ISSN 1018-5593

European Commission

technical coal research

Coal preparation

Identification of coal characteristics indicating problems of swelling

during carbonization

European Commission

é--r i'-.:.. :<■■■■:■. \ : - ¿ ' ,~' Æ* " . , ^Λ-

Coal preparation

Identification of coal characteristics indicating problems of swelling

during carbonization

I. Edwards, K. Thomas, F. Nadji, I. Butterfield

University of Newcastle upon Tyne Newcastle upon Tyne NE1 7RU

United Kingdom

Contract No 7220-EB/839

1 November 1990 to 31 October 1994

Final report

Directorate-General XVII Energy

1996 EUR 17150 EN

A great deal of additional information on the European Union is available on the Internet. It can be accessed through the Europa server (http://europa.eu.int)

LEGAL NOTICE

Neither the European Commission nor any person acting on behalf of the Commission is responsible for the use which might be made of the

following information

Cataloguing data can be found at the end of this publication

Luxembourg: Office for Official Publications of the European Communities, 1997

ISBN 92-827-9227-7

© ECSC-EC-EAEC, Brussels · Luxembourg, 1997 Reproduction is authorized, except for commercial purposes, provided the source is acknowledged

Printed in Luxembourg

CONTENTS

EXECUTIVE SUMMARY 5

PROJECT OBJECTIVES 9 Overall Objectives 11 Specific Objectives 11

INTRODUCTION 15

EXPERIMENTAL TECHNIQUES AND PROCEDURE 23 Coals and Solvents Used 23

The Dynamic Volumetric Swelling 24 Specific experimental conditions 25 Evaluation of diffusion mechanism of pyridine into coals 28 Evaluation of diffusional parameters for pyridine sorption 28 Evaluation of molar amounts of solvent sorbed 30 Oxidation of coals 31

Solvent Extraction Studies on Coals 31 Coal extraction 31 Infra red spectra of coals 32 Oxygen contents of coal extracts 32 Size exclusion chromatography of coal extracts 33

Thermogravimetry of Coals and Extraction

Residues 34

Dilatometry and permeability measurements 34

Dilatation under pressure 35

Carbonization 35

Surface Area Measurements 36

Gas Diffusion into Coals and semicokes 37 Gas adsorption capacity 38 Diffusion rates 39

RESULTS AND DISCUSSION 41 Solvent Swelling 41 Equilibrium swelling ratios 41 Effect of heat treatment temperature on the solvent swelling of coals in pyridine 44 Kinetics of solvent swelling of coals 44 Effect of solvent basicity on the swelling kinetics and equilibrium swelling of coal in solvents 49 Effects of steric properties of solvent on the equilibrium swelling and kinetics of coal solvent swelling 53 Effect of particle size on the solvent swelling of coal 57 Effect of oxidation on swelling of coal in pyridine 59 FTIR of raw and oxidised coals 61 Effect of oxidation on the solvent swelling of different particle sizes of coal 64

Extraction of Coal 66 Extraction yields of coal 66 FTIR of coal extracts 67 Oxygen contents of extracts 72 Size exclusion chromatography of coal extracts with THF as mobile phase 73

Thermogravimetry 75 Effect of extraction on the temperature of commencement of weight loss and rate of volatile. 76 Dilatometry/permeability of semicokes 77 High pressure dilatometry 78 Surface Areas of Coals and Semi-Cokes 79 Gas Diffusion into Coals and Semi-Cokes 79 Review 83

CONCLUSIONS 85

Specific Conclusions 85

Overall Conclusions 93

References 95 Tables 99 Figures 144

EXECUTIVE SUMMARY

During the carbonization of coals in the coke oven some coals generate a very high

internal pressure and swell so excessively that they damage coke oven walls or cause coke

pushing problems known as "stickers". Several methods have been employed over the years

in the assessment of coal for its suitability in the production of metallurgical coke in the coke

oven process. These characterisation methods include the assessment of thermoplastic

properties such as the free swelling index, the Gray-King coke test, the dilatometry,

plastometry and more recently high pressure dilatometry and plastometry. Although these tests

are adequate for assessing the coking properties of coal they have failed to distinguish those

coals that swell excessively during carbonization and those that do not. Hence the need to look

at the problem of predicting dangerously swelling characteristics of coal using a more

fundamental approach. The aim of the present project is to establish a fundamental

understanding of the relationship between the coal structural properties, the carbonization

mechanism and coke properties for coals which exhibit these properties.

Coal is a complex heterogeneous material and has a macromolecular structure. The

generation of coal thermoplasticity during carbonization involves the initial decomposition of

the macromolecular and "mobile" or extractable phases of coal. Therefore the macromolecular

and mobile phases of coal have been characterised. The microporous structure of semi-cokes

prepared at different temperatures within the plastic phase and resolidification temperature

region were also characterised. These measurements were carried out in order to provide

information on the decomposition of the macromolecular phase to form coke via a

thermoplastic phase.

The characterisation of the coal macromolecular structure involved the use of the

solvent swelling technique. The technique describes the macromolecular phase in terms of the

cross-link density which is determined by the extent of coal swelling in basic solvents. High

swelling in solvents indicate low crosslink density, and low swelling in solvents indicate high

crosslink density. However, since the technique was originally designed for the study of

polymers and had not been fully developed for study of coal it was also considered necessary

to study several factors which might affect the swelling of coal in solvents.

It has been found that several factors affect the swelling of coal in solvents. An

increase in temperature has no marked effect on the extent of swelling of coals in pyridine but

increases the rate of swelling. The apparent activation energy of the solvent swelling increases

with coal rank in conformity with increasing structural stability of coal with increasing rank.

An increase in solvent basicity up to pK^ ~8 (a plateau is reached thereafter) increases the

extent of coal swelling in a solvent and also increases the rate of swelling. Therefore solvents

of high basicity should be used to obtain the maximum swelling of coal. Steric properties of a

solvent affect the extent and rate of coal swelling. The solvent swelling ratio passes through a

maximum with increasing solvent size. This represents a balance between an increase in

swelling ratio required to accommodate the solvent molecule in the macromolecule and a

decrease in the accessibility of larger solvent molecules to the active sites in the coal.

Extraction of coal with solvent before solvent swelling causes a reduction in the extent of

swelling but increases the swelling rate. In addition, the solvent swelling ratio for extracted

coals is only weakly dependent on solvent basicity indicating that hydrogen bonds do not

reform to any great extent after removal of solvent after extraction. This reveals the difference

between raw and extracted coal.

The solvent swelling studies have revealed that the macromolecular structure of

dangerously swelling coals are highly cross-linked by covalent bonds compared to the

macromolecular structure of non-dangerously swelling coals. The swelling in pyridine of coals

heated to different temperatures have also been studied in order to investigate the relative

extents to which heat treatment affects the macromolecular structures of safe and dangerously

swelling coals. The results show that while the crosslink density of the macromolecular

structure of safe coals increase with increasing temperature, the crosslink density of the

dangerously swelling coals decrease with increasing temperature and show minimum values at

a stage during decomposition. This shows that the low rank coals decompose but polymerise

quickly by crosslinking due to their high reactivity whereas the high rank coals decompose and

depolymerise over a wide range of temperatures because of their high stability and low

reactivity. It has also been found that the dangerously swelling coals contain low amounts of

oxygen functionalities and therefore low concentrations of hydrogen bonding interactions when

compared to non-dangerously swelling coals.

Characterisation of the mobile phase involved the extraction of coal with pyridine at its

boiling point. Extraction of the rank range of coals with pyridine showed that the dangerously

swelling coals contain very low amounts of pyridine extractable components i.e. mobile

phase. Using FTIR and size exclusion chromatography data it was shown that pyridine

extracts of dangerously swelling coals were more aliphatic and had lower average molar mass

than the extracts of the non-dangerously swelling coals. Elemental analysis showed that while

the extracts of the dangerously swelling coals contain more oxygen than the corresponding

coals, the extracts of non-dangerously swelling coals contain less oxygen than the

corresponding coals. It is considered that because they are richer in oxygen than their residues

the extracts of the dangerously swelling coals will be more reactive and may react during

carbonization to give rise to highly viscous plastic phase of low permeability which will trap

volatiles and cause high internal pressure and swelling. However, it is considered that the

amount, not the nature of the extractables may be the determining factor since they constitute

only - 2 % of the coal structure. Therefore it is reasonable to conclude that the nature of the

macromolecular phase and the way it decomposes has the major influence in the fluidity of the

system.

Gas diffusion kinetics for oxygen and nitrogen and surface area measurements (CO->

273 K) show that the development of microporosity in coals during carbonization reaches a

maximum at about 600°C before decreasing with further increase in temperature. The rates of

diffusion of oxygen and nitrogen into the cokes were slowest for the dangerously swelling coal

indicating that the permeability of these cokes to the transport of gases is low. The

development of microporosity is lowest for the dangerously swelling coals. It is suggested that

the low microporosity and slow diffusion of gases in the semi-cokes of dangerously swelling

coals cause restriction to the escape of volatiles so that gases cannot escape causing high

internal pressure and dangerous swelling. This is corroborated by the results of a combination

of dilatometry and gas permeability study which shows that the plastic phase of dangerously

swelling coals exhibit very low permeability towards gases compared to the safe coals.

Overall, coals which are described as 'dangerously swelling' in relation to

carbonization in a coke oven have high covalently cross-linked macromolecular structures with

low amounts of 'mobile' phase. This macromolecular structure decomposes involving the

breaking of cross-linking to develop a low fluidity thermoplastic phase which has a low

permeability to the release of volatile. Furthermore the coke formed from the dangerously

swelling coals also has the lowest amount of microporosity and rates of adsorption of gases

into the cokes. This also presents a barrier to the release of volatiles during carbonization.

PROJECT OBJECTIVES

BACKGROUND

It has been apparent from many studies that standard coal characterisation data cannot

be used to predict the dangerously swelling behaviour of coals in coke ovens. Some general

guidelines are available for the occurrence of dangerously swelling properties, e.g dangerously

swelling coals usually have vitrinite reflectance values in the range 1.4-1.65%, and low inert

content. However, there is insufficient knowledge to predict, unequivocally, swelling

characteristics in the coke oven.

It must be recognised that coal has a macromolecular structure and therefore the

conversion of raw coal to a plastic mass will involve the decomposition of the macromolecule,

with the initial decomposition of functionality, and the various types of crosslinks and non-

covalent interactions in the coal. The rate and extent of breakage of the crosslinks which will

depend on the nature and density of the crosslinks may affect the swelling characteristics of the

plastic phase. Therefore, this study has involved an investigation of the relationship between

the macromolecular structure of coals belonging to a wide rank range including safe and known

dangerously swelling coals, by investigating their crosslink densities and non-covalent

interactions, especially hydrogen bonding interactions.

It is also known that coal contains some low molecular weight species which are either

weakly bonded by, for example hydrogen bonding, van der Waal's forces etc, to the

macromolecular structure or are physically trapped in the pore system. The low molar mass

species are also extractable by some solvents e.g. pyridine, and their role in the development

of thermoplasticity of coal is important. In this study, the amount and chemical nature of the

pyridine-extractable materials have been determined in a rank range of coals including

dangerously swelling coals.

It is clear that the excessive swelling and inadequate post plastic contraction of coal

during carbonisation which cause pushing problems and damage coke oven walls occur during

the plastic stage of the coking process. It is therefore not surprising that tests so far developed

to evaluate the dangerously swelling propensity of coals utilise measurements taken during the

plastic phase of the heated coal sample. In a similar manner, this study will involve the

characterisation of semicokes produced at different temperatures from a rank range of coals.

Measurements that will be taken on the semicokes are C02(273 K) surface areas, and the

oxygen and nitrogen capacities as well as the rate of diffusion of these gases into the

semicokes.

One of the characteristics of coal which led to its structure being likened to that of a

polymer is its ability to imbibe solvent and swell. High swelling in a solvent signifies low

crosslink density and vice versa. If a basic solvent e.g. pyridine is used hydrogen bonding

interactions will be disrupted so that the coal structure is opened with the result that the

macromolecular structure swells to an extent determined by the crosslink density. Although

there are many repons of solvent swelling measurements on coal in the literature it appears that

in no case has a correlation been sought between the extent of coal swelling in a solvent and the

dangerously swelling behaviour of the coal during carbonisation. In fact, the only attempt to

compare the solvent swelling of coal and its coking property seems to be that of Sanada and

Honda1 in which it was found that Gieseler fluidity shows a maximum in the region of coal

rank where solvent swelling shows a maximum and crosslink density a minimum. They1

therefore concluded that the degree of crosslinking is an important factor in the thermoplastic

properties of coal. In this study, solvent swelling measurements have been carried out on safe

and dangerous coals using pyridine. Prior to this study the solvent swelling technique had not

been well developed completely since the variation of solvent swelling characteristics with

various parameters have not been investigated systematically. Therefore several factors which

are likely to affect the results have also been investigated in order to improve the understanding

10

of the limitations of the technique.

OVERALL OBJECTIVE

The overall objective of the study is to differentiate between safe and dangerously

swelling coals using data from laboratory studies of the semicokes and raw coals.

SPECIFIC OBJECTIVES

Characterisation of coal Structure

Characterisation of coal structure using solvent swelling

The solvent swelling measurements have been carried out in order to study the

macromolecular structure of coals. The following objectives were included in the

investigation.

1. The variation of solvent swelling ratio of coals in pyridine with rank for a

rank range of coals with emphasis on dangerously swelling coals.

2. The effect of heat treatment temperature on the swelling ratio of coals in pyridine in

an effort to study the changes in the macromolecular structure of coals of different

rank during pyrolysis.

3. The effect of temperature on the swelling ratio of coal in pyridine.

4. The kinetics of coal swelling in pyridine and the effect of coal rank on the

kinetics of solvent swelling in pyridine.

11

5. The effect of solvent basicitv on the extent and kinetics of coal solvent

swelling.

6. The differences in swelling behaviour of raw and extracted coals in order

to examine apparent differences in the macromolecular structure.

7. The effect of solvent steric properties on the extent and kinetics of coal

solvent swelling, and hence establish whether there is a limit to which the coal

macromolecule can swell to accommodate solvent molecules.

8. The effect of oxidation on the solvent swelling of safe and dangerously

swelling coals.

9. The investigation of the effect of particle size on the extent of swelling and kinetics

of raw and oxidised coals.

Determination and characterisation of amounts of the pyridine extractable

materials in coals

The mobile or extractable phase of coal is known to play a vital role in the development

of coal thermoplasticity. Therefore a suite of coals covering a wide range of rank have been

extracted with pyridine in order to study the characteristics of the extractable phase. Also

pyridine extracts have been obtained from coals heated to their softening and maximum

contraction temperatures. The following were included in the objectives.

1. An investigation of the variation of extractable materials in coals with rank, with

special reference to dangerously swelling coals. Also an investigation of the

variation of extractable materials as carbonisation progresses.

12

2 Examination of the structural differences in the extracts of different coals using

(a) FTIR spectroscopy, (b) oxygen contents of the coals, extracts and the

residues (c) the molar mass distribution of the extracts using size exclusion

chromatography.

Thermogravimetry of Coal and Extraction Residues

The coals have also been subjected to temperature programmed pyrolysis in order to

determine temperatures of commencement of decomposition, maximum rate of weight loss ,

end of volatile evolution, as well as rate of volatile loss. These parameters have been evaluated

for coals and their pyridine extraction residues.

Dilatometry Tests

The rank range of coals have been subjected to dilatometry tests using the standard as

well as the high pressure dilatometer.

Surface Area of Semicokes

Surface area measurements have been taken on the same coals and semicokes used in

the gas diffusion studies. The aim was to investigate the variation in microporosity of the coals

and semicokes.

Gas Diffusion into Semicokes

In order to investigate the diffusion of gases through semicoke, the coals belonging to a

wide range of rank, and including safe and dangerously swelling coals have been carbonised to

13

450°, 500°, 600°, 800° and 1000°C and the diffusion rates of oxygen and nitrogen gases

measured on the cokes and the coals, using a gravimetric technique. Also, the permeability of

the plastic phase of the rank range of coals towards nitrogen have been studied in a

combination with dilatometry tests. The aims of the gas diffusion studies are to,

1. investigate the development of the porous structure with heat treatment temperature,

and

2. investigate the differences in the transport of gases through the porous structure

using the gravimetric method as well as studies involving gas permeability

apparatus incorporated into a standard dilatometer.

14

INTRODUCTION

Metallurgical coke is one of the principal raw materials used in the production of iron

by the blast furnace method of production. Apart from its function as an energy source coke

supplies the reducing gas which convéns iron oxides to iron and also provides the permeability

which is necessary for the high performance of the blast furnace2'3.

Coke is produced by the carbonisation of coking coals, and a characteristic feature of

such coals is their ability to soften on heating, become fluid, swell with evolution of volatile

matter, and later shrink on solidification to give a porous structure - semi-coke4, which on

further heating undergoes secondary evolution of gas and yields coke. The period between

softening and solidification of coal during carbonisation is known as the plastic phase. This

stage is very important because it is during this plastic phase that porosity and pore wall

material which are considered important in coke structure are formed5.

A critical stage in the plastic phase of coal during carbonisation is the period of swelling

and evolution of volatile matter. This stage is so important in the carbonisation practice that

Foxwell6 suggests that there can be no production of coke unless there is a pressure set up

within the plastic layer. This pressure. Foxwell6 further suggests, is required to force the fluid

material to engulf the solid matter so as to effect a sort of cementing action which was thought

to be necessary for the production of good coke. These suggestions were based on earlier

work by Blayden et al.7 in which it was demonstrated that the application of external pressure

to non-coking and poorly-coking coals during carbonisation improved the quality of cokes

produced from these coals. Although the pressure and swelling generated in the plastic phase

are important for the formation of good coke there are some coals that generate high swelling

pressures that are capable of damaging coke oven walls. The coals that generate high swelling

pressures that damage coke oven walls are known as dangerously swelling coals.

15

The pressure required in the plastic phase for the production of coke, and the swelling

associated with this pressure are caused by the resistance to the escape of volatiles produced by

pyrolysis8. It is suggested that the only mode of escape for trapped volatile matter in the

plastic mass is by diffusion9. The magnitude of the pressure and swelling will depend upon

the rate at which the volatiles escape from the plastic mass. If the resistance to the escape of the

volatiles is so high that the rate of diffusion of gas through the plastic phase is too low, a

pressure build-up will occur in the plastic phase. If excessively high pressure build-up occurs

in the plastic phase the coke oven walls could be damaged or completely destroyed. Therefore

there is a balance to be achieved between the pressure required to produce agglomeration of the

coal particles and excessive pressures which damage the coke ovens or cause pushing

problems. Hence there is an optimum pressure for the production of metallurgical coke.

There are several views about the cause or origin of the dangerously swelling

characteristics of coals. Mott10 proposed that the coals which develop high coking pressure

and dangerously swelling characteristics become excessively plastic during carbonization so

that plastic coal blocks gas outlets thereby causing abnormal internal gas pressure with resultant

swelling. Addes and Kaegi11 are of the opinion that the gases generated within the plastic

layer try to escape but meet some resistance presented by the low permeability, highly viscous

plastic mass resulting in a pressure build-up in the plastic layer. It was also suggested12 that

lack of high fluidity inhibits gas evolution thereby promoting the dangerously swelling

characteristics of coals. Hermann and Schonmuth13 propose that the inner gas pressure

develops when the tar seams are closed and that this pressure is transmitted by the oven charge

especially when coke fails to detach adequately from heating walls.

The problem caused by the excessive swelling of coal during carbonisation did not

manifest until the introduction of the slot-type coke ovens in the early part of the present

century. Prior to that period coke was produced in the bee-hive oven in which the coal was

free to expand upwards towards the heat source14. The arrangement allowed the pressure

16

developed in the coking mass to be dissipated by the unrestricted expansion without causing

any problem to the structure of the oven. However, the introduction of the slot-type oven,

with its unique heating arrangement, and larger dimensions and full by-product recovery

facilities brought with it some complications in the operation of the oven. The increase in the

oven height, and the use of higher flue temperatures to compensate for increase in oven

capacity created some problems. Secondly the introduction of new techniques such as coal

blending, pre-heating, stamp charging, partial briquetting etc. which were adapted to improve

coke quality15, have also increased the complications in the design and operation parameters of

the coke oven. Although the generation of high coking pressure and excessive swelling during

carbonization is to some extent a property inherent in the coal16, there is no doubt that increase

in the dimension of the oven and introduction of these new coking techniques increase the

tendency of the coal charge to generate excessively high swelling pressures during

carbonisation. The result is that sometimes coke oven walls are distorted or completely

destroyed. Coke pushing problems known as "stickers" are also caused by excessive swelling

and little or no contraction of coal during carbonisation15.

Dangerously swelling coals are characterised by low volatile matter contents (17-

25%) 1 5 and maximum vitrinite reflectance. R0max. of 1.4% and above. Although the

swelling characteristics of a coal during carbonisation is largely determined by the inherent

characteristics of the coal, it can also be influenced by coke plant operating variables.

Hermann and Schonmuth13, and Benedict and Thompson16 report the effects of coal

rank, inert content and coking variables such as bulk density coking rate, and oil addition on

the swelling and coking pressure generated by low volatile coals and their blends. From these

studies13-16 the following generations can be made:

(1) Coking pressure increases with coal rank as measured by vitrinite reflectance

and that the increase in coking pressure becomes more rapid above 1.35%

reflectance.

17

(2) Coking pressure decreases with increase in the inert content of the coal.

(3) Expansion of coal as measured by the Bethlehem Steel tester relates to coal rank

and inert content in the same way as coking pressure. In other words

expansion of coal during carbonisation increases with rank and is more rapid

above 1.35% reflectance and decreases with increase in the inert content of the

coal.

(4) For coal blends, the coking pressure varies with the vitrinite reflectance of the

low volatile coal. The higher the reflectance of the low volatile coal the more its

influence on the coking pressure generated by the coal blend.

(5) Coking pressure increases with bulk density of the charge.

(6) Coking pressure increases with coking rate i.e. heating rate.

(7) Reduction of charge panicle size decrease coking pressure.

(8) Addition of inerts such as coke breeze anil anthracites reduce the excessive

coking pressure generated by low volatile coking coals1-.

The reasons for the above generalisations are not fully understood. However it may be

concluded that the factor of inert content is related to their lower volatile matter contents.

It has also been found that the techniques of pre-heating and stamp charging which are

usually employed to improve coke quality increase swelling and coking pressure of coal

charge14·15. The increase in expansion and coking pressure is due to increase in bulk density

which accompanies pre-heating or stamp charging of coal or blends. It therefore appears that

any factor that increases charge bulk density, e.g. reduction of moisture, will increase coking

pressure.

The adverse economic consequences of excessive swelling exhibited by some coals

during the production of coke in the coke ovenshave led to the development of several test

methods to identify those coals which are likely to exhibit this behaviour during carbonisation.

18

The earliest tests used in testing for dangerously swelling characteristics of coals were

the Kopper's14 small scale test, the Baum-Heuser test17 and the Nedelmann's14 test. These

small scale tests were later found to be inadequate for assessing coals for their dangerously

swelling propensity. For example in some instances coals which were known to be "safe"

were labelled "dangerous" based on the results from these tests. The small scale tests failed

probably because they did not take into account the coke oven operational variables which

normally influence the magnitude of the swelling or coking pressure. These variables include

bulk density of charge, moisture content, oven weight, coking rate etc.

Later test methods are the Kopper's large scale test, the Bethlehem tester18, the

moveable wall oven19'20 and most recently the Kopper's-INCAR tests21. They employed

variable bulk density, heating rate, etc. However they also had limitations as tests for

dangerously swelling coals because they could not adequately simulate coke oven conditions

especially the two-directional heating arrangement.

A test apparatus that most closely simulates the coke oven is the moveable wall oven. It

is a pilot scale facility, but like the commercial coke ovens the moveable-wall oven has two

parallel heating walls so that the heat transfer characteristics are similar to those of the industrial

size ovens. Since one of the walls is is moveable any expansion or pressure developed by the

coal charge during carbonisation is measured by a load cell which is usually installed on the

moveable wall. Also, contraction and pressure probes inserted through the charging hole and

the door respectively allow the contraction of the charge and the variation of gas pressure

during carbonisation to be monitored as the process proceeds. However, it must be noted that

the moveable-wall oven is a semi-industrial scale equipment and requires large quantities of test

samples as well as long periods of test time. It is therefore not convenient to adopt it as a

routine test for identifying dangerously swelling coals.

A general criticism of the early test methods for dangerously swelling coals is the

19

existence of little or no correlations between any two of them. For example it was reported22

that only a little conelation exists between results from the moveable-wall oven and those from

the Nedelmann's apparatus. Also, only a very little correlation was found between pressures

recorded in the moveable-wall oven and those obtained from the Kopper's apparatus14. Lack

of conelations in these tests is so much that a coal which generated a pressure of 90 kPa in the

Kopper's test was certified safe, while that which generated a pressure of 4kPa in the moveable

wall oven was condemned as unsafe. A different approach is therefore necessary in order to

distinguish between safe and dangerously swelling coals.

It has been established that swelling of coal and coking pressures during coal

carbonisation are generated in the plastic phase10"13. It is therefore not surprising that all the

early test methods for identifying dangerously swelling coals are designed to test for this

property in the plastic phase of the coal. For this reason, little or no effort has been made to

study the problem from the fundamental point of view.

Development of plasticity by coals during carbonisation occurs only when the coal

macromolecular structure has been partially or totally destroyed by heating in an inert

atmosphere. The fact that coal has a cross-linked polymeric structure23 implies that the partial

or total destruction of this structure will involve the breaking down of various types of cross­

links in the structure. The nature and relative abundance of these cross-links determine the rate

at which plasticity is developed, and the rate at which plasticity is generated will directly or

otherwise determine the suitability of the coal in a given coal conversion process. In fact it is

reported1 that maxima in Gieseler fluidity and percentage tar produced from vacuum pyrolysis

of coal occur in coals of minimum cross-link density.

In this work the macromolecular structure of the coal has been probed using the

technique of solvent swelling to study the variation in the nature and relative abundance of the

various types of cross-links in the coal structure with change in rank of the coal. Solvent

20

swelling tests have also been carried out on coals heated to various temperatures in order to

examine the variation of the crosslink density as heating progresses. Pyridine extracts of raw

and heated coals have also been estimated and characterised using the infra red technique, the

size exclusion chromatography, and oxygen analysis. Coals of different rank have been

subjected to thermogravimetric analysis to investigate the rates of gas release, the temperature at

the beginning of gas release, temperature at which maximum rate of volatile release occurs,

temperature at which volatile release stops and the temperature range within which volatile

release occurs. Dilatation of coals of varying rank have been measured under varying

conditions of pressure and heating rate. Also, permeability of the plastic phase has been

studied in combination with dilatometry in order to monitor the variation of the permeability of

the plastic phases of coals of various rank as carbonisation progresses. It is thought that a

combination of data from these tests will provide some information about the differences

between safe and danuerouslv swell in«: coals.

21

EXPERIMENTAL TECHNIQUES AND PROCEDURE

COALS AND SOLVENTS USED

The coals used cover a percentage reflectance range, R0 of 0.47 and 1.82, and are

listed in Table 1 together with their characterisation data. Rawdon, Wearmouth, Oakdale and

Lady Windsor were obtained from British Gas pic, while Pinnacle, Oakgrove and Line Creek

were received from the coal bank of the Carbon Research Group of the University of

Loughborough. They were all stored in de-oxygenated distilled water. Hucknall and

Barnburgh were taken from NCRL coal bank. They were stored in lump form in a sealed

plastic bag kept in a freezer and only the interior portions of the lumps were used in this study.

Wentz, Ruhrkhole, Buchanan, Norwich and Saraji were received from the coal bank of the

Carbonisation Department of Spanish National Coal Institute (INCAR) in sealed plastic bags

and were on receipt, stored in a freezer. Those coals stored in water were dried in a vacuum

dessicator but all coals were ground to the required particle size, placed in sample tubes and

stored in vacuum dessicator before use. The coals were supplied with their characterisation

data.

Solvents used were pyridine. 2-chloropyridine. 2-fluoropyridine. η-propylamine, n-

butylamine, n-hexylamine. n-octylamine. and n-decylamine. Each of these solvents was above

99% pure. The solvents were obtained from Aldrich Chemical Company and their properties

are shown in Tables 2 and.3.

