructed in a range of different capacities, for use with a variety
of raw materials and reductants/fuels.
(b) Direct reduction iron is an ideal substitute for iron scrap in the electrical-arc furnace where, unlike
scrap, it has a known, uniform composition.
(c) Direct reduction iron can be charged continuously to the electric-arc furnace.
(d) Direct reduction iron can be used as a feed material for the electric-arc furnace, as a supplement
to iron ore for the blast furnace and as a cooler for basic oxygen furnaces.
(e) Direct reduction iron processes use coal instead of metallurgical coke and avoid the need of
construction of the coke oven that is necessary for the blast furnace. The off-gases do not pose
environmental problems as severe as those of coke oven-blast furnace plants.
(f) The off-gas from direct reduction iron processes can be used to generate electricity or chemicals.
(gl Coal gasification units can supply the necessary gas for direct reduction processes.
(h) As air is used instead of oxygen, no oxygen production unit is needed and both investment and
operating costs are reduced.
(i) Lastly, direct reduction iron plants can be erected with a comparatively low capital investment and
can be operated at competitive costs.

9. ADVANTAGES OF IRON DIRECT SMELTING PROCESSES

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

(a) Flexibility in size is improved because the process is faster than the blast furnace and the reactor
can consequently be decreased in size. Iron ore lumps or fines can be used directly, and there is no
need for pre-treatment processes such as pelletizing or sintering.
88 T.ZERVAS, J. T.McMULLAN AND B. C. WILLIAMS

(b) There is high flexibility in raw materials and reductant fuels because iron ore fines, which are
difficult to use in a blast furnace, can be freely employed, and iron sand, scrap or any mixture of
these materials can also be used. As a result, there is a great choice of cheap raw materials with
the smelting reduction process. Concerning reductants/fuels, non-metallurgical coke, low-grade
coal anthracite and lignite can be used.
(c) There is flexibility in the operation because the direct iron smelting process can be stopped and
started with relative freedom.
(d) There is flexibility in the use of the off-gas because of its high calorific value.
(e) Iron direct smelting processes use coal instead of coke. As a result, the capital investment is
reduced, as there is no need to construct a coke oven
(f) Lastly, iron direct smelting process plants can be constructed with comparatively low capital and
operation costs.

10. COMPARISON OF DIRECT REDUCTION WITH DIRECT SMELTING PROCESSES

Direct reduction and direct smelting processes have many similar advantages. In addition, direct smelting
processes have the advantage of allowing the use of different raw materials and reductants/fuels, and a
smelting furnace can also act as a gasifier (Gudenau et al., 1989).
Direct smelting processes usually employ a vertical shaft furnace for pre-reduction of the iron ore
(COREX process). The advantage is based on the fact that the charged iron ore does not need to be
agglomerated. This decreases the investment and operation costs.
Direct reduction has a major disadvantage in that the product is solid sponge iron which has to be
protected against re-oxidation. Melting of the solid sponge requires about 2 GJ of electric energy per
tonne, so that the process is uneconomical in many countries. Additionally, sponge iron has to compete
with scrap which is often cheaper and more readily available (Eketorp, 1987).
The sticking of iron is one of the major problems during the production of sponge iron, but this has
been solved in the direct smelting process as the iron ore is pre-reduced in the reductant shaft furnace
and avoids metallization there, as final reduction, metallization and melting take place in the smelting
furnace (COREX process).
Commercial direct reduction process plants use air instead of oxygen. COREX, the direct smelting
process, uses oxygen.

