The University of New South Wales

School of Materials Science and Engineering of

Masters Research Degree

The Effect of Coal Properties on

Carbonization behaviour and Strength of
Coke Blends

August 2012

Kyung Eun (Jake), JEE

Supervisors

Prof. Veena Sahajwalla

Dr. Sushil Gupta
ACKNOWLEDGEMENTS

First and foremost I would like to sincerely acknowledge the people and organizations
contributed to publish this thesis. I gives grateful thanks to the supervisor, Professor
Veena Sahajwalla for her relentless guidance during the research of study and Dr. Sushil
Gupta, co-supervisor for his suggesting research topic, discussing study procedures and
solving experimental problems. Additionally many great advices from Dr. ByoungChul
Kim and Mr. DongMin Jang, which are important for finishing thesis work, are also
appreciated.

I also would like to be grateful to Dr. WoonJae Lee for his whole support of
experimental planning, Mr. YoungHa Bae, technology development team leader of the
POSCO coke making department who provided all coal samples and all the following
staff members of the coke making technology development team in Pohang Works for
their all efforts in processing experiment: Mr. HeykJin Kim, Mr. TaeMan Kang, Mr.
YongJin Kim, Mr. JongSu Heo, Mr. YongPyo Kim and Mr. BumDuk Han. Special
thanks are given to Mr. JinKook Park and Mr. HyunJun Shin for entire collaboration
involved in the preparation of the coal samples.

I wish to give appreciation to the general manager of the POSCO coke making
department in Pohang Works, Mr. DongWon Kim for his all care to accomplish this
study, Mr. DongKi Son, Mr. Hyung-Ki Seong, Mr. JongYup Choi, Mr. SeaKyoung Oh,
Mr. KiSeog Choi and Mr. SeungOh Yoon for their all instruction of the coke making
technology, Mr. JaeHong Yu and Mr.SangDong Lee for their precious advise, Mr.
JeongHoon Kim, Mr. SangWan Kim, Mr. JaeWon Hur and Miss SuJin Oh for filling the
author’s absence in the cokemaking department of POSCO Pohang Works during this
study.

A sincere debt of gratitude is owed to POSCO executive members for their sponsoring,
Mr. JoonYang Chung, Mr. JinIl Kim, Mr. BongRae Cho and KeeChang Lee.
Additionally, YongOk Choi, Chief Executive Officer of POSFINE, without whom I
never would have been able to come to the Australia and do this study. Finally,
particular thanks to POSCO Australia members: Mr. SungWook Kang, Mr. JongHun
Jong, and Mr. InHyunk Yeou for the supporting life in Australia.

Mostly, I also would like to give my gratitude to my wife Ms. JiSong Yi who always
take care of our family and YeWon Jee, my daughter continuously giving me power to
finish thesis by million dollars of smile.
Table of Contents

1 INTRODUCTION AND OBJECTIVE .........................................................1

1.1 Introduction ..................................................................................................................... 1
1.2 Objectives ....................................................................................................................... 6
2 LITERATURE REVIEW ...............................................................................7
2.1 Iron-making Process ....................................................................................................... 7
2.1.1 Blast-Furnace Process ............................................................................................. 7
2.1.2 Coke-Making Process ............................................................................................ 11
2.2 Coal Characterization ................................................................................................... 14
2.2.1 Coal Composition .................................................................................................. 14
2.2.2 Coal Classification................................................................................................. 15
2.2.3 Physical Properties ................................................................................................ 19
2.2.4 Chemical Properties ............................................................................................... 29
2.2.5 Coal Petrography ................................................................................................... 29
2.2.6 Inorganic Matter .................................................................................................... 34
2.3 Carbonization Phenomena ............................................................................................ 36
2.3.1 Meta-plast Formation ............................................................................................ 36
2.3.2 Resolidification and Semicoke .............................................................................. 45
2.3.3 Carbonization in Industrial Oven .......................................................................... 49
2.4 Factors affecting Coke Properties................................................................................. 50
2.4.1 Coke Quality Parameters ....................................................................................... 50
2.4.2 Coal Rank .............................................................................................................. 57
2.4.3 Coal Maceral.......................................................................................................... 60
2.4.4 Coal Fluidity/Rheology ......................................................................................... 69
2.4.5 Coal Inorganic Matter............................................................................................ 76
2.5 Coke Strength Prediction .............................................................................................. 79
2.5.1 Coke Strength Model ............................................................................................. 79
2.5.2 Effect of Blending ................................................................................................. 83
2.6 Summary....................................................................................................................... 87
3 EXPERIMENTAL ...................................................................................... 89
3.1 Coal Samples ................................................................................................................. 91
3.2 Carbonisation Tests ....................................................................................................... 95
 3.2.1 Carbonisation in 30 kg Operations ........................................................................ 97
3.2.2 Carbonisation in Industrial Oven ........................................................................... 97
3.2.3 Coke Strength Measurement ................................................................................. 98
3.2.4 Coke Texture ....................................................................................................... 100
3.2.5 Physical Structure (SEM/Optical Microscopy) ................................................... 101
4 ASSOCIATION BETWEEN COAL PROPERTIES AND COKE
STRENGTH ..................................................................................................... 102
4.1 Effect of Coal Rank ..................................................................................................... 102
4.2 Effect of Coal Rheology .............................................................................................. 106
4.3 Effect of Coal Rank and Rheology on Coke Strength of Blends ................................ 109
4.4 Coke Micro-texture...................................................................................................... 111
5 EFFECT OF VITRINITE TYPES ON COKE STRENGTH.................... 117
5.1 Effect of Vitrinite Composition ................................................................................... 118
5.2 Effect of Sub-maceral on Coke Strength ..................................................................... 119
6 EFFECT OF BLENDING ON COKE STRENGTH ................................ 123
6.1 Vitrinite Composition and Coal Property .................................................................... 123
6.2 Vitrinite Composition and Coke Strength ................................................................... 124
6.2 Implication on Industrial Coke Strength ..................................................................... 126
7 CONCLUSIONS ....................................................................................... 129
List of Figures
Figure 2.1 Schematic of blast furnace and movement of gas (Hutny, 1991) ........................... 7
Figure 2.2 Relationship between the blast furnace volume and adjusted coke strength
(Nakamura, 1981) .................................................................................................. 10
Figure 2.3 Conventional Coke Making Process ..................................................................... 11
Figure 2.4 Non - recovery coke oven (Bertling, 1999) .......................................................... 13
Figure 2.5 Coalification Process ............................................................................................. 15
Figure 2.6 Estimators of coal rank (Esterle, 2007) ................................................................. 17
Figure 2.7 Correlation of the coal density with carbon contents (Huang, 1995) and
Densities of coal macerals (Krevelen, 1993) ....................................................... 20
Figure 2.8 Microhardness index of coal (Krevelen, 1993) ..................................................... 21
Figure 2.9 Hardgrove Grindability Index of coal (Krevelen, 1993) ....................................... 20
Figure 2.10 Correlation of atomic H/C and O/C for coals of different ranks
(Neavel Richard, 1981) ......................................................................................... 21
Figure 2.11 Model of bituminous coal structure (Shinn, 1984) .............................................. 22
Figure 2.12 Three maceral groups (Stach and Murchison, 1982) ........................................... 31
Figure 2.13 Maceral reflectance vs. Coal rank (Neavel Richard, 1981) ............................... 328
Figure 2.14 SEM from QEMSCAN of a coal sample; 63-90 µm (Gupta, 2007) .................... 35
Figure 2.15 Scheme of general reaction for carbonization (Lewis, 1982) .............................. 37
Figure 2.16 Mechanism of coal aromatization and condensation ........................................... 38
Figure 2.17 Formation and development of the pore structure during carbonization of
coal (Hays, 1976) ................................................................................................. 39
Figure 2.18 Proposed mechanism for the occurrence of coal softening
(Takanohashi,2005) ............................................................................................. 42
Figure 2.19 Development of anisotropic spherules in the resolidification region
(Loison, 1989) ..................................................................................................... 47
Figure 2.20 SEM photographs for the heat-treated (A) Lusca, (B) Pittstone-M, and (C)
Witbank coals (Kidena, 1996) ............................................................................. 48
Figure 2.21 The mechanism of carbonization in coke oven .................................................... 50
Figure 2.22 Drawing of the generation of fissures found in coke pieces (Ragan, 1980) ........ 52
Figure 2.23 Model of micro-texture of coke (Duval et al., 1988) ........................................... 53
Figure 2.24 Increasing of CSR with coke anisotropy (Vogt, 1990) ........................................ 54
Figure 2.25 The factors of influence on coke qualities ........................................................... 55
Figure 2.26 Effect of volatile matter and coal reflectance on coke CRI and CSR
(Zhang et al., 2004) .............................................................................................. 57
Figure 2.27 The effect of the volatile content of coal on the cold strength of coke
(Vanniekerk and Dippenaar, 1991)...................................................................... 58
Figure 2.28 The relation between maximum fluidity and coke strength; M30
(Lin and Hong, 1986) .......................................................................................... 59
Figure 2.29 G Relationship between coal rank and coke wall carbon forms
(Benedict and Thompson, 1980).......................................................................... 60
Figure 2.30 Reflectogram: mean reflectance, scatter and swelling index of 5 coking coal
blends (Stach and Murchison, 1982). .................................................................. 61
Figure 2.31 A Effect of vitrinite content on the total dilatation (Gransden et al., 1991)
and correlation between Tmax and Gieseler plastometry-related
temperatures (Kidena et al., 2002)....................................................................... 63
Figure 2.32 Images of three macerals of vitrinite group (Krevelen, 1993) ............................. 64
Figure 2.33 Reflectogram of pseudovitrinite distribution (Kruszewska, 1998) ...................... 65
Figure 2.34 Gieseler Fluidity of Diagram ............................................................................... 70
Figure 2.35 Curve recorded by the dilatometer during heating (Loison et al., 1989) ............. 71
Figure 2.36 Scale of reference profiles for the crucible swelling test (Loison et al., 1989) .... 72
Figure 2.37 G-Factor Measuring Apparatus ............................................................................ 73
Figure 2.38 Relationship between rheological properties and CRI and CSR (Zhang et al.,
2004) .................................................................................................................... 74
Figure 2.39 A schematic illustration of the dependence of coke strength on the coal rank
and caking properties of coal (Vanniekerk and Dippenaar, 1991) ...................... 74
Figure 2.40 The Coal expansion mechanism; left and the definition of specific dilatation
volume; right (Nomura, 2004) ............................................................................. 75
Figure 2.41 Schematic of effect of interactions on the thermoplastic behaviour of blends
(Sakurovs, 2003) ................................................................................................ 765
Figure 2.42 SEM images illustrating mineral distribution in coke (Gupta, 2008) .................. 77
Figure 2.43 The total percentage of iron, potassium and sodium in the amorphous phase
in coke versus initial apparent rate (Grigore, 2006) ............................................ 78
Figure 2.44 Experimental coke quality prediction model of Bao steel (Zhang et al., 2004)
............................................................................................................................. 82
Figure 2.45 Industrial coke quality prediction model of Bao steel (Zhang et al., 2004) ......... 83
Figure 2.46 Relationship between the experimental and calculated Irsid, CRI and CSR
indices (Álvarez et al., 2007) ................................................................................ 82
Figure 2.47 Fluidity of bituminous coals vs. rank showing specifications of coal
properties for conventional and 30% partially briquetted charges
(Gransden et al., 1991).......................................................................................... 83
Figure 3.1 Experimental procedure of the sample preparation and analysis ......................... 90
Figure 3.2 Polarized light of optical microscope; Leica DM4500P ...................................... 94
Figure 3.3 25kg Movable Wall Pilot Coke Oven ................................................................. 97
Figure 3.4 Drum Index Revolving Tester .............................................................................. 98
Figure 3.5 I-tumble tube for CSR test ................................................................................. 100
Figure 3.6 Scanning Electron Microscope; Hitachi S3400-I............................................... 101
Figure 4.1 Coke strengths vs. mean reflectance as well as volatile matter of coals ............ 103
Figure 4.2 Coke strengths vs. mean reflectance and volatile matter of binary blends ........ 104
Figure 4.3 Percentage distributions of vitrinite vs. reflectance values of component
coals and blends ................................................................................................. 105
Figure 4.4 Coke strength vs. Rheological parameters of single coal .................................. 107
Figure 4.5 Coke strength vs. Rheological parameters of blended coals .............................. 108
Figure 4.6 Fluidity ranges of blended coal and their original coals .................................... 109
Figure 4.7 Contour maps of DI and CSR with coal rank and fluidity parameters of
binary blends ...................................................................................................... 110
Figure 4.8 Coke strength vs. micro-texture of all the tested binary coke blends ................ 112
Figure 4.9 Microscopic textures of cokes in binary blends and their original coals ........... 114
Figure 4.10 Cross sectional SEM images of cokes ............................................................... 115
Figure 4.11 SEM images of coke morphology ...................................................................... 116
Figure 5.1 Microscopic images of vitrinite maceral in single coal ..................................... 117
Figure 5.2 Vitrinite submacerals vs. fluidity parameters of coals ....................................... 118
Figure 5.3 Correlations between the vitrinite indicators and coke strength for single
coals ................................................................................................................... 121
Figure 6.1 Correlations between the indicators of vitrinite and coal parameters of
binary blends. .................................................................................................... 124
Figure 6.2 Vitrinite macerals vs. Coke strength of binary blend. ......................................... 128
Figure 6.3 Vitrinite macerals vs. Coke strength of the industrial test. ................................. 128
List of Tables

Table 2.1 Requirements on chemical composition of coke in Germany (Bertling, 1999) ........ 8
Table 2.2 Requirements on physical properties of coke in Germany (Bertling, 1999) ............. 9
Table 2.3 ASTM D-388 Classification of coal by rank (Krevelen, 1993) .............................. 18
Table 2.4 International codification of higherrank coals (Krevelen, 1993) ............................ 19
Table 2.5 Parameters for coal molecule (Spiro and Kosky, 1982) .......................................... 26
Table 2.6 Petrographic nomenclature (Loison et al., 1989) .................................................... 30
Table 2.7 Different Maceral Standards (Gupta and Shen, 2009) ............................................ 34
Table 3.1 Coal characteristics used for the investigation ........................................................ 92
Table 3.2 Characteristics of the blended coal used for the investigation ................................ 93
Table 3.3 POSCO Standards of maceral devised .................................................................... 95
Table 3.4 Coke characteristics from carbonized single coal ................................................... 96
Table 3.5 Coke characteristics from carbonized coal in blends .............................................. 96
Table 3.6 Coke texture classification of POSCO .................................................................. 100
Table 4.1 Microscopic texture data of individual and blend cokes ....................................... 111
Table 5.1 Vitrinite indicators of single coals and their cokes strengths ................................ 119
Table 6.1 Vitrinite indicators of coal blends and their cokes strengths ................................. 124
Glossary

PCI = Pulverized Coal Injection in blast furnace

COG = Coke Oven Gas from carbonization of coal in coke battery

H/C Ratio = Proportions of Hydrogen compare to the Carbon contents

VM = Volatile Matter, specially in this thesis restricted in coal or coke

MJ = Mega Joule (unit of energy)

ASTM = American Society for Testing Materials

ECE = Economic Commission for Europe

AS = Australian Standards

ICCP = International Committee for Coal and Organic Petrology

JIS = Japanese Industrial Standards

ISO = International Organization for Standardization

RDCIS = Research & Development Centre for Iron & Steel, SAIL in India

SS = Shatter strength of coal or coke

HGI = Hardgrove Grindability Index of coal

wt % = weight percentage

dmmf = Dry Mineral Matter Free basis

daf = Dry Air Free base

MM = Mineral Matter in coal or coke

SEM = Scanning Electron Microscopy

CS = Chloroform Soluble in coal when it was carbonized

CIPS = Chloroform Insoluble Pyridine Soluble in coal when it was carbonized

PIMS = Pyridine Insoluble CS2/NMP Soluble in coal when it was carbonized

MI = CS2/NMP Insoluble in coal when it was carbonized

CSR = Coke Strength after Reaction, stands for the thermal strength of coke
CRI = Coke Reaction Index, stands for the reactivity of coke

DI = Drum Index, stands for the cold strength of coke

DI 15015 = 10kg of +50mm coke is tumbled in a cylindrical drum with 150 revolutions at
a speed of 15 revolutions per minute and the coke is screened on a 15 mm
sieve

DI 150 30 = same as above but screened on a 30 mm sieve

MLD = Modified Linear Drum

DDPM = Dial Division Per Minute, the index of Gieseler fluidity

lg MF, LMF = Logarithmic maximum value of ddpm

TD = Total Dilatation, rheological parameters of coal

FSI = Free Swelling Index, rheological parameters of coal

MFT, Tm = Maximum Fluidity Temperature of coal when it was carbonized

Ts = Softening Temperature of coal when it was carbonized

RT, Tr = Resolidification Temperature of coal when it was carbonized

MOF = Maximum Of Fluidity

Tmax = Maximum value of Temperature

Rvo = Random Reflectance of coal

Rm = Mean Reflectance of vitrinite in coal

GNY = Goonyella, abbreviation of coal name

WIT = Witbank, abbreviation of coal name

AC= Australian Coal, abbreviation

CA = Canadian Coal, abbreviation

CC = Chinese Coal

RC = Russian Coal

XRD = X-Ray Diffraction

GMDH = Group Method Data Handle

SCO = Simulated Coke Oven

MCI = Mineral Catalysis Index

CDQ = Coke Dry Quenching

MO = Molecular Orientation

LMO = Length of Molecular Orientation

R2, r2 = Correlation Factor of Regression Analysis

MINITAB = Analytical Software

Abstract

The composition of a number of coals with varying rank was characterized by a number
of means with emphasis on petrographic examination. The petrographic data of coals
was processed with special emphasis on quantifying vitrinite sub macerals and their
association with coal carbonisation parameters as well as coke strength. The association
between coal petrography data with coke strength of binary blends as well as industrial
blends was also examined. Coke strength parameters including drum index and CSR
were found to be predominantly influenced by coal rank with few exceptions of high
rank and low fluidity coals. The study showed the association between coke strength
and coke micro-texture such that coke strength after reactions was related to increased
anisotropic texture. On the basis of vitrinite sub-components a new coal parameter was
devised. The new parameter, the ratio of difference of collinite and telinite to the mean
vitrinite reflectance, is found to predict coke strength of single coals with high
confidence. However, the same parameter could not predict coke strength of binary
coke as well as coke made in industrial processes. The study suggested that interaction
of sub-vitrinite during carbonisation could be responsible for the limitation of
application of the new parameter to predict coke strength of blends. Further
investigations relating to the interaction of sub-macerals during carbonisation as well as
the role of minerals may improve the confidence of coke strength prediction using the
proposed new index.
1 INTRODUCTION AND OBJECTIVE

1.1 Introduction

Conventionally, a Blast furnace has been used for a long time to produce pig iron
by smelting reaction. Inside the blast furnace, the raw materials such as coke and
iron ore are charged while preheated blast air is blown into the furnace downside,
thus chemical reactions take place throughout the furnaces as raw materials move
downward. Specifically, metallurgical coke, a porous carbon material with high
strength produced by carbonization of coal blends at the temperature about
1,200°C , is utilised to maintain the iron making process in blast furnaces. The
coke not only provides the reducing gases for the iron oxide in iron ore with the
heat for the reaction but also plays an important role as a support in the iron ore
zone with high permeability.

Other alternatives such as oil, gas, plastics and coal can be injected to blast
furnace to generate heat and to increase the reduction rate of the iron ore.
However, there is no other substitute of coke which possesses the ability of
supporting cohesive zone with high permeability in blast furnace. According to
the increasing global demand of the iron production with the cost deduction,
increasing inner volumes of the blast furnace, for instance, from 1000 m 3 in
1960’s to 5000m3 in 2000’s has been a trend. In addition to that, high level of
injected alternatives to blast furnace with thin coke layer can provide cost-
effectiveness to blast furnace operation (Cross, 1994). To keep a thin coke layer in
an increased volume of blast furnace with long residence and high reduction rate
of iron ore, the achievement of the high coke quality is essential by enhancement
of the various key properties such as large coke size with even distribution, high
mechanical strength and high coke strength after reaction. Otherwise, same
quality coke as in smaller size of blast furnace can’t resist in larger size of blast
furnace operation (Nakamura, 1981).

1
However, improvement of the coke quality is very sophisticated process whereas
it is imperative to increase coke quality for the operation of the bigger blast
furnace. This is because coke quality is influenced by various factors such as
carbonization conditions of the coke oven, charging conditions of the coal to coke
oven, coal particle size, a coke quenching system and coal properties. Moreover,
coke oven operation is already conventionalized to the optimum condition as the
coke production rate and qualities for the conservation of the coke oven. To
preserve campaign life of the coke oven, coking temperature in coke oven has to
be everlastingly stable as about 1200 ºC by controlling the calorie balance.
Furthermore, rapid change of operational condition can cause air pollution by
leakage of the coke oven gas or micro-size carryover of the coal. Therefore,
innovative improvement of the coke quality seems to be an unsolved issue due to
complexity of the operational condition.

Coking coal, as a carbonization source, is an element to improve coke qualities

even for slight level by blending various rank coals which have diverse properties.
The better choice of the coking coals can be made with better coking conditions
with coal properties, resulting coke of better qualities. The effect of various coal
properties such as chemical composition, rheological properties, geological
characteristics and optical appearance constituent was examined for an
optimization of a coal selection to carbonize and thus specific conditions of those
properties are already decided and used in a real coking operation. However, the
thermo-dynamic behaviour of a coking coal in a micro-level still remains as black
box, which is a key element to understand the enhancement mechanism. This is
based on the difference of coal properties in between a laboratory scale test and an
industrial operation, because industrial operation can easily be influenced by more
variables.

The coal blending work, requiring the choice of individual coal to blend and the
management of shipping schedules, plays a significant role in the production of
coke in terms of cost and quality. To meet cost-effectiveness of operation, most of
coke plants are using various kinds of coals to produce the most cost-effective
coke by blending coals in wide-range of cost. Blending of coals is also done based

2
on plant logistics and ease of operation, so that most coke making plants are
located on sea shores to give availability on importing many kinds of coals.
Moreover, different types of coals have to be blended in proper percentages to
meet the demand of coke qualities while single coal generally can’t fill up
requirements (Japan, 1979). Besides, the Geological affinity on blended coals can
make synergy effect of coke. Thus, in POSCO Pohang Works, 12 different kinds
of coals are blended at once for coke making, 40 different kinds of coals are used
in a year. The quality of coal blends rely on the characteristics on coking coals to
mix and their interaction making up the blends. This means the blending effect of
coals are not only influencing on achieving high qualities on coke, but also
imperative to understand to know how to control desirable qualities of coke.

To understand the blending effect on coking coals, whole comprehension of coal

properties such as rank, rheology and maceral is essential. The concept of the coal
blending in coke making seems to be comparable with concrete making process.
The raw materials of this process composed of cement, sand, gravel, and water.
The cement, firstly as a main source of concrete, is consistent with carbon texture
of macerals of coal. To increase bonding power between cements and water
mixing which is equivalent to rheology of coal has to be well distributed in
concrete. Besides, sand and gravel are equal to rank of coal to give hardness to the
cement by forming a basic structure. These all of components are to be essential
to cure high quality of cement as all properties of coking coal are also essential to
make high quality coke. Therefore, the carbonization behaviour of a coal effecting
on coke qualities are dependent on the individual properties of source coal and
their maceral proportions as well as their distribution (Loison et al., 1989).