23

THE DYNAMIC VOLUMETRIC SWELLING

The dynamic volumetric method, as has been mentioned in an earlier chapter, has the

advantage of measuring only the swelling caused by the solvent that has diffused into the bulk

structure of coal. Therefore results obtained from the DVS system do not require any

correction for solvent occluded in the pores.

The apparatus is similar to that described by Aida and Squires24, Figure 1. Essentially

it consists of a sintered bottom glass tube into which a coal sample of known weight is loaded.

A sliding piston placed on top of the coal sample is connected to a transducer. During an

experiment the glass tube containing the sample and piston is positioned in a glass jacket into

which the solvent is introduced at the appropriate time. The glass jacket is surrounded by a

thermostated water bath to control the reaction temperature to ±0.1 °C. As solvent is taken up

by the sample by sorption, the coal swells, pushing the piston upwards. The vertical

movement is converted to a voltage change by the transducer and transmitted via an A-D

converter to a computer which records time, swelling, and swelling ratio. The swelling as

recorded by the computer represents the distance moved by the piston and is proportional to the

increase in volume per gram of coal, and therefore to the amount of solvent that has diffused

into the bulk structure of the coal. The final swelling values and swelling ratios are

reproducible to ±2% relative. For example, for six measurements on coal under identical

experimental conditions a swelling ratio of 2.10±0.02 was obtained.

The sintered-bottom glass tube was placed in the glass jacket after which the piston was

lowered into the bottom of the tube and a reading corresponding to that piston position obtained

by means of the computer. 2 g of the coal sample was then loaded into the tube, the piston

placed on top of the coal and a reading corresponding to this piston position obtained. In this

way a factor referred to as packing density and defined as:

24

weight of coal Ρ =

height of coal bed

was obtained for each sample, by compacting the coal bed by continuously dropping the piston

on top of the coal until there was no increase in the packing density. This practice ensured that

uniform packing and equal degree of interparticle contact necessary for reproducibility of

results were achieved. Solvent which had been maintained at the required temperature was

introduced into the glass jacket and the system maintained at the required temperature by means

of a thermostated water bath equipped with water circulating facility.

Swelling ratio is calculated using the equation

Qv = h/h; 2

where h is height of coal bed at time t, and h¡ is initial height of coal bed.

Specific Experimental Conditions

Determination of Equilibrium Swelling Ratio of Coals in Pyridine

This work was carried out on coals of particle size range 355-600 ,um using pyridine,

and at temperatures of 20° and 60°C. However, for swelling kinetic studies on those coals that

show relatively high swelling in pyridine swelling measurements were also carried out at 30.

40 and 50°C.

Effect of heat-treatment on the swelling ratios of coals

This study was carried out on coals heat-treated to a range of temperatures in the

thermoplastic region of the coal. Pyridine was used as solvent.

Effect of Solvent Basicirv on Solvent Swelling Kinetics of Coal ■ ' — — ■ ­ ­ — ^ ■ ■

25

For the study of the effect of solvent basicity on the kinetics of solvent swelling of

coals both raw Wearmouth coal and its pyridine extraction residue in the particle size range of

355-600 μπι were used. The solvents used were pyridine, 2-chloropyridine and 2-

fluoropyridine. The solvents are of varying basicity but they have the same steric properties.

The pkb and molar volumes of these solvents at 20°C are shown in Table 2. With the

exception of the swelling of the raw coal in 2-chloropyridine and 2-fluoropyridine which was

carried out at temperatures in the range of 30-70°C the rest of this experiment was carried out at

temperatues ranging from 20-6()°C.

Effect of Solvent Steric Properties on the Equilibrium Swelling and Kinetics of Solvent

Swelling of Coal

Wearmouth coal of particle size range of 355-600 μιη was used in this study. The

solvents used are η-propylamine n-butylamine. n-hcxylamine. n-octylamine and n-decylamine.

These solvents are of the same basicity but different steric properties. The pk^ and molar

volume of the solvents at 20°C are listed in Table 3.

Effect of Oxidation on the Solvent Swelling of Coal

Rawdon. Wearmouth, Buchanan and Pinnacle coals in the particle size range of 355-

600 μπι were used. They were oxidised and sieved again but there was no appreciable

difference in the panicle size range after oxidation. Rawdon and Wearmouth coals of particle

size range 150-212 μιη, 212-250 μπι, 250-355 μιη and 355-600 μιη were also oxidised in

order to investigate the effect of panicle size on the solvent swelling of oxidised coal. Pyridine

was used as solvent and the experiment was carried out at 20°C.

26

Effect of Particle Size on the Equilibrium Swelling and Kinetics of Coal Solvent Swelling

Rawdon and Wearmouth coals of particle size ranges 150-212 μιη, 212-250 μπι, 250-

355 μιη and 355-600 μπι were used. The swelling experiments were carried out in pyridine at

20°C.

Effect of Extraction on the Solvent Swelling

Coals used were Rawdon, Wearmouth, Ruhrkhole, Buchanan and Pinnacle in the

particle size range of 355-600 μπι extracted with pyridine. The solvent swelling of the

extraction residue was also carried out with pyridine. 2-chloropyridine, and 2-fluoropyridine

for extracted Wearmouth coal at temperatures ranging from 20 to 6()°C.

Evaluation of Diffusion Mechanism of Pyridine into Coals

The diffusional coefficient, n. is evaluated using the empirical Equation ^

Mt = ktn 3

Me

A graph of ln(Mt/Me) against In t for the first 25% of the swelling allows the value of η for the

initial swelling to be determined. For example, η = 1/2 for Fickian diffusion, η = 1 for Case

11 diffusion which involves a well defined solvent front. Values of η in the range 0.5 - 1 are

described as anomalous while values of η > 1 are termed Super Case 11.

Evaluation of Diffusion Parameters for Pyridine Sorption into Coals

For a Fickian diffusion into a spherical particle the total amount of diffusing substance

entering or leaving the sphere at long times is described by26:

27

M. 6 — = 1 - - Σ l/n2 exp ( -DrAtW) 4 Mc π 2 n=i

The solution for the equation for short times is:

. M, /Dt '^ , . ~ na -, Dt — = 6 ( — ) π-'/2 + 2 Σ ierfc — - 3 — 5 Me a 2 ' n=i V(Dt)1 a2

In Equations 4 and 5, Mt and Me are mass uptakes at times t, and at equilibrium respectively, a

is the particle radius and D the diffusion coefficient. Equations 4 and 5 are based on the

assumption that the particle radius does not change during the diffusion process. However, the

assumption does not hold for solvent swelling of coals.

Reanangement of equation 4 gives:

M, 6 1 - — = - Σ l/n2 exp (-Dn27t2t/a2) 6

Mc π - n=i

Since the series of Equation 6 converge quickly then only the first term is important and a

graph of ln[(Me - Mt)/Mc| against t will give a straight line with slope equal to D7t2/a2 from

which D/a2 can be calculated. However, a graph of In|(Mc - Mt)/Mc] against t for theoretical

Fickian diffusion is shown in Figure 2. The graph is slightly curved at the very low values of t

due to the importance of the higher terms at such low t values. On the other hand, the graph is

a straight line for swelling ratios above 50% and the best least-squares straight-line fit for the

data >50% swelling gives a value of D/a2 very close to the theoretical value. This is due to the

rapid convergence of the series of Equation 6: the higher terms in the series expansion

becoming insignificant. This leads to the conclusion that for Fickian diffusion into non-porous

spherical particles a graph of ln[(Mc - Mt)/McJ against t approximates to a straight line with

gradient D7t2/a2. However this approximation is good above about 50% swelling.

For small times equation 5 approximates to:

28

M. D'/2t ''-

M,. 7t1/2a

so that a graph of M/Me against t ^ will give a straight line with gradient 6D Vrc '2a from

which D/a2 the diffusional parameter can be calculated. The diffusional parameter is calculated

instead of the diffusion coefficient because a particle size range of the coal samples is used as

opposed to a specific particle size. Secondly, the particle radius changes as the swelling

progresses and so does not remain constant during the process.

The Case II diffusion process into a spherical particle is described by the

equation25-27.

^ . . . ( . . J ä - O 3

Mc C„a'

where C0 is the equilibrium solvent concentration and k0 the relaxation constant. The

corresponding graph of ln[(Mc - Mt)/Mc| against t is shown in Figure 3 and is, as expected

non-linear.

The swelling process can be empirically described in a simple manner if diffusional

limitations are ignored: and in that case it would be assumed that the coal is reacting with the

solvent, it follows, therefore, that:

ds — = k [solventi" [coal reaction sites|n· dt

Since solvent concentration is always well in excess and essentially constant

ds — = k1 [coal reaction sites]"1 10 dt

29

If the reaction is first order, m = 1 and the fraction of reaction sites available for reaction at time

t is (QVgq-QVj), where QV and QVt are equilibrium swelling ratio and swelling ratio at time t

respectively. Hence a graph of ln(l - QV/Qeq) against time will give a straight line if the

swelling is a first order process.

In subsequent chapters the solvent mass uptake at time t, Mt, and equilibrium solvent

mass uptake, Me , have been replaced by the swelling at time t, St, and equilibrium swelling,

Se respectively because in the Dynamic Volumetric Swelling (DVS) method only the solvent

that diffuses into the coal structure causes swelling, and the amount of solvent uptake is

proportional to the swelling.

It must be recognised that Equations 4 and 5 describe diffusion systems where the kinetics

are governed by the Fickian mechanism and would not be applicable to anomalous sorption

process the mathematical theory of which has not been well developed, or the Case II process

in which the mechanism of diffusion is different.

Solvent sorption into glassy polymers contains contributions from both concentration

gradient controlled diffusion and relaxation controlled swelling. The relative contributions of

these two processes may van' with temperature, particle size, etc. A mathematical model has

been proposed27·28 involving linear superposition of phenomenologically independent first

order relaxation terms on a Fickian diffusion equation. The model yields kinetic and

equilibrium values describing the individual contributions of the diffusion and relaxation

processes.

Evaluation of molar amounts of solvent sorbed by coal

In the studies of solvent properties on the solvent swelling of coals, the molar amounts of

solvent absorbed by coal was calculated using the equation29,

30

mmol absorbed Q-1 11 = χ 1000

g of dry coal Vp

where V is the molar volume of the solvent, Q is the swelling ratio of the coal in the

solvent, and ρ is the density of coal.

Oxidation of coals

The coals were oxidised in a fan-assisted oven. Each of the samples chosen for

oxidation studies was spread in a metallic tray that had been wrapped with aluminium foil and

kept in the oven at a temperature of 2()0°C for 24 hours. The oxidised coal was then left to

cool to room temperature in a vacuum dessicator. Oxidised coals were analysed for oxygen

and studied using the solvent swelling technique and infra red spectroscopy.

SOLVENT EXTRACTION STUDIES IN COALS

Coal extraction

A Sohxlet extractor was used for the extraction of the coals. Extraction of all the coals

studied was carried out with pyridine at its boiling point under nitrogen. The particle size range

of 355-600 μηι was used. The extracts were isolated from solutions by means of a rotary

evaporator and further dried for 24 hours in a vacuum desiccator using a pump.

About 4 g of coal was placed in a previously weighed Whatman extraction thimble and

extracted with pyridine in a Sohxlet extractor until the solvent became clear. Coal residues

were Sohxlet-washed for 6 hours with pentane to remove pyridine, dried in vacuum at room

temperature for 24 hours, and weighed.

31

The weight of extract was obtained by difference and the extraction yield calculated

using the equation30.

Extraction yield (wt% d.a.f.)

total extract

(residue - ash + total extract) χ 100 12

Reproducibility was about 1%.

Infra red spectra of extracts

The apparatus used was a Nicolet 20PC FTIR spectrometer. Semi-quantitative FTIR

spectra were obtained on finely ground coal extracts pressed into KBr discs. To obtain

sample-KBr discs exactly 1.83 mg of each extract was mixed with exactly 320 mg of KBr and

ground in a mini agate mortar for K) mins. transferred to a sample tube and dried for 24 hours

in a dessicator using a vacuum pump. In order to minimise the effect of water absorption by

KBr, each sample-KBr mixture was pressed into a disc and spectra obtained immediately using

the Nicolet instrument. For each sample an average was obtained from thirty two scans.

Extracts of both raw and extracted coals were studied.

Oxygen content of coal extracts

Oxygen contents of the extracts were determined using a Carlo Erba CHN EA 1108

Elemental Analvser.

32

Size Exclusion Chromatography of coal extracts

The pyridine extracts of coals were analysed by size exclusion chromatography in order

to obtain some knowledge about the differences in the molar mass distribution as a function of

coal rank.

The size exclusion chromatographic (s.e.c) technique separates molecules on the basis

of their ability to penetrate a porous gel. As the solute passes through the column in the mobile

phase smaller molecules enter the stationary phase while large molecules that can not enter the

pores are confined to the mobile phase so that they are excluded31. The overall effect is that

large molecules travel with the mobile phase and are eluted before the smaller molecules which

are retarded by the porosity of the gel. the limit being the permeation limit. Therefore large

molecules give low retention volumes while small molecules give higher retention volumes.

Calibration graphs31·32 are usually required in order to obtain the molar mass

distribution of material when using the s.e.c. The variation of retention volume of the extracts

with rank of parent coal shows the molar mass distribution of the extracts. Extracts of both

raw and heat-treated coals were studied. Each of tetrahydrofuran (THF) and chloroform was

used as a mobile phase.

The apparatus used consists of an h.p.l.c. pump, a valve system with a 20 μΐ system

loop, a column of 60 cm length packed with a neutral (polytsyrene/divinylbenzene)

macroporous gel of particle size 10 μιη and average pore diameter 10 nm. a UV monitor, an

electronic flowmeter and a chart recorder. The diagram of the s.e.c apparatus is shown in

Figure 4.

Exactly 2.7 mg of each extract was dissolved in 1 ml of THF and 20 μΐ of the solution

injected into the column. It was observed that the extracts were not 100% soluble in THF,

only about 95% of each extract dissolved in THF. The reproducibility for the retention volume

33

is 1%.

THERMOGRAVIMETRY OF COALS AND EXTRACTION RESIDUES

The instrument used was a Stanton Redcroft STA-780 series thermal analyser. Each of

the coal samples was ground to -212 μητ particle size and 30 mg heated at a fixed heating rate

to a final temperature of 900°C in a stream of nitrogen flowing at the rate of 50 cm3 per minute.

The coal samples and the coal extraction residues were analysed at a heating rate of 5°C min1.

However, in order to study the effect of heating rate on the characteristic tempertures and rate

of devolatilisation a representative group of six coals have also been, in addition, analysed at

10, 15. 20, and 25°C.

DILATOMETRY AND PERMEABILITY MEASLREMENTS

A standard Ruhr dilatometer was used for dilatometer tests but was modified so as to

allow the instrument to be used for the study of the changes in the permeability of the sample

throughout the carbonization process. A 1/8" feed gas line into the coal sample has been fixed

onto the base of the dilatometer tube, through the base of the furnace, and sealed. The mould

plug has also been modified so as to forni a pencil with a hole in the bottom which locates over

the gas pipe

The permeability changes of the coal samples were measured simultaneously with the

standard dilatometrie properties. The rate o\' flow of nitrogen gas was used as a measure of

permeability. Values of gas flow were obtained from a Cole-Parmer float type flow meter and

were recorded together with values of corresponding temperatures.

A graph of gas flow against temperaure was used as a measure of degree of

permeability change throughout the coking process. Samples were heated from room

34

temperature to 650°C at 3K/min.

MEASUREMENT OF DILATATION UNDER ELEVATED PRESSURES

The high pressure dilatometer used is a version of the dilatometer described in BS 1016

Part 12 with the only difference being a modification to allow tests to be carried out at various

elevated pressures. Essentially the apparatus comprised of a stainless steel pressure vessel and

has three independant Eurotherm furnace temperature controllers controlling the top middle and

bottom zones of the furnace. A Baskerville and Lindsay pressure release valve controlled by

presetting a limit on the pressure gauge ensures the system was safeguarded against excessive

pressures. The length transducer and thermocouple are connected to a Bryans 26000 A3 X-Y

plotter. The Sclumberger transducer is connected to the Y-axis and a K-type thermocouple is

connected to the X-axis.

The system was flushed out with nitrogen (oxygen free) gas and the pressure was set

using a pressure gauge. The dilatometer tube and piston, thermocouple and length transducer

were then lowered into a vertically positioned cylindrical furnace. This pencil was then

subjected to heating rate of 3 K/min up to 823K. The samples were studied over a range of 0

to 6MPa.

From the dilatometer trace, the values of softening temperature (Tj), temperature of

maximum contraction (T2). resolidification temperature (T3), percentage contraction (%C) and

percentage dilatation (%D) can be obtained.

CARBONISATION

Coals were ground to pass through a 212 μπι sieve, placed in a silica boat, and

carbonised in a horizontal tube furnace under argon at a heating rate of 3°C min-1 to the

35

required temperature. The sample was held at the final heat treatment temperature for one hour.

All the coals shown in Table 1 were carbonised to 450, 500, 600, 800 and 1000°C. The

semicokes produced were all crushed to pass through 600 μιη sieve before gas diffusion and

surface area measurements were taken on them.

SURFACE AREA MEASUREMENTS

Surface area measurements were made on the coals and semi-cokes by the gravimetric

method using the McBain apparatus and carbon dioxide as adsórbate. A schematic diagram of

the apparatus is shown in Figure 5. Essentially the apparatus consists of four main units, (a)

the vacuum system made up of a rotary pump (Eduards, model E04K) and an oil vapour

diffusion pump (Edwards model E2M2). (b) the gas inlet system consisting of a needle value.

(c) a mercury manometer, (d) sample tubes into which aluminium foil buckets containing the

sample are suspended using a pre-calibrated silica spring.

About 200 mg of each sample was weighed into an aluminium foil bucket which was

then suspended into the sample tube using a silica spring. The apparatus was then evacuated

with the pumps while the samples were being heated at 373 Κ using isomantles surrounding

each tube. Evacuation of the apparatus and heating of the samples were continued for four

hours in order to remove any material that may have been adsorbed on the surface. The sample

tubes were then immersed in ice (273 K) and the adsórbate gas (carbon dioxide) introduced

into the apparatus after which the apparatus was left overnight. By varying the manometer

readings and measuring the spring extension using a cathetometer the amount of gas adsorbed

at a given relative pressure was obtained up to pressures of 0.1 MPa.

The results were.calculated by applying the adsorption data to the Dubinin-

Radushkevich (D-R) equation33·34

36

log W = log W - D log2 (p/p0) 13

where W0 is the total volume of micropores in the sample, and W the volume of micropore that

has been filled at a relative pressure of p/p°. By plotting log W against log2 (p/p°) and

extrapolating the linear portion of the D-R plot to log2 (p°/p) = 0 the micropore volume and

hence total number of gas molecules adsorbed was obtained, so that surface area, A was

calculated using the equation:

A = N m x a χ L 14

where Nm = monolayer capaci ty/mmol g_1

a = cross-sectional area of 1 molecule of carbon dioxide i.e. 2 χ K)"19 m2

L = Avogadro number (6.022 χ IO23 molecules)

Reproducibility was about ±5 m2g-'

GAS DIFFUSION INTO COALS AND SEMICOKES

The apparatus used for measuring the gas uptake of coals and semi-cokes by the

gravimetric method is shown in Figure 6. Essentially it consists of five components:

(a) microbalance head

(b) a glass sample bucket that hangs on the balance and surrounded by a glass tube

through the bottom of which gas is let into the sample by means of a tap. The

tube is positioned within an isomantle for outgassing the sample at 100°C.

(c) a data recording and data collection system consisting of a digital balance

controller, standard chart recorder and an Olivetti computer.

(d) a standard glass vacuum rig.

(e) an evacuating system comprising a rotary pump and a diffusion pump. The two

pumps are the same types described in the case of the VlcBain apparatus.

37

The gases whose capacity and diffusion rates were studied are oxygen and nitrogen.

The gases were 99.9% pure and were dried before introducing into the sample tube by passing

through dehydrated silica gel.

About 150 mg of sample is weighed into the glass sample bucket and the rig evacuated.

The sample is then heated under vacuum for one hour to remove any adsorbed moisture after

which the isomantle and pumps are switched off before helium flowing at the rate of 100 c.c

min-1 is introduced into the sample tube through the top of the tube. Another stream of helium

also flowing at the same rate is passed into the tube through a tap at the bottom and allowed to

flow until a constant weight is achieved. The stream of helium flowing through the bottom of

the tube is then replaced with oxygen or nitrogen also flowing at the rate of 100 c c min-1 and

allowed to flow until a constant weight is achieved. The weight against time were recorded and

stored in the computer for data analysis.

Gas adsorption capacity

The gas adsorption capacity of a microporous carbon gives the volume of a particular

gas taken up by one gram of the outgassed carbon when exposed to that gas. The gas capacitv

is calculated using the equation:

MG - MHc + ΒCG 1000 Capacityc = χ 15

MHc PG

where MG = final mass of sample in nitrogen or oxygen.

MHe = final mass of sample in helium.

BC = buoyancy conection for nitrogen or oxygen

pG = density of nitrogen or oxygen.

38

The buoyancy correction factor corrects the gas uptake for the change in buoyancy of the

bucket and sample caused by introduction of the adsórbate gas after a baseline had been

obtained in helium. The buoyancy effect varies with weight of sample and is obtained by

plotting MHC - MG against weight of sample and taking the value on the y axis conesponding to

the weight of sample in the gas used. Buoyancy correction graphs for oxygen and nitrogen are

shown in Figure 7. Reproducibility was 1.2%.

Gas Diffusion rates

Gas diffusion through coals and semi-cokes can be studied by investigating the uptake

of gas in the particles using the unsteady state diffusion of molecules into or out of the material.

The equations that describe the non-steady state gas uptake in a sphere for a Fickian process are

Equations 4 and 4. Diffusional parameters for the diffusion of oxygen and nitrogen into the

semicokes were obtained using the approximate solutions of Equation 4 as described in a

previous section. The reproducibility was approximately 2%.

39

RESULTS AND DISCUSSION

SOLVENT SWELLING

Equilibrium Swelling Ratios

A typical graph of swelling ratio versus time for swelling of coal in pyridine at 20°C is

shown in Figure 8. This graph is characterised by an initial period of rapid swelling followed

by a progressive decrease in the rate of swelling as equilibrium is approached. Values of the

equilibrium swelling ratios obtained at 20°C and 60°C for the complete rank range of coals

studied are shown in Table 4.

At 20°C high equilibrium swelling ratios of 2.17, 2.26, 2.3, 2.10 and 1.80 were

obtained for low rank coals of vitrinite reflectance 0.47-0.91%. Two medium rank coals of

vitrinite reflectance 1.20% gave swelling ratios of 1.32 and 1.30 respectively, while the higher

rank coals of vitrinite reflectance 1.33-1.82% gave swelling ratios ranging from 1.00 to 1.09 in

pyridine. It is apparent that there is a maximum equilibrium swelling ratio at a rank of about

702 in the British Coal Classification Scheme (Vitrinite Reflectance 0.65%). This equilibrium

swelling ratio then decreases gradually with increasing rank to Wentz ( 0.95%), after which the

decrease becomes very rapid until low values are obtained for the high rank coals with vitrinite

reflectance in the range of 1.33 to 1.82%.

The swelling of coal in basic solvents is thought to be a consequence of the disruption

of hydrogen bonds in the coal structure by the basic solvent35·36. Disruption of hydrogen

bonds opens up the coal structure. The strongly basic character of pyridine enables it to break

virtually all die hydrogen bonds in coal, replacing coal-coal hydrogen bonds with coal-pyridine

hydrogen bonds so that the coal swells if these coal-coal hydrogen bonds were active

41

crosslinks36· The ability of pyridine to disrupt all the hydrogen bonds in coal results in the

swelling of coal tending to attain maximum value in pyridine37. The limit of the swelling is

however determined by the covalent cross-link density. High swelling ratio indicates low

crosslink density, and vice versa.

The trend shown in Figure 9 therefore suggests that after the maximum equilibrium

swelling ratio has been attained at a rank corresponding to R0 ~ 0.65 or 702 in the British Coal

classification scheme the swelling ratio decreases with increase in rank showing that covalent

crosslink density increases with coal rank.

Table 5 provides further details of the equilibrium swelling ratios in the range 20-60°C

for the coals the kinetic aspects of which were studied in detail. It is apparent that there is little

variation in equilibrium swelling ratio in this temperature range. However, a number of the

coals appear to reach maximum at about 40°C. However, for the coals studied in detail which

have a relatively low crosslink density, the swelling ratio varies by less than -15%. Hence this

is relatively small compared with the swelling ratios.

Values of equilibrium swelling ratios obtained for coals in pyridine at 60°C show that

while there is a slight decrease for coals in the percentage reflectance range of 0.47 to 1.337c.

there are increases for those coals in the percentage reflectance range of 1.42 to 1.82. These

differences are highlighted in Figures 9 and 10. However these effects do not alter the overall

trends of equilibrium swelling ratio with coal rank.

The slight decrease in the equilibrium swelling ratio of the low rank coals at 60°C in

pyridine compared with the values at 20°C is noteworthy because earlier studies by Sanada and

Honda38, though on pyridine-extracted coal showed an increase in equilibrium swelling ratio

with temperature. Suuberg et al.39 also reported diminished swelling in pyridine of coals dried

at 60 and 100°C and suggested that thermallv induced crosslinkase could occur at these

42

relatively low temperatures. However, the coals used in the present study, though showing

trends similar to those of Suuberg39 were neither extracted with pyridine nor dried at 60 or

100°C, but were merely dried at 20°C in vacuum. Therefore the decrease in their equilibrium

swelling value when measured at 60°C needs to be explained.

The slight decrease in equilibrium swelling ratio observed in this study may have been

caused by irreversible shrinkage of the coal structure that occurs due to the thermally induced

loss of capillary condensed water as well as water attached to polar groups in the coal

structure40"44.

The decrease in swelling ratio at 60°C caused by thermally-induced water loss

decreases with increase in coal rank. The trend suggests that the amount of water loss and the

shrinkage in coal structure caused by the water loss decreases with increase in coal rank. This

is in agreement with the findings of Mahajan and Walker45 that hydrophylic sites in coal

decrease with increase in coal rank. In general, low rank coals have higher moisture and also

higher oxygen contents.

Obviously the major structural factors determining the swelling of coal in a basic

solvent are the hydrogen bonding interactions and the covalent crosslink density. However,

other intermolecular interactions such as π-π aromatic interactions may exert some minor

effects. The slight increase in the swelling ratios at 60CC in pyridine of higher rank coals of

reflectance 1.42% and above is probably caused by the thermal dissociation of π-π aromatic

interactions which may have become significant due to increased aromaticity of the coals in that

rank region as compared to the lower rank coals. The fact that the swelling of these coals in

pyridine at 60°C slightly increases with rank suggests that the π-π aromatic interactions

increase with coal rank46. This is not unexpected in view of the fact that coal aromaticity

increases with rank47.

43

It is clear that those coals that swell in pyridine at 20°C, Figure 9, contain significant

amounts of hydrogen bonding crosslinks, and also have a good degree of open structure

characterised by low covalent crosslink density which allows the structure to swell when the

hydrogen bonding is disrupted by pyridine. However, the extent of hydrogen bonding and

covalent crosslink density would van' from coal to coal. On the other hand the high rank coals

that have not swollen at 20°C in pyridine contain relatively small amounts of hydrogen

bonding, and more importantly, they must be highly covalently crosslinked.

Effect of heat treatment temperature on swelling of coals in solvent

For each coal swelling ratio showed a maximum value at a given heat treatment

temperature, (see Table 6 ). However, while some coals showed the maximum value near the

softening temperature, others showed their maxumum swelling values near the maximum

contraction temperature. Overall, the heat-treatment temperature at which the maximum

swelling occurs increases with increasing coal rank. This reflects the increasing thermal

stability of coal with increasing rank. Low rank coals initially have a low crosslink density

which increases, i.e crosslinks are formed during heat treatment. High rank (dangerously

swelling coals) have high crosslink densities which decompose to produce low crosslink

densities during the thermoplastic phase.

The trend shown by swelling ratio with heat treatment temperature of coal shows that

the coal macromolecular structure changes as the coal is heated. The trend suggests that the

decomposition of the dangerously swelling coals is highly dependent on the structural stability

of the coal.

