You are on page 1of 201

The University of New South Wales

Faculty of Science

School of Materials Science and Engineering

Interfacial Phenomena and Dissolution of


Carbon from Chars into Liquid Iron During
Pulverised Coal Injection in a Blast Furnace

A Thesis in
Materials Science and Engineering
by
Fiona McCarthy

Submitted in Partial Fulfilment


of the Requirements for the Degree of
DOCTOR OF PHILOSOPHY
August 2004
CERTIFICATE OF ORIGINALITY

I herby declare that this submission is my own work and to the best of my knowledge it
contains no materials previously published or written by another person, nor material
which to a substantial extent has been accepted for the award of any other degree or
diploma at UNSW or any other educational institution, except where due
acknowledgement is made in the thesis. Any contribution made to the research by
others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in
the thesis.

I also declare that the intellectual content of the thesis is the product of my own work,
except to the extent that assistance from others in the projects design and conception or
in style, presentation and linguistic expression is acknowledged.

Signature……………………………..
Fiona Mccarthy

ii
ACKNOWLEDGMENTS

The successful completion of this project was dependent on the project supervisors
Professor Veena Sahajwalla, Mr. John Hart and Dr. Noel Simento. Their expert
guidance, supervision, and encouragement are very much appreciated.

I also wish to acknowledge the support of the Cooperative Research Centre for Coal in
Sustainable Development, which is funded in part by the Cooperative Research Centres
Program of the Commonwealth Government of Australia.

I would like to thank Mr. N. Saha-Choudury for his assistance with experimental
apparatus and my colleagues and friends at UNSW who provided aid, help, and support
throughout the entire project.

I would like to acknowledge Dr. Haiping Sun and Dr. Rita Khanna for their help with
the development of the interfacial model. Finally, I would like to express my
appreciation to Dr. Rita Khanna, whose support, suggestions and help during the writing
of this thesis are deeply appreciated.

Finally, I would like to thank Geoff and my parents for their constant encouragement,
patience and understanding. I could not have completed this project without their
support.

iii
ABSTRACT

Iron/carbon interactions have long been the subject of great interest to ironmakers and
have found application in a large number of ferrous metallurgical processes. With
specific reference to pulverised coal injection in blast furnaces, which is a growing field
due to environmental and economic concerns, this project is focussed on achieving a
comprehensive understanding of the dissolution behaviour of coal-chars into liquid iron.
At high rates of coal injection, full combustion of the injected coal is quite unlikely and
results in the formation of unburnt char. The dissolution of carbon from chars into liquid
metal is one of the important routes for the consumption of the unburnt char. As carbon
dissolution rates have been determined for a few chars only, a systematic and
comprehensive study was undertaken in this project, on the dissolution behaviour of
carbon from non-graphitic materials into liquid iron. In addition to measuring the
kinetics of carbon dissolution from a number of coal chars into liquid iron as a function
of parent coal, liquid iron composition, and coal ash composition, the influence of
chemical reactions between solute/solid carbon and ash oxides was also investigated.
These studies were supplemented with investigations on one metallurgical coke for the
sake of comparison.

Experiments were planned to investigate carbon dissolution under two extreme


conditions. Small-scale experiments using the sessile drop arrangement were carried out
to investigate the dynamic wettability and the formation of interfacial reaction products
under relatively static conditions. This technique has the advantage of having relatively
large amounts of char as compared to liquid iron in the droplet. In addition, a time
dependant view of the reaction can be obtained by quenching samples out of the furnace
at certain times. The larger scale experiments were undertaken using the carburiser
cover technique where both the carburiser and liquid iron are present in excess under
highly turbulent conditions. This technique was chosen to overcome any resistance to
carbon supply for carbon dissolution in liquid iron.

The wettability of coal chars and coke with liquid iron at 1550°C was measured as a
function of time. Measurements were carried out with electrolytic iron and with an iron
alloy containing 2 wt% C and 0.01 wt% S. With electrolytic iron, all materials showed a

iv
non-wetting behaviour with contact angles ranging between 106°-133° in the initial
stages and between 101° to 111° after an hour of contact. For the Fe-C-S alloy, the
contact angles ranged between113°-127° in the initial stages and between 101° to 125°
after an hour of contact. Being essentially non-wetting, only a marginal improvement in
contact angles was observed with time.

A number of reaction products were observed in the interfacial region between the
chars/coke and liquid iron. The accumulation of alumina at the interface was detected
for all materials and was seen to increase with time in all cases. As the carbon dissolves
into the liquid iron, some of the ash material gets left behind at the interface. This is
specifically true in case of alumina, as it does not undergo reduction reactions and is
therefore not removed from the interface. Due to its non-wetting behaviour with liquid
iron (~120°), an increased presence of alumina in the interfacial region is expected to
significantly affect the dynamic wettability of chars/coke with liquid iron. Calcium and
sulphur also appeared to preferentially accumulate at the interface, concentrating at
levels in excess of those expected from the ash composition alone. However it was
observed that calcium sulphide appeared to form much more rapidly with the chars than
for the coke and also had a more extensive coverage for chars. Deposition of reaction
products in the interfacial region also had a significant effect on carbon dissolution.

Despite the high levels of silica in the ash initially, very little silica was detected in the
interfacial region, implying ongoing silica reduction reactions. A small amount of
silicon was however detected in the iron droplets, indicating silica reduction with solute
carbon. The silica in cokes and chars is known to reduce in-situ with the solid carbon,
however it was identified that the reduction reactions can also consume solute carbon in
the liquid iron. As this is occurring simultaneously with carbon dissolution into liquid
iron, the interdependency of silica reduction and carbon dissolution could potentially
limit the observed carbon dissolution rate.

The rates of the carbon dissolution from the chars/coke into liquid iron were determined
using the carburiser cover method. As the area of contact could vary significantly for
these materials, a theoretical model was developed for estimating the interfacial contact
area between chars and liquid iron. Using a force balance approach, the partial
penetration of the particles was calculated numerically and total solid/liquid contact area
v
was estimated. Wettability was found to have a very significant effect on the area of
contact. An improvement in wetting reduces the upward force due to surface tension
and therefore increased downward penetration of particles in the liquid, which results in
an improvement in contact area. A two-step behaviour was observed in the carbon
dissolution behaviour of two chars and coke. Stage I, which corresponds to dissolution
in the times of initial contact, showed a much higher rate of carbon dissolution as
compared to stage II during later times. Slow rates of carbon dissolution in stage II were
attributed to very high levels of interfacial blockage by reaction products leading to
much reduced areas of contact between carbonaceous material and liquid iron.

The first order dissolution rate constants were computed for stage I in four chars/coke
and the observed trend in first order dissolution rate constants (x103 ms-1) was as
follows: 0.02273 (Coke) > 0.01795 (Char 1) > 0.00942 (Char 4) > 0.00606 (Char 3) >
0.00278 (Char 2). These dissolution results compare well with the dissolution rate
constants quoted in the literature. Char 1 has the highest rate constant of the chars. Char
1 also has the lowest silica concentration among all chars. Coke on the other hand has
the highest dissolution rate constant among the materials investigated in this study.
While coke has higher silica concentration as compared to Char 1, it also has a much
higher Lc value. These results therefore indicate that a high Lc value (more structural
order) and low silica concentrations are expected to lead to higher rates of carbon
dissolution. The effect of structure and ash impurities on carbon dissolution rates could
be seen clearly in this study.

vi
TABLE OF CONTENTS
Section Page
Certificate of Originality ...................................................................................................ii
Acknowledgments............................................................................................................iii
Abstract ............................................................................................................................iv
Table of Contents ............................................................................................................vii
List of Figures ..................................................................................................................xi
List of Tables ...............................................................................................................xviii
CHAPTER 1: Introduction...........................................................................................1
1.1 Background .......................................................................................................1
1.2 Importance of Iron/Char Interactions................................................................2
1.3 Interfacial Reactions Between Iron and Char ...................................................3
1.4 Carbon Dissolution from Char ..........................................................................4
CHAPTER 2: Literature Review..................................................................................6
2.1 Pulverised Coal Injection (PCI) ........................................................................7
2.1.1 Partial Replacement of Coke.....................................................................7
2.1.2 Reductants for Blast Furnace Ironmaking ................................................7
2.1.3 Current Status of Pulverised Coal Injection............................................10
2.1.4 Combustion of Pulverised Coal ..............................................................12
2.1.5 Unburnt Char Production ........................................................................15
2.2 Types of Carbonaceous Materials...................................................................19
2.2.1 Coal .........................................................................................................20
2.2.2 Char .........................................................................................................20
2.2.3 Coke ........................................................................................................21
2.2.4 Graphite...................................................................................................21
2.2.5 Characterisation Techniques ...................................................................22
2.3 Structure of Carbonaceous Materials ..............................................................23
2.3.1 Graphite Structure ...................................................................................23
2.3.2 Coal Structure..........................................................................................25
2.3.3 Char Structure .........................................................................................26
2.3.4 Coke Structure.........................................................................................28
2.4 Coal Mineral Matter and Reduction Reactions...............................................29
2.4.1 Silica........................................................................................................30

vii
Section Page
2.4.2 Alumina...................................................................................................32
2.4.3 Iron Oxide ...............................................................................................34
2.4.4 Calcium Oxide ........................................................................................37
2.5 Carbon Dissolution .........................................................................................39
2.5.1 Dissolution of Carbon from Graphite .....................................................41
2.5.2 The Effect of Sulphur on the Dissolution of Graphite ............................43
2.5.3 Dissolution of Carbon from Non-Graphitic Sources ..............................48
2.5.4 Influence of Ash on Carbon Dissolution.................................................53
2.6 Wettability of Carbonaceous Materials...........................................................54
2.6.1 Wettability Phenomena ...........................................................................55
2.6.2 Effect of Chemical Reactions on Wettability..........................................57
2.6.3 Wettability of Carbonaceous Materials with Liquid Iron .......................58
2.7 Summary .........................................................................................................60
CHAPTER 3: Objectives and Study Parameters........................................................62
3.1 Objectives........................................................................................................62
3.2 Study Parameters.............................................................................................65
CHAPTER 4: Experimental Details...........................................................................68
4.1 Objectives........................................................................................................68
4.2 Production and Characterisation of Experimental Materials ..........................68
4.2.1 Iron-based Materials................................................................................69
4.2.1.1 Carbon and Sulphur Analysis using LECO SC244 ............................70
4.2.1.2 Field Emission Scanning Electron Microscopy ..................................70
4.2.2 Carbon-based Materials ..........................................................................71
4.2.2.1 Char Production ..................................................................................71
4.2.2.2 Coke ....................................................................................................73
4.2.2.3 Characterization of Carbonaceous Materials ......................................73
Proximate/Ultimate and Ash Analysis ............................................................73
X-ray Diffraction Analysis..............................................................................74
Visual/FESEM Characterization.....................................................................75
4.3 Experimental Techniques................................................................................78
4.3.1 Sessile Drop Technique ..........................................................................78
4.3.1.1 Dynamic Wettability for interfacial area determination .....................80
4.3.1.2 Study of Interfacial Reaction Products ...............................................82
viii
Section Page
4.3.1.3 Carbon/Sulphur/Silicon Transfer Studies ...........................................84
4.3.2 Carburiser Cover Technique ...................................................................84
CHAPTER 5: Interfacial Phenomena using Sessile Drop Technique: Results and
Discussion 87
5.1 Wettability Measurements ..............................................................................87
5.1.1 Wettability with Electrolytic Iron ...........................................................88
5.1.2 Wettability with Fe-2wt%C-0.01wt%S Alloy ........................................98
5.1.3 Influence of Metal Composition ...........................................................105
5.1.4 Dynamic Wettability Summary.............................................................105
5.2 Study of Interfacial Reaction Products .........................................................107
5.2.1 Coke Interfacial Reaction Products.......................................................107
5.2.2 Char Interfacial Reaction Products .......................................................114
5.2.3 Interfacial Reaction Products Summary................................................120
5.3 Carbon/Sulphur/Silicon Transfer Studies .....................................................122
5.3.1 Coke/Iron System..................................................................................122
5.3.2 Char/Iron System ..................................................................................126
5.3.3 Carbon/Sulphur/Silicon Transfer Summary..........................................131
CHAPTER 6: Carbon Dissolution Studies using the Carburiser Cover Method:
Results and Discussion..................................................................................................132
6.1 Carbon Dissolution Studies...........................................................................132
6.1.1 Char 1 Dissolution Results....................................................................134
6.1.2 Char 2 Dissolution Results....................................................................135
6.1.3 Char 3 Dissolution Results....................................................................137
6.1.4 Char 4 Dissolution Results....................................................................139
6.1.5 Coke Dissolution Results ......................................................................140
6.2 Sulphur Transfer ...........................................................................................142
6.3 Combined Dissolution Rate Constants..........................................................146
6.4 Estimation of Particle Contact Area..............................................................147
6.4.1 Penetration Depth..................................................................................149
6.4.2 Solid/Liquid Contact Area ....................................................................153
6.5 The Calculation of First Order Dissolution Rate Constants..........................154
6.6 Factors affecting Carbon Dissolution from Non-Graphitic Materials ..........155
6.6.1 Atomic structure....................................................................................156
ix
Section Page
6.6.2 Interfacial Phenomena...........................................................................157
6.6.3 Effect of Melt Sulphur ..........................................................................159
6.6.4 Reduction of Silica................................................................................161
6.7 Summary .......................................................................................................163
CHAPTER 7: Summary And Conclusions ..............................................................165
CHAPTER 8: References .........................................................................................170
List of Publications .......................................................................................................179
APPENDIX A: Sample Calculation of the Off-Gas Composition and Calculation of the
Number of Moles of Oxygen Removed from Coke/Iron System .................................180
APPENDIX B: Sample Calculation of Contact Area ....................................................182

x
LIST OF FIGURES
Title Page
Figure 2.1: Evolution of the Average Consumption of Reductants in France from 1970-
1996, from [21]. ........................................................................................................8
Figure 2.2: Schematic diagram of the tuyeres and raceway of a blast furnace, from [26]
.................................................................................................................................10
Figure 2.3: Duration of Long-Term Continuous Operation at Coal Rates in Excess of
200kg/tHM (in italics: average coal rate during the period), from [21]....................11
Figure 2.4: Best coke/coal rates for furnaces injecting PC, to 2003, from [27]..............11
Figure 2.5: Conditions of coal injection at the tuyere and of unburnt pulverised coal
during ascent, from [8]............................................................................................12
Figure 2.6: Burning history of pulverised coal, from [8]................................................13
Figure 2.7: Change of the combustibility of coal, devolatilisation ratio of volatile matter
and gasification ratio of char at the tuyere level with variable VM and PCI rate,
from [9] ...................................................................................................................14
Figure 2.8: Change of the combustibility of coal, devolatilisation ratio of volatile matter
and gasification ratio of char at the tuyere level with variable VM, from [9]. .......15
Figure 2.9: Change of amount of residual unburnt char in hypothetical blast furnace,
from [8]. ..................................................................................................................18
Figure 2.10: The structure of graphite showing the A and B layers. ..............................23
Figure 2.11: Schematic of the carbon crystallite, showing relevant dimensions. ...........24
Figure 2.12: Schematic illustration of the simplified coal structure, from [50]..............25
Figure 2.13: Classification scheme based on measurements of cross-sectional area,
porosity, wall thickness and radial mass distribution, from [51]. ...........................27
Figure 2.14: a) X-ray spectra and profile of the (002) carbon peak for a coke during
heating under neutral conditions; b) relationship between carbon structure (Lc,
angstroms) and temperature, from [54]...................................................................28
Figure 2.15: Change of SiO2 content in coke with increasing temperature in a blast
furnace, from [18]. ..................................................................................................30
Figure 2.16: Arrhenius plots for SiO and SiO2 reactions: A, SiO(s)→SiO(g); B,
SiO2+Si→2SiO; C, SiO2+3C→SiC+2CO; D, SiO+2C→SiC+CO (lid); E,
SiO+SiC→2Si+CO (loose lid); F, SiO+SiC→2Si+CO (tight lid); G,
2SiO2+SiC→3SiO+CO; H, SiO(s)→SiO(g) (lid), from [58]. ...................................32

xi
Title Page
Figure 2.17:Ellingham diagram for oxides, from [59]....................................................33
Figure 2.18:The phase stability diagram for the system Fe-O-Al. Mutual solubility
between Al2O3 and FeAl2O4 is not considered, from [62]. .....................................33
Figure 2.19: Phase diagram for iron and its oxides as a function of oxygen potential
PCO 2
( PCO ), from [64]. .....................................................................................................35

Figure 2.20: Idealised representation of the processes in the reaction of Fe-C droplets
with liquid slag, from [66]. .....................................................................................36
Figure 2.21: Oxygen, sulphur, and silicon profiles in vicinity of CaO particle, from [70].
.................................................................................................................................38
Figure 2.22: Stability regions of CaS and CaO in Fe-Ca-S-O system, from [67]...........38
Figure 2.23: Variation of the Carbon Concentration at the Carbon Hot Metal interface
according to the theory of dissolution. Ci is the carbon concentration at the
interface, C the bulk carbon concentration, and δ is the interfacial layer thickness
(Carbon saturation Cs, is typically greater than Ci) from [5]. .................................39
Figure 2.24: Experimental Methods for Carbon Dissolution Studies, from [5]. ............40
Figure 2.25: Atomic distribution profile across graphite/Fe-C interface before and after
simulation. Z corresponds to the layer number normal to the interface, from [87].43
Figure 2.26: Influence of sulphur concentration on overall mass transfer coefficient in
using graphitic and non-graphitic carbon, (N is the rotational speed), from [76]...45
Figure 2.27: Atomic distribution profile across graphite/Fe-S interface before and after
simulation. Z corresponds to the layer number normal to the interface, from [80].47
Figure 2.28: Sessile drop showing definition of contact angle and the surface tension
forces acting on the drop. ........................................................................................55
Figure 2.29: The variation of the contact angle with time in the sessile drop experiment
for a melt/graphite assembly, from [98]..................................................................58
Figure 2.30: Variation of the contact angle with time for Fe-C-S melts with different
sulphur concentrations (C pct = 2.0), from [98]. ....................................................59
Figure 2.31: Change of contact angle with time for iron on natural graphite substrate,
from [14]. ................................................................................................................59
Figure 4.1: Schematic diagram of drop tube reactor used ..............................................72
Figure 4.2: X-ray diffraction spectra for the chars and the coke. ...................................75
Figure 4.3: SEM image of Char 1 at 100x magnification. ..............................................76

xii
Title Page
Figure 4.4: SEM image Char 2 at 100x magnification. ..................................................76
Figure 4.5: SEM image of Char 3 at 100x magnification. ..............................................77
Figure 4.6: SEM image of Char 4 at 100x magnification. ..............................................77
Figure 4.7:Sessile drop schematic...................................................................................79
Figure 4.8: Horizontal furnace schematic .......................................................................80
Figure 4.9: Sample program output for captured image angle analysis..........................82
Figure 4.10: Schematic of Carburiser cover method. .....................................................85
Figure 4.11: Schematic diagram of Induction Furnace utilising the carburiser cover
method.....................................................................................................................85
Figure 5.1: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Char 1 at 1550°C.....................................................................88
Figure 5.2: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Char 2 at 1550°C.....................................................................88
Figure 5.3: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Char 3 at 1550°C.....................................................................89
Figure 5.4: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Char 4 at 1550°C.....................................................................89
Figure 5.5: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Coke at 1550°C. ......................................................................89
Figure 5.6: Sample analysis of the coke initial contact angle with electrolytic iron using
the Angle software. .................................................................................................90
Figure 5.7: Sample analysis of the coke final contact angle with electrolytic iron using
the Angle software. .................................................................................................90
Figure 5.8: Dynamic contact angle results for Char 1 with electrolytic iron at 1550°C. 91
Figure 5.9: Dynamic contact angle results for Char 2 with electrolytic iron at 1550°C. 91
Figure 5.10: Dynamic contact angle results for Char 3 with electrolytic iron at 1550°C.
.................................................................................................................................92
Figure 5.11: Dynamic contact angle results for Char 4 with electrolytic iron at 1550°C.
.................................................................................................................................92
Figure 5.12: Dynamic contact angle results for Coke with electrolytic iron at 1550°C. 93
Figure 5.13: Dynamic contact angle results for the carbonaceous materials with
electrolytic iron at 1550°C, logarithmic time scale.................................................94

xiii
Title Page
Figure 5.14: Increase of interfacial coverage with time due to formation of interfacial
complex with time, from [14]. ................................................................................97
Figure 5.15: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Char 1 at 1550°C.................................................................98
Figure 5.16: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Char 2 at 1550°C.................................................................99
Figure 5.17: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Char 3 at 1550°C.................................................................99
Figure 5.18: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Char 4 at 1550°C...............................................................100
Figure 5.19: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Coke at 1550°C. ................................................................100
Figure 5.20: Dynamic contact angle results for Char 1 with Fe-2wt%C-0.01wt%S alloy
at 1550°C...............................................................................................................101
Figure 5.21: Dynamic contact angle results for Char 2 with Fe-2wt%C-0.01wt%S alloy
at 1550°C...............................................................................................................101
Figure 5.22: Dynamic contact angle results for Char 3 with Fe-2wt%C-0.01wt%S alloy
at 1550°C...............................................................................................................102
Figure 5.23: Dynamic contact angle results for Char 4 with Fe-2wt%C-0.01wt%S alloy
at 1550°C...............................................................................................................102
Figure 5.24: Dynamic contact angle results for Coke with Fe-2wt%C-0.01wt%S alloy at
1550°C. .................................................................................................................103
Figure 5.25: Dynamic contact angle results for the carbonaceous materials with Fe-
2wt%C-0.01wt%S alloy at 1550°C, logarithmic time scale. ................................104
Figure 5.26: Dynamic contact angle results for Char 1 with electrolytic iron and Fe-
2wt%C-0.01wt%S alloy at 1550°C.......................................................................105
Figure 5.27: Left: Interface of iron droplet after being exposed to coke at 1550oC for 1
minute. Right: EDS spectra of overall interface. ..................................................108
Figure 5.28: Left: Interface of iron droplet after being exposed to coke at 1550oC for 10
minutes. Right: EDS spectra of overall interface..................................................108
Figure 5.29: Left: Interface of iron droplet after being exposed to coke at 1550oC for 30
minutes. Right: EDS spectra of iron drop interface. .............................................109

xiv
Title Page
o
Figure 5.30: Oxygen removal from coke and coke/iron system at 1550 C, calculated
from mass spectrometer data.................................................................................110
Figure 5.31: Left: Interface of iron droplet after being exposed to coke at 1550oC for 60
minutes. Right: EDS spectra of indicated complex phase. ...................................112
Figure 5.32: EDS image map illustrating interfacial layer formed between iron and coke
at 1550oC after 1 hour, a. shows secondary electron image, b. shows iron, c. shows
sulphur, and d. shows calcium. .............................................................................112
Figure 5.33: Still images of liquid electrolytic iron droplet on coke substrate, a. shows
droplet 1 second after melting, b. and c. show reaction occurring between 1 second
and 2 minutes, with bubbles of gas forming within the droplet and a visible gas
surrounding the droplet. ........................................................................................114
Figure 5.34: Left: Interface of metal droplet after 5 minutes contact of iron with Char 4,
x350 magnification. Right: EDS spectra of region indicated. ..............................115
Figure 5.35: Left: High magnification image of interface of metal droplet after 5
minutes contact with Char 4, showing fused ash material and iron surface, x1500
magnification. Right: EDS spectra of region indicated. .......................................115
Figure 5.36: Left: Interface of metal droplet after 3 hours contact with Char 4, showing
high interfacial coverage, x350 magnification. Right: EDS spectra of region
indicated. ...............................................................................................................116
Figure 5.37: Left: Interface of metal droplet after 3 hours contact with Char 4, showing
calcium sulphide complex, and iron surface, x1500 magnification. Right: EDS
spectra of region indicated. ...................................................................................116
Figure 5.38: Interface of metal droplet after 5 minutes contact with Char 1, left x350
magnification, right x1500 magnification.............................................................119
Figure 5.39: Interface of metal droplet after 3 hours contact with Char 1, showing high
amount of interfacial coverage, left x350 magnification, right x1500 magnification.
...............................................................................................................................120
Figure 5.40: Carbon and sulphur transfer from coke into liquid iron at 1550oC ..........122
Figure 5.41: Variation of droplet carbon content with time during natural graphite /iron
droplet interactions, from [14]. .............................................................................124
Figure 5.42: Silicon transfer from coke into liquid iron at 1550oC ..............................125
Figure 5.43: Carbon and sulphur transfer from Char 4 into liquid iron at 1550oC .......127
Figure 5.44: Carbon and sulphur transfer from Char 1 into liquid iron at 1550oC .......127
xv
Title Page
Figure 5.45: Variation of droplet sulphur content with time during natural graphite /iron
droplet interactions, from [14]. .............................................................................130
Figure 6.1: Carbon concentration (wt%) in the melt for dissolution of carbon from Char
1 at 1550°C............................................................................................................134
Figure 6.2: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 1 into liquid iron at
1550°C. Two distinct regions (I and II) can be clearly identified.........................135
Figure 6.3: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 1- region I, into an
iron-carbon-sulphur alloy at 1550°C.....................................................................135
Figure 6.4: Carbon concentration (wt%) in the melt for dissolution of carbon from Char
2 at 1550°C............................................................................................................136
Figure 6.5: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 2 into liquid iron at
1550°C. Only one distinct region (I) can be identified for Char 2........................136
Figure 6.6: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 2 for region I into
liquid iron at 1550°C.............................................................................................137
Figure 6.7: Carbon concentration (wt%) in the melt for dissolution of carbon from Char
3 at 1550°C............................................................................................................137
Figure 6.8: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 3 into liquid iron at
1550°C. Two distinct regions (I and II) can be clearly identified.........................138
Figure 6.9: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 3- region I, into an
iron-carbon-sulphur alloy at 1550°C.....................................................................138
Figure 6.10: Carbon concentration (wt%) in the melt for dissolution of carbon from
Char 4 at 1550 °C..................................................................................................139
Figure 6.11: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 4 into liquid iron at
1550°C. Only one distinct region (I) can be identified for Char 4........................140
Figure 6.12: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 4- region I, into an
iron-carbon-sulphur alloy at 1550°C.....................................................................140
Figure 6.13: Carbon concentration (wt%) in the melt for dissolution of carbon from
coke at 1550 °C. ....................................................................................................141
Figure 6.14: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for coke into liquid iron at
1550°C. Two distinct regions (I and II) can be clearly identified.........................142
Figure 6.15: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Coke- region I, into an
iron-carbon-sulphur alloy at 1550°C.....................................................................142
Figure 6.16: Sulphur dissolution profile for Char 1 at 1550°C.....................................143
xvi
Title Page
Figure 6.17: Sulphur dissolution profile for Char 2 at 1550°C.....................................143
Figure 6.18: Sulphur dissolution profile for Char 3 at 1550°C.....................................144
Figure 6.19: Sulphur dissolution profile for Char 4 at 1550°C.....................................144
Figure 6.20: Sulphur dissolution profile for Coke at 1550°C. ......................................145
Figure 6.21: A schematic representation of particle in contact with liquid. Where a)
represents a wetting scenario (θ<90°), and b) represents a non-wetting scenario
(θ>90°) between the solid particle and liquid surface. .........................................149
Figure 6.22: SEM images of reacted char particles. A number of white deposits,
representing reaction products can clearly be seen. ..............................................152
Figure 6.23: Schematic representation of the various coal structures, from [110]. ......156

xvii
LIST OF TABLES
Title Page
Table 4.1: Iron used in carburiser cover experiments .....................................................69
Table 4.2: Iron used in horizontal tube experiments.......................................................69
Table 4.3: Char/coke compositions, dry basis (in wt%) .................................................73
Table 4.4: Ash composition of chars/coke......................................................................74
Table 4.5: Calculated Lc parameters for the carbonaceous materials .............................75
Table 4.6: Average particle size of char particles ...........................................................78
Table 4.7: Wettability experiments .................................................................................80
Table 4.8: Interfacial Reaction Product Studies..............................................................83
Table 4.9: Sessile Drop Mass Transfer Experiments. .....................................................84
Table 5.1: Initial contact angle for electrolytic iron on different carbonaceous substrates
at 1550°C.................................................................................................................94
Table 5.2: Final contact angle for electrolytic iron on different carbonaceous substrates
at 1550°C.................................................................................................................95
Table 5.3: Initial contact angle for Fe-2wt%C-0.01wt%S alloy on different
carbonaceous substrates at 1550°C. ......................................................................103
Table 5.4: Final contact angle for Fe-2wt%C-0.01wt%S alloy on different carbonaceous
substrates at 1550°C..............................................................................................104
Table 6.1: Combined dissolution rate constants for a range of coal-chars and coke at
1550 °C .................................................................................................................146
Table 6.2: The packing density of carbonaceous particles on the bath surface ............148
Table 6.3: The computation of penetration depth of carbonaceous particles into liquid
iron. .......................................................................................................................150
Table 6.4: Computation of particle/liquid contact area.................................................153
Table 6.5: First order carbon dissolution rate constants for the carbonaceous materials at
1550°C. .................................................................................................................154
Table 6.6: Carbon dissolution rate constants at 1600°C for a range of cokes, from [15,
16]. ........................................................................................................................155

xviii
CHAPTER 1: Introduction

CHAPTER 1: INTRODUCTION

Iron/carbon interactions have long been the subject of interest to ironmakers. These
interactions have application to a large number of ferrous metallurgical processes. This
study is particularly relevant to blast furnace ironmaking with pulverised coal injection,
which is a growing field due to environmental and economic concerns.

1.1 Background

Increasing economic and environmental restrictions have led to alternative ironmaking


technologies based on smelting and direct reduction. Despite the new technologies, blast
furnace ironmaking remains the primary route to produce iron, because the process has
improved to maintain a cost competitive advantage over the emerging alternatives.
However, blast furnace ironmaking is now faced with new challenges. In 1997 the
Kyoto Protocol to the United Nations Convention on Climate Change proposed binding
emission reduction targets for developed countries. Whilst the practical implications of
this agreement are yet to be concluded, it has been identified that reductions of
greenhouse gas (GHG) emissions are a priority for all iron and steel industries [1].

A major reaction to the environmental issues faced by blast furnace ironmaking has
been to reduce the consumption of coke, as coke oven emissions are one of the largest
sources of GHG in the blast furnace ironmaking cycle. Additionally, there is a limited
supply of coking coals, coke making facilities, and a high cost involved in building new
coke ovens. One way to reduce coke consumption is to inject supplementary fuels at the
tuyere level of the blast furnace. Pulverised coal (PC) has emerged as a cheap readily
available auxiliary fuel, and when injected in the tuyeres of a blast furnace (PCI),
supplies heat and reductant gases that partially replace the coke. The pulverised coal
cannot replace the coke entirely as coke is still needed to support the burden and
provide a porous bed that allows permeation of the reactive gases. To maximise the
benefits of lower coke consumption by PCI, high rates of coal injection are required.
This however has proven difficult, with practical limitations forcing ironmakers to
restrict the amount of coal injected. One of the many issues limiting PCI rates are higher
coke quality requirements due to the increased residence times of coke in lower coke

1
CHAPTER 1: Introduction

rate furnaces [2]. The increased residence times for the coke in the furnace means there
is a higher degradation of the coke resulting from mechanical stresses, solution loss
reactions, alkaline and high temperature attack and damage by the high speed hot blast.
Another issue facing ironmakers using high rates of PCI is the formation of char fines
resulting from incomplete combustion of the coal in the raceway region. These char
fines, along with the increased fines produced from coke degradation, can reduce
permeability of the coke bed, and resist iron and gas drainage in the furnace, dirtying of
the deadman (the lower region of the blast furnace) and ultimately lead to a decrease in
the furnace productivity [2-4].

When considering the problems facing the production of iron using high rates of PCI, it
becomes clear that it is very important to determine the factors that control the
formation and accumulation of the fine material in the blast furnace. Current research
trends demonstrate this, with work conducted on coke quality requirements and coke
dissolution studies [2, 5-7], coal char combustion and reactivity with gas in the blast
furnace [8-10], and coal char reactivity with slag in the blast furnace [11, 12]. The
ability of the fines to be consumed by phases within the blast furnace is obviously
important and the capacity of iron to be carburised by the carbon in the char fines is one
area where there is currently very little understanding.

1.2 Importance of Iron/Char Interactions

If the aim of increasing the PCI rate to the highest level possible without sacrificing
furnace stability is to be met, the problem of accumulating unburnt char needs to be
overcome. Several researchers have determined that at higher rates of injection full
combustion of the injected coal is not possible and formation of unburnt char is
inevitable [9, 13]. Once the unburnt char has been produced in the raceway it can
remain within the furnace, be consumed within the furnace or leave the furnace with the
off gases. The issue of most concern is the accumulation of the char fines within the
blast furnace as this may impact on the stability of operation. Knowledge of the rates of
consumption of the char with phases in the blast furnace may allow a coal to be chosen
that is consumed more quickly than others, leading to less accumulation. The possible
consumption mechanisms include carbon dissolution into the liquid iron, consumption

2
CHAPTER 1: Introduction

of the carbon in the char by the reduction of the liquid slag and subsequent assimilation
of the ash into the slag, and reaction with hot gases.

The dissolution of carbon from coke in the blast furnace is usually the source of carbon
in the hot metal. However when injecting coal as an auxiliary fuel, the unburnt char that
is produced at high rates can also carburise the liquid iron. The understanding of carbon
dissolution from non-graphitic carbons such as coke and chars is rather limited, with the
role of carbon type and ash composition not well determined. The carbon dissolution
from char into liquid iron is the focus of this study. The dissolution behaviour of the
char is considered by study of the reactions that occur between the char material and
iron. This work determines the comparative rates of dissolution of the various carbon
sources and undertakes measurements in limiting regimes to identify those factors that
most affect the dissolution reaction.

1.3 Interfacial Reactions Between Iron and Char

Before the dissolution rates of carbon from char into liquid iron can be completely
understood, it is important to identify any reactions that are occurring, other than
dissolution. In previous studies on carbon dissolution from non-graphitic sources the
role of ash has been only briefly considered.

The influence of ash, as a source of reducible material and as a material that may
directly react with liquid iron alloys, has not been fully investigated during studies on
carbon dissolution. Some work has been reported on natural graphite studying the
influence of ash on carbon transfer [14], and several papers [5, 15-17] have reported on
the influence of ash as a physical barrier during coke dissolution. It has been recognised
for some time that silica from the coke ash reduces and enters the blast furnace hot
metal [18]. However, the interdependence of the silica reduction and carbon dissolution
has not been established. A complete understanding of the phenomena that occur at the
interface during carbon dissolution has yet to be gained. In this project the interfacial
reactions will be observed in a small-scale system. The smaller scale studies allow
investigation to focus on the progress and products of interfacial reactions in relatively
static conditions. Here transport-limiting conditions are allowed to dominate, permitting

3
CHAPTER 1: Introduction

the identification of the products of the interfacial reactions. The time dependent
formation of the phases at the interface will be observed, and the influence that these
reactions had on the mass transport of carbon and other elements into the liquid iron
determined.

