182 COAL AND COKE

COAL AND COKE

J A Pajares and M A Dı́ez, Instituto Nacional del Carbón peat, lignite, subbituminous coal, bituminous coal,
(INCAR), Consejo Superior de Investigaciones Cientı́ficas, semianthracite, and anthracite.
Oviedo, Spain The stage of coal alteration gives rise to the con-
& 2005, Elsevier Ltd. All Rights Reserved. cept of rank – the greater the alteration, the higher
the rank of the coal. Thus, lignite and subbituminous
are described as low-rank coals, bituminous are
termed medium-rank coals, whilst semianthracite
Introduction and anthracite are termed high-rank coals. Coals of
higher rank than lignites are often referred to as hard
Coal may be regarded as a combustible sedimentary
coals.
rock. Whereas the majority of rocks are mostly
In addition to rank, two other parameters are im-
inorganically derived, coal consists predominantly of
portant in coal classification, type and grade. Type of
organic matter, largely derived from a variety of
coal refers to the amount and sort of organic con-
plant remains (higher plants, ferns, fungi, algae) and
stituents called ‘macerals’ (see below), whereas grade
different tissues (leaves, stalks, woody trunks, bark,
refers to the relative purity or the amounts of ash-
pollen, spores, sclerotia, resins, etc.) with associated
forming minerals found in coal.
mineral constituents.
Coal mines are spread over 100 countries. Mining
In the first stage of the evolution of coal from
may be surface or underground depending on whe-
vegetable matter – diagenesis – changes in composi-
ther the coal seam is exposed or at high depth. Known
tion were due to biochemical processes, involving the
coal reserves are estimated over 1013 tons and, at
degradation in a humid environment of the cellular
current production levels, workable seams are esti-
structures and of the organic complex of the vegeta-
mated to cover consumption beyond 225 years.
ble matter by activity of fungi and bacteria. Further
About 90% of these reserves are in the Northern
alteration is a geochemical process – catagenesis –
Hemisphere, with three countries – USA, China, and
combination of the effects of pressure (due to over-
the former Soviet Union – accounting for about two-
lying sediments) and mainly of heat (subsidence
thirds of the total.
brings a rise of temperature) over a prolonged peri-
World hard coal production reached 3837 million
od. Evolution in chemical composition results in
tons in 2002, with a 50% rise in the last 25 years.
losses of water and gases (CH4, CO2) and, conse-
Production of lignites was 897 Mt, Germany being
quently, a reduction of hydrogen and oxygen con-
the largest producer (B20% of the total). Data on
tents, a decrease in volatile matter content, and a
the major producer countries are given in Table 1.
progressive enrichment in carbon content. Along
Production of hard coal is increasing in Australia,
these processes – coalification (Figure 1) – the
South Africa, India, and in some new countries in the
original plant remains were progressively altered to
hard coal market, such as Indonesia, Colombia, and
Venezuela. On the other hand, production is gene-
rally decreasing in those countries belonging to the
Volatile
Rank C H O matter European Union.
Swamp

Peat Table 1 Coal production. Major producers of hard coal (2002)

Country Production (Mt)

Lignite
China 1326
USA 917
Bituminous India 334
Australia 276
South Africa 223
Russia 164
Poland 103
Anthracite
Indonesia 101
Ukraine 83
Kazakhstan 71
Figure 1 Scheme of the formation of coal in terms of rank
(coalification series). Data from World Coal Institute.
 COAL AND COKE 183

Table 2 International hard coal trade. Major coal exporters new plants with pressurized fluid bed reactors or
(2002) supercritical condition boilers. About 40% of the
Country Exportation (Mt) world’s electricity production is based on coal; in
countries such as China, Australia, or South Africa,
Australia 198
more than 80% of electricity needs are met with
China 86
Indonesia 73 using coal. Indeed, electricity from coal, with all
South Africa 69 restrictions derived from the Kyoto protocol, still
Russia 45 meets 52% of electricity demands in countries such
USA 36 as Germany and USA.
Colombia 34
The process of iron production in blast furnaces is
Canada 27
Poland 23 the major application of metallurgical coke (about
90% of the coke produced worldwide). The whole
Data from World Coal Institute.
coal consumption in the iron-making industry comes
from two routes. On the one hand, coking coals are
needed to produce metallurgical coke with strict
The recession in coal mining in Europe and the quality specifications to feed the blast furnace. Cur-
need of Southeast Asian countries for coal as energy rently, B700–800 kg of coke are produced from
source has determined a marked increase of the in- 1 ton of coal, so to produce 340 Mt of coke for the
ternational coal trade from some producer countries blast furnace, B450 Mt of coking coal needs to be
to those two neuralgic points. Data for 2002 are carbonized in 250 coking plants around the world.
given in Table 2. Total trade reached 625 Mt. USA’s Excellent coking coals are mined in Australia, Can-
contribution is decreasing because coal companies ada, China, USA, and Poland. On the other hand,
are not interested due to weak coal market condi- coal is also consumed by the steel industry in blast
tions. Australia is the first coal exporter, having furnaces operating with pulverized coal injection
actually a market share of 25% of the European (PCI). As an example, blast furnaces with PCI require
coking coal market and practically 100% of the 700 kg of coal for each ton of hot metal produced,
coking coal market in Japan. 525–600 kg of coking coals to make 350–400 kg of
Coal takes care of 25% of the world’s primary coke, and 100–200 kg of cheaper coal to be injected
energy, with forecasts for the next 15–20 years of 2% via the tuyeres.
per year energy growth and a little higher growth, Among the nonsteel blast furnace applications,
2.2%, for coal production. Productivity depends on foundry iron is the second most important applica-
the grade of mechanization and other factors. Actu- tion of coke produced from coking coals. An esti-
ally, the average mining costs in the USA are in the mation of the global production is between 50 and
order of $5–6 ton
1 in surface and $14–20 ton
1 in 55 Mt per year. Coke for electrometallurgical reduc-
underground mining. Productivity ranges from coun- tion processes represents a minor consumption. In
try to country from less than 1000 ton per person a the nonsteel applications, coke is used as a reducing
year to 220 tons per person and for a shift of 8 h in agent or as an energy source in industrial applica-
the Powder River Basin surface mining. The price of tions such as ferroalloys, lead–zinc smelting, silicon
a ton of coking coal (B$50) is larger than the price carbide production, and stone wool.
of steam coal ($40) in a global market of BUS$125
billions.
Combustion, carbonization, gasification, and liq-
uefaction are considered the four grand processes in
Preparation
the utilization of coal. In general terms, 92% of the In most coal-producing and consuming countries,
coal production is used as fuel and B8% is carbon- national standard methods are available for prepa-
ized to produce metallurgical coke. Coal combus- ration, sampling, and analysis of coal. Through the
tion is carried out in thermal utilities for electricity International Organization for Standardization (ISO)
production, co-generation plants, and cement facto- considerable progress has been made toward the
ries. Coal combines with oxygen from the air giving development of internationally acceptable proce-
carbon dioxide and heat: dures. Most of the ISO procedures are based on
standard methods formulated by National Organ-
C þ O2 -CO2 þ 406:4 kJ mol
1 izations such as the American Society for Testing
Materials (ASTM), the British Standards Institution
Conversion to electric energy reaches efficiencies (BSI), the German Normenausschuss (DIN), and
from 15% to 20% for old installations to 445% for Poland’s Standards Committee (PN). In this article,
184 COAL AND COKE