The optical appearance on coals, macerals which are distinguished by their origin
such as plant, wood, bark and algae, is a key factor to determine coke qualities
such as coke strength, reactivity, composition and size. According to the Vasko et
al. (2005), coal components are divided into two groups, reactive e.g. vitrinite,
exinite, etc. and inert e.g. fusinite, micrinite, etc. In most coal blends, the coals
interact at the temperature range from 400 to 520ºC in a way affecting caking
properties such as fluidity and dilatation (Sakurovs, 2003). These caking

3
properties can be well explained by thermodynamic behaviour of the reactive
group of coal which contains high volatile matter than inert. On the other hand,
inert group, rich in carbon, poor in hydrogen, remain infusible during
carbonization, which does not mean to say that they are completely inert.

Generally, coal macerals of coking coal are divided into three groups, vitrinite,
exinite and inertinite which can be classified by the ratio of light reflection.
Vitrinite group in coking coals, the most part of the maceral, has characteristics
which swelling and agglomerating during the coking stage. Its density increases
with rank from 1.2 to 1.7 (Loison et al., 1989). Exinite group, the most fluid
maceral during coking stage, is derived from the organs that are comparatively
poor in oxygen: algae, spores, pollens, etc. It is chemically distinguished by the
residence of 10 or 20% of aliphatic carbon chains (Loison et al., 1989). Inertinite
group, the most dense maceral, is often originates from the residue of woody
components which undergone long aerobic oxidation or partly burnt. In the coking
process, inertinite remains as inert except for a few semifusinite.

Specifically, vitrinite group which originates from lignocellulose tissues e.g. wood,
roots, bark, etc., appears to be a wall surrounding the other macerals and mineral
matter and also is the most abundant, isolated and homogenous maceral group of
coal. Vitrinite constituent is the majority of normal coals, so that its
thermodynamic behaviour is very similar to that of the whole coal (Loison et al.,
1989). Especially, in high rank coals, the reflectance of the vitrinite is highly used
to measure their ranks. Vitrinite oxygen contents are higher than other macerals
whereas hydrogen and carbon contents are intermediate (Davis, 1976). Therefore,
due to high hydrogen ratio, vitrinite can be an indicator of temperature when coal
start to metamorphose.

Thus, vitrinite reflectance, displaying optical anisotropy, seems to be one of the

significant factors to correlate the coke strength because a plastic phase of coking
coal is a significant stage to determine the quality of coke. Vitrinite reflectance is
also intimately connected with the aromaticity of coal chemistry which is also
related to the coal rank. The determination of the optimum ratio and distribution

4
of vitrinite components in blending coal is a key factor to establish maximum
strength of coke, while optimum ratio of exinite and inert still remains as another
key factor.

Besides, other maceral groups are also important to investigate relationship

between coal petrography and coke quality. Exinite group shows greater plasticity
during coking stage than vitrinite group whereas exinite is a minor component of
coking coal in general. For the same rank coking coal, exinite shows
comparatively low reflectance and the highest hydrogen rate, volatile matter and
aliphatic carbon chain rate (Stach and Murchison, 1982). Inertinite group is also
closely related to coke quality by containing higher aromatic carbons which will
compose basic structure of coke as the gravel and sand in concrete. Excluding few
exceptions of Australian bituminous coals, inertinite is regarded as inert material
when it was carbonized (Diessel, 1983).

However, the petrography of individual macerals is still not well understood due
to the complexity of maceral-specific analysis (Walker, 2005). This complexity is
based on the characteristics of maceral measurement which are time-consuming,
subjective and requiring high level skill. Moreover, coke quality prediction by
vitrinite reflection has become less reliable with increasing usage of non-premium
coal for cost-effective operation. Particularly, further research of the effect of the
sub-vitrinite macerals such as telinite, collinite, etc. on the coke properties still
remains unclear, especially in the matter of their proportions and distribution.

As we discussed, the effect of blended coal properties on coke quality is still

unsolved question while abundant number of studies and experiments has been
processed. Coking coal is a homogenous material but is still composed of various
kinds of macerals originated from various kinds of plants. Even in similar coal
blends, coke qualities are quite different by geographical origin due to the maceral
composition. Vitrinite effect on coke qualities is revealed as a key factor to
understand the plastic phases of carbonization, but the effect of the individual
macerals such as telinite and collinite on coke qualities is still unclear. Therefore

5
the pilot scale coke oven test for coal blends would be necessary to examine the
effect of coal properties on coke quality.

1.2 Objectives

The main objective of this research work is to examine the coking coal effect on
the coke qualities by investigating the key factors: coal rank, rheology, maceral
composition, sub-vitrinite and their implication when they are carbonized. The
specific objectives are to research the effect of the following parameters:

1. To address the effect of coking coal properties on coke qualities by

statistical analysis and its microscopic structure.

2. To investigate the effect of vitrinite composition of coking coal on coke

strength and its implication by statistical method.

3. To validate sub-vitrinite effect on coke qualities and determine new sub-

vitrinite index to apply in coal blending.

In those objectives, we will focus on the examination of the coke strength among
coke qualities, specifically defined as cold strength and CSR (Coke Strength after
reduction Reaction in the furnace). For qualitative investigation, data are
interpreted in a statistical manner to provide specific correlations among the
properties.

6
2 LITERATURE REVIEW

2.1 Iron-making Process

2.1.1 Blast-Furnace Process

A blast furnace has occupied a predominant position as a complex reaction vessel

due to effective process design involving a counter current flow with high heat
and mass transfer such as gas, powder, liquid and solid. The solid lumpy materials
such as iron oxides e.g. iron ore, sinter, pellets, coke etc. are charged as flux into a
tall and vertical shaft furnace, oxygen contents of hot air at 900~1200℃ is blown
into through tuyere in the lower part of furnace. The indirect reduction which
extracts oxygen from iron ores such as hematite and magnetite is continuously
processed in cohesive zone by the carbon monoxides generated in combustion
zones near the base of the packed bed (Burgess, 1985). The direct reduction
produces carbon monoxides and iron in the lower part of furnace (Hutny et al.,
1991). Theses melting process separate the formed iron and gangue materials
which are insoluble in iron as slag phase, the end products are discharged through
the bottom of hearth.

Figure 2.1: Schematic of blast furnace and movement of gas (Hutny et al., 1991)

7
Metallurgical Coke plays a particularly significant role in blast furnace reaction
because it is the only solid substance which was descended from top to bottom.
Coke plays triple roles, specifically a physical, thermal and chemical role
(Bertling, 1999). Coke’s physical role could be explained as strong grid to support
the weight of overlying burden packed layer as well as permeability in cohesive
zone, the region where the molten iron bearing materials everlastingly forms
impermeable layers. The thermal role of Coke appears to be as a fuel providing
the heat required for reductions and producing iron and slag. The last role is as a
chemical agent to produce and regenerate reducing gases to reduce iron oxides
and carbon by the solution loss reaction.

Therefore, coke quality requirement is essential in terms of chemical and physical

properties and are tabulated in Table 2.1 and Table 2.2 (Bertling, 1999). For
instance, alkali contents are to be restricted up to 3.5 kg/ton per pig iron into
German blast furnace. Pig iron has to be de-sulphurized before charging into
oxygen inverter. Coke moisture and chlorine can disturb efficient operation by
increasing dew point of top gas. High coke ash appears to be high slag volume
consequently causing higher energy consumption and lower productivity. Physical
properties are also of greater importance to sustain cohesive zone with permeable
support enduring continuous abrasion.

Table 2.1: Requirements on chemical composition of coke in Germany (Bertling,

1999)

8
Table 2.2: Requirements on physical properties of coke in Germany (Bertling,
1999)

As the global trends of steel production, the production amount of crude steel in
iron-making companies has been increasing by the reliable way of increasing
inner volumes of blast furnaces. Owing to the characteristics of blast furnace
operation, the larger size, it is greater advantageous than other steel making
process in energy efficiency. The recent constructed blast furnaces have become
higher and wider to produce larger amount of pig iron, for example, the volume
has been increased about 5 times from 1000 m3 in 1960’s to 5000 m3 in 2000’s.
However, the increased height and diameter of a furnace inevitably are creating
difficult permeability control problems. When the furnace diameter is large, the
burden in the shaft section has not enough supporting power to maintain a balance
between its own weight and the ascending force of the reaction gas (Nakamura,
1981)

Besides, reducing coke rate by high levels of injection of carbon via the tuyeres,
particularly in pulverized coal injection (PCI), has also been a trend for the cost-
effectiveness of blast furnace operation due to the higher price of coke (Diez et al.,
2002). In addition to that, the steel industry has continuously pressured to reduce
greenhouse gas emissions to prevent global climate change since Kyoto protocol
in 2005. Thus further improvement of energy intensity in iron-making process
would require new approaches: carbon-lean fuel injections or operating the blast
furnace at relatively lower temperatures.

9
Therefore, due to iron-making trends towards increased size, low-cost operation,
energy efficiency, the roles of coke as a permeable support and a catalyst became
greatly significant and so thus further need remains for coke quality that has ‘a
sufficiently strong structure and hard surface to resist abrasion and
fragmentation’(Gupta, 2008). The influence of coke mechanical qualities
increases with increasing hearth diameter and blast volume as shown in Fig. 2.2.
According to the case study by Nakamura (1981), if the same quality cokes were
used for bigger blast furnaces, there would be lots of troubles, such as decreased
permeability, increased thermal load on the furnace walls, unstable tapping work
and irregular slag removals. Besides, for the high level of PCI, strong mechanical
strength of coke is needed to sustain thin coke layer. Lastly, to meet the future
requirement of iron-making process such as low-temperature operation or carbon-
lean fuel injection, the higher reactive cokes without high rate of abrasion and
fragmentation are essential.

Figure 2.2: Relationship between the blast furnace volume and adjusted coke
strength
(Nakamura, 1981)

10
2.1.2 Coke-Making Process

Despite the stress of importance of coke qualities to meet current and future trends
of blast furnace operation, the coke qualities still remains hard to improve. The
one of the main reasons is that coke making process is already optimized for the
best quality of coke. Conventional coke making process roughly can be classified
into three parts: coal & coke handling, coke oven and coke oven gas purification
with by-product producing as shown as Fig. 2.3. Firstly, coal was loaded from
ship to coal yard, transferred from yard to the coke oven through belt conveyor in
coal & coke handling process. During this stage, coal was crushed under about -
3mm by the mechanical hammer and dried through dryer and mixed through
blending bin. The modification of this process costs too much and is complicated
due to the long transferring line. In coke oven process, coals are charged into
coking chamber and indirectly combusted without oxygen for about 15 to 25
hours. Through this process, breeze coals are transformed to lump cokes via
plastic layer. The production factors such as coking time and temperature are
adjusted by the coke demand of blast furnace although low heating rate and long
coking time of the oven operation can guarantee big and strong cokes. However,
gas purification stage is not directly related to the coke qualities when combusted
raw gases from coals are refined to by-products such as Coke Oven Gas (COG),
tar and light oil.

Figure 2.3: Conventional coke making process

11
The restricted logistics of coking coal are also disturbed to the optimization of
coke qualities. Currently, in coke making demands, a coal blend that is low in cost,
produces a high quality coke, and provides a safe oven pushing performance, has
been suggested by the industry due to the limited availability of prime coking
coals and also due to the continued demand for better quality coke for the blast
furnace (Diez et al., 2002). However, to achieve the requirements of this blending
condition, the various kinds of coals are to be supplied from diverse region. Most
of coke plants are located near coal mines, while some of them are located near
sea shores, such as in Korea, Japan and Taiwan. Coke plants close to coal mines
do not have enough choice for their blending sources due to higher transfer fees
from long distant coal mines. On the other hand, other coke plants near sea shores
usually have long term contraction with coal mining companies for the stable coal
supply chain.

Additionally, the fundamental problem to meet requirements of coke qualities is

the characteristics of conventional coke oven operation. Conventional coke oven
cannot be cooled down and modified easily because coke oven was designed for
the 50 years of campaign life unlike blast furnace has 15 years of campaign life.
During this operation period, the relining repair of coke oven is almost impossible
to preserve the steady and stable operation temperature at about 1200 ºC, the
optimum temperature to prevent oven wall from deformation of silica bricks
which are the main materials of coke ovens. However, coke oven temperature is
intimately related to the working ratio whereas working ratio is variable, so that
maintaining steady working ratio is impracticable. Besides, conventional coke
oven is a cartridge composed of numerous coking chambers and the sequences
from coal charging to coke pushing occur everlastingly. Thus, the repairing of
coke ovens should be processed on battery scale, not on the individual slots which
appears to be also unrealistic.

Moreover, as we discussed before, coke making industry is directly related to the

environmental issues such as air pollution and global climate change. Currently,
U.S.A coke industries has pursed further advance in clean coke making system as

12
a non-recovery process of the coke making. The theory of the non-recovery oven
seems to be similar with beehive oven of the 19th century excluding reuse of waste
heat to produce steam and electricity (Bertling, 1999). This system has a benefit
of simplicity of oven design and omission of the by-product so that environment
friendly operation with none of a leakage of coke oven gas and volatile compound.
By the way, this system also has many drawbacks e.g. energy inefficiency,
limitation of coal use and low yield of coke. Large superheated waste gas has to
be supplied through long distance to a central boiler so that loss of heat energy is
substantial. Besides, low-volatile and well-coking coal can only be charged into
oven, whereas various kinds of coal series should be charged for the optimization
of coke strength. However, despite these many disadvantages, Inland Steel in
Chicago have operated non-recovery system due to the demand of ecological
reasons.

Figure 2.4: Non - recovery coke oven (Bertling, 1999)

13
2.2 Coal Characterization

2.2.1 Coal Composition

Coal is organic detrital sedimentary rock generated from a diversity of plant

material e.g. higher plants, ferns, fungi, algae, etc. and various tissues e.g. leaves,
trunks, bark, pollen, spore, etc. A coal deposit is based on the accumulated series
of decaying vegetation, deposition and burying by movements of the earth crust,
sedimentation, erosion and coalification (Krevelen, 1993). The conditions of
decay such as climate, acidity and water are extremely important factors to
determine the characteristics of coal together with the way of deposition and
burying by sediments. However, the most significant factor is the inwards
movement of the earth crust which called settlement and subsidence because the
depth is closely related to the temperature and pressure which are the key factors
in the process coalification.

Coalification is a geological process of the material formation with increasing

content of the element carbon from organic materials. The coalification process,
can be roughly divided into two parts, biochemical coalification, called diagenesis
which is related to coal type, and geochemical coalification called catagenesis
which is related to coal rank (Krevelen, 1993, Loison et al., 1989). The evolution
of coal started from degradation in cellular structures by the action of fungi and
bacteria. However, continuous deposition of new burial covers earlier layer which
generated from bacterial reaction, stops bacteria and fungi. This burial drives out
water and air and forms compressed gelified constituents. These series of stages
are called as diagenesis. The sinking of layers increases temperature and pressure
due to depth and alter the organic material into a coal.

Specifically, coalification can be divided by 5 stages: peatification, lignitification,

bituminisation, anthracitisation and graphitisation (Krevelen, 1993). Peatification
is composed of the bio-chemical reactions of aerobic bacteria, fungi and
actinomyes, which are the condensation and polymerization (Stach and Murchison,
1982). Humification, the most important process in peatification, forms humic

14
substance and aromaticity is increased. In lignitification, dehydration and
decarboxylation is the main chemical reactions, moisture contents are decreased,
and heating values are increased. Bituminisation from lignite to bituminous
include intensified decarboxylation and hydrogen disproportioning. In
bituminisation, reflectance, fluorescence, and extract yields increase to optimize
the condition of coking coals. Anthracitisation is the process of condensation to
small aromatic rings system, in which process reflectance still increases and
fluorescence, molecular weight of extraction and H/C ratio decrease. Throughout
the complete carbonification, graphitisation and H/C ratio also decrease, aromatic
rings are condensed and anisotropy increases. According to the coalification, the
colours of the coals are changed from brown to black, and the shapes are changed
from earthy to shiny.

Figure 2.5: Coalification Process

2.2.2 Coal Classification

Coal is a visibly heterogenous material but not a uniform mixture of carbon,

hydrogen, oxygen, sulphur, etc so that is highly complex in nature. The
heterogeneity of coal is closely connected to the process of coal formation and

15
diversity of source material. Coal classification system is existed by the
parameters that are scientifically valid and useful to the user. Basically coals can
be classified into 3 factors: grade, type and rank (Krevelen, 1993). Coal properties
can be distinguished in extrinsic and intrinsic. Extrinsic properties rely on the
influence of originality of the mineral deposition, in other word, coal impurity
which determines the grade of coal. Intrinsic properties are decided by the organic
matter which is related to the coalification, diagenesis and catagenesis. Diagenesis,
biochemical conversion and compaction are the standard to determine the coal
type by classification in the proportion of microscopically identifiable component
and in the position of coalification. For example, coal type seems to be explained
in terms of variety of petrographic analysis, lithotype and macerals (Suggate,
1998). Catagenesis, how severely geochemical conversion occurred, is connected
with coal rank.

Analysis of coal rank plays a significant role in its usage in industries such as
thermal power plants and steel industries. Coal is addressed on the basis of
relativity of microscopic constituents that affect on coal properties e.g. heat value,
strength, ash content, reactivity, agglomerating characteristics, etc. Rank is not, a
mature property of coal, petrography, measured on the basis of reflectance
generally of a particular organic component called vitrinite (Ghosh, 2002). Rank
was classified according to fixed carbon contents and Volatile Matter (VM) in
coal indicated by metamorphic development from peat to anthracite thus coal rank
is the age indicator of coal. Therefore, according to the coal rank, coal properties
can be varied. For instance, carbon/energy content increase, moisture/ VM
decrease by the coal age increase. Specifically, coal calories increase from 11.7
MJ/kg for wood to 35.2 MJ/kg for anthracite, VM decrease from 65 % dry ash
free base for wood to 2% for anthracite. Moisture decreases from 75 % for peat to
1% for bituminous through anthracite, carbon contents increases from 50% for
wood to 95% for anthracite as shown in Fig. 2.6 (Esterle, 2007).

16
 Figure 2.6: Estimators of coal rank (Esterle, 2007)

It is obvious fact that the main standard of coal rank in all countries is based on
the contents of volatile matter and fixed carbon contents. However, coal with
more than 30% of VM is hard to be classified on the basis of VM so that other
quantities should be included in the classification. A few countries, for instance
Germany and Netherlands have adopted the appearance of coke button as a
second standard to classify. United Kingdom has adopted on Gray-King assay
parameter as a second factor, France on swelling index, Italy on modified swelling
index, Poland on Roga test and U.S.A on calorific value. ASTM D-388
classification system, used in general, was devised in U.S.A as shown in Table 2.3
(Krevelen, 1993). The coal which has fewer than 14 % of VM dry mineral matter
free basis (dmmf) is classified into anthracite and coal has over 14 % of VM with
agglomerating properties is classified as bituminous. Besides, coal having less
than 69% of fixed carbon dmmf and more than 19.31 MJ/kg with non
agglomerating properties is categorized in sub-bituminous. Lignite has almost
same properties as sub-bituminous except the calorific value limit less than 19.31
MJ/kg.

17
Table 2.3: ASTM D-388: Classification of coal by rank (Krevelen, 1993)

However, the ASTM D-388 classification appears to have limitation to explain the
caking properties of coal which is essential for coke making. After World War II,
the United Nations established Economic Commission for Europe (ECE) for the
aid of economic recovery in Europe. Coal Committee is also founded as a sub-
division of ECE to put on end to the confusion prevailing in the field of coal
classification. This committee chose to add caking property index such as
swelling index and Roga index to reveal rheological properties of coal on rapid
heating. However, in addition to that caking properties on rapid heating of coal, it
was also essential to find criterion to reveal the coking property which determine
coke type. Thus, Gray-King assay index by dilatometer test was also added to the
classification of coal type. But, it still had taken long time to modify coal
classification model while ECE added caking and coking parameters to the
classification (Krevelen, 1993).

Consequently, Coal Committee revised a hierarchical system of coded parameters

represented by 14 digit code number. This system, using eight parameters to
classify the coal, is called as codification better than classification as shown in
Table 2.4 (Krevelen, 1993). This codification system is commercial, including
rank, type (petrographic composition), grade and environmental information but

18
only can be applied for medium and high rank coal. Thus, in the combination of
codification of the brown coal, the new ECE codification system was devised in
1988, while this system is still complicated and needing for decoding (Krevelen,
1993).

Table 2.4: International codification of higher rank coals (Krevelen, 1993)

2.2.3 Physical Properties

Physical properties of coal such as density, Shatter Strength (SS), agglomeration

property, microhardness and Hardgrove Grindability Index (HGI) play significant
roles in various industries which demand coal usage. For instance, coal density
can be used for the estimation of the reserve deposit of the coal mine. SS also can
be an indicator to expect percentage of micro coal fine and self-agglomeration of
coal appears to be a useful index to predict strength of agglomerated coal.

The coal density is a significant quantity that should be considered to examine the
structure of coal because all physical properties of coal depend to some extent on
its intermolecular distance based on density. The coal density is correlated with
the coal rank so that it decreases according to increase of carbon content initially

19
but mostly increases with carbon contents after reaching a minimum at about 82
wt % Carbon as shown in Fig. 2.9 (Huang et al., 1995). This tendency is based on
coalification characteristics. Firstly, coal loses relatively heavy oxygen although
hydrogen contents remains as constant so that the density initially drop off. When
nearly all oxygen given off, the lightest hydrogen starts to decrease and the coal
becomes comparatively richer in carbon which implies rapid increase of density.
This correlation is also applied in maceral groups with similar tendency except
slight difference in initial stage as shown in Fig. 2.9 (Krevelen, 1993). In case of
same coal in higher rank, inertinite shows the highest density, vitrinite middle and
exinite the lowest due to the difference of the relative oxygen contents.

Figure 2.7: Correlation of the coal density with carbon contents (Huang et al.,
1995) and Densities of coal macerals (Krevelen, 1993)

The shatter strength of the coal is a parameter to predict the amount of micro coal
fine when the coal has transferred from coal mine to the user. The SS is very
important to the profit of the mining company because the micro coal fines are
directly related to the coal loss. This is the weight percentage of coal above 25mm
size after the breakage of breeze coal (60~100mm) through gravity fall from 2m
height. Inertinite is the most friable maceral group so that it is concentrated on
fine particles. Vitrinite also shatters fairly easily broken under shock compare to
fine particles of exinite or inertinite cemented by vitrinite.

20
The agglomeration property is an index to indicate tensile strength of
agglomerated coal with out bituminous additive. Vitrinite and exinite are
viscoelastic solids so that permanent elastic deformation can occur in medium
rank coals. According to the Loison (1989), a coal crushed to 0.2 mm is
agglomerated by simple pressure in cold state. However, auto agglomeration to
about 0.1 mm would make trouble in a ball mill. Agglomeration also would
increase the strength of coke when coking coal charges to the chamber of the coke
oven so that agglomeration property appears to be a key factor in coking coal.

The microhardness is easy to be determined by the deformation of a polished

surface when an indenter is applied according to a defined operating procedure
and measuring the residual imprint of this indenter. The unit of microhardness can
be used as N/mm2 or Pa. A scale is used which gives values of 10 to 70 for coals
and can attain 200 for anthracites. Microhardness is very higher to a lower rank
coal, anthracite because of hydrogen bonds due to the abundance of hydroxyl
group as shown in Fig. 2.10. Thus microhardness is often used for classify
between anthracite and bituminous coal.