1.4 Carbon Dissolution from Char

The carbon dissolution from chars into liquid metal could be a significant route for the
consumption of the unburnt chars in the blast furnace. However carbon dissolution rates
have been determined for a few chars only, and no systematic studies have been
reported in the literature.

The dissolution of carbon from the char into the liquid iron will be investigated using
the carburiser cover method. This is the only method to study carbon dissolution that
allows the use of a powdered carbonaceous material. Using a powdered form of char is
considered very important in this work to reflect the nature and behaviour of PCI chars
in the ironmaking process. The carbon dissolution experiments will be undertaken in
conditions where both the carburiser and liquid iron were present in excess and highly
stirred. To remove the geometry effects of the experiment from the determined rate
constants, the interfacial area between the chars and the liquid iron will be calculated
using a theoretical model. The model will determine the interfacial contact between
char/liquid iron taking into account the buoyancy forces, surface tension of the liquid,
weight of char, and the contact angle between the chars and liquid iron. Due to a lack of
knowledge of contact angles between char and liquid iron, dynamic wettability
experiments will also be conducted to determine this parameter. Once the carbon
dissolution rates are calculated and the geometrical factors removed, first order carbon
dissolution rate constants for chars will be determined. These rate constants may then be
used to determine which carbon forms – coals/coke/graphite - will dissolve faster than
others. This information will assist in the understanding of PCI coal performance.

4
CHAPTER 1: Introduction

The aims of the present work are briefly summarised below:

• Identification of the reactions occurring between chars and liquid iron,


• Relating the influence of these reactions on the carburising ability of the chars,
• Identification of products forming at the interface, and determining the sequence
of formation,
• Using dynamic wettability measurements to determine contact angle and
through model calculations estimate the iron/char particle contact area,
• Calculation of the first order carbon dissolution rate constants for the chars, and
for a standard reference coke material
• Critical analysis of the experimental results and discussion on carbon dissolution
from chars.

5
CHAPTER 2: Literature Review

CHAPTER 2: LITERATURE REVIEW

A clear understanding of the dissolution behaviour of various carbonaceous materials


into liquid iron and the reactions taking place across the carbon/melt interface is of
fundamental importance in a number of iron making processes. While there have been a
few studies on the dissolution behaviour of cokes and coals, most of the investigations
have focussed their attention on the dissolution behaviour of graphite. These studies,
both experimental and theoretical, have provided a great deal of information about the
reaction kinetics and various factors affecting the graphite dissolution rate. However
significantly less work has been carried out on dissolution behaviour of non-graphitic
carbonaceous materials, and very few consider the influence of the ash and its reactions
on carbon dissolution kinetics. With the focus in this project on carbon dissolution from
chars into liquid iron during PCI and interfacial phenomena, the current status of
literature in this field is critically reviewed and presented in this chapter.

Starting with a brief introduction to pulverised coal injection in blast furnace


ironmaking, a discussion on the combustion of pulverised coal and production of
unburnt char is presented. As the build up of this unburnt char could significantly affect
the stability of the blast furnace burden, an understanding of the consumption of char
through interactions with slag, gas phase and metal is of great importance. This project
is focussed on the interactions of char with liquid metal. In order to develop a
fundamental understanding of these reaction processes/mechanisms, an in-depth
knowledge of basic carbonaceous material characteristics, possible reactions and the
mechanisms of carbon dissolution are required. A literature review examines the
different types of carbonaceous materials, their structural details, ash components and
reduction reactions possible. The basic mechanisms of carbon dissolution and
wettability are also presented in this chapter.

6
CHAPTER 2: Literature Review

2.1 Pulverised Coal Injection (PCI)

2.1.1 Partial Replacement of Coke

While pulverised coal injection has been around for many years, recently it has gained
in importance due to the following reasons:

• Coke making has become increasingly expensive due to the high (and still rising)
cost of coking coal [19],
• Stricter environmental regulations [1] could cause a shut down of all but the most up
to date coke batteries [20],
• Coke batteries world wide are aging, and the cost of replacing batteries is high [20],
• Coke demand increasingly has to be supplied by a lesser number of batteries [20].

All the above factors have resulted in ironmakers reducing the amount of coke used in
their furnaces, and using materials such as coal or gas injection to supply the shortfall in
either heat or carbon resulting from reduced coke usage.

2.1.2 Reductants for Blast Furnace Ironmaking

The blast furnace is the major source of pig iron for steelmaking, and economic
necessity has required increased productivity at reduced cost [20, 21]. Reducing the
costs can be achieved by decreasing the overall coke rate, as coke is one of the main
expenses in ironmaking. One way of accomplishing this is to replace part of the coke
with a tuyere injectant such as oil, natural gas or coal. Each of these injectants are
cheaper than coke, and thus if a commensurate proportion of the coke is replaced, the
hot metal cost will be reduced. A certain quantity of coke is required for burden
stability, but there is significant scope for injection.

7
CHAPTER 2: Literature Review

Figure 2.1: Evolution of the Average Consumption of Reductants in France from 1970-
1996, from [21].

The future supply of coke is uncertain, with declining production of coke in most
developed countries [19]. Coke rate per tonne of hot metal produced has decreased
steadily for the past decade, mainly due to improved ironmaking practices and furnace
additions such as coal, oil, tar, or gas, see Figure 2.1, but coke oven capacity has been
declining too, due to the abovementioned factors.

Decreasing the amount of coke used in the blast furnace is one of the major aims of all
ironmakers [19, 21]. Advantages of decreased coke consumption are:
• Reduced cost of hot metal owing to the price difference between coke and typical
fuel replacements, such as oil or coal.
• Extension of coke battery life, advantageous because of the high cost of
replacement.

Productivity and efficiency from modern blast furnaces has increased greatly in the last
30 years. Before 1960, a coke rate of 600kg of coke/tHM was common in industrialised
countries, [22]. Now, below 500kg of equivalent coke/tHM can be expected, or 450kg of
equivalent coke/tHM for superior furnaces. This reduction was obtained by improving
process performances; a further reduction in the coke quantity can be obtained by
substituting some of the coke with another reducer.

Coke equivalent is one way of quantifying the amount of carbonaceous reductants being
added to a blast furnace [22]. A replacement ratio for each reductant considered could

8
CHAPTER 2: Literature Review

be calculated, by considering the replacement’s reducing potential relative to coke. The


coke equivalent can then be calculated by adding the actual coke rate to the sum of the
other reducers used each multiplied by its replacement capability. This replacement
ratio can vary significantly, depending on the nature of the injectant, blast temperature,
and degree of indirect reduction per unit injectant. The replacement ratio for pulverised
bituminous coals varies little [23], with values ranging from 0.8-0.86 kg/kg. The
replacement ratio of pulverised anthracite also differs only slightly with different
operating conditions, with values around 1.0 kg/kg. This difference between coals and
anthracites reflects their respective carbon:hydrogen ratios, which affects both their
energy and reductant replacement ratios.

Of the available replacements for coke (oil, natural gas, and coal), each has advantages
and disadvantages as possible blast furnace injectants. Recently, waste plastic has also
had limited usage as an injectant. Two main factors used to assess blast additions are
coke replacement ratio (RR) and the change in productivity (∆P) [23]. The coke
replacement ratio has two factors, the coke saving from the decreased amount of direct
reduction and the change in coke consumption due to the amount of heat introduced into
the blast furnace. The influence of injectants on furnace productivity depends on the
change in the quantity and composition of the gases in the furnace and on the relative
amount of coke in the stock column.

Each injectant has different combustibility and replacement ratios, and the decision on
which injectant is to be chosen for a particular furnace often is based around the
following factors:
• Cost in terms of unit energy and carbon delivered to the furnace
• Supply and required treatment of alternative fuel (such as grinding in the case of
coal)
• Cost of add-on installation required for injection
• Other issues such as transport and cost variability of the injectant

Several benefits of injecting natural gas into the blast furnace, as compared with those
arising from coal injection, have been identified [24]. Natural gas allows more oxygen
injection than other fuels because of the different thermal and physical effects caused by

9
CHAPTER 2: Literature Review

the natural gas. Oxygen injection has been seen to increase the furnace productivity.
However many ironmaking companies are choosing coal injection as the solution to
reducing coke consumption [25], over other supplementary fuels such as natural gas,
due to the cost of coal compared to natural gas.

2.1.3 Current Status of Pulverised Coal Injection

Coal is usually injected at the tuyere via an injection lance. The coal begins heating and
combustion within the tuyere, before reaching the raceway, see Figure 2.2.

Figure 2.2: Schematic diagram of the tuyeres and raceway of a blast furnace, from [26]

When coal is chosen as an injectant, the aim of the ironmaker is to gain the maximum
injection rate while maintaining furnace stability and product quality. To date, the aim
of premium furnaces is to reach and maintain in injection rate of 200 kg/tHM. Figure 2.3
shows furnaces that have obtained this goal. Figure 2.4 lists some of the best coal/coke
rates obtained to 2003.

10
CHAPTER 2: Literature Review

Figure 2.3: Duration of Long-Term Continuous Operation at Coal Rates in Excess of


200kg/tHM (in italics: average coal rate during the period), from [21].

Figure 2.4: Best coke/coal rates for furnaces injecting PC, to 2003, from [27].

The decision of which coal is chosen to inject into a blast furnace is usually based on
the specific problems/issues facing each individual furnace. At moderate injection rates,
where technical issues relating to coal injection are few, the best coal to inject would
often be the cheapest readily available coal. However higher rate injection demands
some attention be paid to coal quality. Coal quality factors currently considered during
high rate coal injection are [28]:
• Combustibility (discussed in next section)
• Carbon to hydrogen ratio (C:H) which influences the coke replacement ratio
11
CHAPTER 2: Literature Review

• Coal chemistry (ash and S, P and Cl levels)


• Coal flowability and lance plugging behaviour
• Coal handling behaviour (moisture and fines generation)

2.1.4 Combustion of Pulverised Coal

Coal is injected into the blast furnace by means of a tuyere lance, which allows the coal
and the hot air blast to arrive together into the blast furnace, see Figure 2.5.

Figure 2.5: Conditions of coal injection at the tuyere and of unburnt pulverised coal
during ascent, from [8].

The injected pulverised coal is heated by the hot blast, and then by the heat of
combustion [8]. The combustion of coal begins by devolatilisation, where the volatile
matter (VM) in the coal forms a gaseous product (the formation of VM in an inert
atmosphere is called pyrolysis). This volatile matter can combust in the presence of
oxygen and the remaining solid material is usually known as char. This char can also
combust in the presence of oxygen. The burning history of pulverised coal can be seen
in Figure 2.6. Once the char has fully combusted, all the remaining carbon is consumed,
and the remaining material is ash (oxides derived from the mineral matter contained
within the coal).

12
CHAPTER 2: Literature Review

Figure 2.6: Burning history of pulverised coal, from [8].

From Figure 2.6 it can be seen that the volatile matter combusts first, followed by
combustion of the char. From this study [8], the authors state that there is an ignition
delay of around 500 ms, followed by gaseous combustion taking 200 ms, and finally the
solid combustion taking 900 ms (these times do not correspond to actual combustion
times within the blast furnace, which are typically much shorter). These times will also
be dependent on combustion temperature and particle size. From Figure 2.5 we can see
that the residence time of the PC in the raceway is only 10-30 ms. The currently
accepted knowledge is that the probability of unburnt pulverised coal leaving the
raceway area and moving into other parts of the furnace is high (dependant on PCI rate).
The proportion of char that remains unburnt is a function of the operation of the blast
furnace.

One study states that at high rates of injection (such as 200 kg/tHM) the injected coal is
not able to be combusted fully in the raceway [13], and this leads to devolatilised
unburnt char entering the lower zone and stack. The higher the rate of injection the
higher the proportion of unburnt char produced, due to limited oxygen presence at the
point of combustion.

Yamagata et al. [9] studied the combustion of coal injected into a coke bed at high rate.
Their conclusions were deduced from experimental results from a packed coke-bed
furnace and from mathematical models. The study of Yamagata et al. [9] investigated
13
CHAPTER 2: Literature Review

the effect injection rate had on the combustibility, devolatilisation ratio and gasification
ratio of the coal and coal char, as shown in Figure 2.7.

Figure 2.7: Change of the combustibility of coal, devolatilisation ratio of volatile matter
and gasification ratio of char at the tuyere level with variable VM and PCI rate, from
[9]

It can be seen from the preceding figure that the increasing rate of PCI decreases the
combustibility of the coal, due to the depletion of the oxygen at the tuyere nose. The
devolatilisation ratio of the coal also decreases with PCI rate; however, devolatilisation
is completed within the raceway region (as indicated by the dotted line). The
gasification of the char reduces with injection rate, with the maximum amount of
gasification of the char (at the highest rate examined) being 60%. This would result in
significant amounts of unburnt char accumulation.

The effect of volatile matter in the coal on PCI characteristics was also studied [9]. It
was seen that volatile matter in the coal increases the combustibility, increases the
gasification ratio of the char, and coals with higher VM contents are devolatilised closer
to the tuyere nose, (as shown in Figure 2.8).

14
CHAPTER 2: Literature Review

Figure 2.8: Change of the combustibility of coal, devolatilisation ratio of volatile matter
and gasification ratio of char at the tuyere level with variable VM, from [9].

Yamagata et al. [9] confirmed that the size of the pulverised coal influences the
combustibility, which increases with decreasing particle diameter. It is clear that any
factor that influences combustibility will affect the amount of unburnt char being
formed in the blast furnace.

It was found in a study by Lu et al. [29] that during combustion, the char that was
produced had become structurally more ordered. The level of this increase in ordering
was dependent on the heating rate, temperature and atmosphere.

2.1.5 Unburnt Char Production

Once the coal contacts the hot air blast in the tuyere pipe devolatilisation and
combustion commence. A certain fraction of the char produced is typically not
combusted, and forms a char. Yamaguchi et al. [30] reported that the maximum PCI rate
possible without residual unburnt char was 180 kg/tHM. This is not adjusting for
consumption of char fines by the furnace, and their modified value was 240kg/ tHM. This
upwardly revised figure should be taken loosely however, as the study was conducted
using a small-scale experimental apparatus, and figures calculated via mathematical
modelling. It would be appropriate to say, based on this work that any furnace into
which coal is injected at a moderate to high rate probably contains unburnt char. This
char can be consumed by blast furnace phases, accumulate within the blast furnace,
and/or be expelled from the top of the furnace. If the char does accumulate, it can lead
to furnace instability resulting from a drop in gas permeability through the furnace bed

15
CHAPTER 2: Literature Review

(obviously this accumulation cannot continue indefinitely, and would lead to major
furnace problems if left unchecked).

In work by Ueno et al. [31], the particle shape of the coal before and after pyrolysis was
examined. The effect of devolatilisation caused the blocky coal particles to form
cenospheres with high porosity. Due to the change in the structure of the coal when
pyrolysis occurs, Uneo et al. [31] speculated that the reactivity of the resulting char
particle would be increased (when compared to the initial pulverised coal) due to the
increased surface area. This increase in reactivity was presumed to be higher than for
coke. The volatile matter in the coal influences the final char shape, and thus the
resultant reactivity.

Studies of commercial plants have also revealed unburnt char production from PCI, [32,
33]. The Stelco blast furnace PCI facility injected an average of 121 kg/tHM. Reduced
permeability was one of the problems arising from the start-up of the PCI facility.
Significant char carryover into the blast furnace off-gas was also noted. The reduced
permeability and char carryover were brought under control by manipulation of
operating parameters and increased attention to burden preparation.

In the development of the high PCI technology at Hoogovens Ijmuiden [33] a rate of
215 kg/tHM was trialled followed by a long-term test of 190 kg/tHM. The 215 kg/tHM trial
was abandoned after a short period due to operational problems, but not before a
considerably higher than expected furnace resistance was noted. An increase of the
long-term trial rate from 190 kg/tHM to 210 kg/tHM was undertaken after remarkably
good process stability was noted. The trial showed that the burden resistance increased
with increasing PCI rate and limitation of the pressure difference requires a considerable
increase in oxygen in the blast to maintain productivity. Several problems were
encountered during the high rate trial. The high rate of oxygen leads to a very low top
pressure and flow, with little drying of the burden at the top of the stack, and can result
in heavy chills in the furnace, damaged tuyeres, and uneven burden descent. This trial
accounted for the drop in permeability by the higher amount of ore relative to coke,
thinner coke layers, deposits of unburnt coal particles, and changes in the burden
distribution aiming for increased wall flow (leading to some decrease in the central
flow).
16
CHAPTER 2: Literature Review

The Hoogovens study [33] concluded that almost all of the unburnt coal produced as a
result of incomplete combustion is eventually consumed within the blast furnace and is
not significantly carried over in the top gas. The three possible consumption
mechanisms cited are direct reduction (slag reaction), solution loss (gas reaction) or
carbonisation of the hot metal (metal reaction).

This result is not in agreement with other more recent blast furnace plant studies [28,
34], which found that there was significant deposition of unburnt char within the
furnace, and high concentrations in the off-gas solid traps (dust and sludge). This
highlights the unique nature of each blast furnace plant, which is not only subject to
regional variations in coal chemistry and quality but also the individual vagaries of each
plant operation. Ironmakers have increasingly turned to off-gas solids analysis to
determine if and how much unburnt char is being produced. Such analyses are difficult
to conduct due to the similarities between char and coke (both are often from high rank
coals, have very low VM, experienced thermal annealing to some extent, and often have
similar levels of carbon crystallinity), and typically both forms of carbon can leave in
the furnace off-gas. This may account for some of the inconsistencies between plant
trial reports for high PCI rates.

A study by Iwanaga [8] estimated where the unburnt pulverised coal (PC) resided in the
furnace. The position of maximum char fines was found around the 1400°C zone, see
Figure 2.9. This position is at the interface between the softening zone and the dropping
zone. Thus the phases present are slag, coke, liquid iron, iron ore, and gas. This study
found that unburnt char should be consumed effectively within the blast furnace. It must
be remembered that bench scale experimentation with mathematical modelling, while
effective in investigating a few of the factors influencing the blast furnace, they rarely
can take into account all factors acting in the furnace.

17
CHAPTER 2: Literature Review

Figure 2.9: Change of amount of residual unburnt char in hypothetical blast furnace,
from [8].

Thus from this study [8], it can be seen that to fully model the performance of the blast
furnace during PCI at high rates, it is important to know the rate of reaction coal char
has with each phase in the blast furnace.

Yamagata et al. [9] studied the effect of coal injection rate, coal type, coal size, and the
internal state of the furnace. They stated that it was the accumulation and consumption
behaviour of the unburnt char that was the factor of most interest when considering the
stability of the furnace at high PCI rates. This study extended to injection rates that are
far beyond that currently being employed, even in the very high rate furnaces. This large
injection rate allows problems that could be hidden in lower rate experiments to be
observed. This study modelled the accumulation of char within the blast furnace as a
function of PCI rate, volatile matter content, and pulverised coal size. As shown in the
earlier section, the higher the PCI rate, the more unburnt char produced, and the more
char accumulates within the furnace. Again as seen before, the researchers found
increasing the volatile matter in the coal contributes to burnout, and thus the coals with
the least amount of VM accumulate the most.

From the previous section it is clear that while ironmakers accept that unburnt char does
form within the blast furnace, there is little data to confirm the real state of the furnace
while injecting at moderate to high rates of coal. New techniques such as off-gas solid
analysis (either by petrographic or X-ray diffraction methods) [28, 34] can determine if
unburnt char is being produced, but the inherent difficulties in such an analysis limits its
18
CHAPTER 2: Literature Review

ability to determine the amount. Actual plant trials that have experienced problems
possibly due to char accumulation within the furnace have been able to do limited
sampling, with char being detected in the birds nest (area between raceway and
deadman) and deadman [6, 28].

An associated problem with high rates of coal injection is the corresponding increased
residence time of the coke [6, 35]. This leads to a greater level of coke degeneration and
formation of fines. It is these fines, along with the char generated, which is being
detected in the off-gas, and which contributes to the permeability issues affecting
furnaces operating at high rates of PCI.

As the desired PCI rate increases more complex problems arise from char or the
associated effects on the coke bed. Ironmakers are beginning to study the behaviour of
unburnt char within the furnace, leading to consideration of the state of the furnace
hearth. Studies by Nightingale et al. [3, 36] identified the condition and flow of the coke
bed through the deadman and hearth to be an area requiring more investigation. This is
due to their finding that the state of the coke bed in the deadman and hearth can be
related to the carburising ability of the coke. If the liquid hot metal reaches the deadman
and hearth unsaturated, then it has the ability to consume any fines that may have
accumulated there. If the hot metal is close to saturation then the fines may accumulate,
deteriorating permeability. From this study we can see the importance of also
understanding the carburising ability of carbonaceous materials entering the blast
furnace.

2.2 Types of Carbonaceous Materials

A variety of carbonaceous reducing materials are used for ironmaking. A brief


description of the carbonaceous materials, their properties, and their structure is given
below. In general, these carbonaceous materials are very complex heterogeneous
systems that exhibit a wide variety of physical and chemical properties.

19
CHAPTER 2: Literature Review

2.2.1 Coal

Coal can be considered as an organic rich sedimentary rock [37] with many mineral
components present. It is chemically composed of carbon, hydrogen, nitrogen, sulphur
and inorganic matter, known as mineral matter or ash. Typical criteria that are
considered for coals are volatile matter content, ash content and ash chemistry [38].
Both char and coke are produced from coal, and as such their properties will relate to
the properties of the parent coal material.

2.2.2 Char

Chars are produced when coal is heated to a temperature above which it begins
devolatilising. Chars are often partially combusted as well. This involves two steps [10]:

1. Coal pyrolysis: coal volatile matter is removed, leaving a solid (or possibly
semi-solid at high temperature) residue known as char;

Coal → Char + Volatiles

Equation 2.1: Formation of Char

2. Heterogeneous reactions of char- combustion;

C + O2 = CO2

Equation 2.2: Combustion of char to form CO2

C + 12 O2 = CO

Equation 2.3: Combustion of char to form CO

Char properties are affected by chemical and physical properties of the parent coal,
temperature and time history.

20
CHAPTER 2: Literature Review

2.2.3 Coke

Coke is the source of carbon used in traditional blast furnace ironmaking. It is produced
in ovens called coke batteries. Only a small proportion of coals are suitable for coking
[37]. Cokes are made by heating suitable blended coals in a controlled environment to
produce a porous brittle mass with a cellular structure. High quality coking coals are
often blended with poorer quality coals, to reduce coking costs. Coke is made in this
way to produce a material that is able to provide enough heat to the blast furnace upon
burning, and also with enough strength to support the blast furnace bed and maintain the
gas porosity required [39].

Quality and properties of coke are affected by the coal-rank, fluidity, maceral and
mineral matter composition as well as processing conditions. Traditionally, chemistry,
particle size, gas reactivity (CRI) and strength after reaction (CSR) are considered as
most important properties of metallurgical coke for blast furnace operations [40].

2.2.4 Graphite

Graphite typically is in one of two forms, synthetic or natural. Both forms contain
highly ordered carbon. These materials can contain some mineral content. Typically
graphite is very expensive, and is only used for special applications such as electrodes
in an electric arc furnace.

With implementation of the Kyoto Protocol, renewable and sustainable materials such
as charcoal are more attractive as raw materials than coal based materials, however they
are typically very expensive and use of these materials would only be worthwhile in
countries with large amounts of source wood and a lack of coal suitable for coking [41-
43].

21
CHAPTER 2: Literature Review

2.2.5 Characterisation Techniques

The characterisation of carbonaceous materials is important to a number of industries


and as such, many standardised tests have been created by various standards
organizations (the Standard Association of Australia, the International Organization for
Standardization, and the American Society for Testing and Material) [44]. Some of the
more common techniques are briefly described below.

Chemical analysis of the carbonaceous material can be done to determine the amount of
moisture, ash, volatile matter, and fixed carbon. An elemental analysis can also be
performed for the material.

Mechanical and physical analyses of carbonaceous materials were typically developed


for classifying coals for coke making. Some of the tests usually used to characterise
cokes include: measurements of specific gravity, free-swelling index, grindability, ash
fusion temperature, and coke strength after reaction (CSR).

Tests measuring reactivity are important in assessing the quality of reduction materials
for metallurgy. Typically they are based on the gas phase reaction of the carbonaceous
materials with carbon dioxide, oxygen, air or steam. The percentage of mass lost to
gasification reactions is usually referred to as reactivity [45].

In addition to these techniques, characterisation at an atomic level is also useful. A


number of experimental techniques such as x-ray diffraction (XRD), transmission
electron microscopy and high-resolution transmission electron microscopy (TEM and
HRTEM), scanning electron microscopy and field emission scanning electron
microscopy (SEM and FESEM) are being used. Results obtained from these studies
have provided a better understanding of the structure of carbonaceous materials, their
evolution after heat treatment, oxidation, and interfacial phenomena. The structural
aspect of carbonaceous materials relevant to this study will be discussed in section 2.3.

Previous studies in the field of carbon dissolution have shown that the most important
factors affecting carbon dissolution include the structure of carbonaceous material, ash

22
CHAPTER 2: Literature Review

content and composition, and interfacial phenomena. Chemical reactions of ash


impurities could also play an important role. These aspects are discussed in detail
below.

2.3 Structure of Carbonaceous Materials

There are several ways that the structure of carbonaceous material can be studied. The
atomic-level structure can be determined through transmission electron microscopy
(TEM), and near-atomic scale features can be identified with x-ray diffraction (XRD).
High magnification images can be obtained using scanning electron microscopy (SEM)
and field emission scanning electron microscopy (FESEM).

2.3.1 Graphite Structure

Graphite has been well characterised in terms of structure [46]. It consists of sheets of
polynuclear aromatic condensed rings extending indefinitely in the plane of those
sheets. Each ring is a hexagon of carbon atoms; each atom has three nearest neighbours
1.42 Å apart. Graphite is layered in the form ABABAB (see Figure 2.10).

Figure 2.10: The structure of graphite showing the A and B layers.

When a carboniferous material is exposed to x-rays, the crystal structure of the graphite-
like regions diffract the beam [47]. This causes 001 reflections from the aromatic layers
and (hk) reflections from the two-dimensional lattices. These reflections cause three
main peaks in the carbonaceous spectra, corresponding to the (002) plane and the (10)

23
CHAPTER 2: Literature Review

and (11) two-dimensional band. If the particle size of the material is less than 1000 Å
[48], then that particle size affects the diffraction peak produced. If the small regions of
the carbonaceous material are considered as the crystallite particles, it can be seen that
the extent of the ordering can be calculated. There are two parameters that can be
calculated in this way, Lc and La. The physical significance of these two parameters is
shown in Figure 2.11.

Figure 2.11: Schematic of the carbon crystallite, showing relevant dimensions.

The height of these crystallites can be determined from x-ray diffraction, by using the
Scherrer equation [48]:
0.9λ
Lc =
Bcosθ B

Equation 2.4

Where B is the broadening of the diffraction line measured at half its maximum
intensity in radians, Lc is the diameter of the crystal particle, λ is the wavelength of the
incident radiation and θB is the Bragg angle of the 002 peak (in degrees) or the angular
position of the main carbon peak.

The horizontal size of the crystallites can be similarly obtained [48]:


1.84λ
La =
Bcosθ B

Equation 2.5

24
CHAPTER 2: Literature Review

Where θB is the Bragg angle of the (11) band in degrees and B is the broadening of this
peak at half its maximum intensity.

It has been shown that crystallite size measured by XRD techniques is in good
agreement with the same parameters measured by TEM [47].

Commercially available graphites are not typically 100% carbon. They (like coals) also
contain a portion of oxide materials, although usually in much lower proportions.

The extent of ordering for carbonaceous materials varies widely depending on the
material [44]. Graphite commonly has an Lc value of around 250 Å. Coals have an Lc
value ranging from around 4 to 15 Å (however, Lc is related to coal rank, and thus some
anthracites have an Lc as high as around 20 Å) and chars have Lc values from 7 to 20 Å.
Generally, the process of creating a char from a coal causes the degree of ordering in the
carbon to increase (known as thermal annealing).

2.3.2 Coal Structure

Coal is generally considered to be non-crystalline but it can be thought of as consisting


of small clusters of carbon atoms resembling graphite [49], with the main difference
being the size or extent of ordering of these clusters. It also can contain varying levels
of carbon that is not contained in ordered regions. This type of carbon is known as
amorphous. A schematic of this can be seen in Figure 2.12.

Figure 2.12: Schematic illustration of the simplified coal structure, from [50].
25
CHAPTER 2: Literature Review

Two typical measures of the carbon structure for coals (and all other non-graphitic
carbons) are Lc and amorphous fraction. Both these can be determined from XRD
analysis.

2.3.3 Char Structure

As previously mentioned char is a product of coal combustion. During the charring


process, the carboniferous parts of the coal can change shape and size. Swelling (due to
degassing during devolatilisation), liquefying, and fracture can all occur during char
making [10, 51]. Practically this results in the morphology of the char particles possibly
being quite different to the parent coal particles.

The range of structures possible for a char have been described [51]. The very swollen
structures with large devolatilisation holes are commonly referred to as cenospheres.
The particles having a number of degassing pockets with regular sizes evenly
distributed across their volume are known as networks. Chars having solid structures
with very little porosity are also identified. Figure 2.13 shows a classification scheme
for char particle structure.

26
CHAPTER 2: Literature Review

Figure 2.13: Classification scheme based on measurements of cross-sectional area,


porosity, wall thickness and radial mass distribution, from [51].

The carbon ordering of a char may be somewhat larger than the parent coal [10]. The
level of thermal annealing that occurs will be a function of the thermal history of the
char. Long periods at high temperature will allow for more thermal rearranging than
shorter periods at lower temperatures.

The ordering of the char can also be increased due to an effect known as oxidative
ordering [10]. This is due to the fact that the amorphous carbon in char is more reactive
to gas than the crystalline carbon, so it is selectively removed leaving relatively more
ordered carbon in the sample.

27
CHAPTER 2: Literature Review

2.3.4 Coke Structure

Apart from graphite, coal, and char, there are several other carbon sources. During PCI,
coke competes with char in carburisation of the liquid iron, and as such it is pertinent to
the scope of this study. As coke is made from coals it typically has a similar
composition to chars, with one main difference. The temperature history it has seen
during the coking process can result in thermal annealing and graphitisation of the coke
crystalline structure [39].

Thermal annealing during coking leads to a typically higher degree of ordering of the
carbon than the parent coal(s). It also leads to structural changes of the coke within the
blast furnace itself (due to the thermal profile in the furnace). Coke can reside in the
blast furnace for long enough for significant changes in the crystalline structure of the
carbon to occur. Thermal annealing can also occur for coals and chars, increasing the
degree of ordering [52, 53]. Gas phase reactions in the blast furnace also can alter the
structure of the carbonaceous materials present, by reacting preferentially with the
amorphous carbon in the materials, leaving a relatively higher level of ordering.

A study investigating the thermal annealing behaviour of cokes [54], shows that
significant increases in degree of ordering (Lc) can be obtained by heat treating coke.
Experiments heating raw feed coke show the Lc can be increased 2-3 times in only half
an hour of heating, see Figure 2.14.

Figure 2.14: a) X-ray spectra and profile of the (002) carbon peak for a coke during
heating under neutral conditions; b) relationship between carbon structure (Lc,
angstroms) and temperature, from [54].

28
CHAPTER 2: Literature Review

The same study [54] also examined samples of coke taken from the tuyere level of
several blast furnaces. They found that the coke had indeed been extensively annealed,
with 2-3 times an increase of Lc. Some of the cokes proved more responsive to thermal
annealing than others, the possible reason for this is as yet unknown. The tuyere sample
cokes had slightly higher Lc levels than compared to the corresponding experimentally
annealed cokes, the authors concluding that some gasification may play a role in
determining the final coke structure. This finding may have implications for chars also.
During PCI, unburnt char will also experience conditions that may lead to thermal
annealing.

2.4 Coal Mineral Matter and Reduction Reactions

Coal mineral matter is the non-organic material that is introduced into the coal during
coalification (often as clay minerals). The mineral matter can be contained within the
coal carbonaceous structure (intrinsic) or separate from it (extrinsic) [55]. The most
common form of determining the coal mineral matter composition is to burn the coal,
which removes the carbon, some oxygen, water, and various hydrogen compounds from
the coal; the residual material from the mineral matter is called the ash. This study uses
the ash composition as the standard expression, as weight percentage of the simple
oxide forms. The ten oxides contained in coal ash, which are the most commonly
determined, are SiO2, Al2O3, CaO, Na2O, K2O, Fe2O3, TiO2, MgO, P2O3, and SO3. In
metallurgical studies coal is often considered to contain non-graphitic carbon and ash.
In high temperature applications such as the blast furnace, it is possible for the ash to be
molten, or partially molten. The ash will melt over a range of temperatures, the
temperature where it first starts to melt is known as the fusion point or temperature. The
temperature where the ash is fully molten is known as the liquidus temperature. These
temperatures and the melting range are determined by the composition of the ash.

The reactions that the ash components undergo when exposed to the combination of
high temperature and the conditions prevalent during carbon dissolution are important.
The next section will deal with the ash components that were considered most
influential in this study.

29
CHAPTER 2: Literature Review

2.4.1 Silica

Silica is often a large constituent in coal ash, being common to many clay minerals.
Silica can be present in the mineral matter in the forms kaolinite, illite, quartz, and
orthoclase. The reactions that silica can undergo in high temperature environments, in
the presence of carbon and iron, are of great interest to blast furnace ironmaking [18]

Within the blast furnace, there are several different sources of silicon. There is typically
silica in the slag, in the coke ash, and in any unburnt pulverised coal that may have been
injected. There are also several different routes for removal of silicon: in the slag as
tapped; in the liquid metal as solute silicon; and in the off gas as gaseous SiO [18].
Figure 2.15 shows the relationship between silica content in blast furnace coke and
temperature. From this figure it can be seen that significant amounts of silicon have
been lost from the coke by the time it reaches the tuyere level.

Figure 2.15: Change of SiO2 content in coke with increasing temperature in a blast
furnace, from [18].

Some of the possible reactions that can occur with silica present as an ash constituent
and at temperatures over 1773K are [18, 56, 57]:

SiO2 + C → SiO( g ) + CO( g )

Equation 2.6: Partial reduction to form silicon monoxide vapour.

