See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/305350936

i) Direct Reduced Iron: Production

Chapter · March 2009

DOI: 10.1081/E-EISA-120050996

CITATIONS READS

0 9,415

2 authors, including:

S. K. Dutta
The Maharaja Sayajirao University of Baroda
65 PUBLICATIONS   138 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Development of new alloys View project

Steel plant waste; Low grade iron ore; Coating on MS for corrosion resistance View project

All content following this page was uploaded by S. K. Dutta on 17 September 2016.

The user has requested enhancement of the downloaded file.

 Encyclopedia of Iron, Steel, and Their Alloys

ISSN: (Print) (Online) Journal homepage: http://www.tandfonline.com/doi/book/10.1081/E-EISA

Direct Reduced Iron: Production

Sujay Kumar Dutta , Rameshwar Sah

To cite this entry: Sujay Kumar Dutta , Rameshwar Sah . Direct Reduced Iron: Production. In
Encyclopedia of Iron, Steel, and Their Alloys. Taylor and Francis: New York, Published online: 30
Mar 2016; 1082-1108.

To link to this chapter: http://dx.doi.org/10.1081/E-EISA-120050996

Published online: 30 Mar 2016

Submit your article to this journal

Article views: 3

View related articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=

Download by: [Sujay Kumar Dutta] Date: 29 June 2016, At: 22:29
 Direct Reduced Iron: Production

Sujay Kumar Dutta

Department of Metallurgical and Materials Engineering, M.S. University of Baroda,
Vadodara, India
Rameshwar Sah
R&D and Scientific Services Department, JSW Steel Ltd., Vijaynagar Works, Toranagallu, India

Abstract
The industrial development program of any country, by and large, is based on its natural resources.
Depleting resources of coking coal, the world over, is posing a threat to the conventional (blast
furnace [BF]–basic oxygen furnace [BOF]) route of iron and steelmaking. During the last four
decades, a new route of ironmaking has rapidly developed for direct reduction (DR) of iron ore to
metallic iron by using noncoking coal/natural gas. This product is known as direct reduced iron (DRI)
or sponge iron. Processes that produce iron by reduction of iron ore (in solid state) below the melting
point are generally classified as DR processes. Based on the types of reductant used, DR processes can
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016

be broadly classified into two groups: (1) coal-based DR process and (2) gas-based DR process.
Details of DR processes, reoxidation, storage, transportation, and application of DRI are discussed in
this section.
Cutting—Ductile

INTRODUCTION The second route (direct reduction [DR] ! electric

The blast furnace (BF) has remained the workhorse of
worldwide virgin iron production (i.e., hot metal) for more
than 200 years. Over the years, BFs have evolved into
highly efficient chemical reactors, capable of providing
stable operation with a wide range of feed materials. How-
ever, operation of modern efficient BFs normally involves
sintering and coke making and their associated environ-
mental problems. More than 90% of iron is currently
produced via the BF process, while the rest is coming from
direct reduction (DR) processes, mini blast furnaces (MBFs),
Corex, Finex, Ausmelt, etc. Additionally, the severe short-
age of good-quality metallurgical coal has remained an
additional constraint all over the world. In view of this,
there is an increasing awareness that the BF route needs to
be supplemented with alternative ironmaking processes
that are more environment friendly and less dependent on
metallurgical coal.
There are two major conventional routes for producing
steel. The first route (blast furnace [BF] ! basic oxygen
furnace [BOF]) is an integrated steel plant that consists of
sintering or pelletizing units, coke ovens, BFs, and BOFs.
Such plants require high capital expenses and raw mate-
rials of stringent specifications. Coking coal is needed to
make a coke strong enough to support the burden in the
BF. Integrated steel plants of less than one million tonne
(Mt) annual capacity are generally not economically via-
ble. The coke ovens and sintering plants in an integrated
steel plant are polluting and expensive units. The molten
metal from the BF is subsequently refined in the converter
(BOF) to produce steel.
arc furnace [EAF]) is where recycled/purchased steel
scrap and direct reduced iron (DRI), from DR processes,
are melted in EAFs and further processed into final
products.
During the last four decades, a new route of ironmaking
has rapidly developed for the DR of iron ore to metallic
iron by using noncoking coal/natural gas. This product is
known as direct reduced iron (DRI) or sponge iron. Pro-
cesses that produce iron by reduction of iron ore below the
melting point of the iron are generally classified as DR
processes. DRI is successfully manufactured in various
parts of the world through either natural gas or coal-based
technology. Iron ore is reduced in solid state at 800–
1050 C (1073–1323 K) by either reducing gas (H2 + CO)
or coal. The specific investment and operating costs of DR
plants are low compared to integrated steel plants and are
more suitable for many developing countries where sup-
plies of coking coal are limited.
Steel scrap is an important metallic feed material for
electric steelmaking (EAF/IF), which constitutes 28–30%
of the world’s steel production. The nonavailability of
consistent quality at a reasonable price necessitated the
search for an alternative to scrap for use in secondary steel
sectors. These problems have been overcome with the
help of DRI. DRI is not only a substitute for steel scrap as
a feed material in EAF/IF but also a more suitable melting
stock for good-quality-steel production. It is evident that
DRI has influenced the supply–demand balance for steel
scrap because electric furnace operators regard it as a
high-quality substitute of scrap. DRI has been recognized
as a high-quality, cheaper, and high-purity charge material

Encyclopedia of Iron, Steel, and Their Alloys DOI: 10.1081/E-EISA-120050996

1082 Copyright © 2015 by Taylor & Francis. All rights reserved.
 Direct Reduced Iron: Production 1083

worldwide. In comparison with scrap (residual elements,

Table 1 Direct reduction processes.
which cannot be removed during steelmaking, vary
Type of Type of Type of
between 0.13% and 0.73%), the use of DRI offers consis-
Process reactor ore use reductant Ranka
tency in composition and size, as well as low residual
Midrex Shaft Lump/pellet Gaseous 1
elements (0.02%). A major part of DRI production is used
HyL Retort -Do- -Do- 3
as a substitute of scrap in electric steelmaking furnaces. SL/RN, Rotary kiln -Do- Solid 2
DRI derived from virgin iron units is a relatively pure ACCAR,
material that dilutes contaminants in the scrap and CODIR, etc.
improves the steel quality. HIB Fluidized bed Fine Gaseous 4
The DR process is intrinsically more energy efficient a
Rank is in terms of production and popularity in the world.
than the BF because it operates at a lower temperature.
DRI (containing 90–94% total iron) is an excellent feed-
stock for electric furnaces used by mini steel plants, together with solid reductant (noncoking coal) are charged
allowing them to use lower grades of scrap for the rest of into the reactor. The generation of reducing gas (mainly CO)
the charge or to produce higher grades of steel. Hot takes place in the reduction reactor, and the product has to
briquetted iron (HBI) is a compacted form of DRI be separated from excess reductant, ash, and/or sulfur-
designed for ease of shipping, handling, and storage. absorbing materials (lime, dolomite) by magnetic separation
after discharge at low temperature, which makes product
handling more complicated.[3] Because of the presence of
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016

Definition of Direct Reduced Iron

these substances in DRI, hot briquetting and hot feeding are
not possible for coal-based process. Magnetic separator also
DRI or sponge iron refers to porous iron produced by the
does not work at high temperatures to separate the DRI.

Cutting—Ductile
DR process. The DR process is a solid-state reaction pro-
Coal-based DR processes like SL/RN (1970), ACCAR
cess (i.e., solid–solid or solid–gas reaction) in which
(1973), KRUPP (1973), CODIR, and TDR (1975), using
removable oxygen is removed from the iron oxide, using
rotary kiln as a reactor and coal as a reductant, have been
coal or reformed natural gas as reductants, below the melt-
developed worldwide. India, due to its large reserve of
ing and fusion point of the lump ore or agglomerates of
noncoking coal, has shown keen interest in pursuing these
fine ore. The external shape of the ore remains unchanged.
technologies using such coal as a cheap energy source for
Due to removal of oxygen (about 27–30% reduction in
highest DRI production[4] in the world. These processes
weight occurs), a honeycombed microstructure develops
are highly sensitive to the types of raw materials used.
suggesting the name Sponge Iron.[1] (i.e., solid porous
The following are coal-based processes:
iron; lumps/pellets with many voids filled with air).
DRI is obtained when iron oxide (generally lump iron
1. Rotary kiln–based processes
ore or pellets) is reduced to metallic form in solid state.
SL/RN, CODIR, ACCAR, DRC, TDR, SIIL, OSIL,
Since there is no melting, the external shape is retained.
Jindal
Color changes from red to black. The true density ranges
2. Shaft furnace–based processes
from about 3.5 · 103 to 4.4 · 103 kg/m3. The true density
Kinglor, Metor, NML, Vertical Retort
of pure iron is 7.8 · 103 kg/m3. Thus, there is about
3. Rotary hearth furnace (RHF)-based processes
45–56% reduction in true volume[2] and this is manifested
INMETCO, FASTMET, Comet
in the formation of pores throughout the interior of sponge
iron pieces.
The main advantages of coal-based processes are as
Direct Reduction Processes follows:

Based on the types of reductant used, the DR processes
can be broadly classified into two groups:
1. Using solid reductant, that is, coal-based DR process
2. Using gaseous reductant, that is, gas-based DR process
DR processes are summarized in Table 1.
(i) They do not require high-grade coal that is scarcely
available.
(ii) They use noncoking coal.
(iii) They can be installed at lower capacity.
(iv) They can be easily installed at places where small
reserves of coal and iron ore are available.
(v) Modules of small-scale operation are available.

Coal-based processes Disadvantages of coal-based processes are as follows:

In coal-based DR processes, noncoking coal is used as (i) Lower economy of scale.

reducing agent. In solid reduction processes, iron oxides (ii) High energy consumption (16.0–21.0 GJ/t).
 1084
(iii) Low carbon content in the product (<1.0%).
(iv) Lower productivity (0.5–0.9 t/m3/day).
(v) Hot feeding to the steelmaking furnace and hot
briquetting are not possible due to the presence of
residual char and ash in DRI.
Gas-based processes
Reformed natural gas is used as a reducing agent. Iron ore
lumps or pellets are reduced in the solid state and oxygen
from iron oxide is removed by a gaseous reducing agent.
Natural gas is reformed at 950 C (1223 K), in the presence
of catalysts (Ni or Al2O3), to produce reducing gases CO
and H2. The reducing gases H2, CO, or mixture of H2 and
CO, are introduced into the reactor at elevated tempera-
tures [up to 1000 C (1273 K)] and pressure (up to 5
bars)[3]. If CH4 is present in the reducing gas, it results in
carburization of the reduced product.
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
The processes based on gaseous reduction are confined
to the areas where natural gas is available in abundance
at a reasonable price. Commercialized processes that use
Cutting—Ductile
reformed natural gas as reducing agents are Midrex (1969)
and HyL (II–1957, III–1980).
The following are gas-based processes:
1. Retort processes
HyL I, Hoganas
2. Shaft furnace processes
Midrex, HyL III, Plasmared, Armco, Purofer, NSC,
HyL IV
3. Fluidized bed processes
FIOR, Finmet, Circored
The processes discussed have both merits and demerits.
Gas-based processes have the following advantages:
(i) High productivity
(ii) Lower energy consumption (10.5–14.5 GJ/t)
(iii) Higher carbon content in the product (>1.0%)
Gas-based processes account for 78.8% of the world’s
DRI production; out of that Midrex alone contributed
63.2% in 2013.[5] The gas-based processes offer distinct
advantages over the coal-based processes[3]:
1. Less capital cost: The capital cost per tonne of
installed capacity in case of coal-based plants is 1.8–2
times as high as that of gas-based plants.
2. High productivity: The throughput rates for gas-based
DRI plants are much higher than that of coal-based
plants. The productivity of gas-based plants can be as
high as 11 t/m3/day as against merely 0.5–0.9 t/m3/day
for coal-based plants.
3. Better quality: The quality of the DRI produced in
terms of metallization and carbon content is also
higher for gas-based plants.
Direct Reduced Iron: Production
4. Energy efficiency: The gas-based processes have over
the years improved considerably and are highly energy
efficient.
5. Better plant availability: The gas-based processes are
by now highly standardized. Gas is a clean source of
fuel and there is no problem of ash content, as in the
case of coal, which leads to ring formation and other
attendant problems. Gas-based plants do not suffer
from the maintenance problems as often encountered
in case of rotary kilns, and therefore plant availability
is better.
6. Environmental pollution: Since gas is a clean source of
fuel, it is advantageous from the environmental pollu-
tion point of view.
RAW MATERIALS
DR processes are very sensitive to chemical and physical
characteristics of raw materials used in the process. Iron
ore or pellets, reductant (i.e., noncoking coal or natural
gas), and limestone/dolomite are the main raw materials
for DR technology. For successful operations the process
of DR technology has specified the characteristics of raw
materials to be used in the process.
Iron Ore
Characteristics of iron ore
Lumps or pellets have high iron content, low gangue con-
tent, and good mechanical strength and are readily reduc-
ible and of nondecrepitating variety. Iron ore feed to the
reactors has the following characteristics:
 Chemical composition
 Reduction properties
 Physical characteristics
Chemical composition. The key input material required
for DRI production is the iron ore. Since DRI making is a
solid-state reduction process, the gangue in the ore is
retained in the product. Due to oxygen removal, there
is a reduction in weight by about 30% and the percentage
of gangue material goes up by about 1.4 times[2]. Hence,
it is very important to choose a higher grade of iron ore
input.
In all DR processes, the only noteworthy chemical
change that takes place is the removal of oxygen from
the iron oxide in the charged iron ore. Since no melting
or refining occurs, all impurities in the ore feed get con-
centrated in the reduced product. As a result, in any iron
ore chosen for the DR process, the total iron content
should be as high as possible and the gangue content
should not only be a minimum, but the gangue should
 Direct Reduced Iron: Production 1085

have a composition that is acceptable to the user of the
final product.[4] In general, an ore with iron content higher
than 65% is preferred. The gangue content of the DRI,
that is, its silica plus alumina level, should be as low as
(<5%) possible.
The phosphorus content of the iron ore is extremely
important and should be as low as possible (preferably
below 0.03%) because there is no removal of phosphorus
during the reduction process. Some ores contain phospho-
rous between 0.04% and 0.08%, and if they are used, the
phosphorus content of DRI would be at least 0.055% and
even as high as 0.10%, which is substantially higher than
phosphorus in scrap normally used in steelmaking. Like
phosphorus, the sulfur content of the iron ore has to be low
(below 0.02%), since during reduction there could be
some pickup of sulfur. The sulfur content of many iron
ores available in the world is quite low (0.01–0.02%) and
consequent low sulfur in DRI is a positive feature, partic-
ularly if the phosphorus levels are high.
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
(particularly in the case of pellets). The optimum size of
the iron ore is determined essentially by its reducibility
characteristics. With higher reducibilities, the top size can
be increased. In the case of highly decrepitating-type
ores, a larger size may be necessary so that even after
degradation, the generation of fines (below 3 · 103 m)
is minimum. In the case of lump ores, a size range of
8 · 103–20 · 103 m is normally preferred with a mean
size of around 12 · 103 m, whereas it is the same in the
case of pellets, with 6 · 103 to 22 · 103 m with a
majority of 9 · 103 to 16 · 103 m sized pellets. Fines
in any form are undesirable. In rotary kilns they promote
ring formation inside the kiln, thereby affecting the space
of kiln availability, leading to an increased tendency of the
ore particles to segregate from the rest of the bed. Careful
screening is essential in order to guarantee removal of the
adherent fines that is particularly difficult in the case of
clay bearing lump ores. Poor abrasion property of the ore
can also contribute to generation of fines because of phys-