specific ISO and ASTM standard procedures are flotation cells. The cost is correspondingly higher for a
referred to. higher proportion of fines. Dewatering is done by
In surface mining, the overlying rock (overburden) means of centrifuges and the different products
is removed and the coal extracted. According to from the preparation plant are stored separately in
geological and topographical conditions, different silos. By dosing and blending, ash and moisture
methods (opencast, open pitch, etc.) and equipment contents can be adjusted automatically according to
(shovels, draglines) are used. Conventional room and the market.
pillar mining has been largely replaced by continuous Although cleaning was driven by quality require-
methods – continuous miner and long-wall systems – ments of metallurgical coal, pollution control has
in underground extraction. But, in all cases, the coal increased the need to improve the quality of steam
extracted by any of these techniques will have suf- coal. Actually, the number of coal preparation plants
fered contamination with incombustible rock and is B2200 and they treat B1800 Mt of run-of-mine
water. material with a 73% recovery rate. Although the
Mineral matter is incorporated to the coal seam average plant capacity is 190 tons h
1 biased by the
from different origins along the coalification period small capacity Chinese units (41500 tons), an average
of 1000 tons h
1, are more in line with those estab-
1. as trace elements from decomposition of vegetable lished in Western countries. In the USA, there are
matter; B100 plants with a capacity of more than 900 tons h
1,
2. as detrital minerals from fluvial contribution the largest in the world being Grootegelunk, Repub-
during the diagenesis evolution; lic of South Africa (8200 tons h
1).
3. as mineral inclusions (calcite, pyrite, kaolinite,
etc.) from the neighboring rocks during tectonic
folding in the subsidence period; and
4. through the mechanical contamination cited
Sampling
above. Coal sampling may be defined as the extraction of a
small amount of material from a larger bulk of coal
All the contributions determine percentages in the such that the sample extracted is representative, as
inorganic matter of coal that range from less than far as possible, for all analytes of interest. Sample
2 mass% (some Indonesian ultraclean bituminous preparation, which may involve drying, crushing,
coals) to more than 50 mass% in some run-of-mine and subsampling, is an integral part of the sampling
material. process.
Depending on the intended end-use of coal, The compositional variability of the material is
removal of this extraneous inorganic matter, to a such that coal sampling becomes a complex and dif-
greater or lesser extent, is necessary. Diminishing the ficult operation. The various standardization organ-
ash and sulfur content is important for a better izations around the world have all issued detailed
handling of boilers and to fight air pollution. Limits documents that specify the conditions and methods
in emissions to concentrations of nitrogen and some necessary to obtain representative coal samples for
minor elements, such as mercury, are already estab- analysis.
lished or near to be established. Coking coal is The process of sampling, which may vary from a
required with 10 mass% ash and 1 mass% sulfur 2 ton domestic consignment to a 165 000 ton export
contents as upper limits. To achieve the derived end- shipment, involves obtaining a number of increments
products the run-of-mine mineral is treated in coal (spot samples, the product of a single action of the
preparation plants. sampling device) that are combined to form one or
In a preliminary treatment, the run-of-mine several gross samples. Gross samples may then be
material is crushed to a size of o150 mm. This raw crushed and subsampled (subdivided) to provide the
material, between 2 and 150 mm (named coarse), is analytical sample.
treated through cleaning devices that use the differ- Coal samples may be obtained by either manual or
ence in solid density between coal and mineral mat- mechanical sampling devices, the latter being more
ter. Some heavy media – organic liquids, inorganic appropriate where large tonnages are experienced
salts, and suspended solid particles of magnetite – are (e.g., at coal shipment terminals with throughputs up
used to produce different specific gravities. Smaller to 10 000 tons h
1) and where continuous opera-
mineral fractions (2–100 mm, small, and 0.5–1 mm, tions or operator safety dictate their use.
fines) are separated through hydraulic separation Standard methods for obtaining coal samples
machines (jigs) based on the rate of sedimentation. specify minimum numbers of increments required
Finally, solids o0.5 mm (ultrafine) are treated in to form a gross sample, based on a consignment
 COAL AND COKE 185

(unit) mass of 1000 tons. For consignments of greater 7. Finally, mix and divide to the mass required for
mass, the number of increments may be increased to the laboratory sample.
provide a larger gross sample, or the consignment
may be considered as several subconsignments, each It should be noted that the coking properties of
of 1000 tons. For each of them, a separate gross crushed coal samples deteriorate rapidly with time.
sample is collected, to be individually processed and Therefore, samples intended for such testing should
analyzed. be prepared immediately before analysis.
As well as the mass of the consignment, the re- Coke sampling is marginally less problematic
quired minimum number of increments is influenced because the product from a single source derives from
by coal or blend of coals that have been prepared to a
specification for ash, moisture, particle size distribu-
* the quality of coal – washed or cleaned coals are tion, etc. The final coke produced will be relatively
more homogeneous and therefore require fewer homogeneous in all properties, with the exception of
increments than blended or untreated coals and size distribution. Standard methods are available for
* the sampling process to be used – stopped-belt or coke sampling that reflects the somewhat less rigor-
falling-stream sampling requires fewer increments ous requirements for this material.
to attain the required precision than sampling
from barges or stockpiles.
Petrography and Chemical
Sample preparation involves a series of operations
Composition
such as reduction of size, homogenization, and re-
duction of the mass of the gross sample to that suit- Coal is not a homogeneous rock. In coal seams, pet-
able for analysis. For the determination of size rographers have distinguished with the naked eye
distribution, no sample preparation other than air more or less lustrous thin beds and regular accumu-
drying and reducing the mass of the gross sample is lations that can be divided into four fairly well-
undertaken. For other analysis, the extent of sample defined classes called lithotypes – vitrain (brilliant
preparation is dictated by the intended analysis. For layer), clarain (semibrilliant layer), durain (mat
example, when testing for Hardgrove Grindability layer), and fusain (fibrous layer). When coal is exam-
Index (ISO 5074, ASTM D409 or equivalent), a ined through an optical microscope with reflected
subsample of B1 kg coal prepared to a top-size of light under oil immersion, it is shown that coal is
4.75 mm is required, whereas for general analysis a composed of discrete entities, called macerals, which
final sample of 50–100 g crushed to less than 0.2 mm result from the transformations of the original
is sufficient. vegetal debris. According their optical characteris-
Sample preparation procedures typically involve a tics macerals can be classified into three groups:
fixed sequence of operations:
 The vitrinite group originates from lignocellulo-
1. Extract a subsample for total moisture determi- sic tissues that are gelified by bacterial action. It is
nation. mainly composed of aromatic and hydroaromatic
2. Air dry the gross sample (to ensure that the coal structures and is usually the most abundant maceral
will flow smoothly through subsequent equip- group. It often occurs as a matrix surrounding the
ment). It should be noted that forced drying can other macerals and mineral constituents and has
have adverse effects, especially on coking proper- the property of swelling and agglomerating during
ties; thus heating samples to more than 151C the carbonization of medium-rank coals. Its density
above ambient temperatures should be avoided. increases with rank from 1.2 to 1.7 g cm
3.
3. Reduce the particle size of the sample. The whole  The liptinite group is derived from organisms
gross sample is crushed to some intermediate size and organs that are relatively poor in oxygen such as
(10, 5, and 3 mm) using a mechanical mill or algae, spores, pollens, cuticles, secretions. Chemical-
crusher. ly, it is characterized by a higher content in hydrogen
4. Mix the whole sample thoroughly to ensure and aliphatic structures. It is the lightest mac-
homogeneity. eral with a density between 1.1 and 1.25 g cm
3. By
5. Reduce the mass of the gross sample (sample thermal heating in an inert atmosphere, it is con-
division) to a mass consistent with the present size verted to volatile products, leaving little solid residue.
of the coal, by using a mechanical sample divider.  The inertinite group often originates from
6. Further crush the sample using a high-speed vegetable matter that is partly burnt or has undergone
impact mill to attain the required particle size. lengthy aerobic oxidation before burial. Chemically,
186 COAL AND COKE

it is mostly aromatic with a comparatively higher

Calorific value
(MJ kg
1)
amount of carbon and lower amounts of hydrogen
and volatile matter. It is the densest maceral (1.4–
2.0 g cm
3). Most inertinite remains inert during the

16.61
18.56
19.05

28.47
31.40
33.00
34.42
35.24
34.84
33.96
carbonization process, although the least reflective
ones still retain plastic properties.