30
CHAPTER 2: Literature Review

SiO( g ) + CO( g ) → Si + CO2 ( g )

Equation 2.7: Silicon transfer to carbon-free iron.

SiO( g ) + C → Si + CO( g )

Equation 2.8: Silicon transfer to carbon-containing iron.

SiO( g ) + 2C( s ) → SiC( s ) + CO( g )

Equation 2.9: Production of Silicon Carbide.

Another equation, which is well known in blast furnace ironmaking, is the Boudouard
reaction, and may be needed for the above reactions to occur.

C( s ) + CO2( g ) → 2CO( g )

Equation 2.10: Boudouard Reaction.

Equation 2.6 has been shown to occur readily within a blast furnace [18], with the
generation of a white gaseous product regularly observed.

Equation 2.7 is thermodynamically feasible but liquid iron in a blast furnace always
contains some carbon, and this route for silica reduction is believed to be negligible
compared with Equation 2.8 [18].

Other researchers [56, 58] have suggested an intermediary route within the silica gas
reduction, described in Equation 2.9.

Figure 2.16 illustrates the temperature dependence of the many possible intermediate
steps of carbothermic reduction of silica [58].

31
CHAPTER 2: Literature Review

Figure 2.16: Arrhenius plots for SiO and SiO2 reactions: A, SiO(s)→SiO(g); B,
SiO2+Si→2SiO; C, SiO2+3C→SiC+2CO; D, SiO+2C→SiC+CO (lid); E,
SiO+SiC→2Si+CO (loose lid); F, SiO+SiC→2Si+CO (tight lid); G,
2SiO2+SiC→3SiO+CO; H, SiO(s)→SiO(g) (lid), from [58].

Filisinger and Bourrie [58] determined that all of the equations listed in Figure 2.16
occur, but that equations A and B occur most rapidly, and were detected at the lowest
temperatures. They concluded that carbon does not reduce silica directly to Si, but via a
SiC intermediate. This study was not done in the presence of iron however, and it would
be expected that iron may change the possible routes.

From the Arrhenius plots shown in Figure 2.16 we can see that it is reasonable to expect
char and coke ash will begin evolving SiO(g) when they are heated (Equation 2.6).
Combined with Equation 2.10, the possibility of silica reduction from the ash resulting
in silicon transfer into the liquid iron is apparent. This mechanism has been assumed in
the blast furnace for coke ash [18], but the impact, if any, on the carbon dissolution has
not been reported.

2.4.2 Alumina

Alumina is one of the other major constituents of coal ash, as it is also high in clay
minerals. In respect to the conditions relevant to this study, it was determined that it is
unlikely that any appreciable reaction will occur between the liquid iron and the
32
CHAPTER 2: Literature Review

alumina [59] (see Figure 2.17), due to it’s high stability and oxygen affinity at these
elevated temperatures.

Figure 2.17:Ellingham diagram for oxides, from [59].

However, work studying the reaction and wettability of alumina with iron with various
oxygen potentials [60, 61] suggests that it is possible to form some intermediary phases,
such as FeAl2O4.

Figure 2.18:The phase stability diagram for the system Fe-O-Al. Mutual solubility
between Al2O3 and FeAl2O4 is not considered, from [62].

33
CHAPTER 2: Literature Review

Consideration of Figure 2.18 suggests that the intermediary phase FeAl2O4 will only be
formed at higher oxygen potentials than are likely in this study. Thus alumina may be
expected to act in a physical way only, influencing wettability and creating a physical
barrier to any reaction or mass transfer. It is possible for the ash to melt to form a slag,
with the melting point a function of the composition. Alumina is typically a constituent
of these slags, and can considerably influence the melting point.

2.4.3 Iron Oxide

Iron oxides are present in coal ash typically in lesser amounts than alumina or silica;
there are several possible reactions they can undergo.

The blast furnace has an atmosphere rich in CO(g) and CO2(g) due to the burning of coke
and supplementary fuels (such as PCI) and the Boudouard reaction (Equation 2.10).
This allows the following series of reactions to occur [18]:

Indirect Reduction
3Fe 2 O3 + CO( g ) → 2 Fe3 O 4 + CO2 ( g )

Equation 2.11

Fe3 O 4 + CO( g ) → 3FeO + CO 2 ( g )

Equation 2.12

FeO + CO( g ) → Fe + CO 2( g )

Equation 2.13

It has been shown [18, 63] that the reduction reaction of the iron oxide directly with the
solid carbon is negligible compared with the indirect mechanisms, Equation 2.11 to
Equation 2.13.

34
CHAPTER 2: Literature Review

PCO 2
Figure 2.19 illustrates the thermodynamic equilibrium values of PCO [64]. From this

diagram it can be seen that iron oxide may reduce in the solid or liquid phase, as this
equilibrium falls within the temperature range of interest in this study. Considering that
the atmosphere generated by the coke/char samples when heated would most likely be
in the reducing regime (direct reduction of silica will occur generating CO – see
Equation 2.6), this suggests that the iron oxide will reduce to liquid iron.

1800

LIQUID IRON
LIQUID OXIDE
1600

1400

1200
WÜSTITE
TEMPERATURE ºC

1000

MAGNETITE 2H
2O
METALLIC IRON =2
800 H
2 +O
2

600
HEMATITE

400

B OU
DOU
200 ARD
REAC
TION

-3 -2 -1 0 1 2 3 4 5 6 7
PCO
2
LOG PCO

Figure 2.19: Phase diagram for iron and its oxides as a function of oxygen potential
PCO 2
( PCO ), from [64].

The kinetics of this reduction has been extensively studied, however the studies
generally concentrated on temperature ranges less than those seen in the blast furnace
(<1550°C). It is thought that the reduction of FeO(l) by carbon containing iron occurs
via a two-step process [65, 66]:

CO + ( FeO) = Fe + CO 2

Equation 2.14

CO 2 + C = 2CO

Equation 2.15

35
CHAPTER 2: Literature Review

This gives the overall reaction:


( FeO) + C = Fe + CO

Equation 2.16

It is thought that the reaction occurs via the gas phase, as illustrated in Figure 2.20.

Figure 2.20: Idealised representation of the processes in the reaction of Fe-C droplets
with liquid slag, from [66].

This scenario would be relevant for when the char/coke ash and the liquid metal come
in contact, however as the interface is always more limited than the material available
for reaction, it may be expected that the solid direct reduction of the iron oxide with the
coke/char carbon will also occur.

Reactions between char and liquid slags containing iron oxide [12] suggest that the
reduction of the iron oxide in the slags occurs much more rapidly than the reduction of
the silica in the slags. As suggested by Figure 2.19, iron oxide in ash should also be in
the liquid form, so it is reasonable to expect the reduction of iron oxide in char/coke ash
to occur rapidly in-situ.

36
CHAPTER 2: Literature Review

2.4.4 Calcium Oxide

Calcium oxide is typically a minor constituent of coal ashes, but it is


thermodynamically feasible for reactions to occur between CaO and liquid Fe-C-S
alloys. These reactions are routinely utilized in steelmaking applications.

In the steel industry, lime has long been used to desulphurise steels to produce lower
sulphur contents [67, 68]. This desulphurisation reaction [69] is shown in Equation
2.17:

CaO( s ) + S + C sat → CaS ( s ) + CO( g )

Equation 2.17

It is typically considered that desulphurisation is well described by a first order,


diffusion-limited reaction [70], described by Equation 2.18.

d %S
= − K 1 (% S − % S *)
dt

Equation 2.18

CaO reactions with sulphur will invariably produce oxygen, which is then free to react
with other components in the system, for example silicon [70]. This is shown in a
concentration profile in Figure 2.21. Irons and Celik [70] suggest that the CaO when
reacted with the SiO2 in the system can then also form either 3CaO.SiO2 or 2CaO.SiO2,
both of which will reduce the silica activity and further reduce the oxygen activity. The
authors surmised that both the CaS and silicate reaction product will hinder solid state
diffusion, and slow further desulphurisation. In industry, studies of lime
desulphurisation have thus focussed on fluxing the reaction products with CaF2.

37
CHAPTER 2: Literature Review

Figure 2.21: Oxygen, sulphur, and silicon profiles in vicinity of CaO particle, from
[70].

Figure 2.22 [67] shows the stability of both CaS and CaO. Increasing sulphur in the
system increases the stability of CaS. Increasing calcium content, in the low
concentration regions (<10 ppm), increases the stability of CaS, but above ~10 ppm
calcium level has only a slight impact on CaS stability.

Figure 2.22: Stability regions of CaS and CaO in Fe-Ca-S-O system, from [67].

In carbon dissolution from chars and cokes into liquid iron, if calcium oxide reacts to
form products at the interface, then it could have a significant impact in altering the
interfacial area available for reaction. The calcium sulphide may also alter the
wettability behaviour, as well as desulphurising the liquid metal.

38
CHAPTER 2: Literature Review

2.5 Carbon Dissolution

Many researchers have studied carbon dissolution into liquid iron [5, 15-17, 47, 71-85]
and there is much that is still not understood about the many factors that influence
carbon dissolution rates. Some of the factors that have been identified are sulphur
content in the liquid metal, source carbon type, and source carbon ash composition. The
most thorough body of work is for graphite, with most of the observed effects being
understood [86]. However, for non-graphitic materials the factors relevant to carbon
dissolution are still being discovered. It is relevant to look initially at graphite, as it is
much less complex than the majority of the non-graphitic materials (coal, char, and
coke).

It may be useful at this point to briefly cover the methods typically used by researchers
to study the dissolution of carbon into liquid iron.

Ci

C
δ

Carbonaceous
Material Hot Metal

Figure 2.23: Variation of the Carbon Concentration at the Carbon Hot Metal interface
according to the theory of dissolution. Ci is the carbon concentration at the interface, C
the bulk carbon concentration, and δ is the interfacial layer thickness (Carbon
saturation Cs, is typically greater than Ci) from [5].

A schematic showing the interface during dissolution is shown in Figure 2.23. The
various methods for determining carbon dissolution are illustrated in Figure 2.24. The
method chosen for carbon dissolution is limited by material properties, for example coal
char cannot be made easily into a cylinder without the use of binders, which can alter
the carbon dissolution kinetics.

39
CHAPTER 2: Literature Review

Carburiser Rotating Rotating


Cover Cylinder Disc

Stationary Injection Raining


Rod Method

Figure 2.24: Experimental Methods for Carbon Dissolution Studies, from [5].

In the present study two carbon dissolution techniques were utilised, the carburiser
cover method and the sessile drop technique. The carburiser cover technique was
utilised to determine the apparent rate for carbon dissolution for the coke and char
samples. Of the available well-documented techniques, it is the only one well suited to
examining powdered carbon, and using a powdered form of char was considered
important in this work to reflect the nature and behaviour of PCI chars in the
ironmaking process. The carburiser cover technique provides a highly stirred situation
where both the carburiser and liquid iron are present in excess. It facilitates study of the
apparent rate constants for the dissolution of carbon from the various sources. The
second technique that was utilised to determine the transfer of carbon from chars/coke
to liquid iron was the sessile drop technique. It promotes the study of the interface by
creating transport-limiting conditions, which permitted the identification of the products
of interfacial reaction. The smaller scale sessile drop studies allow investigation of the
progress and products of interfacial reactions in relatively static conditions. This
technique has been used previously to study carbon transfer [14], but is not appropriate
for the determination of rate constants.

40
CHAPTER 2: Literature Review

2.5.1 Dissolution of Carbon from Graphite

Dahlke and Knacke [71] conducted one of the earlier studies into the dissolution of
graphite into Fe-C alloys. They found that the dissolution rate was consistent with the
process being transport controlled.

Olsson, Koump and Perzak [72] studied the rate of dissolution of graphite into molten
Fe-C alloys, with one of the focuses on the hydrodynamic conditions. The study was
undertaken using rotating graphite rods placed in the liquid metal bath. The results were
obtained as a function of the velocity of the cylinder in the melt. The results of this
study indicated that the dissolution of the carbon from the graphite was limited by the
diffusion of the carbon from the interface. They also found that the mass transfer
coefficient was linearly related to the peripheral velocity of the cylinder.

Kosaka and Minowa [73] used the rotating cylinder technique to investigate the
dissolution rate of graphite into molten Fe-C alloys. They also found that the mass
transfer in the liquid controlled the rate of carbon dissolution. Like the studies before
them, this investigation used the first-order rate equation to determine the mass transfer
coefficient, as shown in Equation 2.19.

= k (C s − C b )
dC A
dt V

Equation 2.19

In this equation k is the mass transfer coefficient, t is time, A is the contact area, V is
the volume of the bath, Cs is the saturation concentration, and Cb is the bath
concentration. By integration, Equation 2.19 can be used to calculate the mass transfer
coefficient:
(C s − C o ) A
ln =k t
(C s − Ct ) V
Equation 2.20

41
CHAPTER 2: Literature Review

In this equation Co is the carbon concentration initially, and Ct is the concentration at


time t.

The study by Kosaka and Minowa [73] also was concerned with the physical properties
of the liquid metal and the hydrodynamic conditions. The mass transfer from the
rotating cylinder can be estimated using non-dimensional correlations of the Stanton
(k/U, where U is the relative peripheral velocity of the cylinder), Reynolds (LUρ/µ
where L is the characteristic length, ρ is the density of the liquid, and µ is the viscosity
of the liquid) and Schmidt (µ/ρD, where D is the mass diffusion coefficient) numbers.

The heat of solution of carbon into liquid metal was also studied [73]. The dissolution
of carbon is an endothermic process and as such lowers the bath temperature. The
dissolution occurs according to the following equations:

C ( gr ) → C sat .

Equation 2.21

[
∆H = 22.60 + 24.32 1 − N Fe
2
] (kJ / mol )

Equation 2.22

This endothermic heat will cause a temperature gradient to form in the liquid adjoining
the dissolution interface. For the experiment conducted by Kosaka and Minowa [73],
the temperature drop was found to 7oC or less, and thus the authors considered it could
be disregarded.

The previous studies mainly considered the dissolution of graphite only in forced
convection situations, whereas Kosaka and Minowa [73] also measured the mass
transfer coefficient with a stationary cylinder. They concluded that the rate-controlling
step was still the diffusion of the carbon in the liquid boundary layer generated by
natural convection.

42
CHAPTER 2: Literature Review

The effect of ash in graphite on its dissolution has been illustrated [47]. It was seen that
the ash content had little effect. This study determined the degree of crystalline ordering
in the graphite as a factor that influenced the rate of dissolution, with more-highly-
ordered materials dissolving more quickly than the less-highly-ordered.

Another, more recent study, modelled the carbon dissolution from graphite into liquid
iron using Monte Carlo simulations [34, 87]. The interface between the graphite and the
Fe-C melt was modelled (see Figure 2.25) and it was seen that this model suggests that
there is an interfacial region formed where Fe and C atoms co-exist, and that the
dissociation of the graphite atoms from the carbon matrix is rapid.

Figure 2.25: Atomic distribution profile across graphite/Fe-C interface before and after
simulation. Z corresponds to the layer number normal to the interface, from [87].

2.5.2 The Effect of Sulphur on the Dissolution of Graphite

Once the mechanisms and rate-controlling steps involved in the dissolution of graphite
into liquid Fe-C alloys were confirmed by a majority of researchers, the effect of adding
sulphur into the melt was considered as the next field of study.

Grigoryan and Karshin [74] studied the effect that sulphur has on the dissolution
kinetics of graphite. The rotating disc method was used with rotating discs of graphite
being placed into the liquid melt. The influence of surfactants such as sulphur and
oxygen were investigated. Their findings suggested that when the dissolution of the

43
CHAPTER 2: Literature Review

graphite is not controlled by diffusional mechanisms (low stirring rate), the additions of
surfactants to the melt reduce the dissolution rate. Grigoryan and Karshin concluded
that these results could be satisfactorily explained by an absorption isotherm. When
dissolution is controlled by diffusional mechanisms, the addition of sulphur was found
to have no effect. This implies a different rate-limiting mechanism when the system is
under chemical reaction control, than when it is under diffusional control.

Similar experiments were conducted by Ericsson and Melberg [75], using the rotating
cylinder technique. Loss of carbon from the melt was noted, and assumed to be due to
the formation of carbon sulphides (CS and CS2). This study also found a decrease in the
carbon mass transfer with increasing sulphur content. Ericsson and Melberg found that
the experiments with higher sulphur contents also conformed to Equation 2.20. From
this they suggested that the system was again mass transfer controlled.

Ericsson and Melberg [75] concluded that the effect of sulphur on the carbon
dissolution rate was at least partly to due to the decrease of carbon diffusivity at high
sulphur contents. They postulated that the sulphur influences the interfacial reaction
kinetics of dissolution.

Shigeno et al. [76] studied the influence of sulphur and phosphorus on the dissolution of
two types of electrode-grade carbon (non-graphitic and graphitic carbon). The rotating
cylinder technique was used. Sulphur was again found to decrease the rate of carbon
dissolution. The equation describing the dissolution as found in this study (for Reynolds
numbers 4x103<Re<1.3x105) is:
1

ShSc 3
= 0.051 Re 0.78

Equation 2.23

In this equation Sh is the Sherwood number, Sc is the Schmidt number, and Re is the
Reynolds number.

The effect of the crystallinity of the graphite specimen was noted in this study [76], with
dissolution of the non-graphitic carbon greatly retarded by even small amounts of
sulphur. They suggested that sulphur is selectively adsorbed onto the prism plane of the
44
CHAPTER 2: Literature Review

graphite, which hinders the dissolution from this plane, reducing mass transfer in the
presence of sulphur. The non-graphitic carbon planes corresponding to the prism planes
are more exposed than those of graphitic carbon, thus the reduction of the effective
surface area available for dissolution is much greater. This is shown in Figure 2.26.

Figure 2.26: Influence of sulphur concentration on overall mass transfer coefficient in


using graphitic and non-graphitic carbon, (N is the rotational speed), from [76].

Orsten and Oeters [77] studied the dissolution of graphite into Fe-S melts, using the
rotating cylinder technique. They confirmed the findings of Shigeno et al. [76]. This
study explains the effect of sulphur on the diffusion coefficient of carbon in liquid iron.
Their results on the dissolution of non-graphitic carbon sources illustrate the observed
magnification of the effect of sulphur when present in non-graphitic dissolution.

Wright and Baldock [78] studied particulate graphite injected into Fe-C melts. They
found that the temperature of the bath (1400-1500°C) had little effect on the dissolution
rate. Particle size of the graphite had little effect on the dissolution rate, as the
dissolution kept pace with the injection up to 95% carbon saturation. The effect of
sulphur in the melt was seen to decrease the point at which the dissolution rate follows
the rate. At 0.1% sulphur the dissolution rate keeps pace with the injection rate up till
87% carbon saturation, and at 1.0% sulphur, it kept pace until 62% carbon saturation.
The results agree with previous studies. Wright and Baldock suggest that it is the
decrease in the carbon diffusivity that affects the rate of dissolution. They stated that it

45
CHAPTER 2: Literature Review

has been seen that dissolution can satisfy mass transport formulations, thus the decrease
in k with increasing %S must correspond to a change in the kinematic viscosity or
diffusional properties of the melt. They also suggested at higher levels sulphur might
influence the interfacial kinetics of the process.

Wright and Baldock [78] found that the addition of sulphur also reduces the carbon
saturation point; thus the sulphur will tend to reduce the driving force for the mass
transfer. The carbon saturation point [88] can be taken as:
C s = 1.3 + 0.00257T - 0.31Si - 0.33P - 0.45S + 0.28Mn

Equation 2.24

Here, T is the temperature in degrees Celsius, Si is the % silicon, P is the % phosphorus,


S is the % sulphur, and Mn is the % manganese (all percentages are in weight percent).

Sahajwalla et al. [79] studied the dissolution of carbon into liquid iron, and the
influence wettability has on the dissolution process. The wettability of graphite with Fe
melts of varying sulphur contents was studied, with the conclusion that for graphite the
effect of sulphur is much less significant than for non-graphitic carbon. Increasing the
sulphur in the metal increases the contact angle slightly, thus decreasing the contact
area.

In the many previously studied investigations on the effect of sulphur on dissolution rate
of graphitic carbon in iron melts, the mechanism has not been conclusively determined.
A Monte-Carlo simulation study [80] conducted on the dissolution of carbon from
graphite into liquid Fe-C-S melts found that while increasing sulphur in the melt did
depress the carbon dissolution rate, a model of the graphite-iron interface did not show
sulphur present, see Figure 2.27.

46
CHAPTER 2: Literature Review

Figure 2.27: Atomic distribution profile across graphite/Fe-S interface before and after
simulation. Z corresponds to the layer number normal to the interface, from [80].

The authors suggested that the influence of sulphur on interfacial kinetics was due to
sulphur lowering the carbon dissociation rate. Carbon and sulphur were found to be
mutually repulsive in the melt, and tend to displace each other. The authors also
suggested that there was a significant delay at initial contact, while the carbon
dissociated from the graphite (once this had begun there was no more delay as the
graphite was a single crystal). This study was purely theoretical, and considered a
single-crystal of graphite aligned with the basal plane in contact with the melt. This may
have a significant influence when comparing to experimental situations where the
graphite consists of many small crystallites aligned randomly to the liquid melt.

The above study refutes the site blockage mechanism; however it is known that sulphur
will change the wettability of the liquid with the carbon, reduce the carbon saturation
level, and reduce the diffusivity of carbon in liquid iron. It is possible that these
mechanisms act, to a lesser extent, with the resistance to dissolution caused by the
mutual exclusivity of carbon and sulphur, to reduce the carbon dissolution rate.

47
CHAPTER 2: Literature Review

2.5.3 Dissolution of Carbon from Non-Graphitic Sources

Non-graphitic carbon sources include coke, coal, coal char, and glassy carbon. Most
authors have used a selection of several carbon sources in the dissolution experiments.

Wright and Denholm [81] investigated the dissolution of graphite, black coal char and
petroleum coke, using the injection method. They found that the dissolution rates of the
different materials were similar. The dissolution rates increased linearly with
turbulence, and particle size had little effect on the dissolution rates.

Shigeno et al. [76] found different results, noting that the dissolution rate from non-
graphitic carbon was much less than that from graphitic carbon. They also identified the
enhanced effect of sulphur on non-graphitic sources, which they suggest is due to the
carbon structure.

Orsten and Oeters [16] investigated blowing coal particles into liquid iron. Ash was
found to inhibit dissolution by blocking the surface of the liquid from the carbon source.
If there was insufficient stirring and the temperature was high enough (above the ash
liquidus temperature), the liquid ash could also inhibit carbon transfer. The increased
effect of sulphur on non-graphitic dissolution over graphitic dissolution was observed.
The authors proposed that the effect of sulphur is due to formation of heterocyclic –C-
S-C bonds. With the dissolution of coal, the authors suggested the two steps that take
place:

1. Degassing (devolatilisation) of the coal particle, allowing only partial (or no) contact
of the particle with the melt.
2. Suspension of the particle within the melt and subsequent dissolution.

Ganguly and Reid [82] investigated the dissolution of coals using the rotating cylinder
and rotating disk methods. They deduced the devolatilisation step as limiting the
dissolution and also recognised that the heat transfer from the melt to the particle and
momentum transfer governs the dissolution kinetics.

48
CHAPTER 2: Literature Review

Mourao et al. [15] studied the dissolution of graphites and industrial cokes using the
stationary rod method. Sulphur was found to reduce the dissolution rate. The authors
suggested that the retarding effect was more pronounced in coke than in graphite
because of the differences in the surface structure. It was stated that it was the high
degree of graphitisation in graphite that offered less active sites for sulphur adsorption.
In the light of the previously mentioned study [80], which refutes the possibility of
sulphur accumulating at the interface, this seems unlikely. However, it may be an
indication that the influence of sulphur on carbon dissolution is controlled by carbon
dissociation. As the crystallite size in non-graphitic carbon is typically much lower than
for graphitic carbon, initiation of carbon dissociation will take much longer.

This study [15] also revealed the presence of a thin ash layer formed on the coke surface
after dissolution had partially occurred. It was concluded that both mass transfer and
phase boundary reactions controlled the dissolution rate.

Wright and Taylor [83] formulated a model for the dissolution of particulate
carbonaceous materials injected into an iron bath. The model predicts the experimental
results when mass transfer limitations dominate (for pure Fe and Fe-S up to 1.0% S).
The model does not apply to non-graphitic sources, suggesting that factors other than
mass transfer can be significant in limiting the rate of non-graphitic dissolution into iron
melts. This study’s interpretation of the mechanism by which sulphur reduces the
dissolution rate is that the sulphur (being surface active) inhibits the entrance of carbon
particles into the melt by changing the wetting characteristics of the gas-liquid interface.
Again, this conclusion is based on previous postulations and not on direct evidence in
this experimental study.

Wright and Taylor [83] proposed that when the sulphur levels were low both liquid-side
mass transfer and interfacial effects limit the dissolution rate. At higher levels of
sulphur the rate is limited by mechanisms other than mass transport from solid to liquid.

Sahajwalla, Taylor, and Wright [84] injected anthracites and brown coal char into liquid
Fe-C-S melts. This study found that increasing the bath sulphur content decreased the
dissolution rate, and that each carbon source had a different relationship between bath
sulphur and dissolution rate. A possible explanation for the difference in dissolution
49
CHAPTER 2: Literature Review

rates for graphite and non-graphitic sources, and for the influence of sulphur, is given. A
carbon atom in an aromatic ring is stable, and thus when one is removed from the ring it
is subsequently easier (energetically) to remove further atoms. Thus, with reference to
the earlier section on carbon structure, the larger the layer or region of crystallinity, the
less energy is required for dissolution. With graphite the regions are large, and thus it is
easier to dissolve, but for non-graphitic carbon sources the regions are usually less than
for graphite and thus dissolution is retarded. When the surface of non-graphitic particles
is resistant to dissolution a reinforcing effect of structure and sulphur content can arise.
This is due to the surface activity of sulphur (and subsequent site blockage). This
finding could be reinterpreted, based on the previously mentioned study [80], to replace
the mechanism of sulphur site blockage with the effect sulphur has on dissociation rate.

Sahajwalla et al. [79] used glassy carbon as a carbon source in the study. The injection
method was used to investigate the dissolution performance of graphites and glassy
carbon into liquid Fe-C-S melts. It was found that the dissolution of glassy carbon into
Fe-C-S melts was independent of feed rate, and thus is dictated by interfacial kinetics. A
large variation in the dissolution rate constant was found for various carbonaceous
materials and the dependence of this constant on the sulphur content of the liquid was
seen to vary significantly. It was proposed from this that, at high sulphur levels, it is the
change in wetting characteristics that influences the dissolution kinetics, causing
interfacial kinetics to dominate over liquid phase mass transfer.

Belton and Fruehan [66] proposed a similar explanation on the difference seen between
graphitic and non-graphitic carbon dissolution rates. The authors were of the opinion
that it was the change in the wetting for the carbon/melt system that occurs when
sulphur is present that is the cause of reduction in dissolution rates. This view did not
mention the possible role of the dissociation rate of carbon being one of the limiting
factors in carbon dissolution kinetics. The idea that the initiation of carbon dissolution
from a single crystallite takes more time than the continuing dissolution from the same
crystallite seems plausible. Extending this idea to include the crystallite size of the
different carbons, it follows that the non-graphitic carbons will take longer to dissolve,
as they will require more initiation events than graphitic carbon.

50
CHAPTER 2: Literature Review

Further work has been done on the influence of carbon structure on dissolution
performance [47, 85]. It was determined that as for the graphite case, the degree of
ordering significantly affects the dissolution kinetics.

From the studies discussed in this section (and previous sections) it is clear that there
are three clear mechanisms that have been identified during carbon dissolution (both
graphitic and non-graphitic). These are:
1. The rate of dissociation of carbon, or what can be considered the inherent rate of
carbon dissolution
2. Interfacial effects, including wettability (discussed in section 2.6), interfacial
reactions and new phase formation at the interface
3. Mass transport phenomena
It is this author’s opinion that the rate of carbon dissolution for any given carbon is a
function of the relative strengths of each of the above mechanisms.

The rate of dissociation of the carbon from the solid material into the liquid iron can be
expressed as two steps [34], shown below:

Step 1:
C ( Lattice ) ⎯dissociati
⎯ ⎯⎯ on
→ C (int erface )

Equation 2.25: Dissociation of carbon atoms from its lattice site into the carbon/iron
interface.

Step 2:
C (int erface ) ⎯transfer
⎯⎯→ C ( bulk )

Equation 2.26: Mass transfer of carbon atoms from the interface into bulk liquid
through the liquid boundary layer.

The difference between non-graphitic and graphitic dissolution in one study [34] was
suggested to be because of the relative rates of the two materials for step 1. In coals, the
atomic dissociation rate is expected to be low due to their complex structure, whereas

51
CHAPTER 2: Literature Review

the rate for graphite is faster due to the weak Van der Waals bonds that hold the
graphite planes together.

The second mechanism that can be seen to act during carbon dissolution is interfacial
phenomena. This is a complex field of study as it includes not only the effects
associated with wettability, but reactions that may occur at the interface and also
accumulation of one or more phases at the interface such as solid or liquid ash left after
the carbon has been removed. The influence of wettability may be very strong, as it will
dictate the actual area of interaction between the two materials. Some of the reactions
that may occur at the interface are discussed in section 2.4, however the role of these
reactions has not been considered previously in carbon dissolution literature. The
formation of ash at the interface has been considered somewhat more (and is discussed
in detail in section 2.5.4) however until now, ash has been seen as a physical barrier
only and the reactions between ash and the liquid Fe-C alloy has not been take into
account.

The final mechanism discussed is mass transport. It is agreed in literature that the
dissolution rate for carbon from graphite into pure iron is mass transport controlled (on
the liquid side). It is only when the carbon dissolution rate is high and there are few or
no interfacial effects limiting the rate that the limitations of mass transport become
controlling.

In review of the previous literature regarding the dissolution of non-graphitic materials


into Fe-C and Fe-C-S melts it is clear that a great deal of work and analysis is required
to clarify the observed behaviour. Structure has been proposed as a possible influence,
as has sulphur-related changes to the melt properties, and the change in wetting
properties of the carbonaceous materials. Further work investigating the dissolution
behaviour of one family of carbonaceous materials should help to clarify whether and
how much each of the suggested mechanisms influences the dissolution kinetics.

52
CHAPTER 2: Literature Review

2.5.4 Influence of Ash on Carbon Dissolution

Ash within the carbonaceous material can influence the dissolution of the carbon into
the liquid metal, by either chemically reacting with the liquid iron or physically forming
a barrier to mass transport. The reactions possible for the ash are covered in an earlier
section; however, a few researchers have considered the ash interactions during carbon
dissolution.

A study on the effect of ash composition on coke dissolution rate [16] investigated the
effects of CaO and CaF2 additions on dissolution rate. The researchers found that these
materials increased the dissolution rate of carbon, and attributed this increase to the
reduction of the liquidus temperature of the coke ash by these additives.

Gudenau et al. [5] analysed the interactions of industrial and special cokes with liquid
iron, and found the role of ash to be an important one. It was suggested that phases that
lower the liquidus temperature of the coke ash would aid in carburising by allowing the
ash to be free moving and able to move away from the interface, providing fresh carbon
to the interface. The study found however, that the additions of ash components such as
CaO, MgO, SiO2 and Al2O3 decreased the carbon transfer rate, and that iron oxide
enhanced the carburisation rate. An index called the ash factor was introduced to
indicate the quality of the coke as a carburiser. Ash fusion temperature was cited as an
important issue, but the researchers state that it cannot be the only property influencing
carbon dissolution rate.

Another study of blast furnace coke carburisation [15] that utilised the stationary rod
approach found the formation of an ash/slag layer at the interface to significantly
influence dissolution rate of carbon into hot metal. The analysis of the slag like layer on
the coke/metal interface shows it was mainly alumina and silica, with smaller amounts
of CaO, TiO2, Na2O, K2O and P2O5. An analysis of the initial coke ash chemistry was
not supplied however, so it is not possible to determine if the ash layer has in any way
interacted with the liquid iron or coke material, altering its composition. The study
concluded by saying that the viscous slag layer acted to reduce dissolution rate by

53
CHAPTER 2: Literature Review

reducing the surface area available for dissolution. The possible reactions occurring
between the ash and the liquid iron were not considered.

The idea of the liquidus temperature of the coke ash playing a major role in carbon
dissolution is not new, however experimental results have yet to clarify the exact role
ash has to play in the dissolution of carbon from non-graphitic sources. The addition of
CaO for example is seen to increase dissolution rate in one study [16] and decrease it in
others [5, 17]. These results were found for coke, and suggest that the liquidus
temperature of the coke ash is not the only factor limiting the dissolution rate of the
carbon from the coke. Further work is needed to fully understand these phenomena.

Orsten and Oeters [16] studied the influence of additives to coke, and concluded that it
was the liquidus temperature of the coke ash that limited carbon dissolution. The results
found in this study however, suggest further complexity; cokes studied that had the
same ash liquidus temperature (same parent coals with different additives) were found
to have different dissolution rates. This implies that the nature of the additives
themselves plays a role in determining carbon dissolution kinetics, not just the resulting
ash fusion temperature.

Ash is a factor in carbon dissolution that is poorly understood. Consideration of ashes


role has been previously limited to it’s physical presence, and the idea of chemical
reactions influencing carbon dissolution has not been taken into account. Ash in non-
graphitic carbon is typically of the order of 10-30% by weight, and this level increases
if you consider partially combusted and reacted material. This means that ash reactions
could conceivably have a very significant impact on mass transfer and interfacial
kinetics.

2.6 Wettability of Carbonaceous Materials

In determining the interfacial area of contact between a liquid and a solid, wettability
must be considered. As discussed previously, the carburiser cover method has an
interfacial area that is not consistent but will change for different materials.
Consideration of the wettability will be required to estimate the interfacial area of

54
CHAPTER 2: Literature Review

contact between the carbonaceous materials and the liquid iron. There have been many
studies on the wettability of pure materials with liquid iron; however there have been
fewer on the wettability of more complex systems such as graphite/iron, and none on
non-graphitic carbons.

2.6.1 Wettability Phenomena

Wettability is typically characterised by contact angle or surface tension. The contact


angle is defined as the angle made at the three-phase point of the liquid-solid-vapour
phases measured between the liquid and the solid (see Figure 2.28).

VAPOUR
γLV

LIQUID θ

γLS O γVS
SOLID

Figure 2.28: Sessile drop showing definition of contact angle and the surface tension
forces acting on the drop.

The contact angle is characteristic for any system of solid-liquid-gas, and it is defined
by Young’s Equation [89, 90]:
σ SV + σ SL
cos θ =
σ LV

Equation 2.27

Where σSV, σSL, σLV, are interphasic specific surface energies at the solid-vapour, solid-
liquid, and liquid-vapour interfaces [90]. Temperature, compositions of the solid, liquid
and vapour, time, and the surface finish of the solid can affect contact angle. The
55
CHAPTER 2: Literature Review

conventions for wetting and non-wetting systems are [91] if θ > 90° the system is
considered non-wetting and if θ < 90° the system is wetting.