The specific requirements of lime (CaO) and magnesia ical disintegration arising out of the tumbling action. In
(MgO) in the ore feed, if charged as pellets, are generally order to minimize the generation of fines in the kiln, the
determined by the slag characteristics required for steel- abrasion index of the ore should obviously be as high as

making; up to 2.5% CaO and 1.0% MgO are usually
acceptable. Higher percentages of these constituents (used
as a binder in some pellets) may have an adverse effect on
the reducibility of the pellets. Similarly, though titania
(TiO2) is a slag-forming constituent, in excess amounts, it
may have a deleterious effect on the reducibility thereby
imposing limits on the maximum degree of metallization
that can be achieved in a given reactor. In general, the
titania content should not exceed 0.15% in the ore feed.[4]
Steelmaking grade DRI requires iron ore having a mini-
mum of 65% total Fe, maximum of 4.5–5.0% gangue
(SiO2 and Al2O3), and low sulfur and phosphorous content
(less than 0.05% each).[6]
Cutting—Ductile
possible.
Pellets having a more regular shape and superior reduc-
tion characteristics are definitely superior to lump ores as
the iron oxide feedstock, especially because of a more
uniform size, better chemical consistency, as well as the
guaranteed higher furnace availability. Extra cost of pel-
lets is compensated by smoother operation, higher furnace
availability, higher degree of metallization, and higher net
yield of DRI because lower generation of fines depends on
the relative costs of ore and pellets and would, therefore,
vary from case to case. Tables 2 and 3 show the typical
chemical, physical, and reduction characteristics of the ore
to be used in DR processes.[4]

Reduction properties. The iron ore should have ade- Table 2 Characteristics of iron ore for coal-based process.
quate reducibility and favorable decrepitation character- Chemical constituent Percentage (%)
istics if any DR process has to be techno-economically Fe 62.0 min
feasible. The retention time required to reduce iron oxide SiO2 + Al2O3 7.0 max
to DRI of a specific quality depends, to a large extent, on CaO + MgO 2.0 max
the reducibility of the ore. There exists a direct correla- S 0.03
tion between reducibility and retention time and, hence, P
P 0.08 max
the productivity of any given reactor. The higher the Pb, Zn, Cu, Sn, Cr and As 0.02 max
reducibility, the lower is the retention time required Physical
and, consequently, the higher the productivity. Reduction A. Size range, · 103 m Percentage (%)
degradation is a chemical phenomenon that takes place 5 5 max
during the initial stages of reduction when hematite is +25 10 max
converted to magnetite and expansion occurs. Depending B. Tumbler index Tolerable (%) Preferable (%)
on the type of ore, generation of fines owing to reduc- +5 · 103 m 80 92
tion decrepitation may be as low as 3–4% or as high as 28 mesh 10 5
10–15%. C. Reducibility (40%)
% per min 0.5 0.6
Physical characteristics. The important physical fea- D. Decrepitation, · 103 m
tures of the iron ore feed are its size range, shatter index, 5 20 15
0.5 5 3
tumbling index, abrasion index, and compression strength
 1086 Direct Reduced Iron: Production

Table 3 Characteristics of ore for gas-based process. Table 4 Characteristics of coals for direct reduction processes.
Chemical composition Manufactures’
Weight (%) Pellets Lump ore Criteria Unit ISO standard experience
Fe (total) 67–69 66–68 Fixed carbon % 100–(VM + ash) 40 (min)
SiO2 0.9–1.0 0.5–0.7 Volatile matter % 25–35 32
Al2O3 0.2–0.3 0.3–1.3 (VM)
CaO 1.1–1.2 0.06 max Ash % 24 (max) 22  2
MgO 0.80 0.05 Moisture % 4 (inherent) 8 (total)
P 0.015 0.03–0.06 Reactivity cc/gm. 1.75 2.2
S 0.006 0.005–0.008 sec
Moisture 1.5 max 4.0 max Initial deformation  C (K) >1150 (1423) >1200 (1473)
Physical characteristics temperature (IDT)
A. Size range, (weight %) of ash
· 103 m Calorific value kJ/kg 23,012 21,757
50–75 5 max Coking index >1 >1
30–63 93 max
+15 10 max
8–15 85 max 3. Ash fusion temperature
8 5 max 4. VM and sulfur content
6.3 7 max 5. Coking and swelling indices
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016

B. Bulk density, t/m3 2.0–2.1 2.0–2.6

C. Compression strength, 270 min Reactivity
kg/pellet
Cutting—Ductile

D. ISO tumbler test

The reactivity of the reductant is a measure of its ability to
(weight %) · 103 m
+6.3 95 min react with carbon dioxide to form carbon monoxide
–0.5 4 max according to Boudouard reaction:[4]

CðSÞ þ CO2 ðgÞ ¼ 2COðgÞ (1)