Volatile matter
(mass% db)
The maceral content defines the coal type: sap-
ropelic, with 450% liptinite, or humic, more abun-

70.0
59.0
48.0

32.0
35.5
31.6
24.5
16.1
10.9
5.1
dant, usually presenting a banded structure. On the
other hand, based on the different optical properties
of macerals, the reflectance of vitrinite is an essential

Ash (mass%
characteristic used in coal identification and related
to rank. A good analysis of maceral content provides

db)
knowledge about the chemical composition of a coal,

1.0
3.0
6.0

3.1
5.0
5.9
4.8
2.9
3.0
2.3
their behavior in different conversion processes, and
can also be used as a parameter of coal rank (see
section on Petrographic analysis).

(mass%)
Thus, the chemical composition of the dominant

Moisture
organic part of a coal, mainly the amounts of carbon,

15.0
13.0
16.0

10.7
4.7
1.5
0.8
0.6
1.3
2.9
hydrogen, and oxygen, which together comprise
between 97% and 99% of the mass of pure coal,
depends on the coal type, and has a clear evolution
with coal rank, the carbon content increasing progre- S (mass%
ssively from o65% for lignite to B95% for anthra-
dmmf)

cite. Hydrogen is a minor element, following an 0.5

1.0
2.0

1.4
0.8
1.0
0.7
0.7
0.8
0.8
opposite trend. Oxygen decreases correspondingly
from above 30% in lignites to o1% in the highest
rank coals. No distinct trends are apparent for nit-
rogen (usually between 0.3 and 1.5%) and organic
N (mass%
dmmf)

sulfur (0.3–2%). A progressive decrease in water and

volatile matter contents (from 442 to o4%) goes
0.5
1.5
1.3

1.3
1.8
1.2
1.4
1.4
1.5
1.5

hvb, mvb, and lvb refer to high-, medium-, and low-volatile bituminous coals.
parallel to an increasing rank. Table 3 presents typ-
ical values for elemental analysis of coal precursors

Data from Encyclopedia of Analytical Science, 1st edn, Academic Press.

and coals of increasing rank, together with proxi-
O (mass%

mate analysis and calorific values.

dmmf)

The inorganic constituents of coal are the residues

43.0
34.5
25.2

13.0
7.3
4.8
3.3
2.5
1.3
0.8

of plant components, minerals incorporated in coal

seams, and rocks co-deposited or mixed with the coal
during mining. As such, their chemical composition
is enormously varied. The principal components
H (mass%
dmmf)

found are quartz and minerals of clay, mica, and

Table 3 Composition of different rank coals

chlorite groups. Sulfide, carbonate, and chloride

6.0
6.0
5.5

5.5
5.4
5.5
4.9
4.4
4.0
2.9

minerals, together with feldspars, barites, and titani-

um minerals, make up the remainder of the suite,
along with lesser concentrations of minerals formed
C (mass%

by weathering, such as limonite and sulfates of iron

dmmf)

and calcium. Of particular interest is the content of

50.0
57.0
66.0

78.8
84.7
87.5
89.7
91.0
92.4
94.0

sulfur minerals. Sulfur is an undesirable constituent

of coal, and as such will attract a price penalty if
present at greater than specified levels.
Semianthracite

Virtually all other elements occur sporadically in

Bituminous

Anthracite

coal. These ‘trace’ or ‘minor’ elements rarely consti-

Lignite
Wood

tute more than 1% of the mineral matter and fre-

mvba
Peat
Coal

hvba
hvba
hvba

lvba

quently the proportion is much lower.

a
 COAL AND COKE 187

Analysis and Tests results are given in a dry, ash-free basis (daf). If more
complete analytical data are available they can be
The suite of analyses required for a particular coal given in a dry, mineral matter-free basis (dmmf), re-
mainly depends on its intended end-use. Steam coals sults being a measure of only the organic component
are defined through parameters related to their cal- of coal.
orific value and impurities content having a strong Other bases may be required for the expression of
impact on environmental pollution; metallurgical, or analytical data. Moist, ash-free basis (maf) assumes
coking coals are better classified through specific tests that the sample is free of ash, but with moisture
related to their properties to transform to a good- (ASTM D388, Standard Classification of Coal by
quality coke for a particular application. Testing and Rank, requires calorific value to be expressed on a
assessing of a coal starts with a consideration of its moist, mineral matter-free basis (mmmf)). The var-
rank, washability, hardness, and its inherent moisture ious national standards organizations present differ-
content and is followed by analyses of its chemical, ent formulae for the calculation of mineral matter
physical, rheological, and microscopic properties, (which is not a generally determined value); reference
and, finally, pilot tests. As with coal sampling, na- to their publications is necessary to determine which
tional and international standard methods for the calculation is appropriate in given circumstances.
analysis of coal and coke have been developed in or- An exhaustive discussion of individual methods is
der to define the quality of the material through outside the scope of this article. Thus, the more com-
chemical, petrographic, and empirical tests. These monly reported tests are commented upon below.
enable the producer to monitor variations in the
quality of the product and the purchaser to assess the
Moisture
suitability of marketed coal for a process.
Instrumental methods are now becoming accepted The water in coal is bound in different forms to its
as alternative techniques, provided they are shown to constituents. It can be divided into three types: (1)
give equivalent results to the conventional test. These Free moisture, also referred to as external moisture,
methods can allow a greater throughput of samples, superficial moisture, or the primary moisture frac-
with less operator dependence than conventional tion, which is present in large cracks and capillaries.
techniques, and often generate useful data that were Water bound in this way retains its normal physical
not previously available. In fact, there are some properties. (2) Inherent moisture, also referred to as
instrumental companies, i.e., LECO Corporation, internal moisture or the secondary moisture fraction,
devoted almost exclusively to the development and whose vapor pressure is lower, since it is absorbed
sale of apparatus for chemical analysis – immediate within the pore structure of the coal. (3) Water of
analysis, elemental analysis, sulfur determination – constitution, which is mainly combined with mineral
calorific value, ash fusibility, and other technical tests matter normally present in coal. This water is gene-
of coals. rally driven off only at temperatures higher than
Knowledge of the total moisture of a coal is essen- those normally used for the determination of mois-
tial in commercial activities. Proximate analysis – de- ture content. Standard methods do not make use of
termination of residual moisture, ash, and volatile these terms and define: (1) the total moisture content
matter in a dried sample of coal – is made for all of a coal; and (2) the moisture content of the coal
samples received for coal utilization. Elemental analy- analysis sample. Total moisture determination must
sis (C, H, N, O) and analysis of sulfur are necessary in be made over the sample as received in the labora-
order to have a better knowledge of the quality of a tory, in an air–proof recipient. The determination
coal. Petrographic analysis provides information on consists in drying in an oven at 1051C till constant
the rank, quality, and blending of coking coals. Fria- weight. Its value is of huge interest both in interna-
bility tests are necessary for the behavior of steam tional and domestic coal trade (ISO 589, ASTM
coals, and plasticity and swelling tests for the D3173).
behavior of coking samples. The analysis of mineral This determination must be differentiated from
matter gives information on the possibility of pollu- that of the moisture of the sample for analysis that
tion problems and tests of fusibility on the ash corresponds to a sample of coal equilibrated at the
behavior of combustion coals. laboratory conditions (usually, 24 h). The method-
Most analyses of coal are carried out on air-dried ology is the same and the value is that to be used in
samples and results are normally reported on the dry the calculations at the different bases.
basis (db), as percentage of the coal after the mois- The equilibration of a coal sample in labora-
ture has been removed. When samples are compared tory conditions is absolutely necessary in order to
on the basis of certain properties of the pure coal, facilitate handling of the sample in the following
188 COAL AND COKE