Surface tension (and hence wettability in general) is very sensitive to changes in


temperature. The surface tension of a pure metal is known to decrease with increasing
temperature, becoming zero at a critical temperature. Additions of solute elements into
the liquid iron complicate the determination of surface tension, with the effect of
composition needing to be taken into account. The surface tension of the liquid-vapour
phase boundary between iron and vapour can be estimated by Equation 2.28 below [92]:

γ LV = 2184 − 0.107T ln(1 + k S aS )

Equation 2.28

The term kS from the above equation is defined as:


9960
log k S = − 2.75
T

Equation 2.29

The activity of sulphur, aS is estimated from [93]:


aS = wt % S × γ

Equation 2.30

Here γ is the activity coefficient, defined by:


γ = exp(0.36 × (wt %C − CS ))

Equation 2.31

CS is the saturation carbon level, and can be calculated from Equation 2.24.

Temperature gradients in the droplet can set up flow patterns within the drop, known as
the Marangoni Effect [94]. This liquid motion is more vigorous near the surfaces of the
droplet and can promote mass and heat transport across interfaces.
56
CHAPTER 2: Literature Review

2.6.2 Effect of Chemical Reactions on Wettability

Wettability determinations for non-reactive systems are relatively simple; however the
wettability of reactive systems is more complex. There are two classes of chemical
interaction during wettability, adsorption and chemical reaction. These occur when a
liquid is in mechanical and thermal equilibrium (at the same temperature and pressure)
but is not at chemical equilibrium (that is, the chemical potentials are not equal) with
the solid and gaseous phases.

Adsorption involves the attraction of surface-active species to the two-phase interface.


There is no mass transfer across the interface when the only chemical interaction is
adsorption [91]. The time to (apparent) equilibrium will depend on the diffusion rates of
the active species to the surface. At high temperatures (such as found in liquid metals on
a refractory substrate), the rate of diffusion would usually be quite high. Thus, the
contact angle of the liquid on the solid would vary with time (because the surface
energies, and thus surface tension would vary with changing chemical composition).

There are three types of chemical reaction that could occur in a solid-liquid-vapour
system: a new phase produced at the boundary of the solid-liquid; a dissolution of the
substrate (or some component in the substrate) into the liquid; and a dissolution of the
liquid (or some component of the liquid) into the substrate [95].

When a system is in chemical equilibrium (i.e., the chemical potentials of all


components of the system are equal), the wetting behaviour is relatively simple to
characterise. The change in system properties is only due to the system attaining
mechanical and thermal equilibrium [95]. However, when there is a reaction taking
place in the system, the chemical potentials are not equal, and the system is in a state of
thermal, mechanical and chemical non-equilibrium.

When surface-active elements are present at an interface, they may alter interfacial
energy at that interface [96]. Surface activity is defined as the magnitude of change in
the surface tension brought about by an addition of a unit quantity of the species. The

57
CHAPTER 2: Literature Review

ultimate surface activity of a surface-active species in a binary system generally


decreases with increasing temperature. Additions of more than one solute to a system
complicate the determination of the surface activity, as the effects cannot be predicted
from the individual effects [97]. For example, carbon, which is inactive in liquid iron,
lowers the surface tension of solutions of sulphur in iron, because it increases the
activity coefficient of the dissolved sulphur so that the interface concentration is
increased.

2.6.3 Wettability of Carbonaceous Materials with Liquid Iron

There has been no study to date for the wettability of char or coke materials and liquid
iron. There have been limited studies on the wettability of graphite with liquid iron, and
these are discussed below.

The wettability of graphite with liquid iron was studied [98], with the dynamic wetting
behaviour shown in Figure 2.29.

Figure 2.29: The variation of the contact angle with time in the sessile drop experiment
for a melt/graphite assembly, from [98].

In this study it can be seen that reactions at the interface can change the wettability of
this system. The change in contact angle is dramatic, going from a non-wetting situation
(~100°) to good wetting (~60°). The authors attributed this change to carbon and iron
inter-diffusing into the graphite and the liquid iron. The effect of sulphur on the

58
CHAPTER 2: Literature Review

wettability of iron with graphite was also studied, at it was seen that the sulphur, which
can reduce the surface tension of the melt, acts to increase the contact angle. This can be
seen in Figure 2.30. It can be seen that additions of sulphur up to 0.37 wt% can change
the final wettability significantly, altering the angle from wetting to non-wetting.

Figure 2.30: Variation of the contact angle with time for Fe-C-S melts with different
sulphur concentrations (C pct = 2.0), from [98].

A study on the wettability of natural graphite with iron [14] showed some interesting
variations of contact angle with time.

120

110
Contact Angle (deg)

100

90

80

70

60

50

40
1 10 100 1000 10000
Time(sec)

Figure 2.31: Change of contact angle with time for iron on natural graphite substrate,
from [14].

59
CHAPTER 2: Literature Review

Figure 2.31 shows the change of contact angle with time for a pure iron droplet on a
natural graphite substrate. It is clear that the dynamic wetting behaviour between iron
and natural graphite is very different from that reported previously in iron/pure graphite
wetting system. This indicates that the interactions involved in the two systems are very
different.

As mentioned in the earlier sections, reactions taking place at the interface can
significantly influence the wettability of the system. The system of interest in this study
is quite complex: the carbonaceous material contains ash minerals as well as non-
graphitic carbon, the liquid metal typically contains carbon, sulphur, and quite a few
trace elements, and the composition of the chars and cokes is typically heterogeneous.
In combination, this can lead to the possibility of many reactions occurring at the
interface. Apart from the carbon dissolution, ash reactions and mass transfer from char
to metal can occur, significantly interfering with interfacial analysis.

2.7 Summary

In increasing the rate of pulverised coal injection in the blast furnace, it has become
clear that more understanding is required regarding the interactions that the produced
unburnt char can undergo in the blast furnace. Currently, an appreciation of the
reactions char can undergo with liquid iron is incomplete, with only a few limited
studies relating to carbon dissolution from non-graphitic materials.

Carbon dissolution from graphitic carbons into liquid iron is reasonably well
understood, with the majority of researchers accepting that the mechanism limiting the
rate of dissolution of graphite into liquid iron is the mass transport in the liquid. The
addition of sulphur in the liquid iron was seen to reduce the dissolution rate of carbon,
however the reason for this is not yet resolved. The factors that were considered to
influence carbon dissolution from graphite into liquid iron were the structure of the
carbon material, the composition of the liquid iron with particular emphasis on the
sulphur content, stirring in the liquid (to prevent mass transport considerations from
playing a role) and interfacial phenomena such as wettability.

60
CHAPTER 2: Literature Review

The dissolution of carbon from non-graphitic sources has been examined less
thoroughly. The structure of the carbon or a change in wetting characteristics was
thought to play a role in the dissolution rate. It was seen that the effect of sulphur was
much more pronounced for non-graphitic carbons than for graphite. In the review of the
dissolution studies, it was identified that there were three types of mechanisms that can
limit the dissolution of carbon into liquid iron:

1. The rate of dissociation of carbon, or what can be considered the inherent rate of
carbon dissolution
2. Interfacial effects, including wettability (discussed in section 2.6), interfacial
reactions and new phase formation at the interface
3. Mass transport phenomena

In consideration of the factors that can limit the kinetics of carbon dissolution into
liquid iron, we can see that one or more phenomena may occur simultaneously. As yet
an integrated study of carbon dissolution has not been conducted which takes into
account the above-mentioned factors. The following considerations have not been dealt
with in previous studies:
a. The identification of the interfacial reactions such as reduction of oxides
and the formation of new phases at the interface,
b. Whether there is any interdependence of the reactions occurring with
carbon dissolution and other mass transfer,
c. The extent wettability may influence the interfacial area of contact,
d. Mechanisms that can explain the differences in carbon dissolution rates
seen between non-graphitic materials.

61
CHAPTER 3: Objectives and Study Parameters

CHAPTER 3: OBJECTIVES AND STUDY PARAMETERS

3.1 Objectives

An extensive literature survey was presented in Chapter 2, focussing on pulverised coal


injection, carbon materials and their characterisation, reduction reactions, carbon
dissolution, and wettability. This project aims to achieve a comprehensive
understanding of the dissolution behaviour of coal-chars into liquid iron during PCI.
Within the broad regime of carbon dissolution, the following thrust areas were
identified for a detailed investigation.

1. High injection rates for pulverised coal injection (PCI), desirable for increasing the
efficiency of the furnace, can seriously affect furnace stability with an increasing
amount of unburnt char entering the blast furnace burden. The consumption of
unburnt residual char can take place through reactions with the metal, slag or the
gaseous phases. While Lu [10] investigated the reactions of char with gaseous
phases in detail, Mehta carried out an in-depth investigation on char-slag reactions
[12]. The main focus of this study is on the interactions of coal-chars with liquid
iron, specifically carbon dissolution and interfacial phenomena. The full impact of
coal type on furnace stability at high PCI is not very well known. To avoid a build
up of unconsumed char in the furnace, especially at high PCI rates, it is important to
discriminate between those reactions and char types that favour consumption. The
dissolution of carbon from chars may make significant contributions to the
carburisation of liquid iron and therefore could help in achieving high rates of PCI.
In addition, the reactions between coal char and liquid iron could influence the
dissolution reactions and therefore the accumulation of unburnt char within the blast
furnace.

2. As already noted in section 2.1, char fines can have a strong influence on the blast
furnace operations. A recent study by Nightingale et al. [4] identified the condition
and flow of the coke bed through the deadman and hearth to be an area requiring
further investigation. They observed that the state of the coke bed in the deadman
and hearth can be related to the carburising ability of the coke. If the liquid hot

62
CHAPTER 3: Objectives and Study Parameters

metal reaches the deadman and hearth unsaturated, then it has the ability to consume
any fines that may have accumulated there. If the hot metal is close to saturation
then the fines could accumulate, deteriorating permeability. It was further observed
that the fines that enter the deadman and hearth should be consumed by the liquid
metal (and other phases) in preference to the coke bed, maintaining the area clear of
fines. A clear understanding of factors that may influence preferential consumption
of chars as compared to coke is therefore of vital importance for stable PCI
operation at high injection rates. Experimental investigations were therefore planned
to include coke in this study along with a number of chars, so that an objective
comparison could be made between the two types of materials. As the results on
coke dissolution are more fully documented in the published literature as compared
to coal-chars, experiments on coke would therefore help in standardising results for
coal-chars. Having carbon dissolution rates for coke measured in a manner similar
to chars will assist in validating the techniques used and compare the respective
reaction rates under comparable conditions.

3. In section 2.5.3, it was observed that while some non-graphitic materials had been
investigated, no study had comprehensively covered one type of material in a
systematic way. Previous studies had investigated the comparative performance of
coke, natural graphite, glassy carbon and coals. This approach to the study of carbon
dissolution, while appropriate to the initial stages of development, does not bring
out the finer details involved in the dissolution kinetics and interfacial reactions
occurring in these types of materials. Coal-based materials have additional problems
relating to their innate heterogeneity, complexity of structure and composition, and
the broad disparate fields of study surrounding coal based materials.

The earlier studies have failed to find a convincing mechanism to explain the
differences seen in carbon dissolution rates between cokes and coal chars. While the
theories based on ash content and carbon structures were able to partially account
for some of the differences, these could not explain observed differences in carbon
dissolution kinetics for carbon material having similar ash content and carbon
structure. A systematic and comprehensive study was therefore necessary and was
planned in this project with an aim to get a better understanding of carbon
dissolution from non-graphitic materials.
63
CHAPTER 3: Objectives and Study Parameters

4. As a carbon source material, PCI char is produced in conditions with very high
heating rates, which are difficult to produce and handle. Further, by its nature it
must be in a powdered form, and this precludes the use of the standard dissolution
techniques such as the rotating disc or rod methods. These techniques have the
advantage in that they are easily controlled and the interfacial surface area of the
reaction is easily determined. These techniques however cannot be used with a
material that is in powdered form, unless a binder is used to form the char powder
into a disc or a rod. As binders could interfere with the overall dissolution kinetics,
it was decided to use the carburiser cover method for these studies. This method has
the additional advantage of being large in scale, so the effects due to the
heterogeneity of char materials would be minimised. This technique however has a
few disadvantages. It requires large amounts of carbonaceous materials and the
production of char in a drop tube furnace is a time consuming process. In addition,
the results as determined from this method require an accurate estimation of the
interfacial contact area between the char particles and the liquid iron bath. A
theoretical model will be developed in this project for estimating the interfacial
contact area based on a force balance approach, taking into account the relative
densities of the char and the iron, temperature, surface tension, and size of the char
particles. Once the interfacial area is calculated, the results from carburiser cover
method can be used to produce kinetic rate constants and a comparison with
published literature could be carried out. The utilisation of carburiser cover methods
has in turn necessitated wettability studies to be carried out for the carbonaceous
materials used in this study.

5. With an aim to identify the mechanisms influencing reactions occurring during


carbon dissolution from non-graphitic sources into liquid iron, it was also planned to
identify the reactions occurring in the interfacial region. The carburiser cover
method is not suited to this purpose as it is highly vigorous and there are relatively
large amounts of metal when compared to the amount of char. To expose reactions
at the interface to close scrutiny, the sessile drop method was chosen for these
investigations with a substrate prepared from compacted char particles. This
technique has the advantage of small sample sizes and also having relatively large
amounts of char as compared to liquid iron present in the droplet. In addition, a time
64
CHAPTER 3: Objectives and Study Parameters

dependant view of the reaction can be obtained by quenching samples out of the
furnace at specific times.

To summarise, the broad outlines of the project are as follows:

1. To prepare and characterise chars in terms of particle size, ash content, and
structure. To also characterise coke in a similar manner.
2. To make wettability measurements for all chars/coke with liquid iron as a function
of time and composition of liquid metal
3. To perform an in-depth investigation on the interfacial phenomena between
char/liquid iron and coke/liquid iron using the sessile drop technique. To monitor
the formation of reaction products, carbon and sulphur pick-up by liquid iron as a
function of time.
4. To carry out systematic carbon dissolution studies on chars/coke using the
carburiser cover method
5. To develop a theoretical model for computing interfacial contact area for carburiser
cover investigations and to compute dissolution rate constants for these
carbonaceous materials
6. To acquire a comprehensive understanding of various mechanisms operating during
carbon dissolution from non-graphitic materials and their relative influence in
controlling the overall dissolution kinetics.

3.2 Study Parameters

Keeping the main objectives of the project in mind, the following study parameters
were chosen to complete the project within the allowed time frame. The parameters
chosen for this study are as follows:

1. Coal/Char Type

By choosing several chars to study, a broad range of material parameters such as ash
content and composition, carbon content and structure is covered. Utilising carbon
sources of a similar type (e.g. comparing only coal chars) has the advantage of reducing

65
CHAPTER 3: Objectives and Study Parameters

the differences between the materials, and allowing a clearer picture of the reaction
mechanisms to emerge. A wide range of black coals is available in Australia. A number
of coals, which were identified by industry sources to be of the type injected as PCI
coals, were then chosen for this investigation. Chars were then prepared from these raw
materials.

2. Metal Composition

The composition of the metal plays an important role in the dissolution kinetics, with
sulphur having a large effect on the dissolution kinetics. It also seems to have a strong
influence on carbon dissolution from non-graphitic carbons as compared to that from
graphitic carbons [76]. In addition, the initial carbon level of the liquid metal also plays
a role in limiting carbon dissolution, as it reduces the driving force for carbon
dissolution. A range of liquid iron compositions was used in this study.

3. Experimental techniques

Two experimental techniques were used in this study to identify the interfacial reactions
occurring between iron and char/coke, to determine their influence on the mass transfer
of carbon into the liquid metal, and to measure the first order carbon dissolution rate
constant for a series of chars and a coke.

The carburiser cover technique was chosen to create highly agitated conditions to
overcome mass transport resistance in the liquid phase. By removing this limitation it
was hoped that the reaction rates derived from these experiments would be reflecting
the dissolution rates that may be experienced in the blast furnace conditions.

The static sessile drop experiment was chosen to allow the full development of the
interfacial reaction products. In the carburiser cover technique the high degree of
stirring needed to remove mass transport resistance can also disturb the interfacial layer.
The sessile drop experiments also allowed the time dependent formation of these
interfacial products to be monitored as the reaction can be “frozen” at various stages and
analysed for amount and composition. In the sessile drop technique the carbon transfer
is limited (the drop is static and the only stirring is from Marangoni convection) and the
66
CHAPTER 3: Objectives and Study Parameters

interface is small compared to the char amount. This means that the reaction products
can accumulate at the interface without being disturbed or removed (as with the
carburiser cover method). The reaction products can be seen in-situ after quenching and
the order and mode of the formation examined.

Of the two techniques used, the carburiser cover method is probably more similar to the
blast furnace system than the sessile drop method. It has more metal than char present,
is highly vigorous, and interfacial products that are formed are less likely to play a
dominant role in carbon transfer because of the shorter residence times of the metal in
contact with the char.

67
CHAPTER 4: Experimental Details

CHAPTER 4: EXPERIMENTAL DETAILS

4.1 Objectives

In this study, the rate of carbon dissolution from a number of coal chars into liquid iron
as a function of carbon type (parent coal material) and coal ash composition was
determined. The influence of chemical reactions between solute/solid carbon and ash
oxides has also been investigated. These studies on coal-chars were supplemented with
investigations on metallurgical coke for the sake of comparison.

A number of high temperature experimental facilities were used in this study. A drop
tube furnace was used for preparing chars from raw coals. Small-scale investigations
including wettability and interfacial phenomena investigations were carried out using a
sessile drop approach in a horizontal tube furnace. Large-scale reaction experiments
utilising a carburiser cover method were carried out in an induction furnace. A detailed
description of both these techniques is provided later.

Experiments were chosen to investigate interfacial phenomena in two extreme


conditions. The smaller scale sessile drop studies allow investigation of the progress and
products of interfacial reactions in relatively static conditions. Here transport-limiting
conditions are allowed to dominate permitting the identification of the first products of
interfacial reaction.

The larger scale experiments are undertaken in conditions where both the carburiser and
liquid iron are present in excess and highly stirred. These conditions were chosen to
overcome any resistance to carbon supply for dissolution in liquid iron.

4.2 Production and Characterisation of Experimental Materials

In this study two types of main reaction materials were used: iron to provide liquid iron
for reactions and carbonaceous materials as a source of carbon. Iron used was either in a
pure form or as an alloy containing small quantities of carbon/sulphur. Carbonaceous
materials used were metallurgical coke and four coal-chars. These materials were

68
CHAPTER 4: Experimental Details

characterised using chemical analysis, carbon and sulphur analysis and field emission
scanning electron microscopy. The results are discussed in the following sections along
with some details of characterisation techniques.

4.2.1 Iron-based Materials

For use in the carburiser cover method, which requires a large quantity of liquid iron
(1.2 kg for every experimental run), an iron/carbon alloy was especially prepared by
Quality Castings Pty Ltd to provide enough material for the experiments. The chemical
composition of the iron/carbon alloy is given below in Table 4.1.

Table 4.1: Iron used in carburiser cover experiments

Material/Source Composition (wt%)


2.48 %C, 0.009%S, 0.2%Mn, 0.28% Si,
0.02%Ni, 0.06%Cr, <0.01%Mo,
Cast material/from Quality Castings
0.006%P, 0.01%Cu, <0.01%V,
<0.01%Al, <0.01%Mg, <0.01%Sn

In the small-scale sessile drop experiments in a horizontal tube furnace (wettability and
interfacial reaction studies), which require ~ 0.1 to 1.0 gm of liquid iron per
investigation, a different set of liquid iron compositions were used. Details of these
materials, along with their sources are given below in Table 4.2.

Table 4.2: Iron used in horizontal tube experiments

Material/Source Composition (wt%)


Electrolytic Iron/from Aldrich Chemical Labelled 99.98% pure. Leco analysis
company gives: 0.004%C and 0.002%S
Wettability Sample Iron, 2% C, 0.01%S

The purity of the iron plays an important role in carbon dissolution investigations due to
the influence of impurities on the saturation solubility [88]:

69
CHAPTER 4: Experimental Details

C s = 1.3 + 0.00257T − 0.31Si − 0.33P − 0.45S + 0.28Mn

Equation 2.24 (reproduced from chapter 2)

Care was taken to ensure that the experimental temperature was above the liquidus point
of the liquid iron, which guaranteed that the iron was fully molten. Liquidus
temperature is a function of composition and is strongly influenced by carbon and
sulphur contents.

The techniques used for material characterisation are briefly described below.

4.2.1.1Carbon and Sulphur Analysis using LECO SC244

The LECO™1 CS244 carbon and sulphur analyser combusts the sample of metal using
an induction furnace in excess oxygen. All of the carbon and sulphur in the sample then
exits as an oxidised gas, to be detected by IR cells calibrated for carbon or sulphur. It
allows quick and accurate determination of metal carbon and sulphur levels. For an
accurate analysis the samples should be in the range 0.7-1.0g. Carbon and sulphur can
be measured with an accuracy of ± 0.02wt% carbon and ± 0.001wt% sulphur. The
accuracy of the analysis is further ensured by regular calibration and testing with
calibration samples during analysis - usually once every 5-10 samples.

This analysis machine is used to determine carbon and sulphur levels during dissolution
experiments and is also utilised for the interfacial investigations.

4.2.1.2Field Emission Scanning Electron Microscopy

The field emission scanning electron microscope (FESEM) is essentially a traditional


scanning electron microscope, fitted with a field emission electron source, which
provides a very fine, but highly intense beam. This allows a higher resolution image
than with conventional SEM’s.

1
LECO is a trademark of the LECO Corporation, St Joseph, MI.

70
CHAPTER 4: Experimental Details

The model used in this work was a Hitachi S4500 fitted with an Oxford Isis energy
dispersive x-ray analyser (EDS). This enables a semi-quantitative microchemical
analysis of materials in the sample. The elements detectable are boron and heavier on
the periodic table. This energy dispersive spectroscopy can produce spot microchemical
analysis of a particular area on a sample, or can yield image maps, which are based on a
pre-defined set of elements. The depth of analysis is limited to the interaction volume of
the electron beam (typically ~2-3µm).

4.2.2 Carbon-based Materials

Four coals and one coke were the primary materials used in this study. The coals were
Australian black coals, and may be suitable as PCI coals. The coke is a metallurgical
coke produced for use in a blast furnace. A full analysis of these experimental materials
is provided in section 4.2.2.3.

4.2.2.1Char Production

The coals as received were crushed in a jaw crusher and vibrating grinder and wet
sieved at 38µm and 125µm. The resulting size graded samples were dried at ~80° C in
air in an oven prior to char making.

A schematic of the drop tube furnace (DTF) is shown in Figure 4.1. The carrier gas used
was 23% oxygen in nitrogen, fed at a rate of 1.2L/min. The coal particles were
introduced into the DTF via a screw feeder, with an injection rate of 0.0167 g/s.
Previous experimental work calculated the residence time of the coal particles to
approximately 1 second and a heating rate of around 104°C/sec.

71
CHAPTER 4: Experimental Details

Primary
Gas and
Coal
Water-cooled
Particle in
Injector
Secondary
Working Tube
Gas in

Thermocouple 1
to Controller

Thermocouple 2
to Controller

Reaction
Zone

Heating
Elements

Unburnt
Char and
off gas out Water-cooled
Collector

Figure 4.1: Schematic diagram of drop tube reactor used

The experimental parameters for the DTF were chosen to ensure complete
devolatilisation and partial combustion, to produce char material that may typically be
exposed to liquid iron in a BF. This was confirmed using a thermo-gravimetric furnace
on two samples of char (the coals containing the highest and the lowest volatile
content).

72
CHAPTER 4: Experimental Details

4.2.2.2Coke

The coke material was received from Port Kembla Steelworks. It was received in a
lump form, post coking, but prior to charging into the blast furnace. As such it has had
no thermal history apart from coking.

The lump coke was also ground using jaw crushers and vibrating grinder to a size
fraction +38µm to –125µm. This was performed dry.

4.2.2.3Characterization of Carbonaceous Materials

In this section the techniques used for characterization and the results of the initial
characterization for the carbonaceous materials before experimentation are presented.
Analysis results performed by external commercial companies are also included.

Proximate/Ultimate and Ash Analysis

The chars and coke were sent externally2 for chemical analysis. Proximate, ultimate and
ash analysis was carried out. The composition of the chars and coke can be seen in
Table 4.3 and Table 4.4.

Table 4.3: Char/coke compositions, dry basis (in wt%)

Composition
Char 1 Char 2 Char 3 Char 4 Coke
wt%
C 84.78 80.49 77.86 79.39 85.54
H 1.72 2.91 2.54 2.54 0.4
S 0.35 0.56 0.4 0.36 0.36
Ash 9.04 9.75 12.61 10.88 11.52
Mineral Matter 9.78 11.29 13.99 11.98 12.04

2
BHP Minerals research Newcastle Laboratories, Shortland, NSW.

73
CHAPTER 4: Experimental Details

Table 4.4: Ash composition of chars/coke

Ash
Material Char 1 Char 2 Char 3 Char 4 Coke
(wt%)
Al2O3 37.3 30.6 29.1 25 31
SiO2 41.2 52.6 52.7 53.8 55.8
Fe2O3 6.31 5.91 7.74 7.31 3.46
CaO 4.18 2.29 3.31 2.47 1.95
TiO2 2.16 1.9 1.46 1.34 1.51
K2O 1.11 1.08 1.32 2.98 0.76
MgO 1.08 0.7 1.21 0.88 0.4
P2O5 1.28 1.03 2.08 0.8 1.25
Na2O 1.24 0.49 0.27 0.64 0.38
SO3 2.8 0.72 0.42 0.9 0.18
BaO 0.13 0.06 0.08 0.07 0.08
Mn3O4 0.06 0.05 0.05 0.08 0.02

X-ray Diffraction Analysis

Samples of the chars and coke were analysed using x-ray diffraction (XRD). The
samples before analysis were finely ground using a ring mill. The details of the XRD
analysis technique have been discussed in section 2.3.

Samples for this study were analysed through 2θ angles of 5o to 90o. The step size was
typically 0.02o and the scan rate was 0.5o per minute. The analysis of the spectra was
completed in a Microsoft Excel spreadsheet. Corrections were applied for scattering,
holder intensity, and polarisation. The Lc calculations were determined by fitting two
Pearson curves to the XRD spectra, to either side of the non-symmetrical carbon peaks
and explained in more detail by L. Lu [10].

The spectra for the chars and cokes are shown in Figure 4.2. The primary feature of the
spectra is the prominent carbon peak (seen around 25°). This peak was used to
determine the Lc parameter.
74
CHAPTER 4: Experimental Details

Figure 4.2: X-ray diffraction spectra for the chars and the coke.

Prior to analysis of the spectra the sample holder signal was subtracted from the XRD
trace. The calculated Lc values are listed in Table 4.5.

Table 4.5: Calculated Lc parameters for the carbonaceous materials

Material Lc (Å)
Coke 17.3
Char 1 12.7
Char 2 12.7
Char 3 12.4
Char 4 9.5

Visual/FESEM Characterization

A scanning electron microscope (SEM) was used to visually examine the chars prior to
using these materials in experiments.

Approximately 0.1 gram of well-mixed char was placed onto a double-sided carbon tape
and positioned onto the SEM sample holder. These samples were imaged and the
75
CHAPTER 4: Experimental Details

char/coke particle characteristics examined. SEM images for all four chars are presented
in Figure 4.3 to Figure 4.6.

Figure 4.3: SEM image of Char 1 at 100x magnification.

Figure 4.4: SEM image Char 2 at 100x magnification.

76
CHAPTER 4: Experimental Details

Figure 4.5: SEM image of Char 3 at 100x magnification.

Figure 4.6: SEM image of Char 4 at 100x magnification.

It is apparent from these images that the particle size range varies between the different
chars. This is due partly to the levels of volatile matter present in the parent coals (Char
1 parent coal has the lowest volatile matter content with 12.5% and Char 4 parent coal
has the highest volatile matter content with 32.0%). All the chars studied showed
moderate to large degrees of swelling during the char making process. An average
particle size was determined for these chars, and is listed in Table 4.6.

77
CHAPTER 4: Experimental Details

Table 4.6: Average particle size of char particles

Char Average Particle Diameter (µm)


Char 1 25
Char 2 30
Char 3 100
Char 4 150

The coke material has not been reacted so the particle size range was the sieved size,
38-125µm.

4.3 Experimental Techniques

Two experimental techniques were used to carry out an investigation on the dissolution
of carbon from chars into liquid iron. The phenomena occurring in the iron/carbon
interfacial region, wettability and reaction kinetics was investigated using the sessile
drop method. Further chemical and electron microscopic analyses were also carried out
on the interfacial region after the iron/carbon assembly was removed from the
horizontal tube furnace. The carburiser cover technique was employed to investigate the
macroscopic aspects of carbon dissolution from chars. These experimental techniques
are described below.

4.3.1 Sessile Drop Technique

In the past, the sessile drop technique has primarily been used to study wettability. It is
one of the main techniques used to measure contact angle [99], and it provides a simple
configuration and small sample sizes. In non-reactive systems the equilibrium contact
angle can be measured, but when the system is reactive the contact angle changes as a
function of time [100]. The contact angle measured is then known as a dynamic contact
angle.

The sessile drop technique has also been used to study the interfacial phenomena
occurring between natural graphite and liquid iron [14]. In this study the interface was
observed in several samples produced after various times of reaction at high
78
CHAPTER 4: Experimental Details

temperature. The reaction products were analysed in a field emission scanning electron
microscope using energy dispersive spectroscopy. In this study the sessile drop
technique was used to determine the dynamic contact angle as well as the formation of
an interface between the iron and char or coke as a function of time.

The sessile drop technique was used extensively in this study as it allowed the focus to
be placed on the reactions occurring in the interfacial region. The experimental
arrangement is shown schematically in Figure 4.7.

INE RT ATMOSH P E RE

LIQUID θ
IRON

O
SU BSTRATE

Figure 4.7:Sessile drop schematic

Two sets of investigations were carried out using the sessile drop arrangement. In the
first study, the dynamic wettability/contact angle between the liquid iron and the
char/coke substrates were measured for all chars and coke. Variation of contact angles
with time provides very valuable information on interfacial reactions. The changes in
contact angle can indicate the initiation or conclusion of different chemical reactions
between the substrate and the liquid droplet. Primarily however, the knowledge of the
contact angle between two materials (char/coke and iron for example) can be used to
determine the interfacial area of contact between the two materials.

The second study using the sessile drop technique was focussed on the interfacial
reactions between the carbonaceous substrate and liquid iron. Reaction products at the
interface were monitored as a function of time.

As already noted in the Chapter 2, (section 2.5.4), earlier studies in this field have
neglected to take into account chemical reactions that occur concurrently with the
dissolution of carbon and sulphur and could significantly affect final results. The role of
ash as a physical barrier has been postulated, but the concept of ash as a reactive
component in the char/coke and iron system has not been considered. The sessile drop
79
CHAPTER 4: Experimental Details

technique focuses on the interface, by limiting the amount of iron present and having a
greater amount of carbonaceous material. In addition, melt turbulence is very low as the
droplet is not stirred (apart from Marangoni convection) and this may slow down the
reactions that occur.

4.3.1.1Dynamic Wettability for interfacial area determination

The dynamic wettability study was conducted by placing a sessile drop arrangement of
a char/coke substrate and iron alloy sample (either electrolytic iron or Fe-2wt%C-
0.01wt%S) in a horizontal tube furnace (see Figure 4.8). The substrate was formed by
grinding the powdered coke or char, and pressing to a pressure of 7.75MPa, in a die.
The substrate formed had top face dimensions of 2.4 x 2.4 cm, and contained ~2.5g of
carbonaceous material. The drops weighed ~0.1g, ensuring no weight effects in the
contact angle measurement. The list of experiments conducted has been given in Table
4.7.

Sample Metal Carbon


rod Alumina sample substrate
tube

CCD
Camera

VCR
Cold zone Hot zone

Quartz
Sample window
Tray
Thermocouple Gas Gas
outlet inlet

Figure 4.8: Horizontal furnace schematic

Table 4.7: Wettability experiments

Carbon Type Metal Carbon Metal Sulphur Temperature


Char 1-4, coke 0 wt% 0 wt% 1550°C
Char 1-4, coke 2 wt% 0.01 wt% 1550°C
Char 1 2 wt% 0.01 wt% 1500°C

80
CHAPTER 4: Experimental Details

The sample assembly consisted of a graphite rod with a tray attached to the end. The
metal sample was centred on the top surface of the pressed substrate and placed on the
graphite tray. This assembly was inserted in the cold zone of the furnace and sealed. An
inert atmosphere of argon was fed into the furnace at a rate of 0.5 l/minute and the tube
was purged. The furnace was then heated, and when the furnace reached the desired
reaction temperature (controlled by a B type thermocouple placed below the centre of
the hot zone) the tray assembly was slid into the hot zone.

As illustrated in Figure 4.8, the reaction was monitored by a CCD camera, which was
focussed on the sample in the hot zone of the furnace. The camera was connected to a
television and video recorder, enabling the experiment to be recorded on videotape.
When the sample is first pushed into the hot zone, the temperature in the hot zone drops
(as the sample is initially colder than the surroundings, this drop was around 30°C). The
sample was observed during the heating up period, and when the sample melted the
counter measuring the reaction time was started. The time when the sample melts after
insertion into the hot zone is a function of the melting point of the metal sample. Metal
samples with lower melting points will melt earlier, before the furnace has regained the
reaction temperature or before the sample has fully heated to the reaction temperature.
However the time between the sample melting and the furnace regaining reaction
temperature was measured and was found to be approximately 30 seconds. This time
delay was much shorter for the electrolytic iron as it melts around 1540°C.

A recording of the droplet in profile was started from just before the sample melting
until the sample was removed from the hot zone. Sliding the assembly into the cold
zone of the furnace, which quenches the liquid droplet/coke substrate, effectively
terminates any reaction.

Still images were captured from the videotape recording of the experiment using a
computer. The images were analysed using ANGLE3, software that uses a non-linear
regression calculation procedure to determine the contact angle.

3
Angle- Copyright Electroinf, version 1.4.

81
CHAPTER 4: Experimental Details

Figure 4.9: Sample program output for captured image angle analysis

The output of the Angle software calculates the left and right contact angles. The angles
for the left and right sides of the droplet are then averaged to gain a contact angle for
that time. Images taken from the videotape at various times allow the change in contact
angle as a function of time to be determined. These angles are known as dynamic
contact angles, as they are reported as a function of time and are not the equilibrium
contact angles [101]. The measurement of dynamic wettability is the only way to
determine the interfacial area between the liquid iron and the carbonaceous materials
studied. If the dissolution results are to be used in any situation other than comparison
with other carburiser cover results conducted in an identical manner then the rates
determined must be corrected for interfacial area.