Coal
that is, for conversion of carbon dioxide (CO2) to carbon
The characteristic desired for noncoking coals is that they
monoxide (CO). The reactivity has a significant influence
have high fixed carbon content and high volatile content;
on the reduction process, particularly the operating tem-
ash, sulfur, and moisture content be low; and the ash fusion
perature of the reactor. The reactions of ore reduction as
point be high. Also, the coal should be highly reactive and
well as coal gasification take place simultaneously within
should have low coking and swelling indices. Coal is
a rotary kiln. Reactivity of coal is the main factor affecting
required as a reductant as well as a source of heat. The cost
the kinetics of the process. Coals with high reactivity
of coal is around 50% of the production cost of DRI.
values are preferred. This type of coal makes it possible
Coals of higher reactivity are preferred as they permit
to operate the kiln at lower temperature and to achieve
operation of the furnace at lower temperature with higher
higher kiln productivity.[8] Therefore, the decisive factor
kiln output. Lower-rank coals in the range of bituminous
in rotary kiln is the ability of any coal to react with carbon
high volatile yield char of higher reactivity. Generally,
dioxide (CO2) to form carbon monoxide (CO), a parame-
coals with higher fixed carbons are preferred, since they
ter quantified by the coal reactivity. In general, coals with
provide more carbon for reduction and are of higher calo-
high reactivity values are preferred as these allow kiln
rific value. But higher fixed carbon means lower volatile
operations at relatively lower temperatures. Lower kiln
matter (VM), thereby lowering the char reactivity. Ash
operating temperatures automatically decrease the ten-
content should be as low as possible, as it affects the kiln
dency toward formation of accretions or rings inside the
throughput and heat requirement adversely. High moisture
kiln[4] because the difference between the solid and gas
also reduces kiln productivity and increases the energy
temperature inside a kiln decreases and the chances of fine
requirement. Since sulfur of coals is picked up by DRI, it
materials sticking to the kiln refractory diminish.
should be as low as possible. Further, melting characteris-
tics of ash need to be considered while selecting coals for
DRI production. Table 4 gives the criteria for selection of Ash content
coals for DRI production in rotary kilns.[7]
The following characteristics are of importance in In any reduction process, the ash content of the reductant
selecting coal for DRI production: is an undesirable ballast. With high ash content in coal, a
greater proportion of the furnace volume is occupied by
1. Reactivity waste/inert materials, thereby reducing the volume available
2. Ash content for the iron oxide charge. This affects the productivity
 Direct Reduced Iron: Production
adversely.[4] Further, high ash content in coal also
enhances the tendency toward ring formation in rotary
kilns.
High ash content in coal means higher ratio of coal/iron
ore, which ultimately reduces the kiln capacity and hence
decreases productivity to a certain extent. Hence, the ash
content of coal should be as low as possible. From the
metallurgical standpoint, ash contents up to 35% are toler-
able.[9] In general, the ash content should be preferably
below 20%, and more than 24% ash-bearing coals[4]
should not even be considered if the process is to be
competitive.
Ash fusion temperature
Ash fusion temperatures give an indication of the softening
and melting behavior of coal ash, when the coal is exposed
to high temperatures. Fusion temperatures are typically
measured by heating the coal sample in the form of a cone
at four defined points, under both reducing and oxidizing
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
conditions. These four temperatures are as follows:
(i) Initial deformation temperature (IDT) is the temper-
ature at which the point of the cone begins to get
rounded.
(ii) Softening temperature (ST), also called the spherical
temperature, is the temperature at which the base of
the cone is equal to its height.
(iii) Hemispherical temperature (HT) is the temperature
at which the base of the cone is twice its height.
(iv) Fluid temperature (FT) is the temperature at which
the cone has spread to a fused mass.
Another important characteristic of any noncoking coal is
the IDT of the ash, especially under mild reducing condi-
tions. Normally, the IDT of coals should be at least 100
higher than the maximum furnace operating temperature.
Since under mild reducing conditions, the IDT value can
decrease by 50 –80 , all ash fusion temperatures for
assessing the suitability of coals for DR process should be
determined under reducing conditions.[4] The ash fusion
temperature of the coal is important. Coal having ash ST
above 1300 C (1573 K) is preferred.[8]
Volatile matter and sulfur content
The fuel ratio is the ratio of fixed carbon to VM in natu-
rally occurring coal. The value of fuel ratio should not
exceed 1.8 and should preferably be within 1.5 for
selecting adequately reactive coal.[2] It is desirable to have
a coal[6] with less than 15% ash, about 25–35% VM, and
sulfur content less than 1.0%.
Sulfur in coal can occur as iron pyrite (FeS), calcium
sulfate (CaSO4), and organic sulfur compounds with car-
bon and hydrogen. During heating, approximately half of
the pyritic sulfur is converted into H2S and volatilized at
600 C (873 K), while CaSO4 is reduced to CaS and forms
1087
a constituent of CO–S and H2S and thus gets volatilized.
The organic sulfur, however, is not affected by charring up
to 1000 C (1273 K) and is responsible for sulfur pickup by
DRI.[4] This sulfur can normally be scavenged by dolo-
mite/limestone added in very fine form to the coal feed
into any rotary kiln. In order to ensure sufficiently low
sulfur content in the DRI, the sulfur content of the coal
should not exceed 1.5%. Coals[9] with sulfur content less
than 1.0% are preferred, which ensure sulfur content in
DRI below 0.02%. However, sulfur pickup in DRI can be
effectively controlled[8] by the addition of dolomite to the
charge materials.
Coking and swelling indices
These values should be as low as possible to avoid prob-
lems of formation of large masses with low density within
the reactor. Coals with a free swelling index above 3.0
should be rejected; ideally, the free swelling index of coals
used in the DR process should not exceed 1.0. Similarly,
coals with a coking index of 5.0 (maximum) and prefera-
Cutting—Ductile
bly below 3.0 should only be used[4] as they are strictly
noncoking coals.
Natural Gas
The main advantages of gaseous fuels are as follows:
 It is clean and there is no ash formation (i.e., no residue
after reactions).
 It is easy to handle.
 It has good combustion properties.
In all gas-based processes, the reductant is natural gas.
Natural gas consists primarily of carbon and hydrogen
compounds in the form of hydrocarbons. Natural gas can-
not be used directly for gas-based DR processes, due to the
following reasons[10]:
1. Very slow reduction as compared to H2 and CO
2. Carbon soot formation and consequent process prob-
lems, such as choking, and unfavorable thermal bal-
ance in the reactor owing to endothermic effects
Therefore, natural gas cannot be employed in its virgin
form for DR process. It has to be converted into a mixture,
predominantly of H2 and CO, to increase the calorific
value and to increase the proportion of reducing gas in
relation to the oxidizing gases (like CO2). This step of
conversion into H2 and CO is known as reforming.
Reforming takes care of the problem of thermal balance
since reduction with a 75% H2 and 25% CO mixture has
almost zero heat effect on the system (exothermic CO
reduction is balanced by endothermic reduction by H2 at
this proportion of the two gases).
 1088
The recommended analysis of natural gas to be used in
gas-based DR process is given in Table 5. For efficient gas
reforming, the extent of higher hydrocarbons is limited to
C2H6 below 25%, C3H8 below 4%, and C4H10 below 2%.
Higher amounts of these constituents increase the possi-
bility of carbon deposition on the catalyst used for
reforming.[4] In order to prevent carbon deposition, the
amount of water vapor in the gas mixture exposed to the
catalyst must be increased. Excessive amounts of nitrogen
and carbon dioxide do not harm the reforming process, but
more than 20% carbon dioxide results in excess top gas,
which only has to be unnecessarily heated and cooled.
Every 10% increase in nitrogen results in approximately
2% higher fuel consumption. Sulfur in the natural gas,
either as H2S or COS, can poison the nickel catalyst owing
to the formation of NiS. Therefore, natural gas rich in
sulfur has to be desulfurized prior to reforming.
The main constituent of natural gas is methane (90–
93%). Natural gas is reformed, in the presence of catalysts
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
(Ni or Al2O3) at 950 C (1223 K), to produce reducing
gases CO and H2, which take part in the reduction of iron
oxide. The natural gas required for DRI production should
Cutting—Ductile
be very low in sulfur and low in higher hydrocarbons and
unsaturated hydrocarbons.[6] The gross calorific value and
net calorific value of natural gas[11] vary from 33,494 to
43,543 and 30,145 to 39,356 kJ/Nm3, respectively.
Other Raw Materials
Limestone or dolomite is used in rotary kilns as a
desulfurizing agent. Since most iron ores contain very
low sulfur (0.01–0.02%), the sulfur pickup in DRI is
almost fully from the charged coal; even the best
noncoking coals generally contain 0.5–0.6% sulfur. It has
been reported that dolomite is a better fluxing agent com-
pared with either limestone alone or a mixture of dolomite
and limestone of the same size. However, both dolomite
and limestone are acceptable provided the size is in the
range of 1 · 103 to 4 · 103 m. For minimizing accre-
tion formation inside the kiln without sacrificing efficient
desulfurization, it is essential that most of the desulfurizer
(at least 60%) is 1 · 103 to 2 · 103 m in size with very
little of 1 · 103 m particles.[4] Limestone should con-
tain minimum 46% CaO and maximum 6% SiO2, and
dolomite should contain minimum 28% CaO, 20% MgO
Table 5 Composition of natural gas for gas-based process.
Chemical analysis Volume (%)
CH4 75 min
C2H6 0–25
C3H8 0–4
C4H10 0–2
+ C4 hydrocarbons 0–0.5
CO2 Max 20
N2 Max 20
S Max 20 ppm wt
Direct Reduced Iron: Production
and maximum 6% SiO2. Dolomite is mainly used as a
desulfurizing agent to prevent the pickup of sulfur by the
DRI from the sulfur released by the burning of coal inside
the furnace.
Sizes of Raw Materials
The lump ore can be charged with a particle size of 5 ·
103 to 25 · 103 m. The average size of ore is around 12
· 103 m. The preferred size of pellets for iron ore is 6 ·
103 to 22 · 103 m, with majority in the range of 8 ·
103 to 16 · 103 m. In the initial days the iron ore size
was kept at 5 · 103 to 20 · 103 m and was washed in a
scrubber, but presently it has become a standard norm to
use 5 · 103 to 18 · 103 m ore as feed for a large kiln
without scrubbing and/or washing.[12] This has resulted in
reducing the cost of iron ore fed to the kiln. The optimum
size of noncoking coal is approximately 5 · 103 to 15 ·
103 m. Average sizes of limestone and dolomite is
around 1 · 103 to 4 · 103 m. The initial specifications
for dolomite were 1 · 103 to 4 · 103 m; later it was
found that 4 · 103 to 8 · 103 m dolomite was far more
suitable by which the consumption can be reduced by
50%. This was mainly due to the fact that a lot of dolomite
fines were being lost to waste gases, and with 4 · 103 to
8 · 103 m fraction, this loss was minimized.
Composite Pellets
The two major raw materials required for DRI production
are iron ore and coal/natural gas. The excessive fines gen-
erated from the iron ore crushing units are mostly going to
waste. Fines of coal and coke are also generated during
mining and coking, respectively. Therefore, if composite
pellets are made by the cold-bonding technique from iron
ore fine and coal/coke/char fines, then utilization of fines
takes place. The composite pellets should have sufficient
strength to withstand high temperature and stresses in
reduction furnaces.[13] The main advantage of composite
pellets is utilization of cheap resources (i.e., fines) without
the need for high temperatures such as sintering, coke
making, and conventional pellet making by heat harden-
ing. Reduction is much faster with composite pellets than
with ordinary iron ore pellets or lumps. This lowers the
residence time in rotary kiln by a factor of 6 to 8, thus
improving productivity. Details about iron ore–coal com-
posite pellet are given elsewhere.[14]
The composite pellets are used as feed material in
Rotary Hearth Furnace (RHF), where reduction takes
place within 10 minutes to produce iron nuggets. The
process is known as ironmaking technology mark 3
(ITmk3), jointly developed by Kobe Steel, Ltd., Japan,
and Midrex Direct Reduction Corporation, United States.
Midrex Technologies in partnership with its parent com-
pany Kobe Steel, Ltd. has developed the FASTMET pro-
cess as a carbon-based reduction technology, applicable
 Direct Reduced Iron: Production
for processing iron ore as well as iron oxide–containing
materials such as steel mill wastes.
The advantages of using composite pellets in reduction
processes are as follows[15]:
 Very fast reduction due to intimate contact of iron ore
fines and coal/coke/char fines
 Higher rate of production
 Reduction in energy consumption
 Utilization of fines, which are cheap and are being
generated in large quantities, awaiting suitable dis-
posal and utilization
 Promising prospect for small-scale ironmaking with
higher production rate
REACTIONS OF DIRECT REDUCTION
PROCESSES
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
Reduction Reaction for Solid Reductant
The reduction of iron ore occurs at 900–1000 C (1173–
1273 K) in the presence of coal. Although it is a solid–
solid reaction, it takes place via a gas–solid reaction:
3Fe2 O3 ðsÞ þ COðgÞ ¼ 2Fe3 O4 ðsÞ þ CO2 ðgÞ (2)
Fe3 O4 ðsÞ þ COðgÞ ¼ 3FeOðsÞ þ CO2 ðgÞ (3)
FeOðsÞ þ COðgÞ ¼ FeðsÞ þ CO2 ðgÞ (4)
CO2 ðgÞ þ CðsÞ ¼ 2COðgÞ (5)
Hence, the overall reaction is
Fe2 O3 ðsÞ þ CðsÞ ¼ FeðsÞ þ 3COðgÞ (6)
Carburization
3FeðsÞ þ 2COðgÞ ¼ Fe3 CðsÞ þ CO2 ðgÞ (7)
3FeðsÞ þ CðsÞ ¼ Fe3 CðsÞ (8)
Where the reduction is carried out below approximately
1000 C (1273 K), the reducing agents is CO and the DRI
produced is porous and have roughly the same size and
shape as the original iron ore or pellets. The metallic iron
also absorbs carbon according to reactions (7) and (8) to
form iron carbide (Fe3C).
Reforming of Natural Gas
Reformed natural gas is used as a reducing agent. Iron ore
lumps or pellets are reduced in the solid state and oxygen
1089
from iron oxide is removed by a gaseous reducing agent
(CO and H2) that takes part in the reduction of iron oxide.
Methane (CH4) is not used directly for the reduction of
iron ore because
 The rate of the reaction is very slow
 Due to high temperatures, decomposition of methane
takes place and carbon soot is formed
CH4 ðgÞ ¼ CðsÞ þ 2H2 ðgÞ (9)
Hence, methane (CH4) has to be reformed into CO and
H2 in the presence of a catalyst (Ni or Al2O3) at 950 C
(1223 K). There are four types of reformations:
 Steam reformation
CH4 ðgÞ þ H2 OðgÞ ¼ COðgÞ þ 3H2 ðgÞ (10)
 Oxygen reformation
CH4 ðgÞ þ 1=2 O2 ðgÞ ¼ COðgÞ þ 2H2 ðgÞ (11)
Cutting—Ductile
 Carbon dioxide reformation
CH4 ðgÞ þ CO2 ðgÞ ¼ 2COðgÞ þ 2H2 ðgÞ (12)
 Combination of steam and oxygen reformation
2CH4 ðgÞ þ H2 OðgÞ þ 1=2 O2 ðgÞ ¼ 2COðgÞ þ 5H2 ðgÞ
(13)
Reduction reaction for gaseous reductant
Apart from reactions (2) to (4), a considerable part of
reduction occurs due to H2 gas:
3Fe2 O3 ðsÞ þ H2 ðgÞ ¼ 2Fe3 O4 ðsÞ þ H2 OðgÞ (14)
Fe3 O4 ðsÞ þ H2 ðgÞ ¼ 3FeOðsÞ þ H2 OðgÞ (15)
FeOðsÞ þ H2 ðgÞ ¼ FeðsÞ þ H2 OðgÞ (16)
In gas-based DR process, the reducing agent is a mix-
ture of CO and H2. The interconversion reaction bet-
ween CO and H2O takes place according to the following
reaction:
COðgÞ þ H2 OðgÞ ¼ CO2 ðgÞ þ H2 ðgÞ (17)
This reaction is popularly known as water–gas displace-
ment reaction and occurs at a temperature below 1000 C,
due to the unstability of CO gas. The water–gas displa-
cement reaction becomes less important above 1000 C
 1090
because any water formed from the reduction of iron ore is
rapidly reduced by carbon:
H2 OðgÞ þ CðsÞ ¼ COðgÞ þ H2 ðgÞ (18)
As per the thermodynamic requirements, when a system is
in equilibrium, all parts of the system must be in equilib-
rium. Therefore, whatever (pCO2 =pCO ) ratio is in equilib-
rium with a solid phase must also be in equilibrium with
the gaseous phase [water vapor and hydrogen gas in reac-
tion (17)].
Carburization
3FeðsÞ þ COðgÞ þ H2 ðgÞ ¼ Fe3 CðsÞ þ H2 OðgÞ (19)
3FeðsÞ þ CH4 ðgÞ ¼ Fe3 CðsÞ þ 2H2 ðgÞ (20)
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
Most of the previously mentioned reactions are endother-
mic and hence heat must be provided for these reactions to
Cutting—Ductile
occur. So heat transfer plays a very dominant role in the
DRI production processes.
Thermodynamics of Reduction
The reduction of iron oxide by carbon monoxide or hydro-
gen takes place in three stages at temperature above 567 C
(840 K):
Direct Reduced Iron: Production
I II III
Fe2 O3 ! Fe3 O4 ! FeO ! Fe
Oxygen content 30:1% 27:7% 22:3% 0%
(21)
The important point to be noted is that even after reduction
to the FeO stage, over 22% oxygen (starting oxygen 30%)
still remains, which is removed only in the final stage of
reduction of wustite to metallic iron.
The most common reducing agents used in iron oxide
reduction process are carbon monoxide, hydrogen, and
mixtures of these two gases. The thermodynamics of iron
oxide reduction deals primarily with the equilibrium
between its oxides and these reducing agents. Thermody-
namic calculation shows the feasibility of the reaction,
whether the reaction moves in forward direction or not.
The feasibility of the reaction is indicated by a decrease
in free energy. The term standard state generally con-
siders one atmospheric pressure and a temperature of
25 C (298 K). Fig. 1 is a representation of the Ellingham
diagram of a few key oxides that are of major interest
in iron oxide reduction. The lines of oxides lying on the
higher sides represent a less stable oxide as compared
to the lines on the lower sides. It can be seen that above
727 C (1000 K) the line of CO2 formation from CO
and O2 lies slightly above the line of FeO formation
from Fe and O2. This means that under standard state
the following reaction cannot proceed in the forward
direction:
Fig. 1 Free energy vs. temperature diagram for
main oxides.
 Direct Reduced Iron: Production 1091

can be calculated[18] from the standard data, which is

FeOðsÞ þ COðgÞ⇄FeðsÞ þ CO2 ðgÞ (4)
3.933 kJ/mol of Fe. Similarly the values of ΔG0r, 1273 and
ΔG0r, 1473 for reaction (4) are 9.793 and 14.085 kJ, respec-
Reaction (4) is the key reaction in DRI making and the
tively. By increasing the temperatures the values of the
reaction is made to move in the forward direction by
free energy changes [for reaction (4)] are also increased.
keeping a large excess of CO over CO2; the CO/CO2 ratio
The feasibility of the reaction depends on the decrease in
is required to be over 2.52 at 1000 C (1273 K) for the
free energy. That means reaction (4) cannot proceed in the
reaction to proceed,[16] that is, by decreasing the activity
forward direction according to the thermodynamic calcu-
of CO2.
lation. But the reaction is made to move in the forward
Compared with BFs, the reduction potential of DR is
direction by keeping a large excess of CO over CO2, that
considerably lower and the reduction takes place at a
is, by decreasing the activity of CO2. Hence, reaction (4) is
much lower temperature. The net result is that while in
kinetically controlled rather than thermodynamically.
BF ironmaking, FeO is completely reduced to iron (there
The equilibrium constant and the gas composition at
is no FeO even in the slag phase in BF), the DR process is
equilibrium for reaction (4) can be calculated thermody-
incapable of reducing FeO completely, and some FeO is
namically. The equilibrium gas composition at 727 C
always present in DRI. Gas-based DRI contains less FeO
(1000 K) is 61.61% CO and 38.39% CO2 by volume.
than coal-based DRI because of the presence of hydrogen
Hence, for the reduction of FeO [reaction (4)] at 727 C
gas in reformed natural gas.
(1000 K), the gas composition should consist of more than
61.61% CO by volume (as shown in Fig. 2). If the gas
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016

Reduction by carbon monoxide

phase contains another gas such as nitrogen, the ratio of
CO–CO2 will remain the same (i.e., 1.6), but their percent-
The three reduction reactions by carbon monoxide gas
age composition will decrease by an amount equal to the

with their heats of reaction at 25 C (298 K) are as follows:
3Fe2 O3 ðsÞ þ CO ðgÞ ¼ 2Fe3 O4 ðsÞ þ CO2 ðgÞ
ΔHo 298 ¼ 52:43 kJ=mol of CO
Fe3 O4 ðsÞ þ COðgÞ ¼ 3FeOðsÞ þ CO2 ðgÞ
ΔHo 298 ¼ 40:46 kJ=mol of CO
FeOðsÞ þ COðgÞ ¼ FeðsÞ þ CO2 ðgÞ
ΔHo 298 ¼ 18:54 kJ=mol of CO
Cutting—Ductile
percentage of the nitrogen gas.
For FeO reduction, the temperature and gas composi-
tion must be such that the point on the graph (Fig. 2)
(2) representing the temperature and gas composition must
lie in the area where iron is the stable phase, for example,
at 800 C (1073 K), gas composition must consist of 80%
(3) CO and 20% CO2. Changes in temperature will also affect
the equilibria and the direction in which the reactions will
proceed.
(4) Fig. 2 also contains the curve showing the CO–CO2
composition and temperature equilibrium for the reaction:

The heat of reactions at 25 C (298 K) were calculated by CO2 ðgÞ þ CðsÞ ¼ 2COðgÞ, ΔHo 298 ¼ 172:423 kJ=mol of C
the data taken from the standard reference book.[17] It is (5)
worth noting that for reactions (2) and (4), the heat of
reactions are negative, which means that the reactions are
exothermic, whereas for reaction (3), it is positive, that is,
the intermediate stage of reduction of magnetite to wustite
is endothermic and requires a considerable amount of
heat.
Since wustite (FeO) is metastable below 567 C (840 K),
iron oxide reduction takes place in two stages:

I II
(22)
Fe2 O3 ! Fe3 O4 ! Fe

1=
4 Fe3 O4 ðsÞ þ COðgÞ ¼ 3/4 FeðsÞ þ CO2 ðgÞ,
ΔH o
298 ¼ 3:778 kJ=mol of CO (23)

This reaction is also exothermic. However, most DR pro-

cesses operate at temperatures more than 600 C (873 K)
so that reduction reaction (23) is only of minor interest.
The free energy change (ΔG0r, 1000 ) for reaction (4) [i.e.,
FeO (s) + CO (g) = Fe (s) + CO2 (g)] at 727 C (1000 K) Fig. 2 Fe–O–C system.
 1092
This is commonly known as the Boudouard reaction or
gasification reaction. This reaction has very important
consequences in iron oxide reduction when carbon is used
as the reductant, because it regenerates the reducing gas
by converting CO2 to CO. It is strongly endothermic, and
since the activation energy of an endothermic reaction
must be larger than the heat of the reaction, the rate of
reaction of Eq. 5 is sensitive to the crystalline structure
and impurities of carbon and temperature. To the left of
this curve in Fig. 2, which is at low temperature, CO is an
unstable gas phase. Carbon monoxide tends to decompose
into carbon dioxide and to deposit carbon in the form of
soot [for reverse reaction of Eq. 5]. This carbon soot
accounts for some of the carbon found in DRI. At high
temperatures carbon dioxide reacts with carbon to produce
carbon monoxide. Again this is an endothermic reaction
and is favored by high temperature.
Note that at temperatures above 1000 C (1273 K),
the reaction is essentially completed, that is, 100% CO
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
is formed. This means that thermodynamically, carbon
dioxide cannot exit at temperatures above 1000 C
(1273 K) in the presence of carbon. Therefore, this
Cutting—Ductile
suggests that CO cannot reduce FeO [as reaction (4)]
at temperatures above 1000 C (1273 K). However,
reduction does take place and this appears to be on
account of reactions (4) and (5) taking place separately.
The two reactions can be combined as follows to indi-
cate the overall reaction:
FeOðsÞ þ COðgÞ ¼ FeðsÞ þ CO2 ðgÞ,
ΔHo 298 ¼ 18:535 kJ=mol of CO (4)
CO2 ðgÞ þ CðsÞ ¼ 2COðgÞ, ΔHo 298 ¼ 172:423 kJ=mol of C
(5)
FeOðsÞ þ CðsÞ ¼ FeðsÞ þ COðgÞ,
ΔHo298 ¼ 153:888 kJ=mol of C (24)
Reaction (24) is often referred to as direct reduction (DR)
by carbon, whereas reaction (4) and reaction (5) are
known as indirect reduction by carbon and gasification
reaction, respectively.
It should be noted that the Boudouard equilibrium curve
in Fig. 2 crosses the wustite–iron line at 700 C (973 K) and
the magnetite–wustite line at 650 C (923 K). This means
that thermodynamically,[19] wustite cannot be reduced at
temperatures below 700 C (973 K) and magnetite cannot
be reduced at below 650 C (923 K) because the carbon
monoxide decomposes into carbon dioxide and carbon.
Reduction by CO and H2 mixtures
Reducing gases (CO and H2) made by reforming natural
gas are widely used for DR. Reforming reactions are
shown in section “Reforming of Natural Gas.” Consider
Direct Reduced Iron: Production
that the reactions proceed simultaneously in the same sys-
tem at 927 C (1200 K) as follows:
FeO þ CO ¼ Fe þ CO2 , ΔHo 298 ¼ 18:535 KJ=mol of CO
(4)
FeO þ H2 ¼ Fe þ H2 O, ΔHo 298 ¼ 21:401 KJ=mol of H2
(16)
In the reduction of Fe3O4 to FeO and further to Fe with solid
carbon, a minimum reaction rate has been observed before
the nucleation of iron. The time required for the nucleation of
metallic iron from FeO is largely dependent on the amount
of initial oxygen present in the oxide. The quantity of oxygen
removed in the nucleation period increases with increasing
initial oxygen content. Iron nucleated on FeO grows linearly
with time for reduction with both hydrogen and carbon mon-
oxide. However, the growth rate of iron on FeO, when
hydrogen is used, is approximately 40 times higher than that
at the same partial pressure of carbon monoxide at 800 C
(1073 K). This is because iron forms a thin and dense layer
during hydrogen reduction, while in the case of carbon mon-
oxide, the layer is thick and porous. This is of particular
advantage in the case of gas-based DR processes.[18]
The retarding effect on nucleation occurs only at the
FeO to Fe step and not during the formation stage of
Fe3O4 and FeO from Fe2O3 because both Fe3O4 and FeO
have a cubic structure and are crystallographically simi-
lar. Consequently, the transformation in this stage occurs
without any induction period. However, during the reduc-
tion of Fe2O3 to Fe3O4, the lattice transforms from rhom-
bohedral to cubic; therefore, the reduction behavior
between Fe2O3 and Fe3O4 is different. The nucleation
and growth processes of iron oxide are very important
because they influence the structure of the reduced phase,
which in turn affects the subsequent reduction rate. Reac-
tion (23) shows that Fe3O4 can be reduced to metallic
iron directly below 567 C (840 K). The reduction of
Fe3O4 to FeO below 650 C (923 K) is possible with H2
but not with CO. Thermodynamically, H2 reduces iron
oxide more easily than CO, as revealed by the free energy
changes that occur at 1000 C (1273 K) in both cases:
Fe2 O3 þ 3H2 ¼ 2Fe þ 3H2 O, ΔGo 1273 ¼ 150:62 KJ=mol
(25)
Fe3 O4 þ 4CO ¼ 3Fe þ 4CO2 , ΔGo 1273
¼ 129:70 KJ=mol (26)
Carbon deposition
As stated earlier, carbon forms at low temperatures by the
backward reaction of reaction (4). Although thermody-
namic calculation shows that carbon deposition is highly
 Direct Reduced Iron: Production
favored by low temperatures and high carbon monoxide
concentrations, it is unlikely that this reaction ever goes
completely to a state of equilibrium because reaction rates
are very slow at low temperatures and also because the
deposition of carbon requires the nucleation of carbon on
some catalytic site that may not always be available.
Undoubtedly some of the carbon found in DRI is soot
formed by this reaction. In addition, iron carbide (Fe3C)
can be formed by the following reaction[19]:
3Fe (s) + 2CO (g) = Fe3C (s) + CO2 (g) (7)
Like carbon deposition, reaction (7) is also favored by low
temperatures and high carbon monoxide concentrations.
DIRECT REDUCTION PROCESSES
Coal-Based Processes
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
Coal-based processes are broadly classified into two
categories:
1. Rotary kiln process
2. Rotary hearth process
Rotary kiln process
Rotary kiln processes are SL/RN, ACCAR, KRUPP,
CODIR, TDR, etc.; out of these, the SL/RN process is the
most common. The term SL/RN process stands for Stelco,
Lurgi, Republic Steel, National lead process, which was con-
ceived around 1960 for the production of high-grade DRI.
So many similar types of processes are developed. Funda-
mentally, all processes are the same in principle. Hence, the
discussion is restricted to the rotary kiln process only.
Normally, the rotary kiln is 60–125 m long (L) and
4–6 m in diameter (D), and the L/D ratio is about 15–20.
The size of the biggest kiln is 6.1 m in diameter and 125
m in length. The inclination of the kiln is about 1–4% of
the length and the rotation is about 0.5–4.0 rpm. The kiln
is lined with high alumina refractory (70% Al2O3) having
a thickness of 200 · 103 m, which can withstand 1600–
1700 C (1873–1973 K) and have high abrasion resis-
tance. Fig. 3 shows a flow sheet of the SL/RN process.
Sized lump iron ore (or pellets) and a relatively coarse
fraction of noncoking coal are fed into the kiln from the
feed end. Coal not only acts as a reducing agent but also
acts as fuel to supply the heat required for maintaining
the temperature profile of the charge within the kiln. A
finer fraction of coal is introduced from the discharge end
of the kiln along with the air to complete the reduction,
because by the time the charge travels to around 70% of
the length of the kiln,[10] very little coarse coal is avail-
able to complete the last stages of reduction. Feeding of
raw materials is done from one end that is at a higher
1093
level. It travels under gravity aided by the rotating
motion, through several heating zones, and the reduced
charge comes out from the other end of the kiln. Fifty to
sixty percent of the volume of the kiln is empty, where
the VM of coal and other gases are burnt in the presence
of oxygen from air and produce heat. Forty percent of the
return coke is fed back into the kiln to make the process
economical. Flux (limestone/dolomite) is added along
with coal to control the sulfur level in DRI. CaO in flux
forms CaS by reaction with sulfur. Controlled amount of
air is provided throughout the length of the kiln to get a
uniform temperature.
Most of the reactions are endothermic, so heat should
be provided in the kiln. It is done in two ways:
 Burning of gas, oil, or coal powder in a burner through
the discharge end.
 Exit gas from the kiln is rich in CO and VM. In the
presence of oxygen, partial combustion of these gases
takes place and generates a lot of heat to produce
electricity.
Cutting—Ductile
Part of the coal is introduced from the kiln discharge end to
supply energy at the discharge end, maintaining reducing
atmosphere at the discharge end to prevent reoxidation of
DRI and for controlling the degree of metallization and
carbon content of DRI. The separation of the product is
more or less similar in all the coal-based processes and
involves screening and magnetic separation for removal of
nonmagnetic ash, char, and used desulfurizer.
Fine coal, which contains highly volatile hydrocar-
bons, is injected by compressed air through discharge
end of the kiln. Thus, the volatile hydrocarbon in the
coal is released and cracked [CnHm ! nC + m/2 H2],
acting as an additional and very effective reducing
agent (i.e., H2). The charge moves through the kiln in
counter-current to the hot gases and during the process
gets heated to the reduction temperature.[20] About 45%
of coal is charged with iron ore from the feed end and
the remaining 55% from the discharge end of the kiln.
Depending on the properties of ore and coal, the tem-
perature in the isothermal zone is maintained between
1000 and 1050 C (1273 and 1323 K). After retention
time of 8–10 hours in the kiln, the DRI is discharged
via a transfer chute into a cooler. The discharge product
of the kiln should be cooled in inert or reducing atmo-
sphere since DRI is porous and hot at that time; other-
wise oxidation of DRI takes place. Here the material is
cooled below 100 C (373 K) by indirect cooling of
water in rotary cooler. No water comes in direct contact
with the product in the cooler and hence there is very
little chance of reoxidation of DRI. Since DRI is mag-
netic in nature, it can be easily separated from the
nonmagnetic materials (consisting of coal ash, char,
and fluxes) by using magnetic separators. On the basis
of a C/Fe ratio of 0.42–0.44, the coal input in the
 1094
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
Fig. 3 Flow sheet of the SL/RN process.
process varies between 1.10 and 1.20 t per tonne of
Cutting—Ductile
DRI, depending on the ash content of the coal.
The length of kiln is divided into two zones: 1)
preheating zone and 2) reduction zone. The hot gas heats
up the solid. The difference of the solid charge tempera-
ture and gas temperature is less at the end of the
preheating zone. This temperature difference depends on
the reactivity of coal. In the preheating zone some reduc-
tion will occur. The Fe2O3 ! Fe3O4 and Fe3O4 ! FeO
steps of reduction are completed in the preheating zone.
In this zone, devolatilization and charring of coal and
decomposition of limestone also take place. In the reduc-
tion zone, the FeO ! Fe step of reduction occurs.
The following are problems of the rotary kiln process:
1. Low productivity
2. Building up of accretions or ring formation
3. Erratic operation
4. Heat loss through exit gas
5. Low carbon content in DRI produced
1. Low productivity: Rotary kiln has lower productivity
(0.5–0.7 t/m3/day) than modern BF (2.0–2.5 t/m3/day).
In rotary kiln, the furnace is only partially filled (~40%),
that is, a lot of space is unutilized and the rate of reduc-
tion is low due to the lack of mixing of ore and reductant.
Productivity can be improved by the following:
 Using high reactivity coal.
 Using an iron ore of high reducibility.
 The gasification rate depends on the temperature,
so the higher the temperature of the kiln, the better
will be the gasification. By controlled combustion,
the maximum possible temperature can be achieved
and maintained within the kiln.
Direct Reduced Iron: Production
 Inducing some gaseous reductant (H2), especially
toward the later stages of reduction. At the later
stage, by introducing some H2 in the kiln, the inner
part of the ore can be effectively reduced since the
rate of reduction with H2 is 4–5 times higher than
that of CO.
 Using iron ore–coal composite pellets.
2. Building up of accretions or ring formation: Accretion
or ring formation means the process of growing
together (i.e., separate particles) into one, that is,
depositing on refractory surface. The presence of
fines accelerates the formation of these deposits.
Due to fusion of coal ash and some other reactions,
this buildup occurs. If accretion occurs, then the kiln
must be stopped. This problem is overcome by
adopting proper selection of coal so that ash does
not fuse at low temperature. Coal having fusion tem-
perature of ash above 1300 C (1573 K) is preferred.
If operating temperature is higher, then accretion
will be more. That is why very high temperature in
kiln cannot be maintained to improve the gasifica-
tion reaction.
3. Erratic operation: Erratic means irregular operation
having no certain course. This is due to the lack of
proper mixing.
4. Heat loss through exit gas: Exit gas has a very high
temperature (heat content is about 6.28 GJ/t), so it can
be used to preheat the materials or producing electrical
power.
5. Low carbon content of DRI: Carbon content is less than
1% for rotary kiln product. This can be improved by
injection of hydrocarbon under the solid bed. Carbon
in DRI may go up by 1%. Generally, carbon in DRI
varies between 0.3% and 0.5% without injection of
hydrocarbon.
 Direct Reduced Iron: Production
Rotary hearth process
Some processes based on RHF are as follows:
1. INMETCO process (1970)
2. FASTMET process (1990)
3. Sidcomet process (1990s)
Among all these processes, the FASTMET process has
been a commercial success.
ITmk3 and FASTMET processes are noteworthy DR
processes that involve simple operation, lower unit
consumption, lower production cost, and superior environ-
mental compatibility.[21] The typical features of RHF-
based processes are as follows:

Basically, the commercial scale is 0.5 Mt per annum of
product.
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
 The main raw materials are iron ore concentrate or
iron-bearing waste/dust of iron and steel industries
and noncoking coal fines. Fines are used in the form
of green pellets or briquettes. Though it depends on the
quality, approximately 1.5 t and 0.5 t of iron ore and
noncoking coal are used, respectively.