Figure 2.8: Microhardness index of coal (Krevelen, 1993)

21
The grindability of coal, indicating the level of crushing resistance, may be
determined by diverse methods, the best known of which is hardgorove
grindability test. This method’s merits are simple and reproducible so that many
countries have adopted this process for the grindability index. Hardgrove
developed in 1932 and ASTM accepted in 1951. In this method, a closely graded
(1.18–0.6 mm) air-dried sample of 120 g is prepared, from which 50 g is put into
the hardgrove machine and crushed under specified condition of 60 cycle and 3
minutes. The product is then screened on 74-µm screen and the weight of the
material passing through is determined. HGI is calculated from the relationship
(Sengupta, 2002):

HGI = 13 + 6.93W,
W = weights of - 74- µm materials respectively

Figure 2.11 shows the relationship between grindability and rank, work by
Dryden who found the HGI reaches a maximum at about 90% Carbon.

Figure 2.9: Hardgrove Grindability Index of coal (Krevelen, 1993)

22
2.2.4 Chemical Properties

Coal is a heterogenous organic compound both macroscopically and

microscopically. The macroscopically distinguishable components are classified
as lithotypes. The microscopically various constituents are classified as macerals
which originating from plant tissues. The chemical make up of coal is quite
complex and its analysis is also highly difficult. This is because a sample of coal,
even one taken from a specific site, is already a mixture of three groups of
macerals whose proportions diverse with the sampling point in the seam. The
other difficulty is the variety of macerals from one seam to another within the
same coal field. However, it has been available to suggest the two principles of
macerals of same appearance of the same rank and continuity of various
properties from low rank to high (Loison et al., 1989).

Moisture (air dry base) ISO 331:1983

Air-dried moisture, called inherent moisture is increasing according to the

decrease in coal rank. This analysis is started from the weighing of 1g of sample
then dried at the temperature about 105 to 110 ºC for an hour. After the cooling
stage, dried sample is weighed again. From the comparison of weight of coal
samples inherent moisture of coal is calculated. However, this small amount of
sample is extremely sensitive in ambient air so that repeated tests will increase the
reliability of values.

Volatile Matter ISO 562:1998

The volatile matter is the one of the most significant parameters used in
classification of coal ranks and its determination consists mainly in carbonizing a
given quantity of coal and measuring the resulting loss in weight. During the
process, the volatile matter given off composed of combustible gases, tarry
vapours and incombustible gases. One gram of coal sample is placed in a furnace
preheated to 900 ºC, and a residence time of five minutes with a final temperature
of the range from 895 to 905 ºC. The crucial point is that furnace must be covered

23
by well-fitting lid to avoid partial combustion of the coke residue. The weight loss
is calculated after the cooling stage of sample.

Ash ISO 1171: 1997

The ash yield is calculated by the weight of solid residue left after combustion of
the organic matter. 0.5 gram of coal weighed and charged into oven and heated at
the temperature about 805 to 825 ºC for an hour. Thereafter, ash content is
measured by remaining amount after heating. The ash is approximately 90% of
the mineral matter so that mineral matter present in coal undergoes changes of
which the followings in the ashing process (Loison et al., 1989).

1. Removal of carbon dioxide from carbonates.

2. Loss of Water of constitution from silicates.
3. Transformation of pyrite into ferric oxide.
4. Volatilization of chlorides of alkali metals.

These changes sensitively depend on the ashing conditions so that ash measuring
should be applied strictly by the standard method.

Fixed Carbon

The solid remains after the determination of the volatile matter is composed of the
mineral matter and the non-volatile organic matter termed as “fixed carbon (FC)”.
In the proximate analysis, FC value is calculated by subtracting the percentage of
inherent moisture, volatile matter and ash from the hundred percentages.

Elemental analysis

Analysis for elementary constituents of coal follows techniques similar to those

employed in organic chemistry. However, current automatic apparatus permits
simultaneous analysis with convenience. The elements of coals are analysed by
Vario MACRO elemental analyser. Coal specimens are combusted in a tube with

24
oxygen, and generated gases are derived and flushed by inert helium gas. CO2,
H2O and SO2 gases are separated individually by each adsorption columns and
weighed by thermo conductive detectors. N2 can be passed directly to thermo
conductive detectors for the first time due to its inert properties. Oxygen is
calculated by the substraction amount from C, H, N and S in coals.

Calorie

Calorific value is used in International classification of coal as same as volatile

matter. Parr 6300 Calorimeter is measuring the temperature rise due to complete
combustion of 1g of coal in an atmosphere of oxygen in a constant volume bomb
calorie meter with condensation of water formed. The calorie is considered as
more important in power plants than in coke plants.

Destructive analysis for the elementary constituents of coal such as carbon,

hydrogen and oxygen is essential technique, however, lacks of precision due to
the danger of oxidation. Thus, semi-micro analysis method without very fine
crushing, for example, the sample over 10 mg of size crushing, is widely used to
avoid oxidation. The carbon, hydrogen and oxygen contents are important index
of its rank. Generally, the carbon weight % increases with rank on the contrary to
the loss in weight % of oxygen and hydrogen as shown in Table 2.7. Sulphur in
coal is founded both as mineral and as organic groups attached to the hydrocarbon.
Chlorine concentration is often low but its determination is interesting for
combustion because the hydrochloric acid, corrosion gas is generated by high
incineration of coal. Phosphorus is also only a minor constituent of coal and often
estimated in the ash residue. However, its concentration is significant for
producing some electrometallurgical coke.

25
 Table 2.5: Parameters for coal molecule (Spiro and Kosky, 1982)

In addition to the elementary analysis of a coal, the functional group analysis can
be determined by the reaction with a chemical agent (Krevelen, 1993). Various
determinations of functional groups have been tried by the analytical methods
such as infra-red, ultraviolet and x-ray diffraction. Hydroxyl group (-OH)
percentage in coal is the function of rank. Brown coals may contain up to about 8%
of hydroxyl oxygen and this figures decreases to under the 1% of 90% carbon
contents, according to the increase of carbon contents. The hydroxyl group in
coals are mainly phenolic or at least acidic contrast to the no evidence for the
presence of alcoholic. Carboxyl groups (-COOH) are mostly determined by ion
exchange, by contacting the coal with excessive barium acetate and titrating with
free acetic acid formed. Carboxyl oxygen is existed at 8% in brown coal whereas
this are absent in over 85% of carbon of coal. Methoxyl groups (-OCH3) applied
equally to the carboxyl groups so that methyl oxygen exist in low carbon contents,
absence in 80% of carbon of the coal. Carbonyl oxygen (C=O) is found in all
product of the coalification series, decreasing tendency from an 11.6% in peat to a
1.9% in 90% of carbon of the hard coal.

In Fig. 2.12, containing more than 80% vitrinite on a dmmf basis also reflects the
metamorphic alteration, in other words, coals towards lower H/C and O/C
according to the increase of coal rank. This graphical representation provides a
number of advantages. Firstly, this graphical tendency explains of simple reaction
of the mechanism such as decarboxylation, dehydroxylation, dealkylation,
dehydrogenation and condensation. Coalification is explained by a series of

26
parallel or continuous reactions. Decarboxylation is the reaction of elimination of
carbon dioxide (CO2) from carboxyl group (-COOH). Dehydroxylation is also
elimination of water from hydroxyl group (-OH), dealkylation, removal of methyl
groups (-CH3) mostly in the form of methane (CH4), dehydrogenation, elimination
of hydrogen from naphthenic groups (-CH2-) with transformation of
hydroaromatic rings. Finally, condensation can be expressed as cyclic system with
removal of hydrogen and formation of direct C-C bonds. Besides, the graph shows
to be formed of the structure of the carbon skeleton. The oxygen generally
included in hydroxyl groups, ether groups and heterocyclic oxygen compounds so
that this explains why the H/C ratio is nearly equal to that of hydro carbon with an
identical carbon skeleton which seems to be formed from the elimination of the
oxygen atoms (Krevelen, 1993).

Figure 2.10: Correlation of atomic H/C and O/C for coals of different ranks
(Neavel Richard, 1981)

In addition to the functional group analysis, a coal model can represent the
properties and behaviours of coal such as liquefaction, swelling, coking,
gasification, pyrolysis, etc. Shinn (1984) suggested coal model of Illinois 6 seam
coal, the distribution of atoms in the coal model of C661H561O74N11S6 as shown in
Fig. 2.13. This model indicates about 70% of aromatic carbons, 30% carbons
chiefly combined in hydrogenated rings, mainly alkyl chains. This condensed

27
polycyclic system forms a macro molecule by C-C bonds of which most frequent
are methylene (-CH2-) and ether (-O-). This model also explains the reactivity of
various cross links of carbon under plastic phase according the carbon chain.
Besides, hydroxyl group can have a strong activating effect on theses cross links
so that hydroxyls appear to be important significant as bond activators, potential
condensation site and source of interaction of swelling.

Figure 2.11: Model of bituminous coal structure (Shinn, 1984)

Besides carbonization, the other chemical reactions such as oxidation and

hydrogenation at low temperature are also important part of the chemical
properties of a coal. The oxidation reaction begins to occur even in room
temperature but is very slow as low as 10-12 to 10-15 cm2/sec and for this reason is
difficult to study (Loison et al., 1989). The oxidation rate will be accelerated in
accordance with temperature rise. The coals with richer oxygen can be oxidized
easily so that reaction is faster, the lower the coal rank. Hydrogenation is
generally processed to industrially obtain hydro carbons in liquid phase. The
methods of coal hydrogenation can be generally divided into two categories,
under mild and destructive conditions (Krevelen, 1993). The coal hydrogenation
under mild conditions processed under the temperature of 350℃, the lowest

28
softening temperature of coal. The small chemical destruction of coal can be
occurred in this condition so that originality of carbon skeleton of coal remains.
Thus, this reaction provides structural information of coal so that coal
hydrogenation under mild condition often called as analytical hydrogenation. The
coal hydrogenation under destructive condition, the temperature over 400℃, is
combined with thermal cracking. The purpose of this hydrogenation is only for
the transformation from solid coal into useful liquid fuels.

2.2.5 Coal Petrography

The objectives of application of the petrographical analysis are numerous such as

understanding of coal deposition by geologist, evaluation of coking potential of
coal by technician, monitoring coal quality and detection of contamination factor
in coal by mining company. Generally, coal petrography is defined as a standard
to classify a coal to organic and inorganic components, in other words a
proportion of the maceral composition and a rank defined by vitrinite reflectance
(O'Brien et al., 2003).

In coal seams, petrographers have distinguished coal seam into four defined
classes called lithotype in a Vitrain, Clarain, Durain and Fusain by the brightness
of the layer (Loison et al., 1989). According to the examination of the optical
microscope, these four classes also match Vitrinite with Vitrain, Exinite with
Clarain, Inertinite with Durain and Fusain which has certain homogeneity called
as maceral groups as shown in Table 2.5. In coal utilization process it is often
sufficient to know the relative ratios of these three groups. Furthermore,
petrographers have divided these maceral groups into ‘sub-maceral’ such as
telovitrinite and gelovitrinite between which there are nuances of appearance and
coking capacity. Besides, coal properties depend not only on maceral composition
but also on their distribution so that microlithotype index is also used to indicate
dispersing types of the macerals as shown in Table 2.5.

29
 Table 2.6: Petrographic nomenclature (Loison et al., 1989)

The three maceral groups, vitrinite (huminite in low rank coals and lignites),
exinite and inertinite, have the suffix ‘inite’, and include series of macerals which
can be regarded as belonging together, either due to the similar origin (exinite) or
due to the pattern of conservation (vitrinite and inertinite). They show different
reflectance colour and diverse structures of each group as shown in Fig. 2.7 (Stach
and Murchison, 1982). These macerals of high volatile coal show three grey levels
owing to the spread in their reflectance within the same sample: vitrinite appears
medium-grey, exinite black and inertinite white. These three groups are to a
certain degree characterized by their chemical constituents. In case of the coal
groups of same rank, the vitrinite relatively contains more oxygen, the exinite
more hydrogen and inertinite more carbon. For instance, in maceral groups which
has 84% of carbon in vitrinite, the hydrogen contents would be 7% in the exinite,
5.5% in vitrinite and 3.9% in the inertinite. The volatile matter is also highest in
the exinite of 66.7% against the 35.2% of vitrinite and 22.9% of inertinite. These
chemical characteristics are the significant factors to determine the characteristics
in bituminous coal.

30
 Figure 2.12: Three maceral groups (Stach and Murchison, 1982)

Vitrinite group which is generated through humification from lignocellulose

tissues e.g. woody tissues, roots, bark, leaves etc. appears to be a wall which
surrounds the other macerals and mineral matter, also is the most abundant,
isolated and homogenous maceral of coal, red-brown coloured in thin section.
Vitrinite constituent is the majority of normal coals, so that its thermodynamic
behaviour is very similar to that of the whole coal (Loison et al., 1989). Vitrinite
oxygen contents are higher than other macerals whereas hydrogen and carbon
contents are intermediate (Davis, 1976). Therefore, due to high hydrogen ratio,
vitrinite can be an indicator of temperature when coal start to metamorphose by
heat.

Besides, other macerals are also important to investigate relationship between coal
petrography and coal carbonization. Exinite group, generated through resistance
of pressure from algae, cuticles, resins, etc., yellow coloured in thin section shows
greater plasticity during carbonization stage than vitrinite group whereas exinites
are a minor component of bituminous coal in general. For the same rank
bituminous coal, exinite shows highest hydrogen, volatile matter and aliphatic
carbon chain (Stach and Murchison, 1982). Inertinite group, generated through
degradation of product of the burned or biochemically oxidized plant, opaque

31
coloured in thin section, is also closely related to carbonization process by
containing higher aromatic carbons which will compose basic structure of coke
when it was carbonized. Excluding few exceptions of Australian bituminous coals,
inertinite is mostly regarded as inert material when it was carbonized (Diessel,
1983).

Additionally, the reflectance of the maceral is highly used to measure their coal
ranks. The macerals, microscopically represented by reflected light from about 5µ
of polished coal, has shown a series of reflectance distribution as shown in Fig.
2.8. These distributions are the evidences of the continuous coalification in stages
so that no jump from rank to higher rank (Neavel Richard, 1981). In high volatile
bituminous distribution, the relatively hydrogen-rich constitute liptinite has lowest
reflectance, the relatively oxygen-rich vitrinite has a medium reflectance and
relatively carbon-rich inertinite has the highest reflectance. However, in the stage
of low volatile bituminous, the vitrinite reflectance is surpassed by liptinite.

Figure 2.13: Maceral reflectance vs. Coal rank (Neavel Richard, 1981)

However, typical coking coals are relatively vitrinite-rich so that analyses of

whole coals when appropriately corrected for inorganic content can’t exclude

32
properties of the vitrinite. As a matter of fact, the most important implication of
coal rank should be done with samples of concentrated vitrinite or on samples
where the vitrinite comprises more than about 80% of the organic fraction. This is
because reflectance is intimately correlated with rank- properties and its
determination has also become a widely accepted parameter to designate the rank
of a coal.

Thus, vitrinite reflectance, displaying optical anistropy, seems to be one of the

significant factors to correlate with coke strength because the plastic phase of
coking coal is very significant stage to determine on the quality of coke. Vitrinite
reflectance is also intimately connected with aromaticity of coal chemistry which
is also related to the coal rank. The determination of the optimum ratio and
distribution of vitrinite components in blending coal is a key factor to establish
maximum strength of coke, while optimum ratio of exinite and inert still remains
as another key factor.

However, the petrography of individual macerals is still not well understood due
to the complexity of maceral-specific analysis (Walker, 2005). This complexity is
based on the characteristics of maceral measurement which are time-consuming,
subjective and needing high level skill. Moreover, different standards used in
different countries to characterize macerals, e.g. Australian Standards 2856.2 and
International Committee for Coal and Organic Petrology as shown in Table 2.6, also
decrease the understanding of maceral measurement. Especially, multivariate
vitrinite classifications such as desmocollinite and collodetrinite are confusing the
relationship of vitrinite and prediction of the coal quality. Besides, the effect of
the individual macerals, e.g. collinite, telinite, etc. on properties still remains
unclear in the point of their proportions and distribution. Progress in coal science
can only be established when technological and scientific investigations on coal
are done in a synthesis of data and comprehensive integration.

33
Table 2.7: Different Maceral Standards (Gupta and Shen, 2009)

2.2.6 Inorganic Matter

Coal is a heterogenous mixture of organic and inorganic matter. Mostly the

benefits from coal such as calorific value, metallurgical use and hydrocarbon
source are based on the organic matter, maceral constituents. In contrast,
numerous unwanted problems such as abrasion, corrosion, stickiness and
pollution are caused by inorganic constituents broadly referred to mineral matter
(MM). Such MM is not only predominant constituents of lower rank coals such as
brown coals, lignite and sub-bituminous coal but also key factor of ash formation
in coal. The mineral constituents can be divided into two categories, inherent and
adventitious (Loison et al., 1989). The inherent mineral matter, originating from
the vegetable from coalification, chemically bounded to the organic matter and its
concentration is always very low, usually under 1%. The adventitious MM has
multiple origins of inclusions in the fissures crossing the seam and layers of rock
laid down at the same time as the coal. The classification of two categories is
vague because MM deposited at the same time as the plant may have combined
with the organic matter during coalification and metamorphism.

34
Various experimental methods, including thermal analysis, electron microscopy,
selective leaching processes and x-ray diffraction are used to evaluate the mineral
composition due to the complexity of coal (Gupta, 2007). Figure 2.14 shows the
variety of mineral in coal where black colour indicating coal, whereas other
colours correspond to different minerals. The minerals of coal roughly divided
into groups of aluminosilicates, carbonates, sulphur, oxides, hydroxides,
phosphate and salt. Aluminosilicates, which more than 75% of minerals and
divided into clay minerals and silicates, are composed of microscopic grains,
flakes, thin fibres, etc. The most dominant compound of aluminosilicates is
represented as Illite and kalonite group. Carbonates appear to be form of crystals,
grains and nodules isolated in layers. Calcite, ankerite and siderite are the
common compounds of carbonates. Compounds of sulphur which are rare can be
divided into sulphide (pyrite, marcasite) and sulphate (baryte). Oxides and
hydroxides groups are also very rare but only most common in quartz. Phosphate
contents are important in a production of coke of an electro-metallurgical purpose
because phosphorus has to be low to use in this process. In most coals, salts are
very rare except some coal field of British and Germany, their contents is so high
that salt coal is used. Generally, high contents of salt are the reason of trouble in
coal preparation due to the excessive corrosiveness of water.

Figure 2.14: SEM from QEMSCAN of a coal sample; 63-90 µm (Gupta, 2007)

35
Knowledge of the composition of the MM is of interest for several reasons (Ward,
2002). It determines the ash fusion temperature and indicates the limitations of
coal use. It also an important key factor to understand the coal formation and the
classification in geology. Identification of the mode of occurrence and potential
mobility of mineral elements, understanding potential barriers to gas drainage, and
evaluating the behaviour of different coals are the parts of practical applications of
coal geology.

Recent advances in studies of MM have provided a more definite standard to

evaluate the mineral composition. Due to the significance of MM in different
types of coal use, such study has to be fully applied into advanced coal
characterization as well as coal petrology and chemistry. Besides, in industries of
a steelmaking, mineral constituent has risen as a key word for the cost-effective
operation with decrease of energy by developing high reactivity of coke.

2.3 Carbonization Phenomena

2.3.1 Meta-plast formation

Carbonization which often used in organic chemistry is the conversion of an

organic substance into carbon or a carbon-including residue through pyrolysis or
distillation. This reaction can be addressed as the generation of coal gas or tar or
coke from raw coal. Chemically, carbonization is a polymerization and
aromatization process as shown in Fig. 2. 15 so that small aromatic compounds
are polymerized to an aromatic polymer in which consequently compose the three
dimensional of graphite (Lewis, 1982). It is considerable to explain a
carbonization as parallel and consecutive of individual processes which represent
major reactions involving the pyrolysis of aromatic hydro carbons whereas overall
process of carbonization still exceedingly complicated.

1) C-C, C-H bond decomposition to form free radicals of reactivity.

36
 2) Molecular rearrangement
3) Thermal polymerization
4) Aromatic condensation
5) Elimination of side chains, H2-

Figure 2.15: Scheme of general reaction for carbonization (Lewis, 1982)

Firstly, cleavage of cross links can be addressed as loss of functional group. For
instance, carbon dioxides (CO2) is extracted from carboxyl group (-COOH),
moisture (H2O) from hydroxyl group (-OH). The moisture is released between
100 and 150℃. Between 200 and 350℃ bituminous coals change hardly at all in
weight, but lignites are continuously lose weight. Besides, the cracking reaction
which represented as decomposition of C-C bonds produces compounds that are
less polymerized than the original coal of which a large proportion will be liquid
at the high temperature.

The molecular rearrangement is a very significant step in the early stage of

carbonization because this thermal reaction can transform chemical of structure
from unstable to stable without loss of carbon atoms. In this process that often
makes it difficult to relate the original structure to the structure of consecutive
course of carbonization. The thermal polymerization, in other words
aromatization, composed of the formation of extensive aromatic groups by
dehydrogenation of saturated rings of hydroaromatic groups. Elimination of side

37
chains is the competitive reaction of the aromatization so that these reactions often
occur at the same time.

The aromatic condensation consist in the formation of ever more aromatic groups,
carbon crystallite (coke, char) by recombination of aromatic groups with one
another. The standard example of condensation reaction is in fact

R-OH + R’H → R-R’ + H2O (Loison et al., 1989)

with formation of a molecule R-R’ which is too large to volatilize. These types of
reactions, cracking, aromatization and condensation can well cooperative when
cracking needs condensation for to be fed with hydrogen and, conversely,
condensation will always liberate H2 or H2O or even some larger groups which
can’t enter into the condensed aromatic systems. The first type of reaction
primarily forms tars and the second type transforms these tars into solid residue.
The function of hydrogen is available to combine with oxygen and the carbon in
these two reactions.

Figure 2.16: Mechanism of coal aromatization (left) and condensation (right)

In the view of pore structure development, the microscopic examination of the

stages from carbonization of blackhall coal is graphically illustrated in Fig. 2.17.
This graphical data show that the mean-pore and pore wall sizes and the number
of pores of centimetres being plotted by the temperature diversification. At the
temperature of 360℃, slightly below the softening point, the initial pores
appeared in large particles. Small, but measurable pores speed of weight loss is

38
also processed in this temperature zone but no fluidity was observed. According
to the temperature increase, pore size and number is also increased with the
occurrence of inter particulate void by swelling of the particle and rounding of the
circumference of particle. The swelling of the larger pores are sweeping the
smaller ones and concentrated into the void space so that partial fusion occurred at
the temperature about 390℃. Continuously, small pores are engulfed within the
expanding larger particles as the stage of complete fusion at the temperature about
410℃. After the complete fusion stage, pore and pore wall sizes increased to the
maximum value, the result being occurred to a highly porous region at the
temperature about 440℃. However, as the increase of the temperature, the mean-
pore and pore-wall sizes fell and the number of pores increased. These are
resulted in the compaction zone of the structure at the resolidification temperature
about 470℃. According to the further temperature increase, only small variation
of pore of size is occurred up to the temperature of 600℃ of semi coke stage.

Figure 2.17: Formation and development of the pore structure during

carbonization of coal (Hays et al., 1976)

39
However, the understanding of chemistry of a carbonization still remains complex
due to the innumerable of competitive chemical reactions and its complexity itself.
To measure the rate of carbonization reaction, mass transfer phenomena such as
the traverse of the tars formed into the volatile matter is difficult without
considering the relationship between weight loss and fluidity. In the carbonization
of coal at a low temperature, the definition of final temperature is vague so that
500℃ is tentatively approximated, but not accurate. Therefore, further advance in
technology of carbonization research will make help to approach a complete
understanding of carbonization.