Kinetics of Solvent Swelling of Coals

Values of the diffusional exponent in Table 5 show that the swelling of Rawdon coal in

pyridine at 20°C is a Fickian diffusion process. A graph of ln((Se-St)/Se) against time for the

44

process is shown in Figure 11. The graph is slightly curved in the initial swelling region but is

linear in the region where the swelling is greater than 50%. The initial curvature is

characteristic of Fickian diffusion and is due to the contribution of the higher terms in the

summation of the infinite series given in Equation 4. The linear portion in the region >50%

swelling is a consequence of a rapidly convergent series which at high values of t in the series

of Equation 4 leads to the high terms being insignificant. This is consistent with the shape of

Figure 2 which is a theoretical Fickian diffusion curve. Also a graph of S t/Se against the

square root of time is shown in Figure 12. This shows that the data are consistent with a

Fickian mechanism with the gradient of the graph being 6D1/2M1/2a for M t/Me < 0.25.

A theoretical Case 11 diffusion curve shown in Figure 3 is consistent with Equation 8.

At low values of t corresponding to M t/Me < 0.3 the graph is linear with respect to time, but at

high values of time the higher terms in the expansion of the equation become relatively more

important and they affect the linearity of the graph. Also an empirical relationship, Mt/Me = 1 -

e_kl, has been proposed for relaxation controlled swelling process. A graph of Mt/Me against t

is linear for small time with a gradient approximately equal to 1, (0.92 for M t/Me < 0.25).

Therefore there are similarities between a moving boundary Case 11 situation and a relaxation

controlled swelling process at small times.

Graphs of ln((Se-St)/Se) against time for the swelling in pyridine of Rawdon coal at

50 C and Wearmouth coal at 30 C are shown in Figures 13 and 14 respectively. Diffusional

exponents shown in Table 5 indicate that the diffusion processes are anomalous and Case 11

respectively. However, it is clear that Figures 13 and 14 are straight lines. Therefore the

graph of ln((Se-St)/Se) against time is an experimentally-determined first order rate law which

applies to the process irrespective of the initial diffusion mechanism. The results are consistent

with the swelling controlled by the relaxation process occurring in solvent swelling. It is

possible that die diffusion mechanism changes to Fickian at some stage in the swelling process.

45

Other coals, namely, Barnburgh, Hucknall, and Wentz also gave straight lines for the

plots of ln((Se-St)/Se) against time for all the temperatures at which they were studied. The

solvent swelling data obtained at temperatures between 20-60°C were analysed to evaluate the

rates, k/s'1, diffusional exponent, n, activation energy and the pre-exponential factor. The

diffusional exponent which indicates diffusion mechanism was evaluated by applying Equation

3 to the data obtained from the initial stages of the swelling process. The coals investigated

showed a range of values for n, (0.96 - 1.20). Rate constants were obtained using the

appropriate solution to Equation 4. Rate constants obtained using the graph of ln(QV -QVt)

against time as deduced from Equation 9 are the same as those obtained from the approximate

solutions of Equation 4. Also diffusional parameters D/a2 have been calculated for swelling of

Rawdon coal using the short term and long tenn approximate solutions for Equations 4 and 5.

The rate constants, ratio of the diffusional parameters for Rawdon. diffusional exponents and

the swelling ratios are shown in Table 5. It is clear that although the extent of swelling does

not vary greatly with temperature, the rates vary greatly with temperature. The ratios of the

diffusional parameters calculated from the later parts of the process to those calculated from

initial parts of the swelling process for Rawdon coal give values of approximately 2. However

the activation energies for both the initial ( M t/Me < 0.25) and final ( M t /M e > 50%) are

similar. The differences in the rates of swelling are accounted for by differences in the

exponential factor. Arrhenius plots for the swelling of the five coals in pyridine are shown in

Figure 15. The apparent activation energies obtained for the coals range from 31.6 to

44.9kJmol·1 ' showing a trend of increase with coal rank in accordance with increasing

chemical stability of coal with rank.

The diffusion mechanism for swelling of Rawdon coal at temperatures of 20-40°C is

Fickian but anomalous for 50 and 60°C. On the other hand it is Case 11 for Barnburgh,

Hucknall, Wearmouth, and Wentz coals at all the temperatures studied. Therefore the diffusion

mechanism of pyridine into these four coals is independent of temperature within the

temperature range of 20-60°C. It is clear that the diffusion mechanism and hence structure of

46

Rawdon coal is significantly different from those of the other coals.

Since the diffusion mechanism for Rawdon coal in the temperature range 20-40°C is

Fickian it would be expected that the values of the diffusional parameter, D/a2, obtained from

the long and short term approximate solutions would be similar. However, as has already been

mentioned they differ by a factor of ~2 at each temperature, with the values for the long term

being higher. This shows that the diffusion coefficient varies through the initial stages of the

diffusion process, increasing as the coal swells. The increase in diffusion coefficient as the

coal swells is probably as a result of opening up of the coal structure during the swelling

process. This non-conformity to the expected behaviour is probably connected with the

complexity of the coal-pyridine system as well as the approximations in the treatment of the

solvent swelling data. For example, the diffusion process in the coal pyridine system is

associated with swelling and chemical reaction. Secondly, the particles are not spherical but

are of varying shapes and consist of a range of particle sizes that change during the swelling

process. It is also clear that the suggested variation of diffusion coefficient with swelling

implies a modification of the accessibility of coal to the solvent as the coal structure changes

with swelling. These associated phenomena cause the coal-pyridine diffusion process to

deviate from the ideal situation described by Equations 4 and 5.

Coals have a porous structure and it is reasonable to expect that the porous structure

will influence the kinetics of solvent swelling. Adsorption isotherms of the coals studied

showed that they are largely microporous. Secondly, it might be thought that pyridine-

extractable materials will affect the activity of solvent and thereby affect the solvent swelling

kinetics. However, C0 2 (273 K) surface areas, pyridine extraction yields and swelling ratio

ranges for the coals are shown in Table 7. Also shown in Table 7 are the apparent activation

energies obtained from Arrhenius plots. Figure 15. as well as the pre-exponential factors.

These data, shown in Table 7, indicate that there is no simple relationship between surface

area, pyridine extraction yield, swelling ratio and the kinetic parameters of the solvent swelling

of the coals in pyridine. Evidently, differences in microporosity have not affected the extent

47

and rate of swelling of these coals in pyridine. The pyridine extract yields are low and the

solvent used in the solvent swelling experiment is in large excess, therefore, the extracts do not

affect the solvent activity.

The apparent activation energies show a trend of increase with increasing coal rank.

The trend is expected in view of the increase in chemical stability of coal with increase in rank.

The values of the apparent activation energies are for combined processes of diffusion,

disruption of coal-coal hydrogen bonding by pyridine, structural relaxation and the swelling of

coal. It will be ambiguous to assign the rate-determining step of the swelling process to

diffusion, relaxation or chemical reaction. However, the increase in apparent activation energy

with rank shows that structural stability increases with coal rank.

The apparent activation energies reported in Table 7 are lower than those reported by

Otake and Suuberg48 for coals of similar rank. It must, however, be realised that these

authors evaluated the activation energies using the assumption that the reciprocal of the time

taken to achieve a particular extent of swelling is proportional to the diffusion coefficient in a

Fickian process and to the relaxation time constant in a Case II process. Apparently the

activation energy calculated on the basis of this assumption will account only for the chosen

extent of the solvent swelling process.

From the study of the kinetics of solvent swelling of these five coals it is clear that: (i)

The solvent swelling process obeys an experimentally-determined first order rate law for Se/S t

> 0.5 irrespective of whether the diffusion mechanism is Fickian. anomalous, or Case II. ii)

The rate law is similar for all the diffusion mechanisms determined from the initial swelling

though it also applies to the later part of the process where the mechanism may also change,

iii) In the Fickian diffusi - the diffusion coefficient increases with time and extent of swelling

probably because of uV oening of the coal structure that occurs as swelling progresses, iv)

The apparent activation energy for the overall process increases with coal rank in agreement

48

with the fact that coal stability increases with rank.

Overall, the dangerously swelling coals, unlike the safe coals, showed very low

swelling ratios in pyridine, and this is consistent with very high crosslink density.

Effect of Solvent Basicity on the Swelling Kinetics and Equilibrium Swelling

of Coal in Solvents

The swelling ratio versus time graphs for the swelling of Wearmouth coal in pyridine,

2-chloropyridine and 2-fluoropyridine at 30 C are shown in Figure 16. The solvents were

chosen because of their similar size, shape, and steric properties but different basicity,

therefore eliminating these steric factors. The data in Table 8 show that the rate of swelling

decreases with decrease in solvent basicity. The equilibrium swelling ratio also decreases with

decreasing solvent basicity. This indicates that the basicity of the solvent is an important factor

in breaking the hydrogen bonds present in the coals. It is apparent that the less basic the

substituted pyridine the fewer the hydrogen bonds which are broken. A graph of ln((Se-

St)/Se) against time for swelling of Wearmouth coal in 2-chloropyridine and 2-fluoropyridine

respectively are shown in Figures 17 and 18. It is clear that the swelling of the coal in the

substituted pyridines obeys the experimentally-determined first order rate law.

Swelling ratios of the extracted coal do not vary significantly in the substituted

pyridines, however the values are lower than those for the raw coals, showing that there is a

significant difference between the macromolecular structure of the raw and extracted coals.

The rate constants and swelling ratios for the swelling of raw and extracted Wearmouth

coal in pyridine, 2-chloropyridine, and 2-fluoropyridine at various temperatures are shown in

Table 8. Also shown in Table 8 are the values for the diffusional exponents for the raw

Wearmouth coal at the various temperatures. Diffusional exponents were not calculated for the

49

swelling of extracted Wearmouth coal in these solvents because the initial swelling process was

very rapid, as indicated in Figure 19, preventing the accurate determination of the diffusional

exponents for the initial uptake of solvent. It is clear that the rate constants for the coal is

strongly influenced by solvent basicity and temperature. On the other hand, for the pyridine-

extracted coal only small variation in the rate constants occur with temperature. Arrhenius plots

for the swelling of the coal in the three pyridines are shown in Figure 20. The apparent

activation energies are 37. 31 and 24 kJ mol"1 for pyridine, 2-chloropyridine, and 2-

fluoropyridine respectively for the raw coal. In contrast, apparent activation energies for the

extracted coal in the three solvents are similar (-10 kJ mol·1) for each of the pyridines. This

indicates that solvent basicity is a major factor in determining the extent of solvent swelling in

raw coals. This arises from the need to break the hydrogen bonds in coal in order to cause

swelling. In contrast, in pyridine extracted coals, the hydrogen bonds have been broken in the

extraction process. It is apparent that the hydrogen bonds have not re-formed after the removal

of the solvent at the end of the extraction process. The solvent swelling of the raw and

extracted coal are determined by the covalent crosslink density. Hence the rates of solvent

swelling of extracted coal are not affected by the basicity since the breaking of hydrogen bonds

are not involved to any significant extent.

The molar quantities of each of the pyridines absorbed by raw and extracted

Wearmouth coal have been calculated using the Equation 11. In the calculation, Wearmouth

coal was assigned a dry density of 1.3 g cm3 because it falls within the rank range of coals

with that value of dry density. The extracted Wearmouth coal was also assumed to have the

same dry density value. Values of pKb, molar volumes of solvents, and the molar amounts of

each solvent absorbed by the raw and extracted coal are shown in Table 9. Also shown in

Table 9 are the activation energies and pre-exponential factors for the swelling of the raw and

extracted coal in each solvent. Since the molar volume of the solvents are similar, the molar

quantities absorbed is not unduly affected by the size of the solvent molecule.

50

Swelling of coal in pyridine and other amines involves the disruption of hydrogen

bonds in the coal and the formation of new bonds between the solvent and oxygen

functionalities in coal. Pyridine is capable of disrupting or breaking nearly all hydrogen bonds

in coal because of its strong basic character36. Therefore, when coal which contains

significant amount of hydrogen bonding is exposed to pyridine these hydrogen bonds are

broken so that the coal swells to a limit primarily determined by its covalent crosslink density.

A solvent will disrupt only those coal-coal hydrogen bonds the bond strength of which are

weaker than those of coal-solvent hydrogen bonds36. Solvents having basicities lower than

that of pyridine will disrupt only the weaker hydrogen bonds so that those of higher bond

strength are left intact. This creates a situation where the number of hydrogen bonds broken by

the solvent is lower than that broken by pyridine. Therefore the solvent of lower basicity will

leave the coal with some hydrogen bonding interactions in addition to the covalent crosslinks.

Therefore the swelling ratio obtained in the solvent of lower basicity is lower than that obtained

in pyridine. This suggestion is corroborated by the equilibrium swelling ratios shown in Table

8 for raw Wearmouth coal.

The values of the diffusional exponents shown in Table 8 indicate that while the initial

mechanism of diffusion of pyridine into Wearmouth coal is Case II, the initial mechanism is

anomalous for the diffusion of 2-chloropyridine and 2-fluoropyridine. It is also clear from

Table 8 that the higher the pK^ of a solvent and hence the lower rate and extent of swelling, the

closer the initial diffusion mechanism for this coal approaches Fickian.

In a solvent swelling study of coals using straight-chain amines of same basicity but

different steric properties Green and West29 found that the molar quantity of solvent absorbed

by each coal was almost constant for all amines, suggesting that all the absorbed amine

interacted with specific sites in the coal. However, their results show that the swelling ratio

increased with molar volume of amine. Contrary to the results of Green and West29 the

present study reported here shows that although the molar volumes of the solvents are not very

different there is no correlation between molar volume of solvent and swelling ratio Table 9.

51

Also the coal absorbed different molar quantities of each solvent, the quantity absorbed

increasing with solvent basicity49. The increase with solvent basicity of molar quantity

absorbed suggests that the number of hydrogen bonding sites reacting with solvent molecules

increases with solvent basicity50 and this further explains the observed increase in equilibrium

swelling with basicity of solvent.

A comparison of the results of the present study and those reported by Green and

West29 suggests that while steric effects determine the swelling ratio for solvents of the same

basicity, differences in basicity determine the swelling ratio in solvents of the same steric

properties. Secondly, the number of specific sites for reaction with the solvent in any given

coal increases with basicity of solvent.

The apparent activation energies of 37.4±2.5. 31.7±1.0 and 24.1±1.6kJ mol·1

obtained for pyridine, 2-chloropyridine. and 2-fluoropyridine show that solvents of lower

basicity interact with weaker hydrogen bonds. Therefore strongly basic solvents interact with a

wider range of hydrogen bond strengths thereby causing larger reductions in the crosslink

density. Therefore coals give higher swelling in solvents of higher basicity.

Swelling of pyridine-extracted Wearmouth coal in pyridine. 2-chloropyridine, and 2-

fluoropyridine also obeys the experimentally-determined first order rate law for Sc/St > 50%,

Figure 19. The swelling ratios and the rate constants for the swelling are shown in Table 8.

The swelling ratios and the amounts of each solvent absorbed at 30°C as well as the activation

energy and the pre-exponential factors for the swelling in each solvent are shown in Table 9.

It is apparent that the swelling ratio is not markedly affected by the basicity of the

solvent in contrast to the trend observed in the raw coal. The activation energies and the pre-

exponential factors for the three solvents are very similar. The activation energies are much

lower (-10 kJ mol·1) than for the raw coals suggesting a different type of interaction. The

52

structural differences between the raw and extracted coals manifested itself in the differences in

the rate constants, swelling ratios, activation energies and pre-exponential factors observed in

their swelling in the three solvents. The solvent swelling of the extracted coal does not vary

greatly with the basicity of the pyridines. This suggests that the formation of hydrogen bonds

in the extracted coal on pyridine removal is limited. This suggestion is supported by the faster

rates of swelling and lower activation energies for the extracted coal which also show no

marked dependence on the basicity of the solvent. The extent of swelling of the raw coal in

pyridine is significantly higher than that for extracted coal which indicates structural changes

involving probably covalent crosslinking and π-π interactions. The reduced swelling in

pyridine of the extracted coal may be caused by increased covalent crosslink density if the

extraction process causes decomposition. Secondly, collapse of the macromolecular structure

due to removal of soluble material leads to formation of strong π-π interactions which may act

as effective crosslinks. Although it is difficult to distinguish between the two possibilities but

Larsen et al.51 favour the latter proposal.

Effects of Steric Properties of Solvent on the Equil ibrium Swelling and

Kinetics of Coal Solvent Swelling

Table 10 gives the equilibrium swelling ratios for Wearmouth coal in straight chain

amines. Considering the solvent swelling ratios obtained at 30°C. it shows that a maximum is

reached at n-hexylamine.

A typical graph of ln((Se-St)/Se) against time for the swelling of coal in the amines is

shown in Figure 21. The rate constants, diffusional exponents, and the swelling ratios

obtained for each of n-butylamine. n-hexylamine, and n-octylamine at different temperatures

are shown in Table 10. Rate constants decrease with increasing amine size, showing that

smaller molecules diffuse into the coal structure more rapidly than the larger molecules.

Diffusional exponents calculated from the initial part of the solvent swelling curve also increase

53

with solvent size, becoming Case 11 for the larger amines. The pKb, molar volumes of

solvents, molar quantities of solvents absorbed, swelling ratio at 30°C, and the activation

energy of the swelling process in each solvent are shown in Table 11. The result shows that

initially the extent of swelling increases with solvent molecular size. However there appears to

be a limit to which the coal macromolecule can swell to accommodate large solvent molecules.

The results suggest that the extent of swelling will level off or begin to decrease when that limit

is attained due to restriction to the diffusion and accessibility of the solvent molecule into the

coal structure. The apparent activation energy for the swelling increases with solvent steric size

indicating that there is an energy barrier for the process which increases as the solvent steric

size increases. Clearly, solvent size is a factor in determining the equilibrium swelling ratio

and solvent swelling kinetics. Also the mechanism of solvent swelling varies with the size of

the solvent molecule, becoming Case 11 for longer chain amines.

The ability of a basic solvent to disrupt hydrogen bonds in coal and thereby cause

swelling of coal depends on the basicity of the solvent. However, for the swelling of coal in

straight chain amines of similar basicity but different steric properties it will be inappropriate to

ascribe any differences to the basicity of the solvents. Differences in the solvent swelling

behaviour of coal in these solvents will occur as a result of the differences in solvent steric

properties. A trend of increasing swelling ratio with increasing molar volume of solvent

(amines) has been reported for three coals29.

Aida and Squires24 also reported the steric effects on the swelling of Illinois No. 6 coal

caused by primary amines of different chain lengths. It was shown24 that the coal gave

swelling ratios of 2.45, 2.64 and 2.19 in η-propylamine, n-butylamine. and n-hexylamine

respectively. However, no explanation was given as to why maximum swelling ratio was

obtained in n-butylamine. They24 also showed that primary and secondary amines are more

effective swelling solvents for coal than tertiary amines.

If solvent basicity is the only property of solvent that affects swelling the coal would

54

give the same swelling ratio in the five solvents used in this study. Table 10 shows that the

swelling ratio of Wearmouth coal in the amines increases from η-propylamine, through n-

butylamine to n-hexylamine, and then decreases from n-octylamine to n-decylamine. It is clear

that the swelling of the coal increases with amine size up to n-hexylamine, but the trend

changes for amines of higher size.

The differences in the equilibrium swelling ratios can be explained in terms of the

ability or the tendency of the coal macromolecular network to swell in an effort to accommodate

the large molecules. The tendency of coal macromolecular network to undergo a

transformation from a glassy to a rubbery state in the presence of amines has been

demonstrated by Brenner52 who also attributed this transformation to the breaking of inter

chain hydrogen bonds by the amine. It has been suggested that breaking of these hydrogen

bonds allows the macromolecular chains of the coal network to re-orient and extend themselves

as the solvent is absorbed29. Therefore, for a given coal the extent to which the chains will re­

orient themselves when the hydrogen bonds arc broken will depend upon the number of

hydrogen bonds broken and the covalent crosslink density.

The swelling of Wearmouth coal in η-propylamine has been caused by breaking of a

certain number of coal inter-chain hydrogen bonds by the solvent. If the steric properties of

these five solvents were the same the coal would have given the same swelling ratio in all the

five amines since their basicities are similar. It is hereby suggested that the higher swelling

recorded in n-hexylamine and n-octylamine is due to the rearrangement of the macromolecular

chains of the coal in order to accommodate the larger molecules. It should be expected that the

swelling ratio of the coal would be higher in n-octylamine and n-decylamine than n-hexylamine

since the macromolecular chains are expected to extend more to accommodate the larger n-

octylamine and n-decylamine molecules. But the similarity in the extent to which the coal

swelled in n-hexylamine and n-octylamine suggests that the coal macromolecular chains

extended to about the same extent in the two solvents. In fact the coal macromolecular

55

structure extended to a lower extent in n-decylamine. Table 11 shows that the molar amounts

of solvent absorbed by the coal is highest for η-propylamine. This is not surprising because it

has the highest accessibility into the coal structure because of its small size. However, the

swelling of the coal in η-propylamine is the lowest due to the fact that the coal macromolecular

structure had to extend only to a low extent to accommodate the small n-propylamine

molecules. Table 11 also shows that the coal absorbs about the same molar amounts of each of

n-butylamine and n-hexylamine but absorbs a lower amount of n-octylamine, and much less of

n-decylamine. It is clear that the lower amounts of n-octylamine and n-decylamine absorbed by

the coal is due to the inability of the coal macromolecular structure to extend further to

accommodate the large molecules or the inaccessibility of some coal-coal hydrogen bonds to

large solvent molecules as a result of steric effects.

It is clear from the extents of swelling of the coal in these solvents that increase in

amine size tends to increase the extent to which the coal network extends to accommodate the

amines, giving rise to increase in swelling with amine size. However, there may be. for each

coal, a critical amine size above which increase in amine size will not produce any increase in

swelling. Earlier workers29 predicted from a graph of swelling ratio versus molar volume of

amines, a swelling ratio of 5.8 for Rawhide coal in n-octadecylamine but wondered whether

the existence of covalent crosslinks may not limit the extent to which the coal can expand in

view of the large molecular size of this solvent. They therefore suggested that the swelling

ratio may level off at some critical amine size. The present study demonstrates that a large

increase in amine size may result in a decrease in swelling if the size of the molecule is so high

that it is highly restricted by steric effects from being absorbed by the coal structure as

observed in n-decylamine.

The swelling of Wearmouth coal in these amines also obeys the experimentally

determined first order rate law for Se/St > 50% as shown in Figure 21. Rate constants increase

with decreasing amine size and with temperature. Apparent activation energies obtained for

56

swelling of the coal in three of the amines show a trend of increase with amine size,

n-octylamine > n-hexylamine > n-butylamine

This shows that the barrier for the activated process increases with increasing size of the amine.

It is apparent from this study that although increasing amine size increases the swelling

of coal, it appears that there must be, for each coal, a critical amine size above which increasing

amine size will not increase the swelling of the coal. Secondly, the steric properties of an

amine have significant effects on the kinetics of coal swelling in the solvent. Therefore

molecular size of the solvent must be considered when choosing solvents for coal solvent

swelling studies.

It is interesting to note that the coal swells in pyridine to the maximum extent observed

in the straight chain amines. It is apparent that pyridine has the appropriate balance of basicity

and steric properties to swell coal to its maximum extent.

Effect of Particle Size on the Solvent Swelling of Coal

Table 12 shows the swelling ratios, the diffusional exponents and the rate constants for

the swelling of various panicle size ranges of Rawdon and Wearmouth coals in pyridine. For

Wearmouth coal there is little or no significant difference in the swelling ratios of the various

particle sizes. The same is true for Rawdon coal. Hence it is apparent that particle size is not a

factor which needs to be considered when studying the solvent swelling of coal.

In all the particle size ranges the diffusional exponent remains the same. It was found

that n was in the ranges (0.45-0.51) for Rawdon and (1.02-1.10) for Wearmouth. It is

apparent that within the particle size range studied ((150 - 212 .urn) to (355 - 600 μπι)) the

57

diffusion mechanism is not affected by the particle size. On the other hand the rate constant

increases with decrease in particle size for both coals.

The results also show that the swelling ratios for Rawdon are more strongly dependent

on particle size compared with the Wearmouth coal. This can be explained by the different

mechanisms. Fickian rates are proportional to the reciprocal of the square of the particle size

while the relaxation controlled process is proportional to the reciprocal of the particle size.

Neither of these relationships is followed precisely since there are a number of assumptions

which are not strictly valid in the theoretical descriptions. These results indicate that the

swelling ratio does not vary greatly with particle size. Hence, it is appropriate to select a single

particle size for comparing swelling ratios.

It is shown that the rate constants increase with decrease in particle size for both coals.

The trend can be explained by the fact that the smaller the particle size of the sample the shorter

the time taken by solvent to get to the centre of the particle giving rise to a shorter equilibrium

time. Although the swelling ratios did not vary so much with particle size there is a discernible

trend of slight decrease in swelling ratio with particle size. Enscore et al.27 determined the

mechanisms of n-hexane sorption into two different particle sizes of polystyrene spheres (184

μπι diameter and 5340Å diameter) and found that it was Fickian for the small particle size and

Case II for the large particle size. These findings do not apply to Wearmouth coal as the

diffusional exponent, n. indicates Case II mechanism for the particle size ranges studied.

There are several reasons to explain the constant nature of the diffusional exponent for pyridine

sorption into the particle size ranges shown for Wearmouth coal. Probably the particle size

ranges are not wide enough to show any changes in the diffusion mechanism. It may also be

that the combined processes of diffusion, chemical reaction and swelling involved in the coal-

pyridine system are more complex compared with the simple polystyrene-n-hexane systems

studied by Enscore et al.27 to give a change in diffusion mechanism with particle size. It is not

surprising that Rawdon coal showed Fickian diffusion for all the particle size ranges because

even the largest particle size has been previously shown to exhibit Fickian mechanism during

58

pyridine sorption53.

Effect of Oxidation on Swelling of Coal in Pyridine

The oxygen contents and swelling ratios in pyridine at 20°C for the raw and oxidised

Rawdon, Wearmouth, Buchanan and Pinnacle coals are shown in Table 13. Also shown in

Table 13 are the values for Wearmouth coal heat-treated in an inert atmosphere to the same

temperature of oxidation for twenty four hours. It is clear that oxidation, as expected,

increases the oxygen content of these coals. In addition, oxidation increases their swelling

ratio in pyridine. On the other hand the oxygen content and swelling ratio of the blank

(Wearmouth coal heated in vacuum for 24 lirs. at 2(K)°C) were not affected.

The swelling ratio in pyridine and the oxygen contents for the raw and oxidised coals

are also shown in Table 13. The swelling ratio in pyridine as well as oxygen contents of coal

increased as a result of oxidation. Each of these oxidised coals showed a two-stage process

during swelling in pyridine, as illustrated in Figures 22 and 23. It is clear that the coal

macromolecule is affected by oxidation.

The activation energy for the first stage in the swelling process of oxidised Wearmouth

coal is 60 kJ mol·1 and that of the second stage is 65 kJ mol·1. The pre-exponential factors are

1.77 χ 107 and 4.47 χ IO8 s-1 respectively. The increase in rate is mainly related to the

increase in pre-exponential factor.

The swelling in pyridine of the Wearmouth coal heat-treated to simulate temperature

effects in the absence of oxygen shows only a one-stage process like the raw coal. Figure 24.

with an activation energy of 43 kJmol·1. and pre-exponential factor of 7.3 χ IO3 s"1. The

rates, diffusional exponents and swelling ratios for the heat-treated and oxidised Wearmouth

coal are in Table 14. Values of the pre-exponential factors show that the improved accessibility

59

between the reacting species are higher for the second stage giving rise to higher reaction rates

irrespective of the slightly higher activation energy involved. This is probably due to the

difference in structure of the material at the later staee but could be a change in solvent

accessibility. It is apparent that heating of coal in the absence of oxygen does not significantly

affect the coal macromolecular structure to the extent it does in the presence of oxygen. The

very similar solvent swelling ratios for raw and heat treated Wearmouth coal indicates that the

crosslink densities are very similar. In other words, the changes in the coal macromolecular

structure were caused by oxidation. However, heat treatment does produce a small overall

increase in the solvent swelling and halves the rate of swelling for temperatures in the range 20

- 60°C. Clearly, there are some minor changes in the coal structure.

Buchanan and Pinnacle coals which exhibit dangerously swelling characteristics during

carbonisation in the coke oven were also investigated. These coals exhibit very low solvent

swelling ratios which increase slightly with increase in temperature. On oxidation the swelling

ratios increased dramatically by -0.8. These results are consistent with a decrease in the

crosslink density with oxidation. It is possible that oxidation has broken some covalent

crosslinks thereby causing a decrease in the covalent crosslink density which has reduced

structural constraints to the swelling of the coal in pyridine.