11. COMPARISONS IN ENERGY CONSUMPTION

Comparison of the energy consumption of the various process routes should be based on molten iron as a
comparable uniform product. When comparing the gas and solid material reduction processes, the energy
consumption for the gas process is always lower. This can be off-set, however, by the price of cheap
non-cokeable coal compared to natural gas. The vertical shaft reactor process (Midrex, HyL 111) has
achieved lower energy consumptions than the rotary kiln reactor process (SL/RN).
Comparison of the Midrex and COREX processes, as shown in Table 2, shows Midrex to be preferable
in energy terms, It is less so when account is taken of the energy needed to melt the sponge iron. The
new COREX process (using a fluidized bed melter gasifier reactor) is the most promising commercial
process for the future because it uses cheap non-cokeable coal and produces molten iron similar to that
from a blast furnace.

12. CONCLUSION
Direct-reduction iron increased its share of the pig-iron market to about 30Mt in mid-1994, following an
average annual growth of over 9% over the last ten years. The new figure was 15% higher than that of
 DEVELOPMENTS IN IRON AND STEEL MAKING 89

Table 2. Classification of commercial processes by reactor type and energy comsumption.

For comparison of the various process routes, 2GJt-' must be added to allow for the
production of molten iron.

Fluidized Fluidized
Reactor Vertical shaft bed Rotary kiln Retort bed melter
~~ ~~ ~

Process Midrex HyLIII FIOR SL/RN HYLI COREX

Energy (GJ t-l of 9.5 10 16.7 14.6 11 13.4
molten iron)
Pressure (bar) 1 45 10 1 45 5
Product Sponge Sponge Sponge Sponge Sponge Molten iron
Reductant Gas Gas Gas Coal Gas Coal

1992. Direct reduction iron usage is increasing either as a complement to or as a substitute for scrap in
both electric arc and blast furnaces.
New plant proposals have been advanced for the developing countries as stricter environmental
regulations and modernization in the industrialized countries are leading to the elimination of old blast
furnace plants.
As the concept of environmentally friendly processes gains ground, new direct smelting processes such
as the cokeless COREX process are being developed and overcome the earlier capacity limitations. The
first commercial iron carbide process plant is completed, with start-up in Trinidad during 1995 (Munson,
1995).
The Midrex process is still the dominant commercial process as its share is increasing dramatically,
from 38.5% in 1979 to 48% in 1986 to 54% in 1988, to 60% in 1990 and to 67% in 1993. Over the same
period, coal-based plant has fallen from 11.1% in 1979 to 9% in 1988, and the SL/RN process has fallen
from 7% in 1988 to 3% in 1993 (Chatterjee, 1994; Scott, 1994; Steffen et al., 1988; Steffen, 1989;
Steinmetz, 1986).
This can be explained because the gas-based DRI processes are generally easier to control, their
energy efficiency is higher, the product is cleaner, it contains a higher percentage of carbon and a lower
percentage of sulphur compared with DRI from coal based processes in which process control is difficult,
the product contains coal ash and char, and a higher sulphur content related to the sulphur content of
the coal.
The plant capacity of direct reduction iron gas-based shaft reactor can be varied between 100 kt to
1 Mt of sponge iron per year. In contrast to direct reduction, the coal-based rotary reactor has an upper
limit of 150 kt of sponge iron per year.
Energy consumption is higher in coal-based compared to gas-based direct reduction iron processes.
The most successful direct reduction processes are those based on a vertical counter current shaft
reactor. Because vertical shaft reactor processes use a mixture of oxide pellets and lump iron ore, the
cost of raw materials is higher than that for a fluidized bed process (which can use fines). However,
production of the high-H, reducing gas needed for the fluidized bed process requires more natural gas
per tonne of iron sponge than for the vertical shaft reactors, and a fluidized bed plant must include a
pressing unit for hot briquetting, so increasing the operating cost.
The advantage of fluidized bed reactor plant in being able to use fine ore rather than pellets or lump
iron ore, may result in favourable overall economics, particularly where cheap natural gas is available.
However, in the future, cheap hydrogen may be available by electrolysis of water and the use of hydrogen
as a reductant in current and new processes of iron and steel making will increase. A demonstration
project within the Europe-Quebec Hydro-Hydrogen Power Project is intended to include the use of
hydrogen for iron ore reduction in a plasma-arc process (Gretz et al., 1993).
The COREX process is the only commercial process for the direct smelting of iron ore operating
today and plants with capacities between 100 and 6OOkty-' of molten iron can be offered as an
90 T.ZERVAS, J. T.McMULLAN AND B.C. WILLIAMS