operations, such as sizing – humid coal samples do strictly maintained in order to obtain comparable
not answer to riffling – or the Hardgrove grindability results. The ash content differs from the mineral
test (see below). matter content, both in their composition and in
their mass (always lower). The mineral matter
Sizing content is often calculated following the empirical
formula of Parr:
Size analysis is useful in assessing problems in coal
transportation and handling. For larger sizes coals, Mineral matter ð%Þ ¼ 1:08Ash þ 0:55Spyritic
hand or mechanical sieving or screening is under-
taken; fines (o1 mm) are better screened in humid The difference between both contents – ash and
medium; while for ultrafines (o0.06 mm) optical or mineral matter – can lead to considerable errors,
electrical field effect techniques are more appropriate. larger when the contents are larger. For less than
Size reduction is carried out on the samples for 30 mass%, the error is B1.3%, which rises to 4%
analysis, depending on the specific analysis or test when the ash content is around 25% and to 28% for
(i.e., o0.212 mm for chemical analysis, o4.75 mm inorganic contents of 70%.
for Hardgrove; o1 mm for petrographic analysis). The whole analysis can be done in an automatic
apparatus designed to reproduce the analytical con-
Hardgrove Grindability Index ditions of the established procedures. The apparatus
This index gives a value for the grindability of coal follows a program and can work simultaneously
and coke. The numerical value is related to the with 20 samples. The conditions for determination of
number of revolutions needed to reduce a given moisture are well accomplished, but some details on
amount to a determined particle size, relative to a the determination of volatile matter (time) and ashes
standard (see ISO 5074 and ASTM D409). Grind- (temperature, 8151C) are not exactly the same and
ability is related with rank, Hardgrove index rea- the tolerance (2%) is higher than that of the standard
ching a maximum for coals with 80–90% of carbon. method (1%). Anyhow they are largely used in elec-
High values of Hardgrove index (50–80) are indi- trical utilities that need to make hundreds of analyses
cative of easier to grind materials while values in the from different suppliers each day.
order of 20–30 indicate a higher hardness of a coal.
Calorific Value
The gross calorific value is the amount of heat
Chemical Analyses released by burning with oxygen a coal sample in a
Proximate Analysis calorimeter under controlled conditions. A correc-
tion for the heat absorbed by the remaining ashes
The analyses of the air-dried equilibrated sample for must be calculated. If the correction includes the
moisture, ash, and volatile matter are collectively latent heat of vaporization, the net calorific value,
termed the proximate analysis. Fixed carbon is, by important in the coal market, is determined.
definition, the difference between 100 and the sum of The measurement can be done both in isothermal
the analytes (moisture, ash, volatile matter). The or adiabatic calorimeters, the latter being preferred.
proximate analysis gives information on the classifi- For isothermal measurement (see ASTM D3286),
cation of coal by measuring the relative percentage of the temperature of the calorimeter jacket is held
volatile and nonvolatile organic matter as those of constant and a correction for heat transfer from the
moisture and noncombustible mineral matter. calorimeter is applied, while in the adiabatic meas-
Determination of moisture was discussed above urement (see ISO 1928 and ASTM D2015), the tem-
(ISO 589, ASTM D3173). Volatile matter – param- perature of the calorimeter jacket is continuously
eter used in coal classification – can be defined as the adjusted to approximate that of the calorimeter itself.
percentage of gaseous components of the coal, except The calorific value can be correlated with the fixed
moisture, loss at a high temperature in an inert at- carbon content of the coal. It is a good parameter for
mosphere. Its determination is carried out under coal classification of specific coal types and as a price
prescribed conditions (i.e., 7 min at 9001C) starting determination index ($ per MBtu) for commercial
from an amount of normalized material around a steam coals.
mass of 1 g (ISO 562, ASTM D3175).
Ash is the solid residue from inorganic material
Ultimate or Elemental Analysis
after the complete combustion of coal. Its composi-
tion and amount depends on the analytical condi- Analysis for the elementary constituents of coal fol-
tions (ISO 1171, ASTM D3174) that must be lows techniques similar to those employed in organic
 COAL AND COKE 189

chemistry. It comprises analyses for carbon, hydro- H2O, and N2, monitored by a thermal conductivity
gen, nitrogen, sulfur, and oxygen, and altogether detector.
give the composition of the organic matter of a coal. The oxygen content of a coal is an important rank
Corrections must be made because, apart from nit- parameter, the younger coals being richer in this
rogen, carbon, sulfur, and oxygen can be found in element than the more mature coals. Oxygen is tra-
impurities (as carbonates, sulfides, sulfates, clays), ditionally calculated by difference mainly for high–
and hydrogen and oxygen are present in the sample rank coals from a knowledge of the amount of other
moisture. chemical components (C, H, N, and S), moisture,
Carbon and hydrogen are determined through and ash contents. Techniques involving direct deter-
complete combustion at 12001C. The gases (CO2 mination of oxygen avoid the effect of cumulative
and H2O) are absorbed, respectively, on anhydrone error in the analysis by difference.
and NaOH and measured gravimetrically (see ISO The direct determination of oxygen has been and
609, ASTM D3178). continues to be a problem as it does not have an easy
Nitrogen is determined by the Kjeldahl method. solution. Laboratory microanalyzers for C, H, N, S,
Nitrogen is transformed in ammonium sulfate by and O have been developed. They work over smaller
treatment with concentrated sulfuric acid at 9001C. coal samples (B1 mg) and the oxygen determination,
Afterwards it is liberated as ammonia in a steam based on chemical methods available, is carried
stream that passes through a diluted acid solution out in a supplementary device, where the sample is
and a titration with alkali (see ISO 332, ASTM pyrolyzed at 13501C in a helium stream and redu-
D3179). cing atmosphere. The resulting gases are passed
Sulfur occurs in both organic and inorganic com- over activated carbon, which converts oxygenated
binations. For certain purposes a knowledge of total products into carbon monoxide. The CO is conver-
sulfur content is adequate; however, for coal prepa- ted catalytically to CO2 and then quantified by
ration and conversion processes, determination of means of an infrared detector. The method is increa-
the forms of sulfur (organic, pyritic, and sulfate) is singly used mainly in analysis of low-rank coals, in
valuable. coal weathering research, and analysis of carbon
The total sulfur content may be determined by one materials.
of several methods that convert it to sulfate by wet
chemical analysis. One of these, the Eschka method,
Other Chemical Analyses
involves combustion of coal at 8001C in the presence
of alkaline/oxidant medium (e.g., two parts of cal- The chemical analyses of ash for major elements and
cined MgO and one part anhydrous sodium carbon- a test in ash fusibility temperature give guidelines on
ate); all sulfur is converted to sulfate that by the coal utilization, defining the suitability of coal for
addition of barium chloride precipitates as barium different combustion or conversion systems. Analysis
sulfate, which is calcined to BaO and measured for minor elements is gaining impetus as greater
gravimetrically (see ASTM D3177). This is a stand- interest is taken in the environmental consequences
ard method in many countries. Another is the ‘high– of coal uses. Typical ranges of concentration for
temperature method’ where the coal is burned in major, minor, and trace elements are summarized in
oxygen at 13501C, converting all sulfur present into Table 4.
SO2. The SO2 is then converted to sulfuric acid for Multielement analytical techniques – atomic
titrimetric determination. absorption spectrometry, inductively coupled plasma
Pyritic sulfur may be determined by the estimation mass spectrometry, X-ray fluorescence, neutron
of pyritic iron, which involves a pretreatment with activation analysis, etc. – are used. The experimen-
hydrochloric acid to eliminate the nonpyritic iron tation can be done directly on the mineral matter of
and then dissolving the pyritic iron in nitric acid. the coal sample after the removal of the organic
Sulfate sulfur is estimated by solution in hydro- matter by a prolonged treatment of activation with
chloric acid followed by gravimetric estimation of oxygen plasma (low–temperature ashing). Neutron
the dissolved sulfates. activation is also applied to online analyses of coal
Procedures to determine the sulfur forms – sul- and fly–ashes on feeding-belts in order to provide
fates, pyritic, and organic sulfur – are well described information on a continuous basis.
in ISO 157 and ASTM D2492. Chlorine can also occur in coal as organic and
Automatic systems have been developed and are inorganic compounds, but only the total chlorine is
largely used in the simultaneous analysis of carbon, normally determined. Formulae for calculating min-
hydrogen, and nitrogen. Samples (up to 20) of eral matter contents arbitrarily assume 50% of the
50–100 mg are burned and the evolved gases, CO2, chlorine to be inorganic.
190 COAL AND COKE