The values of the apparent activation energies for the swelling of the oxidised coal

shows that the oxygen functionalities in the oxidised coal are more chemically stable than those

in the parent coal. In addition, slower swelling rates and higher activation energies for the

swelling in pyridine are observed. This is not surprising since the coal may have been oxidised

to higher temperatures than it experienced during formation.

It is not very clear why the swelling of the oxidised Wearmouth coal in pyridine

showed a two-stage process. It is possible that the first stage serves to open up the structure so

that swelling proceeds faster at a later stage. Another possibility is that if the outer parts of the

60

coal particles are more oxidised than the inner parts, the outer parts would contain more

oxidised coal and perhaps a greater extent of hydrogen bonding. This would give rise to a

situation where the outer parts of the particles will swell at a slower rate than the inner parts of

the particle, giving rise to a two-stage swelling process.

In order to further understand the structural differences that led to differences in the

solvent swelling of raw and oxidised coals FTIR spectra have been obtained on both raw and

oxidised coals. The idea to oxidise the coals at 200°C stemmed from the fact that oxidation of

coals at temperatures above 170°C is known to cause a drastic decrease in the aliphatic

structures and increase in oxygen-functional groups54·56. Esters and ether groups which

were usually prominent in low temperature oxidised coals to cause increase in crosslinkage and

loss of coking properties57'58 were not detected in coals oxidised at temperatures above

1 7 0 o C 5 4 - 5 6

Fourier-Transform Infra-red Spectra of Raw and Oxidised Coals

Infra red spectroscopy has been used to study the changes that occur in coal structure at

the molecular level as a result of oxidation in an effort to discern the causes of the differences in

the solvent swelling behaviour of coals of different rank. Wearmouth and Buchanan coals

have been chosen for this study because they vary significantly in rank. Wearmouth has a rank

of 502 in the British Coal Classification scheme, whereas Buchanan has a rank of 301a and is

known to be a dangerously swelling coal in coke oven carbonisation.

The FTIR spectra of most coals show several resolved bands between 2800 and 3000

cm"1 and a well resolved band between 3000 and 3100 c m ' l The former are assigned to the

aliphatic C-Η stretching modes of methyl or methylene groups while the later is assigned to the

aromatic C-Η stretching mode59·60. The FTIR spectra of the raw and oxidised coals show

similar absorption bands in this region. On deconvolution the bands assigned to the aliphatic

C-Η stretching can be resolved into six bands which appear at 2830, 2853. 2870, 2890, 2925,

61

and 2960 cm-'. In accordance with an earlier work by Griffiths et al the bands at 2853

2870, and 2890 cm-1 have been assigned to the aliphatic C-Η symmetric stretching of

methylene, methyl, and tertiary CH groups respectively, while the 2925 and 2960 cm-1 bands

are respectively assigned to asymmetric C-Η stretching of methylene and methyl groups. The

band at 2830 cnr1 is assigned to the C-Η stretching vibration of the methoxy group.

The infra-red spectra of the raw and oxidised Wearmouth and Buchanan coals are

shown in Figures 25. 26. 27 and 28. It is clear that these spectra exhibit the characteristic

absorption bands of coal. Although it is difficult to visually detect differences between the

spectra of raw and oxidised coals, it is possible to highlight differences using the difference

spectra, ( see Figures 29 and 30) obtained by subtracting the spectral data of raw coals from

those of the oxidised coals. The original spectra were obtained from specimen containing equal

amounts of sample in equal amounts of KBr. therefore the comparison is valid.

The difference spectra, show negative peaks at 2920 and 1450 cnr1 indicating decrease

in aliphatic groups as a result of oxidation, a prominent positive peak at 1700 cnr1 show a

general increase in carbonyl groups, an increase in carboxylate ions is indicated by the positive

peaks at 1580 and 1395 cm-1. Also negative peaks at 1600 cnr1 and 3050 cnr1 show a

decrease in aromatic structures while the positive peak at 3500 cnr1 indicates increase in

hydroxyl groups.

The method of Christy et al.62, and Sobkowiak and Painter63 has been used to analyse

the bands which comprise the C-Η aliphatic stretching mode. The area under each band has

been taken as the area of the corresponding functional group in the sample. Keuhn et al64

showed that the selection of individual n(C-H) stretching mode shows a better correlation with

coal rank than if the complete set of the n(C-H) aliphatic are considered as a whole. Brown59,

and Durie et al6:\ obtained values of aromatic to aliphatic ratios using the intensities of the

aliphatic C-Η stretching band at 2923 cnr1 and the aromatic C-Η stretching band at 3050 cnr1

62

The areas corresponding to the various functional groups are shown in Tabic 15. It is

clear that both total aliphatic bands as well as aliphatic to aromatic ratios are reduced by

oxidation. While raw Wearmouth coal has an aliphatic to aromatic ratio of 3.17, the oxidised

coal has a value of 1.32. Similarly, the raw and oxidised Buchanan coals have aliphatic to

aromatic ratios of 0.78 and 0.36 respectively. An increase in the swelling of these oxidised

coals in pyridine also indicates that the crosslink density has decreased due to the destruction of

aliphatic crosslinks. Obviously, the aliphatic contents of coal are drastically reduced by

oxidation.

Changes also occur in the oxygen functional groups. Table 15 shows that the band due

to hydrogen bonded OH groups in the 3100 - 3600 cnr1 of the spectra59·65'66 is increased by

oxidation in both coals. It will be recalled that in order to eliminate the effect of moisture

adsorption on the KBr disc each ground mixture of coal and KBr was evacuated for 24 hours

using a vacuum pump and pressed into pellet only at the time of measurement of the spectra.

The bands due to several oxygen functional groups such as the carboxylate ion.

ketones, quiñones, carboxyl groups, and esters67-69 also increase with oxidation. The

solvent swelling studies show that the swelling ratio increases on oxidation for all four coals

studied. This is indicative of a decrease in crosslink density on oxidation. Taking these facts

into consideration it is apparent that aliphatic covalent crosslinks are being broken during

oxidation. It is also clear from Table 15 that oxidation increased the oxygen functional groups.

Increase in oxygen functional groups led to increase in hydrogen bonding interactions.

Buchanan and Pinnacle coals swell dangerously during carbonisation but do not swell

in pyridine. Evidently, swelling of oxidised Buchanan and oxidised Pinnacle coals in pyridine

was caused by the reduction of aliphatic crosslinks as a result of oxidation. Therefore the

density of covalent crosslinks is a major difference between high rank and low rank coals.

63

Dangerously swelling coals have a high crosslink density, and therefore it is proposed that this

is a structural factor which is of significance in assessing the carbonisation characteristics of

coals.

Within the rank range of coals used in this study, those coals that are lower in rank than

Buchanan do not swell dangerously during carbonisation in the coke oven. These coals swell

in pyridine at 20°C because they have low covalent crosslink densities as well as significant

amounts of hydrogen bonding interactions which would be disrupted by pyridine. The

observation that Buchanan coal and those of higher rank do not swell in pyridine at 20°C

indicates that they have high covalent crosslink densities. Buchanan and Pinnacle are known to

be dangerously swelling coals in carbonisation, and they swell in pyridine like non-

dangerously swelling coals after oxidation has reduced their covalent crosslink density and

introduced some hydrogen bonding interactions. This shows that the macromolecular structure

of the dangerously swelling coals differs from that of non-dangerously swelling coals in being

highly covalenti) crosslinked. This result corroborates that of Alvarez et al.70 who measured

the changes in swelling and contraction of a rank range of coals caused by oxidation at 140°C.

using the Koppers-INCAR apparatus and found that after 36 hours of oxidation two

dangerously swelling coals did not show any dangerously swelling behaviours.

Effect of Oxidation on the Solvent Swelling of different Particle Sizes of Coal

The swelling ratio in pyridine of different particle sizes of oxidised Rawdon and

Wearmouth coals are higher than those of the corresponding particle sizes of raw coals. The

various oxidised particle sizes of Rawdon also show a two-stage process during swelling in

pyridine. Rate constants for oxidised Wearmouth are higher than those for oxidised Rawdon.

The rate constants for oxidised Wearmouth coal increase with decrease in particle size.

However, for oxidised Rawdon coal, the smallest particle size ranges studied (212-250 μπι,

and 150-212 μηι) the rate constants for the first and second stages are similar. In fact it

appears that the swelling process for these small particle size ranges of the oxidised Rawdon

64

coal begin to approach a one-stage process. This effect is not significant in oxidised

Wearmouth coal. It is noteworthy that although the oxygen content, after oxidation increases

with decrease in particle size for both coals the percentage increase in oxygen due to oxidation

shows a different trend. While the percentage increase in oxygen content on oxidation

increases progressively but gradually with decrease in particle size for Wearmouth coal. For

Rawdon it decreases from particle size range 355-600 μιη to 250-355 μιη and remains constant

with decrease in particle size. The reason for this trend is not clear. It is also apparent from

Table 16, that the relative increase in oxygen content after oxidation is significantly higher in

Wearmouth coal than in Rawdon coal. Probably, due to its lower oxygen content than

Rawdon, Wearmouth coal has more oxidisable entities than Rawdon.

Table 17 shows the extent of swelling at which the second stage of the swelling begins

for each particle size. Also shown in Table 17 are the depths of penetration of the solvent

before the beginning of the second stage for each particle size. This calculation of the depth of

penetration was based on the assumption of a well-defined solvent front.

The extent of swelling at which the second stage begins increases with decrease in

particle size. The depth of penetration of the oxidation decreases with particle size for Rawdon

coal but does not show any clear trend for Wearmouth coal. In fact, it is almost constant for

Wearmouth coal. This suggests that the oxidation of the two coals involve two different

mechanisms. For Wearmouth coal the constant depth of penetration suggests that oxidation is

proceeding with a well-defined front. It is noteworthy that Wearmouth coal exhibited a Case

11 diffusion mechanism (well-defined solvent front ) during solvent swelling in pyridine.

Therefore it is not surprising that oxidation occurs by an oxidation front moving towards the

centre of the coal particle. The trend shown by the depth of oxidation in Rawdon coal suggests

that the oxidation front is not well-defined as in Wearmouth coal.

65

EXTRACTION STUDIES ON COAL

Extraction Yields of Coal

The percentage yields of the pyridine extraction for the rank range of coals are shown in

Table 18. It is clear that the extraction yield decreases as coal rank increases, as shown in

Figure 32. This shows that the pyridine soluble components of coal which are thought to be

responsible for coal fluidity during carbonisation decrease as coal rank increases. Variation of

extraction yield with coal fluidity is shown in Figure 34. It is clear that extract yield shows

maximum values for coals that show maximum fluidity. However, there are other factors

which may also affect the development of fluidity in the thermoplastic phase.

It is clear that for the rank range of coals studied extraction yield initially increases with

rank, reaching a maximum in the vitrinite reflectance range of -0.8% and decreases rapidly

thereafter with increase in rank, reaching minimum values in the high reflectance region from

1.33% reflectance, as shown in Figure 32. The pyridine extraction yield is low and almost

constant within the volatile matter range of 15.4-21.1% (daf). but increases thereafter with

increase in volatile matter. Figure 33. This is expected from structural consideration, since

extractable material is likely to be volatile. It is striking that pyridine extractability shows a

similar trend with coal rank as does swelling ratio in pyridine. The extraction yield is high for

coals that show high values of fluidity. Figure 6.34. The trend in the variation of extraction

yields with coal fluidity is in agreement with the findings of Wynne-Jones et al.71 who

showed for several coals that the maximum yield of pyridine extracts of coals coincided with

maximum coal fluidity. It is clear that the low volatile, low fluidity bituminous coals which

exhibit dangerous swelling characteristics during carbonisation contain very low amounts of

pyridine extractable materials.

It is thought that the liquid fraction in the pyrolysing system produced by the pyridine

66

extractables determines the fluidity of the system72"74. It is apparent that the coals which

swell excessively in the coke oven during carbonisation exhibit low fluidity and have

correspondingly very low contents of pyridine extractables. The decomposition of the highly

crosslinked macromolecular structure in association with very small amount of mobile phase

leads to the development of a thermoplastic phase of very low fluidity and low permeability.

FTIR of coal extracts

The infra-red spectra of the coal extracts have been studied in order to provide

information regarding the composition of the extracts. The spectra are stacked in order of

increasing rank with that from coal of the highest rank at the bottom in Figures 35 and 36.

Although the spectra feature the characteristic absorption bands expected in the spectra of coal

extracts it is clear that there are variations in the relative intensities of the absorption bands in

the spectra of extracts from different coals.

The FTIR spectra of the coal extracts show similar absorption bands in the same

regions as the coals. Band assignments are the same as in the previous section although minor

shifts occur as expected. On deconvolution of the 2800 - 3000 cm-1 region of the spectra it can

be resolved into six bands which appear at 2X30. 2853. 2870. 2890. 2925. and 2960 cm"'.

The infra-red spectra of pyridine extracts of coals show that the absorption peaks at

2850 and 2925 cnr1 assigned to the aliphatic CH2 and CH3 stretching modes are better

resolved in the extracts of high rank coals from Buchanan to Lady Windsor than in the lower

rank coals ranging from Line Creek to Rawdon. Figure 35 also shows that the aromatic CH

stretching band at 3050 cnr1 is more prominent in the spectra of extracts of the coals ranging

from Rawdon to Line Creek, almost negligible in the extracts of Buchanan, Saraji and

Oakgrove, but appear again in spectra of the extracts of higher rank coals Oakdale, Norwich

and Lady Windsor. The peales at 1454 and 1375 cm"1 which are due to the aliphatic CH3 and

67

CH2 bending modes are very prominent in the spectra of extracts from Buchanan, Saraji and

Oakdale compared to Oakdale, Norwich and Lady Windsor. These peaks are very small in the

spectra of extracts from the low rank coals from Rawdon to Line Creek. It is also evident from

Figure 36 that the aromatic ring vibration absorption peaks at 1600 cnr1 are very low in the

spectra of extracts of Buchanan, Saraji and Oakgrove compared to the extracts of other coals

both higher and lower rank. Hence the extracts of these three coals are more aliphatic than

those of others.

Specific regions of the FTIR spectra of the extracts have been deconvoluted in order to

obtain information on band intensities for the extracts. Using the aliphatic C-Η asymmetric

stretching bands at 2923 and 2960 cnr1 and the aromatic C-Η stretching band at 3050 cnr1 it is

shown in Table 20 that the areas of the aromatic stretching band do not show any trend with

coal rank. On the other hand the areas for the asymmetric C-Η stretching of methylene groups

are low for low rank coals from Rawdon to Line Creek but high for the extracts of higher rank

coals ranging from Buchanan to Lady Windsor. Sobkowiak et al.78 have shown, using the

nuclear magnetic resonance technique that hydrogen content as aromatic C-Η in the pyridine

extract of coals shows no trend with coal rank for coals of 70.3 -87.8 % C. They also showed

that the percentage hydrogen content as aliphatic C-Η decreases with increase in rank of the

parent coal though the decrease was not continuous with rank increase. However, the present

study reported here covers a wider range of coal rank and shows that for coals higher in rank

than those studied by Sobkowiak et al.78 the C-Η aliphatic of the extracts are higher than those

of the extracts of lower rank coals.

Figure 37 and Table 20 show that while the percentage of the area of the methylene C-

H asymmetric stretching frequency increases in the higher rank coals, the percentage of the

area of the methyl C-Η asymmetric stretching decreases. This indicates that for the extracts of

the high rank coals the length of the aliphatic chain probably increases with rank.

The aliphatic to aromatic ratios of the extracts obtained using the aliphatic asymmetric

68

C-Η stretching vibration, shows no clear trend with coal rank for low rank coals from Rawdon

to Line Creek but increases with coal rank for higher rank coals and shows maximum values

for Buchanan, Saraji, and Oakgrove (Vitrinite reflectance 1.33-1.42 %). This is illustrated in

Figure 6.31. The aliphatic to aromatic ratios based on the methyl C-Η asymmetric stretching

frequency shows the same trend with rank of parent coal although the ratios are lower. Using

infra red spectra Durie et al.79 showed that the aromatic to aliphatic ratios of pyridine coal

extracts increase with increase in coal rank for coals of 81.7-89 % C though the increase was

not continuous. In other words, the aliphatic to aromatic ratios decrease with increase in coal

rank for the coals studied. Again it must be highlighted that they studied only a narrow region

of coal rank. The present study shows that the aliphatic to aromatic ratio increases with

increasing rank for the pyridine extracts of coals which are higher in rank than those studied by

Durie et al79.

Although the areas for the asymmetric C-Η stretching for methylene and methyl groups

do not show any clear trend with coal rank for low rank coals from Rawdon to Line Creek, it is

clear that the values are higher for high rank coals from Buchanan to Lady Windsor.

The FTIR spectral region from 700-900 cnr1 has been deconvolved in order to study

the aromatic C-Η out-of-plane deformation modes. It was necessary to fit as many as nine

peaks in some cases in order to obtain a good fit. The pyridine band at 704 cnr1 was

eliminated from each spectra Also the bands at 720. 830. and 840 cm-1 assigned to the

rocking modes of aliphatic CH2 groups63·75 were not required except for the spectra of

Buchanan, Saraji. and Oakgrove which have the 720 cnr1 band. Table 21 shows the areas

obtained for the bands at 750, 770. 801, and 816 cnr1 which correspond to absorptions by

five, four, three, and two adjacent aromatic C-Η groups respectively: and the areas for the

bands at 860, 874, and 888 cm"1 each of which corresponds to an isolated C-Η group in a

different chemical environment61·63·76. The higher the wavenumber of the vibrational band,

the greater the substitution of the aromatic ring which it represents.

69

There is no clear trend between the area of any of these bands with coal rank. Rather it

is possible to separate the spectra into five different groups based on the similarities between

members of each group according to the data in Table 21. The groups comprise of Rawdon,

Barnburgh, Hucknall, Wearmouth and Wentz in group a; Ruhrkhole and Line Creek in group

b; Buchanan, Saraji, and Oakgrove in group c; Oakdale and Norwich in group d; and Lady

Windsor which is quite different from others. It is clear that Buchanan, Saraji, and Oakgrove

have the least peak areas for these aromatic contents as estimated from the n(C-H)arom

intensity. In other words , they are more aliphatic, and this supports the earlier findings in this

work that they have the highest aliphatic to aromatic ratios. The grouping of the spectra

according to similarity based on the data of Table 21 is also observed in Figure 39 (a), (b), (c),

(d) and (e) where the spectra have been stacked according to the groups. It is clear that there is

no marked difference between members of the same group. The spectra also show that extracts

of Buchanan. Saraji, and Oakgrove show the lowest intensity.

It is clear that the spectra of the extracts of Buchanan, Saraji. and Oakgrove differ from

others in that they show a distinct band at 720 cnr1. The band can be assigned to the rocking

modes of the methylene C-Η. This band is very small or entirely absent in the spectra of the

extracts of other coals. This observation lends further support to the conclusion that the

pyridine extracts of these three coals are more aliphatic than those of the others.

The region of absorbance of oxygen functional groups in the FTIR spectra of the coal

extracts (1550-1800 cm'1) is dominated by the broad band at 1600 cnr1, but on deconvolution

of this region of the spectra it is resolved into four bands at 1600. 1655. 1700, and 1720 cnr1 .

The areas of the bands at 1655. 1700. and 1720 cnr ' have been standardised using the band at

1600 cnr1 as an internal standard. In accordance with previous work60 '75·77 the bands have

been assigned to n(C=0) of conjugated carbonyls {e.g. quiñones), carboxylic acids, and

esters respectively. Table 22 shows that the area for the conjugated carbonyls increase with

70

coal rank.

It is clear that the areas of the bands at 1700 and 1720 cm-1 corresponding to carbonyls

of carboxylic acids and esters respectively do not show any trend with rank of parent coal.

However, the area of the band at 1655 cm-1 corresponding to conjugated carbonyls (e.g.

quiñones) show an increase with rank of parent coal, Figure 40. This trend contrasts with a

trend of decrease in total oxygen content with increase in coal rank observed for the extracts.

Probably, the extracts of lower rank coals contain more of their oxygen in the groups that do

not absorb in the C=0 absorption region. It is apparent that the extracts of high rank coals

starting from vitrinite reflectance of 1.33 % have a higher percentage of their oxygen content in

form of conjugated carbonyls, e.g. quiñones. It is clear that the extracts of the dangerously

swelling coals Buchanan. Saraji. and Oakgrove are more aliphatic than those of the other coals,

and also have more of their oxygen content in the form of conjugated carbonyls. probably

quiñones.

Total areas for the oxygen functional group bands show a trend of increase with coal

rank in contrast to a trend of decrease with increasing coal rank shown by values determined by

elemental analysis. This may be as a result of the fact that the major part of the oxygen

contents of the extracts of the lower rank coals occur in other forms that do not absorb in the

n(C-O) stretching vibration region considered in the FTIR spectra. This is supported by the

appearance of high intensity bands in the n(C-O) stretching vibration region in the extracts from

the low rank coals.

Care must be taken in the use of the curve resolution data. An example of this comes

from the curve resolution of the FTIR spectra in the d(C-H)bend region where Hucknall shows

distinct differences based on curve resolution but the FTIR spectrum does not look out of place

when compared with coals of similar rank, see Figure 39(a).

The C-Η stretching and deformation peaks obtained from extracts of coals heated to

71

softening point are shown in Table 23. There seems to be no relationship between areas of the

aromatic and aromatic bands and coal rank. The data also does not show any difference for

dangerously swelling coals. Obviously the decomposition of coal is a very complex process,

involving various pyrolytic and condensation reactions the combined effect of which may not

show any trend with coal rank.

Oxygen Contents of Extracts

Oxygen contents of coals, extracts, and residues are shown in Table 24. They decrease

with increase in coal rank. Figures 41-43. Oxygen content of extracts increases compared with

those of the corresponding coals, Figure 44. However, oxygen contents of the lower rank

coals are higher than those of their extracts whereas for the higher rank coals it is the opposite.

It is clear that while low rank coals in the reflectance range 0.47 to 1.20% show

negative values for (% oxygen in extract - % oxygen in coal), coals of higher rank from 1.33 to

1.82% reflectance show positive values, Figure 45. This trend suggests that the extractable

"mobile" phase of the lower rank coals contain less oxygen than the macromolecular phase

while reverse is the case for the high rank coals. In other words, the extractable mobile phase

of the high rank coals from reflectance of 1.33% and above contain more oxygen than their

macromolecular phase while the extractable phase of those low rank coals below -1.2%

reflectance contain less oxygen than their macromolecular phase. The mobile phase of the

higher rank coals may then be more reactive than the corresponding macromolecular phase

since they are richer in oxygen. The reactivities and relative amounts of the macromolecular

structure and the extracts are important in determining the softening and decomposition

temperatures of the coals. This is of major significance in the development of the thermoplastic

phase. If they are more reactive they will react very fast on heating and even undergo

crosslinking reactions early in the carbonisation process with the result that they will not

develop a high degree of thermoplasticity. Hence these high rank coals produce highly viscous

72

thermoplastic phases which do not allow easy escape of volatiles thereby causing high internal

pressure and excessive swelling during carbonisation in the coke ovens. However it is likely

that the lower functional group concentration as shown by the lower oxygen contents and the

lower total amounts of extracts in high rank coals are the major influences on the

thermoplasticity. However the quantity of the mobile phase needs to be considered in

conjunction with chemical and physical characteristics of the mobile phase.

For lower rank coals the mobile phase of which have lower or about the same oxygen

content with the macromolecular phase, the two phases may be of the same order of reactivity

so that they can react and crosslink in the thermoplastic phase leading to the eventual formation

of semicoke.

Values of the oxygen contents of residues and extracts of coals heated to their softening

and maximum contraction temperatures. Table 25, show that they give the same trend with

rank as those of the raw coals. This is illustrated in Figures 46(a) and 46(b). This shows that

up to these temperatures the compositions of the macromolecule and the extractable phase do

not change greatly indicating only small amounts of decomposition. It is also clear that there is

no definite relationship between oxygen contents of the extracts of the coals heated to their

softening and maximum contraction temperatures and those of the raw coals. Obviously,

decomposition products are chemically different from the raw materials.

Size Exclusion Chromatography of Coal Extracts

Retention volume in tetrahydrofuran (THF) of the extracts from various coals are

shown in Table 26. It is apparent that a weak trend exists. ( see Figure 47 ) of increase in

retention volume of the extracts with reflectance of the parent coal. Extracts from coals of

higher rank show higher values of retention time which is a reflection of lower molecular sizes.

It is clear from Figure 47 that extracts from coals of high rank from about 1.33% and above

73

show the highest values of retention time. Therefore extracts from coals of this rank have the

lowest molar mass distributions in the rank range of coals studied. Earlier studies71 have

shown that coals having maximum molar mass of extracts also show maximum fluidity. It is

therefore not surprising that coals of reflectance 1.33% and above which show low fluidity

also have extracts of low molar mass. It is possible that the decomposition of the higher molar

mass extracts to smaller units gives rise to greater amount of liquid fraction that promotes

fluidity in the coals having extracts of high molar mass distribution. Whereas low molar mass

extracts on decomposition produce limited amount of lower molar mass species and hence low

amounts of liquid fraction and low fluidity.

It is clear from the extraction studies that within the rank range of coals studied those

having reflectance of 1.33% and above have very low quantities of pyridine-extractable

materials. Extracts from some of those coals within the reflectance range of 1.33-1.82% are

more aliphatic than the extracts from other coals. Extracts of coals within this range of rank

contain more oxygen than the corresponding coals while the remaining have lower oxygen

contents than their parent coals. The molar mass of the extracts from this rank range are low

compared to those of others. It is noteworthy that dangerously swelling coals are found within

this rank range of coals.

Table 26 also shows the retention volumes in tetrahydrofuran of pyridine extracts of

coals, and extracts of the coals heated to their softening and maximum contraction

temperatures. The data show that the retention volumes of the extracts of coals heated to the

softening point are higher than those of the extracts of raw coals. On the other hand the

retention volumes of the extracts of those coals heated to maximum contraction temperatures

are lowest. This suggests that at the softening point, the extracts decompose and later

polymerise near the maximum contraction temperature.

The retention volumes in chlorofonn. ( see Table 27 ) show a different trend. It is clear

74

that the retention volume decreases with increasing heat treatment temperature. This shows that

there is a continous decomposition of the extracts as temperature of the coal increases. This

trend appears to be more credible than that obtained in tetrahydrofuran. It is also clear that the

extracts of the dangerously swelling coals show high retention volume in chloroform. The

same trend is maintained in the extracts of the heated coals.

THERMOGRAVIMETRY

The thermogravimerry data obtained from a wide range of coals are shown in Tables 28

and 29. Four characteristic thermogravimetric temperatures and rate of volatile loss in

milligrams per minute are shown for five coals at heating rates of 5, 10, 15, 20, 25 and 40°C

per min, Table 28. The rates of decomposition and temperatures of initial decomposition for

5°C mirr1 heating rate for extracted and raw coals covering a wide rank range are shown in

Table 29. Also shown in Table 29 are the softening temperatures for the rank range of coals.

The temperature at which decomposition begins increases with rank. Lower rank coals

contain higher amounts of hydrogen-bonding interactions which dissociate at low temperatures

to cause initial softening and evolution of volatile matter at lower temperatures. Secondly, the

lower rank coals are more reactive because of higher oxygen contents so they decompose more

easily and at lower temperatures than do the high rank coals. The higher rank coals are highly

covalently crosslinked, and these crosslinks have to be broken in order to bring about initial

softening.

The temperature at which volatile release ends increases with coal rank. Higher rank

coals have very low oxygen contents and hence lower hydrogen-bonding interactions. Higher

rank coals therefore, have more stable structures which require more thermal energy for their

decomposition than would be required by lower rank coals.

For heating rates of 5 and 10°C/min the temperature range of weight loss decreases

75

with increase in rank, reaches a minimum and then increases with rank. On the other hand, it

shows a trend of increase with rank for heating rates of 15, 20 and 25°C. The high values for

lower rank coals at the heating rates of 5 and 10°C may be due to prolonged release of volatiles

caused by the high values of volatile content as well as the low temperature at which it starts.

The high value for high rank coals may have been caused by the prolonged decomposition of

the coal caused by high crosslink density and chemical stability. At higher heating rates the

chemical stability of the coal seems to be the overriding factor.

Rate of volatile loss decreases with increase in coal rank as would be expected from the

decreasing volatile matter content because volatile matter, oxygen functionality and hydrogen

bonding interactions are all decreasing with increase in coal rank, so that the coals are

becoming more chemically stable and less reactive. Secondly, increase in covalent crosslink

density with coal rank may also result in a higher stability.