Direct use of coal as a reducing agent in pellets. The
heat requirement is met through firing gas to heat the
furnace.

These processes are developed in order to supply iron
nuggets or hot metal for those countries endowed with
noncoking coal.

These processes produce iron nugget or hot metal
that is the premium quality feedstock for EAF and/or
BOF.

Plant can be located next to the area of EAF or BOF in
steelmaking facilities for the efficient charging of hot
metal.
FASTMET process. The success of the RHF-based DRI
technology has led to the development of the FASTMET
process. Midrex Technologies in partnership with its parent
company Kobe Steel, Ltd. has developed the FASTMET
process as a carbon-based reduction technology, applicable
for processing iron ore as well as iron oxide–containing
materials such as steel mill wastes. From both economical
and environmental points of view, the FASTMET process
is very attractive as a proven technology for dust recycling.
This process was developed by Midrex Corporation in
1991. Rotary hearth of 2.5 m diameter was installed at the
Midrex Technical Center to simulate the reduction process.
Midrex and Kobe Steel, Ltd. constructed a 2.5 tonne per
hour (tph) FASTMET plant at Kobe Steel, Ltd., Kakogawa
Works, Japan, in 1994.
Fig. 4 shows the flow sheet of the FASTMET process.
The FASTMET process is a solid reductant-based RHF
process. Iron ore concentrate fines along with reductant
1095
fines (such as charcoal or other carbon-bearing solid) are
pelletized to form iron oxide–carbon composite pellets,
dried, and then charged directly to the RHF. If the mill
scale or other wastes of steel plant are used as feedstock of
iron oxide, then they are briquetted along with a solid
reductant before charging to the RHF. As the hearth
rotates, the pellets/briquettes are heated to 1250–1400 C
(1523–1673 K) by combustion of natural gas, oil, or pul-
verized coal. The pellet layers on rotary hearth are of one
to three pellets deep, and burners and postcombustion of
CO provide the heat to raise the pellets to reduction tem-
perature. CO is generated from the carbon present in the
charcoal or other carbon-bearing solid in the composite
pellets as well as by the combustion of liquid or gaseous
fuels in burners installed above the rotating hearth. The
agglomerates containing the solid reductant get reduced to
metallic iron. The reduction is accomplished by intimate
contact between the carbon and iron oxide particles within
the pellets/briquettes at relatively high temperature. The
rate of reduction is very fast; hence the residence time
of the charge in the hearth is typically as less as 6–12
minutes, during which 90–95% of the iron oxide is
Cutting—Ductile
reduced. The product can be cold DRI, hot DRI, HBI, or
hot metal depending on end use requirements. The DRI
produced is continuously discharged at around 1000 C
(1273 K) from the furnace, either into refractory-lined
container for hot charging to the melt shop or into
briquetting machines for the production of HBI or directly
cooled in inert atmosphere to get directly reduced pellets.
The carbon content in the product can be controlled
between 1.0% and as high as 6.0%, if required. Additional
heat from the exit gas of the process can be recovered by
producing electrical power. FASTMET DRI (HBI) is
melted by electric energy to produce hot metal, hence
another name of the FASTMELT process.
ITmk3 process. The ITmk3 process is considered to be
the third generation of ironmaking technology, the first two
generations comprising BF and DR processes. Represen-
ting the next generation of modern ironmaking technology,
ITmk3 processes iron ore fines into almost pure pig iron
nuggets within 10 minutes. The result is a conveniently
sized, slag-free material ideally suited for further pro-
cessing by conventional technologies into high-quality
steel products and foundry iron castings. The Mesabi Nug-
get demonstration plant, commissioned in July 2004,
achieved continuous, reliable production of pig iron nug-
gets under commercial operation conditions. Ten thousand
metric tonnes of quality pig iron nuggets were produced
during the four test campaigns. The produced nuggets
were consumed in the steelmaking operations at various
North American locations.[22] Fig. 5 represents the ITmk3
process.
In ITmk3, the iron ore fines and noncoking coal are
formed into green iron ore–coal composite pellets. ITmk3
uses the same type of mixing and agglomeration steps and
 1096 Direct Reduced Iron: Production

Fine ore/Mill waste

Offgas
Coal Binder

Feed preparation Dryer air

Combustion air Gas cleaning

Pelletizer system Fan

Mixer Heat exchangers Stack

Dust
recovery
Flue gas conditioner
Dryer
Briquetter
Offgas treatment
Burners
Product discharge Fuel gas
Hot DRI oil
pulverized coal
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016

Rotary hearth furnace

Reduction
Cutting—Ductile

Briquette machine
DRI container
DRI

Liquid Solid
HBI
Hot DRI FASTIRON* FASTIRON*

Fig. 4 Flow sheet of the FASTMET process.

RHF as FASTMET. The composite pellets are fed to an
RHF and heated to 1300–1450 C (1573–1723 K); at this
temperature range, the pellets are reduced to form iron
nuggets. The temperature of RHF is raised thereby melt-
ing the reduced iron and enabling it to easily separate from
the gangue. This ironmaking process takes only 10
minutes against 10 hours in the BF process and 8 hours in
Fig. 5 Flow sheet of the ITmk3 process.
the rotary kiln process. Iron and slag get separated and the
product is called nuggets. Iron nuggets can be fed directly
into BOF or EAF as a pure iron source and a substitute of
scrap. By substituting scrap, it can dilute tramp elements
like Cu, Pb, Sn, and Cr.
Some of the key features of ITmk3 are as follows[23]:
 The process produces iron nuggets within a very short
reduction time (10–12 minutes) in an RHF.
 Impurities are removed to some extent in the form of
slag.
 The iron nuggets contain high carbon (2–3%) and low
sulfur, but in solid form.
 Coal consumption is reported to be 500 kg/t iron nug-
gets, which is fairly low.
 The process is flexible as far as the types of raw mate-
rials are concerned.
 ITmk3 reactors emit at least 20% less carbon dioxide
than BF. The overall NOx, SOx, and particulate matter
emissions are also lower.
 The iron nuggets can be fed directly to the EAF/IF to
make steel along with steel scrap or DRI.
 The iron nuggets are easy to transport and handle. High
in density, they do not reoxidize or generate fines.
 Direct Reduced Iron: Production 1097

Gas-Based Processes a total capacity of 300,000 tpa. The average energy con-
sumption of this early plant was approximately 15 GJ/t of
Gas-based processes are broadly classified as follows: DRI.[24]
The flow sheet for Midrex is shown in Fig. 6. The main
1. Midrex process components of the process are the shaft furnace, the gas
2. HYL process reformer, and the cooling gas system. The temperature and
3. Purofer process composition of the gas to the shaft furnace are controlled
4. Finmet process to maintain optimum bed temperature for reduction,
5. High iron briquette (HIB) process degree of metallization, carbonization level, and efficient
utilization of the reducing gas.
The charge of the Midrex shaft consists of around 60%
Midrex process pellets and 40% lump ore of a particular type. Pellets are
the preferred feedstock owing to their superior physico-
The Midrex process was developed by the Surface Com- chemical characteristics compared with lump ores. As a
bustion Division of Midland-Ross Corporation in the result, in most cases, a minimum amount of pellet usage
mid-1960s. The Midrex Division became a subsidiary of becomes mandatory.[10] The process essentially consists
Korff Industries in 1974. Midrex was subsequently of reducing the iron ore pellets and lump by reformed
acquired by Kobe Steel, Ltd. in 1983. The first commer- natural gas in a vertical shaft furnace. It is a continuous
cial Midrex plant was installed near Portland, Oregon, counter current gas–solid reactor. Iron ore lump and pel-
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016

and started production in 1969. The plant included two lets are fed in the shaft from the top and the product DRI is
shaft reduction furnaces of 3.4 m inside diameter and had removed from the bottom.

Cutting—Ductile
Iron oxide

Flue
gas NAtural Shaft
gas
furnace
Process gas Top gas
compressors scrubber

Reformer
Reducing gas
Main air
blower
Fuel
gas
Natural gas
O2
Flue
gas
Hot transport
Natural
gas
Ejector
Feed gas
Stack HDRI

Combustion air HDRI

Hot transport conveyor
Distance greater than 100m
Heat recovery DRI
Briquette
cooler
machine
Electric arc furnace
@
HOTLINK
Distance less than 40m
DRI Hot transport vessel
storage Distance greater than 100m
HBI
Discharge options storage

Fig. 6 Flow sheet of the Midrex process.