Recently, the phenomena of the plastic state have been a great significance with
the aspect to coke making process due to the lacks of premium metallurgical coal.
The majority of coals become coke through plastic state when heated to around
400℃ in the absence of air, specifically oxygen. Other coals do note melt
considerably and are called infusible. Coals are softened, melted and swelled with
evaporating of Volatile Matter (VM) in the plastic state. According to the Kim
(2004), the viscosity of the melting phases is so high, at least 10000 poise so that
these phases cannot be recognized as the normal liquid. In these phases, fluidity is
induced by breaking the coal bonds and forming a plastic state where nucleation
occurs, volatiles evaporate and flow as bubble form, diffuse, and rupture in a
complex combination of reactions that lead to the transforming structural
evolution of the heated coal particle. Therefore, these plastic sates are considered
as significant for coke qualities. In order to understand the phenomena of plastic
state of a coal, it is essential to understand the properties which govern this
behaviour. Theories and extrinsic factors play an important part in the plasticity
of coal. Therefore, review of these theories and factors will be presented to
determine fluidity of characteristics and correlation with extrinsic parameter.

Theories of Plastic State

Physical Theories:
a) Homogenous Melt:

40
The homogenous melting theory proposes that coal undergoes initial softening
which is similar to the homogenous substance. This softening, which is stressed
above the temperature of 350℃, is a purely physical phenomena unconnected
with thermal decomposition (Loison et al., 1989). The initial softening
temperature relies little on chemical constituent in the range from 20% to 38% of
volatile matter. This temperature also depends little on heating rate, although the
phenomena intimately related to the pyrolysis. Concurrently, the melting of a coal
is continued to the decomposing state, producing gases and infusible solids with
resolidification. The homogenous melting theory of coal defines this fluid
behaviour of coal as the difference between the softening and decomposition point
in the assumption of a coal as a homogenous substance (Read, 1982). However,
this assumption has some limitations itself. Firstly, certain coking coals fails to
soften under the condition of slow heating or vacuum. Secondly, some coals show
softening under the condition of constant heating and time increase.

b) Partial Melt:

The partial melting theory suggests that only a fraction of coal melts on heating
(Read, 1982). This fraction of coal lubricates and melts the remaining portion of
coal. This fusible fraction is often referred as “bitumen” such as benzene, pyridine
and chloroform and is in general found in extracts of coals and considered to be
responsible for caking phenomenon. Coal under extraction often shows that coal
residues show no plastic properties whereas the extracts are extremely plastic.
Takanohashi (2005) suggested the coal model which is composed of Chloroform
Soluble (CS), Chloroform Insoluble Pyridine Soluble (CIPS), Pyridine Insoluble
CS2/NMP Soluble (PIMS), and CS2/NMP Insoluble (MI) with different degrees of
heaviness as shown in Fig. 2.18. This bitumen forms an associated structure
through variety of coexistence and lighter component act as solvent. The
increased mobility of the lighter constituent, CS gradually enhances the mobility
of the heavier mobility from CIPS to MI.

41
 Figure 2.18: Proposed mechanism for the occurrence of coal softening
(Takanohashi et al., 2005)

Micellular Partial Melt Theory:

According to the micellular theory, coal can be presented as micellular type of

structure such as “cogel” in which the micelles and other particles had a similar
colloidal type of dispersion or “isogel” which assemblage of large and small
molecules having similar composition. All these molecules are combined by weak
van der waals forces and this bonding power easily reduced when it was heated so
that the smaller molecules become mobile. Spiro (1982) has suggested the space-
filling models of coal structure as the mechanism for thermal plasticity and
decomposition. Based on this theory, coal is composed of planar structures with
occasional protrusions of small aliphatic rich regions. In the process of pyrolysis,
protrusions of aliphatic chains cleaves the weakest bonds, the remaining planar
aromatic structures were locked into place by the aliphatic chains, begin sliding
over one another in two dimensions so that the sliding due to lubrication
responsible for the thermoplastic phenomena.

Metaplast Theory:

The Metaplast theory was modelled of thermoplastic behaviour of coal by

Fitzgerald (1956). “Metaplast,” in other words “thermo-bitumen”, is the product

42
of heating, unstable intermediate product and not pre-exist in coal. The metaplast
then decomposed by cracking reactions into primary gases and solid-residues
which forms semi-coke.

Coal → Metaplast → Primary Gas +Semicoke → Secondary Gas + Coke

This theory satisfies a hypothesis that the metaplast is not formed in low rank
coals due to the relative thermal instability of metaplast. Besides, rapid removal of
metaplast under vacuum condition destroys the plasticity. It also explains the
reason that fluidity increases under isothermal heating of condition. However, this
doesn’t have explained that the extracted residue from the coking coals fail to
develop plastic behaviour.

Physico-Chemical Theory:

This theory models the coal as the highly polymerized macromolecules

interconnected with hydrogen bonding with weak Van deer Waals forces (Read,
1982). The number of hydrogens which mostly in the ethylene (C2H4) bridges
determines the number of breakable bridges (Solomon et al., 1992). These
aliphatic bridges are generally richer in bituminous coal than other ranks of coals
so that bituminous coal has high fluidity. Especially, low rank coals are appears to
be extensively cross-linked with functional groups whereas bituminous coal has
less extensive cross-linked density and upon heating a greater degree of mobility
is attainable. This is because bituminous coals are used for coking in general.

Internal hydro-liquefaction Theory:

The liquefaction behaviour of coal can be understood as donation of internal

hydrogenation so that the development of the coal plasticity is a process of the
hydrogen donation (Read, 1982). The bitumen is acting as hydrogen donor and
lubricant with the increasing of the temperature. The continual increase in fluidity
can be sustained only as long as those free radicals form abstract hydrogen. Once

43
lack of hydrogen donor has occurred the free radicals attacks the neighbour
molecule and forms a semi-coke.

Extrinsic factors of Plastic State

Coal thermo-plasticity depends on the extrinsic parameters e.g. heating rate,

pressure, particle size, oxidation, etc. and occurs with a complex combination of
phenomena. Generally, the heating rate of a coal ultimately effects on its plastic
properties. There is an increase in maximum fluidity, plastic zone, maximum fluid
temperature and resolidification point without softening point when it increases
heating rates (Loison et al., 1989). Conversely, a coal moderately fusible at a
normal heating rate and become infusible at very low heating rate. The extension
of the plastic zone with higher temperature is readily explained by the following
reasons. Firstly, homologous state of coal can be occurred at a temperature which
is higher, the greater the heating rate. Other secondary reason is that tar can’t
escape if the particles are large and heating is rapid. Finally, in low-rank coals rich
in oxygen, the high heating rate enables a degree of plastic fusion before pyrolysis
reaction whereas normal heating rate prevents any possibility of fusion.

A coal is slightly fusible at atmospheric temperature and low pressure

considerably reduces the plasticity and coking properties. Macroscopically, the
effect of pressure can be explained as reduced volatilized of tar and increase in the
amount of metaplast. Microscopically, when the pressure is increased, it aids in
closure of the surface pores and thus restricts the volatile release, promoting
formation and coalescence of the bubbles (Strezov et al., 2005). Volatile
compounds are addressed with a high resistance to release and only can be
liberated from the particle when the inter particle pressure is as much high as
enough to compete with these forces under higher pressure. The pressure of coal
during pyrolysis may be the weight of particles themselves so that they will be
weaker in laboratory test which uses small amount of coal whereas the pressure in
the real operation will be higher when the height of coal above the particles
considered is greater. Besides, Read (1985) found that the lowest rank coal has the
greatest increase in their respective fluidities when subjected to pressure under

44
pressured condition of helium gas. The constituents in low rank coals such as total
sulphur, pyritic and organic sulphur, nitrogen contents and reactive/mineral matter
ratios are may be greatly affected by the condition of pressure.

How the plasticity of coal is influenced by the particle size is a contradictory to

understand the effect of plasticity of coking coal on the coke properties due to the
cases of variety (Read, 1982). The smaller the dimensions of the particles in
lorrain coking coal during pyrolysis, the more easily the tar escapes by diffusion
without intra-granular pores and the lower is the plasticity. In contrast, within the
larger particles, bubbles of spheroidal form arise and grow without easy exit due
to the formation of gases at the centre of material thus the plasticity of coal won’t
be decreased. However, this effect is relatively unimportant with bright coals,
because the fusion itself removes the initial state of size distribution.

Coals richer in oxygen are, at the same rank, less fusible and the oxygen
introduced by oxidation in air appears to be active than that natural state. When a
coal begins to oxidize, the softening temperature increase and the resolidification
temperature decrease. The maximum plasticity decreases rapidly and the plastic
zone narrows at first and then more slowly with increased time of exposure to air.
For instance, the 0.1% of weight loss of oxygen is enough to reduce the plasticity
and within the 2% of loss of oxygen will eliminate the coking property of
bituminous coal. Part of oxygen that has reacted with the coal is released in the
form of CO, CO2 and H2O during oxidation of coal. Some coals are oxidized
easily even at room temperature when a coal are stored at fine sizes, other coals
can be stored for month without change due to the oxidation. In order to prevent
the oxidation with maintaining the maximum fluidity of coal, the freezing the coal
fines or storing the coal at the temperature below -10℃ is required.

2.3.2 Resolidification and Semicoke

A coal can’t be maintained for an extended time in its plastic state so that the
plastic state of coals appears to be made into coke is maintained from around at

45
the temperature from 350 to 500℃ for almost an hour. After an hour, the
plasticity of coal ceased and the coal irreversibly resolidifies into coke. The
resolidification temperature is better defined than softening temperature because
the fluidity decreases more rapidly than appears. The coke properties dominantly
depend on the resolidification point because the mechanical strength causing
fissures to coke during further carbonization depend on what occurs around
resolidification stage. The texture of the coke by the gas bubbles is established in
the last part of the plastic layer, fixed by resolidification and remains as identical
at the moment.

The metaplast of coal is progressively transformed by pyrolysis into coke and

volatile matter (VM), so it is natural that the plastic state has to end due to the
absence of coal. In the case of bituminous coals, the resolidification state shows
rough tendency of disappearance of soluble constituents whereas a small amount
of metaplast still remains (Loison et al., 1989). At the moment at which fluidity
reaches peak point, around 450℃, the segregation of another liquid phase
insoluble in the plastic state occurs. As it were, the new anisotropic phase formed
in the initial plastic state which is perfectly isotropic. This anisotropic phase
appears in the form of spherical globules which gradually grow and coalescence
when they come to contact with one another. Consequently the isotropic phase
wholly substituted with anisotropic phase and resolidification then occurs by an
increase of the viscosity of the anisotropic phase as shown in Fig. 2.19.

After resolidification, a solid residue is obtained often called semicoke as an

intermediate of the coke before the temperature reaches the region of 1000℃.
Kidena (1996) studies the semicoke formation by the pyrolysis of the three kinds
of coal as shown in Fig. 2.20. Yields of semicoke are 73-86 % at resolidification
temperature, 37-61% at 1000 °C, and 72-81% at 764 °C. The size of the particles
of the semicoke developed and larger particles could be seen than the case of raw
coal. The PM coal (B) shows the clear case of ordered structure of semicoke
formation. Especially, the unit structure of PM coal was fairly developed at the
temperature of 1000℃.

46
Compared with the plastic phase at the temperature range from 350℃ to 500℃,
the coking conditions are less important above 1000℃ because the most of the
properties of the coke are determined by the state at 500℃. However, the
thermodynamic behaviour of semicoke is still important to evaluate the optical
properties on coke. The semicoke form a layer on the contact material and the
thickness is getting grows according to the increase in temperature from 500℃ to
1000℃ (Brown et al., 1966). According to the increase in the temperature, the
semicoke will have increased maximum reflectance and greater optical anistropy.
Besides, on heating between 500 and 800℃, semicoke evolves VM and their
gasification leaves the voids in the heart of the solid and those reactions develop
the surface of coke with the numerous of the pores. The release of VM instantly
causes the contraction of semicoke.

Figure 2.19: Development of anisotropic spherules in the resolidification region.

Heating Rate 0.5℃/min. Temperature: (a) 450℃; (b) 460℃; (c) 470℃; (d) 510℃,
(a)-(c), x 500; (d), x 200 (Loison et al., 1989)

47
Figure 2.20: SEM photographs for the heat-treated (A) Lusca, (B) Pittstone-M,
and (C) Witbank coals (Kidena et al., 1996)

48
2.3.3 Carbonization in Industrial Oven

Coke making is the consecutive carbonization process of heat transmission within

the coke oven chamber so that the blends of bituminous coal are transformed to
cokes. Coke oven chamber is a parallel piped a longitude of 12-18 m, a height of
4-8 m and width of 400-600 mm. And then heat is applied into the two facing
walls at isothermal temperature. The industrial operation is much more complex
than the simple experiment in laboratory so coals go through a number of stages
over a period of 15 to 25 hours as the temperature rises to around 1100 ºC. Heat in
transmitted not simply by conduction so that the migration of water and tar can
play an important role.

The mechanism of carbonization in coke oven is well displayed in Fig. 2.21.

When a coal particle reaches a temperature of about 100℃, its moisture, CO and
CO2 evaporate and the water condenses on the interior colder particles with slight
increase in pressure. As the temperature increase to 350 ºC, the charged coal starts
to expand, and then after that temperature the outer layers of the coal particles in
the coking chambers start to establish plastic layers characterized as envelop form
of a metaplast. The plastic layer is often regarded as buffer zone due to the
relative high resistance to the passage of gases. The VM evaporate and the plastic
layers move inwards leaving the semi cokes behind according to the increase in
temperature. The coke oven gases such as CH4, H2 and CO evaporate until the end
of the carbonization at the temperature about 1000 ºC. The most of gases around
77-90% escapes at the hot side and passes through the plastic layer and travels
along the oven wall. According to the inward movement of plastic layer, the semi
cokes shrink and real cokes are established. When the two plastic layers meet in
the middle of coking chambers, the carbonization in coke oven ends with the
maximum pressure towards walls and last large amount of evaporating gases. As
the temperature increases towards 1100 ºC hydrogen gases are evolved, the semi
cokes usually shrink, and true cokes are established.

49
 Figure 2.21: The mechanism of carbonization in coke oven

2.4 Factors affecting Coke Properties

2.4.1 Coke Quality Parameters

The Coke, as a fuel, chemical reducing agent and permeable support, is utilised to
maintain the iron making process in blast furnaces by not only to provide heat for
reaction and gases for the reduction of iron oxides but also to support the iron ore
zone with permeability. As a fuel, reducing agent and permeable support, a
calorific value of coke, an adequate reactivity to CO2 and minimal breakdown
when the coke pass through blast furnaces are should be maximized.

The reactivity of the coke is the rate of reduction with CO2 in blast furnaces. In
blast furnace operation, cokes should be provided continuously from the top to the
bottom in blast furnaces with reduction. Coke should maintain a certain extent of
coke strength after reaction (CSR) as the range from 60 to 67. If the CSR is too
low, cokes are easily burned off in the upper part and the degradation of coke
besides the tuyere will be accelerated so the pressure of tuyere can be changed as
unstable. If the CSR is too high, the reduction rate of iron ore will be decreased
and the remained coke in the bottom will increase the viscosity of melted iron and
slag. Therefore, out of the range of CSR will be connected to the decrease in the

50
production amount of the iron ore, the increase in the cost of fuel and the
fluctuation of the quality of iron ore. A number of factors, e.g. tensile strength,
catalytic effect, specific surface area and graphitization appear to influence the
CSR.

Besides, cokes should suffer minimal breakdown as they pass through blast
furnaces for the permeable support. In other words, the coke needs to have enough
strength to sustain the burden pressures of the continuous layers of the iron ore
and lime stone in the blast furnaces. Specifically, coke has to have well distributed
in size for the permeability of counter current of gas flow and drainage at the
lower part of the blast furnace. In addition to that, the physical strength against the
thermal shock, the abrasion and the fragmentation is also essential condition of
the appropriate coke for the optimum operation of the blast furnace.

With the operational trend of increase in the inner volume of the blast furnace, the
physical qualities of coke especially fragmentation and abrasion resistance need to
be improved. Big size of coke will provide good permeability to the layers in blast
furnace and strong coke will make a lesser fine particle which decreases the
permeability of cohesive and melting zone. Thus, coke needs to have enough
physical strength to sustain the burden pressures of the continuous layers of the
iron ore and lime stone in the blast furnaces. Coke macroscopically resembles
fissured crystallised foam as shown in Fig. 2.22 (Ragan and Marsh, 1980). It has
pore size and shape with heterogeneity but continuous micro-crack propagation
will be generated by gasification when the coke pass through blast furnace as
illustrated in this picture. Thus, coke should suffer minimal breakage against the
abrasion and the fragmentation for the essential condition of the appropriate coke
for the optimum operation in blast furnace. The physical strength of coke appears
to be classified into resistance of fragmentation and abrasion. Fragmentation
stands for breakage from lump to small particles. Abrasion is wearing away from
surfaces to breeze particles by scarping or rubbing. In case of the breakage of coke,
fragmentation is affected by big cracks and abrasion is affected by small cracks.
To support the upper cohesive zone of Blast Furnace, the resistance against the
abrasion is more import than fragmentation.

51
 Figure 2.22: Drawing of the generation of fissures found in coke pieces (Ragan
and Marsh, 1980)

Therefore, the test of physical strength of coke such as ASTM (Tumbler), ISO
(Micum) and JIS (Drum Index) is considered to be a main indicator to evaluate
the supporting power of cohesive zone to make a lesser fine particle which
decreases the permeability of cohesive and melting zone in Blast Furnace. In the
estimation of resistance of fragmentation, the big size screen of test is appropriate
whereas the small size screen is well matched with the estimation of resistance
against abrasion. Therefore, in the American and European Standards for coke
quality in cold states, 40 mm sieve is used to measure the fragmentation resistance
(M40), and 10 mm sieve is used to measure the abrasion resistance (M10).
However, Japanese Standards (JIS) Drum Index (DI) in which the sieve size is 15
mm is mainly used in POSCO to emphasize the abrasion resistance as a main
factor of the operation of blast furnace.

Petrographic analysis of coke, specifically the optical coke textures can also
explain the coke qualities. Preparation process of coke sample for microscopic
analysis is the similar to coal’s one so that 25g sample of crushed coke under 12
to 60 meshed is moulded with hardening agent and specimens are prepared
through three stages of grinding and two stages of polishing. Then, the samples

52
are examined under the microscope with the 380 magnification. The optical
textures of coke samples show arrangement, distribution and length of molecular
orientation (MO). Coking coals are composed of plane poly-aromatic structural
units, with sizes less than 1 nm, maintained in mutual disorder by different hetero-
atoms. During pyrolysis, these groups are eliminated and reoriented themselves
locally parallel to form MO. Coke schematically shows the structure where each
pore wall is formed by a MO as shown in Fig. 2.23 (Duval et al., 1988). All cokes
have the same micro-texture but the size of MO and the size of pore are various.

Figure 2.23: Model of micro-texture of coke (Duval et al., 1988)

MO appeared as aggregates of bright domains and the length of MO can be

measured (LMO). The microscopical isotropic and anisotropic regions of coke are
exposed to surfaces in the form of appearance, molecular ordering, MO
distribution and size. Generally accepting that for coals that become fluid during
pyrolysis, the anisotropy textures are developed during the plastic stage and they
seem to increase in intensity. The anisotropic regions can be classified broadly
into mosaic and flow type, and these classes can be subdivided according to the
dimensions of unit area. Coals can be transformed to the strong anisotropic coke
textures, coarse mosaic and flow type when the heavy carbon composites are
melted enough due to high fluidity at a high temperature. Also, coals can be
transformed to weak anisotropic coke textures, fine mosaic and medium mosaic
when the light carbon composites are melted enough due to high fluidity at a low
temperature. Besides, the insufficient plastic phases remain iso-vitrinite and inert

53
which are the partially melted heavy carbon components during pyrolysis. The
isotropic regions roughly are classified into iso-vitrinite and inert.

The comparative proportions of coke textures can present the coke properties such
as coke reactivity and strength. According to Sharma (2005), increase in isotropic
textures will make a coke weak and reactive. This is because the higher surface
area of LMO available for reaction and the intrinsic reactivity may be higher for
molecule. On the other hand, the strength of coke increases with anisotropy.
Figure 2.24 shows the increasing CSR according to the increase proportion of
anisotropic mosaics (Vogt and Depoux, 1990). The growths of anisotropic
textures mainly depend on the fluidity of source of coal. Therefore, petrographic
analysis of coke is one of the important key to reveal the correlation between
source of coal petrology and coke textures.

Figure 2.24: Increasing of CSR with coke anisotropy (Vogt and Depoux, 1990)

However, the improvement of coke qualities is sophisticate process whereas it is

imperative to increase coke qualities for the operation of the bigger blast furnace.
This is because coke qualities are influenced by various factors such as charging
condition of coal to coke oven, carbonization condition, condition of a coke oven
and coal properties as shown Fig. 2. 25.

54
 Figure 2.25: The factors of influence on coke qualities

Initially, charged coal conditions can be divided by bulk density and moisture in a
coking chamber. The distance between coals is dominantly depended by the bulk
density so that high bulk density in coking chamber can make a close distance
between coal particles which can be fused easily. The initial state of moisture is
also an effective factor to control the bulk density of a coal. The high percentage
of moisture in coal coats on the surface of particle and this coating will make void
volume between coal particles so that bulk density of coal will be decreased. The
void volume of coal particles caused by moisture can be remained during
carbonization and will make the bad influence on coke qualities although the
moisture will evaporates around 100 ºC. Moreover, an increase in moisture
contents also involves a reduction in heating rate and a waste of the calorific value.
According to the coal moisture control process, the moisture of coals can be
reduced from 2 to 7 % with increasing the charging coal bulk density. Besides, a
small addition of sufficiently fluid oil importantly increases the density of wet
blends whereas reduces the dry blends (Loison et al., 1989). However, lower
contents of coal moisture are worse for an environment and a campaign life of
coke oven because these can cause severe emission problems and high coking
pressure.

55
The carbonization and oven conditions such as heating rate, coking time, soaking
time and temperature distribution are also significant factors of coke qualities
present and after plastic phases whereas the condition of charged coal is set before
plastic phases. These coking conditions can affect the agglomerating properties of
coals in the plastic phases. Besides, these properties are also directly related to the
gasification process which can make coke harder with splitting off the residual
gases. The slow heating rate and long coking time make the long duration of
plastic phase and enough release of residual gases which are making high strength
of coke. According to the Loison (1989), when the residence time extension, the
mechanical properties of coke go towards a state of thermally stabilized

However, coke oven operation is already conventionalized to the optimum

condition of the coke production rate and qualities for conservation of the coke
oven. To preserve campaign life of the coke oven, coking temperature in coke
oven has to be everlastingly stable as about 1200 ºC by controlling of the heating
temperature. Furthermore, rapid change of operation condition can cause air
pollution by leakage of the coke oven gas or micro-size particle of the coal.
Therefore, conditions of coke oven can’t be changed easily and the innovative
improvement of the coke quality seems to be impossible due to the complexity of
operation conditions.

Coking coal, as a carbonization source, is a source to improve coke qualities even

for slight level by blending various rank coals which have diverse properties. The
better choice of the coking coals can be made with better coking conditions with
coal properties, so that obtain a coke of better qualities. The effect of varied coal
properties such as rheology, rank, maceral and mineral constituent are used for an
optimization of a coal selection to carbonize. However, it still remains as black
box that thermo-dynamic behaviour of a coking coal in a micro-level whereas coal
properties are already used for a real coking operation. This is based on the
difference of coal properties in between a laboratory scale test and an industrial
operation, because industrial operation can be influenced by more variables.