It is clear that all the characteristic thermogravimetric temperatures as well as rate of

volatile loss are dependent upon coal rank, increase as the coal rank increases due to the

increase in chemical stability of the system.

Effect of Extraction on the Temperature of Commencement of Weight Loss and

Rate of Loss of Volatiles at a heating rate of 5CC.

The aim of studying the thermogravimetric properties of extracted coal is to establish

the role of the macromolecular structure on the thermal decomposition of both safe and

dangerously swelling coals.

The temperature at which release of volatile matter begins increases with coal rank for

both extracted and raw coals although the value for extracted coals are higher. The difference

is larger with low rank coals than for higher rank coals. This observation is expected because

76

the removal of extractable materials of coal will leave the coal with mainly covalent crosslinks

and some non-covalent interactions that may have reformed on removal of solvent. Extractable

materials which usually are the sources of the initial volatiles are absent. Under such situation

only the thermal breakage of covalent crosslinks will cause volatile release and this occurs at

high temperatures than it would if extractables are not removed from the coal. The increase in

temperature of initial decomposition due to extraction is more pronounced in lower rank coals

because they contain higher amounts of extractable materials the effect of the removal of which

will be more pronounced than in the higher rank coals that have lower amounts of extractables.

The rate of weight loss decreases with increase in coal rank for both raw and extracted

coals, but the rates are higher for extracted than for raw coals. The increase in rate of volatile

loss as a result of extraction is caused by the reduction in the additional structural stability

conferred upon the coal by the original hydrogen bonding interactions. These original

hydrogen bonding interactions are virtually absent in the extracted coals. It is also possible that

the extractable material in the raw coal participates in crosslinking reactions during pyrolysis

thereby producing large molecules which would cause a delay in the decomposition of the coal.

This effect will be absent in extracted coal, hence the increased rate of decomposition observed

for extracted coal. It is clear that rates and temperatures of decomposition do not show any

special characteristics of dangerously swelling coals.

Dilatometry/permeability of semicokes

Tables 30 and 31 show the dilatometrie and permeability data obtained from the rank

range of coals. The dilatometry data did not show any differences between safe and

dangerously swelling coals. The data show that the permeability of the semicokes continue to

increase even after the plastic phase has re-solidified. In other words the porosity of the

semicokes continue to change in the secondary devolatilisation region. Figures 48(a) to 48(h)

show the variation of permeability with temperature for Rawdon, Wearmouth, Wentz,

77

Buchanan, and Saraji coals. These graphs show that the permeability of the coal charge

decreases with increasing temperature after softening and attains minimum values and increases

again as temperature increases. It is clear from Table 31 that the temperature after the minimum

at which the permeability starts to increase is rank dependent. It increases with increase in the

rank of the parent coal and are in the range of the temperature of maximum contraction. It is

also shown in Figure 48(h) that the rate of increase of permeability after the minimum value

increases with rank of parent coal, with the dangerously swelling coals Buchanan and Saraji

showing very slow increase in permeability after the minimum values. This shows that escape

of volatiles are markedly restricted in the plastic phase of the dangerously swelling coals, and

this property gives rise to accummulation of volatiles which results in high internal pressure

and swelling pressures that may be capable of damaging coke oven walls.

High pressure dilatometry

Table 32 shows the variation (if dilatometrie parameters with applied pressure for a rank

range of coals. The data shows that there are different effects of pressure on the dilatation of

coals and the effect appears to be different for coals of different rank. Apan from Rawdon. the

dilatation of which is negative and is more or less independent of pressure the dilatation of the

other low rank coals decrease with increasing pressure. On the other hand the dilatation of the

higher rank coals from vitrinite reflectance of 1.33 increase with increase in applied pressure

and show maximum values at varying values of applied pressure. The pressure at which the

maxima occurs is in the region of 15-25 bar. It appears that dangerously swelling coals are

characterised by such maxima in dilatation with pressure.

The dilatation of coals at elevated pressures is controlled by two factors. Increase in

pressure decreases the volume of gases thereby causing a decrease in the swelling of the plastic

layer. On the other hand, increase in pressure increases the residence time of the volatiles so

that they undergo secondary decomposition which leads to increase in fluidity and swelling80.

As already mentioned elsewhere the low rank coals show a gradual decrease in dilatation with

78

increase in pressure while each of the dangerously swelling coals shows a maximum dilatation

at a given value of pressure. This indicates that for the dangerously swelling coals there is a

particular range of pressure within which the effect of increased residence time is more

predominant than the effect of reduced volume of volatiles, whereas the effect of the decrease

in gas volume controls the dilatation of the low rank coals for all the pressures at which the

dilatations were measured.

SURFACE AREA OF COALS AND SEMI-COKES

The C0 2 (273 K) surface areas of the coals and semi-cokes are shown in Table 33. It

is clear that surface areas of the coals and semi-cokes are dependent upon the rank of the parent

coal. The surface areas increase with heat treatment temperature and attain maximum values at

about 600°C and then decrease with increase in temperature. This shows that open

microporosity development during carbonisation attains a maximum value at about 600 C. For

the raw coals as well as the semicokes there is a trend of decrease in the surface area with

increase in rank. The gas capacities and diffusion rates reach a minimum in the medium rank

range except in the HTT 50()°C semicokes. Again, the maximum surface area for the lowest

rank coals Rawdon, Barnburgh and Hucknall do not show maximum values for HTT 600'C

but for HTT 800°C. There appears to be a close relationship between surface area ( C O T .

273K) and gas capacity and diffusion kinetics into the semicokes. The data also show that

microporosity in the semicokes show lower values for the dangerously swelling coals.

GAS DIFFUSION INTO COALS AND SEMICOKES

The two gases, oxygen and nitrogen used in this study are slightly different in

molecular dimensions. Oxygen has a molecular diameter of 0.346 nm while the molecular

diameter of nitrogen is 0.364 nm. A typical uptake versus time curve for oxygen uptake by

semicoke is shown in Figure 49. It is characterised by initial rapid uptake, followed by

79

progressive decrease in rate of uptake until equilibrium is attained.

Oxygen capacities of semicokes prepared at 450, 500, 600 and 800°C are shown in

Table 34. Coals and semicokes prepared at 1000°C did not take up significant quantities of

oxygen. The pore sizes of the coals and semicokes of 1000°C are sufficiently small that the

oxygen molecule cannot diffuse into them. Table 34 shows that for the semicokes made at

450°C only those of the low rank coals Rawdon, Barnburgh, Hucknall, Wearmouth and Wentz

take up oxygen, and the capacity decreases with increase in coal rank. For the rest of the

semicokes the oxygen capacity also decreases with increase in rank of parent coal except that of

600° and 800°C semicokes in which minimum values tend to occur in the region of rank

comprising of Buchanan, Saraji and Oakgrove, see Figure 50. Also, for all coals there appears

to be a trend of maximum capacity at 600°C semicokes except for the low rank coals Rawdon,

Barnburgh and Hucknall. With the exception of these three coals there appears to be a general

trend of increase in microporosity of the semicokes as heat treatment temperature (HTT)

increases with maximum values occurring at HTT of 600°C after which the microporosity

begins to decrease.

A high capacity for oxygen implies high microporosity and if the semicoke shows high

microporosity it means that it has enough channels through which volatiles can escape by

diffusion during carbonisation. If the semicoke has enough channels for escape of volatiles by

diffusion, accumulation of gases and high internal pressure will not occur. However, if the

semicoke is of low porosity, volatiles will have inadequate amount of channels through which

they could escape by diffusion, and under such a situation a large amount of the volatiles will

be retained in the macroporosity. resulting in high internal pressure and high swelling

pressures.

Those coals the semicokes of which have ven' low gas capacities will have a higher

resistance to the diffusion of gases and will exhibit high internal pressure during carbonization.

80

Therefore those coals are likely to develop excessive swelling during carbonization.

Nitrogen capacities for the semi-cokes are shown in Table.35. The coals as well as the

semi-cokes produced at 1000°C did not show any significant adsorption capacity for oxygen or

nitrogen. The semi-cokes produced at 450, 500, and 600°C from Rawdon, Barnburgh,

Hucknall, Wearmouth and Wentz did not adsorb nitrogen. Those produced at 600°C adsorbed

both oxygen and nitrogen with the exception of those from Buchanan and Pinnacle which

adsorbed only oxygen. This is consistent with the restricted access of nitrogen to the porous

structure.

The fact that only HTT 600°C semicokes showed capacities for nitrogen indicates that

pore structure development during carbonization of coals occurs in such a way that the plastic

phase exhibits maximum microporosity at the temperature of about 600°C. The fact that

semicokes of HTT 450, 500 and 1000°C do not take up nitrogen but take up oxygen shows

that their pore diameters are so small that they are inaccessible to nitrogen. It is clear that for

the HTT 600°C semicokes which take up both gases, the oxygen capacity is higher for each

semicoke, showing that the porosity accessible to oxygen is higher than that accessible to

nitrogen due to the larger molecular diameter of nitrogen. This indicates a well-defined narrow

microporous structure with molecular sieving effects. Nitrogen capacity also shows the same

trend with rank as oxygen capacity. Figure 51. Nitrogen capacity of the semicokes also

decreases with increase in rank of parent coal, and shows minimum values for the dangerously

swelling coals, see Figure 51.

A typical graph of ln((Me-Mt)/Me) against time for oxygen uptake into the semicokes is

shown in Figure 52. It is clear that the shape of the graph is consistent with that of the

theoretical Fickian curve, see Figure 2. Similar graphs were obtained for nitrogen diffusion

into those semicokes which showed some adsorption capacity for nitrogen. Also application of

the empirical diffusion equation to the uptake data gives values of diffusional exponent of about

81

0.5, Figure 53. It is clear that the diffusion of oxygen and nitrogen into these semicokes is a

Fickian diffusion process.

It was demonstrated by Schröter and co-workers81 in a study of the influence of pore

structure on the diffusion of Krypton into activated carbons that for spherical particles under

conditions for removed from equilibrium the diffusional parameter increases with increasing

pore diameter. A similar observation was also noted by Juntgen et al.82. It is therefore clear

that higher diffusional parameters indicate larger pore diameters.

The diffusion rates denoted by the diffusional parameter D/a2 for the diffusion of the

gases into the semi-cokes are shown in Table 36 for oxygen and Table 37 for nitrogen. The

diffusion rates for oxygen are higher than those for nitrogen. The trend is not surprising in

view of the larger molecular size of nitrogen. The diffusional parameter for oxygen decrease

with increase in coal rank. Figures 54. The diffusional parameters for nitrogen also show a

trend of decrease with increase in coal rank. Figure 55. with minimum values occurring in the

mid rank region where dangerously swelling characteristics are observed. However, it is clear

that oxygen diffusion rates are higher than those of nitrogen, due to differences in their

molecular size. Hence it is apparent that molecular sieving effects are occuring.

Pore structure of cokes derived from coals which exhibit dangerously swelling

properties show some distinct differences. The cokes have low adsorption capacities and the

diffusion of gases into the structure is slow in comparison to the cokes derived from lower

rank coals. As is normal in coal science the trend has some anomalies. Oakdale is the

exception but the dangerously swelling characteristics of this coal are not known. Buchanan

and Pinnacle are the coals which exhibit the most severe dangerously swelling characteristics.

It is clear that because of the low diffusional parameters associated with small pore

sizes those coals the semicokes of which show low diffusional parameters will develop plastic

phases of low permeability during carbonization. The low permeability of the plastic phase

82

will give rise to accumulation of volatiles due to the restrictions in their escape. This will cause

high internal pressure and excessive swelling.

Review

The dangerously swelling behaviour of coal is caused by the restricted escape of

volatiles through the plastic layer and coke during carbonisation in the coke oven38'39. The

restriction causes accumulation of gases in the macroporosity with consequent generation of

internal pressure which causes excessive swelling. It is the high viscosity and low

permeability of the plastic layer of these coals which is involved in the restriction to the escape

of gases. Dangerously swelling coals develop a plastic layer of low fluidity during

carbonisation unlike the "safe" coals which develop a highly fluid plastic layer during

carbonisation.

Coal consists of the macromolecular phase and the extractable phase, and the pyridine-

extractable component is thought to be involved in the development of fluidity during

carbonisation. This study reveals that the macromolecular structure of the dangerously

swelling coals are highly crosslinked. Secondly, the dangerously swelling coals contain very

low amounts (~ 2 %) of pyridine extractable materials.

Although the nature of the extracts do not vary greatly with coal rank, it is easily

discernible that the extracts of the dangerously swelling coals

(1) are more aliphatic than those of other coals in the rank range studied

(2) have low molar mass distribution

(3) have more oxygen than their parent coals in contrast to the lower rank coals that

have less oxygen than their parent coals

In view of the fact that the extracts of the dangerously swelling coals constitute only

83

about 2% of the coal and their composition only varies to a limited extent, it is clear that the

amount of extract rather than its nature affects the dangerously swelling property of coals.

Secondly, because the macromolecular phase makes up about 98 % of the dangerously

swelling coals it is suggested that the nature of the macromolecular phase and the way in which

it decomposes is a major factor in determining the dangerously swelling properties of coals.

Therefore the decomposition of the highly crosslinked macromolecular structure with a small

amount of extractable phase leads to the formation of highly viscous plastic phase which shows

low permeability to the escape of volatiles.

The low microporosity, low surface area, and low gas diffusion rates measured on the

semicokes of the dangerously swelling coals, when compared with the values obtained from

the semicokes of other coals show that the semicoke / coke structures in dangerously swelling

coals are capable of restricting the escape of volatiles. This restriction to the escape of volatiles

will cause high internal pressure and excessive swelling.

A recent study83·84 of the variation in the pressure exerted by several coals on the

walls of a laboratory test oven shows that Pinnacle. Buchanan, and Oakgrove coals are among

the coals that generated the four highest wall and internal gas pressures recorded in the test.

These three coals are known to exhibit dangerously swelling characteristics in industrial

carbonisation. The results of the wall pressure tests therefore corroborate the findings from the

present study which has highlighted the distinct characteristics of dangerously swelling coals

includine these three coals.

84

CONCLUSIONS

This study was aimed at identifying the fundamental properties of coals that exhibit

dangerously swelling characteristics during carbonisation in the coke oven. In pursuit of that

objective the project was approached from two view points, namely, the study of coal

structure; and the study of the pore structure development in the plastic stage of coal

carbonisation. The specific conclusions drawn from the results obtained have been presented

in line with this approach.

SPECIFIC CONCLUSIONS

Characterisation of Coal Structure

Coals are complex heterogeneous materials and this presents a major difficulty as far as

structural characterisation is concerned. Basic coal characterisation involves the the following:

ultimate and proximate analysis, petrology, caking and swelling properties, calorific value and

ash analysis. This has been shown to be insufficient for the characterisation of dangerously

swelling characteristics in carbonization in the coke oven. A more detailed characterisation of

the coal structure is required. A more detailed characterisation of coal structure was divided *

into two parts, namely, study of the coal macromolecular structure using the solvent swelling

technique, and the estimation of the quantity and characterisation of the extractable materials in

the coal.

Characterisation of the Coal Macromolecular Structure using Solvent Swelling

The study of coal macromolecular structure involved the use of the solvent swelling

technique. The solvent swelling technique was originally designed for the study of

conventional polymers, and had previously not been well developed for the study of coals. It

85

was therefore considered necessary to develop and establish the technique for the study of coal

by way of studying the effects of the various experimental variables on the solvent swelling of

coal. It is thought that standardisation of the experimental conditions is necessary if

reproducible and comparable results are to be obtained. The specific conclusions drawn from

the study of the effects of the changes in the experimental variables are presented below:

1. The effects of temperature on the swelling ratio

Increase in temperature has a relatively small effect on the extent of swelling of coals in

pyridine. An increase in temperature slightly decreases extent of swelling of low rank coals,

but slightly increases extent of swelling of high rank coals in pyridine. However, these slight

variations in the swelling ratio do not affect the overall trend shown by swelling ratio with

rank.

2. The effect of temperature on the kinetics of solvent swelling

The rate of coal swelling in pyndine increases with increasing temperature. The solvent

swelling versus time graphs obey an experimentally determined first order rate law for swelling

>50% irrespective of the mechanism of solvent diffusion into the coal. The rates of swelling

decrease with increase in coal rank. Also the activation energy of the swelling process

increases with coal rank showing that the structural stability of coal is rank dependent.

3. The effect of solvent basicity on the solvent swelling of coal

The rate and extent of coal swelling increase with solvent basicity. Also the mechanism

of solvent sorption into coal is dependent on solvent basicity. It is clear that a wide range of

hydrogen bond strengths exist in coal and the extent to which the hydrogen bonds are broken

depends on the basicity of the solvent. The wider range of bond strengths disrupted the higher

86

the swelling of the coal in that solvent. Therefore, to obtain the maximum value of'swelling on

a coal only strongly basic solvents are recommended. Solvents of lower basicity will only

break the weaker hydrogen bonding, leaving the coal with an apparent higher cross-link

density than it would have if all the hydrogen bonding had been disrupted. This situation

results in a lower swelling in the solvent. It is therefore recommended that only solvents of

high basicity should be used in order that all the hydrogen bonds are broken and that the extent

of swelling is determined by the covalent cross linked density.

4. Differences in the solvent swelling behaviour of raw and extracted coals

Raw and extracted coals show different solvent swelling behaviour. The raw coal

shows higher swelling ratios in pyridine compared with the extracted coal. On the other hand

the rate of swelling is higher for the extracted coal. The reduced swelling and increased rate of

swelling for the extracted coal in pyridine indicate structural differences between raw and

extracted coals. The extent of swelling of the extracted coal in the substituted pyridines is

virtually independent of the basicity. This suggests that extraction with pyridine alters the

original hydrogen bonding in the coal, and it is probable that only a small fraction of the

original hydrogen bonds reform on removal of pyridine. It is also possible that extraction at

the boiling point of pyridine causes slight decomposition which is enough to give rise to

increase in cross-link density. A collapse of the coal structure on removal of pyridine could

also cause an increase in π-π interactions which can act as effective non-covalent crosslinks,

resulting in diminished swelling in solvents. The results suggest that the solvent extraction

process has modified the macromolecular structure of the coal. Therefore it is most appropriate

for the solvent swelling studies to be carried out on the raw coal despite the problems

associated with such measurements.

87

5. Effect of solvent steric properties on solvent swelling of coal

The steric properties of the solvent significantly affects the solvent swelling of coal.

The rate of swelling increases with decrease in molecular size of solvent. The extent of

swelling increases with molecular size of solvent reaching a maximum and then decreases with

further increase in the size of the molecule. The maximum swelling ratio obtained for the suite

of straight chain amines is very similar to that obtained for pyridine. This supports the

proposal that pyridine breaks virtually all the hydrogen bonds in coal. This shows that

although the coal macromolecule extends and re-orientates itself when the hydrogen bonding is

disrupted in order to accommodate the solvent molecule, there appears to be a limit to the ability

of the coal macromolecule to swell. This ability will be dependent on the molecular size of

solvent and also the density of crosslinks. It is therefore very important to consider solvent

steric properties when selecting solvents for solvent swelling studies on coal. It is apparent

that pyridine is a suitable solvent for the study of the solvent swelling of coal in that it appears

to swell the coal to its maximum extent.

6. The effect of particle size on the solvent swelling of raw and oxidised coal

For raw coals only a slight decrease in swelling ratio is observed with decrease in

particle size. Rate constants increase with decrease in particle size while the diffusional

exponent remains the same. For oxidised coal, swelling in pyridine is a two-stage process

with the second stage being faster than the first stage. Swelling ratio also shows a slight

decrease with particle size. It is clear that there is a significant difference between the structures

of raw and oxidised coals. Therefore coals for solvent swelling studies must be adequately

protected from oxidation.

Oxidation causes an increase in solvent swelling ratio corresponding to a decrease in

cross-linked density. This change in the macromolecular structure is associated with a decrease

in the aliphatic/aromatic ratio and an increase in the OH and C=0 functionalities.

88

7. The variation of the swelling ratio with coal rank

The low rank coals in the suite of coals studied have high swelling ratios in pyridine.

Although a maximum value occurs at a reflectance of about 0.77 there is a trend of decrease in

swelling ratio with increases in rank. This shows that covalent cross-link density increases

with rank. Since the dangerously swelling coals do not swell in pyridine, it may be concluded

that they have highly covalent crosslinked structures.

8. The effect of heating on the swelling ratios of coals in pyridine

The low rank coals undergo progressive crosslinking reactions during decomposition,

resulting in progressive increase in their crosslink density. On the other hand the dangerously

swelling coals decompose during pyrolysis in such a way that their crosslink densities decrease

initially, reaching maximum values before crosslinking reactions begin and eventually result in

re-solidification. Clearly, different mechanisms are involved in the decomposition of safe and

dangerously swelling coals.

9. Macromolecular structure of oxidised and dangerously swellinu coals

Raw and oxidised coals which do not exhibit dangerously swelling characteristics

during carbonization in the coke oven usually swell extensively in pyridine. On the other hand

while raw dangerously swelling coal does not swell in pyridine they do swell after oxidation.

Swelling in pyridine of oxidised dangerously swelling coal is caused bv a decrease in covalent

crosslinks. The infrared spectra suggest that this is due to reduction in aliphatic crosslinkages.

Therefore the structural difference between the macromolecular structure of safe and

dangerously swelling coals is associated with the amounts of covalent crosslinks and then

decomposition during carbonisation. Dangerously swelling coals have a high cross-link

89

density and this is associated with low volatiles, oxygen and hydrogen contents.

Determination and Characterisation of Pyridine-Extractable Components of

Coal

The concentration of the pyridine-soluble component of coal shows a maximum value

at a reflectance of 0.77 and above this reflectance value decreases with increase in coal rank

until it attains a very low, but almost constant values in the reflectance region of dangerously

swelling coals. The dangerously swelling coals Buchanan, Saraji, Oakdale, and Norwich have

very low amounts of pyridine extractables. Although the oxygen content of the extracts show a

trend of decrease with increase in coal rank in a similar manner to the raw coals the extracts of

the dangerously swelling coals contain more oxygen than the parent coals. On the other hand

extracts of the non-dangerously swelling coals contain less oxygen than the parent coals.

Extracts of the dangerously swelling coals are more aliphatic and have lower molar mass

distribution than those of the lower rank coals. There does not seem to be any clear differences

in the aromatic and aliphatic contents of extracts of heat treated coals. Extracts of the coals

heated to their softening and maximum contraction temperatures contain less oxygen than those

of the raw coals. Evidently, some of the oxygen has been lost in volatile gases. Like those of

the raw coals, extracts of the dangerously swelling coals that were heat treated to their

softening and maximum contraction temperatures have lower molar mass distributions than

those of the safe coals.

In general the study of coal structure using solvent swelling and solvent extraction

techniques reveals that dangerously swelling coals have a highly crosslinked macromolecular

structure. In addition dangerously swelling coals contain low amounts of extractable materials

which contain more oxygen than the corresponding coals, are more aliphatic and of low

molecular mass distribution compared with non-dangerously swelling coals.

90

In dangerously swelling coals the higher oxygen content of the extracts suggest that the

extracts are more reactive than the macromolecular phase. The reverse situation is true for the

lower rank safe coals where the macromolecular phase has a higher oxygen content. This

corresponds with the differences in the changes in crosslinking with heat treatment

temperature. In the case where the macromolecular phase has high oxygen content, and higher

than the corresponding mobile phase, crosslinking increases with heat treatment temperature.

In the case where the macromolecular structure has a low oxygen content, the phase must

decompose with the accompanying decrease in crosslink density before eventual crosslinking

when carbonisation is complete. The higher reactivity of the mobile phase is likely to result in

"pore blocking" which is likely to lead to low permeability.

Thermogravimetry of Coals and Extraction Residues

The temperatures of initial decomposition, maximum weight loss and rate of weight

loss all depend on coal rank. They also vary with heating rate. There is no relationship

between the thermogravimetric data and dangerously swelling properties of coal.

Dangerously swelling coals are more stable and the decomposition temperatures of the

raw and extracted coals are very similar. The low rank coals have lower decomposition

temperatures than the extracted coals. This corresponds to the larger amounts of extract found

in the low rank coals.

High pressure dilatometry

As a result of their ability to restrict the escape of gases dangerously swelling coals

encourage secondar)' decomposition which results in the increase in the volume of trapped

volatiles thereby increasing coking pressure.

91

Development of Porous Structure

Variation in Microporosity (C0 2 , 273 K surface area)

Surface areas (C02 , 273 K) of the semi-cokes show that those of the dangerously

swelling coals have low values. This confirms that microporous structure of the cokes derived

from the dangerously swelling coals is less extensive.

Gas Permeability of semicokes

The dangerously swelling coals exhibit low permeability to gases compared to the safe

coals. This property is probably the main cause of high coking pressures.

Gas Transport through Porosity of Semicokes

The gas capacities of the semi-cokes show that micropore volume increases as

temperature increases and reaches a maximum at 6()()°C. though for low rank coals Rawdon.

Barnburgh and Hucknall it reaches a maximum at 800°C. The gas capacities also decreases

with increase in rank of parent coal showing low values for the dangerously swelling coals.

The rate of gas transport also decreases with increase in coal rank for each carbonization

temperature, but also shows a maximum value at 600°C for each coal. It is therefore clear that

the semi-cokes of the dangerously swelling coals are very low in microporosity and diffusion

of gases into the cokes is slow. This indicates that molecular sieving occurs in these carbons.

Therefore gases cannot easily escape by diffusion but will be trapped by the coke and this will

cause a high internal pressure and excessive swelling during the carbonization process.

92

OVERALL CONCLUSIONS

1. The various experimental factors which affect the extent of swelling of coal in solvents

have been established. Therefore solvent swelling measurements of coal can now be made

under standard conditions and comparisons of the cross-linked density can be obtained. It is

clear from the results of this study that temperature and particle size have little or no significant

effects on the equilibrium swelling ratios. They affect only the rates of the process. On the

other hand solvent properties such as basicity and steric properties affect both the extent and

rates of solvent swelling. It is also recommended that for solvent swelling studies in basic

solvents those solvents of high basicity e.g. pyridine, are preferred because they can break

almost all hydrogen bonds in coal. Solvent molecular size should not be too high otherwise

diffusion and accessibility into the coal structure may be impaired. Pyridine is considered to be

the most appropriate solvent. Also coal samples to be used in solvent swelling studies must be

protected from oxidation since the macromolecular structures of raw and oxidised coals are

essentially different.

2. This study reveals some of the fundamental characteristics of dangerously swelling

coals. They have a macromolecular structure that contain high density of covalent crosslinks

and low amounts of hydrogen bonding interactions. They also contain very small amounts of

extractable material which are richer in oxygen and more aliphatic than the macromolecular

phase. The extracts also have a lower average molar mass than the non-dangerously swelling

coals.

However it must be recognised that the extractable materials are about only 2% of the

coal structure of the dangerously swelling coals. Therefore it is possible that it is the ven' low

amounts rather than the nature of these extractables that affect the dangerously swelling

characteristics of coals. Moreover, the macromolecular phase of the dangerously swelling

coals make up about 98% of the coal structure, therefore the nature of the macromolecular

93

structure as well as the way in which it decomposes during carbonisation may be a major

factor.

During carbonization the dangerously swelling coals produce a coke which shows

molecular sieving characteristics and therefore impairs escape of volatiles by diffusion. This

leads to high internal pressure and excessive swelling during the carbonization of coal in coke

ovens. Therefore a detailed characterisation of the macromolecular and mobile phases of the

coals provides an insight into the carbonization characteristics of the coal. Further work is

required to obtain a detailed understanding of the differences in swelling pressures obtained for

dangerously swelling coals.

94

References

1. Sanada Y. and Honda H., Fuel 1966, 45, 451.

2. Goleczka J., Tucker J. and Everitt G., Yearbook of the Coke Oven Managers

Association, Mexborough UK, 1983, 148.

3. Marshall J.R. Yearbook of the Coke Oven Managers Association, Mexborough, UK,

1979, 135.

4. Marsh H. and Clarke D.E., Erdöl und Kohle 1986, 39(3), 113.

5. Forrest M. and Marsh H. in "Coal and Coal Products: Analytical Characterisation

Techniques". Ed. E.L. Fuller Jr.. ACS Symposium Series No. 20. Amere. Chem.

Soc, Washington D.C. 1982. 1.