4.3.1.2Study of Interfacial Reaction Products

This study was carried out with an aim to identify the reactions that are occurring at the
interface of the char or coke and iron. The sessile drop technique was employed as the
technique allows for the study of the developing interfacial products with time. These
experiments were quite similar to the wettability experiments (section 4.3.1.1), with a
few differences.

In these experiments, to enlarge the interface between the iron and the char/coke
substrate, the drop size was increased to 0.8-1.0g. This increase in sample size also
increased the accuracy of subsequent compositional analysis of the metal after reaction.
The metal sample used was electrolytic iron; this ensured that all elements (other than
iron) come from the substrate. The atmosphere was Ar, with a flow rate of 0.5 l/minute.

82
CHAPTER 4: Experimental Details

The droplets were monitored by CCD camera, and the reaction time was measured
using a counter started when the metal sample melts. However, contrary to the dynamic
wettability investigations, the aim of this experiment was to produce many samples of
droplets that had been reacted with the char/coke substrates for varying times. Once the
sample was pushed in, and the counter started after the metal melted, the time was
closely monitored and the droplet and substrate quenched (by withdrawing the tray into
the cold zone of the furnace) after a fixed period of time. The solidification of the
droplet was visually monitored and generally took 5-10 seconds. Samples were
produced for the following times: ½, 1, 3, 5, 7, 10, 15, 30, 60 and 180 minutes. Once
these droplets were quenched, the droplet was removed from the substrate.

Two sets of samples were produced for each of the times indicated and for each of the
substrates studied. One set was used to examine the new phases formed at the interface.
The underside of the droplet (which is effectively the iron/substrate interface) was
inspected using a Field Emission Scanning Electron Microscope (FESEM, see section
4.2.1.2), and analysed with Energy Dispersive Spectroscopy (EDS). This type of
analysis allowed the time dependant growth of the interfacial material to be determined.

The series of experiments conducted using this technique is listed in Table 4.8.

Table 4.8: Interfacial Reaction Product Studies.

Carbon Sulphur
Carbon type
concentration of concentration of Temperature
studied
metal metal
Coke 0 wt% 0 wt% 1550°C
Char 1 0 wt% 0 wt% 1550°C
Char 4 0 wt% 0 wt% 1550°C

Mass Spectrographic analysis of the gases produced during coke/iron interaction was
also performed. This was conducted to monitor the reduction of oxides in the ash by the
carbon in the coke and by the solute carbon in the liquid iron. The experimental
arrangement was similar to that used above for the study of the interfacial reactions,
however the graphite rod was replaced with an alumina rod, ensuring only the carbon
83
CHAPTER 4: Experimental Details

from the coke sample was measured in the off gas analysis (mainly CO and CO2 and
Ar). An experimental run of a pressed coke substrate without the liquid iron droplet was
also conducted as comparison as a blank test. This also generated CO and CO2 due to
in-situ reduction of the silica in the coke ash.

4.3.1.3Carbon/Sulphur/Silicon Transfer Studies

This series of experiments using the sessile drop technique examined the transfer of
carbon and sulphur across the interface from the char or coke into the liquid iron. The
experimental detail for the production of the samples was the same as described in the
above section (section 4.3.1.2).

The sets of droplets produced (for the same time periods listed in section 4.3.1.2) were
used to determine the carbon and sulphur contents at the different times investigated
(analysed using a CS244 carbon sulphur analyser, see section 4.2.1.1). In the coke/iron
case, another series of droplets were produced to analyse the metal for silicon content.
These results produced graphs for carbon, sulphur and silicon (when measured) contents
as a function of time. The silicon content was determined in the droplets by external
analysis conducted by BHP Minerals research Newcastle Laboratories, Shortland,
NSW. The details of the experiments conducted are listed in Table 4.9.

Table 4.9: Sessile Drop Mass Transfer Experiments.

Metal
Carbon Temperature Analysis
Composition
Coke Electrolytic iron 1550°C C, S, and Si
Char 1 Electrolytic iron 1550°C C and S
Char 4 Electrolytic iron 1550°C C and S

4.3.2 Carburiser Cover Technique

In previous studies, a number of different techniques have been used to measure the
carbon dissolution rate from solid carbon into liquid iron (these are listed in section
2.5). The carburiser cover method, while not the most common method, is the only
84
CHAPTER 4: Experimental Details

method for which powdered carbon can be used effectively. Using a powdered form of
char was considered important in this work to reflect the nature and behaviour of PCI
chars in the ironmaking process. The carburiser cover method involves placing
pulverised or small lump material onto the top of a molten metal bath, and monitoring
the change in composition with time. The schematic for this is shown in Figure 4.10.

Figure 4.10: Schematic of Carburiser cover method.

The furnace utilised in this study was an induction furnace, shown in Figure 4.11.
Thermocouple
Quartz Tube
Crucible

Molten Iron
H2O
Gas Inlet
(N2)

Carburiser
Cover

Figure 4.11: Schematic diagram of Induction Furnace utilising the carburiser cover
method

85
CHAPTER 4: Experimental Details

This furnace is capable of attaining temperatures in excess of 1600oC, and the induction
currents induced in the melt provide stirring within the melt around 0.33m/s [102]. The
furnace atmosphere is purged with nitrogen (high purity) and vented using a suction
system. The temperature of the liquid bath is maintained using an immersed
thermocouple, which is attached to the controller for the furnace. This allows for a much
more consistent bath temperature, as the furnace can compensate for thermal losses due
to opening the furnace lid or the high heat loss associated with the addition of room
temperature carbonaceous material and the subsequent dissolution process.

Typically 150grams of carburiser is added to the top of the melt, and the melt itself
weighs around 1.2kg. Bath composition prior to adding the carburiser is altered using
high purity graphite. Initial bath compositions were normally 2wt% carbon and
0.01wt% sulphur. Any slag that may have formed during melting and composition
alteration is removed prior to the experiment. Small deviations were observed in the
initial carbon level in some cases, and were attributed to carbon burn off resulting from
adding the carbonaceous material to the furnace.

Once the initial composition was reached the carbon source is added to the top of the
melt, and the time is started. Iron samples were then removed from the bath
periodically, with the use of a silica glass pipette. These samples removed are of the
order of 1-2 grams, and were considered not to influence bath size significantly.

For non-graphitic carbon sources (like coke or char), the dissolution time is typically
20-40 minutes. Samples of the melt were taken every 30 seconds for the first 5 minutes,
and subsequently less frequently (carbon and sulphur mass transfer occurs less rapidly
after this time, such that the rate of change is not as great). The removed samples were
later analysed for carbon and sulphur using the LECO CS244 (described in section
4.2.1.1).

86
CHAPTER 5: Results and Discussion – Sessile Drop Technique

CHAPTER 5: INTERFACIAL PHENOMENA USING SESSILE DROP


TECHNIQUE: RESULTS AND DISCUSSION

The sessile drop technique has been used extensively to study wettability. However this
technique can also be used for an in-depth study of the interface and interfacial products
that form. This technique was used to determine the dynamic wettability for chars and
coke with liquid iron, as this fundamental information about chars is currently not
available. This wettability data will be used in a model to determine the contact area
between the liquid iron and the carbonaceous materials during dissolution. The
reactions occurring at the interface between the chars/coke and liquid iron were studied
using the sessile drop technique. This technique was chosen for these investigations
because it has the advantage of small sample sizes and also having relatively large
amounts of char as compared to liquid iron present in the droplet. The influence of
interfacial phenomena is enhanced due to the relatively static nature of the drop, which
only has stirring due to Marangoni convection, and the interface cannot be easily
refreshed due to stirring that is the case in the carburiser cover technique. In addition, a
time dependant view of the reaction can be obtained by quenching samples out of the
furnace at specific times. Finally, the transfer of carbon and sulphur was determined for
the sessile drop arrangement and correlations made between the reactions occurring and
the observed composition change due to these reactions.

5.1 Wettability Measurements

The sessile drop technique has been used to determine the dynamic wettability of chars
and coke with liquid iron at 1550°C. Currently very little information is available on the
wetting behaviour of these materials, mainly due to difficulties associated with
measuring contact angles for fine powders. In these studies, a solid substrate was
prepared by grinding the powdered coke or char, and pressing to a pressure of 7.75MPa,
in a die. No binder was used in the preparation of the substrate. Apart from generating
fundamental wettability data on these systems, these results will also be used to
calculate the interfacial area of contact between the liquid iron and the carbonaceous
material using a theoretical model. Wettability was measured as a function of time for a

87
CHAPTER 5: Results and Discussion – Sessile Drop Technique

range of liquid iron compositions. The experimental results are presented in the
following section.

5.1.1 Wettability with Electrolytic Iron

The wettability for the chars and coke was determined using the approach detailed in
section 4.3.1.1. The results showed a small variation in contact angle between the chars
and coke, with the contact angle changing slightly with time. The wetting images of the
various chars with electrolytic iron are shown in Figure 5.1 to Figure 5.5.

Figure 5.1: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Char 1 at 1550°C.

Figure 5.2: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Char 2 at 1550°C.

88
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.3: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Char 3 at 1550°C.

Figure 5.4: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Char 4 at 1550°C.

Figure 5.5: Initial (left) and final (right) images of the contact angle assembly for
Electrolytic iron on Coke at 1550°C.

89
CHAPTER 5: Results and Discussion – Sessile Drop Technique

A sample of the analysis of the angles is shown in Figure 5.6 and Figure 5.7.

Figure 5.6: Sample analysis of the coke initial contact angle with electrolytic iron using
the Angle software.

Figure 5.7: Sample analysis of the coke final contact angle with electrolytic iron using
the Angle software.

The results of the wettability studies on the different carbonaceous materials with
electrolytic iron are given in Figure 5.8 to Figure 5.12.

90
CHAPTER 5: Results and Discussion – Sessile Drop Technique

150

140

130
Contact Angle (°)

120

110

100

90

80
0 500 1000 1500 2000 2500 3000 3500 4000
Time (s)

Figure 5.8: Dynamic contact angle results for Char 1 with electrolytic iron at 1550°C.

150

140

130
Contact Angle (°)

120

110

100

90

80
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Time (s)

Figure 5.9: Dynamic contact angle results for Char 2 with electrolytic iron at 1550°C.

91
CHAPTER 5: Results and Discussion – Sessile Drop Technique

150

140

130
Contact Angle (°)

120

110

100

90

80
0 1000 2000 3000 4000 5000 6000
Time (s)

Figure 5.10: Dynamic contact angle results for Char 3 with electrolytic iron at 1550°C.

150

140

130
Contact Angle (°)

120

110

100

90

80
0 1000 2000 3000 4000 5000 6000
Time (s)

Figure 5.11: Dynamic contact angle results for Char 4 with electrolytic iron at 1550°C.

92
CHAPTER 5: Results and Discussion – Sessile Drop Technique

150

140

130
Contact Angle (°)

120

110

100

90

80
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Time (s)

Figure 5.12: Dynamic contact angle results for Coke with electrolytic iron at 1550°C.

From Figure 5.8 to Figure 5.12 it can be seen that contact angles for the different
materials show a scatter of ~10-15° after approximately 1000 seconds (~16 minutes
contact). Given the error (±4°) associated with the measurement of these angles [103],
this is a significant variation. The largest variation seen in these results occurs at the
initial stages of contact. The same results can be plotted on a logarithmic time scale
(Figure 5.13) to more clearly see the trends apparent in the initial stages of contact.

93
CHAPTER 5: Results and Discussion – Sessile Drop Technique

150

140

130
Contact Angle (°)

120

110

100 Char 1
Char 2
Char 3
90 Char 4
Coke
80
1 10 100
Time (s)

Figure 5.13: Dynamic contact angle results for the carbonaceous materials with
electrolytic iron at 1550°C, logarithmic time scale.

In the above figure, it is clear that the nature of carbonaceous material plays an
important role in wetting behaviour. The variation between the different substrates is
approximately 30-40° in the first 10-100 seconds. Although this period corresponds (at
least for the first 30 seconds, see section 4.3.1.1) to the heating up period for the sample,
it can be assumed that the thermal state of the different iron/substrate series is similar.
This then leaves the substrate material as influencing the initial contact angle. For this
study the initial contact angles (for the purposes of modelling contact area) were taken
as the average contact angle in the period of wetting up to 100 seconds. From Figure
5.13, it can be seen that the angles were reasonably stable in this time frame. These
results are summarised in Table 5.1.

Table 5.1: Initial contact angle for electrolytic iron on different carbonaceous
substrates at 1550°C.

Char/coke Initial Contact angle (°)


Char 1 137°
Char 2 133°
Char 3 117°
Char 4 106°
Coke 107°
94
CHAPTER 5: Results and Discussion – Sessile Drop Technique

The final contact angles for the different substrates were more similar than the initial
contact angles. It can be seen that the final angles are all in the range 100-120°. These
angles, taken as an average of the final 5 measurements, are listed in Table 5.2.

Table 5.2: Final contact angle for electrolytic iron on different carbonaceous substrates
at 1550°C.

Char/coke Final Contact Angle (°)


Char 1 111°
Char 2 107°
Char 3 119°
Char 4 101°
Coke 111°

In complex systems such as chars and cokes, several mechanisms influence wettability
with iron [103]. Factors that may play a role in the wetting phenomena include carbon
content, carbon structure (atomic structural factors like crystallinity and amorphous
carbon levels), ash content, ash composition, formation of other phases at the interface
due to reactions, and sulphur content [103].

Previous work on the wettability of natural graphite with liquid electrolytic iron [14]
found that the contact angle varied quite significantly with time. Starting from an initial
contact angle of 104°, it reduced to ~60° after 10 seconds of contact, and then increased
in the next 200 seconds to stabilise at 102° and was 102° for the remainder of the
experiment. The contact angles of iron with alumina and calcium sulphide have been
reported in literature and are 122° and 87° respectively. The dynamic changes on the
natural graphite/iron interface were interpreted in terms of deposition of products such
as alumina/calcium sulphide at the interface. This study brought out the importance of
the nature and composition of the interfacial layer, particularly for contact angles
reported after significant time of contact.

In an attempt to explain the observed wettability results for chars/coke, a number of


aspects were considered. Examining the role of the carbon content and carbon structure
95
CHAPTER 5: Results and Discussion – Sessile Drop Technique

on wettability for chars and coke in the initial stages (0-100 seconds), no specific
dependence on carbon content could be observed. Char 1 and the coke both have similar
carbon contents (section 4.2.2.3) and yet they appear at the extremes of the initial
contact angle range. It is quite difficult to compare char results with coke due to a
number of inherent differences. Even while comparing results only from chars, no well-
defined trends could be seen. The carbon contents for the materials are not significantly
different, however, there are differences in the carbon structure. While Char 1 has the
highest Lc value (12.7Å), Char 4 has the least Lc value (9.5Å) and Char 3 has a Lc value
similar to Char 1. This suggests that the carbon structure of the chars probably does not
play a dominant role in iron/char wettability in the initial stages. This is also supported
by the study on natural graphite wettability [14], where the initial contact angles are
similar to the values seen for char 4 and coke.

Considering the ash content and composition, the materials that have the lowest ash
content (Char 1 and Char 2) had the highest contact angles in the initial stages. This
trend is not continued for the other materials however, with Char 3 and Char 4 not
showing this trend. Ash composition did not lead to any observable trends in the initial
wettability in this study. One ash material that may have some influence on wettability
is alumina [104]. Previous work with graphite/alumina substrates suggests that
increasing the alumina content in the substrate will increase the contact angle. This
general trend is not observed in current results with the bulk alumina concentration in
the char. However, this amount may have little to do with the actual amount of alumina
present at the interface, as it will be a function of the rate of carbon removal from the
interface by carbon dissolution.

The formation of other phases at the interface has been observed for natural graphite
[14], with the accumulation of alumina and calcium sulphide detected after iron/natural
graphite interaction. This interfacial layer was seen to grow with time, and reached
around 45% coverage after approximately 1 hour.

96
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.14: Increase of interfacial coverage with time due to formation of interfacial
complex with time, from [14].

This layer [14] also changed in composition over the time of contact, with aluminium,
oxygen and calcium being detected up to 7 minutes contact. Subsequent analysis of the
samples from 10 minutes to 60 minutes showed only iron, calcium and sulphur. The
authors of this study suggest that up to 7 minutes the composition of the interface
corresponds to the composition of the ash in the graphite, and the interfaces measured
after longer periods of contact show evidence of reactions occurring, one of which
produces calcium sulphide. It is interesting to note that the alumina is not detected at
later times of contact. As it seems unlikely that the alumina can be removed from the
interfacial area in such a static system as the sessile drop it may be that the alumina in
later times of contact was present at the interface. It is possible that the separation of the
droplet and graphite substrate for the analysis leads to the alumina phase being retained
on the graphite side (it is the droplet side of the interface that was analysed). Wu et al.
[14] also suggest that the lack of silicon at the interface may be due to the reduction of
silica from the ash material.

From this previous study, it is evident that the role of the interfacial layer may play an
important role in determining the contact angle at longer interaction times. The nature of
the interfacial layer formed during char/coke interactions with iron had not previously
been determined, and the experiments detailed in later sections (section 5.2 and 5.3)
97
CHAPTER 5: Results and Discussion – Sessile Drop Technique

were aimed at determining the nature of the interfacial reaction products and identifying
the reactions that produced them.

After contact with iron for a long period of time, it is to be expected that the
carbonaceous materials would develop some form of interfacial material between the
iron and substrate. The nature of this interfacial layer will strongly influence the final
contact angle, as it is this material that the iron is in quasi-equilibrium with (and not the
char/coke material). Whether this layer is continuous or fragmented, and the nature of
the layer and its composition will determine the contact angle. It is also important to
note that the final contact angles are all in the range 100-120°, suggesting that alumina
could be playing an important role in controlling the dynamic wetting of these
materials. Other factors such as carbon transfer and formation of calcium sulphide are
also expected to play an important role.

5.1.2 Wettability with Fe-2wt%C-0.01wt%S Alloy

The wetting images of the various chars with an iron 2wt%C 0.01wt% sulphur alloy are
shown in Figure 5.15 to Figure 5.19.

Figure 5.15: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Char 1 at 1550°C.

98
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.16: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Char 2 at 1550°C.

Figure 5.17: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Char 3 at 1550°C.

99
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.18: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Char 4 at 1550°C.

Figure 5.19: Initial (left) and final (right) images of the contact angle assembly for Fe-
2wt%C-0.01wt%S on Coke at 1550°C.

The results of the wettability studies on the different carbonaceous materials with the
Fe-2wt%C-0.01wt%S alloy are given in Figure 5.20 to Figure 5.24.

100
CHAPTER 5: Results and Discussion – Sessile Drop Technique

150

140

130
Contact Angle (°)

120

110

100

90

80
0 1000 2000 3000 4000 5000 6000 7000
Time (s)

Figure 5.20: Dynamic contact angle results for Char 1 with Fe-2wt%C-0.01wt%S alloy
at 1550°C.

150

140

130
Contact Angle (°)

120

110

100

90

80
0 1000 2000 3000 4000 5000 6000 7000 8000
Time (s)

Figure 5.21: Dynamic contact angle results for Char 2 with Fe-2wt%C-0.01wt%S alloy
at 1550°C.

101
CHAPTER 5: Results and Discussion – Sessile Drop Technique

150

140

130
Contact Angle (°)

120

110

100

90

80
0 1000 2000 3000 4000 5000 6000
Time (s)

Figure 5.22: Dynamic contact angle results for Char 3 with Fe-2wt%C-0.01wt%S alloy
at 1550°C.

150

140

130
Contact Angle (°)

120

110

100

90

80
0 1000 2000 3000 4000 5000 6000
Time (s)

Figure 5.23: Dynamic contact angle results for Char 4 with Fe-2wt%C-0.01wt%S alloy
at 1550°C.

102
CHAPTER 5: Results and Discussion – Sessile Drop Technique

150

140

130
Contact Angle (°)

120

110

100

90

80
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Time (s)

Figure 5.24: Dynamic contact angle results for Coke with Fe-2wt%C-0.01wt%S alloy at
1550°C.

The results of the contact angle determination for iron containing carbon and sulphur
are much less scattered than those for electrolytic iron. Due to the increased level
carbon and sulphur in the iron droplet there would be a lower driving force for mass
transfer. The initial contact angles are observed clearly in Figure 5.25, and the results
are summarised in Table 5.3.

Table 5.3: Initial contact angle for Fe-2wt%C-0.01wt%S alloy on different


carbonaceous substrates at 1550°C.

Char/coke Initial Contact angle (°)


Char 1 113°
Char 2 123°
Char 3 115°
Char 4 125°
Coke 127°

103
CHAPTER 5: Results and Discussion – Sessile Drop Technique

150
Char 1
Char 2
140
Char 3
Char 4
130 Coke
Contact Angle (°)

120

110

100

90

80
1 10 100
Time (s)

Figure 5.25: Dynamic contact angle results for the carbonaceous materials with Fe-
2wt%C-0.01wt%S alloy at 1550°C, logarithmic time scale.

The final contact angles are given in Table 5.4.

Table 5.4: Final contact angle for Fe-2wt%C-0.01wt%S alloy on different


carbonaceous substrates at 1550°C.

Char/coke Final Contact Angle (°)


Char 1 111°
Char 2 110°
Char 3 101°
Char 4 102°
Coke 125°

It is quite possible that a number of processes/reactions are occurring simultaneously in


the interfacial region and could affect the wetting behaviour. Various processes are:
• Transfer of carbon and sulphur across the interface
• Formation of an enriched interfacial layer containing calcium sulphide and
alumina

104
CHAPTER 5: Results and Discussion – Sessile Drop Technique

• Reduction of reducible oxides such as silica and iron oxides, and possible
transfer of these elements into the liquid iron

5.1.3 Influence of Metal Composition

The influence of the metal composition on char 1 can be seen in Figure 5.26. The two
experiments shown in Figure 5.26 demonstrate that the contact angle with electrolytic
iron is higher than the contact angle with the iron alloy for this char. This continues
until around 2500 seconds (~41 minutes) where the contact angles converge to similar
values.

150
Char 1 Wettability
Electrolytic iron at 1550°C
140 Fe-2wt%C-0.01wt%S at 1550°C

130
Contact Angle (°)

120

110

100

90

80
0 500 1000 1500 2000 2500 3000 3500 4000
Time (s)

Figure 5.26: Dynamic contact angle results for Char 1 with electrolytic iron and Fe-
2wt%C-0.01wt%S alloy at 1550°C.

5.1.4 Dynamic Wettability Summary

Using the Sessile drop method, the wettability of carbonaceous materials (four coal
chars/coke) with liquid iron at 1550°C was measured as a function of time.
Measurements were carried out with electrolytic iron and with an iron alloy containing
2 wt% C and 0.01 wt% S. These systems are inherently very complex, and there have
105
CHAPTER 5: Results and Discussion – Sessile Drop Technique

been no previous studies on the contact angles for non-graphitic carbon with liquid iron.
With electrolytic iron, all materials showed non-wetting behaviour with contact angles
ranging between 106°-133° in the initial stages and between 101° to 111° after an hour
of contact. For the Fe-C-S alloy, the contact angles ranged between113°-127° in the
initial stages and between 101° to 125° after an hour of contact. Being essentially non-
wetting, only a marginal improvement in contact angles was observed with time.
Several factors were identified that would influence the wettability:

• Transfer of carbon and sulphur across the interface


• Formation of an enriched interfacial layer containing calcium, sulphur and
alumina
• Reduction of reducible oxides such as silica and iron oxides, and possible
transfer of these elements into the liquid iron

It is expected that the sulphur content will have a particularly strong influence on the
wettability of these systems, as it has been previously reported that the contact angle for
graphite and iron will dramatically change with even small additions of sulphur in the
iron.

The formation of an interfacial layer changes the chemistry of the substrate locally (but
not necessarily changing the bulk composition). The formation of an interfacial layer
rich in calcium sulphide may act to reduce the contact angle.

The reduction of metal oxides in the ash can also remove elements from the liquid
metal, and may also involve some gas phase reactions. The interfacial products that
form between the iron and chars/coke are of great interest, not only for wettability but to
understand more fully what is occurring during carbon dissolution. A study of the
interfacial reaction products was conducted and the results presented and discussed in
section 5.2. The contact angles that were determined from these wettability studies
allow estimation of the area of contact between chars/coke and liquid iron during carbon
dissolution in the carburiser cover technique using the developed model (discussed in
section 6.4).

106
CHAPTER 5: Results and Discussion – Sessile Drop Technique

5.2 Study of Interfacial Reaction Products

In the studies of the interfacial reaction products formed between chars/coke and liquid
iron, several previously unreported phases or reactions were observed. The novel
technique used allows the formation of the interfacial layer to be observed. Several
general observations can be made for the two types of carbonaceous materials (chars
and coke) examined:
• Both types of material developed similar interfaces, with similar reaction
products being observed,
• The rates of formation of these reaction products varied according to
carbonaceous type,
• Interfacial coverage became almost complete within the time frames studied (up
to 3 hours contact),
• Whilst silica is a principal component of the ashes in the materials studied, very
little silicon was detected at or around the interface,
• Calcium and sulphur appeared to preferentially accumulate at the interface,
concentrating at levels in excess of that expected from ash composition.

5.2.1 Coke Interfacial Reaction Products

The FESEM images and the EDS spectra for samples (the underside of the droplets
interfacial region) quenched at various times are shown in Figure 5.27 through to Figure
5.29 and Figure 5.31.

107
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.27: Left: Interface of iron droplet after being exposed to coke at 1550oC for 1
minute. Right: EDS spectra of overall interface.

Figure 5.27 shows the interface after 1 minute of contact. The white phase is primarily
alumina. The interface is highly covered with the layer, even after one minute.

Figure 5.28: Left: Interface of iron droplet after being exposed to coke at 1550oC for 10
minutes. Right: EDS spectra of overall interface.

The interface at 10 minutes is shown in Figure 5.28. More ash components are present
at the interface, and there is significantly less iron showing through, indicating a
decrease in iron/carbon contact area.

108
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.29: Left: Interface of iron droplet after being exposed to coke at 1550oC for 30
minutes. Right: EDS spectra of iron drop interface.

Figure 5.29 shows the interface at 30 minutes contact. Again the majority of the ash
material appears to be alumina, and there is less iron seen.

Despite the high levels of silica in the coke ash initially (based on the analysis of the
unreacted ash), very low levels were detected during the EDS analysis of samples after
even 1 minute of contact. Previous studies suggest that the iron oxide within the ash is
likely to reduce to iron. During the EDS analysis, alumina was the oxide identified at
the interface; this suggests the other reducible oxides, such as silica and iron oxide, are
undergoing reduction reactions.

Based on the initial ash content of the coke, it can be seen from the EDS analysis of the
droplets quenched after various times, that alumina, sulphur and calcium are enriched
and silicon is depleted at the substrate interface. This can be seen as evidence for the
preferential reduction of silica. The silica is being depleted from the initial stages of the
interaction (from 1 minute contact), resulting in the alumina in the coke ash
accumulating, as these two materials are the main constituents of the ash. Notably, the
calcium and sulphur are starting to accumulate at the interface after 30 minutes.

109
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Thermochemical calculations using a computer package (FactSage™4 software) suggest


a liquidus temperature for the coke ash to be in excess of 1700°C, thereby severely
limiting the amount of liquid oxides present. The reactions involving slag are therefore
unlikely to be the main reactions in this system. However, even if some of the ash
materials are liquid or semi-liquid, it is possible that the reduction reactions could still
occur (as discussed in section 2.4). It can be seen from this study that measurement of
the liquidus temperature of the ash (at the initial composition) is not meaningful, as the
composition of the ash material is changing dynamically during the experiment.

Oxygen removal as calculated from mass spectroscopy data can be seen in Figure 5.30.
A sample calculation of this data is shown in Appendix A.

0.0045

0.004
0.0035
Oxygen removed (moles)

0.003

0.0025

0.002

0.0015

0.001
Moles of O2 removed-blank
0.0005 Moles of O2 removed-iron
0
0 10 20 30 40 50 60
Time (min)

Figure 5.30: Oxygen removal from coke and coke/iron system at 1550oC, calculated
from mass spectrometer data.

The levels of CO and CO2 in the off gas were measured from the mass spectrometer
data, which was then converted to moles of oxygen removed from the system. Figure
5.30 shows the difference between the total volume of oxygen removed from the

4
Thermochemical Software for Windows developed by Thermfact Ltd. in Montréal, Canada and GTT-

Technologies in Herzogenrath, Germany.

110
CHAPTER 5: Results and Discussion – Sessile Drop Technique

samples of coke, with and without a liquid iron droplet. The volume of oxygen removed
from the system due to the reduction of the iron oxide (calculated from a mass balance
based on the amount in the coke ash) equals the amount removed after approximately
10-20 seconds. Thus it is reasonable to assume that the majority of the removed oxygen
is due to the reduction of one or more reducible oxides, such as silica.

The primary routes for silica reduction are listed in Equation 2.6 and Equation 2.8
(reproduced from section 2.4.1 below) [18]:

SiO2 + C → SiO( g ) + CO( g )

Equation 2.6: Partial reduction to form silicon monoxide vapour.

SiO( g ) + C → Si + CO( g )

Equation 2.8: Silicon transfer to carbon-containing iron.

There are two sources of carbon for the reduction of silica.


a. The solid carbon in the coke substrate and/or gaseous carbon monoxide,
b. The solute carbon in liquid iron, resulting from the dissolution of the
carbon from the coke into the liquid iron
These carbon sources can reduce the silica to silicon monoxide, which can be further
reduced to silicon in the metal.

In Figure 5.30, the lower curve indicates oxygen removed during the reduction of silica
by the solid carbon source. The upper curve (which has a greater slope) indicates
oxygen removed during reduction of silica, not only with solid carbon, but possibly with
solute carbon as well. These curves show that the reduction of silica by solute carbon is
possibly occuring, as suggested by the differences in the oxygen removal rates.

111
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.31: Left: Interface of iron droplet after being exposed to coke at 1550oC for 60
minutes. Right: EDS spectra of indicated complex phase.

At longer periods of contact (more that 30 minutes) another phase is detected,


consisting of approximately equal parts calcium and sulphur. Figure 5.31 shows this
phase and the EDS determination of its composition. This phase is semi-continuous, but
occurred regularly on the interface. EDS analysis did not detect this phase prior to 60
minutes contact, so it can be deduced that this phase formed some time after 30 minutes
of reaction. Figure 5.32 shows an EDS map of the semi-continuous phase occurring at
the interface after 60 minutes of contact. It can be seen to cover large portions of the
interface, and is a complex phase containing calcium sulphide.

a. b.

c. d.

Figure 5.32: EDS image map illustrating interfacial layer formed between iron and
coke at 1550oC after 1 hour, a. shows secondary electron image, b. shows iron, c. shows
sulphur, and d. shows calcium.

112
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Calcium present in the ash as CaO is a minor component of the ash (comprising 1.95%
of the total coke ash content). EDS analysis shows that it is present at the interface to a
significant degree, in the later stages of contact after 30 minutes. Calcium appears to be
associated with an increase in the local sulphur levels. This suggests the
desulphurisation of the liquid iron by the calcium oxide in the ash (see section 2.4.4), or
more specifically:

CaO( s ) + S + C sat → CaS ( s ) + CO( g )

Equation 2.17 (reproduced from Chapter 2)

This reaction has the following equation for the Gibbs free energy [69]:
∆G o = 115060 − 113 .60T J.mole-1 with T (K)

Equation 5.1

At a temperature of 1550oC, this gives a Gibbs free energy of –92kJ.mole-1.

This supposition is supported by a previous study between natural graphite and liquid
iron [14], which found the calcium sulphide phase formed at the interface and a drop in
sulphur levels was detected in the droplet corresponding with the formation of the
calcium sulphide. The sulphur levels in this natural graphite study were much less than
in the present study. This may mean that the detection of the desulphurisation of the
liquid metal may not be as easily detected in this case because of the higher levels of
sulphur present tend to obscure smaller amounts of sulphur removed during the
desulphurisation.

The equation above suggests that the desulphurisation reaction also consumes solute
carbon in liquid iron. However, as this reaction can be seen to occur only at longer
contact times (>30 minutes contact time), and therefore it probably is not responsible
for the consumption of solution carbon seen in the early contact times (<30 minutes
contact).

113
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Visual observations, which were made during the experiment, of the liquid droplet on
the coke substrate are shown in a series of still images in Figure 5.33. This figure
illustrates the reaction occurring at the interface during contact. Gas has evolved at the
interface, which bubbled up through the liquid iron. This gas formation is further
evidence suggesting reduction reactions occurred.

a. b. c.

Figure 5.33: Still images of liquid electrolytic iron droplet on coke substrate, a. shows
droplet 1 second after melting, b. and c. show reaction occurring between 1 second and
2 minutes, with bubbles of gas forming within the droplet and a visible gas surrounding
the droplet.

The reduction of silica and to a lesser extent iron oxide, as it proceeds via the gas phase
route, may occur at the interface and be responsible for the bubbling observed in Figure
5.33. Moreover, the droplet has no external stirring, and only the influence of
Marangoni flow is assisting mass transfer of the carbon away from the interface. It is
likely that it is silicon monoxide gas that is present surrounding the entire droplet,
visible as a halo in Figure 5.33b and c. White crystalline deposits identified as SiO2
were found in the gas outlet filter from the furnace showing further evidence of SiO(g).

5.2.2 Char Interfacial Reaction Products

The interfacial region between liquid iron and two chars (Char 1 and Char 4) was
investigated using FESEM and EDS analysis. The result for Char 4 after 5 minutes of
contact with liquid iron is shown in Figure 5.34.

114
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.34: Left: Interface of metal droplet after 5 minutes contact of iron with Char 4,
x350 magnification. Right: EDS spectra of region indicated.

Aluminium, calcium, sulphur, oxygen and iron were detected in the interfacial layer.
The white phase was mainly alumina. A higher magnification image of this region is
shown in Figure 5.35.

Figure 5.35: Left: High magnification image of interface of metal droplet after 5
minutes contact with Char 4, showing fused ash material and iron surface, x1500
magnification. Right: EDS spectra of region indicated.

From these figures, ash deposits and iron can be clearly observed. Small concentrations
of magnesium and silicon were also detected indicating the partial presence of the
parent ash. While the ash in Char 4 contained 53.8wt% silica, the detected silicon peak

115
CHAPTER 5: Results and Discussion – Sessile Drop Technique

at the interface was very small and could be found only in some of the areas at the
interface.