alternative both to the blast furnace process and to direct reduction of iron (iron sponge), and particular
to the Midrex process. The COREX process will undoubtedly be the main competitor to today's
dominant Midrex process.
In conclusion, local operating conditions, such as environmental protection legislation, energy costs
and taxes, will play a key role in the choice of the new process. In some countries, these conditions could
rule out the commercial operation of a number of new, technically viable processes. Another important
factor in the selection of processes will be their flexibility in adaptation to meet market needs.
With the possible exception of the developing countries, it is unlikely that any more large blast furnace
plant will be built in the foreseeable future, because of the heavy capital cost, lack of flexibility,
environmental impact and market limitations. At the same time, quality considerations and the availabil-
ity and cost of both scrap and electricity can be important factors in determining the most attractive way
forward.

APPENDIX DEFINITIONS RELATED TO CONVERSION EFFECTIVENESS

The iron and steel industry uses a number of terms which require definition for the more general reader.
These act as measures of the effectiveness of the reduction process. They relate to the impurity content
and the iron content of the ore and the product, and to the chemical and physical state of the refined iron
(Davis et al., 1982; Lankford et al., 1985).

The main parameters and their defining equations are as follows.

Per cent total iron (Fe,). This gives the total amount of iron in a sample. All forms of iron including
metallic iron and iron chemically combined with other elements is included.
Weight of iron
FeT = Total weight of sample
x 100
Per cent metallic iron (Fe'). This refers to iron either chemically uncombined in the sample or combined
as cementite (Fe,C).
Weight of metallic iron
Feo = Total weight of sample x 100
Per cent metallization. This refers to that portion of the total iron which is present as metallic iron. It is
not the same as per cent reduction.
Total weight of metallic iron Feo
%Metallization = x100=100x-
Total weight of iron Fe,
Per cent reduction. This refers to the oxygen that has been removed during the process. It is defined
assuming that all of the iron was originally present as hematite (Fe,O,).

%Reduction =
( %Oxygen )Fez(),-
FeT
%Oxygen
( FeT)
%Oxygen Product
xlOO

( Fe, )Fe203
= Fe,
[1-2.327( %Oxygen )product] 100X

Note. Some investigators define per cent reduction with reference, not to cementite, but to the actual
oxide state of the ore. This defines the amount of reduction performed during the process.

Per cent gangue. This is a measure of the non-iron compounds (SiO,, Al,O,, CaO, MgO, etc.) present in
the product.
 DEVELOPMENTS IN IRON AND STEEL MAKING 91

%Gangue = 100 - %FeO + Feo - %C

Per cent oxygen. This gives the amount of oxygen combined as iron oxides. Oxygen combined with the
gangue constituents is excluded.
%Oxygen = %FeO + Feo - Fe,
Per cent Wustite. This gives the amount of wustite (FeO) present in the sample.
Total weight of FeO 100
%Fee = Total weight of sample