Table 4 Concentration ranges of major, minor, and trace ele- Vitrinite Reflectance Analysis
ments in coal

Major and minor constituents Trace elements The reflectance of vitrinite is a rank parameter that
has well-established relationships with other chemi-
Ash analysis Ash Element Concentration
cal rank parameters such as volatile matter, C, and H
(mass%) (mg per g coal)
contents. Among the advantages of the use of vitri-
SiO2 40–90 Be 0.1–15 nite reflectance as a rank parameter are that it fol-
Al2O3 20–60 Cr 0.5–60
lows a regular increase over the whole coalification
Fe2O3 5–25 Mn 5–300
CaO 1–15 Co 0.5–30 scale and is not affected by maceral composition
MgO 0.5–4 Ni 0.5–50 since it is recorded on individual particles. Reflect-
Na2O 0.5–3 As 0.5–80 ance is defined as the proportion of perpendicularly
K2O 0.5–10 Se 0.2–10 incident light reflected from a component compared
P2O5 o1 Cd 0.1–3
to that reflected from a standard of known reflect-
TiO2 o2 Sb 0.05–10
Hg 0.02–1 ance. Readings are taken with monochromatic light
Pb 2–80 (546 nm) and oil immersion objectives (1.518 refrac-
tion index at 231C) following the standard ISO
7404-5 procedure (see also ASTM D2798). Strict
rank determination requires recording the readings
Petrographic Analysis on a single maceral from the vitrinite group – col-
lotelinite – although for most industrial applications
The most widely used petrographic analyses of coal readings recorded on any maceral of vitrinite group
are maceral analysis and vitrinite reflectance analysis. might be sufficient.
Both are performed on representative samples Two types of readings can be taken on vitrinite
ground to o1 mm in size and embedded in resin. particles: (1) random reflectance (Rr) using nonpo-
The polished surfaces are then examined under a larized light; and (2) maximum reflectance (Rmax)
white reflected light microscope. using polarized light and turning the stage till the
maximum reflectance is achieved. For low-rank
Maceral Analysis coals, Rr and Rmax are equivalent. As coal rank in-
The relative proportion of petrographic components creases vitrinite develops anisotropy yielding higher
in a coal can be performed at a maceral level or Rmax values than Rr . In both cases 100 readings se-
maceral group level depending on the degree of detail lected using a regular frame are averaged to calculate
desired. For most industrial applications the maceral mean reflectance values and to construct the re-
group analysis is usually enough. For coals to be used flectogram with distribution of vitrinite classes.
in coking processes it might be desirable distingui- Standard deviation must also be provided. Reflect-
shing between low- and high-reflecting inertinites ance values typically approach to a Gaussian distri-
since the former still retains plastic properties (see bution for low-rank coals and standard deviation
below). For a better discrimination of liptinite group increases with the rank of the coal. Rmax is recom-
macerals in low-rank coal the use of fluorescence mended for detailed studies on burial history of coal
light might be convenient. As with other coal analy- basins, whereas for most industrial applications ran-
ses, several international and national standard pro- dom reflectance is preferred since its analysis is less
cedures are available with minor differences among time consuming.
them and they describe with some degree of detail Petrographic analysis is the only procedure to de-
the procedure of pellet preparation, the total number termine the various components of a coal blend,
of points counted, the magnification employed, the which is of foremost importance in coal utilization.
repeatability and reproducibility values, etc. (i.e., In this case reflectance readings are taken on any
ISO 7404-3 and ASTM D2797). Oil immersion vitrinite within the particle regardless the maceral on
objectives are normally used to provide best identi- which the crosswire is landed. With this procedure
fication conditions. For standard maceral analysis the histogram plot reflects directly the proportion
typically 500 points are recorded by point counting of the coals in the blend. The analysis yields the rank
displacing the sample with a mechanical stage at of the single-component coals and their amount. Coal
regular intervals (0.5 mm). The results are expressed discrimination is easy for blends of coals of rather
in volume percent based on the principle that the different rank. Calculations may require histogram
surface occupied by a given component in a ran- deconvolution for coals close in rank. A higher
domly selected section is proportional to the volume amount of readings is recommended for coal blend
of that component in the sample. analysis.
 COAL AND COKE 191

60 cokemaking, but criteria for selecting coals in coke

manufacture are based on certain properties of coals
50
that, when heated in an inert atmosphere, cause them
Frequency (%)

40 to soften, become plastic, and coalesce into a coher-

Rmax = 1.11%

t = 0.07 ent mass that swells and resolidifies to form a solid
30
Volatile matter = 22.5 wt% db porous carbon material (coke). These particular
20 types of coal are generally known as ‘caking coals’.
10 The physical changes occurring in coal when heated
in the temperature range of 350–5501C in the
0 absence of oxygen are known as ‘plastic or
0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
thermoplastic properties’. The importance of such
(A) Reflectance (%)
properties is related to the capacity of agglomeration
30 between the reactive maceral constituents of the coal
and the inert particles (some part of the inertinite
25
maceral group and mineral constituents). On the
Frequency (%)

Rmax = 1.11%
20

t = 0.24
other hand, the remaining coals are referred to as
Volatile matter = 22.7 wt % db ‘noncaking’ coals and they produce a weakly coher-
15
ent or noncoherent char. It is important to point out
10 the difference between ‘caking coal’ and ‘coking
coal’. The term ‘caking’ is reserved for coals that
5
possess plastic (or agglutinating) properties in labo-
0 ratory tests, while the use of ‘coking’ is related to
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
caking coals in which plastic properties are so
(B) Reflectance (%)
strongly expressed as to make them suitable for
Figure 2 Histograms of the distribution of vitrinite in reflectance conversion in metallurgical and other industrial
classes for a bituminous coal (A) and a complex coal blend (B). cokes. Coking coals are therefore strongly caking
coals with a volatile matter content ranging between
As an example, Figure 2 displays the histograms of 19 and 32 wt%. Caking behavior can be assessed
the distribution of vitrinite in different reflectance during rapid heating (1001C min
1), whereas coking
classes for a bituminous coal and a complex coal behavior is assessed during slow controlled heating at
blend with the same volatile matter content and a rate of 3–51C min
1.
mean vitrinite reflectance. An appraisal of caking and related rheological
Both maceral and reflectance analysis may be per- properties is consequently not simple and usually re-
formed by automated image analysis systems that quires measurement of several parameters. Distinct
process images acquired with a camera. They usually laboratory tests to assess these properties, swelling
yield an average reflectogram of all components in and caking, have been developed to simulate the coal
the sample where thresholds must be established in a behavior and the experimental conditions in coke
further step either manually or using more or less ovens. Some of them have been standardized and are
complex computer routines. The maceral group per- widely used to evaluate the suitability of a coal or
centages are calculated as the area under the curve coal blend for coke production. It has been reported
within two thresholds and the vitrinite reflectance that swelling and caking are therefore mutually rein-
value is based on the position of the peak assigned to forcing phenomena and that swelling is simultane-
vitrinite. Automated systems yield results similar to ously the cause and the consequence of caking. It
manual analysis for medium-rank vitrinite-rich coals, should be pointed out that coal blending has been
but problems remain with resin–mineral matter–lip- adopted by the industry because of the limited avail-
tinite thresholding for low-rank coals and inertinite– ability and high cost of good coking coals and the
vitrinite thresholding for high-rank coals. continued demand for better quality coke for the
blast furnace. Coal blends must be low in cost, pro-
duce a high-quality coke, and provide a safe oven
pushing performance. They are composed of coals
Thermal and Rheological Properties (four or more) differing in rank, rheological proper-
It is well established that only coals within a specific ties, geographical origin, and their proportion in the
range of rank and type are suitable for coke produc- blend. As aids to coal selection and coke quality
tion. Chemical analyses of the single coals, includ- prediction, several mathematical models have been
ing ash and sulfur, are important parameters in developed and extensively reported. Most of them
192 COAL AND COKE