56
2.4.2 Coal Rank

During the carbonization of a coking coal, the range between the maximum
fluidity temperature (MFT) and resolidification temperature (RT) is the most
important stage to effect on the coke qualities because the weight loss with
generating tar and texture change from coal to coke appears to be happened
mainly between MFT and RT. However, the similar mean values of fluidity of
coals e.g. coal blend with high volatile and low volatile, single medium volatile
coal can have different coke qualities because the fluid temperature range is not
enough to represent the coke qualities without the consideration of the coal rank
indicator.

Therefore, the degree of coalification, ascribed by the rank of the coal, plays a
very important role in the properties of the coke made from the coal. The coal
rank, in general, is expressed by the reflectance or the contents of volatile matter.
Zhang (2004) shows the relationship between coke thermal properties and coal
rank indicator such as volatile matter (Vd) and random reflectance (Rvo) in Fig.
2.34. These two graphs illustrate that a coke with good thermal strength can be
made from a coal with an Rvo of 1.1–1.2 and Vd of 22–26%.

Figure 2.26: Effect of volatile matter and coal reflectance on coke CRI and CSR
(Zhang et al., 2004)

57
The relationship between the mechanical strength of the coke and rank indicator is
also addressed as shown in Fig. 2.35. According to Vanniekerk (1991), the
correlation coefficient of the plant data between the Drum index strength (DI 150 30)
of the coke and the volatile contents of coal is 0.74. This result, to a certain extent,
confirms the intimate relationship between coke strength and rank indicator of
coal whereas the plant data is not experimentally monitored and another large
number of variables influence the operation on the coke quality.

Figure 2.27: The effect of the volatile content of coal on the cold strength of coke
(Vanniekerk and Dippenaar, 1991)

Figure 2.36, according to Lin (1986), shows the existence of the rank dominated
region by showing the relationship between coke strength and maximum fluidity
with increasing fluidity. It illustrates that with the addition of high fluidity of coal,
the coke strength increases with increasing coal blend fluidity when the maximum
fluidity is lower than 200 ddpm. Thereafter, the strength of coke is dominated by
fluidity of coal blends when it is lower than 200 ddpm, due to lack of caking
component. However, after 200 ddpm, the strength of coke is dominated by the
rank of coal so that the further addition of high fluidity coals cannot improve the
coke mechanical strength, due to excessive caking property. Addition of inert

58
material such as coke breeze can improve the coke strength by optimising the
maximum fluidity in the stage of excess fluidity of coal blends (Lin and Hong,
1986).

Figure 2.28: The relation between maximum fluidity and coke strength; M30 (Lin
and Hong, 1986)

Microscopically, the carbon forms of types produced from the vitrinite of

increasing reflectance within the range of coking coals are illustrated in Fig. 2.37.
This figure shows the close relationship between the carbon forms in coke and the
vitrinite reflectance of the initial coals. The amorphous forms of carbon structure
of coke which themselves are low in reflectance are produced by the coals from
the Midwestern coal basin with the 0.6% of reflectance. On the other hand, the
cokes from coals of higher reflectance exhibit the development of the graphite
structure of carbon forms so that fused carbon structures of coke are changed from
the isotropic to the anisotropic domain between 0.80 and 1.4% of reflectance
increase of coal origin. Therefore, these images from the electron microscope
shows that reflectance of the carbon forms of the higher-rank coals are

59
considerably higher than that produced from the low-rank coals (Benedict and
Thompson, 1980).

Figure 2.29: Relationship between coal rank and coke wall carbon forms
(Benedict and Thompson, 1980)

2.4.3 Coal Maceral

The possibility of producing good quality of coke from a given bituminous coal
depend on its coking power which are based on rank, oxidation, oil addition,
selective crushing and maceral composition (Stach and Murchison, 1982). Among
the factors which effect on coke quality, reflectance of the maceral analysis is still
the most important part of the coal petrology. It is apparent that for a given rank
of coal there are differences in the dilatation due to the variety of contents of
volatile matter. However, there is a question that difference of coking power of

60
blends having the same yield of volatile matter. In Fig. 2.38 five coals of about
same volatile matter but differing in dilatation, are presented. The reflectogram in
the middle of the figure can only explain the extreme difference in coking power
as characterized by the swelling index. Thus, for the reliable expectation of the
coking power, analyses of maceral groups should be carried out in addition to the
coal rank.

Figure 2.30: Reflectogram: mean reflectance, scatter and swelling index of 5

coking coal blends (Stach and Murchison, 1982)

Generally, coal macerals of coking coal are divided into three groups, vitrinite,
exinite and inertinite which can be classified by ratio of light reflection. Vitrinite
group in coking coals, the most part of the maceral, has characteristics which
swelling and agglomerating during the coking stage. Its density increases with
rank from 1.2 to 1.7. Exinite group, the most fluid maceral during coking stage, is
chemically distinguished by the residence of 10 or 20% of aliphatic carbon chains.
Inertinite group, the most dense maceral, is remains as inert material when it
coked except for a few semifusinite (Loison et al., 1989).

Vitrinite group which originates from lignocellulose tissues e.g. wood, roots, bark,
etc. appears to be a wall which surrounds the other macerals and mineral matter,
also is the most abundant, isolated and homogenous maceral group of coal.
Vitrinite constituent is the majority of normal coals, so that its thermodynamic
behaviour is very similar to that of the whole coal (Loison et al., 1989). Vitrinite
oxygen contents are higher than other macerals whereas hydrogen and carbon

61
contents are intermediate (Davis, 1976). At maximum fluidity, the vitrinite
constituents generate similar proportions of fluid material as the whole coal,
accounting for 30% of the hydrogen, when the semifusinite fractions yield only 15%
of mobile hydrogen (Maroto-Valer et al., 1998). The fluidity of the mobile
material is excessively higher for the vitrinite concentrates than inertinite fractions.

The graph about total dilatation plotted against the vitrinite content of six
Cretaceous Canadian and four Carboniferous Appalachian coals is illustrated in
Fig. 2.39 (left). Dilatation for any of the Canadian coals was less than half of the
Appalachian coals at high vitrinite contents. The differences in total dilatation (c +
d) between Appalachian and Canadian coals have a larger tendency with the
vitrinite content increased (Gransden et al., 1991). Additionally, according to the
study of Kidena (2002), vitrinite rich group have higher fluidity and lower
physical density. In this study, two kinds of maceral concentrates of coal, one is
vitrinite-rich Goonyella (GNY) and the other contains about 50% inertinite,
Witbank (WIT), were prepared. Moreover, further classification was also
processed by the maceral concentration of inertinite-rich fractions (“A” samples)
and vitrinite-rich fractions (“B” samples: higher fluidity) from both coals. As
shown in Fig. 2.39 (right), GNY coal has wider of plastic range than WIT and
vitrinite-rich “B” samples have wider range of fluidity than “A” samples in the
same kinds of coal. In the case of sample “B,” the maximum temperature (Tmax) is
located between maximum fluidity temperature (MFT) and resolidification
temperature (RT). This implies that effective evolution of tars occurs during this
temperature range while the maximization of tar evolution in “A” sample occurs
after resolidification. This is because vitrinite contents have a greater number of
aliphatic moieties and alicyclic part which connected with relative small aromatic
rings although inertinite have relatively large aromatic clusters.

62
 Figure 2.31: Effect of vitrinite content on the total dilatation (Gransden et al.,
1991) and correlation between Tmax and Gieseler plastometry-related temperatures
(Kidena et al., 2002)

Thus, the reflectance of vitrinite group, displaying optical anistropy seems to be

one of the significant factors to correlate with coke strength due to the most
significance of plastic phases of coking coal in carbonization. Vitrinite group is
classified three sub-groups of vitrinite, telinite, collinite and vitrodetrinite (Stach
and Murchison, 1982). Telinite is the cell wall originated from cellular structures
of trunks, branches, stem, leaves and roots. Collinite is the structure less
constituents of vitrinite originated from all parts of any plants. Stach (1982) also
claims that telinite has a better coking power than collinite. However, often
telinite cell walls are filled with collinite or resinite, micrinite or clay. In such
cases, it is difficult to distinguish of two types of macerals, telinite and collinite in
colour or refractive index between cell walls and collinite filling due to the
rareness of original telinite. Thus, new nomenclature term ‘telocollinite’ which
characterized by a normal reflectance has been introduced.

Vitrodetrinite can occur in the form of fragment mostly originated from humic
peat particles or plants which were degraded in early stage. This degraded or
broken matter of vitrinite is common in bituminous coal and can not be
distinguished easily from each other through the examination in normal reflected
light. This type of vitrinite occurs in the same seam in the layers of clarite,
duroclarite and clarodurite. They are grouped together as the group of
‘desmocollinite’, having the somewhat weaker reflectance.

63
Figure 2.32: Images of three macerals of vitrinite group (Krevelen, 1993)

There is particular type of vitrinite which is called ‘pseudovitrinite’ which has a

lower hydrogen contents than telocollinite. According to the Kruszewska (1998),
pseudovitrinite is characterized as the slitted structure, marked fissurization, the
absence of pyrite, higher relief and higher reflectance than telocollinite. This at
least 0.025% higher reflectance indicates that pseudovitrinite should be regarded
as part of oxidized telocollinite, which underwent a mild oxidation. The slitted
structure is considered to be the most predominant feature of the pseudovitrinite.
The existence of pseudovitrinite in coking coal is generally blamed for poor
coking power due to the delay of pore formation and the increased amount of ash
particles. However, Figure 2.41 shows that this inert property of pseudovitrinite is
untrue. This figure shows that reflectance distribution of pseudovitrinite (2)
associate with telocollinite (1). The diverse presence of pseudovitrinite particles at
450℃ can be examined, but lack of theses particles at 800℃ indicates that
pseudovitrinite has reactive potential. Stach (1982) also claims that there is no
denying of coking power of pseudovitrinite, to some extent, inferior to that of
normal vitrinite. Besides, according to the test processed by Benedict (1980), the
West Virginia-Pennsylvania coals, which contain low percentages of an
essentially unaltered pseudovitrinite, produced cokes with comparatively thick
walls and relatively low reactivity. On the other hand, the Eastern Kentucky coals
contain a large percentage of severely altered pseudovitrinite, produced cokes
with the reduced wall thicknesses and high reactivity of the resulting cokes. As
previously noted, the carbon forms of coke from pseudovitrinite are highly
reactive and amorphous.

64
 Figure 2.33: Reflectogram of pseudovitrinite distribution (Kruszewska, 1998)

Besides, other maceral groups are also important to investigate relationship

between coal petrography and coke qualities. Exinite group shows greater
plasticity during coking stage than vitrinite group whereas exinites are a minor
component of coking coal at the ratio about 5~15% in general. For the same rank
coking coal, exinite shows comparatively low reflectance and highest hydrogen,
volatile matter and aliphatic carbon chain (Stach and Murchison, 1982). Exinite
group composed of macerals, sporinite, cutinite, resinite, alginite, etc. originated
from fats and oils of vegetable origin. The exinite included petrographic concept
‘protobitumina’ which is opposed to the chemical term ‘bitumen’. They produce
much tar and gas during carbonization due to higher hydrogen contents.

Sporinite originated from the skins of spores and pollens in lignite and bituminous
coal is the most important maceral of the exinite group (Stach and Murchison,
1982). This is the lightest constituents of the coal and its density increases with
rank. An important physical property of sporinite is its toughness contrast to the
brittleness of surrounding vitrinite. They act in analogous way to iron rods to
reinforce concrete in holding together a spore-rich durite layer. Particularly,

65
sporinite yield exceptionally high amount of tar when it was carbonized owing to
its chemical composition of sporine, aliphatic aromatic skeleton without fatty acid.

Cutine is similar to sporine but not identical to its structure which forms the outer
layers of leaves or cuticles (Stach and Murchison, 1982). The resistant cutine
layers are termed ‘cutinite’, appears in the brown and bituminous coal as an
accessory maceral as a form of layers. The physical properties of cutinite are
similar to those of sporinite in the colour, reflectance and toughness with slight
difference so that cutinite usually associates with sporinite. The cutine, basic
substance of cutinite is closely related to suberine (cork).

The behaviour of sporinite and cutinite depends on their association with vitrinite
macerals so that they are inter-grown with the tar, decomposition products of
vitrinite, act as softeners and cause fusibility of vitrinite to increase. Well fused
coke with large pores is the result of thermal effect of sporinite and cutinite on
carbonization. However, inertinite are not influenced by the decomposition of
sporinite or cutinite. This is proved by the coke form in discordance of the size
and shape as well as the number of pores after carbonization of inertinite with the
sporinite and cutinite.

Resinite is originated from not only resins of plants but also secretions such as oils
occurring in the leaves which have been converted into resin with the form of
small rounded bodies (Stach and Murchison, 1982). In coals, it appears mostly as
cell fillings as well as layers of finely dispersed. Resinite bodies usually appear as
small spherical or spindle-shaped bodies and darker than sporinite and cutinite in
vertical section. The reflectance of the resinite of low-rank coals varies at the
range from 0.13 to 0.20 but generally lower than those of the corresponding
vitrinite. Resinite contains more hydrogen than do sporinite of the same carbon.
The resinite alerted chemically, becoming poorer in hydrogen and richer in carbon
according to the increase of coal rank so that reflectance of resinite also
corresponds with the rank of coal. The thermal behaviour of resinite is different
from sporinite and cutinite. In coking coal, resinite doesn’t form a coke but
enhances the plasticity of vitrinite when it was carbonized.

66
Alginite originated from algae is found in certain specific coal and distinguished
from high hydrogen ration compare to other exinite group (Stach and Murchison,
1982). In low rank of coals, alginite illustrates the weakest reflectance and the
strongest fluorescence of all macerals. Its colour is black to dark grey on polished
surfaces and white to yellow in thin sections.

Inertinite group is also closely related to coke qualities by containing higher

aromatic carbons which will compose basic structure of coke as the gravel and
sand in concrete. The fusible property of inertinite is very low whereas the
macerals of the vitrinite and exinite groups in coking coals will soften and melt
during carbonization. Thus, excluding few exceptions of Australian bituminous
coals, inertinite is regarded as inert material with high reflectance when it was
carbonized (Diessel, 1983). Inertinite group is also composed of several macerals
such as fusinite, semifusinite, sclerotinite, micrinite, e.g.

Fusinite originated from char coal is characterized as the richest in carbon of all
the constituents of coal. Fusinite occurs in peat, brown coal and bituminous coal
in general, but its proportion is low and does not exceed few percent. The colour
of fusinite is yellow-wish white to white in reflected light and black in transmitted
light. Its reflectance is highest of all macerals except the vitrinite in anthracite.
The formation of cell-wall fragment of fusinite illustrated as pointed, sharp-angled
and needle-shaped splinters called as ‘fusinite needles as shown in Fig. 2.7.
Chemically, fusinite contains high carbon and low hydrogen contents and it
volatile matter is decreasing with the increase in rank of coal.

Semifusinite is the intermediate of transition between fusinite and telinite. The

colour of semifusinite is light-grey to white in reflected light and brown to black
in transmitted light. Their density in bituminous coal varies in a range of high
degree from 1.35 to 1.45 due to their anisotropic structure. Their chemical
composition is richer in carbon and poorer in hydrogen than vitrinite and richer in
hydrogen than fusinite so that their plasticity is less than vitrinite. According to
the test by Maroto-Valer (1998), the fluid material in the semifusinite fractions in

67
medium volatile bituminous coal only accounts for 15% of the hydrogen at
maximum fluidity whereas the vitrinite concentrates from same coal produce 30%
of mobile hydrogen.

Sclerotinite is consisted in the fungal remains of brown and hard coals in

Carboniferous and Permian times so that it can be termed as ‘funginite’ (Stach
and Murchison, 1982). They occur in the form of cellular, non-cellular or tubular
hyphae and easily recognizable as fungal remains. Its reflectance is remarkably
high about 0.4% in soft brown coals and 6% in bituminous coals. Its colour is
white to pale grey on polished surfaces and black to brown in thin sections.
Sclerotinite does not soften until the temperature about 600℃, but dissolved at the
temperature range from 600 to 1000℃.

Micrinite is particular maceral which appears in the close vicinity of sporinite as

not only in the form of cell fillings but also in cell walls. Besides, micrinite often
occurs in finely dispersed in collinite. This is characterized as rounded shape and
the very small size of its grain. The colour of micrinite is pale-grey to white in
reflected light and black or dark brown in transmitted light. Micrinite behave as an
inert material at least up to 600℃ but it is differ from other inertinite maceral due
to the large amount of volatile matter contents.

Previously, reactivity of inertinite has been ignored due to the volumetric

unimportance of transition between vitrinite and semifusinite. However, this
conventional notion is changed by the fact that some coals contain more reactive
inertinite such as reactive semifusinite, semimicrinite and semi-inertinite.
Therefore, inertinite can be classified to the 4 types by its size and reactivity
(Diessel, 1983). Firstly, highly reactive inertinite has equal or exceeding
reflectance that of associated vitrinite and property of partial melting to complete
textural integration, large to small degassing pores and mosaic to granular flow
anisotropy. The second type of moderately reactive inertinite also has equal to or
exceeding reflectance that of associated vitrinite with weak plasticity but
reasonable to good bonding and basic anisotropy. Another non-reactive inertinite
of small size which mainly originated from micrinite and inertodetrinite can

68
strengthen coke fabric by good integration. Lastly, non-reactive inertinite of large
size weakens coke fabric by poor integration.

Consequently, the analysis of the effect of individual maceral and distribution is

another important factor to evaluate the coking power of bituminous coal with the
rank of coal. Vitrinite effect on coke qualities is revealed as a key factor to
understand the plastic phases of coal carbonization because vitrinite is the most
abundant maceral groups in coking coal. Exinite is also significant due to its
extreme high plasticity and synergy effect with other macerals. Effect of inertinite
on coke qualities is certainly useful to reveal the relationship between inertinite
and coke strength because it can form a basic structure of coke. Moreover, some
plasticity of inertinite is still not well known but it can be new clue to analyse the
effect of maceral on coke. However, coke quality prediction by vitrinite reflection
has become less reliable with increasing usage of non-premium coal in cost-
effective blends. The further research of the individual maceral such as telinite
and collinite can reveal this unproved question.

2.4.4 Coal Fluidity / Rheology

Gieseler Fluidity ASTM D 2639

Plastometer is an instrument to measure the resistance to rotation of a mobile

device immersed in a mass of coal. There are two distinct types of plastometers.
One is plastometer using a constant torque, with the measuring of the speed of
rotation, another is constant-speed plastometer, in which the torque opposing
rotation is measured (Loison et al., 1989). The standard form of constant torque-
plastometer was devised by Gieseler. This plastometer measures the thermo-
plastic behaviour of coal and determine the temperature of the maximum fluidity.
A paddle type stirrer is placed in the five grams of coal which ground under 400
meshes and pressed with a 1 kg weight for ten times. The test starts from 350 ºC
and is heated at a rate of 3 ºC per minute. Meanwhile, the stirrer start to stir the
coals in the retort is driven with a constant torque of 10-5Nm. The stirrer’s

69
revolutions are measured on a dial which is divided by 100 divisions and reported
as dial divisions per minute (DDPM). When the coal is cold, the stirrer is
immobile. It begins to revolve when the coal acquires certain fluidity. The initial
temperature of softening (Ts) is recorded when the dial reads 1 DDPM, and
stirrer’s speed increases with fluidity, passing through maximum of temperature
(Tm). Then its speed decreases and stops when the coal resolidifies (Tr). The
temperature range from Ts to Tr often called plastic range of coal which is
significant in blending of coking coal. Generally, high volatile coals have wide
range of plastic range so that its overlapping of other coal’s plastic range is
advantageous to increase coke strength. Logarithmic maximum fluidity (lg MF) is
a logarithm value of fluidity which concisely expresses the value of Gieseler
maximum fluidity.

Figure 2.34: Gieseler Fluidity of Diagram

Ruhr Dilatation ASTM D 5155

In dilatation test, five grams of coal which ground under 70 meshes are wetted
and placed into a pencil of 60 mm long by pressing with 2 Ton. The pencil is
placed in a tub and a sliding fit steel rod placed over it. The tub is placed in a
special furnace at 300 ºC and heated at 3 ºC per minute like Gieseler Plastometer.
The probe on the tub is moved with a contraction and dilatation of a coal pencil.
The moving downward and upward of probe is measured and reported as shown
in Fig. 2.27. This thermodynamic movement can explain the behaviour of coal
during fusion and softening as well as the contraction of semicoke beyond the

70
temperature of resolidification (Loison et al., 1989). Before the coals enter the
plastic zone, the contraction observed is due to the softening of particles, which
agglomerate with one another. The initial temperature of shrinkage is recorded as
dilatation softening temperature (θs), and when the probe goes the lowest height
the temperature is reported as maximum contraction temperature (θc), lastly the
temperature point when it meets the highest height is determined as end of
swelling temperature (θe). The total dilatation value (g) is calculated by
summation of contraction percentage (c) and dilatation percentage (d).

Free Swelling Index ASTM 720 - 91

A finely ground 1 gram of sample under 60 mesh of fusible coal is placed in a

crucible and then heated under the temperature about 820 ºC for 150 seconds. The
given off volatile matter causes the viscous mass to swell and after the
resolidification the residue then resembles a porous mass of coke which is light
and larger in volume than the original sample. This resulted coke ‘button’ is
compared with a series of standard profiles numbered 1 to 9 as shown in Fig. 2.28.
Free Swelling Index permits quick classification of coking coals having mediocre
or average coking properties, but it makes hardly any distinction between good or
bad coking coals because it does not measure the plasticity but the swelling
properties.

Figure 2.35: Curve recorded by the dilatometer during heating (Loison et al.,
1989)

71
 Figure 2.36: Scale of reference profiles for the crucible swelling test (Loison et
al., 1989)

G-Factor Value

This is the index to present the characteristics of strength of coke after coking
which evaluate the bonding power between inert and reactive. It was mainly used
in China and devised from the Roga Index which is devised to estimate the caking
properties in Russia. 1 gram of coking coal sample is placed in crucible with 5
grams of anthracite and heated at constant temperature of 850ºC for 15 minutes.
Then specimen after carbonization is weighed and put into drum for the revolution
of 50 rpm for 5 minutes. After the revolution, residue of specimen screened and
over 1mm of amount is weighed. G-Factor value is the weight ratio of this final
residue compare to the weight before revolution. However, this value is merely
used as an indicator to predict coke quality due to the lack of correlation between
theoretical and measured value.

72
 Figure 2.37: G-Factor Measuring Apparatus

The rheological properties of coals such as Gieseler fluidity (DDPM), caking

index G (G-value), total dilatation (TD) and maximum thickness of plastic layer Y
play a significant role in coke quality whereas these are not always proportional to
the coking power of a coal (Zhang et al., 2004). Figure 2.30 show that all the
properties of the rheology have the similar effects on the coke thermal quality.
These results are also show proportional relationship between the rheological
factors and Coke Strength after Reaction (CSR) so that CSR has the tendency of
increase with the increase of the rheological factors except the excessive high
value of rheology of a coal. The high rheological index can agglomerate the coal
particle with one another as well as with the inert constituents so that high fluidity
of coal determines the strong bonding process during carbonization.
Microscopically, the caking properties of coal can be characterized as the
existence of the gas saturated zone in the plastic layer (Zubkova, 2005). The
appearance of this zone gives rise to the transportation of the plastic layer to a hot
side of coal with the formation of semi-coke and consequently increases its carbon
density.