6. Foxwell G.E., J. Institute of Fuel 1939. 304.

7. Blayden H.E. and Noble W. and Riley H.L.. Jouni. Iron and Steel Inst. 1938.

8. Seyler E.A.. Journ. Inst, of Fuel. 1939. 301.

9. Hays D.. Patrick J.W. and Walker Α., Fuel 1976. 55. 297.

10. Mott R.A. J. Inst. Fuel 1940. 189.

11. Addes V.l. and Kaegi D.D. Ironmaking Conf. Proc. of the AIME, Detroit, USA 1990,

663.

12. Marsh H.. Fuel. 1971. 50. 280.

13. Herman W. and Schonmuth F.. Ironmaking Conference Proc. of AIME 1990, Detroit,

USA, 145.

14. CRE Final report on ECSC Project No. 7220-EB 823. 1990.

15. Eisenhart W. in "High Tern. Carbonisation". Chemistry of Coal Utilisation, Ed. M.A.

Elliot, John Wiley & Sons. N.Y.. 1981. Vol. 2. 892.

16. Benedict L.G. and Thompson R.R., Ironmaking Proceedings of AIME, 1976, 35,

276.

17. Baum K. and Henser P.. Fuel 1931, 10. 51.

95

18. Brown W.T., Coal Expansion Proc. Am. Gas. Ass. 1938, 640.

19. Koppers H. and Jenkner Α., Fuel 1931, 10, 232.

20. Soth G.C. and Russell I.C.C, Trans. AIME 1944, 157, 281.

21. Alvarez R., Pis J.J., Barriocanal C. and Lazaro M., Cokemaking International, 1991,

1, 37.

22. BCRA, Technical Paper No. 2, 1948.

23. van Krevelen D.W. " Coal " Elsevier, Amsterdam. 1961.

24. Aida T. and Squires G.. Amer. Chem. Soc. Div. of Fuel Chem. Preprints 1985.

30(1), 102.

25. Ritzger P.L. and Peppas N.A.. Fuel 1987, 66, 1379.

26. Crank J., Mathematics of Diffusion, Clarendon Press. U.K., 1975.

27. Enscore D.J., Hopfenberg H.B.. Stannet V.T., Polymer 1977. 18, 793.

28. Berens A.R. and Hopfenberg H.B. Polymer. 1978, 19. 489.

29. Green T.K. and West T.A.. Fuel 1986. 65. 298.

30. Nishioka M.. Fuel 1991. 70. 1413.

31. Evans N., Haley T.M.. Mulligan M.J. and Thomas K.M., Fuel 1986. 65, 694.

32. Bartle K.D.. Mulligan M.J.. Taylor N.. Martin T.G. and Snape CE. , Fuel

1984, 63. 1556.

33. Dubinin M.M.. Proc. First Coni, on Industrial Carbon and Graphites, Lond.. SCI..

1958. 219.

34 Marsh H., Carbon 1987. 25. 49.

35 Hall P.J., Marsh H. and Thomas K.M.. Fuel 1986. 67. 863.

36 Larsen J.W.. Green T.K. and Kovac J.. J. Org. Chem. 1985. 50, 4729.

37 Hall P.J., Thomas K.M.. and Marsh H..Fuel 1992. 71, 1273.

38 Sanada Y. and Honda H., Fuel 1966, 45. 451.

39 Suuberg E.M., Otake Y., Deevi C. and Yun Y.. Proc. Int. Conf. on Coal Science

1991, Univ. of Newcastle upon Tyne, Butterworth - Heinemann, Oxford, UK, p36.

40. Schäfer H.N.S., Fuel 1972. 51. 4.

96

41. Allardice D.J. and Evans D.G., Fuel 1971, 50, 236.

42. Deevi S.C. and Suuberg E.M., Fuel 1987, 66, 454.

43. Gorbaty M.L., Fuel 1987, 57, 796.

44. Evans D.G., Fuel 1973, 52, 155.

45. Mahajan O.P. and Walker P.L.Jr., Fuel 1971, 50, 308.

46. Nishioka M. and Larsen J.W., Energy and Fuels 1990, 4, 101.

47. Whitehurst D.D. in "Organic Chemistry of coal" 1978, Ed. J.W.Larsen, ACS

Symposium Series, Washington D.C.. pi.

48. Otake Y. and Suuberg E.M., Fuel 1989. 68. 1609.

49. Ndaji F.E. and Thomas K.M., Fuel 1993,72, 1531.

50 Ndaji F.E. and Thomas K.M., Proc. of the 7 t h Int. Conf. on Coal Science. 1993,

Banff, Alberta. Canada, p441.

51 Larsen J.W. and Muhammadi M.. Energy and Fuels 1990,4.107.

52 Brenner D., Fuel 1984, 63, 1325.

53. Ndaji F.E. and Thomas K.M.. Fuel 1993. 72. 1525.

54. Conrow R.B., Durie R.A., Shannon J.S. and Sternhell Α.. Fuel 1963. 42, 275.

55. Friedman L.D. and Kinney CR.. Ind. Eng. Chem. 1950. 42(12), 2525.

56. Clemens Α.Η. Matheson T.W. and Rogers D.E.. Fuel 1991. 70. 215.

57. Liotta R., Brons G. and Isaacs J.. Fuel 1983. 62. 781.

58. Larsen J.W.. Lee D.. Schmidt T. and Grint Α.. Fuel 1986. 65. 595.

59 Brown J.K.. J.Chem. Soc. 1955. 752. 744.

60. Cannon CG., and Sutherland G.B.B.M.. Trans. Faraday Soc. 1945. 41. 279.

61. Wang S. and Griffiths P.R.. Fuel 1985. 64. 229.

62. Christy A.A.. Liang Y.Z., and Kvalheim O.M.. Fuel 1992. 71. 125.

63. Sobkowiak M., and Painter P.. Fuel 1992. 71. 1105.

64. Kuehn D.W., Snyder R.W.. Davis Α.,and Painter P.C., Fuel 1982, 61, 682.

65. Durie R.A.. Shewchyk Y., and Sternhell S.. Fuel 1966, 45. 99.

66. Painter P.C.,Sobkowiak M.. and Youtcheff J., Fuel 1987, 66, 973.

97

67. Painter P.C., Snyder R.W., Pearson D.E.. and Kwong J., Fuel 1980, 59, 282.

68. Bellamy L.J., The Infra Red Spectra of Complex Molecules, Chapman and Hall,

London, 1975.

69. Conrow R.B., Durie R.A., Shannon J.S. and Sternhell Α., Fuel 1963, 42, 275.

70. Alvarez R., Pis J.J., Borriocanal C. and Lazaro M., Cokemaking International, 1991,

1, 37.

71. Wynne-Jones W.F.K., Blayden H.E. and Shaw F., Brenstoff-Chemie, 1952, 33,

201.

72 Brown H.R., and Waters P.L., Fuel 1966. 45, 17.

73. Fitzgerald D., Trans. Faraday Soc, 1956. 52. 362.

74. Dryden I.G.C. and Pankhurst K.S., Fuel 1955. 34. 363.

75. Painter P.C., Starsinic M.. Squires E.. and Davis Α.. Fuel 1983, 62, 742.

76. Williams D.H. and Flemming I.. Spectroscopic Methods in Organic Chemistry. 4 t n

Edition. McGraw-Hill. London. 19X9.

77. Brown J.K.. Fuel 1959, 38. 55.

78. Sobkowiak M., Reisser E.. Given P.. and Painter P.. Fuel 1984, 63, 1245.

79. Durie R.A..Shewchyk Y., and Sternhell S. Fuel 1966, 45, 99.

80. Green P.D. and Thomas K.M.. Fuel 19X5.64.1423.

81. Schröter D.. Juntgen H. and Peters W.. Carbon 1973.11.93.

82. Juntgen H.. Knoblauch Κ.. Munzner Η.. Schröter D., and Zudorf D., Proc. 4 t n i m .

Carbon and Graphite Cnof. SCI. London. 1976.441.

83. Jordan P. Patrick J.W.. and Walker Α.. Cokemaking International, 1992.4,12.

84. Psomiadu E., M.Phil. Thesis. Loughborough Universitv of Technology. 1993.

98

Rawdon

Barnburgh

Hucknall

Wearmouth

Pittsburgh No. 8

Wentz

Ruhrkohle

Line Creek

Woodside

PD

Buchanan

CC

Saraji

Oakgrove

Oakdale

K

Norwich

Pinnacle

Ladv Windsor

0.47

0.65

0.77

0.85

0.87

0.92

1.20

1.20

1.26

1.32

1.33

1.39

1.42

1.42

1.46

1.48

1.49

1.55

1.82

Table 1

Coals used and their Characterisation Data

Proximate Analysis

% Reflectance Volatile Matter Ash

(d.a.f.)

37.1 5.62

38.5 4.9

38.7 3.8

35.9 4.2

32.1 16.0

31.8 5.9

23.4 7.4

23.3 10.2

28.3 8.1

20.6 9.8

21.1 7.4

21.4 9.3

19.1 9.5

21.7 9.X

19.7 10.0

18.6 10.0

17.8 9.2

17.2 5.2

15.4 8.9

Maceral Analysis

V E I

66

78

78

74

n.s

81

73

n.s

n.s

n.s

84

n.s

75

n.s.

65

n.s

77

n.s.

70

12

10

9

12

n.s

5

3

n.s

n.s

n.s

-

n.s

0

n.s.

-

n.s

0

n.s.

_

22

12

13

14

n.s

15

14

n.s

n.s

n.s

10

n.s

19

n.s

35

n.s

13

n.s

30

n.s. = not supplied.

99

Table 2

Properties of Solvents used in the Studv of the Effects of Solvent

Basicitv on Coal Solvent Swelling

Molar Volume (cm3 mob1)

80.9

94.6

86.1

Solvent

pyridine

2-chloropyridine

2-fluoropyridine

pkb

8.6

13.5

14.4

100

Table 3

Properties of Solvents used in the Studv of the Effects of Solvent

Steric Properties on the Solvent Swelling of Coal

Solvent

n-propylamine

n-butylamine

n-hexylamine

n-octylamine

n-decvlamine

pkb

3.3

3.2

3.4

3.4

3.6

Molar Volume (cm3 mob1)

82.2

98.8

132.1

165.3

199.8

101

Table 4

Equilibrium Swelling Ratios of Coals at 20 and 60°C

Coal Reflectance / % Swelling Ratio in Pyridine 20°C 60°C

Rawdon 0.47

Barnburgh 0.65

Hucknall 0.77

Wearmouth 0.85

Wentz 0.92

Ruhrkohle 1.20

Line Creek 1.20

Buchanan 1.33

Saraji 1.42

Oakgrove 1.46

Oakdale 1.46

Norwich 1.49

Pinnacle 1.55

Lady Windsor 1.82

2.17

2.26

2.13

2.10

1.80

1.32

1.30

1.09

1.00

1.02

1.02

1.03

1.01

1.02

2.06

2.16

2.03

2.04

1.77

1.30

1.26

1.05

1.04

1.18

1.18

1.20

1.17

1.21

102

Table 5

Rate Constants. Ratios of Diffusional Coefficents* Diffusional Exponents and Swelling Ratios

Coal

Rawdon

Barnburgh

Hucknall

Wearmouth

Wentz

Temp. (°C)

20 30 40 50 60

20 30 40 50 60

20 30 40 50 60

20 30 40 50 60

20 30 40 50 60

K (s-1)

4.84x10-4 7.09x10-4 1.31x10-3 1.62x10-3 2.23x10-3

2.64x10-4 3.34x10-4 5.57x10-4 9.61x10-4 1.26x10-3

2.96x10-4 5.82x10-4 7.53x10-4 1.13x10-3 1.70x10-3

3.57x10-4 5.15x10-4 8.84x10-4 1.34x10-3 2.25x10-3

1.02x10-4 2.46x10-4 2.70x10-4 2.56x10-4 1.09x10-3

D i / D 2

1.98 1.98 2.26 2.23 2.32

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

n

0.51 0.54 0.54 0.66 0.66

0.97 1.01 0.96 0.97 0.97

1.06 0.98 0.96 0.97 0.96

1.10 0.99 0.95 0.96 1.06

1.15 1.14 1.12 1.20 1.20

Swelling Ratio

2.17 2.22 2.21 2.13 2.06

2.26 2.09 2.16 2.10 2.16

2.13 2.14 2.18 2.09 1.98

2.10 2.15 2.10 2.09 2.04

1.76 1.76 1.77 1.81 1.76

DT for 50% swelling, D2 for initial swelling.

103

Tnhle 6:

Swelling ratios of coals carbonised to varying temperatures

Coal

Rawdon

Coal

Barnburgh

Coal

Wearmouth

about their plast

Temperature

carbonisation

25

340

370

400

430

460

490

520

555

575

ic range

of

(°C)

Temperature of

carbonisation

25

370

390

410

435

455

480

500

530

555

Temperature

carbonisation

25

340

378

400

425

450

465

480

520

555

(°C)

of

(°C)

Swelling

Ratio (QV)

2.17

2.38

2.62

2.55

2.25

2.03

1.92

1.42

1.14

1.02

Swelling

Ratio (QV)

2.26

2.40

2.60

2.20 ■

1.78

1.75

1.60

1.40

1.12

1.04

Swelling

Ratio (QV)

2.10 Ί

2 2 1 1 1 1 1 1

15 27 26 88 41 53 88 13 00

104

Table 6 continued...

Coal

Line Creek

Coal

Woodside

Coal

CC ,

Temperature of carbonisation

25 370 390 415 440 465 490 515 540 560

(°C)

Temperature of carbonisation

25 375 405 435 460 485 515 535

(°C)

Temperature of carbonisation (CC)

25 330 350 375 395 415 440 460 485 510 530 555

Swelling Ratio (QV)

1.30 1 1 1 1 1 1 1 1 1

35 42 71 69 54 50 21 07 02

Swelling Ratio (QV)

1.04 1.86 1.51 1.5X 1.46 1.40 1.10 1.02

Swelling Ratio (QV)

1.02 1.06 1.36 1.63 1.48 1.14 1.44 1.45 1.47 1.41 1.12 1.02

105

Table 6 continued...

Coal

PD

Coal

Buchanan

Temperature of carbonisation

25 350 370 415 435 455 475 495 515 545

Temperature carbonisation

25 415 430 450 470 490 520

(°C)

ï of (°C)

Swelling Ratio (QV)

1.04 1.08 1.14 1.83 1.47 1.50 1.41 1.40 1.19 1.09

Swelling Ratio (QV)

1.09 1.29 1.66 1.84 1.84 1.53 1.43

Coal

Saraji

Temperature of carbonisation

25 385 400 420 440 460 480 500 530 555

(°C) Swelling

Ratio (QV)

1.00 1.04 1.10 1.33 1.63 1.62 1.34 1.37 1.14 1.02

106

Table 6 continued...

Coal

Oakgrove

Coal

K

Coal

Pinnacle

Temperature of carbonisation

25 400 420 445 470 490 515 540 565

(°C)

Temperature of carbonisation

25 440 460 480 500 520 540 560

Temperature carbonisation

25 450 470 490 510 530 555 585

(°C)

•of rC)

Swelling Ratio (QV)

1.01 1.05 1.13 1.34 1.20 1.32 1.30 1.32 1.03

Swelling Ratio (QV)

1.02 1.16 1.30 1.23 1.16 1.10 1.04 1.02

Swelling Ratio (QV)

1.01 1.15 1.36 1.22 1.23 1.22 1.17 1.01

107

Table 6 continued... Temperature of

Coal carbonisation

Lady Windsor 25 380 430 470 500 530 560 580 610

(°C) Swellir

Ratio (Ç

1.02 1.03 1.06 1.08 1.12 1.07 1.05 1.03 1.02

108

Table.7

Surface Areas. Pyridine Extract Yields. Swelling Ratios. Activation Energies and Pre-exponential Factors for Swelling

Coal Surface Extract Qv range Ea A (m2g-1) (wt%) 20-60°C (kJ mol-1) (s"1)

Rawdon

Barnburgh

Hucknall

Wearmouth

Wentz

129

105

111

83

68

17.3

19.0

25.4

17.8

2.06-2.22

2.09-2.26

1.98-2.18

2.04-2.15

1.76-1.81

31.6

33.9

33.8

37.5

44.9

2.3xl()2

2.48xl02

3.16xl()2

1.50x10?

9.40x103

109

Table 8

Rate Constants and Swelling Ratios for Raw and Extracted Wearmouth Coal in Pyridine. 2-chloropvridine and 2-fluoropvridine. Also Diffusional

Exponents for the Raw Coal in the Three Solvents

Temp / «C 'yridine

20 30 40 50 60

2-chloropyridine 20 30 40 50 60 70

1 Kis- 1 )

3.57xl0-4

5.15x10-» 8.84χ1(Η 1.34x10-3

2.25x10-'

-

7.67x10-s

1.25xl()-4

1.74x1 (H 2.45x1 (H 3.43x1 (H

Raw Coal n

1.10 0.99 0.95 0.96 1.06

-

0.86 0.86 0.88 0.91 0.93

Qv

2.10 2.15 2.10 2.09 2.04

-

1.64 1.62 1.61 1.69 1.70

Extracted Kis · 1 )

4.59xl()-3

4.86xl0-3

5.5xl()-3

6.12x10-3

7.5x10-3

3.26x10-3 3.56x10-3 4.01x10-3 5.09x10-3 5.23x10-3

_

Coal Qv

1.92 1.89 1.90 1.88 1.85

1.93 1.86 1.86 1.79 1.81

_

2-fluoropyridine 20 . . . 2.59x10-3 1.82 30 3.36xl0-5 0.63 1.27 3.27x10-3 1.70 40 5.16xl0-s 0.66 1.34 3.63x10-3 1.68 50 6.05xl(r5 0.70 1.41 4.34xl0-3 1.70 60 8.17xl()-5 0.76 1.41 4.90x10-3 1.65 70 1.07xl0-4 0.71 1.38

110

Table 9; Basicities, molar volumes, and amounts of solvents absorbed at 3QÜC. also apparent

activation energies and pre-exponential factors for the swelling of the raw and extracted Wearmouth coal in pyridine. 2-chloropvridine. and 2-nuoropvriiline

Solvent Properties Raw Wearmouth Coal

pKb Molar volume Amount absorbed Ea A QV

(cm'mol 1 ) (mmol g i) /kJ mol ' (s >) al 3()'C

Extracted Wearmouth Coal

Amount absorbed Ea A QV

mmol g ' /kJ mo l 1 (s1) at 30X

Pyridine 8.6 80.9 10.93 37.4 1494 2.15 8.46 9.7 0.97 1.89

2-chloropyridine 13.5 94.6 5.20 31.7 21 1.64 7.00 10.5 1.04 1.86

2-fluoropyridine 14.4 86. 2.41 24. 0.14 1.27 6.25 10.3 0.97 1.70

7.34x10-4

1.14x10-3

1.68x10-3

2.5x10-3

3.24x10-3

0.69

0.69

0.61

0.62

0.52

1.92

1.84

1.89

1.77

1.77

Table 10

Rate Constants. Diffusional Exponents and the Swelling Ratios for Swelling

of Wearmouth Coal in Butvlamine. Hexvlamine and Qctvlamine

Temp/. ( 'Q K n Qv

(S"1)

η-propylamine

30 1.89x10-3 0.67 1.77

n-butylamine

20

30

40

50

60

n-hexylamine

20

30

40

50

60

n-octylamine

30

40

50

60

70

n-decylamine

30 5.22xl0"6 1.10 1.82

1.98X1Í)-4

3.52x10-4

6.10x10-4

8.04xl0-4

1.10x10-3

1.1

1.1

1.1

1.1

1.0

2.18

2.16

2.19

2.08

1.90

2.58x10o

4.5 lx IO5

6.01 χ 10-s

1.08x104

1.29.x 1(H

1.1

0.97

1.10

1.1

1.1

2.08

2.10

2.17

1.97

1.93

112

Table Π

pKo and Molar Volumes of Amines. Swelling Ratios of Coal in each Amine and Molar Quantity of Amine Absorbed at 30°C. and the Activation Energy of the Swelling Process

Solvent

n-propylamine

n-butylamine

n-hexylamine

n-octylamine

n-decylamine

pKb

3.3

3.2

3.4

3.6

3.6

Molar Vol (cm3inol-

82.2

98.8

132.1

165.3

199.8

ume 1)

Swelling Wearmoi at 30°C

1.77

1.84

2.16

2.08

1.82

Ri ah

itio coal

Molar c]ii antity of solvent absorbed per gram (mmol g-

7.2

6.5

6.7

5.0

3.2

of coal

1)

Activation Energy for the swelling process Ea(kJmol-l)

nd

30.3

34.5

35.2

nd

n.d - not determined

Table 12

Rate Constants. Diffusional Exponents and Swelling Ratios in Pyridine at 2()°C for Various Particle Sizes of Raw Rawdon and Wearmouth Coals

Particle Size

(μ) K(s-i)

Rawdon Weannouth η Swelling Ratio K(s-l)

Qv Swelling ratio

QV

355-600

250-355

212-250

150-212

4.83xl()"4 0.5

8.98x10- 0.5

1.63xl()-3 0.51

1.72xl()-3 0.45

2.17

2.14

2.08

2.08

3.57xl()-4 1.10

6.19xl0-4

7.4xl()-4

.10

.02

8.77xl0"4 1.1

2.10

2.02

2.06

2.02

Table 13

Oxygen Contents and Swelling Ratios in Pyridine at 20°C for Raw and Oxidised Rawdon. Wearmouth. Buchanan and Pinnacle Coals

U1

Oxygen content wt% d.a.f.

Rawdon Wearmouth Buchanan Pinnacle Raw Oxidised Raw Oxidised Blank Raw Oxidised Raw Oxidised

13.9 21.28 8.23 16.71 9.08 3.72 9.20 4.25 5.58

Swelling Ratio in Pyridine at 20°C 2.17 2.37 2.10 2.33 2.12 1.09 1.87 1.01 1.81

Increase in Swelling Ratio 0.20 0.23 0.78 0.80

Table 14

Rate Constants. Diffusional Exponents and Swelling Ratios for Swelling of Oxidised Wearmouth Coal and its Blank in Pyridine

Rate Constant (ks-1) n 1st Stase 2nd S tace

Qv

Oxidised Wearmouth (355 - 600 μπι) Oxidised at 2(X)°C for 24 hours in an air oven 20 30 40 50 60

7.57 χ IO"6

1.65 χ I0-5

3.80 χ Κ)5

6.27 χ IO-5

1.60 χ Κ)-4

2.33 5.92 1.28 -.

0.96 0.76 0.77 0.65 0.56

2.33 2.34 2.26 2.38 2.32

Heated Wearmouth (blank) (355 - 600 μιη) Heat-treated at 200°C for 24 hours in vacuum

20 30 40 50 60

1.32 χ HH 2.41 χ 10-* 3.42 χ IO"4

7.46 χ IO"4

1.12 χ IO"3

0.91 0.92 0.97 1.1 1.2

2.12 2.24 2.19 2.13 2.15

Raw Wearmouth coal QV 2.10 2.15 2.10 2.09 2.06

116

Table 15:Showtng the areas of the various absorption bands in the FTIR

spectra of raw and oxidised Wearmouth

and Buchanan coals

Vibrational frequency

Wearmouth

Raw Oxidised

Buchanan

Raw Oxidised

Hydrogen bonded OH groups

3100-3600 cm" 1 8.14 10.66 7.47 9.82

Aromatic C-Η stretching

3050 cm-1 0.20 0.29 1.74 1.38

Aliphatic C-Η stretching

2830 cm-1

2853 cm"1 v(CH2)sym

2870 cm"1 v(CH3)sym

2890 cm"1 v(CH)aliph (tertian')

2923 cm"1 v(CH2)asym

2960 cm"1 v(CH3)asym

Total

Aliphatic / Aromatic (2923/3050)

0.20

0.22

1.38

0.39

1.27

0.91

4.37

6.35

0.06

0.07

0.50

0.15

0.39

0.35

1.52

1.34

0.25

1.10

0.64

0.45

1.37

0.27

4.08

0.78

0.12

0.25

0.13

0.23

0.50

0.20

1.43

0.36

Oxygen functional groups

1580 cm"1 carboxylate

1620 cm'1 hydroxy ketones

1635 cn r 1 hydroxy ketones

1653 cm"1 quiñones

1685 cm"1 conjugated ketones

1700 cm"1 carboxyl

1720 cm"1 esters or carboxyl

1730 cm"1 esters

Total

0.17

0.32

0.06

0.07

—

0.01

-

0.02

0.65

1.59

0.81

0.50

0.11

0.48

0.04

0.02

0.33

3.88

0.32

0.22

0.01

—

—

—

—

-

0.55

0.47

0.40

-

-

0.12

0.15

0.03

0.02

1.19

117

Tahle 16: Showing rate constants, oxygen contents, increase in oxviien contents and swelling ratios lor various particle sizes of oxidised Rawdon and Wearmoiilh coals

Oxidised Rawdon Wearmouth Coal Particle ks" · Oxygen % Increase in Swelling ralio k(s"') Oxygen % Increase in Swelling

Size Range 1st stage 2nd stage Content (wt%) oxygen content Ratio 1st stage 2nd stage content (wt%) oxygen content Ratio - (μπι) oo

355-60« 7.41x10-6 2.94x10-5 21.28 53 2.37 7.56x10-6 2.33x10-5 16.71 103 2.33

250-355 8.01x10-6 1.72xl0"5 23.95 72 2.27 1.61x10-5 6.01x10-5 17.74 115 2.30

212-250 1.40x10-5 1.66x10-5 24.04 73 2.25 2.27x10-5 6.18xl()-5 18.48 124 2.27

150-250 1.46x10-5 1.55x10-5 24.10 73 2.23 3.24x10-5 9.00x10-5 18.81 128 2.25

Table 17

Extents of Swelling at which the Second Stage begins and also depth of Penetration of Solvent at the same point

Ox. Rawdon Ox. Wearmouth

change over Depth of point (%) penetration

μιπ

55 56

81 64

90 62

150-212 89 45 93 53

Particle size range /μπι

355-600

250-355

212-250

change over point

69

86

89

(%) Depth of penetration μπι

77

72

59

119

Table 18

Percentage Pyridine Extraction Yields of Coals

Coal

Rawdon

Barnburgh

Hucknall

Pitts 8

Wearmouth

Wentz

Ruhrkohle

Line Creek

Buchanan

Saraji

Oakgrove

Oakdale

Norwich

Pinnacle

Ladv Windsor

Yield (%) Raw coal

17.3

19.0

25.4

23.3

17.8

8.5

6.6

2.1

2.0

1.5

2.0

1.0

1.4

2.2

Yield (%) Softening temp.

46.1

23.5

23.7

22.8

16.9

19.9

9.2

7.2

5.4

Yield (%) Max. contraction Temp

33.2

44.4

45.1

25.4

19.0

9.6

22.3

16.9

5.1

120

NJ

Coal

% Refi.

Rawdon Barnburgh Hucknall Weannouth Wentz Ruhrkholc Line Creek Buchanan Saraji Oakgrove Oakdale Norwich Lady Windsor

0.47 0.65 0.77 0.85 0.92 1.20 1.20 1.33 1.42 1.42 1.46 1.49 1.82

Table 19: Table showing the areas of the aromatic and aliphatic stretching frequencies

observed in the ΙΗΓΙΚ spectra of pyridine extracts of coals.

FHR band intensities / arbitrary units

Aliphatic

2830 cm"1 2853 cm-1

V(.CIb)Sym

Aromatic

3050 cm" I

V(C-H)arom

0.96

1.09

0.96

0.85

0.85

1.35

1.73

0.68

0.60

0.63

1.34

1.37

0.95

1.62(282'))

1.34(2824)

2.00(282'))

1.98(2830)

1.38(2828)

2.50(2831)

1.94(2832)

2.85(2832)

2.81(2830)

2.84(2836)

2.90(2832)

2.58(2835)

2.96(2834)

2.45(2852)

3.47(2855)

3.20(2856)

3.37(2854)

2.07(2851)

3.80(2853)

3.02(2856)

4.38(2853)

4.53(2852)

5.47(2854)

3.34(2853)

3.68(2853)

3.20(2852)

2870 c n r I

V(CH3)sym

1.24(2870)

0.09(2867)

0.66(2873)

0.86(2870)

1.43(2869)

0.80(2868)

0.15(2866)

1.57(2869)

1.51(2870)

1.40(2870)

0.15(2868)

1.26(2870)

1.70(2873)

2890 cnr I

v(CH)

3.54(2895)

4.50(2893)

2.25(2829)

2.24(2893)

2.63(2895)

4.78(2894)

4.32(2893)

5.14(2892)

5.42(2893)

6.07(2893)

6.03(2893)

3.72(2892)

3.12(2895)

2923 cm"1

V(CH2)asyni

4.94(2923)

5.68(2924)

5.52(2924)

6.17(2922)

5.57(2924)

4.96(2923)

4.80(2922)

13.86(2924)

14.43(2924)

16.45(2924)

8.95(2923)

9.90(2923)

9.93(2923)

2960 cm"1

v(CH3)a s y m

4.92(2963)

4.78(2955)

3.40(2958)

3.64(2954)

3.42(2957)

2.21(2954)

4.41(2953)

5.57(2958)

5.92(2960)

5.04(2957)

3.93(2953)

4.32(2954)

3.27(2957)

Coal % Refi.