Figure 5.36 and Figure 5.37 show the interfacial region after contacting metal droplets
with Char 4 for 3 hours.

Figure 5.36: Left: Interface of metal droplet after 3 hours contact with Char 4, showing
high interfacial coverage, x350 magnification. Right: EDS spectra of region indicated.

Figure 5.37: Left: Interface of metal droplet after 3 hours contact with Char 4, showing
calcium sulphide complex, and iron surface, x1500 magnification. Right: EDS spectra
of region indicated.

From these two figures, it is clear that the interfacial coverage has increased
considerably over time. Significantly less iron was measured indicating a further

116
CHAPTER 5: Results and Discussion – Sessile Drop Technique

decrease in iron/carbon contact area. EDS spectra (Figure 5.36 and Figure 5.37) showed
a significant increase in the concentrations of both calcium and sulphur at the interface.
Levels of Ca detected in this region were quite high considering the concentration of
CaO was only 2.5wt% of the ash in Char 4. This enrichment appeared to be linked with
increased sulphur levels at the interface, and was seen in all samples. Char was the only
significant source of sulphur in this system, which is picked up by the metal during
contact with char. Enrichment of calcium along with the association of calcium and
sulphur in the interfacial region suggests that the calcium oxide in the char ash was
removing sulphur from the liquid iron (see Equation 2.17 and data supporting this is
discussed in section 5.3). This desulphurisation reaction produces calcium sulphide,
which was detected on the interfacial region between Char 4 and iron.

In Figure 5.36 and Figure 5.37, the aluminium peak had become much smaller as
compared to the one observed in Figure 5.34 and Figure 5.35 after 5 minutes of contact.
The thickness of the interfacial layer showed an increase with time. With carbon
dissolving in liquid iron, alumina present in ash remains at the interface, as the carbon is
removed from the char preferentially. However due to non-wetting between alumina
with liquid iron, it has a tendency to stay closer to the char substrate away from liquid
iron if possible. In non-wetting situations the contact area between the two materials,
iron and alumina, is minimised. If another phase is also present, which has better
wetting with the iron, it would be expected that this phase would be in contact more
with the iron. This would be particularly the case if this phase were actually being
produced via reaction with the liquid iron. When the iron droplet was physically
removed from the char substrate after quenching, it is likely that alumina deposits
remained on the char substrate and therefore were not detected in the interfacial layer
(underside of the iron droplet). Due to the nature of the char substrates, it was not
practical to analyse the char remaining in the region of the interface as the char was too
friable and generally could not be handled after the experiment.

An important result emerging from Figure 5.34 to Figure 5.37 relates to the detection of
very small levels of silicon in the interfacial region. Despite high levels of silica in the
char initially (53.8 wt% of the ash present), very low levels of Si were detected in the
EDS spectra even after three hours of contact. This result suggests that silica was
undergoing reduction reactions and was not present in the interfacial region. In review
117
CHAPTER 5: Results and Discussion – Sessile Drop Technique

of the studies on coke/silica reaction considered in the literature survey (section 2.4.1),
at temperatures around 1550°C a large proportion of the silica in the coke is believed to
be reduced. Based on thermodynamic considerations, SiO2 present in ash will
eventually get reduced by carbon at high temperatures, leading to the evolution of CO
and CO2 gases. The reaction can be seen in Equation 2.6 and Equation 2.8:

SiO2 + C → SiO( g ) + CO( g )

Equation 2.6: Partial reduction to form silicon monoxide vapour.

SiO( g ) + C → Si + CO( g )

Equation 2.8: Silicon transfer to carbon-containing iron.

Silica present in char could either escape the system as SiO(g) or dissolve in liquid iron
as silicon. In the study presented in section 5.2.1 on the interfacial reactions between
coke and liquid iron, a significant transfer of silicon into liquid metal was recorded.
With the penetration of a 20 KeV electron beam being limited to ~2 microns, it was
probably unable to probe deeper regions of the interfacial layer and could therefore not
detect the presence of silicon in the metal droplet. Another factor worth mentioning is
that the silica reduction does not have to be limited to the interface area. The silica
reduction can proceed via the gas phase and remove carbon from solution in the liquid
iron at the gas/liquid interface also.

Similar trends as described for Char 4 were seen with Char 1 (see Figure 5.38).

118
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.38: Interface of metal droplet after 5 minutes contact with Char 1, left x350
magnification, right x1500 magnification.

The concentrations of ash deposits were higher for Char 1, probably due to this char’s
higher concentrations of alumina and calcia in the ash. Alumina concentration in Char 1
ash was 37.3 wt% as compared to 25 wt% in Char 4 ash. In addition, the initial CaO
concentration in Char 1 ash was ~1.4 times the concentration in Char 4 ash. Char 1 also
had a higher sulphur level in the ash. It could be seen that the calcium was once again
associated with sulphur and was present as CaS (possibly in a complex phase). Regions
rich in Fe and CaS could clearly be identified. The morphology of the CaS phase
appeared the same across both chars. However, a comparison between two chars after 5
minutes of contact (Figure 5.34 and Figure 5.38) suggests that the Char 1, with a higher
CaO level in the ash, appeared to form CaS earlier in the reaction. Figure 5.39 shows
the interface of the metal droplet after 3 hours contact with Char 1.

119
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Figure 5.39: Interface of metal droplet after 3 hours contact with Char 1, showing high
amount of interfacial coverage, left x350 magnification, right x1500 magnification.

The interfacial region, after three hours of contact (Figure 5.39), once again showed a
higher concentration of interfacial deposits as compared to Char 4. The EDS analysis
results for Char 1 were similar to those for Char 4 and very small levels of silicon were
detected in the metal contacting Char 1 as well.

5.2.3 Interfacial Reaction Products Summary

In the previous two sections it can be seen that the interfacial products formed for the
two chars and the coke were very similar. In comparison, the following points of
interest can be seen:
• Calcium sulphide appears to form earlier with the chars than for the coke,
• Calcium sulphide coverage is much more extensive for the chars than for the
coke at corresponding times,
• Silica reduction appears to start at around the same time for both the chars and
the coke,
• The formation of the alumina layer is seen for all the materials studied.

The formation of the calcium sulphide layer appears to be related to the amount of
calcium oxide present in the carbonaceous material. From the FESEM observations of
the interface, Char 1 forms calcium sulphide earliest and Coke the latest. In section
5.2.2 we can see that the chars both show evidence of calcium sulphide formation at 5

120
CHAPTER 5: Results and Discussion – Sessile Drop Technique

minutes contact, the coke did not have calcium sulphide detected until around 60
minutes of contact. Comparisons of the compositions of the chars/coke reveal that Char
1 has ~1.4 times the amount of CaO in the ash initially compared to Char 4 and ~1.7
times that of coke. This compositional difference appears to play a significant role in the
formation of this phase and the speed of its formation, as one might expect. The two
chars and coke in this study have very similar sulphur levels, and in the next section it is
seen that the sulphur pickup by the metal in contact with these materials is also similar.
Thus it is believed that the difference seen in initiation time for the formation of calcium
sulphide is more likely due to the differences in calcium oxide composition in the
carbonaceous materials.

The silica may be reduced by the carbon in liquid iron solution, and also in-situ by the
solid carbon. Iron oxide and silica reduction reactions are significant at the interface,
with the silica having the greatest effect due to the high levels in the ash. The
differences in the silica levels in the materials (Char 4 has 1.4 times the silica and Coke
has 1.7 times that compared to Char 1) does not appear to have an influence on starting
time for the reaction, but it may be considered to have an impact on the amount of
silicon produced.

The formation of the alumina layer is seen clearly in the coke, and also observed in the
interfacial layer after 5 minutes contact for the chars. Whilst the layer formation in the
chars is not seen at later periods of contact it may be inferred. This layer, in combination
with the calcium sulphide layer, will with time act to restrict interfacial area available
for reaction. Practically this may bring about a significant reduction in the apparent rate
of carbon dissolution, particularly after longer periods of contact, or in systems with
little or poor stirring that may not allow the products formed at the interface to be
removed.

121
CHAPTER 5: Results and Discussion – Sessile Drop Technique

5.3 Carbon/Sulphur/Silicon Transfer Studies

In the sessile drop studies between chars/coke and liquid iron, a number of interesting
features were noted and are given below:
• The carbon levels remained very small after contact with the carbonaceous
material for a long period of time
• Some evidence of carbon transfer was noticed after the initial period of low
carbon levels
• The sulphur transfer occurred rapidly for all materials.

5.3.1 Coke/Iron System

The droplets produced after varying the time at temperature, were analysed for carbon
and sulphur content, with the results shown in Figure 5.40.

2.5 0.2

0.18

2 0.16

Sulphur Concentration (wt%)


Carbon Concentration (wt%)

0.14

1.5 0.12

0.1

1 0.08

0.06

0.5 0.04
Carbon
Sulphur 0.02

0 0
0 50 100 150 200
Time (min)

Figure 5.40: Carbon and sulphur transfer from coke into liquid iron at 1550oC

The results for carbon show a marked delay in carbon transfer, to approximately 30
minutes, after which carbon transfer occurs. This transfer is quite slow however, and
122
CHAPTER 5: Results and Discussion – Sessile Drop Technique

even after 3 hours contact, the droplet has not reached carbon saturation (calculated
after Neumann et al. [88] to be approximately 5.3wt% carbon). The transfer of
significant amounts of carbon does not occur until 30 to 60 minutes.

Several factors could be limiting the carbon transfer and/or accumulation. The
production of a solid (or semi-fused) barrier of ash at the interface would reduce the
contact area available for mass transfer. Reduction reactions may also be consuming the
carbon in solution in iron. Some possible reduction reactions that could consume the
carbon are the reduction reactions involving iron oxide and silica in the coke ash.
Carbon transfer could also be limited by slower rates of carbon dissolution from non-
graphitic materials compared to that from graphites.

In the sessile drop arrangement, it is likely that the carbon accumulation is being
hindered by the consumption of the carbon in solution in the small droplet as formation
of an ash barrier would be more likely after longer periods of contact. Reactions such as
the reduction of iron and silicon oxides, for example, are consuming the carbon almost
as fast as it can be transferred to the droplet. It is not until these reactions slow due to
depletion of reactants at the interface, that appreciable carbon can accumulate within the
iron droplet.

Further evidence for this can be seen when a comparison is made between the
phenomena seen in this study and one between iron and natural graphite [14]. This
study uses a similar technique, and the natural graphite has a roughly comparable
composition to the chars examined (8.8% ash, however the ash contains much less
alumina and more silica and iron oxide than the chars). When the carbon transfer with
time is examined, the behaviour seen is very different, see Figure 5.41.

123
CHAPTER 5: Results and Discussion – Sessile Drop Technique

Carbon% (wt)
4

0
0 10 20 30 40 50 60 70
Time (min)

Figure 5.41: Variation of droplet carbon content with time during natural graphite
/iron droplet interactions, from [14].

Even from the initial stage, carbon transfer into the iron droplet is observed [14], after 1
second there is 4wt% carbon in the droplet, which was initially electrolytic iron. The
natural graphite had a similar amount of ash, with higher silica and iron oxide contents
than the char but with less alumina, and the carbon transfer rate is much higher. The
system did exhibit the formation of an interfacial layer, but the transfer of carbon
occurred, to a significant extent, mostly before this could form. This highlights the
differences in carbon dissolution rate between the two materials. The non-graphitic
material has an intrinsic carbon transfer rate that is much lower than that for natural
graphite. Comparison of this study with the current work highlights the role of ash
composition on the nature of the interfacial layer. The interfacial layer in the natural
graphite study contained calcium sulphide but had little alumina (due to the lower levels
in the ash). There was also significant iron in the interfacial layer, which was not
observed in the coke/iron studies. The faster intrinsic rate of carbon dissolution from the
natural graphite means that the influence of ash on carbon dissolution is not significant
(because the layer is formed after the majority of carbon transfer), whereas in the coke
studies, the slower rate of carbon dissolution means that the role of ash is very
significant.

124
CHAPTER 5: Results and Discussion – Sessile Drop Technique

In this study, the second set of droplets that were produced were analysed for silicon
content, and the results of this are shown in Figure 5.42. The transfer of silicon can be
clearly seen.

2.5 0.35

0.3
Carbon Concentration (wt%)

Silicon Concentration (wt%)


2
0.25

1.5
0.2

0.15
1

0.1
0.5
Carbon 0.05
Silicon
0 0
0 10 20 30 40 50 60
Time (min)

Figure 5.42: Silicon transfer from coke into liquid iron at 1550oC

This graph shows the accumulation of silicon in the liquid iron as a function of time and
compares this with the carbon level seen in the droplet. The pick up of silicon into the
liquid iron occurs reasonably rapidly, and appears to be virtually complete after
approximately 30 minutes contact. That silicon is transferring into liquid iron suggests
that the route of reduction via solute carbon does also occur together with the in situ
reduction of the silica. Calculations based on total silicon in the system initially only
present as silica in the coke ash, and the total amount accumulated in the liquid iron,
suggest that only a small proportion (less than 3%) of the available silicon transfers into
the liquid iron. It is likely that the majority of the silicon is removed as silicon
monoxide gas in the off gas stream.

Analysis of the droplet’s silicon content (Figure 5.42) suggests that the silica reduction
is the major cause of the lack of carbon accumulation as silica is the main reducible
oxide in the coke ash with iron and other oxides being present only in small quantities.

125
CHAPTER 5: Results and Discussion – Sessile Drop Technique

There is further confirmation of this hypothesis: it is noted that the time to reach a stable
silicon content in the liquid metal, is approximately the same time as when carbon
begins to accumulate within the droplet. After approximately 30 minutes, the silicon
content levels off, coming to a stable value of approximately 0.3wt%, and the carbon
levels increase from the very low levels. The consumption of the solute carbon
combined with low carbon transfer rates together ensure that the carbon concentration in
the metal remains low until the rate of silicon transfer into iron drops to negligible
levels.

Combining this with the data from the off gas analysis (Figure 5.30), we can see that in
relatively short time intervals, the reduction of silica can have a large impact on the
carbon transfer behaviour into liquid iron. The transfer of silicon into the iron will
influence the dissolution kinetics. Increases in the silicon levels in solution will decrease
the carbon saturation level, decreasing the overall driving force for carbon dissolution.
This will also mean that for the same weight percent of carbon in the liquid metal, the
subsaturation (difference between level of carbon and the theoretical saturation limit)
will be less.

5.3.2 Char/Iron System

Figure 5.43 and Figure 5.44 show carbon and sulphur transfer from, respectively Char 4
and Char 1 into iron droplets at 1550°C as a function of contact time.

126
CHAPTER 5: Results and Discussion – Sessile Drop Technique

0.6 0.2
0.18
0.5

Sulphur Concentration (wt%)


Carbon Concentration (wt%)
0.16
0.14
0.4
0.12
0.3 0.1
0.08
0.2
0.06
0.04
0.1
Carbon
0.02
Sulphur
0 0
0 50 100 150 200
Time (min)

Figure 5.43: Carbon and sulphur transfer from Char 4 into liquid iron at 1550oC

0.6 0.2
0.18
0.5
Carbon Concentration (wt%)

0.16

Sulphur Concentration (wt%)


0.14
0.4
0.12
0.3 0.1
0.08
0.2
0.06
0.04
0.1
Carbon
0.02
Sulphur
0 0
0 50 100 150 200
Time (min)

Figure 5.44: Carbon and sulphur transfer from Char 1 into liquid iron at 1550oC

For Char 4, the carbon levels showed a sharp peak with a maximum of 0.48 wt% after 5
minutes of contact followed by a sharp dip and approximately stable levels (~ 0.1 wt%)

127
CHAPTER 5: Results and Discussion – Sessile Drop Technique

up to 3 hours. The corresponding results for Char 1 were slightly different. The carbon
levels for Char 1 showed a small peak after ~ 5 minutes of contact and then tended to
decrease slightly for some time followed by a small increase later reaching a maximum
of 0.28 wt%. Carbon levels in liquid iron droplet for both chars were much below the
saturation level of ~ 5.4 wt% at these temperatures.

From Figure 5.43 and Figure 5.44, it can be seen that the transfer of sulphur from both
chars into the liquid iron occurred quite rapidly in the initial stages. While the sulphur
content in the liquid iron in contact with Char 4 reached a steady level by 30 minutes
(0.16%), the liquid in contact with Char 1 continued to accumulate sulphur for up to 3
hours, eventually reaching 0.2 wt%.

Several mechanisms could be responsible for the observed carbon and sulphur transfers
from chars into liquid iron. It appears that transfer of carbon from chars into the liquid
iron took place in the initial stages of contact and accumulated until reduction reactions
started consuming solute carbon. Reduction reactions of ash oxides such as SiO2 and
Fe2O3 are quite likely at temperatures around 1550°C [18]. Iron oxide will reduce
rapidly [66] under these conditions and is expected to be totally reduced within one
minute of contact. Both chars, however, contain significantly less iron oxide than silica,
so the relative impact of iron oxide reduction on liquid metal carbon levels is expected
to be less than that of silica reduction.

Reduction of silica, (Equation 2.6 and Equation 2.8), is likely to consume some of the
solute carbon from the iron droplet lowering its carbon content, and could therefore
explain the decrease in solute carbon concentration. The peak in carbon at the 3-5
minute mark also suggests that the two reactions, transfer of carbon and the reduction of
silica, probably do not initiate at the same time. There are a number of possible reasons.
It may take some time for silica to form silica monoxide due to small delays in its
heating up to the reaction temperature. Flow of SiO(g) may also take some time to reach
the liquid iron interface. The reduction of the silica in the char ash is expected to have
an impact on blast furnaces using PCI coal, as it can consume some of the carbon in
solution. In contrast, the transfer of the sulphur reaches a reasonably steady level fairly
quickly; at least half of the final sulphur level is transferred in the first minute of
contact.
128
CHAPTER 5: Results and Discussion – Sessile Drop Technique

From Figure 5.43 and Figure 5.44, it can be seen that liquid iron doesn’t seem to pick
up carbon even after long exposure to Char 4, whereas a marginal increase was
observed for Char 1. This may be explained in terms of the char ash composition. Char
1 has a lower concentration of reducible oxides, and if the reduction of these oxides
occurred at a similar rate for both chars, Char 1 will have carbon consuming reactions
occurring for a shorter time than Char 4. This will result in a shorter period of no net
carbon transfer to the liquid iron from the char. The formation of reaction products/ash
was much more in the case of Char 1 as compared to Char 4. After 3 hours of contact, it
appeared to have formed a continuous deposit on the iron/char interface. This would
lead to a much-reduced contact between carbon and liquid iron and could inhibit carbon
transfer. The consumption of solute carbon by reducible oxides and the subsequent
formation of a dense interfacial layer are seen to significantly affect carbon dissolution
from these chars.

Similarly for the coke/iron mass transfer study (section 5.3.1) we can compare these
results with a sessile drop study on iron and natural graphite [14]. The carbon transfer
into the liquid iron droplet (see Figure 5.41) reached 4wt% carbon after 1 minute of
contact. The natural graphite used has a similar weight percentage of silica in the ash as
these chars, and also comparable levels of carbon. However the carbon transfer rates
were much higher. The system did exhibit the formation of an interfacial layer, but the
transfer of carbon had occurred, to a significant extent, mostly before this could form.
This highlights the differences in carbon dissolution rate between the two materials. The
non-graphitic material has an intrinsic carbon transfer rate that is much lower than that
for natural graphite [79]. If the rate at which carbon dissolves from chars into liquid iron
is comparable to the rate at which solute carbon is consumed by reducible oxides,
carbon levels in liquid iron would remain at low levels, far below saturation limit.
Comparisons with the sulphur transfer in the natural graphite study [14] also highlight a
major difference. In natural graphite, the sulphur transfer occurs rapidly, and peaks at
around 1-2 minutes and is seen to significantly drop off, see Figure 5.45.

129
CHAPTER 5: Results and Discussion – Sessile Drop Technique

0.03

0.025

0.02

Sulfur%(wt)
0.015

0.01

0.005

0
0 10 20 30 40 50 60 70
Time (min)

Figure 5.45: Variation of droplet sulphur content with time during natural graphite
/iron droplet interactions, from [14].

A comparison of the two studies shows two main points of interest. Firstly, the sulphur
levels transferred in the natural graphite study are significantly lower than in the char
study, despite the sulphur levels being similar for the different materials. Secondly they
both show a drop in sulphur levels after an initial peak. The drop in sulphur seen in the
graphite study was attributed to the formation of calcium sulphide, and as calcium
sulphide was observed forming in the char study, it is reasonable to assume the sulphur
drop in the char study is also caused by this formation. The higher levels of sulphur
transferred in the char study ensure that the amount of sulphur removed is not as
significant overall, than it is in the natural graphite case. The comparison of these two
sets of results emphasises an interesting phenomenon that may be seen when you
compare two different materials that transfer carbon very differently. Early work by
Moriss et al. [105] suggested that carbon and sulphur in solution in liquid iron tend to be
mutually exclusive. Later research [106] confirmed this finding with respect to graphite
dissolution, and found that the carbon and sulphur atoms tend to displace each other,
resulting in two immiscible regions, with one region rich in carbon but low in sulphur,
and the second region low in carbon but high in sulphur. Higher levels of carbon in the
liquid would therefore result in a higher resistance to sulphur transfer into the liquid.
This finding may begin to explain the difference in sulphur levels in the droplets found
when reacted with chars and natural graphite. The carbon level in the drop when in
contact with natural graphite increases very rapidly, with near saturation levels after
only 1-2 minutes contact. In the case of the chars, the carbon levels are very different,

130
CHAPTER 5: Results and Discussion – Sessile Drop Technique

with carbon much below the saturation level for the full period of contact. The transfer
of sulphur in the case of graphite must therefore dissolve into carbon saturated iron,
which based on the previously mentioned research will cause a greater resistance to
mass transfer. The iron in contact with the chars contains much less carbon, and thus
does not have the same resistance.

5.3.3 Carbon/Sulphur/Silicon Transfer Summary

Focussing on the carbon/sulphur/silicon transfer in the iron droplet, the results on char
and coke are summarised below:
• For all materials investigated there appears to be a period of low/no carbon
accumulation, and this has been related to silica and iron oxide reduction
reactions consuming solution carbon as quickly as it is able to be transferred,
• The reduction of the silica was faster for coke than for the chars and the coke
demonstrates appreciable carbon accumulation after 30 minutes contact,
• The reduction of silica did not seem to be related to silica concentration for all
the materials, as coke had the highest level of silica,
• Char 1 appeared to start picking up carbon after 3 hours contact, as compared to
Char 4. This could be a due to Char 1 having lower silica levels. The transfer of
carbon after the completion of reduction reactions is likely to be hindered
however due to blockage of interface by interfacial reaction products,
• Sulphur can be seen to transfer rapidly.

131
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

CHAPTER 6: CARBON DISSOLUTION STUDIES USING THE CARBURISER


COVER METHOD: RESULTS AND DISCUSSION

Systematic investigations were carried out on four chars and one coke using the
carburiser cover method at 1550°C. Given below are the detailed experimental results, a
theoretical computation of interfacial contact area and dissolution rate constants
followed by a discussion on factors affecting carbon dissolution in these carbonaceous
materials.

6.1 Carbon Dissolution Studies

Carbon dissolution experiments were carried out on four chars and one coke at 1550 °C.
The initial liquid iron bath contained 2 wt% of carbon and 0.01 wt% of sulphur. The
carbonaceous material was added to the top of the melt and marked the start of time of
contact. Iron samples were removed from the bath periodically, with a silica glass
pipette. These samples were of the order of 1-2 grams only, and are not expected to
influence the bath size of 1.2kg. The total time of investigation was around 40 minutes.
Samples from the melt were taken every 30 seconds for the first 5 minutes, and
subsequently less frequently. The carbon and sulphur mass transfer occurs less rapidly
after this time and the rate of change is not expected to be high. The removed samples
were analysed for carbon and sulphur using the LECO CS244 (described in section
4.2.1.1). The bath velocity under induction stirring was estimated to be approximately
0.35 m/s, as measured by filming particle movement on the surface of the iron bath
using a CCD camera [107]. This result is comparable to the one reported by Kosaka and
Minowa [73].

A number of researchers have [5, 15-17, 47, 71-85] used a first order kinetic equation to
describe the dissolution of carbon from graphite into liquid iron, which was based on
the carbon concentration gradient (in weight percent).

dC t Ak '
= (C S − C t )
dt V

Equation 6.1
132
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

In this equation, Cs and Ct represent the saturation solubility and carbon concentration
(wt%) in the iron melt at time t, and k’ (m.s-1) is the first order dissolution rate constant.
A represents the interfacial contact area and V is the bath volume. The standard Gibbs
energy for the dissolution of carbon into liquid iron is [108]:

∆G o = 22593.6 − 42.2584T ( J / mol )

Equation 6.2

At 1550°C (1823K) ∆G°= -54.44 kJ/mol. At this temperature there is a strong


thermodynamic driving force for carbon dissolution. Integrating Equation 6.1 we have:

(C s − C t )
ln = − Kt
(C s − C 0 )

Equation 6.3

In this equation C0 is the initial carbon concentration in the melt. A combined


dissolution rate constant, K = Ak’/V, can be easily measured from the negative slope of
a ln((Cs-Ct)/(Cs-Co)) plot versus time. This parameter K has the dimensions of s-1. The
carbon saturation level, Cs was calculated as a function of time taking into account the
sulphur pick up by liquid iron and the initial silicon content of the metal [88]. The initial
value of Cs was taken as approximately 5.28 wt%.

The experimental results obtained for the four chars and coke are shown in Figure 6.1 to
Figure 6.15. These figures include carbon concentration in the melt as a function of
time, plots of ln((Cs-Ct)/(Cs-Co)) versus time, and curve fitting plots to determine
dissolution rate constants. The experimental results are discussed below for each
material.

133
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

6.1.1 Char 1 Dissolution Results

From Figure 6.1 it can be noted that the carbon concentration in the melt picked up
slowly and reaches a maximum of ~ 3 wt% after 15 minutes of contact. There is little
subsequent carbon transfer. This level is significantly below the saturation level of 5.28
wt% for liquid iron at 1550°C.

4
Carbon Concentration (wt%)

3.5

2.5

1.5

1
0 10 20 30 40 50
Time (min)

Figure 6.1: Carbon concentration (wt%) in the melt for dissolution of carbon from
Char 1 at 1550°C.

The plot of ln((Cs-Ct)/(Cs-Co)) versus time (shown in Figure 6.2) showed some
interesting features. Two distinct regions, marked I and II, can be clearly identified.
Carbon dissolution from Char 1 in region I (from 0.5 to 12 minutes) is significantly
more rapid as compared to region II (from 12 – 45 minutes). The gradients for the data
representing the first of these regimes were obtained from linear regression as shown in
Figure 6.3. The data points show a good fit to a straight line, thereby indicating the
validity of assuming first order kinetics (Equation 6.1 and Equation 6.3). From this an
apparent dissolution rate constant K may be computed (for the initial rate only) and if
the surface area is known the results may yield an intrinsic rate constant for the
controlling reaction.

134
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

-0.1

ln ((Cs-Ct)/(Cs-Co))
-0.2
I
-0.3

-0.4

-0.5
II

-0.6
0 10 20 30 40 50
Time (min)

Figure 6.2: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 1 into liquid iron at
1550°C. Two distinct regions (I and II) can be clearly identified.

0
y = -0.0382x
-0.05 2
R = 0.9529
-0.1
ln ((Cs-Ct)/(Cs-Co))

-0.15

-0.2

-0.25

-0.3

-0.35

-0.4
0 2 4 6 8 10 12
Time (min)

Figure 6.3: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 1- region I, into an
iron-carbon-sulphur alloy at 1550°C.

6.1.2 Char 2 Dissolution Results

From Figure 6.4 it can be noted that carbon concentration in the melt picked up slowly
for Char 2 and kept on increasing slowly even up to 50 minutes. This behaviour is quite
different from the behaviour observed for Char 1. The final carbon level is once again
significantly below the saturation level of 5.28 wt% for liquid iron.

135
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

3.5

Carbon Concentration (wt%)


3

2.5

1.5

1
0 10 20 30 40 50
Time (min)

Figure 6.4: Carbon concentration (wt%) in the melt for dissolution of carbon from
Char 2 at 1550°C.

The plot of ln((Cs-Ct)/(Cs-Co)) versus time (shown in Figure 6.5) also showed a trend
different from char 1. Only one region can be clearly identified, with no evidence of any
change in the rate of carbon dissolution. Through a curve fitting exercise (Figure 6.6)
the slope of curve, representing combined dissolution rate constant K, was accurately
evaluated. The data points showed a good fit to a straight line, thereby indicating the
validity of first order kinetic equations (Equation 6.1 and Equation 6.3).

-0.05
ln ((Cs-Ct)/(Cs-Co))

-0.1
I
-0.15

-0.2

-0.25

-0.3
0 10 20 30 40 50
Time (min)

Figure 6.5: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 2 into liquid iron at
1550°C. Only one distinct region (I) can be identified for Char 2.
136
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

y = -0.0049x
2
-0.05 R = 0.9743

ln ((Cs-Ct)/(Cs-Co))
-0.1

-0.15

-0.2

-0.25
0 10 20 30 40 50
Time (min)

Figure 6.6: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 2 for region I into
liquid iron at 1550°C.

6.1.3 Char 3 Dissolution Results

The behaviour of Char 3 was quite similar to that observed for Char 1. From Figure 6.7,
it can be noted that the carbon concentration in the melt picked up rapidly and reached a
maximum of ~ 3 wt% after 45 minutes.

4
Carbon Concentration (wt%)

3.5

2.5

1.5

1
0 10 20 30 40 50
Time (min)

Figure 6.7: Carbon concentration (wt%) in the melt for dissolution of carbon from
Char 3 at 1550°C.

137
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

The plot of ln((Cs-Ct)/(Cs-Co)) versus time (shown in Figure 6.8) showed features
similar to ones observed for Char 1. The final carbon level was significantly lower than
the carbon saturation level of 5.28wt%. Two distinct regions, marked as I and II, can be
clearly identified. The slopes of the straight-line region I, representing the combined
dissolution rate constant K, was accurately evaluated, see Figure 6.9. The data points
showed a good fit to a straight line, thereby indicating the validity of first order kinetic
equations (Equation 6.1 and Equation 6.3).

-0.1

-0.2
ln ((Cs-Ct)/(Cs-Co))

-0.3 I
-0.4

-0.5 II
-0.6

-0.7
0 10 20 30 40 50
Time (min)

Figure 6.8: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 3 into liquid iron at
1550°C. Two distinct regions (I and II) can be clearly identified.

0
y = -0.013x
-0.05 2
R = 0.9865

-0.1
ln ((Cs-Ct)/(Cs-Co))

-0.15

-0.2

-0.25

-0.3

-0.35
0 5 10 15 20 25 30
Time (min)

Figure 6.9: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 3- region I, into an
iron-carbon-sulphur alloy at 1550°C.
138
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

6.1.4 Char 4 Dissolution Results

From Figure 6.10 it can be noted that carbon concentration in the melt picked up very
slowly for Char 4 and increased slowly even up to 50 minutes. The behaviour of Char 4
was quite similar to that observed for Char 2. The final carbon level was once again
below the saturation level of 5.28 wt% for liquid iron at 1550°C.

3.5
Carbon Concentration (wt%)

2.5

1.5

1
0 10 20 30 40 50
Time (min)

Figure 6.10: Carbon concentration (wt%) in the melt for dissolution of carbon from
Char 4 at 1550 °C.

The plot of ln((Cs-Ct)/(Cs-Co)) versus time (shown in Figure 6.11) also showed a trend
similar to Char 2. Only one region can be clearly identified with no evidence of region
II. The slope of the curve, representing the combined dissolution rate constant K, was
accurately evaluated in Figure 6.12. The data points again showed a good fit to a
straight line.

139
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

-0.1

-0.2

ln ((Cs-Ct)/(Cs-Co))
-0.3 I

-0.4

-0.5

-0.6

-0.7
0 10 20 30 40 50
Time (min)

Figure 6.11: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 4 into liquid iron
at 1550°C. Only one distinct region (I) can be identified for Char 4.

0
y = -0.0132x
2
-0.1 R = 0.9824

-0.2
ln ((Cs-Ct)/(Cs-Co))

-0.3

-0.4

-0.5

-0.6

-0.7
0 10 20 30 40 50
Time (min)

Figure 6.12: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Char 4- region I, into an
iron-carbon-sulphur alloy at 1550°C.

6.1.5 Coke Dissolution Results

From Figure 6.13 it can be noted that the carbon transfer from coke into the melt was
relatively slow and reached a maximum of ~ 2.8 wt% after 7 minutes of contact. There
was very little carbon transfer after this. This final carbon content, after 40 minutes of

140
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

contact with the liquid iron was also significantly below the saturation level of 5.28
wt% for liquid iron at 1550°C.

3.5
Carbon Concentration (wt%)

2.5

1.5

1
0 5 10 15 20 25 30 35 40 45
Time (min)

Figure 6.13: Carbon concentration (wt%) in the melt for dissolution of carbon from
coke at 1550 °C.

The plot of ln((Cs-Ct)/(Cs-Co)) versus time (Figure 6.14) showed two distinct rate
regions, marked as I and II, a behaviour similar to the one observed for Char 1 and Char
3. However region I was limited to only 7 minutes as compared to 15 minutes for Char
1 and 25 minutes for Char 3. The slope of the plot shown in Figure 6.14 was accurately
determined, see Figure 6.15. The data points showed a good fit to a straight line for
coke.

141
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

-0.1
I
-0.2

ln ((Cs-Ct)/(Cs-Co))
-0.3

-0.4
II
-0.5

-0.6

-0.7
0 5 10 15 20 25 30 35 40 45
Time (min)

Figure 6.14: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for coke into liquid iron at
1550°C. Two distinct regions (I and II) can be clearly identified.

0
y = -0.0318x
R2 = 0.9618
-0.05
ln ((Cs-Ct)/(Cs-Co))

-0.1

-0.15

-0.2

-0.25
0 1 2 3 4 5 6 7 8
Time (min)

Figure 6.15: Plot of ln((Cs-Ct)/(Cs-Co)) with respect to time for Coke- region I, into an
iron-carbon-sulphur alloy at 1550°C.

6.2 Sulphur Transfer

The pick-up of sulphur by liquid iron from all five carbonaceous materials is shown in
Figure 6.16 to Figure 6.20. These sulphur levels in liquid iron were used to calculate the
carbon saturation value Cs. Final sulphur levels were quite small for all materials and
appeared to stabilise around 0.025 wt% in most cases.

142
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

0.03

0.025

Sulphur Concentration (wt%)


0.02

0.015

0.01

0.005

0
0 10 20 30 40 50
Time (min)

Figure 6.16: Sulphur dissolution profile for Char 1 at 1550°C.