 1098
Reduction of iron ore by reducing gases, in a shaft
furnace, is according to the counter current principle. The
main characteristic is given by the production of reducing
gas by recirculation of the top gas from the shaft and using
part of it as a reforming agent for natural gas. The process
can be broken down into three main steps that are
interdependent and influence each other[25]:
1. Production of reducing gas
2. Reduction of iron ore
3. Cooling of end product
Production of reducing gas. Reducing gas is produced
by continuous catalytic reforming of natural gas with CO2
and H2O contents in the recirculated shaft furnace top gas,
thus allowing the reforming process (in the presence of Ni
catalyst) to work without additional input of external
reforming media like O2, air, or steam. The reforming
reactions take place at a temperature of 950 C (1223 K).
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
The product is a reducing gas with about 85–90% (CO +
H2) and 10–15% CO2 + H2O. The reforming reactions are
Cutting—Ductile
Major : CH4 ðgÞ þ CO2 ðgÞ ¼ 2COðgÞ þ 2 H2 ðgÞ,
ΔHo 298 ¼ 11, 033 kJ=Nm3 CH4 (27)
Minor : CH4 ðgÞ þ H2 OðgÞ ¼ COðgÞ þ 3H2 ðgÞ,
ΔHo 298 ¼ 9196 kJ=Nm3 CH4 (28)
The gas volume is increased as per reactions (27) and (28).
The reducing reactions in the shaft do not affect gas vol-
ume, so that the surplus gas together with additional fuels
can be used for heating the reformer. Heating of the
reformer is necessary for endothermic reforming reac-
tions, temperature losses of the reformer, and supplying
sensible heat to the reducing gas for heat requirements in
the shaft furnace.
Reduction of iron ore. Iron ore reduction is performed
in the vertical, cylindrical shaft furnace by the H2 and CO
components of the reducing gas while the iron ore is con-
tinuously charged from the top. The reduction gas enters
the shaft furnace at the bottom of the reduction zone with a
temperature of 800–900 C (1073–1173 K). Higher tem-
perature could cause formation of clusters and cracking in
the descending ore burden. The charge takes about six
hours to pass through the reduction zone, where it is
preheated and reduced by counter-current contact with
the reducing gas. The residence time for obtaining 95%
metallized DRI that varies depending upon the gas com-
position, flow rate, type of iron ore, etc., is between 5 and
6 hours. The carbon present in the DRI depends on
the (CO + H2)/(CO2 + H2O) ratio. The higher the ratio,
the more the carbon is deposited from the CO and CH4
decomposition.
Direct Reduced Iron: Production
Cooling of end product. As the metallized material
passes down through the conical section of the shaft, it is
cooled by inert gas. Because the hot DRI leaving the
reducing zone is a good reforming catalyst, a controlled
amount of natural gas mixed with process off-gas is added
to the cooling gas. The reforming reaction, which is endo-
thermic, helps to cool the hot DRI and provides additional
hot reducing gas, some of which enters the reduction zone
where its sensible heat and reducing power are recovered.
In this manner the amount of reducing gas required from
the reformer can be decreased by 8–10%. The DRI is
cooled to 50 C (323 K) before it is discharged. The gas is
withdrawn at the top of the shaft. This gas then passes
through a scrubber and cooler, after which it contains about
70% hydrogen and carbon monoxide.[19] It is then used for
firing the reformers and for mixing with natural gas to
make the process gas mixture that is fed to the reformers.
The top gas, at 350–400 C (623–673 K), leaves the
reduction furnace containing about 30% (CO2 + H2O)
and enters a five-stage venture cooler scrubber where it is
cooled to about 80 C (353 K), steam is condensed, and the
dust is removed to less than 0.02 grains of solid/m3. Part of
this gas (60–70%) goes to the compressor and then to a
mixing chamber where it is mixed with fresh natural gas.
The mixture passes through a reforming furnace where it
is converted to reducing gas containing 95% (CO + H2).
Low-carbon DRI (<1.5% C) is directly cooled using a
circulating stream of cooled exhaust gas introduced in the
bottom section, before cold DRI is discharged. When
higher-carbon DRI (up to 4.0% C) has to be produced,
natural gas is introduced along with the cooling gas into
the bottom. Natural gas readily decomposes in the pres-
ence of highly reactive metallic iron, thereby generating
nascent carbon, which gets absorbed in the product before
it is discharged. In both cases, the final product is DRI
with 93–94% metallization, with the desired carbon con-
tent.[10] Midrex technologies is commercially the most
successful DR technology today.
HyL process
In the HyL process, developed by Hojalata y Lamina
South Africa (Hylsa) of Monterrey, Mexico, in 1953, lump
ore and fired pellets are reduced in fixed-bed retorts by
reformed natural gas. This development came up in 1957
to a four (fixed-bed) reactor prototype unit, using steam
reforming of natural gas for generation H2 and CO gases
for reduction. The first commercial HyL plant was
installed at Monterrey and started production late in
1957. This plant has a capacity of 200 tpd of DRI, and the
reactors are approximately 2.5 m in diameter and hold
approximately 15 t of ore in a 1.5 m deep bed. The reac-
tors in the more recent plants are 5.4 m in diameter and 15
m high. Design capacity is approximately 1900 tpd of DRI
having an average reduction of approximately 90%. The
energy consumption in the most recent plants is 14.9 GJ/t
 Direct Reduced Iron: Production
of 90% reduced DRI.[25] Two-stage reduction of lump ore
or pellets in retorts used a mixture of H2 and CO gases
obtained by catalytic reforming of natural gas by steam.
Fig. 7 shows a flow sheet of the HyL I/II process.
Each HyL plant consists of a gas reformer and four
reactors; the reactors are identical. In a 12 hours cycle
each reactor goes through four steps of about 3 hours each.
In step 1, reactor loading follows reactor discharging. In
step 2 (secondary reduction), valves are opened to pass
partially depleted reducing gas through the gas fixed pre-
heater. From the preheater the reducing gas enters the
burner section where preheated air is injected to raise the
reducing gas temperature to 1040 C (1313 K). This hot
gas flows down through the charge, reducing ores/pellets,
and exits through the quencher (about 40% of combined
oxygen is removed). The off-gas from this step is not
recycled but is used as fuel gas throughout the plant. In
step 3 (primary reduction), valves are remotely operated to
switch from partially depleted reducing gas to fresh
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
reformer gas coming from the reactor being cooled. This
gas is also preheated and this strong reducing gas reduces
almost all of the FeO present in the charge to metallic Fe.
The quenched off-gas from this step is piped to step 2
above. In step 4 (cooling and carburizing), cold reducing
gas directly from the reformer is passed through the reac-
tor without preheating. The passage of the cold and fresh
reducing gas through the product helps in cooling it and
introducing carbon into it. The off-gas from this cooling
step is quenched and is piped to a reactor in step 3. Mostly
cooling of the gas to condense the product water and then
heating before entering the reactor are followed. Every
HyL reduction plant has a fixed volume input of reducing
gas per hour (80–90% of CO + H2).
Fig. 7 Flow sheet of the HyL I/II process.
1099
Production of reducing gas: The reducing gas is pro-
duced in a steam reformation. The high-pressure material
gas is first desulfurized with activated carbon to prevent
poisoning of the catalyst in the reformer. The gas is then
mixed with superheated steam. This mixture is preheated
and then passed through heated Cr–Ni alloy tubes
containing a Ni catalyst. The tubes are heated to provide
the endothermic heat required in the following reaction:
CH4 ðgÞ þ H2 OðgÞ ¼ COðgÞ þ 3 H2 ðgÞ,
ΔHo 298 ¼ 9196 kJ=Nm3 CH4 (28)
Reducing gas temperature and pressure: The reducing
gases for both the primary and secondary steps are first
preheated to about 800 C (1073 K) in tubular heaters.
After that the gas is elevated to a temperature of 980–
1230 C (1253–1503 K) by injecting a controlled quantity
of heated air into the reducing gas steam (for partial
combustion).
The pressure zones are as follows:
Cutting—Ductile
1st retort—4.5 kg/cm2 abs
2nd retort—3.5 kg/cm2 abs
3rd retort—2.5 kg/cm2 abs
HyL III process. Hojalata Y. Lamina of South Africa
(HyL) developed the moving bed HyL III technology and
converted a few of the plants based on the earlier fixed-
bed or HyL I technology. After many years of research,
HyL announced a new process called HyL III (III signify-
ing the third generation of HyL reactors, though II was not
 1100
a separate process, but only an improvement of HyL I).[10]
The HyL III process is shown schematically in Fig. 8.
The main distinguishing features of the HyL III process
are as follows[25]:
1. High-pressure operation (5.5 kg/cm2) for better pro-
cess control and higher throughput.
2. Independent operation of the reforming and reduction
areas and independent recycling circuits in the reduc-
tion and cooling sections of the reactor; thus the
reformer and reformer catalyst are unaffected by sulfur
and dust.
3. The energy consumption varies between 9.0 and 10.0
GJ/t of DRI.
The principal change introduced (over HyL I) was the
replacement of the four fixed-bed reactors by a single
moving bed reactor. The HyL III process can operate with
100% pellets, 100% lump ore, or mixtures thereof; nor-
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
mally 70% pellets and 30% lump ore are charged. In some
cases, it is extremely important to add 5% nonsticking ore
in the feed,[10] in order to check the sticking tendency of
Cutting—Ductile
the pellets and thereby improve the furnace performance
in terms of uniform descent of the burden.
Reducing gas generation: Natural gas is preheated to
370 C (643 K) in a coil in the reformer convection section
and is then passed through a desulfurization system to
reduce the sulfur content to permissible levels for the
reformer catalyst operation. Desulfurization is necessary
because the nickel reformer catalysts are very sensitive to
even low concentrations of sulfur.[26] Now this gas is
mixed with a predetermined amount of superheated steam.
Fig. 8 Flow sheet of the HyL III process.
Direct Reduced Iron: Production
The steam to carbon ratio depends on the catalyst used, the
composition of gas, and the operating temperature. Typi-
cal values of this ratio range between 2.1 and 2.4. This
ratio has been chosen in order to prevent carbon deposit on
the catalyst.
The steam–natural gas mixture is preheated to about
620 C (893 K) in the reformer convection section and then
enters the reformer section. The reformer consists of a set
of stainless steel pipes packed with a nickel-based cata-
lyst. The reforming reactions are carried out at 830 C
(1103 K):
Cn Hð2nþ2Þ þ nH2 O⇄nCO þ ð2n þ 1Þ H2 (29)
At the same time the water–gas shift equilibrium is
established:
CO þ H2 O⇄CO2 þ H2 (30)
The overall reaction proceeds to equilibrium, which
depends on the temperature, pressure, and steam to carbon
ratio. The reformed gas leaves the reformer via the efflu-
ent chamber at 830 C (1103 K) and 7.8 bar pressure and
enters the reformer quench boiler where steam is gener-
ated by recovering sensible heat from hot reformed gas.
Here, the temperature of the reformed gas is lower than
300 C (573 K). Quenching of the hot reformed gas is
necessary to get a gas with a low content of water vapor.
The reformer product gas has typically the following com-
position (by volume): 73% H2, 17% CO, 6% CO2, 1%
H2O, and 3% CH4.
 Direct Reduced Iron: Production
Reduction: Before the fresh produced cold reformed
gas enters the reduction gas heater, it is mixed with the
stream of recycled gas and is heated up to the desired gas
temperature of 900–950 C (1173–1223 K). This gas is fed
directly to the reactor at a pressure of 5.5 bars. While the
process operates at high pressure, the feeding of the iron
ore to the reactor is achieved by means of the charging
scaling mechanism, which consists of a bin and a set of
special valves designed for the mass flow of lump ore or
pellets. The iron ore is fed through the top of the reactor
and flow downward by the effect of gravity. The hot
reducing gas is injected at the middle of the reactor and
flows upwardly that is counter-current to the solids’ flow.
The product DRI is cooled by counter-current cooling gas
in the cooling section from where the cooled product is
removed continuously.
The reducing gas leaves the reactor as the top gas at
approximately 450 C (723 K) and 5.0 bar. From the
reduction part of the reactor it is then scrubbed in the
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
reduction quench tower: (1) for condensating the vapor in
the top gas and (2) cleaning the gas from dust. Then it is
sent to the recycle gas compressor.
HyL IV M process. It is a shaft furnace process with
self-reforming, in which reforming of natural gas occurs
within the reduction reactor itself and the metallic iron in
DRI acting as the catalyst. This allows in situ reforming
of natural gas to proceed in parallel with the reduction of
iron oxide and carburization of DRI. No separate gas
reformer is required.[10] Fig. 9 is a schematic representa-
tion of the HyL IV M process. The HyL IV M process
employs a moving bed shaft furnace (similar to HyL III)
to reduce the iron ore pellets and lump ore at normal
reduction temperature and at intermediate pressure. The
process can produce cold/hot DRI as well as HBI. The
first HyL IV M plant started production in 1998 in Mon-
terrey, Mexico, hence the suffix M.
Heat
recuperator
Iron ore
PG
H2O
compressor
HYL
PG reactor
CO2 heater
Oxygen
Humidifier
Fuel
DRI
NG
1101
The process starts with the injection of natural gas,
which together with recycled gas, enters a humidifier,
where the required amount of steam is supplied. This gas
mixture then enters a heater where its temperature is
increased to above 900 C (1173 K). The hot gas mixture
along with oxygen is introduced at the bottom of the shaft
furnace. Oxygen helps in partial combustion of the reducing
gas to increase the temperature above 1020 C (1293 K). In
the lower part of the reduction zone, in situ reforming
reactions occur, when the hot gas comes in contact with
metallic iron in DRI. The reacted exit gas leaves from the
top of the reactor. Steam is condensed and removed before
the gas is compressed and sent for CO2 removal, after
which it is mixed with natural gas for further use.
Purofer process
Counter-current reduction of ores in a shaft furnace with
gases is obtained by reforming hydrocarbons with oxidiz-
ing gases. In catalytic reforming of natural gas with air,
the temperatures required is reached by the stored regen-
Cutting—Ductile
erative heat of the reforming chamber and the catalyst as
well by the reaction heat of the partial oxidation of meth-
ane. Reforming with air has the advantage[25] that N2
contained in the reducing gas acts as the heat carrier for
supplying heat to the reduction shaft.
Reforming of natural gas with recycled top gas gives
better overall utilization of the reducing gas. CO + H2
components of the top gas are heated in the reformer and
recycled to the shaft; moreover, CO2 replaces O2 of air as
the reforming medium. Reforming reactions with natural
gas and oxygen (air), steam, or recycled top gas as the
reforming medium are as follows:
ðAirÞ CH4 ðgÞ þ 1=2 O2 ðgÞ ¼ COðgÞ þ 2 H2 ðgÞ (31)
ðTop gasÞ CH4 ðgÞ þ CO2 ðgÞ ¼ 2COðgÞ þ 2 H2 ðgÞ (27)
H2O
CG
compression
unit
Fig. 9 Flow sheet of the HyL IV M process.
 1102
CH4 ðgÞ þ H2 OðgÞ ¼ COðgÞ þ 3 H2 ðgÞ (28)
The two reformers operate alternately. One reformer is
heated by burning part of the top gas to a temperature of
about 1400 C (1673 K). The heat is stored in the mass of
the catalyst.
Finmet process
The fluidized iron ore reduction (FIOR) process was
developed for the continuous reduction of iron ore fines
by reformed natural gas in a train of fluidized bed reactors.
The original FIOR process was substantially improved,
resulting in the development of the Finmet process. The
Finmet process uses a train of four fluidized bed reactors
in which the gas and solids moving in counter-current
directions come into contact throughout the entire reactor
train. The feed concentrate (1.2 · 103 m iron ore fines)
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
is charged to the reactor train at the topmost reactor via a
pressurized lock hopper system. The ore fines are fed
continuously to the reactors out of the lower lock hopper,
Cutting—Ductile
which is always maintained at the reactor pressure of
11–13 bar. In the topmost reactor, the feed is preheated to
550–570 C (823–843 K) by the reducing gas coming from
the lower reactor. Preheating, dehydration, and reduction
of different stages take place as the feed is transferred
downward to the subsequent reactors. The gas required
for reduction is a mixture of recycled top gas and fresh
reformer makeup gas provided by a standard steam
reformer. The composition of the recycled gas entering
the last reactor is adjusted in accordance with the desired
carbon content of the product. The temperature in the last
reactor is maintained at around 780–800 C (1053–1073
K) so that the final reduction to 93% metallization takes
place.[10] Hot, fine DRI at around 650 C (923 K) is
transported to the briquetting machine by a sealed system
for hot brequetting to HBI.
HIB process
The HIB fluidized bed DR process is a modified version of
the Nu-iron process that was developed by the United
States Steel Corporation in 1953. The HIB plant was
started at Venezuela in 1968, and in 1971 the plant was
ready to go into operation.[19] In the HIB process, fine iron
ore is reduced in a two-stage fluidized bed reactor and then
the product in a hot briquetted form. Steam natural gas
reformers are used to generate the reducing gas. These
are followed by heat exchangers to cool the gas and
remove the unreacted water vapor. The gas is then heated
to the desired temperature in a direct fired tubular heater.
The off-gas from the reduction process is cleaned and used
as a fuel for the reformer and the reducing gas heater.
The reduction reactor is divided into two stages.
Reduction of Fe2O3 to FeO is accomplished in the first
Direct Reduced Iron: Production
stage by contact with spent gas from the second stage.
Reduction of FeO to Fe is accomplished in the second
stage by contact with fresh reducing gas. The fine reduced
product from the reactor is carried by an inert gas pneu-
matic lift to a sealed surge bin above the briquetting rolls.
Screw feeders extract the fines from this bin, precompact
the fines, and feed them into the briquetting rolls. Under-
sized material is screened out and recycled to the surge
bin, and the briquettes are transported to a shaft cooler
where they are cooled with inert gas.
CHARACTERISTICS AND USES OF DRI
Characteristics of DRI
DRI has been fast gaining ground throughout the world
since the 1980s mainly because of the shortage of coking
coal for BF and steel scrap in steelmaking. DRI is pro-
duced by the DR of iron ore by using noncoking coal/
natural gas. The major part of DRI production is used as a
substitute for scrap in steelmaking. DRI is consumed in
three primary product forms, namely, lump, pellets, and
hot briquettes (as shown in Fig. 10). The other secondary
product form is cold briquettes made from DRI fines. The
hot briquette form is popularly known as HBI. HBI is a
combined solid form of DRI lump and pellets, hot pressed
at 700–800 C (973–1073 K), immediately after its pro-
duction. The most important characteristics of HBI are its
high density and lower specific area, which improves the
resistance to reoxidation and makes it easier to handle.
Due to high density, the charging of HBI in furnace is
much easier and melting is faster. Tables 6 and 7 compare
the chemical composition and physical properties of dif-
ferent forms of DRI.[27] Being uniform in size, DRI is easy
(A) (B)
(C)
Fig. 10 Different forms of DRI: (A) lump, (B) pellets, and
(C) HBI.
 Direct Reduced Iron: Production 1103