73
 Figure 2.38: Relationship between rheological properties and CRI and CSR
(Zhang et al., 2004)

Additionally, the mechanical properties of coke also depend on the plasticity and
the period during which they remain in the plastic state. Vanniekerk (1991) shows
the schematic relationship between the cold strength of coke and caking property
through the diagram as shown in Fig. 2.31. It is evident that regions are divided
into the domination by caking property and rank of coal. An increase in caking
property will result in an increase in the cold strength of the coke at the stage of
weak caking property (Region 1). However, in the rank-dominated region, no
further increase in strength will be resulted and the caking property will even
result in a decrease in the cold strength during the state of the excessive caking
property (Region 2).

Figure 2.39: A schematic illustration of the dependence of coke strength on the

coal rank and caking properties of coal (Vanniekerk and Dippenaar, 1991)

74
As the Fig. 2.31 illustrates, too high fluidity is not favourable because it facilitates
the rapid release of the gases from the mass of coal before the resolidification.
This rapid rate of gas release will decrease the swelling rate of coal so that the
strength of coke quality will be also decreased. This is because the swelling
appears to be better criterion of caking than plasticity so that total dilatation (TD)
has a better relationship with a mechanical strength of the coke. The reason why
the TD greatly influences the cold strength of a coke is considered as follows.
Nomura (2004) shows the schematic mechanism of the pore formation in coal
particles during carbonization in Fig. 2.32. In this Figure coal particles are
charged in the oven chamber where partly come to contact with each other. At the
softening temperature, the coal particles start to soften and expand into the inter-
particle space so the fused coal particles come into contact with each other, stick
together and finally form strong bonds when the volume of coal expansion is
larger than that of inter-particle space. Otherwise, the fused coal particles can’t
stick together well and weak bonds are formed when the volume of coal
expansion is smaller than that of inter-particle space. Therefore, it seemed that the
effect of coal expansion on cold strength of coke is the significant factor to
determine the filling of the inter-particle space with bulk density. Then, to
evaluate the coal expansion and the bulk density on coke strength, a new concept,
‘specific dilatation volume’, the volume of expanded coal per unit mass of coal
can be introduced as shown in Fig. 2.32. This index is acquired by dividing the
volume of expansion of a coal by the mass of the test specimen (coal pencil) for
dilatometry. The volume of expanded coal is obtained by multiplying the cross-
section area of the dilatometer retort by the height of the expanded coal sample.

Figure 2.40: The Coal expansion mechanism; left and the definition of specific
dilatation volume; right (Nomura, 2004)

75
The rheological properties come to be more complicated in a blending of a coal.
Some blended coals can have lower rheological properties, whereas other blended
coals can have higher. The blending work, however, can change the effect of the
inherent properties of the rheology of a single coal. This apparent effect of the
interactions on the plastic range is shown in Fig. 2.33. The value of the maximum
fluidity of blended coal is slightly greater than expected value in the absence of
interactions whereas the plastic range is reduced. There are no correlations
between the difference between the measured and predicted softening temperature
and the difference between measured and predicted solidification temperature
(Sakurovs, 2003). The decrease in plastic range is not symmetrical about the
temperature of maximum fluidity because the difference in coal rank for the blend
makes the measured plastic range narrower than the predicted values are.

Figure 2.41: Schematic of effect of interactions on the thermoplastic behaviour of

blends (Sakurovs, 2003)

2.4.5 Coal Inorganic Matter

In addition to the analysis of organic matter of coal, inorganic matter, specifically

mineral matter in coal is also important factor to estimate the quality of the future
coke. This is because coke reactivity is mainly influenced by its individual
mineral and distribution whereas minerals occupy a small volume of coke about
less than 15 wt% of total coke. Thus, the conventional analysis coke reaction rate

76
has to be improved by the mineral analysis of using advanced technology such as
scanning electron microscopy (SEM) and quantitative X-ray diffraction (XRD).

Coke minerals were physically classified into three groups, agglomerate, coarse
mineral and fine as illustrated in Fig. 2.42 (Gupta, 2008). Agglomerate mineral
characterized as the size varied with the range from 100 to 1000μm, discrete
distribution and refractory reactivity. Coarse mineral has a size of 50-100μm with
disseminated distribution and moderate reactivity. Fine mineral has a size less
than 50μm and it is distributed as pore inclusions with reactive thermal nature.

Figure 2.42: SEM images illustrating mineral distribution in coke: (a) typical
discrete mineral agglomerate (b) disseminated (c) pore inclusions (Gupta, 2008)

Additionally, mineral matter in coke plays a major role to determine the coke
strength after reaction (CSR) because thermodynamic behaviour of mineral can be
addressed to the coke degradation pathways, both directly and indirectly (Grigore
et al., 2007). Directly, mineral causes the carbon consumption due to reduction of
the oxides in the coke above 1400℃ and expands the meta-aluminosilicates in the
lower zone of blast furnace. Indirectly, some metallic iron such as pyrrhotite, iron
oxides and calcium oxide facilitate the catalyst effect of gas phase reaction. The
mineral matter in the coals is mostly composed of aluminosilicates (clay minerals
and, less commonly, feldspar), carbonates, quartz and pyrite as written in chapter
2.2.6. The principal clay minerals presented as kaolinite, illite and
montmorillonite, and the latter containing cations as Ca, Fe, Mg, K and Na. Iron
existed as carbonate and sulfide. Calcium and magnesium can also be found in
carbonate.

77
According to the study of coal mineralogy by Gupta (2008), the reactivity of the
kaolinite, quartz and feldspar were described as least-reactive, illite and
montmorillonite as moderately reactive, and carbonate, sulfide and chloride as the
reactive phases. Similarly, coke minerals were also defined as refractory,
moderately reactive and reactive phases on the basis of the presence of relatively
high amounts of fluxing elements such as K, Ca and Fe. The reactivity of ash
(mineral) can be determined by mineral fusibility at high or low temperature.
Vassilev (1995) classified the coal ashes to the low-melting temperature ashes
which mostly existed in lower rank coals with increased concentration and high
melting temperature ashes which were contained in high rank coal. The low-
melting temperature ashes composed of S, Ca, Mg, Fe, and Na and, respectively,
sulphates, carbonates, sulphides, oxides, montmorillonite, and feldspars. The
high-melting temperature ashes contain Si, Al, and Ti and, respectively, quartz,
kaolinite, illite, and rutile, as well as some Fe oxides and siderite. Figure 2.43
shows the relationship between low melting temperature ashes and coke reactivity.
This appears that the initial apparent rate of coke reaction is increased with the
increase in the amount of total iron, potassium and sodium in the amorphous
phase of coke.

Figure 2.43: The total percentage of iron, potassium and sodium in the
amorphous phase in coke versus initial apparent rate (Grigore et al., 2006)

Minerals distribution in coke may be affected by the coking conditions such as

heating rate, coking temperature, bulk density of the charge and cooling procedure.

78
According the study processed by Grigore (2007), water quenching of coke didn’t
produce any of the hydrated sulfates. Besides, slower heating rate allows the
metallic iron produced to interact with other minerals such as cristobalite to form
ferrosilicide at low temperature. However, the effect of coking conditions on
mineral phases of coke in still not well known because the mineral matter of coal
undergoes complex reactions during carbonization from coal to coke. Further
study of the mineral matter transformation during coking can improve the better
control over coke quality.

2.5 Coke Strength Prediction

2.5.1 Coke Strength Model

According to the numerous studies to predict the quality of coke such as cold
strength and hot strength, a number of prediction models are developed.
Vanniekerk (1991) suggested correlation of prediction model of coke strength in
his paper by linear regression technics. Experimental data of coke qualities such
150 150
as the cold strength parameters (DI 15, DI 30) and hot strength parameters
(SARI-10, CO2-RI) are determined.

150
DI 15 = - (0.308 x volatile content, %) + (0.018 x temperature,℃) – 0.473 x
moisture content, %) + 69.68

150
DI 30 = - (1.562 x volatile content, %) + (0.026 x maximum fluidity, ddpm) -
(1.642 x moisture content, %) + 118.48

CO2-RI = 165.335 - (22.55 x mean maximum reflectance, Rm) – 0.0913 x

temperature, ℃
SARI-l0 = 90,529 + (0,302 x organic inerts, %) -(0.576 x moisture content, %).

79
In case of cold strength of coke as presented by Drum Index, 10 kg of +50 mm
coke is tumbled in a cylindrical drum (1.5m by 1.5m). After 150 revolutions at a
speed of 15 revolutions / min, the coke is removed and screened on a 15 mm or 30
mm sieve. The percentage by mass of coke remaining on the sieve is expressed as
the DI 15015 or DI 15030. Additionally, MLD (Modified Linear Drum) dynamic test
developed at Iscor to determine the hot strength of coke differs from the NSC test.
In the MLD method, the carbon dioxide reaction occurs in a rotating drum at a
temperature of 1100°C, but a 1500 g sample between 30 and 25 mm in particle
size is used. The drum rotates at a speed of 3 revolutions / min for 120 minutes
while the carbon dioxide is supplied constantly at a rate of 37 l/min. The CO2
reactivity index is expressed as follows:

CO2- RI = [(1500 - mass after reaction) / (1500)] x 1000,

and the strength-after-reaction index by

SARI-10 = [(oversize fraction remaining on 10 mm sieve)/(mass after reaction)] x

100.

Another coke strength prediction model based on extensive pilot coke oven
studies is developed by Research & Development Centre for Iron & Steel
(RDCIS), SAIL in India (Gupta, 2007). In this model, the quality of blast furnace
coke can be predicted in terms of ash contents and volatile matter contents. These
correlations are validated using monthly coke oven data and by analysing samples
of resultant coke and coal blend. Particularly, this following relationship excludes
some of coal index such as ddpm or reflectance whereas these parameters stand
for fluidity or rank of coking coal.

Ashcoke = Ashblend x 100 / (100 - VMblend + 0.9) R2 = 0.89

M10 = A x Ashblend + B x VMdaf - C R2 = 0.82
CSR = D - E x VMdaf + F x (VMdaf)2 – G x Ashblend R2= 0.77
CRI = H – I x CSR R2=0.86

80
Where, Ashcoke is % of ash in coke, Ashblend is % of ash in coal blends, VMblend is %
coal blend volatile matter, M10 is mecum index of coke, CSR is coke strength
after reaction, CRI is coke reactivity index, VMdaf is dry ash free volatile matter of
coal blends and A~I are constants.

However, these models above are too simple to predict exact coke qualities hence
the industrial operation of coke making is complex. Industrially, Bao steel
developed sophisticate model of coke quality prediction by using the evolved
Group Method Data Handle (GMDH) method which apply regression self-study
self-organization; adaptive and genetic algorithm to choice model parameter form
the expert experience as shown in Fig. 2.46 (Zhang et al., 2004). Experimental
data are used from the Baosteel’s 76-kg simulated coke oven (SCO), has the
interior dimensions of 500mm in height, 700mm in width and 400mm in depth.
Carbonization conditions of SCO oven are moisture of 10%, coking time of 18.5
hr, dry oven bulk density of 0.72~0.78 ton/m3, final coke temperature of 1050℃
and dry quenching time of 5.5hr. Where MCI is mineral catalysis index, TI is inert
contents, I is organic inert component calculated form automatic petrographic
measurement, R is random reflectance of coal, Vd is volatile material content, MF
is maximum fluidity, and G is caking index.

In addition to this theoretical prediction, more practical regression equations are

obtained by 259 plant coking tests under the controlled conditions of blend
particle size of (- 3 mm) 85% and dry-quenched. Models of prediction equation of
coke quality as a function of coal properties were established as a result of this
statistic analyses as shown in Fig. 2.47. The coke quality predicted by this model
meets with high correctness, with a deviation of < 2% for DI, < 3% for CRI and <
4% for CSR. For stable operation of Baosteel blast furnace, the coke quality
should be maintained above the level of Ad < 11.5%, Std < 0.54%, DI 15015 >87%,
CRI < 26%, and CSR>66%. Specifically, mineral catalysis index (MCI) has
proven to useful indicator for the model of CRI and CSR so that the role of
mineral constituent of coal also proved as important indicator in determining coke
product quality.

81
Figure 2.44: Experimental coke quality prediction model of Bao steel (Zhang et
al., 2004)

82
Figure 2.45: Industrial coke quality prediction model of Bao steel (Zhang et al.,
2004)

2.5.2 Effect of Blending

The relationship between coke qualities and coal properties is a key factor to
predict the coke qualities as written in chapter 2.4 as above. Additionally,
selection of coals is the most important factor to meet coke qualities. In the
operation of coke making, adjustment of the rank and fluidity of coal blending by
choosing the coals to use is the first step. To make strong coke, longer aromatic
carbon chain and higher agglomerating surroundings during carbonization are
essential. The high rank coal having longer aromatic carbon chains usually has
fewer branched aliphatic chains so that the high rank coal remains stable as solid

83
phases during pyrolysis. Reversely, the low rank coals with high fluidity has more
volatile matter which can escape through out plastic layers during pyrolysis, thus
they leave more pores in fluidized states and the pore wall thickness becomes
thinner. Both of these coals make cokes to weaker.

Therefore, it is general practice to blend high-volatile and low-volatile coals to

produce a medium-volatile rank of blend coal for coke production. The usage of
true medium-volatile coal alone or in large percentages is restricted because coal
supplies are conditioned by economic exigencies. High-volatile coal generally has
a lower softening temperature when it compare to others. The blends of High-
volatile coal and low-volatile coal should be carefully proportioned because
excessive amounts of high-volatile coal will reduce coke strength and excessive
amounts of low-volatile coal can cause high wall pressures to damage coke oven
and coke-sticking problem on the wall.

Additional rank and rheological properties of coal in blends also has to be

optimized when the coking coals are blended. Vitrinite reflectance and Strength
Index should be considered as the rank parameters of coal, when Gieseler fluidity,
Dilatation in Ruhr Dilatometer and Free Swelling Index has to be controlled to the
rheological parameters. However, the problem of optimization of coal blend for
coke making and blast furnace operations is complex to solve due to the nonlinear
characteristics of the transformation from coal to coke.

According to Álvarez (2007), these non-linear trends of coke qualities between

the experimental and predictive often appeared in mechanical strength of coke
rather than thermal strength of coke. Álvarez (2007) compared many coke
strength data from coals of binary blends as shown in Fig. 2.44. Mechanical
strength indices of coke, Irsid 10 and 20 data are scattered without correlation
whereas coke reaction index (CRI) and coke strength after reaction index (CSR)
show high correlated relationship with r-value of 0.983 and 0.966. This approach
for predicting coke qualities would be useful in industrial practice although only
binary coals in blends are used.

84
Figure 2.46: Relationship between the experimental and calculated Irsid, CRI and
CSR indices (Álvarez et al., 2007)

Coal properties of blends are slightly different from their individual coal
properties due to the interaction between different kinds of coal particles. Many
blends of coking coal interact in a way that effect fluidity when they heated. For
instance, significant interactions occurred at the temperature range from 400 to
520℃ with increasing difference in rank and fluidity (Sakurovs, 2003). This
interaction can be both positive and negative. The positive type of interaction
increases with increasing rank difference and maximum fluidity of the low rank
coal in blends. This interaction can’t be found when the ranks of the two coals are
the same, thus wouldn’t be observed in the mixing of same coal which have same
concentrates of vitrinite and inertinite. The interaction meets its most positive
value about 470℃, above the mean temperature of maximum fluidity of the lower
rank coals in blends. The negative interaction increases at high temperature with
increase in difference in rank and fluidity of the higher rank coal in blends. The
lower rank coal acts as sorbent and absorb the plasticizer released from the higher
rank of coal when it solidify itself. Consequently, it will decrease the
resolidification temperature and the net fluidity of coal blends.

85
Generally, target for blends of coal has to be aimed to maintain a maximum
fluidity (Gieseler fluidity) in the range from 200 to 1000 dial division per minute
(ddpm) and a mean maximum reflectance in the range from 1.2 to 1.3. Several
models developed by Japanese steel making company to find the ideal blending of
coal for coke making. The MOF diagram developed by NKK Corp. shows the
specific combination of the range of vitrinite reflectance and logarithm of fluidity
(Gransden et al., 1991) as shown in Fig. 2.45 (left). This model suggesting that the
specific range of fluidity and reflectance above is exactly appropriate for the ideal
30/15
quality of coke exceeding 90 of JIS Drum Index (DI ). Figure 2.45 (right)
compares the several methods to predict the ASTM stability according to the
vitrinite reflectance and the total dilatation of coking coal. This diagram also
suggests standard blending area of vitrinite reflectance in the range of 1.1-1.2 as
well as dilatation in the range of 80-120.

Figure 2.47: Fluidity of bituminous coals vs. rank showing specifications of coal
properties for conventional and 30% partially briquetted charges; Comparison of
various models used to predict coke strength. Estimates of ASTM stability values
are shown on contour lines (Gransden et al., 1991)

86
2.6 Summary

Literature review highlighted following points.

● Coke quality is critical for blast furnace process and a higher strength of coke
is more important at higher pulverized coal injection (PCI) rate operation. It is
challenging to meet coke quality for the current and future operations such as low-
temperature operations which requires high reactivity coke without compromising
coke strength.

● Coke making process has been optimized for the best quality of coke. Reason
is the restricted logistics of coking coal disturbed to the optimization of coke
qualities. Furthermore, rapid change of operation condition can cause air pollution
by leakage of the coke oven gas or micro-size particle of the coal.

● Coal appears a source to further improve coke qualities through suitable

blending options of a variety of coals. Choice of the coking coals can be improved
by optimising coking conditions and coal properties. This requires improved
understanding of rheological properties, coal rank, maceral and mineral
constituents of coals on coke blends.

● A number of rheological parameters of coals have been used to relate with

coke quality of blends, but often do not vary with the coking power of individual
coals. High rheological index facilitates agglomeration of coal particles including
those of inert constitution. However, excessive high fluidity is not favourable
because it facilitates the rapid release of the gases from the mass of coal before the
re-solidification.

● Generally, high strength coke requires a narrow range coal with an Rvo of 1.1–
1.2 and VM of 22–26%.

● Mineral matter affects coke strength after reaction (CSR) due to carbo-thermal
reduction

87
● Vitrinite is well known indicator of plastic stage in terms of softening and
resolidification temperature due to its high oxygen contents. Exinite is also
significant due to its extreme high plasticity and synergy effect with other
macerals. The effect of inertinite on coke qualities is certainly useful to reveal the
relationship between inertinite and coke structure. However, the petrography of
individual macerals particularly vitrinite types is still not well understood due to
the complexity of maceral-specific analysis. This complexity is based on the
characteristics of maceral measurement which are time-consuming, subjective and
needing high level skill. Besides, coke quality prediction by vitrinite reflection has
become less reliable with increasing usage of non-premium coal in cost-effective
blends. The further research of the vitrinite types such as telinite and collinite can
reveal this unproved question.

From the literature review above, it is appeared that current studies are focused on
the effect of varied coal properties for an optimization of the selection of coking
coals to carbonize. However, the effect of coking coal properties on coke in
micro-level still remains as complex whereas coking coal of properties are already
used for an industrial operation. Specifically, the effect of blended coal properties
including vitrinite macerals on coke quality is still unproved question while the
abundant number of studies and experiments were processed about this issue.
Coking coal is a homogenous material but still composed various kinds of
macerals which are originated from various kinds of plants. The rank parameter,
vitrinite reflectance on coke qualities is revealed as a key factor to understand the
plastic phases of carbonization, but vitrinite types such as telinite and collinite are
still unclear for their effect on coke qualities.

Objectives: The specific objectives of the thesis are.

1. To investigate the effect of vitrinite types on coke strength and implications.

2. To improve understanding the effect of blending on coke strength and micro-

structures.

88
3 EXPERIMENTAL

To examine coking coal effect on coke qualities, the samples were prepared,
processed through the experimental procedure and analysed via microscopic
analysis. In this chapter we explain the experimental details of the sample
fabrication and analysis procedures, which is illustrated in Fig. 3.1. Specifically,
the preparation of the coal samples and coke samples are addressed with the
analytical methods used to investigate their properties.

89
 Coal Samples (single x12, blending x6)

Dry Oven (2Hr)

Grinding (60 mesh)

Hygrostat (24Hr)

Heating Heating Heating Gasification Burning

(107 ºC, 1Hr) (900 ºC, 5min) (815 ºC, 7min) Adsorption
Moisture Volatile Ash CHONS Calorie
Matter

Grinding (400 mesh) Grinding (70 mesh) Grinding (60 mesh)

Press (10kg) Press (2Ton) Heating (820ºC, 150s)
Heating (3 ºC/min) Heating (3 ºC/min) Free Swelling Index
Fluidity Dilatation

Grinding (18 to 100 mesh)

Drying (Moisture 6%)
Moulding
Box Charging
Polishing
Carbonization (850 to 1050)
ºC)
Optical Distinction Vitrinite (150point) Quenching (N2)

Maceral Reflectance Screening (75 to 25mm)

SEM Drum Grinding Crushing

(150revolution)
Morphology Drum Index Moulding Sampling (20mm)
Polishing Charging (1100 ºC)

SEM Optical Distinction N2 (7.5ml/min,

Pore 30min)
Coke Texture CO2 (5ml/min, 2hr)
Distribution
Coke Factor Quenching (N2)

Drum
(600revolution)
CSR Reactivity

Figure 3.1: Experimental procedure of the sample preparation and analysis

90
 3.1 Coal Samples

Eight Australian (AC01~08), two Canadian (CA01~02), one Chinese (CC01) and one
Russian (RC01) coals were selected, based on the different ranks, rheological factors
and macerals. Proximate and ultimate analysis of coking coal is provided in Table 3.1
with the rheological factors including logarithmic maximum fluidity (lg MF), total
dilatation (TD), and free swelling index (FSI) and G-factor. Additionally, the optical
distinctions of coking coals of maceral classification and rank parameters of mean
reflectance of vitrinite are given, too.

Chemical components of coal were measured by these standards below.

- Moisture (air dry base) ISO 331:1983

- Volatile Matter ISO 562:1998
- Ash ISO 1171: 1997

Rheological parameters of coal were measured by these standards below.

- Gieseler Fluidity ASTM D 2639

- Ruhr Dilatation ASTM D 5155
- Free Swelling Index ASTM 720 – 91

91
Table 3.1: Coal characteristics used for the investigation

Twelve Coking coals are sampled from the industrial blending bin so that all of
coal samples are already finely crushed under the size of 3mm. Figure 2.3 in
chapter 2 illustrates the location of blending stage in coke making process. To
prevent the accidental oxidation, all of the samples are instantly corrected whereas
a few hours of oxidation chance still exist during the transferring stage of coal by
belt conveyor. However, this few hours of oxidation chance appeared to be
ignorable. Particularly, among these twelve samples, four semi-soft coking coals
of two Australian (AC07, 08), one Canadian (CA01) and one Russian (RC01) are
selected for the study of blending effect on coke properties. Six blended coal
samples (BC01~06) are prepared by the certain blending ratio as shown in Table
3.2. AC07 is used as reference coal and provide enough fluidity to coal blends.

92
AC08, CA01 and RC01 are used as blending coal to provide enough character of
higher rank which is also a significant factor for the coke strength.

Table 3.2: Characteristics of the blended coal used for the investigation

In industrial test, forty kinds of coal in blends are used for one year’s test of real
operation. Average value of volatile percentage is 24.98, lg MF is 2.50, TD is 89,
mean reflectance is 1.11, vitrinite percentage is 66.27, collinite percentage is
39.54 and telinite percentage is 26.73 with minimum standard deviation.