0.03

0.025
Sulphur Concentration (wt%)

0.02

0.015

0.01

0.005

0
0 10 20 30 40 50
Time (min)

Figure 6.17: Sulphur dissolution profile for Char 2 at 1550°C.

143
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

0.03

0.025

Sulphur Concentration (wt%)


0.02

0.015

0.01

0.005

0
0 10 20 30 40 50
Time (min)

Figure 6.18: Sulphur dissolution profile for Char 3 at 1550°C.

0.03

0.025
Suphur Concentration (wt%)

0.02

0.015

0.01

0.005

0
0 10 20 30 40 50
Time (min)

Figure 6.19: Sulphur dissolution profile for Char 4 at 1550°C.

144
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

0.03

0.025

Sulphur Concentration (wt%)


0.02

0.015

0.01

0.005

0
0 5 10 15 20 25 30 35 40 45
Time (min)

Figure 6.20: Sulphur dissolution profile for Coke at 1550°C.

The pick-up of sulphur by the liquid metal for the different chars/coke was found to be
similar, with the rate of transfer and the final levels being quite comparable. The initial
levels of sulphur in the chars/coke were also quite similar and ranged from 0.35 to
0.56wt%. The presence of sulphur in the liquid iron is known to have a strongly
retarding effect on carbon dissolution [16, 75, 103]. The effect of sulphur was even
more pronounced for non-graphitic carbon dissolution, with small additions of sulphur
into the liquid metal resulting in large depressions in carbon dissolution rate. Sulphur is
well known to be surface active and is likely to be present in the interfacial region in the
form of sulphur based products such as CaS. A recent atomistic Monte Carlo
investigation by Sahajwalla and Khanna [80] has thrown some new light on the role of
sulphur. Experimental investigations on liquid iron containing C and S have shown that
the liquid had a tendency to separate into two immiscible layers, one rich in C but low
in S, and other rich in S but low in C. The presence of sulphur therefore has a tendency
to displace C from its immediate neighbourhood. Sulphur in liquid iron would therefore
oppose a build-up of local concentration of carbon in the liquid, resulting in a reduction
in carbon dissolution. Along with affecting carbon dissolution rates, chemical reactions
of S with CaO would also lead to the deposition of CaS in the interfacial layer.

145
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

6.3 Combined Dissolution Rate Constants

The combined dissolution rate constants K, as measured from the slopes of ln((Cs-
Ct)/(Cs-Co)) versus time plots, are tabulated below in Table 6.1. While Char 1, Char 3
and the Coke clearly showed two regions with distinct slopes, only one region was
observed for both Char 2 and Char 4. The rate constants for region I, as computed from
the experimental data, are given below in Table 6.1.

Table 6.1: Combined dissolution rate constants for a range of coal-chars and coke at
1550 °C

Carbonaceous Material K x103 s-1 (Region I)


Char 1 0.64
Char 2 0.08
Char 3 0.22
Char 4 0.22
Coke 0.53

Combined dissolution rate constants, as determined using Equation 6.3, showed a wide
variation. These values depend on the physical characteristics of each material and the
interfacial area of contact. Interfacial area of contact could vary significantly for these
materials, based on wettability, and needs to be evaluated before a detailed comparison
between these materials. The role of geometric factors also needs to be established
before the results obtained in this study could be compared with published data.
Researchers using the carburiser cover method have tended to use the area of cross-
section of the crucible as the estimated area of contact between the carburiser and the
liquid metal. This simple assumption however does not take into account the effect of
wettability, particle size/packing and other material properties.

As the computation of interfacial contact area is rather difficult in the case of the
carburiser cover technique, a theoretical model was developed using a force balance
approach, taking into account a number of material characteristics. This model was
developed with support from Dr. R. Khanna and Dr. H. Sun and their help is gratefully
acknowledged. Basic model and contact area calculations are given in the next section.
146
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

6.4 Estimation of Particle Contact Area

In the carburiser cover method, a known weight of particulate carbonaceous material


(~150 gm) was introduced on top of a bath with cylindrical cross-section, containing 1.2
kg of liquid iron. The first order dissolution rate of carbon from the particles depends on
the interfacial area of contact with liquid iron, which could vary significantly for the
completely or partially engulfed particles. An exact computation of solid/liquid contact
area is difficult and is rather complex due to the interplay of a large number of factors,
such as partial penetration of particles in the liquid, particle size distribution, packing
density of particles on the liquid surface, total weight of carbonaceous materials,
penetration of liquid iron in the crevices on the surface and into carbonaceous material,
and volatile release. Based on a number of simplifying assumptions, an attempt has
been made to estimate the solid/liquid contact area for the carburiser cover method in
the initial stages of contact. These assumptions are briefly discussed below.

The first assumption was with regard to the particle size distribution and total weight of
the material. The introduction of carbonaceous material on top of a turbulent liquid iron
bath generally results in the release of volatile matter and loss of fine particles leading
to a weight loss. Exact weight loss is difficult to calculate and was assumed to be ~ 20%
for these preliminary calculations. Neglecting particle size distribution, an average
particle size, as determined from microscopy and particle size measurements, was
associated with each char/coke particle.

The second assumption was with respect to the packing density of particles on the liquid
surface. When char/coke particles were added on top of the liquid iron bath in the
cylindrical cross-sectioned crucible in the induction furnace (internal diameter: 11 cm),
these reached an average height of ~ 2 cm, giving a total particle containment volume of
190 cm3. In view of the experimental configuration, these particles were assumed to
pack in layers in a random configuration. The total number of particles (N) in the
reaction cell was computed from the total weight (W) of the carbonaceous material and
mass of an individual particle (mp), expressed in Equation 6.4.

147
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

W W
N= =
mp 4 3
πr p ρ p
3

Equation 6.4

In this equation, rp and ρp are respectively the particle radius and density. The total
number of layers was then computed by dividing the total sample height by the average
particle diameter. This data was then used to compute average number of particles per
layer (nl) and fractional packing (expressed as a percentage - total particle area/liquid
surface area x100). The detailed results of these computations for four chars and coke
are given below in Table 6.2. A sample calculation for this is given in Appendix B.

Table 6.2: The packing density of carbonaceous particles on the bath surface

Carbonaceous Particle Particle Total No. Total No. No. of Fractional


Material Diameter Density of Particles of Layers particles packing
(µm) (Kg/m3) x 108 /layer (%)
-3 6
x 10 x 10
Char 1 25 1.78 82.2 800 10.35 53.5
Char 2 30 1.59 53.33 667 8.0 59.5
Char 3 100 1.65 1.39 200 0.695 57.4
Char 4 150 1.67 0.41 135 0.304 56.5
Coke 50 1.72 10.49 400 2.62 54.1

As seen clearly from Table 6.2, the packing density of the carbon particles over the bath
surface is only partial. This computation however was carried out based on particle
diameters and did not take into account the penetration depth of particles in the liquid.
When particles partially penetrate in the liquid, the surface area of contact is given by
2πrpx where x is the penetration depth of particle. Penetration of particle will therefore
result in an increased solid/liquid contact area. The possibility of particle entrainment
was discounted based on previous studies [109], which suggest that a significant
velocity is required to engulf a particle into a liquid melt, particularly when the particle
is non-wetting with the melt. Based on forces acting on the particle, the penetration
depth will be calculated in the following section.
148
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

6.4.1 Penetration Depth

When a spherical particle comes in contact with a liquid surface and tries to penetrate it,
the total force acting on the particle in a direction normal to the liquid surface can be
written as:

Ftotal = FG + FS + FB

Equation 6.5

The gravitational force, FG, acting vertically downwards depends on the mass of the
particle. In the present experiment, the layers of carbonaceous particles above the
particle in contact with liquid will also contribute to the downward gravitational force.
Total gravitational force is assumed to be equal to the mass of particle (mp) times the
number of layers (nl) above it. The force due to surface tension, FS, depends strongly on
the surface tension and wetting characteristics. Under non-wetting conditions (θ >90°),
it pushes the particle away from the liquid surface. Under wetting conditions (θ < 90°),
FS pulls the particle towards the liquid. The buoyancy force FB acts on the immersed
part of the particle and depends on the penetration depth x. A schematic representation
of a partially immersed particle is shown below in Figure 6.21.

a) b)

Figure 6.21: A schematic representation of particle in contact with liquid. Where a)


represents a wetting scenario (θ<90°), and b) represents a non-wetting scenario
(θ>90°) between the solid particle and liquid surface.

149
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

Under equilibrium conditions, the net force on the particle is zero. The downward
gravitational force balances the upward forces due to surface tension and buoyancy
[109]. The force balance equation can be written as:

4 3 1 x
πr p ρ p gn l = (3r p x 2 − x 3 ) ρ l g + 2πr p γ lv (1 − + cos θ )
3 3 rp

Equation 6.6

In this equation, ρl is the liquid density, γlv the liquid surface tension and θ the contact
angle. Equation 6.6 was solved numerically and the penetration depth of particles and
interfacial contact area was computed for all chars/coke. As contact angles for these
materials showed a variation with time, two sets of contact angles (I and II) were used
in these computations in an attempt to probe the effect of contact angles. This range is
expected to cover the possible variation in contact area caused by the variation in the
contact angle. The density of liquid was taken as 6.8 x10-3 kg/m3 and surface tension as
1.7288 Joule/m2. The computation results are given below in Table 6.3.

Table 6.3: The computation of penetration depth of carbonaceous particles into liquid
iron.

Fractional
Penetration
Particle Contact angles θ penetration (%)
Carbonaceous depth x (µm)
Diameter x/2rp . 100
Material
(µm)
I II I II I II

Char 1 25 113° 111° 7.7 8.1 30.8 32.4


Char 2 30 123° 110° 6.9 9.9 23.3 33.2
Char 3 100 115° 101° 29.1 40.7 29.1 40.7
Char 4 150 125° 102° 32.4 59.8 21.6 39.8
Coke 50 127° 125° 10.0 10.8 21.6 20.0

Some interesting trends are discernible from these numerical results. The wettability
played a very important role in determining the penetration depth of particles. A
reduction in contact angles led to a significant increase in the penetration depth of
150
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

particles. This result is to be expected as an improvement in wetting would result in a


lower upward force due to surface tension and therefore increased downward
penetration of particles in the liquid. While penetration depth x increased with particle
size, no well-defined trend was observed for fractional penetration.

It should be noted that the results presented above are expected to be valid only during
the initial stages of contact. As dissolution proceeds, a number of changes could take
place in the interfacial region. Sahajwalla and Khanna [87] have carried out Monte
Carlo simulations on carbon dissolution from graphite into liquid iron and have reported
on the morphology of a dissolving particle. It was observed in these simulations that the
liquid had a tendency to penetrate the particle and dissolve carbon. In this study, the
overall shape and size of the particle remained intact until they eventually disintegrated.
SEM images on reacted char particles collected from the liquid surface after the
carburiser cover experiment also show similar characteristics (a large number of pores)
(Figure 6.22).

151
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

Char 1 Char 2

Char 3 Char 4

Figure 6.22: SEM images of reacted char particles. A number of white deposits,
representing reaction products can clearly be seen.

The deposition of reaction products can also be seen as white material on the outer
surface of the char particles. It is noticed from these images that the contact area
between a particle and liquid could increase considerably with time with liquid
penetration in the pores. It would therefore be quite difficult to model/compute
solid/liquid contact area in the later stages of contact. In addition, the interfacial
blockage due to reaction products would also begin to play an important role and could
significantly affect contact area. Also the increased levels of sulphur due to sulphur
transfer from the chars/coke into the liquid may play a large role in limiting the surface
area available for interaction due to its surface-active nature.

152
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

6.4.2 Solid/Liquid Contact Area

Based on the data given Tables 6.2 and 6.3, total solid/liquid contact area was computed
for four chars and coke and is given below in Table 6.4. The corresponding particle
sizes, initial contact angles and penetration depths have also been shown. The liquid
bath surface area (the cross-sectional area of the crucible) was 95.05 cm2. Fractional
contact area (total contact area/bath surface area x 100) was also computed and is also
shown in Table 6.4.

Table 6.4: Computation of particle/liquid contact area

Fractional
Contact area
Particle Total
Carbonaceous Contact Penetration (%)
Diameter contact area
Material angle depth (µm) (contact
(µm) (cm2)
area/area of
crucible).100
Char 1 25 113° 7.7 62.6 65.86
Char 2 30 123° 6.9 51.96 54.67
Char 3 100 115° 29.1 63.09 66.37
Char 4 150 125° 32.4 41.23 43.38
Coke 50 127° 10.0 41.16 43.31

The results reported above once again show a wide variation. Char 1 and Char 3, with
the respective sizes of 25µm and 100µm, show a similar area of contact. Both these
particles however had a very similar contact angle with liquid iron. Both Char 2 and
Char 4, with the respective sizes of 30µm and 150µm and similar contact angles,
showed much smaller areas of contact. It needs to be noted that these areas correspond
to the initial stages of contact and should be used to compute first order dissolution rate
constants in the initial stages of contact only.

153
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

6.5 The Calculation of First Order Dissolution Rate Constants

Using the data given in Table 6.1 and Table 6.4, first order dissolution rate constants
were computed for chars and coke in contact with liquid iron (with a starting
composition of 2wt%C and 0.01wt%S) at 1550°C. The volume of liquid bath was taken
as 0.0001764 m3. This was done using Equation 6.7 below.

V
k' = K
A

Equation 6.7

This calculation removes the contribution of the geometric factors such as bath volume
V and bath/carbon contact area A, and produces a rate constant that is only a function of
the inherent dissociation rate of carbon and the contribution that any reactions may
make to the carbon dissolution rate. The results of this calculation have been given
below in Table 6.5.

Table 6.5: First order carbon dissolution rate constants for the carbonaceous materials
at 1550°C.

First order dissolution rate constant


Carbonaceous material
k’ x 103 (m.s-1)
Char 1 0.01795
Char 2 0.00278
Char 3 0.00606
Char 4 0.00942
Coke 0.02273

The observed trend in the dissolution rate constants was: Coke (50 µm) > Char 1 (25
µm) > Char 4 (150µm) > Char 3 (100 µm) > Char 2 (30µm). The particle sizes have
been given in parenthesis. No well-defined trend could be observed with particle size.
These results can be compared with dissolution rate constants quoted in the literature.

154
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

Experimental results from Orsten and Oeters [16] for cokes and from Mourao et al. [15]
for graphite are listed in Table 6.6.

Table 6.6: Carbon dissolution rate constants at 1600°C for a range of cokes, from [15,
16].

Dissolution rate constant


Carbonaceous material
k’ x 103 (m.s-1)
Coal Cliff coke [16] 0.0019
Blair Athol coke [16] 0.01
Great Greta coke [16] 0.01
Ruhrkohle coke [16] 0.12
Graphite [15] 0.184

While there is a wide variation in results for different coals, the overall range of results
reported in Table 6.5 compare well the results reported in Table 6.6 and are of a similar
magnitude.

In the following section, a discussion on factors affecting carbon dissolution from non-
graphitic materials has been carried out and an attempt will be made to identify the
dominant rate controlling mechanisms for the materials under investigation.

6.6 Factors affecting Carbon Dissolution from Non-Graphitic


Materials

A number of factors could affect the overall rate of carbon dissolution from chars/coke.
These include factors affecting the dissociation of carbon from the solid carbon (host
lattice), interfacial blockage due to reaction products, changes in melt composition, and
consumption of solute carbon by reducible oxides. These will be discussed in detail
below. Due to highly turbulent liquid bath in the induction furnace, the mass transfer
effects are not expected to dominate carbon dissolution in these experiments.

155
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

6.6.1 Atomic structure

All the carbonaceous materials under investigation are non-graphitic in nature with
small Lc values (Table 4.5) when compared with graphite that has Lc values of around
250 Å [44]. While Coke has the largest Lc value of 17.6Å, Char 4 had the smallest Lc
value of 9.5Å. Chars 1 to 3 had a similar Lc value of around 12.7Å. These Lc values are
representative of the degree of ordering and crystallite size seen in char materials [10].
Coal-chars are considered to have some carbon atoms arranged in small clusters
resembling the graphite structure [34]. The inter-layer spacing is fairly similar for all
carbonaceous materials ranging typically between 3-4 Å. The bonding of carbon atoms
in coals is highly complex compared to that in graphite. According to Hirsch [110], the
lamellae in coals containing up to 85% carbon are randomly oriented and are connected
by 3-D cross-links (see Figure 6.23).

Figure 6.23: Schematic representation of the various coal structures, from [110].

These cross-link bonds can be quite strong and a much higher energy will be required to
dissociate the carbon atoms from their complex network in coals than the energy that is
required for the carbon in graphite. The rate at which carbon atoms dissociate from the
host lattice would therefore depend on their degree of ordering. Due to its high Lc value,
it is to be expected that carbon atom dissociation rate will be highest for Coke. With
rather low and similar Lc values for chars, no definitive comments can be made on the
observed trends, except that carbon atom dissociation rates are expected to be slow for
chars.

As compared to both synthetic and natural graphite, the carbon dissolution rates
observed for the chars were found to be much slower. With the Lc value for chars
ranging from 9.5 Å to 12.7 Å, the corresponding Lc values for graphites are around 250
156
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

Å. It is therefore to be expected that the rate at which carbon atoms dissociate from the
host lattice in synthetic/natural graphite will be much faster as compared to the
corresponding rate in chars. The slow rates of carbon atom dissociation in chars will
therefore be an important factor in affecting carbon dissolution rates in chars. As
discussed in the literature review, the dissolution of carbon from a carbonaceous
material into a liquid iron melt can be considered to consist of two steps [34] shown
below:

Step 1:
C ( Lattice ) ⎯dissociati
⎯ ⎯⎯ on
→ C (int erface )

Equation 2.25: Dissociation of carbon atoms from its lattice site into the carbon/iron
interface, (reproduced from Chapter 2 section 2.5.3).

Step 2:
C (int erface ) ⎯transfer
⎯⎯→ C ( bulk )

Equation 2.26: Mass transfer of carbon atoms from the interface into bulk liquid
through the liquid boundary layer, (reproduced from Chapter 2 section 2.5.3).

6.6.2 Interfacial Phenomena

Due to the high ash content of all carbonaceous materials under investigation, the
deposition of reaction products in the interfacial region is expected to affect the kinetics
of carbon dissolution. Detailed investigations on the interfacial region, reported earlier
in Chapter 5, indicated that the interfaces were similar for both Coke and Chars with
similar type of reactions taking place. A number of reaction products were observed at
the interface, with interfacial coverage becoming almost complete with extended
periods of contact. A very interesting feature regarding the interfacial region, was the
detection of very small levels of silica, considering that silica was a principal
component of ash in the materials investigated. The role of silica reduction will be
discussed in detail in the next section.

157
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

The accumulation of alumina at the interface was detected for all materials. This is to be
expected, as the carbon is being dissolved into the liquid iron the remaining ash material
would be left behind at the interface. The amount of alumina at the interface was seen to
increase with time. Due to its non-wetting behaviour with liquid iron (θ ~120°), the
increased presence of alumina in the interfacial region is expected to significantly affect
the dynamic wettability of chars/coke with liquid iron. As shown in the model
calculations, an increase in contact angle resulted in a significant reduction in contact
area between the carbonaceous material and liquid iron, which would inturn lead to a
slow down in the kinetics of carbon dissolution. Even when the amount of alumina in
the parent ash is quite small, its concentration in the interfacial region makes it an
important factor affecting carbon dissolution.

Calcium and sulphur also appeared to preferentially accumulate at the interface,


concentrating at levels in excess of those expected from the ash composition alone.
However it was observed that calcium sulphide appeared to form much more rapidly
with the chars than for the coke and also had a more extensive coverage for chars. A
study by Wu et al [14] on interfacial reactions between natural graphite and liquid iron
monitored the deposition of reaction products in the interfacial region with time. While
the EDS results on the interface showed the presence of Al, Ca and Fe oxides in the
initial stages of contact, sulphur was not identified in the interfacial layer in the first
seven minutes of contact. After this time, a sulphur peak started to appear, probably due
to a desulphurising reaction between sulphur in liquid iron and calcium oxide in the
substrate transferring sulphur from the droplet into the interfacial layer.

CaO( s ) + S + C sat → CaS ( s ) + CO( g )

Equation 2.17 (reproduced from Chapter 2)

FESEM investigations on chars/coke also showed an increasing deposition of CaS in


the interfacial region with time. It is postulated that the two-stage behaviour observed in
carbon dissolution from Char 1 and Char 3 may be associated with the level of CaO
present in the carbonaceous material. The second stage seen in the two chars is
characterised by a large change in the rate after contact for longer times. The total CaO

158
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

present in Chars is given in Table 6.7 along with the corresponding dissolution rate
constants.

Table 6.7: Total CaO composition (wt%) for chars, and correlation with two-stage rate
behaviour

Carbonaceous Two stage rate


Total CaO k’ x 103 (m.s-1)
material behaviour
Char 2 0.2233 0.00278 No
Char 4 0.2687 0.00942 No
Char 1 0.3779 0.01795 Yes
Char 3 0.4174 0.00606 Yes

It is noted that the levels of CaO were much lower for both Char 2 and Char 4 as
compared to Char 1 and Char 3. Both Char 2 and Char 4 did not show a two stage
behaviour in carbon dissolution. The increased deposition of calcium based reaction
products with time probably resulted in enhanced interfacial blockage for Char 1 and 3,
leading to a significant reduction in carbon dissolution (stage II). Interfacial phenomena
are therefore expected to be a dominant rate controlling mechanism for carbon
dissolution from chars during stage II. Due to chemical reactions taking place in the
interfacial region and associated changes in local compositions, the role played by ash
fusion in interfacial blockage has not been taken into account.

6.6.3 Effect of Melt Sulphur

In addition to the build up of interfacial reaction products at the interface after longer
periods of contact between the carbonaceous materials and liquid iron, it is expected
that the amount of sulphur in liquid iron will also play an important role in the kinetics
of carbon dissolution. Previous studies of carbon dissolution have found a strong
dependence of dissolution rate with sulphur composition. Ericsson and Melberg [75]
believed that the reduction in carbon dissolution rate was due to the reduction in carbon
diffusivity with increasing sulphur concentrations. This view was supported by Orsten
and Oeters [77] who found a correlation with the reduction in rate with the reduction in
diffusivity when sulphur was added. Shigeno et al. [76] postulated that the reduction in
159
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

carbon dissolution rate was due to the surface active nature of sulphur, and that it was
selectively absorbing on the prismatic plane of graphite, thereby reducing sites for
carbon dissolution. Other researchers have reported on the surface active nature of
sulphur [93, 111] stating that high surface coverage can occur with even small amounts
of sulphur.

Not only does sulphur reduce carbon diffusivity in liquid iron, it also will reduce the
carbon saturation level. The carbon saturation point [88] can be taken as:

C s = 1.3 + 0.00257T - 0.31Si - 0.33P - 0.45S + 0.28Mn

Equation 2.24

From the above equation, we can see that sulphur has a strong influence on carbon
saturation levels in liquid iron. Atomistic simulation studies on liquid Fe-C-S system
[80] have also shown that the presence of sulphur has a tendency to displace C from its
immediate neighbourhood. Sulphur in liquid iron would therefore oppose a build-up of
local concentration of carbon in the liquid, resulting in a reduction in carbon dissolution.

In the current study on char dissolution, there was a marginal difference between the
sulphur transfers into liquid iron from the different chars. However there was a
significant amount of sulphur transfer into the liquid iron, which could result in an
increased resistance to carbon transfer with time.

The increasing levels of blockage at the interface from alumina and reaction products,
and the effect of sulphur in the melt depressing the carbon dissolution rate with
increasing sulphur levels, are most likely to be more significant with increasing contact
times. These multiple factors are therefore expected to play an important role in carbon
dissolution rate.

160
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

6.6.4 Reduction of Silica

The reduction reactions of silica, due to its high levels in ash, could have significant
implications for the dissolution of carbon. The reduction of silica can take place in-situ
by the solid carbon in the chars/coke or with solute carbon in liquid iron solution [18].

SiO2 + C → SiO( g ) + CO( g )

Equation 2.6: Partial reduction to form silicon monoxide vapour.

SiO( g ) + C → Si + CO( g )

Equation 2.8: Silicon transfer to carbon-containing iron.

As seen in Chapter 5, reduction reactions of silica can consume some of the solute
carbon in liquid iron. In this situation, part of the carbon dissolved from carbonaceous
material into liquid iron will be removed through silica reduction. The experimentally
measured rate of carbon dissolution would therefore be slower than the actual rate of
carbon dissociation into the liquid metal. With reduction reactions significantly altering
the measured carbon levels in liquid iron, it may not be possible to determine the
inherent rate constant for dissociation of carbon from carbonaceous materials. With
simultaneous carbon dissolution and consumption through reduction reactions generally
taking place in these materials, the experimentally determined carbon dissolution rate
constant into liquid iron therefore includes the contribution of other reactions occurring
such as silica reduction.

Total silica present in materials under study is given below in Table 6.8. It is to be noted
that Char 1, with the highest rate constant, also has a significantly lower silica
concentration.

161
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

Table 6.8: Comparison of apparent rate constant with silica content in the
carbonaceous materials.

Total SiO2 (wt%) in


Carbonaceous material k’ x 103 (m.s-1)
carbonaceous material
Char 2 5.1285 0.00278
Char 3 6.64547 0.00606
Char 4 5.85344 0.00942
Char 1 3.72448 0.01795
Coke 6.42816 0.02273

An approximate calculation was carried out to estimate the amount of solute carbon
removed through silica reduction. Assuming silica reduction takes place only with
solute carbon, a rough estimate of carbon removed through silica reduction is given in
Table 6.9. An exact calculation taking into account reduction of silica by solid carbon or
CO is beyond the scope of this study.

Table 6.9: Amount of carbon that could be removed due to silica reduction.

Amount of carbon that can be theoretically removed


Carbonaceous material
due to silica reduction (wt%)
Char 1 0.149
Char 2 0.205
Char 3 0.266
Char 4 0.234
Coke 0.257

Taking the rough estimates in Table 6.9 into consideration, the rate constant for Char 1
is enhanced from 0.01795 m/s to 0.01833 m/s. For Char 2, with the slowest rate of
carbon dissolution and a silica level comparable to other Chars (except Char 1), the rate
would be enhanced from 0.00278 m/s to 0.00948 m/s. Though difficult to exactly
quantify, reduction of silica is expected to significantly affect the measured rates of
carbon dissolution.

162
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

It can be seen in Equation 2.24 that the presence of silicon in the melt will also reduce
the carbon saturation level in liquid iron. While the reduction reactions imply that some
of the solute carbon is being consumed from the carbonaceous materials, the driving
force for carbon dissolution is also being reduced with increased silicon levels in the
metal, which will inturn slow down the rate of carbon dissolution.

6.7 Summary

A detailed investigation on carbon dissolution from non-graphitic carbonaceous


materials has clearly brought out the important role of interfacial reactions. With high
levels of melt turbulence in the induction furnace, mass transfer in the melt was not
expected to be rate controlling.

The effect of structure on carbon dissolution rates can be clearly seen in this study.
Slow rates of atomic dissociation from the host lattice are expected from these less
ordered forms of carbonaceous materials. The differences in dissolution rates between
chars and graphitic carbons are very large and could largely be due to differences in
their atomic structure. The structure of coke is more ordered as compared to chars and
this was reflected in its dissolution rates.

The consumption of solute carbon through silica reduction is also expected to be


significant in determining carbon dissolution rates. Char 1 with the lowest silica levels
among the chars was observed to have the highest carbon dissolution rate. Interfacial
reactions are also expected to have a significant effect on dissolution kinetics. During
the initial stages of contact (stage I), there is only partial blockage of interface by
reaction products. However an increase in alumina concentrations in the interfacial
region results in an increase in contact angle (poor wetting). Wettability was found to
play an important role. The interfacial area of contact was a function of contact angle,
decreasing significantly with increasing contact angle. A mixed control is expected to
dominate in stage I. During the later stages of contact (stage II), the interfacial effects
tend to slow down carbon dissolution considerably and are therefore expected to be a
dominant rate controlling mechanism. Sulphur is also expected to play an important part
in determining the carbon dissolution rate.

163
CHAPTER 6: Carbon Dissolution Studies: Results and Discussion

From this study we conclude that a number of factors such as carbon structure, silica
reduction, sulphur content in the liquid iron, and interfacial reaction products affect
carbon dissolution from chars into liquid iron and should be taken into account while
determining the rate controlling mechanism.

164
CHAPTER 7: Summary and Conclusions

CHAPTER 7: SUMMARY AND CONCLUSIONS

An in-depth investigation has been carried out in this project on the dissolution behaviour of
coal-chars into liquid iron. Focussing specifically on the behaviour of carbonaceous materials
during PCI, a range of experimental techniques was used to obtain detailed information on
reactions occurring during carbon dissolution from non-graphitic sources. Investigations were
carried out at 1550°C on four chars and one coke for a number of liquid iron compositions.
Using the sessile drop and carburiser cover methods, results were obtained for wettability,
interfacial phenomena and the kinetics of carbon dissolution. A number of factors were found
to play an important role in carbon dissolution from these non-graphitic carbonaceous
materials. These are discussed below in detail.

1. Silica, being one of the main ash components in coal-chars/coke under investigation,
can undergo in-situ reduction with solid carbon in the chars/coke and also with solute
carbon in liquid iron. Despite the high levels of silica in the ash initially, very little
silica was detected during the EDS analysis of the interfacial region, implying ongoing
silica reduction reactions. A small amount of silicon was however detected in the iron
droplets, indicating silica reduction with solute carbon. In sessile drop studies on
Coke/liquid iron, the droplets showed a marked delay in carbon transfer
(approximately 30 minutes), after which carbon transfer started to occur rather slowly.
Even after three hours of contact, carbon in the iron droplet did not reach saturation. It
was also noticed that the silicon in metal had levelled off after approximately thirty
minutes. It is therefore quite likely that the carbon accumulation was being affected
through the consumption of solute carbon in the droplet. Similar results were obtained
in sessile drop investigations on two coal-chars.

With the simultaneous carbon dissolution and consumption through reduction


reactions generally taking place in these materials, the inherent carbon dissolution rate
constant into liquid iron would be difficult to measure. As a part of the carbon
dissolved from carbonaceous material into liquid iron gets removed through silica
reduction, experimentally measured rates of carbon dissolution would therefore be
slower than the actual rates of carbon dissolution into the liquid metal. The reduction
of ash silica therefore has important implications for these non-graphitic materials,

165
CHAPTER 7: Summary and Conclusions

which have inherently slow carbon dissolution rates. This was one of the main
findings of this project.

2. Deposition of reaction products in the interfacial region also had a significant effect on
carbon dissolution. The accumulation of alumina at the interface was detected for all
materials. As the carbon dissolves into the liquid iron, some of the ash material gets
left behind at the interface. This is specifically true in case of alumina, as it does not
undergo reduction reactions like silica and is therefore not removed from the interface.
The amount of alumina at the interface was seen to increase with time in all cases.
Calcium and sulphur also appeared to preferentially accumulate at the interface,
concentrating at levels in excess of those expected from the ash composition alone.
However it was observed that calcium sulphide appeared to form much more rapidly
with the chars than for the coke and also had a more extensive coverage for chars.

3. Using the Carburiser cover method, the combined dissolution rate constants K were
measured from the slopes of ln((Cs-Ct)/(Cs-Co)) versus time plots. While Char 1, Char
3 and the Coke clearly showed two regions with distinct slopes, only one region was
observed for both Char 2 and Char 4. The rate constants for region I (5-40 minutes)
followed the trend: Coke > Char 1 > Char 3 & Char 4 > Char 2. These rates however
depend on the physical characteristics of each material and the interfacial area of
contact. The role of geometric factors has to be established before the results obtained
in this study could be compared with published data. As interfacial area of contact
could vary significantly for these materials, depending on the respective wettabilities,
a theoretical model was developed for estimating contact area for the Carburiser cover
technique.

4. Taking into account the geometry of the induction furnace, a theoretical model was
developed assuming an average particle size and random packing of carbon particles
in layers on top of liquid surface. Total number of layers and average number of
particles per layer were computed taking into account total weight, particle size and
containment volume. Using a force balance approach, the partial penetration of the
particles was calculated numerically and total solid/liquid contact area was estimated.
Wettability had a very significant effect on interfacial area of contact. A reduction in
contact angles led to a significant increase in the penetration depth of particle. An

166
CHAPTER 7: Summary and Conclusions

improvement in wetting reduces the upward force due to surface tension and therefore
increased downward penetration of particles in the liquid. While the penetration depth
increased with particle size, no well-defined trend was observed for fractional
penetration and total area of contact.

5. Using the Sessile drop method, the wettability of carbonaceous materials (four coal
chars/coke) with liquid iron at 1550°C was measured as a function of time.
Measurements were carried out with electrolytic iron and with an iron alloy containing
2 wt% C and 0.01 wt% S. With electrolytic iron, all materials showed non-wetting
behaviour with contact angles ranging between 106°-133° in the initial stages and
between 101° to 111° after an hour of contact. For the Fe-C-S alloy, the contact angles
ranged between113°-127° in the initial stages and between 101° to 125° after an hour
of contact. Being essentially non-wetting, only a marginal improvement in contact
angles was observed with time. There was an increased deposition of sulphur in the
interfacial region in the form of CaS with time. Due to the surface-active nature of
sulphur, an increase in sulphur concentration will tend to reduce the surface tension of
the liquid metal, resulting in an improved wettability.

6. Using the wettability data for chars/coke under investigation, interfacial area of
contact was estimated for the carburiser cover experiments. Computed values of
fractional contact area (the ratio of particle contact area to crucible cross-sectional
area) showed a wide variation. Char 1 and Char 3, with the respective sizes of 25µm
and 100µm, showed a similar area of contact (~66%). Both these particles however
had a very similar contact angle with liquid iron. Both Char 2 and Char 4, with the
respective sizes of 30µm and 150µm and similar contact angles, showed much smaller
areas of contact (43% and 55% respectively). These interfacial areas correspond to the
initial stages of contact and should be used to compute apparent dissolution rate
constants in the initial stages of contact. Estimated contact area in all cases was much
less than the crucible surface area. As some of the dissolution rate constants reported
in literature were calculated using crucible surface area as the contact area, these rate
constants are quite likely to be underestimated.