are based on parameters reflecting the rank, maceral undesirable for cokemaking, because the coke is too
behavior during carbonization (the concept of reac- weak with thin walls of its pores.
tive and inert), and plastic properties of coal.
Some of the most common methods to evaluate the Gray–King Assay
suitability of a coal or coal blend for coke production
are the free-swelling index test, Gray–King assay, It is a predominantly visual assessment of agglome-
Roga assay, various dilatometer tests, and Gieseler ration and swelling properties when an unconfined
plastometer test. In addition, research using larger sample of coal (with or without inert diluent) is
amount of coal sample is also performed by means of heated in a horizontal oven. The coal sample is
a pilot coke oven. slowly heated in a tube at 51C min
1 up to 6001C.
The appearance of the resulting coke is examined
visually and classified by comparison with a series of
Free-Swelling Index reference cokes designated as A, B, C,y, G (ISO
The free-swelling index (FSI), also called the crucible- 502). The coal may remain in powder form, either
swelling index, is the simplest test to evaluate whether granular or weakly coherent (coke types A, B, C), or
a coal is of potential value for coke manufacture. It it may be fused, but reduced in volume (coke types
can be taken as a preliminary separation of coking D, E, F) – Figure 4. Coal given a well-fused coke is
and noncoking coals and it provides some measure of classified as type G. If coal has a very high swelling
relative coking characteristics. It does not reflect, how- power, it is tested with variable proportions of
ever, essential coking characteristics such as plasticity. electrode carbon. The proportion of added carbon
This test involves rapid heating of a small sample of
crushed coal in a standardized crucible to a temper-
ature of B8001C. After heating, a small coke ‘button’ Type Profile
remains in the crucible. The cross-sectional profile of
this coke ‘button’ is compared to a series of standard
profiles numbered 1–9 with 1/2 increments – Figure 3 A
(ISO 501, ASTM D720). The lower the profile of
the coke ‘button’ formed, the less the free-swelling
power and/or caking power of coal. An index of 0 B
is assigned to a noncoherent and pulverulent coke
‘button’ and indices of 1–3 are usually taken as the
coal is only weakly caking. A high FSI (8–9) is also
C

D
1 11/2 2 21/2 3
E

31/2 4 41/2 5
G

G1

G2
51/2 6 61/2 7
G3

G9

71/2 8 81/2 9
Figure 3 Reference profiles of the free-swelling index test. Figure 4 Gray–King coke types and profiles.
 COAL AND COKE 193

Table 5 Comparison between free-swelling index, Gray–King Audibert-Arnu dilatometer, and the experimental
coke type, and Roga index conditions used are given in the International Standard
Free-swelling index Gray–King coke type ISO 349.
The Ruhr dilatometer test (ISO 8264), used in
0–1/2 A–B
1–4 C–G2 Germany and many other countries, is a modification
4 1/2–6 F–G4 of the Audibert–Arnu dilatometer test. By this test,
6 1/2–8 G3–G9 the coking capacity G or G-factor, used for coke
8 1/2–9 G7 and above quality prediction, is calculated.

Free-swelling index Roga index Gieseler Plastometer Test

0–1/2 0–5 In coke production, crushed coal passes through a
1–2 5–20
fluid or plastic state to become a fused porous solid
2 1/2–4 20–45
44 445 carbon material (semicoke). The temperature over
which coal exists in this fluid state and the extent of
Data taken from Ward CR (1984) Coal Geology and Coal Tech-
fluidity development are critical factors in coke man-
nology. Oxford: Blackwell Scientific.
ufacture, because they are responsible for good ad-
hesion between reactive components of coal and
inert particles. These characteristics of a coal or a
characterizes the coal and the coke type is designated
coal blend (softening, reduction in viscosity, resolid-
as G1–G8. Although it is not determined under the
ification) are tested by using a constant-torque
same conditions, the Gray–King coke type corre-
Gieseler plastometer (a specific type of viscosimeter)
sponds broadly to the FSI (Table 5).
as described in the ASTM D2639 standard proce-
dure. In this test, the coal sample (5 g with a particle
Roga Index
size less than 0.425 mm) is heated from 3001C to
This test, originally developed in Poland, indicates 5501C at a rate of 31C min
1. The parameters
the caking power of coal under standardized condi- derived from this test are: (1) softening temperature,
tions (ISO 335). It is based on pressing a coal sample the temperature at which the coal starts to be fluid;
mixed with a standard anthracite under a small load (2) the temperature of maximum fluidity, the tem-
at 8501C for 15 min in a standard crucible. The co- perature at which the fluidity of the coal reaches a
hesion of the resultant coke is then tested in a drum maximum; (3) resolidification temperature, the tem-
tumbler, and the Roga index determined as the per- perature at which the fluid mass resolidifies; (4) plas-
centage of coarse material remaining. A general cor- tic range, which is defined as the difference between
respondence exists between the FSI and the Roga the resolidification and softening temperatures; and
index. A Roga index of 45 indicates a coking coal (5) maximum fluidity, obtained as dial divisions per
with an FSI of B312 to 4; the stronger coking coals minute (ddpm). Maximum fluidity may also be re-
have Roga indices above 45 (Table 5). ported as the logarithm (base 10) of that value. The
maximum fluidity is one of the most important pa-
Dilatometer Tests rameters used to characterize coals for cokemaking.
Parameters maximum fluidity and plastic range are
Dilatometers allow the study of two types of phe-
of great relevance in the compatibility of coals for
nomena: (1) the behavior of coal during fusion and
blend preparation.
softening; and (2) the contraction of semicoke or
green coke beyond the temperature of resolidifica-
tion. The dilatometry-based tests measure the vo-
Coking Pressure Measurements
lume changes when a compressed pencil of coal,
confined by a piston, is heated in a vertical retort During carbonization, the volatile matters that
(ISO 349, ISO 8264, and ASTM D5515). This type evolve as a result of the thermal coal decomposition
of test differs from those described above in that try to escape through the different layers formed in
swelling is not free, either because the coal has first coke ovens. There is certain resistance to the passage
been agglomerated by compression and because the of the volatile products and, consequently, a pressure
piston is under load. It uses slow heating (usually, is developed against the chamber walls, which is
31C min
1) in an attempt to simulate the behavior known as coking pressure. Certain coals generate
of coal in a coke oven. The most commonly used high gas pressures in coke ovens that cause opera-
dilatometer is that developed by Audibert-Arnu. A tional difficulties during coke pushing and coke-oven
detailed description of the sample preparation, wall damage, thereby shortening coke oven life. This
194 COAL AND COKE