Preparation of a coal sample for the maceral analysis was processed by ISO 7404-
2. A 15g sample of dried coal is ground under 18 to 100 meshes and moulded
with agglomerating and hardening agent. Then, specimens are prepared through 5
stages of polishing. The polished surfaces should be flat and free from scratches
and relief. The surface area should be at least 600 mm2. The coal particles must be

93
evenly distributed and make up at least 60 % of the polished cross-sectional area
of the block.

Detailed maceral analysis was processed by ISO 7404-3. Macerals are the
microscopically recognizable individual constituents of coal and depending on their
quantitative participation. For the classification of maceral, polarized light of optical
microscope is operated as shown in Fig. 3.2. Each maceral group is examined under
oil immersion lens of which magnification range from 250 to 2000 and is
distinguished by their relative reflectance, morphology, size and shape of the surface
of a polished specimen after directly incident on the surface. Continuous movement
of point counting of 0.5 mm intervals determines the proportion of the maceral groups
into the three, vitrinite (huminite in low rank coals and lignites), exinite and
inertinite. These macerals of coking coal show three grey levels owing to the
spread in their reflectance within the same sample: vitrinite appears medium-grey,
exinite black and inertinite white. However, distinction of sub-maceral from
maceral groups is very complicated due to its technological limitation and also
there is no prominent classification standard of individual maceral though there
are various ones (eg. American Society of Testing Materials, International
Committee for Coal and Organic Petrology, Australian Standards, Japan Standards,
etc). Thus, we devised POSCO Standards classification of maceral for the
simplification and practical use as shown in Table 3.3. Here in our research only
telinite and collinite are measured as the proportion among the individual
macerals since they are dominantly used macerals compared to other ones.

Figure 3.2: Polarized light of optical microscope; Leica DM4500P

94
 Table 3.3: POSCO Standards of maceral devised

Vitrinite of Reflectance, which is indicative of the rank or maturity of coals, was

measured by ISO 7404-5. Maceral reflectance is not the only most informative
parameter in the coals, but also it can be used to correlate or predict certain key
physical and chemical properties of coals. However, analysis of full maceral
reflectance is time-consuming and complicated so that analysis of the vitrinite is
examined. Vitrinite reflectance addresses optical anisotropy. Under polarized of
monochromatic light with a wavelength of 546 nm, the 150 point of reflectance value
of vitrinite varies from a minimum to a maximum. These 150 points are usually
taken on different vitrinite particles evenly distributed over the polished surface of
the specimen. Each vitrinite values are determined via comparison with the standard
sample of vitrinite reflectance. Average percentages of 150 reflectance values are
expressed as mean reflectance and also in the form of a reflectogram which shows
the vitrinite-class distribution. The reflectogram of individual coal can provide
inherent information about the rank of the coals so the analysis of reflectogram of
coal blends can track which coal series are used in coal blend.

3.2 Carbonisation Tests

Proximate analysis of coke from single coal is provided in Table 3.4 with the strength
index and size distribution. Additionally, strength data and texture analysis of coke
from blended coal are given in Table 3.5.

95
Chemical components of coke were measured by these standards below.
- Moisture (air dry base) ISO 331:1983
- Volatile Matter ISO 562:1998
- Ash ISO 1171: 1997

Table 3.4: Coke characteristics from carbonized single coal

Table 3.5: Coke characteristics from carbonized coal in blends

96
 3.2.1 Carbonisation in 30kg Operations

All the cokes in this test were carbonized through the 30 kg moveable wall pilot
oven as shown in Fig. 3.3. The specification of pilot oven is the length of 405 mm,
the height of inner space of 600 mm, and the width of 210 mm. To have similar
conditions with the industrial operation, the oven has two doors and one charge
hole on the top. Besides, for the stable charging density of 730 kg/m3, the wooden
boxes were used to fix charging amount of coal. The test coke oven has movable
wall to adjust coking pressure. Final coking temperature is set as 1100 ºC and
heated by electricity with increasing temperature at 3 ºC per minute for 7 hours
before pushing cokes. The carbonized cokes are cooled in a box with nitrogen for
the same effect of Coke Dry Quenching (CDQ) in the industrial process.

Figure 3.3: 25kg movable wall pilot coke oven

3.2.2 Carbonisation in Industrial Oven

Conventional type of coke oven battery in POHANG works, POSCO, was used in
real operational test. These oven chambers are composed of 68 chambers having
29.5m3 per chamber of valid volume with charging density of 20.8 ton per
chamber. The specification of oven chambers is the length of 15.5m, the height of
5.15m, and the width of 0.43m. Coal is gravitationally charged to each chamber
by charging car. Chambers are heated to 1,100ºC for 15 to 25 hours by

97
carbonisation of mixed gas with coke oven gas and blast furnace gas which
calorie is 1,100 kcal per Nm3. Through the semi-coke stage, coal was carbonised
to coke and pushed to the transfer car by the pusher car. Cokes in transfer car are
cooled by wet quenching.

3.2.3 Coke Strength Measurements

Drum index is considered as the most important index in coke in POSCO so that
every industrial condition of coke making is aimed to the stretch target value of
drum index. However, the breakage of coke in the revolving drum test is complex.
The level and shape of dominant breakage mechanism is influenced by drum type,
duration of test and particularly the number, size and angle of lifters. In this study,
Japanese style Drum Index tester is used. Diameter 1.5m drum with 6 lifters
inside is used to measure abrasion and fragmentation resistance of cokes as shown
in Fig. 3.4. The 10kg of cokes of which size is over 25mm are charged in the
drum, and they are rotated at 150 revolutions with 15 rpm. Due to its many lifters,
breakage in the JIS drum is dominated by impact when the coke is continuously
lifted to maximum heights by the 6 lifters. The abrasion products remained in the
drum tests may have two principal origins. Some would result from impact
crushing of coke and some by breakage along macro-fissures. The remained coke
proportion of over 15mm size after breakage in the drum is expressed as DI 15 150.
Afterwards, the coke size distribution and the mean size of cokes can be
calculated.

Figure 3.4: Drum Index Revolving Tester

98
Coke reactivity is the reduction rate with CO2 in blast furnaces. In blast furnace
operation, cokes should be provided continuously from the top to the bottom in
blast furnaces with reduction. Coke reduction reaction continuously occurs
because cokes are placed at a high temperature contacting with carbon dioxide
and metal oxide. Rank, rheology and alkali are known as the indicators affecting
on reactivity and strength after reaction. In this test, the JIS Coke Reactivity Index
(CRI) and Coke Strength after CO2 Reaction (CSR) are applied.

The Coke Reactivity Index (CRI) is the proportion of weight loss of cokes after
reaction with carbon dioxide. 200g of crushed size of dried coke in a size range
from 19mm to 21mm is heated in an electrical retort to 1100 ºC at nitrogen
atmosphere. Then, the atmosphere is changed to carbon dioxide for exactly 2
hours for the reduction. After the reaction period, the sample is allowed to be
cooled down to 50 ºC at nitrogen atmosphere. The proportion of remained sample
of weight after the reaction is determined as CRI.

For the optimum condition of operation of blast furnace, coke should posses a
certain extent of coke strength after reaction (CSR). If the CSR is too low, cokes
are easily burned off in the upper part and the degradation of coke besides the
tuyere will be accelerated so the pressure of tuyere can be changed as unstable. If
the CSR is too high, the reduction rate of iron ore will decrease and the remained
coke in the bottom will increase the viscosity of melted iron and slag. Therefore,
out of the range of CSR will be connected to the decrease in the production
amount of the iron ore, the increase in the cost of fuel and the fluctuation of the
quality of iron ore.

The CSR test is consecutively processed with the use of the reacted coke formed
in the CRI test. This is treated in a specially designed tumbler drum called an “I-
tumble tube” which allows the coke particles to tumble from one end to the other
for 600 revolutions during 30 minutes as shown in Fig.3.5. The CSR index is
determined by sieving and weighing the amount of cokes remained over a 10mm
size. This CSR value is intimately connected to optical isotropic characteristics on
coke.

99
 Figure 3.5: I-tumble tube for CSR test

3.2.4 Coke Texture

The coke texture classification used in this study, which is based on the standards
of POSCO, is illustrated in Table 3.6.

Table 3.6: Coke texture classification of POSCO

The comparative proportions of coke textures are able to present the coke
properties such as coke reactivity and strength. The increase in isotropic textures
of coke will make a coke weaker and reactive due to the higher surface area of

100
LMO available for reaction which is directly correlated with the intrinsic
reactivity of the molecule. On the other hand, the strength of coke increases with
anisotropy. The growths of anisotropic textures mainly depend on the fluidity of
the source of coal. Therefore, petrographic analysis of coke is one of the important
key to reveal the correlation between source of coal petrology and coke textures.

3.2.5 Physical structure (SEM/Optical Microscopy)

Scanning Electron Microscope (SEM) is essential to investigate the pore amount,

pore shape, pore distribution, pore size, wall of pore thickness, crack and inert.
Hitachi S3400-I in University of New South Wales (UNSW) Electron Microscope
Unit Analytical Centre is used to take images of the coke surfaces as shown in Fig.
3.6. Coke pieces of size between 5 to 10mm were embedded into epoxy resin,
grinded through four stages and polished through two stages with diamond paste.
All samples are gold coated for the electron conduction. Qualitative and semi-
quantitative analyses were performed on particles larger than 10 µm using an
accelerating voltage of 15 kV. Cross-sectional images of coke are observed to
show pore distribution and thickness of pore walls. Besides, the morphology of
the coke is also investigated via SEM three dimensional images. Small particles of
crushed coke sample in the size between 2 and 3 mm was investigated via SEM
and the various properties of pores were investigated. We examined the pore
shape, structure of pore across the radius and properties of pore such as viscid or
crumby. Thick pore wall of coke has high resistance against the breakage and high
rank of coking coal is advantageous to intensify the pore wall thickness of coke.

Figure 3.6: Scanning Electron Microscope; Hitachi S3400-I

101
4 ASSOCIATION BETWEEN COAL PROPERTIES
AND COKE STRENGTH

In this chapter, we will discuss about the effect of most important coal properties
i.e. coal rank and rheological properties of coals. The degree of coalification,
ascribed by the rank parameters such as volatile matter and reflectance of coal, is
known to play a significant role on the properties of the coke. High rank coals
have more aromatic carbons which are more stable than low rank coals while
branched aliphatic carbons which are rich in low rank coals react easily and have
more hydrogen. The rheological properties of coals such as Gieseler fluidity
(DDPM), caking index G (G-value), total dilatation (TD) and maximum thickness
of plastic layer are considered. A high rheological index represents igh degree of
agglomeration of coal particles with one another as well as with the inert
constituents so that high fluidity of coal determines the strong bonding tendency
during carbonization. The interaction of two single parameter will be ascribed by
the analysis of coke samples of binary coal blends. Seven Australian (AC01~07),
two Canadian (CA01~02), one Chinese (CC01), one Russian (RC01), six blended
coals (BC01~BC06) were used to prepare cokes. Specific properties of the selected
coals were given in Table 3.1 and Table 3.2.

4.1 Effect of Coal Rank

Coal rank is one of the vital factors to affect coke strength due to the intimate
relationship between carbon density of coal and coal rank. Coal rank, in general,
is expressed by the reflectance or the volatile matter. Often coals form single
geological source indicate increasing coke strength in accordance with the
increase in coal rank. Figure 4.1 shows correlations between cold and thermal
strength of coke with mean reflectance as well as volatile matter. This graph
shows the increasing tendency of cold and thermal strength of coke by the
increasing mean reflectance of vitrinite(a,b) while illustrating the decrease in coke
strength by the increase in volatile matter which means increase in fixed carbon in

102
coking coal (c,d). These figures show that thermal strength of cokes generally
increase with mean reflectance (b) at least among coals with reflectance values
less than 1.4 with the exception of two coals.

Table 3.1 shows that the Drum Index of coke AC05 (79.09) is higher than coke
AC07 (69.72) even the parent coals have similar range of logarithmic maximum
fluidity (LMF) 3.75 and 3.69 respectively. This highlights the predominance of
effect of coal rank on coke strength when fluidity is of the same order. On the
other hand, the CSR value of coke AC05 (55.19) is much higher compared to the
CSR value of coke of AC07 (17.81). This can be explained on the basis of higher
vitrinite reflectance of coal AC05 (0.90) compared to the vitrinite reflectance of
AC07 (0.72).

Higher cold and thermal strength of coke CA02 (80.79 and 69.68) compared to
those of coke AC04 (72.58 and 30.92) can be attributed to higher vitrinite
reflectance values of parent coals 1.14 and 0.94 respectively as both have similar
fluidity range of the order of two. Figure 4a and 4b show unexpected low values
of cold strength and CSR values of cokes CA01 and RC01 respectively, which
happen to be made of coals with extremely low fluidity. It seems the intensity of
the effect of rank on coke strength is less significant for coals with extremely low
fluidity.

Figure 4.1: Coke strengths vs. mean reflectance as well as volatile matter of coals

103
Figure 4.2 plots coke strength values against mean reflectance as well as volatile
matter. Unlike the trend observed in single coal cokes, in case of the binary blend,
coke strength did not show any trend with the coal rank (Fig. 4.2). This can be
related to comparative narrow range of the rank of the blended coals (0.85 to 1.15).
It may be noted that binary blends were selected to have similar rank range but
different vitrinite composition and levels.

Figure 4.2: Coke strengths vs. mean reflectance and volatile matter of binary
blends

Reflectogram of vitrinite of coal blends are illustrated in Fig. 4.3 to investigate the
overlapping effect of coal rank in binary blends. In binary blends, the
reflectogram of vitrinite couldn’t illustrate any interaction of reflectance between
coal rank and coke strength. For instance, blend BC02 coke (f), 50% mixture of
AC07 and AC08 doesn’t show 1.0 of reflectance value due to its absence of 1.0

104
reflectance in original coal (a, b). Thus, rank parameter of coals seemed to be
independent factors in binary blends.

(a) (b)
AC07 AC08

40 40

30 30

20 20

10 10

0 0
0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88 0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88

(c) (d)
CA01 RC01

40
40
30 30
20 20
10 10

0
0
0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88 0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88

(e) (f)
BC01 BC02

40 40

30 30

20 20

10 10

0 0
0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88 0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88

(g) (h)
BC03 BC04

40 40

30 30

20 20

10 10

0 0
0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88 0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88

(i) (j)
BC05 BC06

40 40

30 30

20 20

10 10

0 0
0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88 0.53 0.68 0.83 0.97 1.13 1.28 1.43 1.58 1.73 1.88

Figure 4.3: Percentage distributions of vitrinite vs. reflectance values of

component coals and blends

The results suggest that coke strength values generally show an increase in
accordance with the increase in rank of all the tested single coals particularly the

105
CSR values. However, high end of vitrinite reflectance, the CSR values become
steady. Coal ranks dominate the coke strengths particularly when fluidity of
coking coals is of the similar order. At extremely low fluidity, the effect of coal
rank on coke strength becomes less significant. The exact reason is not clear but
could be related to differences in parent coal geology such as the Canadian and
Russian coals used in present study coke strength of the selected binary coal
blends could not be related to rank or the geological origin. .

4.2 Effect of Coal Rheology

The rheological properties of coals are commonly measured in terms logarithmic

maximum fluidity (LMF), total dilatation (TD), free swelling index (FSI) and G-
factor. Figure 4.4 shows that coke strength values (DI and CSR) increases with
the increase in fluidity (FSI) as the fused coal particles can agglomerate and stick
together well. Strong bonds of carbon matrix are requiring sufficient plasticity and
expansion. The FSI seems to have better correlations with drum index and CSR (e)
values when compared to other rheological parameters. A better FSI correlation in
case of single coal can be related to better scale of reference profiles of inherent
swelling properties of individual coals. Notably, G-Factor could display any
correlation with coke strength (g, h) values of single coal cokes which suggests
that it is an indicator of degree of mixture of reactive and inerts and could be more
suitable for blends rather than individual coals.

106
 Figure 4.4: Coke strength vs. Rheological parameters of single coal

In binary blends coke strengths did not show any clear correlations with the
rheological parameters including FSI. Fluidity of binary blend coals was limited
to short range of Gieseler fluidity values (1.5 – 2.75) as well as total dilation
values 5 -50) compared to wide range of all tested coals.

107
 Figure 4.5: Coke strength vs. Rheological parameters of blended coals

The plastic range of coal blends is not symmetrical along temperature because the
blending can modify the inherent rheology of individual component coals
(Sakurovs, 2003). The fluidity ranges of coal blends (b, c, e, f, h, i) and those of
their original single coals (a, d, g) are represented in Fig. 4.6. In these
comparisons, change of the rheological graphs caused by blending is not apparent
because blending coal’s plastic ranges (AC08, CA01, and RC01) are included in

108
the plastic range of reference coal (AC07). Increased drum index of blend coke
suggests high overlapping or modification of rheological properties of individual
coals after blending. Specifically, the average drum index of BC series cokes is
80.62 although the average value of the individual component coals (AC07, AC08,
CA0, and RC01) is 71.46.

(a) (b) (c)

AC07 AC08 BC01 (AC07 75% + AC08 25%) BC02 (AC07 50% + AC08 50%)

4
4 4
3 3 3
lg MF

lg MF

lg MF
2 2 2
1 1 1
0 0 0
300 350 400 450 500 550 300 350 400 450 500 550 300 350 400 450 500 550
Temperaure (℃) Temperaure (℃) Temperaure (℃)

(d) (e) (f)

AC07 CA01 BC03 (AC07 75% + CA01 25%) BC04 (AC07 50% + CA01 50%)

4 4 4
3 3 3
lg MF

lg MF
lg MF

2 2 2
1 1 1
0 0 0
300 350 400 450 500 550 300 350 400 450 500 550 300 350 400 450 500 550

Temperaure (℃) Temperaure (℃) Temperaure (℃)

(g) (h) (i)

AC07 RC01 BC05 (AC07 75% + RC01 25%) BC06 (AC07 50% + RC01 50%)

4 4
4
3 3 3
lg MF

lg MF
lg MF

2 2 2
1 1 1
0 0 0
300 350 400 450 500 550 300 350 400 450 500 550 300 350 400 450 500 550
Temperaure (℃) Temperaure (℃) Temperaure (℃)

Figure 4.6: Fluidity ranges of blended coal and their original coals

4.3 Effect of Coal Rank and Rheology on Coke

Strength of Blends

In all the tested binary blends, the coke strength variation is narrow. This
complexity to investigate the effect of individual coal properties on coke strength
of binary blend requires considering the interaction of rheological factors and coal
rank. A statistical analysis of drum index and CSR as a function of rank and

109
rheological indicator was carried out using MINITAB. Regression equations are
as follows.

1) DI 15015 = 30.5 + 38.2 x Mean Reflectance + 2.75 x FSI R2 = 0.66

150
2) DI 15 = 107 - 1.12 x Volatile Matter (%) + 0.256 x TD R2 = 0.52
3) CSR = 167 - 94.2 x Mean Reflectance - 22.2 x LMF R2 = 0.67
4) CSR = 113 - 3.62 x Volatile Matter (%) + 5.01 x FSI R2 = 0.88

The equations of DI 15015 and CSR with correlation factor over 50% were selected
to estimate the optimum range of required coal properties for high strength of
coke.

DI 150/15(%) 대 Mean Reflectance의 등고선도, FSI DI 150/15(%) 대 Volatile

DI 150/15(%) Matter(%)의
대 Volatile 등고선도,
Matter(%)의 Total
등고선도, TotalDilatation
Dilatation
33 DI 33 DI
DI
150/15(%) 150/15(%)
1.10 150/15(%)
< 78
< 78
32 78 –< 7978
32
78 – 79 78 –– 8079
79
79 – 80 79 –– 8180
Volatile Matter(%)

80
1.05
Mean Reflectance

80 – 81 31 81
80 –– 8281
Volatile Matter(%)

81 – 82
31 81 >– 8282
> 82 > 82
30
1.00
30
29

0.95 29
28

0.90 28 10 15 20 25 30 35 40 45
Total Dilatation

2 3 4 5 6 10 15 20 25 30 35 40 45
FSI
CSR(%) 대 Mean Reflectance의 등고선도, LMF Total Dilatation
CSR(%) 대 Volatile Matter(%)의 등고선도, FSI
33
CSR(%) CSR(%)
< 24 < 24
1.10
24 – 27 24– 27
32
27 – 30 27– 30
30 – 33 30– 33
1.05
Volatile Matter(%)

33– 36
Mean Reflectance

33 – 36
31 > 36 > 36

1.00
30

0.95 29

0.90 28

1.6 1.8 2.0 2.2 2.4 2.6 2 3 4 5 6

LMF FSI

Figure 4.7: Contour maps of DI and CSR with coal rank and fluidity parameters
of binary blends

In Fig. 4.7, four contour maps of coke qualities and coal properties exhibit the
optimum ranges of coal properties such as mean reflectance from 1.0 to 1.15,
volatile matter from 26 to 27%, LMF from 1.6 to 1.8, total dilatation from 22 to
30% and FSI from 2.5 to 4.5. The optimum ranges of coal rank parameters is
consistent with past experimental reported data (Zhang, 2004) such that the
identified rank range is commonly used to produce high strength industrial coke.

110
However, the estimated optimum fluidity ranges are only effective in six restricted
blends and slightly different than an optimum fluidity range preferred in industrial
practice.

4.4 Coke Micro-texture

Table 4.1 provides the summary of BC series coke strength and microscopic
textures of all tested blend cokes as well individual cokes of coals AC07, AC08,
CA01 and RC01 POSCO Classification of coke textures is used quantify the
proportions of anisotropic and isotropic textures (Table 3.6). In addition, a coke
factor index is also calculated using following equation which presents the coke
anisotropy.

* Coke Factor = {(Fine Mosaic % x 2.5) + (Medium Mosaic % x 3.0) + (Coarse

Mosaic % x 4.0) + (Fibrous % x 3.5) + (Leaflet % x 5.0) + (Iso-Vitrinite % x 1.0)
+ (Aniso-Fusinite % x 2)} / (Aniso Texture % + Iso-Vitrinite % + Aniso-
Fusinite %)

Table 4.1: Microscopic texture data of individual and blend cokes

111
To examine the correlations between microscopic textures of coke and its strength,
coke strength is assumed to be function of anisotropic, isotropic texture and coke
factor. Fig. 4.8 shows that the CSR of cokes increase with increasing anisotropic
texture (b, f) in accordance with past studies (Vogt and Depoux, 1990). The CSR
decrease with the increase in isotropy (d) a better correlation and showed a better
correlation with coke factor (f). The drum index of cokes showed no correlations
with micro-textures of the all the tested cokes (a, c, e).

Figure 4.8: Coke strength vs. micro-texture of all the tested binary coke blends

112
Fig. 4.9 shows optical images of cokes based on polarized light from Leica
DM4500P microscope. Leaflet type is a typical fibrous type which is known to
have positive influence on coke strength. This texture present a bright green
polarized light and reflectance range from 1.0 to 1.2. RC01 coke shows the
highest proportion of leaflet texture (Table 4.1) and hence high drum index (a).
Fibrous type of coke shows the brightest reflectance value of more than 1.2 with
elongated iso-chromatic areas. This texture often positioned curved area of pores
representing that these textures mainly participate in pore generation. Therefore,
coke from CA01 which is the most abundant fibrous textures shows the weak
value of drum index in spite of high ratio of anisotropic coke textures (b). Coarse
mosaic of coke is distributed in the region of reflectance value from 0.9 to 1.0 and
is the mostly affecting mosaic structure on coke strength. Sometimes, high
proportion of coarse mosaic can be the indicator of high strength of coke. The
BC01 coke shows the high drum index due to the high proportion of coarse
mosaic in spite of lower percentage of anisotropic texture (c). Medium mosaic
texture has the middle size ranging from 1.8 to 5 μm which is recognized as the
medium size of fine and coarse mosaic. This texture represents the reflectance
range from 0.8 to 0.9. BC04 coke has the highest proportion of medium mosaic
texture.