167
CHAPTER 7: Summary and Conclusions

7. The first order dissolution rate constants were computed using the combined
dissolution rate constants (evaluated from the slopes of ln((Cs-Ct)/(Cs-Co)) versus time
plots) and interfacial areas of contact. The observed trend in first order dissolution rate
constants (x103 m.s-1) was as follows: 0.02273 (Coke, 50 µm) > 0.01795 (Char 1, 25
µm) > 0.00942 (Char 4, 150µm) > 0.00606 (Char 3, 100 µm) > 0.00278 (Char 2,
30µm). The particle sizes have also been given in parenthesis. No well-defined trend
could be observed with particle size. The dissolution results compare well with the
dissolution rate constants quoted in the literature. In the results quoted above, Char 1
has the highest rate constant of the chars. Char 1 also has the lowest silica
concentration among all chars. Coke on the other hand has the highest dissolution rate
constant among the materials investigated in this study. While coke has higher silica
concentration as compared to Char 1, it also has a much higher Lc value. These results
therefore indicate that a high Lc value (more structural order) and low silica
concentrations are expected to lead to higher rates of carbon dissolution.

8. It is to be noted that dissolution rate for char 1 is higher than some of the published
rates for cokes and this could have important implications for PCI. In the blast furnace
both char and coke are able to carburise the liquid metal. If the rate of carbon
dissolution for the injected char is higher than that for the coke in the furnace, the
chars will dissolve faster into the liquid metal than the cokes. This may be desirable
for higher rates of PCI as the unburnt char will be consumed, and the coke will be
relatively less consumed.

9. Apart from the main results reported above, another interesting result was the
observed two-stage dissolution behaviour for Char 1, Char 3 and Coke. It is postulated
that this two-stage behaviour may be associated with two mechanisms. The first
mechanism is associated with the level of CaO present in the carbonaceous material. It
is to be noted that the levels of CaO were lower for both Char 2 and Char 4 as
compared to Char 1 and Char 3. Both Char 2 and Char 4 did not show the two-stage
behaviour in carbon dissolution. The increased deposition of calcium based reaction
products with time probably resulted in enhanced interfacial blockage for Char 1 and
3, leading to a significant reduction in carbon dissolution (stage II). The second is
associated with the changing composition of the liquid iron due to sulphur transfer

168
CHAPTER 7: Summary and Conclusions

from the carbonaceous materials. The increasing sulphur in the liquid iron reduces the
ability of the liquid iron to dissolve carbon into solution. The combination of these
two factors, both of which tend to be more important at longer periods of dissolution,
will alter the controlling mechanisms for carbon dissolution. These interfacial
phenomena are therefore expected to be a dominant rate controlling mechanism for
carbon dissolution from chars during stage II.

In conclusion, a detailed investigation on carbon dissolution from non-graphitic carbonaceous


materials has been carried out in this study and has produced important fundamental data on
wettability and dissolution rate constants for four chars and one coke. These in-depth studies
have also clearly brought out the important role of interfacial reactions. Along with the
consumption of solute carbon through silica reduction and slow rates of atomic dissociation,
the deposition of reaction products in the interfacial region also strongly inhibited carbon
dissolution in char. The observed dissolution rate behaviour could be well explained taking
into account these rate-controlling factors. With the main focus being on coal-chars, only one
coke was investigated in this study. Due to a large variation in commercial cokes in terms of
mineral matter, carbon content and structure, systematic studies reported in this project could
be extended further to investigate carbon dissolution in a variety of cokes. It will then allow a
more in-depth understanding of the competing carburising behaviour of chars and cokes
during PCI.

169
CHAPTER 8: References

CHAPTER 8: REFERENCES

1. I. Hidalgo, S. Laszlo, J. C. Ciscar and A. Soria, "Technological Prospects and


CO2 Emission trading Analyses in the Iron and Steel Industry: A Global Model",
Energy, Article in Press (2004).

2. H. W. Gudenau, D. Senk, K. Fukada, A. I. Babich and C. F. Froehling, "Coke


Behaviour in the Lower Part of the Blast Furnace with High Injection Rate".
Presented at International Blast Furnace Lower Zone Symposium, Wollongong,
Australia, 25-27 November, 2002. 2002. 11:1-11:12.

3. R. J. Nightingale, S. J. Chew, P. R. Austin and J. G. Mathieson, "Assessment of


Blast Furnace Deadman Condition". Presented at International Blast Furnace
Lower Zone Symposium, Wollongong, Australia, 25-27 November, 2002. 2002.
20:1-20:12.

4. R. J. Nightingale, R. J. Dippenaar and W.-K. Lu, "Developments in Blast


Furnace Process Control at Port Kembla Based on Process Fundamentals",
Metall. Mater. Trans. B, 31B [No. 5] 993-1003 (2000).

5. H. W. Gudenau, J. P. Mulanza and D. G. R. Sharma, "Carburisation of Hot


Metal by Industrial and Special Cokes", Steel Res., 61 [3] 97-104 (1990).

6. H. W. Gudenau, L. Meier and V. Schemmann, "Coke Quality Requirements for


Modern Blast Furnace Operation". Presented at 2nd International Congress on
the Science and Technology of Ironmaking and 57th Ironmaking Conference,
Toronto, Canada, 22-25 Mar. 1998. 1067-1072.

7. F. Didelon, P. Lacroix, J. M. Libralesso and D. Sert, "Conditions for Achieving


Very Low Coke Rate". Presented at 2nd International Congress on the Science
and Technology of Ironmaking and 57th Ironmaking Conference, Toronto,
Canada, 22-25 Mar. 1998. 411-416.

8. Y. Iwanaga, "Behaviour of Unburnt Pulverised Coal in the Blast Furnace",


Ironmaking Steelmaking, 19 [1] 36-44 (1992).

9. C. Yamagata, S. Suyama, S. Y. U. Horisaka, K. Takatani, Y. Kajiwara, S.


Komatsu, H. Shibuta and Y. Aminaga, "Fundamental Study on Combustion of
Pulverised Coal Char Injected into Coke Bed at High Rate", Iron and Steel
Institute of Japan International, 32 [6] 725-732 (1992).

10. L. Lu, "Coal/Char Structure and its Influence on Char/Gas Reactions During
Pulverised Coal Injection in a Blast Furnace", Ph.D. The University of New
South Wales, 2000.

11. A. S. Mehta, V. Sahajwalla, J. J. Poveromo and T. F. Wall, "Fundamental


Investigation of Slag/Carbon/Gas Interactions Occurring in a Blast Furnace
During Pulverised Coal Injection". Presented at 2nd International Congress on
the Science and Technology of Ironmaking and 57th Ironmaking Conference,
Toronto, Canada, 22-25 Mar. 1998. 1459-1470.

170
CHAPTER 8: References

12. A. S. Mehta, "Wettability and Reactions Occurring at the Slag/Carbon Interface


during Pulverised Coal Injection in a Blast Furnace", Ph.D. School of Materials
Science and Engineering, The University of New South Wales, Sydney,
Australia, 2000.

13. R. J. Haywood, J. S. Truelove and M. J. Mccarthy, "Modelling of Pulverised


Coal Injection and Combustion in Blast Furnaces". Presented at 53rd
Ironmaking Conference, Chicago, Illinois, USA, 20-23 Mar. 1994. 437-442.

14. C. Wu, R. Wiblen and V. Sahajwalla, "Influence of Ash on Mass Transfer and
Interfacial Reaction between Natural Graphite and Liquid Iron", Metall. Mater.
Trans. B, 31B [5] 1099-1104 (2000).

15. M. B. Mourao, G. G. Krishna Murthy and J. F. Elliott, "Experimental


Investigation of Dissolution Rates of Carbonaceous Materials in Liquid Iron-
Carbon Melts", Metall. Trans. B, 24B 629-637 (1993).

16. S. Orsten and F. Oeters, "Behaviour of Coal Particles Blown into Liquid iron".
Presented at W.O. Philbrook Memorial Symposium, Toronto, Ontario, Canada,
17-20 Apr. 1988. 27-38.

17. M. Peters, H. W. Gudenau, M. Scheiwe and R. Sieger, "Stresses Exerted on


Coke in the Lower Region of a Blast Furnace". Presented at 2nd European
Ironmaking Conference, Glasgow, Scotland, UK, 15-18 Sept. 1991. 205-223.

18. Blast Furnace Phenomena and Modelling. Edited by the Committee on


Reactions within the Blast Furnace: Joint Society on Iron and Steel Basic
Research: The Iron and Steel Institute of Japan, Elsevier Applied Science
Publishers, Essex, England, 1987.

19. J. Agarwal, F. Brown, D. Chin, G. Stevens, R. Clark and D. Smith, "The Future
Supply of Coke", New Steel, 12 88 (1996).

20. R. J. Fruehan, "Future Ironmaking in North America". Presented at 2nd


International Congress on the Science and Technology of Ironmaking and 57th
Ironmaking Conference, Toronto,Canada, 22-25 Mar. 1998. 59-66.

21. J.-M. Steiler, "State of the Art of Blast Furnace Ironmaking". Presented at 2nd
International Congress on the Science and Technology of Ironmaking and 57th
Ironmaking Conference, Toronto, Canada, 22-25 Mar. 1998. 161-173.

22. N. Ponghis, "Reduction of Blast Furnace Coke Rate: Coal Injection". Presented
at 2nd International Congress on the Science and Technology of Ironmaking and
57th Ironmaking Conference, Toronto, Canada, 22-25 Mar. 1998. 1998. 365-
372.

23. G. Tovarovskii, V. N. Khomich and G. P. Boyarovskaya, "Effectiveness of


Using Blast Additions in Blast Furnace Process", Steel in the USSR, 10 [11] 573-
579 (1980).

24. J. Agarwal, F. Brown, D. Chin, G. Stevens and D. Smith, "Injecting Coal and
Natural Gas: Which one? How much?" New Steel, 12 70-72 (1996).
171
CHAPTER 8: References

25. "Coal Injection Project Under Way at Burns Harbor", Iron and Steel Eng., 71
122 (1994).

26. P. Bennett and D. Holcombe, "Report Prepared for Australaian Coal Research
Limited", 1994.

27. J.-M. Steiler and F. Hanrot, "Present State and Innovative Issues for
Ironmaking". Presented at Science and Technology of Innovative Ironmaking
for Aiming at Energy Half Consumption, Tokyo, Japan, November 27-28. 2003.
229-236.

28. R. J. Kruse, P. Chaubal, J. B. Moore and H. S. Valia, "Blast Furnace PCI Coal
Selection at Ispat Inland". Presented at Iron & Steel Society International
Technology Conference and Exposition 2003, Indianapolis, USA, 27-30 Apr.
2003. 787-796.

29. L. Lu, V. Sahajwalla, C. Kong, D. Harris and A. McLean, "Novel Equations to


Establish Influence of Char Structure on Reactivity during PCI in a Blast
Furnace". Presented at 61st Ironmaking Conference, Nashville, TN, USA, 10-13
Mar. 2002. 277-292.

30. K. Yamaguchi, H. Ueno and K. Tamura, "Maximum Injection Rate of


Pulverised Coal into Blast Furnace through Tuyeres with Consideration of
Unburnt Char", Iron and Steel Institute of Japan International, 32 [6] 716-724
(1992).

31. H. Ueno, K. Yamaguchi and K. Tamura, "Coal Combustion in the Raceway and
Tuyere of a Blast Furnace", Iron and Steel Institute of Japan International, 33
[6] 640-645 (1993).

32. J. B. Hyde, W. R. Brown and S. W. Hutchinson, "Start-up of Stelco Blast


Furnace Pulverised Coal Injection Facility", Iron and Steel Eng., 75 34-43
(1998).

33. W. Koen, H. L. Toxopeus, E. E. Schoone, C. V. D. Vliet and H. V. D. Berg,


"The Development of the High PCI Technology at Hoogovens Ijmuiden".
Presented at 53rd Ironmaking Conference, Chicago, Illinois, USA, 20-23 Mar.
1994. 429-436.

34. V. Sahajwalla and R. Khanna, "Fundamental Investigation of Basic Mechanisms


of Carbon Dissolution in Molten Iron". Presented at Yazawa International
Symposium: Metallurgical and Materials Processing: Principles and
Technologies, San Diego, USA, 2-5 March. 2003. 825 - 840.

35. E. Beppler, B. Gerstenberg, U. Janhsen and M. Peters, "Requirements on the


Coke Properties, Especially when Injecting High Coal Rates". Presented at 51st
Ironmaking Conference, Toronto, Ontario, Canada, 5-8 Apr. 1992. 171-184.

36. R. J. Nightingale, R. J. Dippenaar and W.-K. Lu, "Developments in Blast


Furnace Process Control at Port Kembla Based on Process Fundamentals".
Presented at The Belton Memorial Symposium, Sydney, Australia, 2000. 227-
240.
172
CHAPTER 8: References

37. G. Fawcett, Coal Terminology and Technology. Government Printer, NSW,


Sydney, 1980.

38. S. J. Todd and H. Hansen, "Nature and Origin of Coal and its Petrographic
Characteristics". Presented at 53rd Electric Furnace Conference, Orlando, USA.,
1995. 3-13.

39. J. Gibson and D. H. Gregory, Carbonisation of Coal. Edited by J. Gordon Cook,


Mills & Boon Limited, London, 1971.

40. P. C. Pistorius, "Reductant Selection in Ferro-Alloy Production: The Case for


the Importance of Dissolution in the Metal", Journal of the South African
Institute of Mining and Metallurgy, 102 [1] 33-36 (2002).

41. F. G. Emmerich and C. A. Luengo, "Babassu Charcoal: A Sulfurless Renewable


Thermo-Reducing Feedstock for Steelmaking", Biomass and Bioenergy, 10 [1]
41-44 (1996).

42. I. Areklett and L. P. Nygaard, "The Ferroalloys Industry: Back to Charcoal?"


Greenhouse Issues, 60 (2002).

43. R. S. Sampaio, M. E. Resende, L. P. Almeida and R. R. Junqueira,


"Environmentally and Metallurgically Clean Ferroalloy Producing System".
Presented at 55th Electric Furnace Conference, Chicago, IL. USA, 9-12 Nov.
1997. 281-287.

44. V. Sahajwalla, M. Dubikova and R. Khanna, "Reductant Characterisation and


Selection - Implications for Ferroalloys Processing". Presented at Tenth
International Ferro Alloys Congress, Johannesburg, South Africa, 1-4 February.
2004.

45. K. Farrell, L. Lu, V. Sahajwalla and D. Harris, "Atomic Scale Structural


Features of Some Coals and Chars". Presented at 8th Australian Coal Science
Conference, Sydney, Australia, 1998.

46. W. F. Sheehan, Physical Chemistry. Allyn and Bacon, Inc., Boston, 1961.

47. V. Sahajwalla, K. Gao, C. Wu, J. Xia and M. Burdett, "XRD and TEM Studies
of Carbon Structure to Understand its Dissolution Performance during
ironmaking". Presented at The AusIMM Annual Conference, Perth, Australia,
24-28 March. 1996. 207-212.

48. B. D. Cullity, Elements of X-ray Diffraction. Addison-Wesley Publishing Co.,


Reading, 1978.

49. R. Diamond, "A Least Squares Analysis of the Diffuse X-ray Scattering from
Carbons", Acta Cryst., 11 129-138 (1958).

50. L. Lu, V. Sahajwalla, C. Kong and D. Harris, "Quantitative X-ray Diffraction


Analysis and its Application to Various Coals", Carbon, 39 [12] 1821-1833
(2001).

173
CHAPTER 8: References

51. D. Alvarez, A. G. Borrego and R. Mendez, "Unbiased Methods for the


Morphological Description of Char Structures", Fuel, 76 [13] 1241-1248 (1997).

52. R. E. Franklin, "The Structure of Graphitic Carbons", Acta Cryst., 4 253-261


(1951).

53. A. Oberlin, "Carbonization and Graphitization", Carbon, 22 [6] 521-541 (1984).

54. S. K. Gupta, V. Sahajwalla, Y. Al-Omari, N. Saha-Chaudhury, F. Rorick, G.


Hegedus, P. Chaubal, J. Burgo, M. Best and F. Hyle, "Atomic Structure of Coke
Fines in Blast-Furnace Dust and their Origin in Operating Blast Furnaces."
Presented at Iron & Steel Society International Technology Conference and
Exposition 2003;, Indianapolis, USA, 27-30 Apr. 2003. 841-856.

55. D. W. Van Krevelen, Coal: Typology-Physics-Chemistry-Constitution. Edited


by Elsevier, Amsterdam, 1993.

56. A. Agarwal and U. Pal, "Influence of Pellet Composition and Structure on


Carbothermic Reduction of Silica", Metall. Mater. Trans. B, 30B [2] 295-306
(1999).

57. A. Chatterjee, B. N. Singh and S. N. Asthana, "Mechanism of Silicon Transfer


to Hot Metal and Its Application to Indian Blast Furnaces". Presented at
International Symposium on Blast Furnace Ironmaking, Jamshedpur; India., 15-
16 Nov. 1985. 1985. 265-274.

58. D. H. Filsinger and D. B. Bourrie, "Silica to Silicon: Key Carbothermal


Reactions and Kinetics", J. Am. Ceram. Soc., 73 [6] 1726-1732 (1990).

59. N. J. Themelis, Transport and Chemical Rate Phenomena. Edited by Gordon and
Breach Science Publishers, Switzerland, 1995.

60. E. Kapilashrami, A. K. Lahiri and S. Seetharaman, "Corrosion of Dense


Alumina by Molten Iron Containing Oxygen", Steel Grips 1, [2] 126-132
(2003).

61. E. Kapilashrami, A. Jakobsson, A. K. Lahiri and S. Seetharaman, "Studies of the


Wetting Characteristics of Liquid Iron on Dense Alumina by the X-Ray Sessile
Drop Technique", Metall. Mater. Trans. B, 34B [April 2003] 193-199 (2003).

62. E. Kapilashrami, "Investigation of Interactions Between Liquid Iron Containing


Oxygen and Aluminosilicate Refractories", Ph.D. Royal Institute of Technology,
2003.

63. R. J. Fruehan, "The Rate of Reduction of Iron Oxides by Carbon", Metall.


Trans. B, 8B [June] 279-286 (1977).

64. A. Muan, "Phase Equilibria at High Temperatures in Oxide Systems involving


Changes in Oxidation States", Amer. J. Sci., 256 171-207 (1958).

65. D.-J. Min and R. J. Fruehan, "Rate of Reduction of FeO in Slag by Fe-C Drops",
Metall. Mater. Trans. B, 23B [February, 1992] 29-37 (1992).
174
CHAPTER 8: References

66. G. R. Belton and R. J. Fruehan, "Rate Phenomena in Iron Bath Smelting",


Proceedings of the Ethem T. Turkdogan Symposium- Fundamentals and
Analysis of New and Emerging Steelmaking Technologies., 3-22 (1994).

67. R. Inoue and H. Suito, "Calcium Desulfurisation Equilibrium in Liquid Iron",


Steel Res., 65 [10] 403-409 (1994).

68. M. A. T. Andersson, L. T. I. Jonsson and P. G. Jonsson, "A Thermodynamic and


Kinetic Model of Reoxidation and Desulphurisation in the Ladle Furnace", Iron
and Steel Institute of Japan International, 40 [11] 1080-1088 (2000).

69. A. K. Biswas, Principles of Blast Furnace Ironmaking: Theory and Practice.


Cootha Publishing House, Brisbane, 1981.

70. G. A. Irons and C. Celik, "Comparison of Iron Desulphurisation using Calcium


Carbide and Lime-Magnesium Mixtures", Ironmaking Steelmaking, 19 [2] 136-
144 (1992).

71. V. O. Dahlke and O. Knacke, "Die Auflösung von Kohlenstoff in flüssigem


Eisen (The Dissolution of Carbon in Liquid Iron)", Arch. Eisenhuttenwes., 26
373-378 (1955).

72. R. G. Olsson, V. Koump and T. F. Perzak, "Rate of Dissolution of Carbon in


Molten Fe-C Alloys", Trans.TMS-AIME, 236 426-429 (1966).

73. M. Kosaka and S. Minowa, "On the Rate of Dissolution of Carbon into Molten
Fe-C Alloy", Trans. Iron Steel Inst. Jpn., 8 [6] 392-400 (1968).

74. V. A. Grigoryan and V. P. Karshin, "Influence of Surfactants on the Dissolution


Kinetics of Graphite in Liquid Iron", Russian Metallurgy, 1 57-59 (1972).

75. S. O. Ericsson and P. O. Melberg, "Influence of Sulphur on the Rate of Carbon


Dissolution in Liquid Iron", Scand. J. Metall., 10 15-18 (1981).

76. Y. Shigeno, M. Tokuda and M. Ohtani, "The Dissolution Rate of Graphite into
Fe-C Melt Containing Sulphur of Phosphorus", Transactions of the Japan
Institute of Metals, 26 [1] 33-43 (1985).

77. S. Orsten and F. Oeters, "Dissolution of Carbon in Liquid Iron". Presented at


Process Technology Proceedings, Washington, D.C., USA, 6-9 Apr. 1986. 143-
155.

78. J. K. Wright and B. R. Baldock, "Dissolution Kinetics of Particulate Graphite


Injected into Iron/Carbon Melts", Metall. Trans. B, 19B 375-381 (1988).

79. V. Sahajwalla, I. F. Taylor, J. K. Wright and G. J. Hardie, "Dissolution of


Carbon into Iron Melts: The New Direct Ironmaking Perspective", pp. 715-730
in Metallurgical Processes for the Early Twenty First Century, Edited by H. Y.
Sohn. The Minerals, Metals, & Materials Society, 1994.

175
CHAPTER 8: References

80. V. Sahajwalla and R. Khanna, "Effect of Sulfur on the Dissolution Behaviour of


Graphite in Fe-C-S melts: A Monte Carlo Simulation Study", Scand. J. Metall.,
32 [1] 53-57 (2002).

81. J. K. Wright and W. T. Denholm, "Dissolution of Particulate Carbon in a


Turbulent Iron Bath". Presented at The AusIMM Melbourne Branch,
Symposium on "Extractive Metallurgy", November. 1984. 323-329.

82. A. Ganguly and K. J. Reid, "Carbon Utilisation in Direct Smelting Systems",


Steel Res., 63 [7] 281-290 (1992).

83. J. K. Wright and I. F. Taylor, "Multiparticle Dissolution Kinetics of Carbon in


Iron-Carbon-Sulphur Melts", ISIJ Int., 33 [5] 529-538 (1993).

84. V. Sahajwalla, I. F. Taylor and J. K. Wright, "Rates of Carbon Transfer from


Sources Injected into Fe-C-S Melts". Presented at 52nd Ironmaking Conference,
1993. 355-365.

85. V. Sahajwalla, B. Waugh, J. M. Langley, K. Farrell, C. Wu and K. Gao,


"Influence of Atomic Structure of Carbonaceous Material on its Dissolution into
Fe-C-S Melts", The AusIMM Proceedings, 301 [1] 17-21 (1996).

86. H. Sun, K. Mori, V. Sahajwalla and R. D. Pehlke, "Carbon solution in liquid


iron and iron alloys", High Temperature Materials and Processes, 17 [4] 257-
270 (1998).

87. V. Sahajwalla and R. Khanna, "Monte Carlo Simulation Study of Dissolution of


Graphite in Iron-Carbon Melts", Metall. Mater. Trans. B, 31 [6] 1517-1525
(2000).

88. F. Neumann, H. Scheneck and W. Patterson, Giesserei, 47 25-95 (1960).

89. T. Young, Trans. R. Soc. Lon., 94 65 (1805).

90. J. V. Naidich, "Wettability of Solids by Liquid Metals", Prog. Surf. Mem. Sci.,
14 353-484 (1981).

91. B. J. Keene, "Contact Angle and Work of Adhesion between Ferrous Melts and
Non-Metallic Solids", pp. 513-539 in Slag Atlas, Edited by V. D. E. (VDEh).
Verlag Stahleisen GmbH, Düsseldorf, 1995.

92. L. Jimbo, A. Sharan and A. W. Cramb, "Recent Measurments of the Surface and
Interfacial Tensions in Steels". Presented at 76th Steelmaking Conference,
Dallas, Texas, USA, 28-31 Mar. 1993. 485-494.

93. D. R. Sain and G. R. Belton, "The Influence of Sulphur on Interfacial Reaction


Kinetics in the Decarburisation of Liquid Iron by Carbon Dioxide", Metall.
Trans. B, 9B 403-407 (1978).

94. K. Mukai, "Wetting and the Marangoni Effect in Iron and Steelmaking
Processes", Iron and Steel Institute of Japan International, 32 [1] 19-25 (1992).

176
CHAPTER 8: References

95. I. A. Aksay, C. E. Hoge and J. A. Pask, "Wetting under Chemical Equilibrium


and Non-Equilibrium Conditions", J. Phys. Chem, 78 [12] 1178-1183 (1974).

96. B. J. Keene, "Review of Data for the Surface Tension of Iron and its Binary
Alloys", Int. Mat. Rev., 33 [1] 1-37 (1988).

97. F. D. Richardson, "Interfacial Phenomena in Metallurgical reactions", Trans.


Iron Steel Inst. Jpn., 14 [Special Lecture] 1-8 (1974).

98. C. Wu and V. Sahajwalla, "Influence of Melt Carbon and Sulfur on the Wetting
of Solid Graphite by Fe-C-S Melts", Metall. Trans. B, 29B 471-477 (1998).

99. Wettability. Edited by J. C. Berg, Marcel Dekker, Inc., New York, 1993.

100. B. J. J. Zelinski, J. P. Cronin, M. Denesuk and D. R. Uhlmann, "High


Temperature Wetting Behaviour of Inorganic Liquids", pp. 465-524 in
Wettability, Edited by J. C. Berg. Marcel Dekker Inc., New York, 1993.

101. T. D. Blake, "Dynamic Contact Angles and Wetting Kinetics", pp. 252-309 in
Wettability, Edited by J. C. Berg. Marcel Dekker, Inc, New York, 1993.

102. B. McDonald, C. Wu, V. Sahajwalla, K. Farrell and T. Wall, "Dissolution


Studies of Carbonaceous Materials in Blast Furnace Hot Metal During
Pulverised Coal Injection". Presented at 2nd International Congress on the
Science and Technology of Ironmaking and 57th Ironmaking Conference,
Toronto, Canada, 22-25 Mar. 1998.

103. C. Wu, "Fundamental Investigations of Molten Iron/Carbon Interactions in


Direct Ironmaking", Ph.D. The University of New South Wales, 1998.

104. L. Zhao and V. Sahajwalla, "Interfacial Phenomena During Wetting of


Graphite/Alumina Mixtures by Liquid Iron", ISIJ Int., 43 [1] 1-6 (2003).

105. J. Moriss and D. Buehl, Journal of Metals, 188 317 (1950).

106. V. Sahajwalla and R. Khanna, "Influence of Sulphur on the Solubility of


Graphite in Fe-C-S Melts: Optimization of Interaction Parameters", Acta Mater.,
50 663-671 (2002).

107. C. Wu and V. Sahajwalla, "Dissolution Rates of Coals and Graphite in Fe-C-S


Melts in Direct Ironmaking: Influence of Melt Carbon and Sulphur on Carbon
Dissolution", Metall. Mater. Trans. B, 31B [2] 243-251 (2000).

108. R. I. L. Guthrie, Engineering in Process Metallurgy. Clarendon Press, Oxford,


UK., 1992.

109. S. K. Khanna, C. Wu and V. Sahajwalla, "Estimation of Critical Carbon Particle


Engulfment Velocity in Metallurgical Melts", High Temperature Materials and
Processes, 17 [3] 193-201 (1998).

110. P. B. Hirsch, Electron Microscopy of Thin Crystals. Butterworths, Washington,


1965.
177
CHAPTER 8: References

111. R. J. Fruehan, "The Effect of Sulphur and of Phosphorus on the Rate of


Decarburization of Solid Iron in Hydrogen", Metall. Trans. B, 3 1447-1453
(1972).

178
List of Publications

LIST OF PUBLICATIONS

Refereed Journal Papers

1. F. McCarthy, V. Sahajwalla, J. Hart and N. Saha-Chaudhury, "Influence of Ash


on Interfacial Reactions between Coke and Liquid Iron", Metall. Mater. Trans. B, 34B
[5] 573-580 (2003).

Refereed Conference or Symposium Publications

2. F. McCarthy, V. Sahajwalla, J. Hart and N. Saha-Chaudhury, "Coke/Iron


Interfacial Phenomena". Presented at The Mills Symposium, Metals, Slags, Glasses:
High Temperature Properties and Phenomena, London, U.K. 2002. 397-406.

3. F. McCarthy, V. Sahajwalla and J. Hart, "Metal/Char Interactions During


Pulverised Coal Injection in a Blast Furnace". Presented at 61st Ironmaking Conference,
Nashville, TN, USA, 10-13 Mar. 2002, 313-324.

4. V. Sahajwalla, F. McCarthy, S. T. Cham, J. Hart, R. Sakurovs and N. Saha-


Chaudhury, "Influence of Carbonaceous Materials on Carburisation of Liquid Iron".
Presented at International Blast Furnace Lower Zone Symposium, Wollongong,
Australia, 2002.

5. V. Sahajwalla, A. S. Mehta, F. McCarthy, T. Wall and J. Hart, "Pulverised Coal


Injection: Slag/char and Metal/char Interactions". Presented at Steel Technology
International 2000, 17-20.

6. F. McCarthy, C. Wu, V. Sahajwalla and J. Hart, "Dissolution Studies of Black


Coal Char in Blast Furnace Hot Metal During Pulverized Coal Injection". Presented at
58th Ironmaking Conference, Chicago, IL, USA, 21-24 Mar. 1999, 21-24 Mar. 1999.
605-612.

179
Appendix A

APPENDIX A: SAMPLE CALCULATION OF THE OFF-GAS COMPOSITION


AND CALCULATION OF THE NUMBER OF MOLES OF OXYGEN REMOVED
FROM COKE/IRON SYSTEM

The mass spectrometer was calibrated to analyse for methane, water, argon, carbon
monoxide, carbon dioxide, hydrogen, oxygen, hydrogen sulphide, and sulphur dioxide.
The analysed amounts of argon (purge gas/inert atmosphere), carbon monoxide, and
carbon dioxide were considered as the other constituents were measured at very low
amounts. By utilising the knowledge that one mole of a standard gas occupies 24465
cm3 litres at 25°C and at 1 atmosphere and that the input of argon into the furnace was
0.492 L/min, the output of the mass spectrometer, which is in weight percent of the total
input stream, can be converted to number of moles for each of the gases of interest, CO
and CO2.

The two sample runs consisted of a coke substrate without an iron droplet (or blank) and
a coke substrate with the iron droplet. From this the difference the contribution of iron
in reduction reactions can be observed.

The number of moles of carbon monoxide and carbon dioxide was converted to number
of moles of oxygen, with one mole of carbon monoxide converted to one mole of
oxygen and one mole of carbon dioxide converted to two moles of oxygen (as O). This
calculation was converted to O2 for the comparison shown in section 5.2.1. A sample of
the calculation, which was completed in an excel spreadsheet, is shown in Table 1.A on
the next page.

180
Table 1.A: Sample calculation for determination of oxygen removal from mass spectrometer calculations.

Appendix A
L/min
Volume Volume of Volume Volume of Volume Volume Volume
Time CO2 Argon (based on
CO wt% of CO CO (L) of CO2 CO2 (L) O2 (from O2 (from of O2
(sec) wt% wt% input and
(L) cumulative (L) cumulative CO) CO2) removed
Ar)
0 0.6130 2.3263 96.4220 0.5103 0.0020 0.0020 0.0005 0.0005 0.0010 0.0005 0.0015
10 0.6194 2.3197 96.4171 0.5103 0.0020 0.0040 0.0005 0.0010 0.0020 0.0010 0.0030
20 0.6258 2.3116 96.4299 0.5102 0.0020 0.0059 0.0005 0.0016 0.0030 0.0016 0.0045
30 0.6212 2.2793 96.4647 0.5100 0.0019 0.0079 0.0005 0.0021 0.0039 0.0021 0.0060
40 0.6135 2.2394 96.5192 0.5097 0.0019 0.0098 0.0005 0.0026 0.0049 0.0026 0.0075
50 0.6042 2.1895 96.5951 0.5093 0.0019 0.0116 0.0005 0.0031 0.0058 0.0031 0.0090
60 0.5895 2.1419 96.6598 0.5090 0.0018 0.0134 0.0005 0.0036 0.0067 0.0036 0.0104
70 0.5810 2.0926 96.7317 0.5086 0.0018 0.0152 0.0005 0.0041 0.0076 0.0041 0.0117
80 0.5667 2.0271 96.8123 0.5082 0.0017 0.0169 0.0005 0.0046 0.0085 0.0046 0.0131
90 0.5521 1.9695 96.9081 0.5077 0.0017 0.0186 0.0005 0.0051 0.0093 0.0051 0.0144
100 0.5353 1.9062 96.9923 0.5073 0.0018 0.0204 0.0005 0.0056 0.0102 0.0056 0.0158

181
Appendix B

APPENDIX B: SAMPLE CALCULATION OF CONTACT AREA

The total number of particles (N) in the reaction cell was computed from the total
weight (W) of the carbonaceous material and mass of an individual particle (mp),
expressed in Equation 6.4.

W W
N= =
mp 4 3
πr p ρ p
3

Equation 6.4

Taking rp = 12.5 µm and ρp = 1.78 g/cm3, and the weight, W = 120 g, we calculate the
number of particles N, to be = 82.2x108. If the sample height is 2 cm, then the total
number of layers is equal to this divided by the particle diameter, giving the total
number of layers equal to 800. The average number of particles per layer is found by
dividing the total number of particles by the number of layers, resulting in the number
of particles per layer being 10.35x106. The fractional packing was determined by
calculating the area occupied by the particles and dividing this by the total area
available (cross-sectional area of the crucible), and this was determined to be 53.5% of
the available area contains particles.

The penetration of the particle can be calculated by using the penetrated area of contact
as being equal to 2πrpx where x is the penetration depth of particle. When a spherical
particle comes in contact with a liquid surface and tries to penetrate it, the total force
acting on the particle in a direction normal to the liquid surface can be written as:

Ftotal = FG + FS + FB

Equation 6.5

In this equation, FG is the gravitational force acting vertically downwards and the layers
of carbonaceous particles above the particle in contact with liquid will also contribute to
the downward gravitational force. Total gravitational force is assumed to be equal to the

182
Appendix B

mass of particle (mp) times the number of layers (nl) above it. The force due to surface
tension is FS. The buoyancy force FB acts on the immersed part of the particle and
depends on the penetration depth x. Under equilibrium conditions, the net force on the
particle is zero. The downward gravitational force balances the upward forces due to
surface tension and buoyancy. The force balance equation can be written as:

4 3 1 x
πr p ρ p gn l = (3r p x 2 − x 3 ) ρ l g + 2πr p γ lv (1 − + cos θ )
3 3 rp

Equation 6.6

In this equation, ρl is the liquid density, γlv the liquid surface tension and θ the contact
angle. Equation 6.6 was solved numerically and the penetration depth of particles and
interfacial contact area was computed for all chars/coke.

183

You might also like