Iahte 2Q; PvrcenlíUK' contents of the various stretching modes of CIH and Cll¿ groups in the total aliphatic stretching modes in the FTIR spectra of pyridine extract of coals.

Total 2830 cm-1 2853 cm"1 2870 cm"1 2890 cm"1 2923 cm"1

aliphatic unknown viClbisym v(CH3)sym v(CII) v(CH?)asym

Rawdon

Barnburgh

Hucknall

Wearmouth

M Wentz

Ruhrkohle

Line Creek

Buchanan

Saraji

Oakgrove

Oakdale

Norwich

Lady Windsor

0.47

0.65

0.77

0.85

0.92

1.20

1.20

1.33

1.42

1.42

1.46

1.49

1.82

18.71

19.86

17.03

18.26

16.50

18.84

18.64

33.37

34.62

37.27

25.30

25.46

24.18

8.66

6.75

11.74

10.84

8.36

13.27

10.40

8.54

8.12

7.62

11.46

10.13

12.24

12.34

17.47

18.79

18.45

12.55

20.17

16.20

13.12

13.08

14.68

13.20

14.45

13.23

6.62

0.00

3.88

4.71

8.67

4.25

0.01

4.70

4.36

3.76

0.01

4.95

7.03

18.92

22.66

13.21

12.27

15.94

25.37

23.17

15.40

15.65

16.29

23.83

14.61

12.90

26.40

33.35

32.41

33.79

33.76

26.35

25.75

41.53

41.68

44.14

35.38

38.88

41.07

2960 cm"1

V(CH3)asym

26.30

24.07

19.96

19.93

20.73

11.73

23.66

16.69

17.10

13.52

15.49

16.97

13.52

rvj

Table 21: Table showing the areas of the aromatic out of plane deformation bands in the

FTIR spectra of the pyridine extracts of coal.

Infra red absorption peaks

Coal

Rawdon

Barnburgh

Hucknall

Wearmouth

Wentz

Ruhrkhole

Line Creek

Buchanan

Saraji

Oakgrove

Oakdale

Norwich

Lady Windsor

% Rell

0.45

0.65

0.77

0.85

0.92

1.20

1.20

1.33

1.42

1.42

1.46

1.49

1.82

750 cnr 1

0.70

0.56

1.38

0.59

0.72

1.54

1.84

0.51

0.72

0.75

1.77

1.71

1.44

770cnr '

0.94

1.41

0.43

1.48

1.04

1.97

0.43

0.31

0.64

0.22

0.33

0.69

0.37

801cm"1

1.12

1.26

4.45

0.7')

0.17

--

2.05

-

-

0.20

1.19

1.04

2.91

816 cm"1

0.19

0.89

-

I l l

0.79

1.24

1.08

0.87

0.72

0.57

0.93

0.75

—

860 cnr >

0.23

1.89

1.36

1.07

0.71

1.33

0.90

--

--

0.19

0.45

0.32

0.35

874 cm"1

~

-

-

-

-

0.95

2.09

0.38

0.27

0.27

1.00

0.84

0.23

888 cm"1

0.85

0.87

0.55

1.09

0.77

0.39

0.31

0.10

0.03

0.07

0.53

0.40

0.23

Total

4.03

6.88

8.17

6.16

4.20

7.42

8.70

2.17

2.38

2.27

6.20

5.75

5.53

Table 22: Table showing the areas of the bands for the oxygen functional groups in the FTIR soectra of pyridine extract of coals after standardisation using the areas

of the 1600 cm-1 band as an internal standard.

4^

Coal Rawdon Barnburgh Hucknall Wearmouth Wentz Ruhrkhole Line Creek Buchanan Saraji Oakgrove Oakdale Norwich Lady Windsor

% Reflectane 0.47 0.65 0.77 0.85 0.92 1.20 1.20 1.33 1.42 1.42 1.46 1.59 1.82

1655 c 0.11(1 0.12(1

0.17(1

0.22(1

0.25(1

0.35(1

0.40(1

0.63(1 0.74(1 0.48(1 0.63(1

0.30(1

0.48(1

n-1 54) 54)

50) 53) 55) 56) 58) 54) 52) 54) 59) 52) 56)

1700 cm-1 0.26(1703) 0.18(17.02)

0.20(1698) 0.40(1701) 0.18(1698) 0.07(1703) 0.81(1700) 0.15(1697) 0.42(1703) 0.09(1705) 0.24(1703) 0.38(1703)

1720cnr' --0.08(1722) ---0.01(1725) 0.52(1722) 1.37(1721) -

0.08(1730) 0.06(1721) 0.38(1721)

Extract /% 17.3 19.0 25.4 23.3 17.8 8.5 6.6 2.1 2.0

1.5 2.0 1.0 2.2

01% 12.2 10.5 9.6 7.5 7.3 6.1 5.1 5.0 4.7 5.4 5.4 6.7 5.7

Table 23: Areas of Aromatic and Aliphatic Stretching frequencies in the FTIR spectra of extracts of coals heated to their temperature of maximum contraction (arbitrary units)

3()50(cm1) 2830(cm·) 2855(αιι') 2875(cm ') 290()(cnrl) 2928(cm') 2928(cm1) 2960(cm ·) 296()(cm-') (C-ll) (aliph/arom) (aliph/arom)

aromatic

Coal

tn Rawdon Wearmouth Pittsburgh 8 Wentz Woodside Buchanan CC Saraji Oakgrove K Pinnacle

0.54 0.72 0.77 0.76 1.16 0.70 1.13 0.79 0.60 1.17 1.05

2.68 1.50 3.35 1.50 1.21 0.58 1.23 1.69 0.77 2.25 2.25

4.07 0.75 0.89 0.75 I.I 1 1.25 0.69 0.96 0.32 0.53 0.79

0.18 2.4 3 1.12 2.43 1.79 0.30 0.94 3.62 2.35 1.41 4.17

5.52 1.44 2.36 1.44 1.57 0.81

-

1.21 0.44

-

0.64

7.77 2.30

--

2.30 2.01 1.44 1.99 5.09 1.8.3 2.90 4.49

14.39 3.19

-

3.02 1.73 2.06 1.76 6.44 3.05 2.48 4.43

4.55 2.20 1.99 2.20 2.38 1.04 1.51 2.54 1.17 1.88 2.58

8.43 3.06 2.58 2.89 2.05 1.49 1.34 3.22 1.95 1.60 2.46

Table 23 continued: Areas of the out of plane deformation deformation hands in the FTIR snectra of extracts of coals heated to their temperature of maximum contraction (arbitrary units)

690(cm1) 72()(cm') 74()(cm ') 765(cm ') 780(cnri) 8(X)(cm>) 830(cm·) 86()(cm>) 885(cm->)

Coal

Rawdon an Weannouth

Pittsburgh 8 1.08 Wentz Woodside Buchanan CC 0.71 Saraji Oakgrove K 0.54 Pinnacle

0.28 --

0.10 ~

~ 0.25 0.62

--—

0.50

0.20 -

0.72 --

----

2.05 0.10 1.14 2.00

—

0.56 1.28 0.55 0.5 1 1.02 1.02

--

1.34 0.50

-

2.30

----

1.43 1.75 0.76

--

-0.42

-—

0.72 0.42 1.68 0.17 0.01

--

3.44 0.76 1.02 2.81

—

1.47 1.50

--

1.28 1.89 0.82

--

0.78 0.54 0.12 2.72

0.12 0.30 1.56 0.42 0.41 0.30 1.71 0.78 1.30 1.18 0.19

1.16 1.26

--

1.25 1.96 1.05

--

1.20 ----

1.53

Table 24

Oxvgen Contents of Coal. Extracts, and Residues (wt% daH

Rawdon

Barnburgh

Hucknall

Wearmouth

Wentz

Ruhrkohle

Line Creek

Buchanan

Saraji

Oakgrove

Oakdale

Norwich

Pinnacle

Ladv Windsor

Oxygen Content of Coal

13.9

10.7

9.8

8.2

6.4

6.3

5.1

3.7

4.2

4.4

5.1

4.0

4.2

3.6

Oxygen Content of Extract

12.2

10.5

9.6

7.5

7.3

6.1

5.1

5.0

4.7

5.4

5.4

6.7

4.5

5.7

Oxygen' of Resid

13.7

12.2

9.8

8.9

7.8

6.8

5.6

3.7

3.5

4.1

4.0

3.8

3.5

2.7

127

Table 25: Values of Oxvgen content of the coal residues and the coal extracts for samples carbonised to Tl and T2.

Coal %Oxygen of coal %Oxygen of coal %Oxygen of coal %Oxygen of coal residue to Tl extract to Tl residue to T2 extract to T2

Rawdon 14.59

Pittsburgh 8 11.96

Wearmouth

Wentz

8.70

7.76

6.64

8.86

8.15

5.33

12.35

8.85

8.46

7.47

8.85

7.54

6.30

6.64

Ruhrkohl 5.73

Woodside

Buchanan

5.43

4.68

4.31

5.10

C. C.

Saraji

Oaksrrove

K.

Pinnacle

Ladv Windsor 3.94

3.94

4.67

-

4.29

5.71

-

-

4.63

5.31

6.10

5.54

4.90

4.90

4.99

5.00

4.20

5.20

7.44

4.73

4.03

5.70

3.37

128

Table. 26: Values of Retention volume and normalised Area under the curve of coal extracts of raw coal, coal carbonised to T; and coal carbonised to T2 with tetrahydrofuran as the mobile phase.

Retention volume (cm3) of raw coal extracts

Nonnalised area of raw coal extracts

OlllTr/nig)

Retention volume (cm3) of Τ| extracts

Normalised area ol"T| extracts

(mm2/mg)

Retention volume (cm3) of Ί'2 extracts

Normalised area of T2 extracts

(mtmVmg)

Rawdon Barnburgh Hucknall Pittsburgh 8

-* Weannouth 10 Wentz

Ruhrkohl Line Creek Woodside Buchanan C C . Saraji Oakgrove Oakdale K. Norwich Pinnacle Lady Windsor

15.9 15.5 16.1

--

15.3 15.5 15.0 16.1

--16.4 15.3 18.4 16.9 16.3 15.6 16.5 16.4 16.9

765 515 456

1009 945 775 1168

791 46 662 504 619 241 172 442 1166

17.

15.2 15.4 17.7 15.8

15.3 16.5

16.5 17.1

122

71 I 613 219 607

314 493

1286 1588

17.0 222

15.1

14.2 14.6 15.0

15.6 16.2 15.9 15.9 15.9

16.2

16.4 18.4

500

929 671 980

578 1119 1684 1782 1642

1549

1220 115

Table. 27: Values of Retention volume and normalised Area under the curve of coal extracts of raw coal-

coal carbonised to T¿ and coal carbonised to T2 with chloroform as the mobile phase.

Retention volume (cm.3)

of raw coal extracts

Normalised area

of raw coal extracts

(mm2/mg)

Retention volume (cm.3)

of Τ| extracts

Normalised area

of Τ| extracts

(mm2/mg)

Retention volume (cm3)

of T2 extracts

Normalised area

of T2 extracts

(mm2/mg)

Rawdon

Barnburgh

Hucknall

Pittsburgh 8

_, Wearmouth 0 Wentz

Ruhrkohl

Line Creek

Woodside

Buchanan

C C .

Saraji

Oakgrove

Oakdale

K.

Norwich

Pinnacle

Lady Windsor

16.46

15.75

16.13

-

16.27

16.58

16.35

16.77

-

16.46

16.95

16.73

16.91

16.43

16.80

-

16.8

16.7

20

20

4X

25

68

102

96

38

26

155

52

172

36

68

39

5.90

16.01

16.24

15.76

16.06

16.17

16.28

16.24

16.58

30

48

25

50

74

42

65

165

239

5.61 45

15.54

15.66

15.86

15.69

16.50

16.11

16.35

16.15

16.21

16.47

16.24

17.59

90

53

56

9 S

204

124

112

206

134

198

76

27

Heating Rate 5°C/mTn Ti

T2

T3

T4

R(mg/min)

Table 28

Thermogravimetric Data

Rawdon

Heating Rate 10°C /min

τ, To

τ3 T4

R(mg/min)

Heating Rate 15°C /min Τι T-,

τ; T4 R(mg/min)

Heating Rate /min Τι To

τ: T¡ R(mg/min)

Heating Rate /min

Τι To

T3" T4

T(mg/min)

Ti To

T~,

T4

R

20X

25°C

=

=

= —

304 424 354 120 0.32

298 433 357 135 0.76

304 438 368 134 1.04

308 451 379 143 1.51

313 439 381 126 1.85

Wearmouth

319 429 366 110 0.41

321 437 382 116 0.72

331 465 402 134 0.99

329 481 405 152 1.37

337 474 418 137 1.78

on Çoals

Buchanan

371 466 432 95 0.14

372 485 425 113 0.42

367 518 445 151 0.47

372 538 447 166 0.70

369 577 451 208 0.97

Oakdale

389 516 416 127 0.12

349 493 425 144 0.31

367 526 442 158 0.43

363 535 447 172 0.54

361 554 461 193 0.77

Temperature of commencement of volatile release Temperature of end of volatile release

Temperature at the

Temperature range

Lady Windsor

398 532 430 134 0.08

366 544 430 177 0.23

387 559 459 172 Ò.27

388 554 465 166 0.37

391 594 469 203 0.50

maximum rate of volatile release.

of volatile loss. Rate of volatile loss (mg/min).

131

Table 29

Comparison of Temperature of initial Weight Loss and Rate of Weight Loss for Raw and Extracted Coals at at

Heating Rate of 5°C min1

Raw Coal Extracted Coal Temp, of initial Soft. Rate of Weight Temp, of initial Rate of weight weight loss Temp. loss mg/min weight loss loss mg/min CO CC) CC)

334 0.70

343 0.62

355 0.60

355 0.60

363 0.62

368 0.50

371 0.52

390 0.42

383 0.36

390 0.36

390 0.30

389 0.29

389 0.29

Lady Windsor 398 438 0.08 405 0.24

*n.a = not available

Rawdon

Barnburgh

Hucknall

Wearmouth

Wentz

Ruhrkohle

Line Creek

Buchanan

Saraji

Oakgrove

Oakdale

Norwich

Pinnacle

304

308

312

319

322

367

365

371

390

387

389

390

390

339

n.a*

n.a*

345

351

375

n.a*

398

410

412

-

426

410

438

0.32

0.58

0.52

0.52

0.41

0.35

0.40

0.14

0.22

0.18

0.12

0.18

0.18

0.08

132

Table 30: Dilatometrie data of coals used in this studv

Rn Wh Wz Rk Bn Si Lw

Tl 339 345 351 375 398 410 438

T2 - 419 420 440 448 457

T3 525 450 458 470 490 489

%c 19 27 24 20 23 15 14*

%d - 46 88 45 71 40

Contraction measured at 500°C

Table 31: Permeability data of coals used in this studv

Rn Wh Wz Rk Bn Si Lw

Ta 360 340 360 385 402 420 445

Tb - 412 414 440 449 460

Tc 450 440 440 446 459 477 490

Td 480 485 545 530 - 530 490

133

Table 32: Thermoplastic data of coals carbonised under varying gas pressure

Rawdon coal

Pressure (bar) %C %D T1(°Q T2(°C) T3(°C)

357 -- 530 355 530 337 -- 450 340 - 475 325 - 425

Wearmouth coal

0 15 25 40 60

24 26 34 33 32

Pressure (bar)

0 5 10 15 20 30 40 50 60

%C

13 8 13 13 16 16 19 22 25

%D

51 87 89 101 87 79 76 63 55

T1(°C)

375 375 365 362 362 350 330 325 320

T2(°C)

424 419 415 417 416 420 425 418 420

T3(°C

537 537 525 550 540 550 550 550 550

Pittsburgh 8 coal

Pressure (bar) %C %D T1(°C) T2(°C) T3(°C)

0 5 10 15 20 30 40 50 60

10 21 22 27 25 26 24 22 18

176 167 163 148 143 128 109 97 79

395 365 375 355 345 322 320 320 340

420 410 404 403 410 390 404 400 400

492 490 485 463 471 455 495 468 475

134

Table 32: continued...

Wentz coal

Pressure (bar) %C %D T1(°C) T2(°C) T3(°C)

0 20 40 60

Line Creek coal

25 16 25 25

163 125 91

350 337 329 321

418 416 410 408

475 475 487 475

Pressure (bar) %C %D T1(°C) T2(°C) T3(CC)

0 10 25 45 60

Woodside coal

13 12 13 14 14

-1 2 5 8

385 389 375 360 358

— 470 462 462 460

500 500 500 505 500

Pressure (bar) %C %Ό T1(°C) T2(°C) T3(°C)

0 10 20 30 45 60

25 26 28 25 25 24

206 157 126 109 55 48

352 349 342 325 325 310

418 417 411 403 402 399

476 482 490 475 502 500

PD coal

Pressure (bar) %C <7cD T1(°C) T2(°C) T3(°C)

0 20 40 60

16 15 18 16

43 45 34 30

388 375 370 350

447 435 447 430

483 500 512 500

135

Table 32: continued...

Buchanan coal

Pressure (bar)

0 10 15 20 25 30 40 50 60

C C coal

Pressure (bar)

0 10 20 30 40 50 60

Saraji coal

%C

24 20 20 16 16 24 25 25 24

%C

23 25 25 25 28 26 29

%D

23 72 85 81 86 73 76 70 69

9rD

53 61 51 44 40 36 28

T1(°C)

405 400 393 397 387 383 387 377 375

T1(°C)

375 370 368 350 350 337 337

T2(°C)

469 439 435 450 434 439 436 434 435

T2(°C)

446 435 435 430 430 425 435

T3(°C)

506 492 502 520 520 502 513 508 512

T3(°C)

495 485 488 475 483 491 500

Pressure (bar) %C %Ό T1(°C) T2(°C) T3(°C)

0 5 10 15 20 30 40 50 60

15 18 13 17 20 14 18 23 24

57 74 83 85 76 74 59 62 54

400 400 387 393 384 377 375 372 355

449 439 434 437 437 433 430 425 428

502 500 500 503 503 498 500 510 525

136

Table 32: continued...

Oakgrove coal

Pressure (bar) •

0 10 20 30 40 50 60

%C

20 19 24 21 27 21 24

%D

112 115 105 95 86 85 75

T1(°C)

399 383 383 383 375 373 365

T2(°C)

435 434 430 425 420 422 420

T3(°C)

495 510 505 505 510 512 522

K coal

Pressure (bar)

0 10 20 30 40 60

Pinnacle coal

%C

29 16 29 15 22 29

%D

13 68 71 74 62 58

T1(°C)

407 407 400 405 400 380

T2(°C)

477 459 465 462 463 457

T3(°C

512 513 513 518 513 513

Pressure (bar) %C %D T1(°C) T2(°C) T3(°C)

0 5 10 15 20 30 40 50 60 75

20 17 25 19 25 17 22 21 23 25

10 50 48 80 58 65 64 68 67 57

430 430 425 410 403 406 403 400 390 385

487 485 462 455 458 456 453 448 447 440

525 524 525 525 524 525 525 525 525 525

137

Table 32: continued...

Lady Windsor coal

Pressure (bar) %C %D T1(°C) T2(°C) T3(°C)

442 420 410 410

0 20 40 60

19 25 16 26

138

Table 33

Surface Areas (C02 . 273 K) of Coals, and of their Semi-cokes produced at 450. 500. 600 and 800°C

Coal Surface area/m2^-1

450°C 500°C 600°C 800°C

103 102 214 265

120 140 212 247

87 141 182 239

93 117 166 146

80 81 160 132

72 85 163 132

75 90 165 130

61 77 140 55

58 79 130 57

50 62 122 107

32 114 142 138

64 67 138 110

52 60 138 60

64 61 143 83

Rawdon

Barnburgh

Hucknall

Wearmouth

Wentz

Ruhrkohle

Line Creek

Buchanan

Saraji

Oakgrove

Oakdale

Norwich

Pinnacle

Ladv Windsor

Raw coal

129

110

111

113

68

51

69

46

42

53

57

66

77

75

139

Table 34

Oxygen Capacities of Semi-cokes (cc/g) produced at 450°C. 500°C. 600°C and 800°C

Coal 450'C 500*C 600°C 800'C

Rawdon

Barnburgh

Hucknall

Wearmouth

Wentz

Ruhrkohle

Line Creek

Buchanan

Saraji

Oakgrove

Oakdale

Norwich

Pinnacle

Lady Windsor

1.01

2.01

0.90

0.39

0.02

-

-

-

-

-

-

-

-

1.61

1.52

1.39

1.28

0.94

0.34

0.34

0.16

0.28

0.02

0.79

0.06

-

0.06

2.73

1.36

2.64

2.53

2.38

1.75

1.35

1.01

1.09

1.07

1.26

1.08

0.95

1.36

6.19

4.97

4.77

2.27

0.58

0.23

0.28

0.15

0.26

1.03

1.08

1.39

-

1.22

All semi-cokes produced at 1000° C did not take up oxygen. Also semi-cokes produced at 450°C from Ruhrkohle to Lady Windsor did not take up oxygen.

140

Table 35

Nitrogen Capacities of Semi-cokes produced at 600°C

Coal Nitrogen Capacity (cc / g)

Rawdon 1.96

Barnburgh 1.13

Hucknall 1.40

Wearmouth 1.17

Wentz 0.68

Ruhrkohle 0.65

Line Creek 0.72

Buchanan 0.00

Saraji 0.26

Oakgrove 0.14

Oakdale 0.30

Norwich 0.12

Pinnacle 0.00

Ladv Windsor 0.95

Semi-cokes produced at 450. 500. 800 and 1000°C did not take up any nitrogen.

141

Table 36

Diffusional Parameters (xl(H) for Oxygen into Semi-cokes produced from Coals at 450. 500. 600 and 80fl°C

Diffusional parameters (D/r2 χ 10^) / s~l Coal 450°C 500°C 600°C 800°C

Rawdon

Barnburgh

Hucknall

Wearmouth

Wentz

Ruhrkohle

Line Creek

Buchanan

Saraji

Oakgrove

Oakdale

Norwich

Pinnacle

Ladv Windsor

2.34

0.48

0.39

0.19

0.34

-

-

-

-

-

-

-

-

_

0.98

2.35

3.42

1.88

2.01

1.69

1.35

0.91

1.23

1.28

1.39

1.21

-

5.03

9.15

16.5

8.42

7.00

6.15

6.42

5.42

4.53

1.33

3.18

4.75

4.01

3.84

5.31

3.32

3.26

1.40

0.34

0.30

1.03

1.80

0.42

0.99

0.99

3.47

1.55

-

4.20

142

Table 37

Diffusional Parameters (xlQ-4) for Nitrogen into Semi-cokes produced from Coals at 600°C

Coal Diffusional parameters (Ρίχΐ χ IO-4) / s'

Rawdon 4.11 Barnburgh 8.44 Hucknall 4.76 Wearmouth 1.37 Wentz 1.22 Ruhrkohle 1.27 Line Creek 1.24 Buchanan 0.00 Saraji 0.86 Oakgrove 1.09 Oakdale 4.08 Norwich 1-05 Pinnacle 0.00 Ladv Windsor 1.11

143

TO TRANSDUCER AND COMPUTER

Thermometer

Sinter

Water

Solvent

Fig.l: Schematic diagram of the Dynamic Volumetric swelling Apparatus.

144

CJ

I

0.00

Fl·' "»·

0.10 0.20

Time, DIT2 t / a

2

Theoretical Fickian curve

0.30

• 0 . 63

■ 0 . 87

■ 0 .95

_ (30

"e

. 0 . 9 8

o

I

5

0.00 0.20 0.40

Time, k o t / C 0 a

0.60

■ 0 .63

- 0 .87

= 0 .95

J- 0 . 98

_ CO

Fi«.3: Theoretical Case 11 curve

145

J >

f Solvent Λ V Reservoir J

I H.P.L.C. 1 Pump

S

Column

^ SAMPLE Λ ^ J* .

Ν

Plotter -

UV Detector

, s SAMPLE 1 1 l U W I I I L I C I 1

OUT

Fig. 4: Schematic ((ingrani of the s.e.c apparatus.

Piiani Vacuum Gauge

j >

Rotary Pump

.·■ 7 ; .· ■■■ — \ S N \ ' —

/ / / / • \ \ \ \ -— ■ ■ / . ' / S — .

S \ \ S < / / / / / \ \ \ \ ' V ( < ( (

Adsorption Tubes

Diffusion Pump Mercury Manometer

Fig. 5: McBatn spring balance.

j^-y^

To now meters Micro balance head

Gas to be adsorbed or helium

Helium Puree flow

Water outlet

Helium or gas to be adsorbed

Heater

Sinter

Gas inlet

Fig.6: Apparatus for measuring gas uptake by gravimetric method.

148

SO

Ξ 0.30

- 0.26 ι ' r

40 60 80

n Oxygen

• Nitrofen

τ — ■ — τ — ' — τ

100 120 140 160

Weight of sample in helium

Fig.7: Buoyancy correction graphs for oxygen and nitrogen in helium

149

a u zc

ti •ï

'S!

100 200 time (mins)

400

Fig.8: Variation of swelling ratio with time for the swelling of Rawdon coal in pyridine at 2()°C.

150

0

u

ti ζ

ã

¿.o ­

2 . 0 ­

1.5­

1.0­

B

1 τ­

Ώ

aB

a

I

i

, E f f i a — □

0.0 0.5 1.0 1.5

Reflectance of coal (%)

2.0

Fig.9: Variat ion of swelling ratio wi th coal rank for swelling of coal in pyridine at 20

eC.

51

¿.¿ ­

2.0 ­

1.8­

1.6­

1.4­

1.2­

1.0­

D

> Γ

□

α □

Q

ι

D

Q B ■ 1

□

0.0 0.5 1 .0 1 .5 2.0

Reflectance of coal (%)

Fig.lO: Var iat ion of swelling ratio wi th coal rank for the swelling of coal in pyridine at 20

eC.

151

0>

tn ι

Qi C/}

time (mins) Fig.11: Variation of ln((Se-St)/Se) with time for the swelling of Rawdon coal in pyridine at 20°C (Fickian process ).

tn

tn

Vt

Fig.l2: Variation of Se/St with square root of time for the swelling of Rawdon coal in pyridine at 20eC. ( A Fickian process)

152

tn

tn ι

ZJ

tn

O 10

t ime (mins)

20

Fig.l3: Variation of ln((Se-St)/Se) with time for the swelling of Rawdon coal in pyridine at 50eC.( Anomalous process)

_ -1 -«j

C/2

tn I

O tn

0 20 40 60 80 100

time (mins)

Fig.l4: Variation of ln((Se-St)/Se) with time for the swelling of Wearmouth coal in pyridine at 30eC (Case 11 process).

153

α In k(R)

♦ In k(B)

α In k(H)

o In k(We)

■ In k(Wn)

0.0029 0.0031 0.00 . 1 J 0.0035

ι/τ

Fig.15: Arrhenius plots for the swelling of Rawdon, Barnburgh, Hucknall, Wearmouth and Wentz coals in pyridine.

154

D pyridine

♦ 2-chloropyridine

■ 2-fluoropyridine

100 200 3 0 0

time (mins) Fig.16: Variation of swelling ratio with time for Wearmouth coal in pyridine, 2-chloropyridine, and 2-fluoropvridine.

tn

en I

tn

time (mins)

Fig.l7: Variation of In((Se-St)/Se) with time for the swelling of Wearmouth coal in 2-chloropyridine at 70°C.

155

- 0 . 5 -

tn

tn ■

tn

800

time (mins)

Fig.l8: Variation of ln((Se-St)/Se) with time for the swelling of Wearmouth coal in 2-fluoropyridine at 40° C.

tn

tn ι

o tn

time (mins)

Fig.19: Variation of ln((Se-St)/Se) with time for the swelling of extracted Wearmouth coal in 2-fluoropyridine at 30°C.