Fine mosaic is composed of millet like grains with the finest sizes and reflectance
range from 0.75 to 0.85. This texture can be further classified into very fine and
general fine mosaic. This texture is also often observed in the coke from high
proportion of semi-soft coking coal. Iso-vitrinite is the standard texture of weak
strength of coke which has numerous pores as shown in Fig. 4.9 (f). These micro
pores which are the evidences of volatilized matter broaden the reactive surface
areas of coke so that high proportion of iso-vitrinite enhances the reactivity of
coke but deteriorates the CSR of coke. This texture also has weak reflectance
value of less than 0.8.

113
 Figure 4.9: Microscopic textures of cokes in binary blends and their original
coals

Figure 4.10 shows SEM images of coke cross-section. Three cokes, AC07, 08,
CA01 (a,c,d) show similar DI from 66 to 70 but AC07 shows extremely low CSR
(17.8). Coke from AC07 has high fluidity (LMF; 3.7) of its original coal and
lower rank. Coke AC07 shows high proportions of big pores which can be
attributed to excessive fluidity (a). Big pores also have many micro pores inside

114
which contribute to increased reactive surface and hence weaken the CSR values
(b). Besides, coke AC07 also shows molten carbon phases due to its high fluidity
while coke AC08 (Rm; 1.3), CA01 (Rm; 1.5) show significant low amount of
fused carbon matter (c, d).

Figure 4.10: Cross sectional SEM images of cokes

Figure 4.11 shows morphology of cokes. Coke CA01 (Rm; 1.5) shows the crumby
surface with small pores which is related to indicative of insufficient melting due
to high rank and low fluidity (a). Coke AC01 (b) shows fine and well developed
structure due to the combination of high rank and some extent of fluidity. Coke
AC06 shows the viscid structures with big pores (c) and its drum index is in
medium range (DI; 74.1) which was caused by the medium range of fluidity.
However, its CSR (21.5) is extremely low because of big pores, which cab related
to low rank of parent coal. Coke AC05 has viscid surface due to high fluidity (d).
It’s CSR (55.2) value is medium due to the numerous pores. Two blend cokes

115
BC02 and BC05 show viscid– crumby structures, which are typical characteristics
of high DI and low CSR cokes (e, f).

Figure 4.11: SEM images of coke morphology

Coke texture and morphology is influenced by fluidity and rank. High CSR cokes
were characterized with high anisotropic textures. Coke Factor indicative of
anisotropic texture showed a better correlation with the CSR values.

116
5 EFFECT OF VITRINITE TYPES ON COKE
STRENGTH

POSCO classification of vitrinite subgroups is consistent with the Stach (1982).

As far as the compatible is concerned telinite and collinite follows similar
behaviour as those of telocollinite and pseudovitrinite in other standards. Telinite
of AC07 which has comparatively high proportion of telinite than collinite of
POSCO standards shows the typical appearance of telocollinite mixed with
telinite and collinite (5.1a). The collinite image of AC08 which include higher
contents of collinite than telinite shows clean and high reflectance of surface
which is the characteristics of pseudovitrinite (5.1b). During coalification, volatile
matter of collinite is oxidized and small pores are generated as seen in CA01 coal
sample which has higher collinite content compared to coal AC08 (51c).
Excessively higher portion of collinite of RC01 shows the big pores and slitted
structures which is typical shape of pseudovitrinite (5.1d). In POSCO standards,
vitrodetrinite is not classified due to its small proportion and low significance.
Until now, the effect of sub-vitrinite group remained as black-box and its
microscopic mechanism during carbonization is still unknown. Therefore, in
following section, effect of interaction of various vitrinite sub-groups on coke
strength is examined.

Figure 5.1: Microscopic images of vitrinite maceral in single coal

117
 5.1 Effect of Vitrinite Composition

Fig. 5.2 plots the portions of telinite and collinite as eight coal properties namely
mean vitrinite reflectance, volatile matter, carbon content, hydrogen content, L-
MF, total distillation, FSI and G-factor. The trend lines in plots are fitted by
MINITAB software for all eleven coals to investigate the correlations between
vitrinite and coal properties. The carbon, hydrogen contents and rank parameters (a,
b, c, d) show reasonable correlations whereas rheological parameters (e, f, and g,
h) did not show any correlations.

118
 Figure 5.2: Vitrinite submacerals vs. fluidity parameters of coals

Coal rank parameter such as Rm and VM show better correlations sub-vitrinite

contents. It is matter of course that rank parameters show strong correlations
because the collinite and telinite can be distinguished by the degree of oxidation.
Further oxidized collinite possess higher carbon contents and lower hydrogen
contents which is typical of high rank. Hydrogen contents of coals decrease with
the increase in collinite contents and decrease in telinite contents. Collinite and
telinite are distinguished by the degree of oxidation so that collinite have similar
characteristics of pseudovitrinite which is slightly further oxidized than
telocollinite. Therefore, due to relatively higher oxidation level of collinite than
telinite, its carbon contents increase with the increase in collinite proportion of
vitrinite in single coals. However, rheological properties of coals show no
correlations to sub-vitrinite contents, therefore a sub-vitrinite index may only be
suitable for assessing coke strength of individual coals.

119
 5.2 Effect of Sub-maceral on Coke Strength

Table 5.1 shows new vitrinite indicators to predict coke strength by combining
the basic indicators. Mean reflectance and FSI are selected as basic parameters of
new vitrinite indicators to represent rank and rheology influence. Particularly, FSI
was chosen among rheological parameters due to its high correlation with sub-
vitrinite content of coals. Other indicators of sub-vitrinite are determined to use as
basic indicators such as vitrinite, collinite and telinite.

Table 5.1: Vitrinite indicators of single coals and their cokes strengths

Fig. 5.3 plots sub-vitrinite indicators of coals with coke strengths. Generally, coke
strengths increase with increase in the proportions of collinite and decrease in the
proportion of telinite (5.3a, 5.3b). The CSR appears to show better correlation
compared to the drum index. Unexpectedly, higher rank CA01 and RC01 coals
shows low coke strength despite containing high proportions of collinite (5.1a,
5.1b) which can be attributed to insufficient rheological properties of coal. Due to
high proportion of collinite in vitrinite, collinite - telinite is plotted with coke
strength and shows a reasonable correlation (5.1c, 5.1d).

Collinite percentages in vitrinite (e, f), telinite percentages in vitrinite (g, h),
(collinite - telinite) percentages in vitrinite (i, j), (collinite - telinite) / mean
reflectance (k, l) and (collinite - telinite) / FSI (m, n) are also related with coke
strength. Among these indicators, collinite percentages in vitrinite (e, f), telinite
percentages in vitrinite (g, h) and (collinite - telinite) percentages in vitrinite (i, j)

120
are showing similar degree of correlation with drum index as well as CSR. The
(collinite - telinite) / FSI (m, n) also illustrates higher value of correlations but
comparatively lower than sub-vitrinite percentages in vitrinite. These figures show
that (collinite – telinite) / mean reflectance (k, l) provides the best correlation with
coke strength.
(a) (b)

(a) (b)

CA01
RC01

CA01
RC01
(c) (d)

(c) (d)

(e) (f)
(e) (f)

(g) (h)

(g) (h)

121
 (i) (j)

(k) (l)

(m) (n)

Figure 5.3: Correlations between the vitrinite indicators and coke strength for
single coals
Due to the comparative higher oxidation level of collinite than telinite, its rank
and carbon contents increase with the increase in collinite proportion of vitrinite
in single coals. Coke strengths increase with the increase of the proportions of
collinite whereas decrease of the proportion of telinite. The CSR shows slightly
better correlations with these indicators compared to drum index with the
exception of higher rank of CA01 and RC01. This would be related to lack of
sufficient rheological properties or possibly due to mineral effect. The collinite –
telinite) / mean reflectance is seen to provide the best indicator of coke strength of
single coals.

122
6 EFFECT OF BLENDING ON COKE STRENGTH

6.1 Vitrinite Composition and Coal Property

Six binary blended coals (BC01~06) are selected and all plots are fitted by
MINITAB software. Fig. 6.1 shows that rank and fluidity parameters provide better
correlation with coke strength of blends when compared to single coals. It may be
noted that properties of binary blend coals have a narrow range. Coal rank parameters
including carbon, hydrogen contents and vitrinite reflectance (Figure 6.1a, b, c, d)
show a good correlation compared to rheological parameters (Figure 6.1e, f, g).

123
Figure 6.1: Correlations between the indicators of vitrinite and coal parameters of
binary blends.

In binary blends, the correlation trend of subvitrinite parameters with coal

parameters is better when compared to single coal correlations due to greater
degree of oxidation of collinite maceral. However, rheological properties were
decreased by increase in collinite due to decreasing hydrogen contents. These
results suggests that sub-vitrinite content of coals are intimately related to coal
properties, particularly rank for both single as well as binary coal blends.

6.2 Vitrinite Composition and Coke Strength

Table 6.1: Vitrinite indicators of coal blends and their coke strengths

124
Unlike ns between coke strengths and sub-vitrinite composition in binary coal
blends while there was relatively higher correlations observed in the case of single
coals. Thus, it is hard to distinguish the contribution of collinite and telinite on
coke strength of the tested binary blends. This may be attributed to non-linear
changes of thermo-plasticity and possibly to some impact of mineral interactions
during carbonization.

125
 Figure 6.2: Vitrinite macerals vs. Coke strength of binary blend
listed in Table 6.1

6.3 Implication on Industrial Coke Strength

Coal properties particularly rank was correlated with sub-vitrinite new sub-
vitrinite index showed good correlation with coke strength in case of single coals
but does not relate with coke strength of all binary cokes tested. Therefore, the
effect of vitrinite compositions on coke strengths needs to be validated through
industrial tests. Mean reflectance, total dilation, vitrinite, collinite and telinite are
also suggested as suitable indicators to assess effect of sub-vitrinite macerals.
Majority of indicators were correlated with coke strength of industrial coke
blend samples for one year operation. . Figure 6.3 shows correlations between
coke strengths and sub-vitrinite composition in industrial cases. All coal
parameters were kept same except sub-vitrinite composition.

Figure 6.3 shows that the correlation between coke strengths and sub-vitrinite
composition in industrial samples is poor as noted in case of binary coke trends.

126
127
Figure 6.3: Vitrinite macerals vs. Coke strength of the industrial test

128
7 CONCLUSIONS

The petrography of a number of coals was conducted to examine the effect of coal
blending on coke strength parameters with emphasis to clarify the role of vitrinite
sub-maceral. Following conclusions were made.

 The study demonstrated the predominant effect of coal rank on coke

strength compared to coal rheological properties. In general, coke strength
including the CSR of single coals is shown to increase with increasing coal rank
with few exceptions.

 None of the rhelogical parameters of single colas showed a clear

correlation with coke strength except free swelling index which provided the best
association with both cold and coke strength after reaction. A better correlation of
free swelling index with coke strength can be related to better representation of
inherent swelling properties of coal as a consequence of better calibration of
reference profile.

 A coal indicator based on the rank and rheological properties was devised
to relate with the drum index and the CSR values. The statistical analysis
identified the optimum coal mean reflectance range (1.0 to 1.15) and volatile
matter range (26 to 27%) to design high strength coke which is in agreement with
current industrial practice of high quality coking coals.

 Micro-texture of coke particularly anisotropic texture is shown to be

strongly influenced by coal properties such that high CSR cokes displayed high
anisotropic texture. The Coke Factor, indicative of anisotropic texture was
correlated coke strength (CSR). Low CSR cokes were found to be associated with
high frequency of pores as well as large pore size. The study further clarified the
association of coal rank and fluidity on morphology and pore structure of cokes.

129
 The composition of vitrinite maceral types varied with coal types such that
vitrinite reflectance and volatile matter of coals could be related to the sub-
vitrinite contents. This can be attributed to better identification of sub maceral
components due to differences in the oxidation degree of collinite and telinite sub-
macerals. The variation in rheological properties of coals could not be
distinguished by their sub-vitrinite contents. The collinite content of coals is found
to increase with coal rank and carbon contents whereas rheological properties and
hydrogen contents decreased. A new coal parameter based on the collinite,
tellinite content and vitrinite reflectance is proposed such that the ratio of
differences of collinite and telinite contents to mean reflectance of coals showed a
strong association with coke strength.

 The proposed new coal parameter based on collinite and tellinite could not
be related to the coke strength of the tested binary blends. The study further
showed that new sub-maceral index could not be applied to predict coke strength
data of industrial coke blends samples. It is hard to distinguish the contribution of
collinite and telinite on coke strength.

 The proposed sub-vitrinite approach seems to be suitable means to predict

coke strength in single coal but may be limited in case of high rank and low
fluidity coals in blends. This may be attributed to non-linear modification of
thermo-plasticity of coal blends during carbonization. Future research may focus
to clarify the factors affecting the thermoplastic nature of coal bends including the
role of minerals. Improved understanding of effect of coal properties on thermo-
plastic nature of blends may improve the application of sub-vitrinite index for
industrial application.

130
REFERENCES

ÁLVAREZ, R., D EZ, M. A., BARRIOCANAL, C., D AZ-FAES, E. &

CIMADEVILLA, J. L. G. 2007. An approach to blast furnace coke quality
prediction. Fuel, 86, 2159-2166.
BENEDICT, L. G. & THOMPSON, R. R. 1980. Coke/carbon reactions in the
study of factors affecting coke quality. International Journal of Coal
Geology, 1, 19-34.
BERTLING, H. 1999. Coal and coke for blast furnaces. Isij International, 39,
617-624.
BROWN, H. R., HESP, W. R. & TAYLOR, G. H. 1966. CARBONS OBTAINED
BY THERMAL AND CATALYTIC CRACKING OF COAL TARS.
Carbon, 4, 193-&.
BURGESS, J. M. 1985. Fuel combustion in the blast furnace raceway zone.
Progress in Energy and Combustion Science, 11, 61-82.
CROSS, W. B. 1994. The future and the European steel industry and its demand
for coal, Mexborough.
DAVIS, A., SPACKMAN, W. & GIVEN, P. H. 1976. The Influence of the
Properties of Coals on Their Conversion into Clean Fuels. Energy Sources,
Part A: Recovery, Utilization, and Environmental Effects, 3, 55 - 81.
DIESSEL, C. F. K. 1983. Carbonization reactions of inertinite macerals in
Australian coals. Fuel, 62, 883-892.
DIEZ, M. A., ALVAREZ, R. & BARRIOCANAL, C. 2002. Coal for
metallurgical coke production: predictions of coke quality and future
requirements for coke making. International Journal of Coal Geology, 50,
389-412.
DUVAL, B., GUET, J. M., RICHARD, J. R. & ROUZAUD, J. N. 1988. COKE
PROPERTIES AND THEIR MICROTEXTURE .3. 1ST RESULTS
ABOUT RELATIONSHIP BETWEEN MICROTEXTURE AND
REACTIVITY OF SOME COKES. Fuel Processing Technology, 20, 163-
175.
ESTERLE, J. 2007. Coalification Processes and Coal Rank, Grade and Type.

131
FITZGERALD, D. 1956. The Kinetics of Coal Carbonization in the Plastic State.
Fuel, 35.
GHOSH, A. 2002. AuroView 2000 - An optical image processing system in rank
and grade analysis of coal. Journal of the Geological Society of India, 60,
143-150.
GRANSDEN, J. F., JORGENSEN, J. G., MANERY, N., PRICE, J. T. & RAMEY,
N. J. 1991. APPLICATIONS OF MICROSCOPY TO COKE MAKING.
International Journal of Coal Geology, 19, 77-107.
GRIGORE, M., SAKUROVS, R., FRENCH, D. & SAHAJWALLA, V. 2006.
Influence of mineral matter on coke reactivity with carbon dioxide. Isij
International, 46, 503-512.
GRIGORE, M., SAKUROVS, R., FRENCH, D. & SAHAJWALLA, V. 2007.
Effect of carbonisation conditions on mineral matter in coke. Isij
International, 47, 62-66.
GUPTA, R. 2007. Advanced coal characterization: A review. Energy & Fuels, 21,
451-460.
GUPTA, S., FRENCH, D., SAKUROVS, R., GRIGORE, M., SUN, H., CHAM,
T., HILDING, T., HALLIN, M., LINDBLOM, B. & SAHAJWALLA, V.
2008. Minerals and iron-making reactions in blast furnaces. Progress in
Energy and Combustion Science, 34, 155-197.
GUPTA, S. & SHEN, F. 2009. Advanced Characterization of Coal Petrography
for Cokemaking. Sydeny: Centre for Sustainable Materials Research &
Technology University of New South Wales.
HAYS, D., PATRICK, J. W. & WALKER, A. 1976. Pore structure development
during coal carbonization. 1. Behaviour of single coals. Fuel, 55, 297-302.
HUANG, H., WANG, K., BODILY, D. M. & HUCKA, V. J. 1995. Density
Measurements of Argonne Premium Coal Samples. Energy & Fuels, 9,
20-24.
HUTNY, W. P., LEE, G. K. & PRICE, J. T. 1991. FUNDAMENTALS OF COAL
COMBUSTION DURING INJECTION INTO A BLAST-FURNACE.
Progress in Energy and Combustion Science, 17, 373-395.
JAPAN, T. I. A. S. I. O. 1979. Handbook of Iron and Steel: The Manufacture of
Iron and Steel, Tokyo, Maruzen.

132
KIDENA, K., KATSUYAMA, M., MURATA, S. & NOMURA, M. 2002. Study
on plasticity of maceral concentrates in terms of their structural features.
Energy & Fuels, 16, 1231-1238.
KIDENA, K., MURATA, S. & NOMURA, M. 1996. Studies on the chemical
structural change during carbonization process. Energy & Fuels, 10, 672-
678.
KIM, M. J. & OH, M. 2004. Evaluation of Coal Viscosity in Thermoplastic State.
Korean Chemical Engineering Research, 42, 696-705.
KREVELEN, D. W. 1993. COAL: TYPOLOGY-PHYSICS-CHEMISTRY-
CONSTITUTION, Amsterdam, ELSEVIER SCIENCE PUBLISHERS B.V.
KRUSZEWSKA, K. J. 1998. The reactivity of pseudovitrinite in some coals. Fuel,
77, 1655-1661.
LEWIS, I. C. 1982. Chemistry of carbonization. Carbon, 20, 519-529.
LIN, M.-F. & HONG, M.-T. 1986. The effect of coal blend fluidity on the
properties of coke. Fuel, 65, 307-311.
LOISON, R., FOCH, P. & BOYER, A. (eds.) 1989. Coke Quality and Production,
Paris: Butterworth & Co.
MAROTO-VALER, M. M., TAULBEE, D. N., ANDRESEN, J. M., HOWER, J.
C. & SNAPE, C. E. 1998. The role of semifusinite in plasticity
development for a coking coal. Energy & Fuels, 12, 1040-1046.
NAKAMURA, N. 1981. RECENT PROBLEMS IN IRON-MAKING
TECHNOLOGY. Transactions of the Iron and Steel Institute of Japan, 21,
75-91.
NEAVEL RICHARD, C. 1981. Coal Structure and Coal Science: Overview and
Recommendations. Coal Structure. WASHINGTON, D. C.: AMERICAN
CHEMICAL SOCIETY.
NOMURA, S., ARIMA, T. & KATO, K. 2004. Coal blending theory for dry coal
charging process. Fuel, 83, 1771-1776.
O'BRIEN, G., JENKINS, B., ESTERLE, J. & BEATH, H. 2003. Coal
characterisation by automated coal petrography. Fuel, 82, 1067-1073.
RAGAN, S. & MARSH, H. 1980. A CRITIQUE OF INDUSTRIAL METHODS
OF MEASUREMENT OF STRENGTH OF METALLURGICAL COKE.
Journal of Physics D-Applied Physics, 13, 983-&.

133
READ, R. B. 1982. THE INFLUENCE OF COMPOSITION, RHEOLOGY, AND
PRESSURE ON THE FLUID BEHAVIOR OF BITUMINOUS COALS.
Ph.D Dissertations & Theses, University of Kentucky
READ, R. B., REUCROFT, P. J. & LLOYD, W. G. 1985. The effects of pressure
on the isothermal fluid behavior of bituminous coals. Fuel Processing
Technology, 11, 133-151.
SAKUROVS, R. 2003. Interactions between coking coals in blends. Fuel, 82,
439-450.
SENGUPTA, A. N. 2002. An assessment of grindability index of coal. Fuel
Processing Technology, 76, 1-10.
SHARMA, R., DASH, P. S., BANERJEE, P. K. & KUMAR, D. 2005. Effect of
coke micro-textural and coal petrographic properties on coke strength
characteristics. Isij International, 45, 1820-1827.
SHINN, J. H. 1984. From coal to single-stage and two-stage products: A reactive
model of coal structure. Fuel, 63, 1187-1196.
SOLOMON, P. R., BEST, P. E., YU, Z. Z. & CHARPENAY, S. 1992. An
empirical model for coal fluidity based on a macromolecular network
pyrolysis model. Energy & Fuels, 6, 143-154.
SPIRO, C. L. & KOSKY, P. G. 1982. Space-filling models for coal. 2. Extension
to coals of various ranks. Fuel, 61, 1080-1084, IN1-IN2.
STACH, E. & MURCHISON, D. 1982. Stach's Textbook of Coal Petrology,
Berlin, Gebruder Borntraeger.
STREZOV, V., LUCAS, J. A. & WALL, T. F. 2005. Effect of pressure on the
swelling of density separated coal particles. Fuel, 84, 1238-1245.
SUGGATE, R. P. 1998. Analytical variation in Australian coals related to coal
type and rank. International Journal of Coal Geology, 37, 179-206.
TAKANOHASHI, T., YOSHIDA, T., IINO, M., KATO, K., FUKADA, K. &
KUMAGAI, H. 2005. Structure and Thermoplasticity of Coal. Structural
Changes of Coal Macromolecules during Softening. Nova Publishers, Inc.
VANNIEKERK, W. H. & DIPPENAAR, R. J. 1991. BLAST-FURNACE COKE
- A COAL-BLENDING MODEL. Journal of the South African Institute of
Mining and Metallurgy, 91, 53-61.

134
VASKO, F. J., NEWHART, D. D. & STRAUSS, A. D. 2005. Coal Blending
Models for Optimum Cokemaking and Blast Furnace Operation. The
Journal of the Operational Research Society, 56, 235-243.
VASSILEV, S. V., KITANO, K., TAKEDA, S. & TSURUE, T. 1995.
INFLUENCE OF MINERAL AND CHEMICAL-COMPOSITION OF
COAL ASHES ON THEIR FUSIBILITY. Fuel Processing Technology,
45, 27-51.
VOGT, D. & DEPOUX, M. 1990. COKE REACTIVITY PREDICTION BY
TEXTURE ANALYSIS. Fuel Processing Technology, 24, 99-105.
WALKER, R. 2005. Insights into the coking behavior of southern Indiana coals:
Bulk coal properties and individual maceral chemistries Ph.D dissertation,
Indiana University.
WARD, C. R. 2002. Analysis and significance of mineral matter in coal seams.
International Journal of Coal Geology, 50, 135-168.
ZHANG, Q., WU, X., FENG, A. & SHI, M. 2004. Prediction of coke quality at
Baosteel. Fuel Processing Technology, 86, 1-11.
ZUBKOVA, V. V. 2005. Some aspects of structural transformations taking place
in organic mass of Ukrainian coals during heating. Part 1. Study of
structural transformations when heating coals of different caking capacity.
Fuel, 84, 741-754.

135