Table 6 Typical chemical composition of direct reduction iron. usually give low carbon content (<0.5%), and this is
Product forms inherent to the process itself. However, with the gaseous
Chemical Coal based Gas based reduction processes, the carbon level can be adjusted,
composition Lump Pellets HBI within limits, to any desired value. The carbon content of
Fe (m) (%) 82–87 83–88 85–88 gas-based DRI is varied between about 1.0% and 3.0%,
Fe (t) (%) 91–93 91–93 91–93 corresponding to a degree of metallization of 85% to 95%,
Metallization (%) 88–93 92–95 93–95 respectively.[28] The carbon as Fe3C is more desirable than
C (%) 0.15–0.20 1.0–2.5 1.0–1.5 loose carbon fines or soot that may not be useful to the
S (%) 0.01 0.005–0.015 0.002–0.006 process. Oxygen present in the DRI is in the form of FeO,
P (%) 0.05 0.02–0.04 0.02 which reacts vigorously with carbon in DRI within the
Total gangue 4.7 1.8–4.0 3.5–4.0
molten bath and improves heat transfer, slag metal contact,
materials (%)
and homogeneity of the molten bath. Thus, higher percent-
age of carbon is required in DRI, and hence steelmakers
prefer gas-based DRI (which contains 1.0–2.5% C) instead
to transport and handle and permits continuous charging in of coal-based DRI (0.2% C).[6] Any unreduced iron oxide
electric furnace steelmaking. present in DRI, during steelmaking, enters the slag.
The DR process consists of removing the larger part of
the oxygen contained in the iron oxide and some sulfur,
Gangue content
but otherwise no other refining and melting occur in the
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016

process. As a consequence, all of the original impurities of

The gangue in DRI that substitutes for normal slag-pro-
the ore, that is, all of the gangue, remain in the product,
ducing agents does not penalize the steelmaking operation
which stresses the importance of having high-quality, feed

materials.
The main chemical characteristics of DRI are[28]
1. Metallization
2. Carbon content
3. Gangue content
4. Impurities (S, P, etc.) and residual elements
Metallization
The degree of reduction of DRI is usually expressed as the
percent metallization of the product. It is the ratio of the
metallic iron present in DRI divided by the total iron pre-
sent in DRI. The degree of metallization depends greatly on
the type of reduction process being used. The degree of
metallization varies from 85% to 95% depending on the
process adopted for DRI production.[28] Low degree of
metallization leads to economic disruption such as higher
energy consumption, higher slag volume, more heat time,
and lower yield[6] during steelmaking.
Carbon content
The control of the carbon content of DRI depends on the
DR process being used. Those based on solid reductants
Cutting—Ductile
and is usually between 2% and 4%. However, the actual
amount depends on the proportion of silica (SiO2) in the
gangue that must be fluxed and on the percentage of iron
units in the charge that is derived from DRI.[24] All the
gangue content of the iron ore fed to the DR plant is finally
landed to the DRI. Since DRI has become competitive
with steel scrap, the gangue content of the DRI had to be
reduced appreciably. The iron ore feeds for DR processes
usually contain less than 2% SiO2 and 1% Al2O3. More-
over, certain additions of either limestone or dolomite are
made to improve the behavior of the iron ore pellets dur-
ing reduction.[28] Therefore, some lime and magnesia are
found in DRI.
Impurities and residual elements
The residual elements usually found in steel scrap (Cu, Zn,
Pb, Sn, As, Cr, Ni, and Mo) are not often present in any
noticeable quantities in iron ore deposits, and therefore,
these elements are found only as traces (<0.02%) in DRI.
The impurities that can be found in DRI are the alkalis
(Na2O, K2O), titanium oxide, sulfur, and phosphorus. The
first three are usually present in quantities, which are small
enough not to have any bearing on the steelmaking prac-
tice [e.g., (Na2O + K2O) < 0.1% and TiO2 < 0.1%].[28]
The sulfur content of DRI is also relatively low, depending

Table 7 Physical properties of direct reduction iron.

Product Carbon Metallization Apparent density Bulk density Shape and size Weight Relative fine
forms (%) (%) (kg/m3) ( · 103) (t/m3) (m) ( · 103) (kg) ( · 103) generation
Lump 0.2 max 93 max 3.0 1.7 Irregular, 3–25 3–5 More
Pellets 1.0 min 92 min 3.5 1.6–2.0 Spherical, 4–20 3–4 Less
HBI 1.0–1.5 93.5 5.0 2.6–2.7 Pillow-like, 35 · 50 · 450–750 Minimum
110
 1104 Direct Reduced Iron: Production

on the amount of sulfur in the fuels and reductants used.
Sulfur contents of DRI vary from less than 0.005% in the
DR processes employing sulfur-free gas to approximately
0.02% in DR processes employing sulfur-bearing coal and
limestone together with iron ore in the charge mix.[24]
Phosphorus is not eliminated during the DR process and
therefore the quantities found in DRI will be directly a
function of those contained in the iron ore. Phosphorus
content[28] normally found in DRI ranged between 0.01%
and 0.04%.
Quality of DRI
The portions of iron oxide and gangue in DRI above cer-
tain minimums increase the power requirements in the
EAF compared to an equivalent quantity of scrap. A por-
tion of the iron oxide reacts with carbon in the EAF to
produce metallic iron and carbon monoxide according to
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
processes and are usually not a factor in electric steelmak-
ing operations.
Reoxidation, Storage, and Transportation of DRI
DRI is produced, in the solid state, by gaseous or solid
reduction of iron ore. Due to the removal of oxygen, a
honeycombed structure remains that has a very large sur-
face area per unit weight. Hence, DRI has an inherent
tendency to reoxidize back to its native stage. The main
reasons for such behavior are: (1) extremely high surface
area to volume ratio and (2) poor thermal conductivity due
to porosity and gangue content of DRI (approx 2.092 kJ/
mhK).[29] Freshly prepared DRI is highly prone to oxida-
tion whenever it comes into contact with air. The heat
generated in the oxidation reaction increases the suscepti-
bility of oxidation, thereby starting a sort of chain reaction
and ultimately leading to the burning of DRI.[30] Hence,

the following reaction: storage and handling of DRI is a major concern of DRI.

FeO þ C ¼ Fe þ CO ΔHo 298 ¼ 153:888 kJ=mol of C Reoxidation of DRI

Cutting—Ductile

(32)
This is an endothermic reaction. Furthermore, the gangue
and the associated flux require energy for melting. How-
ever, a metallization that is too high decreases both the
fuel efficiency and the productivity of the DR process. The
metallization of DRI normally ranges between 85% and
95% depending on the process and on the reducibility of
the original iron oxide. Based on the level of metalliza-
tion, the carbon content of the DRI is controlled in some
DR processes to be between 1.0% and 2.5% to facilitate
reduction of FeO by carbon during melting. This increases
iron recovery and the CO generated promotes foamy slag
practice in the EAF. Another impact on the operation of
the EAF pertains to fines in the DRI that affect iron recov-
ery, increase the dust loading, and contaminate electrical
parts. The allowable fines are usually approximately 5% at
5 · 103 m. A major advantage of DRI in EAF opera-
tions is the absence of contaminating residuals such as
copper and tin. Phosphorus, manganese, and vanadium
contained in the ore remain in the gangue in the gas-based
The reoxidation of DRI, caused by the influence of water
and oxygen, is an exothermic reaction. Thus, an oxygen
pickup of 0.1% would mean a rise of about 35 C (308 K)
under adiabatic conditions. This makes DRI highly reac-
tive in oxygen containing atmosphere. It is even more
prone to reoxidation in the presence of moisture. Due to
the poor thermal conductivity of DRI,[29] the heat gener-
ated (as shown in Table 8) during reoxidation reaction
cannot be dissipated away, ultimately leading to the devel-
opment of hot spots or even autoignition of DRI piles.
DRI is chemically reactive and this makes it dangerous
to store or ship in bulk. In the case of some stockpiles that
have wet DRI buried deeply, high-temperature reoxidation
reaction has occurred. The heat generated by the exother-
mic reoxidation reaction (Table 8) within the pile does not
have an opportunity to dissipate. As a result, increase in
temperature causes a corresponding increase in reaction
rates, leading to autoignition. Overseas transportation of
DRI has therefore often resulted in corroding the cargo,
even sometimes to the extent of setting ships ablaze, due

Table 8 Heat effects during reoxidation of DRI.