type of coking coals, namely dangerous coals, are not Movable Wall Oven
carbonized individually, but are used as components
Direct measurement of coking pressure can be ob-
in industrial coking blends in order to adjust the
tained in a movable wall oven, which is widely used
volatile matter and fluidity of the blend and to im-
to estimate coking pressure of a coal and a coking
prove coke mechanical strength. It is therefore nec-
blend. A characteristic feature of these ovens is that
essary to be able to predict and assess the danger of a
one wall is mounted on runners such that it can
coal, and keep the coking pressure below certain move, or tend to move, away from the other oven
limits in order to prolong the life of the coke oven.
wall. Its movement is nowadays restrained by a load
Two basic methods of approach can be undertaken
cell that measures the force necessary to prevent the
to determine the suitability of a coal or coking blend:
wall movement. By monitoring this force during a
(1) indirect coking pressure measurements at labo-
coking cycle, the pressure exerted by the charge on
ratory scale and (2) direct coking pressure measure-
the wall can be calculated. There is no standard oven
ments by larger-scale tests, using a few hundreds of
and procedure, although all movable wall ovens
kilograms of coal while trying to reproduce indus-
work on the same basic principles. These ovens with
trial conditions. 250–400 kg capacity have similar width to commer-
Generally, the first type of tests involves heating of
cial ovens, so that heating regimes can be accurately
a coal sample while it is compressed by a piston acted
reproduced on the pilot scale. In addition to coking
by a standard load. As carbonization proceeds, the
pressure measurements, the amount of coke pro-
pressure generated in the coal charge causes move-
duced allows to run full-scale coke tests in terms of
ment of the piston, which is monitored. Depending
physical and mechanical properties and reactivity to
on the expansion and/or contraction values observed,
carbon dioxide.
coals can be classified as very dangerous, dangerous,
or nondangerous (safe). Among the laboratory-scale
tests developed to monitor dangerous coals, the Kop-
pers test and its variants and the sole–heated oven
Coke
test are briefly described. The principal advantages of Metallurgical coke is a macroporous carbon material
such tests are the shorter time, lower cost, and a of high strength and relatively large lump size pro-
smaller amount of coal required for testing. duced by the carbonization of coals with a specific
rank or of coal blends at temperatures up to 1400 K.
In conventional coke production, coal blends crushed
Koppers Test
to B80% less than 3 mm particle size are usually
The Koppers test uses an oven with unidirectional carbonized in batteries of coke ovens. The ovens are
heating from one side. By this test, carbonization is indirectly heated through the side walls at a temper-
performed with a sample of 80 g of coal that is sub- ature of B13001C over a period of 18–20 h. About
jected to a constant pressure of 10 kPa by means of a 90% of the coke produced from coal in the world is
piston and the change in volume is measured. Mod- used to maintain the process of iron production in
ifications to the early Koppers test have been intro- the blast furnace where it has three major roles: (1) as
duced by INCAR (Spanish Patent No. 524.258, a fuel, it provides heat for the endothermic require-
1983) based on Mott and Spooner’s modifications. ments of chemical reactions and the melting of slag
They include the heating system and experimental and metal; (2) as a chemical reducing agent, it pro-
conditions such as pressure on the charge, bulk den- duces gases for the reduction of iron oxides; and
sity, and the rate and duration of heating. As a result (3) as a permeable support, it acts as the only solid
of the research conducted on several coals and coal material in the furnace that supports the iron–
blends, a criterion was established to classify coals. bearing burden and provides a permeable matrix
Coals giving a contraction greater than 10 mm and necessary for slag and metal to pass down into the
no expansion can be considered to be not dangerous hearth and for hot gases to pass upwards into the
during carbonization. stack.
Of these three roles, the first two can be substi-
tuted by oil, gas, plastics, and coal, which are inject-
Sole-Heated Oven Test
ed at the tuyeres as generating energy and a carbon
An ASTM procedure specifies the conditions under source. Such a substitution brings about a reduction
which the expansion or contraction of coal or coal in coke rates for the blast furnace. (Coke rate is the
blends during carbonization can be measured (ASTM weight of coke required to produce 1 ton of
D2014). This type of coking oven has unidirectional iron). However, there is no other satisfactory mate-
heating from one side (the sole). rial available, which can replace, fully or partially,
 COAL AND COKE 195

metallurgical coke as a permeable support of blast- Table 6 Required chemical properties of blast furnace coke
furnace charge. Chemical property European range
An assessment of the coke performance in the blast
Moisture (mass%) 1–6
furnace operating with or without injection tech-
Volatile matter (mass% db) o1.0
nology should include those properties of coke that Ash (mass% db) 8–12
reflect its resistance to degradation under the chem- Sulfur (mass% db) 0.5–0.9
ical and thermal environments of the blast furnace. Phosphorous (mass% db) 0.02–0.06
Such properties are related to lump size, shape Alkalies (mass% db) o0.3
and size uniformity, chemical composition, mecha- Data taken from Leonard DC, Bonte L, Dufour A, Ferstl A, Rai-
nical strength, and thermal and chemical stabilities. pala K, Scmole P, Schoone P, Verduras JL, and Willmers RR
Thus, coke for the blast furnace needs to be a (1996) Coke quality requirements of European blast furnace
engineers (joint EBFC-Paper). In: Proceedings of the Third
successful compromise between structure and prop-
European Cokemaking Congress, pp. 1–10. Gent, Belgium,
erties. To ensure good blast furnace performance, CRM–VDEh.
coke should be moderately large, with a narrow
size range, and have a high mechanical strength and
a high resistance to abrasion and thermal shock in Coke is a porous carbon material consisting of a
the blast furnace. Because of the many unknown network of pores of various dimensions and shapes,
factors, it is not possible to establish universal quality some of which are closed, but the majority of which
indices common to all blast furnaces, although typ- are interconnected. Generally, porosity of industrial
ical specifications for metallurgical coke quality are cokes is determined indirectly from the ratio of
available. It could be said that each blast furnace apparent and true relative densities, according to
depending on design and operation requires a tai- the equation
lored coke.  
Impurities present in coke (moisture, volatile mat- apparent density
porosity ð%Þ ¼ 100 1

ter, ash, sulfur, phosphorous, and alkali contents) true density
affect its performance in the blast furnace by
decreasing its role as a fuel in terms of amounts of A full description of the determination of apparent
carbon available for direct and indirect reduction and true specific densities of lump coke is given in the
roles and also its role as a permeable support. ISO 1014 and ASTM D167 procedures.
As in the case of coal, all determinations included Empirical mechanical strength tests, commonly
in proximate analysis (moisture, ash, and volatile used to measure resistance to size degradation,
matter contents) are the subject of national and in- involve dynamic loading either in the form of shat-
ternational standards (ISO 579 for the determination ter tests (ISO 6161, ASTM D3038), where breakage
of total moisture content; ISO 687 and ASTM occurs by impact, or revolving drum tests such as
D3173 for the determination of moisture in the ASTM Tumbler (ASTM D3402), MICUM, half- and
analysis sample; ISO 1171 and ASTM D3174 for the extended-MICUM, IRSID, and JIS (JIS K2151) tests,
determination of ash; ISO 562 and ASTM D3175 for where attrition takes place by a combination of bre-
the determination of volatile matter content). Table 6 akage and abrasion. In Europe, the MICUM and
summarizes typical coke chemical properties for IRSID (ISO 556 and ISO 1881) tests, which use the
some operating blast furnaces in Europe. same equipment, are dominant. The JIS test is widely
used in Japan and Australia, while the ASTM Tum-
Physical Tests bler is commonly used in North America. Table 7
The importance of the physical properties of coke is summarizes the cold mechanical strength methods
linked to the need to support the ferrous burden and for coke testing.
to give a permeable matrix through which reducing All these tests are based on the mechanical treat-
gases can flow and molten material can percolate in ment of a specific amount of coke (10–50 kg) with
the lower blast-furnace region. These physical prop- a defined size (460 to 420 mm) performed in a
erties are related to its size (mean and distribution) rotating drum under well–defined conditions (num-
and its resistance to breakage and abrasion. Coke ber of revolutions and rate). Afterwards, the coke is
size is mostly controlled by screening (ISO 728, sieved and the different size fractions weighted.
ASTM D293). A large mean size with a narrow size During mechanical treatments, coke fragmentation
distribution maintains adequate permeability. Most takes place by fissuring, cohesiveness, and abrasion.
operators consider a mean optimum size to be in the Two indices are normally derived from these tests:
range of 50–55 mm with a lower limit B20–30 mm one referred to fissuring or cohesion and the other to
and an upper limit B70–100 mm. abrasion.
196 COAL AND COKE

Table 7 Standard tests for assessing mechanical strength of coke

MICUM Half-MICUM IRSID ASTM Tumbler JIS

Standard procedure ISO 556 ISO 556 ISO 1881 ASTM D294 JIS K2151
Coke characteristics
Weight (kg) 50 25 50 10 10
Particle size (mm) 460 460 420 51–76 450
Type of sieve Rounded hole Rounded hole Rounded hole Square hole Square hole
Drum
characteristics
Length (m) 1 0.5 1 0.46 1.5
Diameter (m) 1 1 1 0.91 1.5
Test characteristics
Drum (rpm) 25 25 25 24 15
Duration (min) 4 4 20 58 2 and 10
Total revolutions 100 100 500 1400 30 (2 min); 150
(10 min)
Strength indices
Breakage M40 ¼ mass% As for MICUM I40 and I20 ¼ mass% Stability DI30/15 ¼ mass%
440 mm 440 and factor ¼ mass% 415 mm; DI150/
420 mm 425 mm 15 ¼ mass%
415 mm
Abrasion M10 ¼ mass % As for MICUM I10 ¼ mass% Hardness
o10 mm o10 mm factor ¼ mass%
6.3 mm

Adapted from Loison R, Foch P, and Boyer A (1989) Coke. Quality and Production. London: Butterworth.