REFERENCES
Aukrust, E. (1992). ‘Results of the AISI/DOE direct steelmaking program’, Int. Symp. on Smelting, Quebec, Canada, pp. 591-610.
Bold, A. and Evanst, N. (1977). ‘Direct reduction down under: the New Zealand story’, Iron & Steel hat., 145-152.
Chatterjee, A. (Ed.) (1994). Beyond the B h t 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-reductiontechnology and economics’, Ironmaking & Steelmaking, 9, 93-129.
Delport, H. (1991). ‘The CORJZX process’, 2nd European Ironmaking Congress, Glasgow, U.K., pp. 289-302.
Eketorp, S. (1987). ‘Energy considerations of classical and new iron and steel-making technology’, Energv, 12, 1153-1168.
Eriksson, S., Johansson, B. and Santen, S. (1987). ‘Applications of plasma technology to steelplant waste treatment’, in Plasma
Technology in Metallurgical Processing, Iron & Steel Society, U.S.A.
Faucher, A., Marquis, A, and Dancy, T. (1979). ‘Review of the sidbec-dosco direct reduction operations’, Iron Steel Maker, 35-40.
Feichtner, H., Maschlanka, W. and Helten , F. (1989). ‘The COREX process’, Shillings’ Mining Reu., 14, 20-27.
Fey, M. (1985). ‘Plasma technology for the steel industry’, Iron & Steel Engineer, 38-45.
Fruehan, R. (1990). ‘Iron bath smelting-current status and understanding’, Elliot Symp. Proc., Cambridge, Massachusetts,. U.S.A,
pp. 1-10.
Fruehan, R., Ito, K. and Ozturk, B. (1989). ‘Analysis of bath smelting processes for producing iron’, Steel Res., 60, 129-137.
Gretz, J., Drolet, B., Kluyskens, D., Sandmann, F. and Ullmann, 0.(1994). ‘Status of the hydro-hydrogen pilot project (EQHHPP)’.
Int. J. Hydrogen Energy, 19, 169-174.
Gudenau, H., Fang, J., Hirata, T. and Gebel, U. (1989). ‘Fluidizedbed reduction as the prestep of smelting reduction’, Steel Res, 60,
138-144.
Lankford, W., Samways, N., Craven, R. and McGannon, H. (Eds) (1985). The Making, Shaping and Treating of Steel, United States
Steel, U.S.A.
Manson, M. (Ed.) (1995). ‘Events and trends in metal and mineral commodities’, JOM, 47, (April), 54-58.
Mitsubishi (1991). ‘Apparatus for manufacturing iron by melting and reduction’, European patent publication number: 0 488 429
A2.
Nabi, G. (1993). ’Method and apparatus for steelmaking’, Iron & Steel Maker, 71-77.
Nijhawan, B. (1991). ‘Current status and commercial potentials of smelt-reduction technologies’, 2nd European Ironmaking Congress,
Glasgow, U.K., pp. 337-346.
Oeters, F. and Saatci, A. (Eds) (1987). Mass and Heat Balances During the Reduction oflron Process, (Eds) Stahleisen, Germany.
Quintero, R. and Elias, F. (1989). ‘HyLIII latest project’, Ironmaking ConfProc, Chicago, USA, pp. 155-160.
Rosenqvist, T. (Ed.) (1983). Principles of Emactiw Metallurgy, McGraw-Hill, U.S.A.
Rouillier, L. and Boote, A. (1978). ‘Second reduction plant at the Sidbec-Dosco Contrecoeur works’, Iron & Steel Eng, 55,29-34.
Santen, S., Bentell, L., Johansson, B. and Westerlund, P. (1987). ‘Applications of Plasma Technology in ironmaking’, in P h m a
Technology in Metallurgical Processing, Iron & Steel Society, U.S.A.
Scott, D. (Ed.) (1994). Dewlopments Aflecting 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 Cont 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’, MetalL Plant Technol., 20-23.
Steffen, R. and Lungen, H. (1988). ‘Current state of the direct reduction of iron ores to sponge iron’, MetallurgicalPlant & 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’, Metall Plant & Technol., 24-39.
Stephenson, R. and Smaller, R. (Eds) (1980). Direct Reduced Iron, (Eds)The Iron & Steel Society of AIME, U.S.A.
Tanaka, K. (1988). ‘Future trends in the technological evolution of the Japanese steel industry’, Iron & Steel Maker,11-15.
Wright, J. Taylor, I. and Philip, D. (1991). ‘A review of progress of the development of new ironmaking technologies’, Minerals
Engng, 4, 983-1001.