Reaction Heat of formation at 25 C (298 K) and 1.0 atm Amount of heat produced per kg DRI
I Exothermic kcal/mole (kJ/mol) kcal/kg (kJ/kg)
0.95 Fe + ½ O2 = Fe 0.95 O 63.2 (264.43) 1188 (4970.59)
3Fe + 2 O2 = Fe3O4 266.9 (1116.71) 1589 (6648.38)
2Fe + 3/2 O2 = Fe2O3 196.3 (821.32) 1753 (7334.55)
Fe + ½ H2O + 3/4 O2 = FeOOH 133.6 (558.98) 2386 (9983.02)
Fe + H2O + ½ O2 = Fe(OH)2 135.8 (568.19) 2425 (10146.20)
Fe + 3/2 H2O + 3/4 O2 = Fe(OH)3 197.2 (825.08) 3521 (14731.86)
II Endothermic
3 Fe + 4 H2O = Fe3O4 + 4 H2 +6.6 (27.61) +39.3 (164.43)
Fe + 2H2O = Fe(OH)2 + H2 +0.8 (3.35) +14.3 (59.83)
 Direct Reduced Iron: Production
to the initial reaction between DRI and seawater yielding
hydrogen and heat.
Chemical reactions involved in the reoxidation of DRI
are as follows[29]:
(i) Oxidation in dry air
4Fe þ 3O2 ¼ 2Fe2 O3 (33)
(ii) Oxidation in the presence of moisture
2Fe þ 3H2 O ¼ Fe2 O3 þ 3H2 (34)
(iii) Oxidation in the presence of dissolved oxygen
2Fe þ 2H2 O þ O2 ¼ 2FeðOHÞ2 (35)
The following recommendations were made to prevent
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
DRI burning[30]:
 DRI should be kept dry at all times. DRI should be
naturally aged for at least 72 hours before loading into
the ship. Natural passivation is caused by the formation
of an oxide layer on the surface of the DRI that stops
further oxidation. DRI should not be loaded if the
temperature is greater than 65 C (>338 K).
 Naked flame should be removed from the storage
areas.
 Nitrogen purging should be done in the ship hutch.
 Hydrogen and oxygen concentration should be
measured.
 A higher degree of metallization for DRI should be
preferred.
 Precautions should be taken during loading:
 Loading, steering, and unloading during rain must
be avoided; this is extremely important.
 Holds must be dry and clean and shall be sealed and
watertight.
 Holds must not contain steam lines or other heat
sources.
 Holds must have watertight hutches to avoid
ingress of water during voyage; seawater is partic-
ularly very aggressive.
Storage and transportation/shipment of DRI/HBI
The International Maritime Organization (IMO) Sub-
Committee on Dangerous Goods, Solid Cargoes and Con-
tainers (DSC) recommended certain amendments to the
entry for DRI in the IMO Code of Safe Practice for Solid
Bulk Cargoes (the BC Code).[31] The new amendments
were agreed in September 2008 and were adopted by
Maritime Safety Committee (MSC) resolution in Decem-
ber 2008 as part of the new International Maritime Solid
Bulk Cargoes (IMSBC) Code, which was published in
1105
2009. The code was recommended on January 1, 2009,
but became mandatory on January 1, 2011.
DRI (usually in the form of pellets or lumps) has a
hugely porous structure, which makes the material
extremely reactive and prone to reoxidation on contact
with air and/or moisture. These oxidation reactions cause
self-heating in the stow, which can lead to auto-oxidation
where cargo temperatures in excess of 900 C (1173 K)
can be generated. Moreover, contact with moisture
evolves hydrogen, an extremely flammable and sensitive
gas that has caused explosions in the holds of several ships
following its ignition.
Directly reduced iron is highly susceptible to oxidation
and rusting if left unprotected and is normally quickly
processed further to steel. The bulk iron can also catch fire
since it is pyrophoric. The material, if wetted, may oxidize
rapidly and generate heat over a period of time. In addi-
tion, hydrogen may be generated, which could possibly
form an explosive atmosphere.
Molded briquettes and specially processed materials
minimize certain risks. However, every effort should be
made to prevent the ingress of water into cargo compart-
Cutting—Ductile
ments. Flammable gases should be prevented from enter-
ing adjacent enclosed spaces. Prior to shipment, DRI
should be aged for at least 72 hours or treated with
some passivation technique to reduce its activity to at least
the same level of the aged product. Holds should be
maintained under inert atmosphere (less than 5% oxygen
and less than 1% hydrogen) throughout the voyage.
The hazards of DRI. DRI is classified as a Group B
cargo according to the BC Code, that is, a cargo that
possesses a chemical hazard that could give rise to a dan-
gerous situation on a ship.[32] The principal hazards of all
cargoes of DRI and its derivatives are twofold. First, they
react with the oxygen present in the air, thereby producing
heat and depleting the atmosphere of oxygen that is essen-
tial for life. With DRI (B), the former effect can run away
in spectacular fashion, leading to auto-oxidation of the
iron in which the stow becomes incandescent as the tem-
peratures approach 1000 C (1273 K). The cargo is then
said to be on fire and burning much in the way that char-
coal behaves.
While self-heating is dangerous and alarming, it is a
gradual and progressive event that can often be diagnosed
early, affording masters time to consult the association,
obtain advice from suitably qualified experts, and institute
suitable safety measures. Self-heating to dangerously high
temperatures can be successfully prevented in most prac-
tical applications by compressing the DRI into HBI at the
manufacturing stage, DRI (A). Association is not aware of
any fine cargoes that have undergone self-heating to the
auto-oxidation stage, although there is no theoretical rea-
son why 4 · 103 m pieces of DRI will not self-heat to
auto-oxidation.
 1106
The second hazard is again related to the reactivity of
iron, this time with moisture or water. The result is the
generation of hydrogen gas, which is explosive over a very
wide range of concentrations and, in practical situations,
displays an alarming readiness to be ignited. Explosions of
hydrogen in air are extremely violent and rapid: Flame
speeds may exceed 1000 m/sec and pressures exceeding
8 bar may be generated in a fraction of a second. An
unfortunate master has no time in which to react to a
hydrogen (or any other gas) explosion.
All three types of materials will evolve hydrogen to a
greater or lesser extent on contact with water. Seawater is
more aggressive than freshwater, producing a greater
degree of self-heating and higher evolution rates of
hydrogen.
In 2003, the Adamandas (cargo ship) was deliberately
sunk by the French Authorities following overheating of
its cargo of 21.0 Mt of DRI pellets. On February 28, 2004,
hydrogen explosions in four of its cargo holds caused the
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
total loss of Ythan (the cargo ship) off Colombia, together
with the death of six of its crew. The hydrogen had been
produced from a cargo of damp DRI fines.
Cutting—Ductile
The following is a list of those countries known to have
exported DRI and whether they are known to be currently
shipping (Tables 9 and 10).
Classification of DRI according to the IMSBC Code.
The IMSBC Code now categorizes three types of DRI[33]:
1. The first type, DRI (A), is the less reactive, high-density
variety of DRI that is known as HBI, or hot molded
briquettes (HMBs). A briquette will typically measure
about 120 · 103 · 90 · 103 · 30 · 103 m thick
and should have a density greater than 5000 kg/m3.
2. The second type, DRI (B), is a highly reactive, low-
density DRI in the form of lumps and pellets (6 · 103
to 25 · 103 m in size) and cold molded briquettes.
This material was originally entered in the BC Code
under BC No. 015. It may now be carried under an
inert gas atmosphere.
3. The schedule now included a new entry, DRI (C),
which is described as by-product fines and is intended
Table 9 High-density type, DRI (A).
Country Product Comments
India Briquettes Fire in 1994, excess fines present
Russia/Black Briquettes Of concern because of excessive
Sea load ports hydrogen production when cargo
loaded wet in 2001
Venezuela Briquettes Fire in 1985, possibly due to excess
and fines of fines. Also exporting fines that
give serious concern following a
number of hydrogen explosions
that have caused several deaths and
the total loss of one vessel
Direct Reduced Iron: Production
Table 10 Low-density type, DRI (B).
Country Product Comments
Germany Pellets Passivated but serious fire broke out
in a cargo of the passivated product
in 1981
Nigeria Pellets One known incident of self-heating
during loading in 2006
Russia/Black Pellets Of great concern because there have
Sea load ports been a large number of fires since
1990 in this product, most recently
to our knowledge in 2003, despite
claimed passivation
Trinidad Pellets Of concern following total loss of
and fines vessel after losing inert atmosphere
on voyage in 2003 and explosion
while carrying fines in 2004
United States Fines Of concern because of fatal
explosions in 1999
Venezuela Fines Serious concern following a number
of hydrogen explosions that have
caused deaths and the total loss of
one vessel
Germany Pellets Passivated but serious fire broke out
in a cargo of the passivated product
in 1981
to include all the materials generated as by-products in
the manufacture and handling process of DRI (A) and/
or DRI (B).
The following recommendations should be acted upon
before loading all forms of DRI:
a. Shippers should certify that the material conforms to
the requirements of the BC Code, 2005 edition.
b. The hatches should be inspected and hose tested before
loading to ensure they are watertight. All cargo spaces
should be clean and dry.
c. The ship shall be provided with a detector suitable for
measuring hydrogen in an oxygen-depleted atmosphere
and for use in a flammable atmosphere. Closed cargo
spaces containing any form of DRI should be assessed
for safe oxygen levels prior to personnel entry.
d. Only certified safe electrical equipment and associated
wiring should be installed in any cargo space.
e. The temperature of the cargo should be monitored
during loading and recorded in a log. If the temperature
is in excess of 65 C (338 K), sailing should be post-
poned until it is clear that it is falling.
f. Cargo temperatures and hydrogen concentrations in
hold atmospheres are to be monitored on voyage.
g. During cargo handling, “No Smoking” signs are to be
posted on decks.
h. Loading and transfer between ships during rain are
unacceptable.
i. All records of measurements are to be retained on
board for two years.
 Direct Reduced Iron: Production
DRI (A) and (B) cargoes typically contain about 85%
metallic iron, whereas in blends containing DRI (C), it
can be as low as 1% or 2%. Such blended cargoes should
be regarded as the hazardous commodity DRI (C). Cargo
blends containing DRI (C) can be identified by their chem-
ical composition. The chemical composition must include
the total iron content (TFe), the metallic or free iron con-
tent (Fe), and the moisture content. This information
should preferably be supported by a certificate that should
state the methods and standards that have been followed
when obtaining the samples that have been tested (prefer-
ably ISO 10835:2000) and the standards that have been
followed to determine the metallic iron content (prefera-
bly ISO 5416:2006).
The entry for DRI (C) specifies the following:
Prior to shipment, the cargo shall be aged for at least 30
days, and a certificate confirming this shall be issued by a
competent person recognized by the national administra-
tion of the port of loading stating that the cargo is suitable
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
for shipment. After loading, a certificate shall be issued
confirming that throughout the whole consignment of
fines and small particles, the moisture content has not
exceeded 0.3% and the temperature does not exceed 65 C
(>338 K).
The following are recommendations for the loading
and carriage of high-density DRI (A):
HBI or HMBs
a. The maximum limit on the moisture content is 1%.
b. The main hazard arises from the evolution of hydrogen
if water enters the stow. For this reason, effective venti-
lation should be carried out whenever possible. When
mechanical ventilation is used, the fans shall be certified
as explosion proof and shall prevent spark generation.
c. The concentration of hydrogen in the hold atmosphere
should be regularly monitored using a suitable instru-
ment. Advice should be sought if the hydrogen concen-
tration exceeds 1% by volume in air.
d. Any dust accumulated on decks or elsewhere during
loading or discharge should be washed off as soon as
possible to prevent adhesion.
e. On completion of loading and prior to sailing if
weather and circumstances permit, the hold should be
left open to allow the cargo to cool and to allow the
dissipation of any moisture or hydrogen that may have
been evolved.
f. Slight emissions of steam are normal and should clear
with adequate ventilation.
The following are recommendations for the loading and
carriage of low-density DRI (B):
Lumps, pellets, cold molded briquettes
a. The average particle size is from 6.35 · 103 to 25 ·
103 m.
b. Loading conveyors are to be dry.
1107
c. Prior to loading, an ultrasonic test or another equiva-
lent method shall be conducted to ensure weather tight-
ness of the hatch covers and closing arrangements.
d. The moisture content must be less than 0.3% and must
be monitored during loading.
e. Any wet cargo shall be discharged without delay.
f. The ship shall be provided with reliably measuring the
temperature at several points within the stow, deter-
mining the concentrations of hydrogen and oxygen.
g. The ship shall be provided with the means to ensure
that the oxygen concentration below 5% can be
achieved throughout the voyage.
h. The ship shall not sail until the master and a competent
person are satisfied that
 All loaded cargo spaces are correctly sealed and
inerted
 The cargo temperatures have stabilized at all mea-
suring points and are less than 65 C (338 K)
 The concentration of hydrogen in the free space has
stabilized and is less than 0.2% by volume (i.e., 5%
of the lower explosive limit)
i. The oxygen concentration shall be maintained at less
Cutting—Ductile
than 5% throughout the duration of voyage
The following are recommendations for the loading and
carriage of the DRI (C):
By-products and fines
a. The average particle size is less than 6.35 · 103 m,
and no particles should be greater than 12 · 103 m in
size.
b. The reactivity of the cargo is extremely difficult to
assess due to the nature of the material that can be
included in the category.
c. The IMSBC Code schedule for DRI (C) sets the max-
imum allowable moisture content as 0.3% and the
maximum permissible temperature of 65 C (338 K).
Usages of DRI
With the high prices of steel scrap in the international
market, coupled with the difficulty in the importing pro-
cess for small-scale units, DRI has become the most viable
alternative to steel scrap. In view of shortage of steel scrap
and high prices of scrap, the secondary steel producers
have to depend on DRI as the metallic feed material to
produce steel. It is felt that a considerable quantity of DRI
will be indigenously available to meet the demand of
metallic feed material of secondary steel producers. Using
DRI in EAF increases furnace productivity and reduces
both electrode and power consumption. Attaining the best
melting results also requires control of the rate at which
DRI is fed to the EAF. The feed rate is primarily depen-
dent upon the effective power input but is also influenced
by the DRI composition, bath temperature, and thermal
losses. The use of DRI is not confined to only EAF/IF as
 1108
feed material. Further, charging of DRI in BF shows dras-
tic changes in the yield and furnace efficiency, creating
more space for DRI[34] and resulting in increased produc-
tivity and reduced coke rate. DRI is also being charged in
BOF steelmaking as coolant. DRI with a lower degree of
metallization has more cooling effect. The use of DRI has
yielded better results compared to scrap cooling. If scrap
gives 1.0 unit of cooling effect, DRI gives 1.2 units of
cooling effect, that is, it is 20% more efficient.[35] DRI
also partially substitutes the pig iron in cupola for making
cast iron. DRI is also used for diluting the residual ele-
ments, in steelmaking, which are present in the steel scrap,
to produce good-quality steel.
The advantages of DRI as a feed material can be sum-
marized as follows:
1. DRI has a known and uniform chemical composition.
2. It has a uniform size and is capable of continuous
charging.
Downloaded by [Sujay Kumar Dutta] at 22:29 29 June 2016
3. It has very low levels (0.02%) of residual elements
(like Cu, Sn, Ni, Cr, Mo), whereas scrap and pig iron
contain 0.13–0.73% and 0.06% residual element,
Cutting—Ductile
respectively.
4. It maintains low levels of sulfur.
5. Unreduced iron oxide in DRI reacts vigorously with
carbon (i.e., carbon boil) in molten bath that improves
heat transfer, slag metal mixing, homogeneity, and low
content of dissolved gases of the bath.
6. It is capable of forming a protective cover of the arc by
foamy slag in the EAF’s bath.
7. It has a potential of sensible heat recovery from waste
gases.
8. DRI has a predictable price structure than scrap
because DRI is a product and scrap is a by-product.
REFERENCES
1. Dutta, S.K.; Roy Chowdhury, P.J. IE(I) J.—MM 1986,
66(2), 91.
2. Prasad, K.K. Steel, Metall. 2004, 6(8), 29.
3. Mehta, P. Special Report, 1987, Personal communication.
4. Chatterjee, A. In Proceedings of International Conference
on Alternative Routes of Iron and Steelmaking, Perth, Aus-
tralia, 1999; 3.
5. 2013 World Direct Reduction Statistics, Direct from
Midrex, 2nd Quarter 2014, 2.
6. Dutta, S.K.; Roy Chowdhury, P.J.; Gadgeel, V.L. IE(I) J.—
MM (Spl) 1988, 68(2), 52.
7. Vijayavergia, R.K.; Chauhan, G.I.S. Awakening of the
Indian Steel Industry: 2005–2020; Chatterjee, A., Ed.;
IIM: Jamshedpur, India, 2006; 27.
8. Chatterjee, A.; Sathe, A.V.; Singh, R.; Chakravarty, P.K.In
Proceedings of Symposium on Direct Reduction Processes
in Iron and Steelmaking, 36th ATM of IIM: Rourkela,
India, 1982; 133.
9. Schlebusch, D.; Choudgury, L.R.; Dayal, R. Proceedings of
Symposium on Direct Reduction Processes in Iron and
View publication stats
Direct Reduced Iron: Production
Steelmaking. 36th ATM of IIM: Rourkela, India, 1982;
166.
10. Ghosh, A.; Chatterjee, A. Ironmaking Steelmaking: Theory
and Practice; Prentice-Hall of India Pvt Ltd: New Delhi,
India, 2008, Chap. 13, 225.
11. Dutta, B.B. IE(I) J.-CH 1993, 73, 69.
12. Pandit, A.; Sarangi, B.M.; Kesava Babu, A.; Sheshadri, M.
K. Steel World, 2003.
13. Dutta, S.K.; Ghosh, A. Trans. Indian Inst. Met. 1995,
48(1), 1.
14. Dutta, S.K.; Sah, R.; Chokshi, Y. Iron Ore–Coal/Coke
Composite Pellets; LAMBERT Academic Publishing,
Saarbrücken: Germany, 2013.
15. Dutta, S.K.; Sah, R. Proceedings of Asia Steel International
Conference, Jamshedpur, India, 2003. Vol. I, 1.d.4.1.
16. Prasad, K.K.; Ray, H.S. Advances in Rotary Kiln Sponge
Iron Plant; New Age International Publishers: New Delhi,
India, 2009, Chap. 3.
17. Kubaschewski, O.; Alcock, C.B. Metallurgical
Thermochemistry, 5th Ed.; Pergamon Press: Oxford, U.K.,
1979.
18. Dutta, S.K.; Sah, R. Alternate Methods of Ironmaking
S. Chand & Co Ltd: New Delhi, India, 2012.
19. Stephenson, R.L.; Smailer, R.M., Direct Reduced Iron:
Technology and Economics of Production and Use; The
Iron & Steel Society of AIME: USA, 1980.
20. Thatoi, K.C.; Sen Singh, V.; Sohal, G.D.S. Proceedings of
Conference on Direct Reduction and Direct Smelting.
2001. Jamshedpur, India, 2011, 11.
21. Kosuke, S. Proceedings of International Seminar on Alter-
native Routes for Ironmaking in India, Kolkata, India,
2009, 39.
22. Klawonn, R.M.; Hoffman, G.E. Direct from Midrex, 2nd
Qtr. 2005; 4.
23. Chatterjee, A. Proceedings of International Seminar on
Alternative Routes for Ironmaking in India, Kolkata, India,
2009, 5.
24. MSTS Iron making; The AISE Steel Foundation:
Pittsburgh, PA, 1999; Chap. 11, Vol. I, 741.
25. Ganguly, A.; Gupta, K.N.; Aswath, H.S. Symposium on
Direct Reduction Processes in Iron and Steelmaking, 36th
ATM of IIM, Rourkela, India, November 1982, 90.
26. Gerigk, E.; Knop, K.; Pena, J.M. Symposium on Direct
Reduction Processes in Iron and Steelmaking, 36th ATM
of IIM, Rourkela, India, November 1982, 192.
27. Dutta, S.K.; Lele, A.B. Miner. Met. Rev. 1997, 23(1), 79.
28. Taylor, C.R. Ed. Electric Furnace Steelmaking, Iron &
Steel Society: AIME, Chap. 10, 115.
29. Dutta, S.K.; Lele, A.B Tool & Alloy Steels, 1993, 27(11),
349.
30. Paswan, M.; Mukherjee, C. IIM Metal News February
2009; 6.
31. http://www.skuld.com/topics/cargo/dangerous-cargo/dri/
Loss-Prevention-Advice-on-the-Carriage-of-Direct-Reduced-
Iron-and-its-Derivatives.pdf.
32. http://www.intercargo.org/pdf_members/intercargo on dri.
pdf.
33. http://www.cargohandbook.com/index.php title Direct_
Reduced_Iron.
34. Dutta, S.K. Perform. Rev. Iron Steel, JPC 2007–2008; 31.
35. Dutta, S.K. JPC Bull. Iron Steel 2006, VI(5), 17.