High-Temperature Tests there is a high degree of correlation between the two

indices, CRI and CSR both determined by NSC
The importance of high-temperature properties of
(Nippon Steel Corporation) method. It has been es-
coke was established by the Japanese Steel Industry
tablished that CRI is strongly influenced by the rank
from the dissection surveys of three blast furnaces
and type and quantity of mineral matter of the parent
that were quenching whilst operating. As a result the
reactivity of coke to carbon dioxide was considered coal and CSR is also primary controlled by coal
properties and in general by coking conditions and
one of the most important characteristics to assess
oven design.
the quality of coke and the suitability of its use in the
The range of optimum values in coke quality
blast furnace. A high reactivity of cokes is dis-
parameters is wide and strongly dependent upon
advantageous because of a higher specific consump-
the characteristics and operational conditions of
tion of coke in the blast furnace. In addition, the
the blast furnaces. As an example, typical coke
reactive coke is mechanically less resistant, affecting
size, mechanical strength before and after reaction
the conditions of the gas flow in the blast furnace.
Coke reactivity is commonly determined using a with CO2, and reactivity index of cokes used in
different blast furnaces worldwide are given in
method developed by the Nippon Steel Corporation
Table 8.
(NSC), which has been adopted as an ASTM stand-
ard procedure (ASTM D5341). The method meas-
Coke Petrography
ures the solution loss reaction of coke by carbon
dioxide at 11001C for 2 h on a dried coke sample of Optical microscopy, using polarized reflected light,
200 g (2071 mm in size) under standardized gas flow has been recognized as a very useful tool and has
rate conditions. The coke residue is cooled in nit- played a major role in studies of carbonization
rogen and the weight loss after reaction is defined as mechanisms and coke quality properties. Examina-
the coke reactivity index (CRI). The gasified coke is tion of a coke polished surface under a microscope
subjected to a mechanical strength test in a drum reveals the presence of certain parts without optical
(600 revolutions at 20 rpm). The coke postreaction activity (isotropic coke) and others with optical
strength or coke strength after reaction index (CSR) activity (anisotropic coke). The anisotropic carbon
is defined as the percentage of coke 49.5 mm or exhibits yellow, blue, and purple areas of various
410 mm remaining after the mechanical treatment. sizes, o1–200 mm, and shapes (optical texture) with
For cokes produced under similar coking conditions, an interchange of color on rotation of the specimen,
 COAL AND COKE 197

Table 8 Required physical and high-temperature properties of blast furnace coke in current operation

European range Australian BHP Port Kembla American range Japan range

Mean size (mm) 47–70 50 50 45–60

M40 478–488 85 n.a. n.a.
M10 o5–o8 6.5 n.a. n.a.
I40 53–55 n.a. n.a. n.a.
I20 477.5 n.a. n.a. n.a.
DI150/15 n.a. 84.4 n.a. 83–85
ASTM stability n.a. 63.6 60 n.a.
CSR 460 74.1 61 50–65
CRI 20–30 17.7 23 n.a.

n.a.: not available.

After Dı́ez MA, Álvarez R, and Barriocanal C (2002) Coal for metallurgical coke production: Predictions of coke quality and future
requirements for cokemaking. International Journal of Coal Geology 50: 389–412.

in association with the use of a retarder plate. Each Overview. Microscopy Techniques: Light Microscopy;
color represents a given orientation of the aromatic Sample Preparation for Light Microscopy; X-Ray Micros-
lamellar molecules that constitutes the carbon ma- copy. Sample Handling: Comminution of Samples. Sa-
trix. To describe the different components of the op- mpling: Theory; Practice. Sulfur. X-Ray Fluorescence
and Emission: Energy Dispersive X-Ray Fluorescence.
tical texture of coke, different classifications and
nomenclatures have been developed by research cen-
ters and industry. Most of them make a distinction
between isotropic, mosaics of various sizes, flow– Further Reading
type anisotropy of various sizes and shapes, and in- Alpern B and Lemos de Sousa MJ (2002) Documented in-
erts. Quantification of the optical textural compo- ternational enquiry on solid sedimentary fossil fuels;
nents can be conducted by means of a point counting Coal: definitions, classifications, reserves-resources and
technique on a sample embedded in resin (e.g., energy potential. International Journal of Coal Geology
ASTM D3997 and ASTM D5061). 50: 3–41.
By combining this microscopic technique with an Bertkowitz N (1979) An Introduction to Coal Technology.
image analyzer, the analysis program allows the de- London: Academic Press.
termination of total porosity of coke, the number, the Couch GR (2001) Metallurgical Coke Production. London:
IEA Coal Research.
total perimeter, the average diameter, and the wall
Davidson RM and Clarke LB (1996) Trace Elements in
thickness of pores.
Coal. London: IEA Coal Research.
The use of optical microscopy for qualitative and Elliot MA (1981) Chemistry of Coal Utilization, Second
quantitative analysis of the optical texture of coke is supplementary volume. New York: Wiley.
performed rather in the field of research on carbon- Kural O (ed.) (1994) Coal. Resources, Properties, Utiliza-
ization than in a full characterization of coke by the tion, Pollution. Turkey: Orhan Kural.
industry. Marsh H and Rodrı́guez-Reinoso F (eds.) (2000) Sciences
It is established that the rank and chemistry of the of Carbon Materials. Spain: Secretariado de Publicaci-
parent coal strongly influence the optical textures of ones, The University of Alicante.
cokes. Several studies show that the development of Stach E, Mackowski MTh, Teichmüller M, et al. (1982)
anisotropy (size, shape, and intensity) during car- Stach’s Textbook of Coal Petrology, 3rd edn. Berlin:
Gebrüder Borntraeger.
bonization varies mainly with (1) coal rank; (2) pet-
Scott AC (2002) Coal petrology and the origin of coal
rographic composition of the coals; (3) plasticity of
macerals: A way ahead. International Journal of Coal
the parent coal; (4) carbonization conditions such as Geology 50: 119–134.
rate of heating, soak time, and gas overpressure; as Scott DA (1994) Development Affecting Metallurgical
well as (5) the nature of additives used in the coal Uses of Coal. London: IEA Coal Research.
blends. Van Krevelen DW (1993) Coal, Typology–Physics–Chem-
istry-Constitution, 3rd edn. Amsterdam: Elsevier.
See also: Activation Analysis: Neutron Activation. Ward CR (1984) Coal Geology and Coal Technology. Ox-
Atomic Absorption Spectrometry: Principles and In- ford: Blackwell Scientific.
strumentation. Atomic Mass Spectrometry: Inductively Zimmerman RE (1979) Evaluating and Testing of Coking
Coupled Plasma. Carbon. Geochemistry: Soil, Organic Properties of Coal. San Francisco: Miller Freeman
Components. Humic and Fulvic Compounds. Microscopy: Publication.