156

-C

s

D Pyridine

♦ 2-chloropyridine

■ 2-fluoropyridine

τ ' 1 r

0.0029 0.0030 0.0031 0.0032 0.0033 0.0034 0.0035

l/T

Fig.20: Arrhenius plots for the swelling of Wearmouth coal in pyridine, 2-chloropyridine, and 2-fluoropyridine.

tn

tn I

tn

time (mins)

Fig.21: Variation of ln((Se-St)/Se) with time for the swelling of Wearmouth coal in n-hexylamine at 30eC.

157

υ tn

tn ι

tn

1000 2000 3000

time (mins) Fig.22: Variation of ln((Se-St)/SE) with time for the swelling of oxidised Wearmouth coal in pyridine at 20°C.

tn

tn I

SJ

tn

450

time (mins) Fig.23: Variation of ln((Se-St)/Se) with time for the swelling of oxidised Buchanan coal in pyridine at 20 °C.

158

tn

tn ι

ZJ

tn

time (mins)

Fig.24: Variation of ln((Se-St)/Se) with time for the swelling of Wearmouth coal heat-treated for 24hrs at 200eC in vacuum (blank).

159

s

Ü . 4 1 Í »

0 . 3 9 0

untici ε

MAUEMJnBEfiS 408.θ

Fig.25: FTIR spectra of Wearmouth coal

σι

α.a im

α.340

0 . 3 2 0

0 . 3 0 0 ­

snnPi ε

S ¡o a:

s!

0 . 2 0 0 ­

0 . 2 4 0 ­

0 . 2 2 0 ­

0 .1977

•IlìQU τ 1 r 20U0

MAUEinmBERS

lOUQ 400.0

Fig.26: FTIR spcclra of oxidised VVearmoulh coal

6íHiPi ε

σι ISJ

ø.eø-

0 . 5 5 -

a . S ä ­

lj 0.45­s

I « 0 . 4 0 -

0 . 3 5 -

0 . 3 0 -

Ö.gC 0.192 i 1 1 Γ Ί , , {

4000 3QU0 - j Ι Γ euua

UAUEHUnBERS

IOU0 400 .θ

Fig.27: FTÏR spectra of liiiclinnnn coni

σι UU

0 . 4 2 0 - r

0.392

0 . 3 8 4 ­

0 . 3 3 5 -

0 . 3 0 7 -

εηπριε

o 0 . 2 7 9 ­

0 . 2 5 0 ­

0.222

0 . 1 9 4 ­

Ø. IBS­

4000 3UUB UAUEUUflflERS 400.0

Fig.28: FTÏR spectra of oxidised Buchanan coal

σι

0 . 0 7 2 7 ­

0 . 0 7 0

Ø.BGO­

0 . 0 5 0 ­

0 . 0 4 0 ­

HESIII Τ

4QØQ

Fig.29: nifferenco spectra for oxidised nnd raw Wenrmouth coal

σι ui

β.0(131 I

0 . 0 6 0

0 . 0 5 0 ­

0 . 0 4 0 ­

IILÍitH Τ

0 . 0 3 0

0 . 0 2 0 ­

0 .010

UAUEHUriBERS

Fig.30: Difference speclra for oxidised and raw Buchanan coal

400 .0

α tn

tn ι

tn

1000 2000

time

Fig.31: Variation of ln((Se-St)/Se) with time for the swelling of oxidised Wearmouth coal of particle size range 212-250u.m in pyridine at 20°C.

166

χ.

*9

JU ­

2 0 ­

ïo­

0 1

Q

■ ■ Γ­

α

a

G D

ι

α

□

— ι 1

α

0.0 1.0 1.5 2.0 0.5

% Reflectance of coal

Fig.32: Variation of percentage extraction with coal reflectance.

JU ­

2 0 ­

10 ­

o­

□ D

ι 1 1

Β

"I

Β

Β

1 o 20 30 40

Volatile matter ( % d.a.f).

Fig.33: Variation of extract yield with volatile matter.

167

*

u

log ddpm

Fi».34: Variation of extraction vield with coal fluiditv.

168

σι io

1 I I ι ι ρ l i l t i . la

Norwich

Lady Windsor

Fitt.35: FTIR spcclm (1700-3600 cm" 1 ) of pyridine evlrncls of coals.

VJ

o

Oakgrove

Oakdale

Norwich

Lady Windsor I * ' 1 1 1 ' ' ' r · ' · 1 ' τ ιΛα ι ( . ο ι . t o ι Λ ο ι ο ο ο

M a v ì ι» ».ι)ι IIS

Fig.36: FTIR speclra (10(1(1­20(10 cm'1 ) of pyridine extracts of coals.

<£

sn Í3

■s: CS

C/5

Γ^

C.

<

Β % CH2

♦ %CH3

2.0

Reflectance of coal (%)

Fig.37: Variation of (CH2)asym and (CH3)asym in the aliphatic stretching bands of the FTIR spectra of pyr idine extract of coals with coal rank.

J U ­

20 ­

10 ­

0 ­

■

1 —■ Γ­

α π

α '

* ♦ ♦ ♦

■ ■ τ

D

α

□

•

e * •

ι

α

♦

Β CH2/CHar

♦ CH3/CHar

0.0 0.5 1.5 2.0

Reflectance of coal (%)

Fig.38: Variation of the ratios of CH2 and CH3 asymmetric bands to the aromatic CH stretching band wi th rank of parent coal.

171

Rawdon

Barnburgh

Hucknall

Wearmouth

Wentz

7£a aoo í £ 0 sea 9 3

­­ι

7Ξ2 SCO B5Q

1

900

Ruhrkhole

Line Creek

ÇÇQ

wavenum herstem )

Fig.39: The 700-900 cm-1 region of the FTIR spectra of pyridine extracts of coals.

172

700

•«i

750

750

SCO

8C3

BSO

SSO

wavenumbers(cm -1 )

173

Buchanan

Saraji

Oakqrove

sea

Oakdale

Norwich

300 950

SOO

Lady Windsor

950

CS

.2

IT)

TZ

U.Ö ­

0.6 ­

0.4 ­

0.2 ­

Β

' 1

El

□

1

□ Β

1

Β

Β □

□

α

ι

D

■

0.0 0.5 1.0 1 .5 2.0

Reflectance of coal (%)

Fig.40: Variation of the area of the 1655 cm-1 band with rank of parent coal.

174

<£ —

£ 13 s

o —

s CU

o υ c o

14 -

'

1 2 -

"

10 -

"

8 -

" 6 -

.

4 -ZI

χ

C 2 0.0 2.0 0.5 1.0 1.5

Reflectance of coal (%)

Fig.41: Variation of oxygen content with reflectance of coal.

~ 14

!/5

il

X Zi

S

o Î X >> χ

I 4 ­

12 -

10 -

8 -

6 -

*+ ι

Β

I

O

□

QB

1 1

Β

□ QQ

B Q 0 Β

r­­ —ι

Β

1

0.0 0.5 1.0 1.5 2.0

Reflectance of coal (%)

Fig.42: Variation of oxygen content of extracts with reflectance of coal.

175

s

V3 CU

o u

C eu

c o υ

s cu C£ >-> X

1 4 -

1 2 -

1 0 -

8 -

6 -

4 -

2 -

B

' I

□

QB

□

' r-

B

B

° B ^ B

■ 1

B

c o.o 0.5 2.0

Reflectance of coal (%)

Fig.43: Variation of oxygen content of residues with reflectance of parent coal.

- 1 4 -

- 1 2 -%i tz

8 -

6 -

4 -

2 -

0

cu

c o u

c o CX) >> χ

-" r 2

Β

Β Ξ

ΕΒ

π — ' 1 ' 1 ' 1 <—

6 8 10 12 14

Oxygen content of coal (wt %).

Fig.44: Variation of oxygen contents of extracts with those of coals.

176

«5 O cu

CU Zl

>-> χ

c

Γ5

X Zi

Zi

zi χ

o

ó ­

2­

1 -

ο­

ι ­

2 ­

Β

■>■­■ ' ι ■

α α

Β

■ Ι

σ

□ Β

Β Β Β

α □

■■ ■ ' ι "■

α

0.0 0.5 1.0 1.5 2.0

~ Reflectance of coal (%)

Fig.45: Variation of differences in oxygen contents of extracts and coals with percentage reflectance of coals.

177

o cu

cä « u

cu — CU C. s

£ S ­

eu ν Π 8 C

O g = c j ­ g

4» S3 S3 S U O S >> CJ C

C o o 0.0 0.5 1.0

Reflectance of coal (%)

Fig.46(b): Variation of oxygen contents of extracts of coals heated to maximum contraction temperatues with rank of parent coal.

W3

es

3 o

ce .J

cu —

c ­ Sí β .g — 'Ξ = -cu C "Ξ O Β χ

CU w

Reflectance of coal (9c)

Fig.46(a): Variation of oxygen content of extracts of coal heated to softening temperature with rank of parent coal.

178

.2 SX) υ s

2*5 £ cu XC « O

e— w

» O

c« *­

C "β

I­S CU u

3·= s 3 α ä sx) o Q . X . B

/—. » " cu w O —

2 4 6 8 U) 12 14 16

Oxygen contents of raw coals

Fig.46(c): Variation of oxygen contents of extracts of coals heated to softening temperatures with those of the raw coals.

« 5 * Ï3 — ­w U S — . _ ς; Χ X U Zi S3 3

«S = 2 • O cu c« — c s-β E c* c» ej s « *" C cu c eu - = C CU s : C5 C£ C ■-> > CU —

Γ

10 12 14

Oxygen contents of raw coals (wt %)

Fig.46(d): Variation of oxygen contents of extracts of coals heated to maximum contraction temperatures with rank of parent coal.

179

co g υ

o Q

IV

18-

17-

16-

15 -

14 -1

Β

, r

□

Β α Β

ι

Β

α □

° Q B

Β

Ξ

ι

Β

0.0 0.5 .0 2.0

Reflectance of coal (%)

Fig.47: Variation of retention volume of extract in tetrahydrofuran with rank of parent coal.

180

Fig. 48(a): 160

Permeability of Rawdon coal

c ε « υ

5 o

LL

en (Β

O

200 4 0 0 600

Temp (°C)

8 0 0

Fig. 48(b)

200

Permeabil ity of Wearmouth coal

c E o

o li­

en to O

100 -

Temp (°C)

181

Fig. 48(a): 160

Permeability of Rawdon coal

ε o υ

o LL

en m

C

200 400 600

Temp (°C)

8 0 0

Fig. 48(b): Permeability of Wearmouth coal

200

£ E u

o LL

m to O

100 -

200 4 0 0 600 Temp (°C)

800

182

Fig. 48(c): Permeability of Wentz coal 200

e | u

5 o t i ­en co O

100

200 400 600

Temp (°C)

800

Fig. 48(d): Permeability of Ruhrkohle coal 200

c E u υ

3 o

en es

100 -

800

183

Fig. 48(e): Permeability of Buchanan coal

200

E υ υ

3 o

re

100 -

800

Fig. 48(f): Permeabil i ty of Saraji coal

200

c

E u o^

3 o

LL

tn ¡o

α

100-

200 400 600

Temp (°C)

800

184

Fig. 48(g): Permeability of Lady Windsor coal 160

o u

3 o

en CO

ϋ

200 400

Temp (°C)

600 800

Fig.48(h): Latter half of permeabil i ty curves 120

Q Wearmouth

♦ Wentz

B Ruhrkohle

I • Saraji

■ Buchanan

c

E u o

3 o

LL

tn

re ü

185

2O00 4000 6000

time (secs)

Fig.49: A typical graph showing the variation of fractional uptake with time for the diffusion of oxygen into semicokes.

186

ZI

E cu

cu cz e . C5

CU

CX

X

2.5-

1.5-

0.5-

□

' ί ­

α

B Q

ι

Region of dangerously swelling coals

B

B

Ψ

Β

ι

Β

0.0 0.5 1 .0 1.5 2.0

Reflectance of parent coal (%) Fig.50: Var ia t ion of oxygen capaci ty of semicoke produced at 600°C with rank of pa ren t coal.

2.0

·"- 1.5 03

.^: 1.0 Η

5 0.5 Η sr

0.0 0.0

Region of dangerously swelling coals

—ι— 0.5

— ι — 1.0

B B

i

m* Q B

a po-—r 1.5 2.0

Reflectance of parent coal (%)

Fig.51: Variation of nitrogen capacity of semicokes produced at 600eC with rank of parent coal.

187

CI

o.o-l

-1.0 -

-2.0 -

■

-3.0 -

^ E L · ^

, . J ·τ- - τ - - f — -ι 1 r—

□

time (sec.) Fig.52: Typical graph of the variation of ln((Me-Mt)/Me) with time for the diffusion of oxygen into semicokes.

CU

-1 -

- 2 -

4.0748 + 0.50443X FT 2 = 0.987

DDE] u i n

B BB

mu-mi

ΕΠ3

In t

Fig.53: Variation of ln(Mt/Me) with In t for the diffusion of oxygen into semicoke, a typical Fickian diffusion.

188

t ­ t

'í/3 ■ — .

·—

ro · "■

Χ

u υ SJ

Ξ C3

'_ τζ #—»

"α Β Ο

"35 2 ι—

^

¿ υ ­

15 -

10 -

5 -

0 -

G Region of dangerously

swelling coals

Β

Q PI

Β Q

1 1 1 1 J

V

Q Qn B Q

Β

Β

1 — '

Β

0.0 0.5 1 .0 1.5 2.0

Reflectance of parent coal (%)

Fig.54: Variat ion of diffusional parameter with rank of parent coal for the diffusion of oxygen into semicokes prepared at 600°C.

10

TL 8 -

χ

— 6 u cu

Ξ 2 Η 5

"55

= 0 -

S o

Region of dangerously

swell inu coals

—ι—

1 .0

I Β Άη

Q [-Q-1 ρ

0.5 1.0 1.5

Reflectance of parent coal (%)

2.0

Fig.55: Variat ion of diffusional parameter wi th rank of parent coal for the diffusion of nitrogen into semicokes prepared at 600°C.

189

0 CORDIS

The Community Research and Development Information Service

Your European R&D Information Source CORDIS represents a central source of information crucial for any organisation - be it industry, small and medium-sized enterprises, research organisations or universities - wishing to participate in the exploitation of research results, participate in EU funded science and technology programmes and/or seek partnerships.

CORDIS makes information available to the public through a collection of databases. The databases cover research programmes and projects from their preparatory stages through to their execution and final publication of results. A daily news service provides up-to-date information on EU research activities including calls for proposals, events, publications and tenders as well as progress and results of research and development programmes. A partner search facility allows users to register their own details on the database as well as search for potential partners. Other databases cover Commission documents, contact information and relevant publications as well as acronyms and abbreviations.

By becoming a user of CORDIS you have the possibility to:

• Identify opportunities to manufacture and market new products

• Identify partnerships for research and development

• Identify major players in research projects

• Review research completed and in progress in areas of your interest

The databases - nine in total - are accessible on-line free of charge. As a user-friendly aid for on-line searching, Watch-CORDIS, a Windows-based interface, is available on request. The databases are also available on a CD-ROM. The current databases are:

News (English, German and French version) - Results -Partners - Projects - Programmes - Publications -

Acronyms - Comdocuments - Contacts

CORDIS on World Wide Web The CORDIS service was extended in September 1994 to include the CORDIS World Wide Web (WWW) server on Internet This service provides information on CORDIS and the CORDIS databases, various software products, which can be downloaded (including the above mentioned Watch-CORDIS) and the possibility of downloading full text documents including the work programmes and information packages for all the research programmes in the Fourth Framework and calls for proposals.

The CORDIS WWW service can be accessed on the Internet using browser software (e.g. Netscape) and the address is: http://www.cordis.lu/

The CORDIS News database can be accessed through the WWW.

Contact details f or further Information If you would like further information on the CORDIS services, publications and products, please contact the CORDIS Help Desk :

CORDIS Customer Service Telephone: +352-401162-240 B.P. 2373 Fax: +352-401162-248 L-1023 Luxembourg E-mail: helpdesk@cordis.lu

WWW: http://www.cordis.lu/

European Commission

EUR 17150 — Coal preparation Identification of coal characteristics indicating problems of swelling during carbonization

/. Edwards, K. Thomas, F. Nadji, I. Butterfield

Luxembourg: Office for Official Publications of the European Communities

1997 — 189 pp. — 21.0 χ 29.7 cm

Technical coal research series

ISBN 92-827-9227-7

Price (excluding VAT) in Luxembourg: ECU 31.50

During the carbonization of coals in the coke oven some coals generate a very high internal pressure and swell so excessively that they damage coke oven walls or cause coke pushing problems known as 'stickers'. Several methods have been employed over the years in the assessment of coal for its suitability in the production of metallurgical coke in the coke oven process. These characterization methods include the assessment of thermoplastic properties such as the free swelling index, the Gray-King coke test, the dilatometry, plastometry and more recently high pressure dilatometry and plastometry. Although these tests are adequate for assessing the coking properties of coal they have failed to distinguish those coals that swell excessively during carbonization and those that do not. Hence the need to look at the problem of predicting dangerous swelling characteristics of coal using a more fundamental approach. The aim of the present project is to establish a fundamental understanding of the relationship between the coal structural properties, the carbonization mechanism and coke properties for coals which exhibit these properties.

Venta · Salg · Verkauf · Πωλήσεις · Sales · Vente · Vendita · Verkoop · Venda · Myynti · Försäljning

BELGIQUBBELGIE

Moniteur beige/Belgisch Staatsblad

Rue de Louvain 40-42/ Leuvenseweg 40-42 B-1000 Bruxelles/Brussel Tél. (32-2) 552 22 11 Fax (32-2) 511 01 84

Jean De Lannoy

Avenue du Roi 202/ Koningslaan 202 B-1060 Bruxelles/Brussel Tél. (32-2) 538 51 69 Fax ¡32-2) 538 08 41 E-mail: jean.de. lannoy@infoboard.be

Librairie européenne/Europese Boekhandel

Rue de la Loi 244/ Wetstraat 244 B-1040 Bruxelles/Brussel Tél. (32-2) 295 26 39 Fax (32-2) 735 08 60

DANMARK

J. H. Schultz Information A/S

Herstedvang 10-12 DK-2620 Albertslund "ΠΙ. (45) 43 63 23 00 Fax (45) 43 63 19 69 E-mail: schultz@schultz.dk URL: www.schultz.dk

DEUTSCHLAND

Bundesanzeiger Verlag

Breite Straße 78-80 Postfach 10 05 34 D-50667 Köln Tel. (49-221)20 29-0 Fax (49-221 ¡20 29 278

Nur für Veröffentlichungen des Gerichtshofes

Carl Heymanns Verlag KG Luxemburger Strasse 449 D-50939 Köln Tel. (49-221)94 373-0 Fax (49-221) 94 373-901

GREECE/ΕΛΛΑΔΑ

G.C. Eleftheroudakis SA International Bookstore Panepistimiou 17 GR-105 64 Athens Tel. (30-1)331 4180/1/2/3 Fax (30-1) 323 98 21 E-mail: elebooks@netor.gr

ESPAÑA

Mundi Prensa Libros, SA

Castella, 37 E-28001 Madrid Tel. (34-1) 431 33 99/431 32 22 Fax (34-1) 575 39 98 E-mail: mundiprensa@tsai.es URL: www.tsai.es/mprensa

Boletín Oficial del Estado

Trafalgar, 27-29 E-28071 Madrid Tel. (34-1 ) 538 22 95 (Libros)/

384 17 15 (Suscripciones) Fax (34-1) 538 23 49 (Libros)/

384 17 14 (Suscripciones) URL: www.boe.es

Mundi Prensa Barcelona

Conseil de Cent. 391 E-08009 Barcelona Tel. (34-3) 488 34 92 Fax (34-3) 487 76 59

FRANCE

Journal officiel

Service des publications des CE 26, rue Desaix F-75727 Paris Cedex 15 Tél. (33-1)40 58 77 01/31 Fax (33-1)40 58 77 00

IRELAND

Government Supplies Agency

Publications Section 4-5 Harcourt Road Dublin 2 Tel. (353-1)66131 11 Fax (353-1) 475 27 60

ITALIA

Licosa SpA

Via Duca di Calabria. 1/1 Casella postale 552 1-50125 Firenze Tel. (39-55) 64 54 15 Fax (39-55) 64 12 57 E-mail: licosa@ftbcc.it URL: icl382.dlea.iWirtual_IJbrary/biblioWetrina/ Ibosa/ttl.htm

GRAND-DUCHE DE LUXEMBOURG

Messageries du livre Sari

5, rue Raifleisen L-2411 Luxembourg Tél. (352)4010 20 Fax (352) 490 661 E-mail: mdl@pt.lu

Abonnements:

Messageries Paul Kraus

11, rue Christophe Plantin L-2339 Luxembourg Tél. (352) 499 88 88 Fax (352) 499 888 444 E-mail: mpk@pt.lu URL: www.mpk.lu

NEDERLAND

SDU Servicecentrum Uitgevers

Christoffel Plantijnstraat 2 Postbus 20014 2500 EA 's-Gravenhage Tel. (31-70)378 98 80 Fax (31-70) 378 97 83 E-mail: sdu@sdu.nl URL: www.sdu.nl.

OSTERREICH

Mañz'sche Verlags­ und Universitäts­buchhandlung GmbH

Siebenbrunnengasse 21 Postfach 1 A-1050 Wien Tel. (43-1)53 161 334/340 Fax (43-1) 53 161 339 E-mail: auslieferung@manz.co.at URL: www.austria.EU. nel:81/manz

PORTUGAL

Imprensa Nacional­Casa da Moeda, EP

Rua Marquês de Sá da Bandeira, 16 A P-1050 Lisboa Codex Tel. (351-1)353 03 99 Fax (351 -1 ) 353 02 94/384 01 32

Distribuidora de Livros Bertrand Ld.'

Rua das Terras dos Vales, 4 A Apartado 60037 P-2701 Amadora Codex Tel. (351-1) 495 90 50/495 87 87 Fax (351-1)496 02 55

SUOMI/FINLAND

Akateeminen Kirjakauppa /

Akademiska Bokhandeln Pohjoisesplanadi 39/ Norra esplanaden 39 PL/PB 128 FIN-00101 Helsinki/Helsingfors Tel. (358-9) 121 41 Fax (358-9) 121 44 35 E-mail: akatilaus@stockmann.mailnet fi URL: booknet.cultnet.fi/aka/index.htm

SVERIGE

BTJAB

Traktorvägen 11 PO Box 200 S-22100Lund Tel. (46-46) 18 00 00 Fax (46-46) 18 01 25 E-mail: btj_tc@mail.btj.se URL: www.btj.se/media/eu

UNITED KINGDOM

The Stationery Office Ltd (Agency Section)

51, Nine Elms Lane London SW8 5DR Tel. (44-171)873 9090 Fax (44-171) 873 8463 URL: www.the-stationery-office.co.uk

ICELAND

Bokabud Larusar Blöndal

Skólavõrdustig, 2 IS-101 Reykjavik Tel.(354)5515 650 Fax (354) 55 25 560

NORGE

NIC Info A/S

Østenjoveien 18 Boks 6512 Etterstad N-0606 Oslo Tel. (47-22) 97 45 00 Fax ¡47-22) 97 45 45

SCHWEIZ/SUISSE/SVIZZERA

OSEC

Stampfenbachstraße 85 CH-8035 Zürich Tel. (41-1)365 53 15 Fax (41-1)365 54 11 E-mail: urs.Ieimbacher@ecs osee.inet eh URL: www.osec.ch

ISRAEL

CESKÁREPUBLIKA

NIS CR - prodejna

Konviktská 5 CZ-113 57Praha1 Tel. (42-2) 24 22 94 33 Fax (42-2) 24 22 94 33 E-mail: nkposp@dec.nis.cz URL: www.nis.cz

CYPRUS

Cyprus Chamber Of Commerce & Industry

38, Grivas Digenis Ave Mail orders: PO Box 1455 CY-1509 Nicosia Tel. (357-2) 44 95 00/46 23 12 Fax (357-2) 361 044 E-mail: cy1691_eic_cyprus@vans.infonet.com

MAGYARORSZAG

Euro Info Service

Europa Haz Margitsziget PO Box 475 H-1396 Budapest 62 Tel. (36-1)11 16061/11 16216 Fax (36-1) 302 50 35 E-mail: euroinfo@mail.matav.hu URL: www.euroinfo.hu/îndex.htm

MALTA

Miller Distributors Ltd

Malta International Airport PO Box 25 LOA 05 Malta Tel. (356) 66 44 88 Fax (356) 67 67 99

POLSKA

Ars Polona

Krakowskie Przedmiescie 7 Skr. poeztowa 1001 PL-00-950 Warszawa Tel. (48-2)26 12 01 Fax (48-2) 26 62 40

TURKIYE

Dünya Infotel A.S. Istiklál Cad. No: 469 TR-80050 Tünel-Istanbul Tel. (90-212)251 91 96

(90-312)427 02 10 Fax (90-212) 251 91 97

BALGARIJA

Europress­Euromedia Ltd

59, Bid Vitosha BG-1000 Sofia Tel. (359-2) 80 46 41 Fax (359-2) 80 45 41

HRVATSKA

Mediatrade Ltd Pavia Hatza 1 HR-10000 Zagreb Tel. (385-1)43 03 92 Fax (385-1 ) 44 40 59

ROMANIA

Euromedia

Str. G-ral Berthelot Nr 41 RO-70749 Bucuresti Tel. (40-1 )2104401/6140664 Fax (40-1 ) 210 44 01 /312 96 46

SLOVAKIA

Slovenska Technlcka Kniznica

Nãm estie slobody 19 SLO-81223 Bratislava 1 Tel (42-7)53 18 364 Fax (42-7) 53 18 364 E-mail: europ@tbb1 sltk.stuba.sk

SLOVENIA

Gospodarski Vestnlk

Zalozniska skupina d.d. Dunajska cesta 5 SI-1000 Ljubljana Tel. (386)61 133 03 54 Fax (386) 61 133 9128 E-mail: belicd@gvestnik.si URL: www.gvestnik.si

R.O.Y. International

17, Shimon Hatarssi Street PO Box 13056 61130 Tel Aviv Tel. (972-3)54614 23 Fax (972-3) 546 14 42 E-mail: royrl@netvision.net.il

Sub-agent for the Palestinian Authority:

Index Information Services

PO Box 19502 Jerusalem Tel. (972-2)27 16 34 Fax (972-2) 27 12 19

RUSSIA

CCEC

60-letiya Oktyabrya Av. 9 117312 Moscow Tel.(095)135 52 27 Fax (095) 135 52 27

AUSTRALIA

Hunter Publications

PO Box 404 3167 Abbotsford, Victoria Tel. (61-3)9417 53 61 Fax (61-3) 9419 71 54

CANADA

Uniquement abonnements/ Subscriptions only:

Renouf Publishing Co. Ltd

1294 Algoma Road K1B 3W8 Ottawa. Ontario Tel. (1-613)741 73 33 Fax (1-613) 741 54 39 E-mail: renouf@fox.nstn.ca URL: fox.NSTN.Ca/-renouf

EGYPT

The Middle East Observer

41, Sherif Street Cairo Tel. (20-2) 39 39 732 Fax (20-2) 39 39 732

JAPAN

PSI­Japan

Asahl Sanbancho Plaza #206 7-1 Sanbancho, Chiyoda-ku Tokyo 102 Tel. (81 -3) 3234 69 21 Fax (81-3) 3234 69 15 E-mail: psijapan@gol.com URL: www.psi-japan.com

SOUTH AFRICA

Saffo

5th Floor Export House. CNR Maude & West Streets PO Box 782 706 2146 Sandten Tel. (27-11)883 37 37 Fax (27-11)883 65 69

UNITED STATES OF AMERICA

Beman Associates

4611 -F Assembly Drive MD20706 Lanham Tel. (301) 459 2255 (toll free telephone) Fax (800) 865 3450 (toll free fax) E-mail: query@bernan.com URL: www.beman.com

MEXICO

Mundi­Prensa Mexico, SA de CV

Río Panuco, 141 Delegación Cuauhtemoc ME-06500 México DF Tel. (52-5) 553 56 58/60 Fax (52-5) 514 67 99 E-mail: 104164.23compuserve.com

REPUBLIQUE DE CORÉE

Kyowa Book Company

1 F1. Phyung Hwa Bldg 411-2 Hap Jeong Dong, Mapo Ku 121-220 Seoul Tél. (82-2) 322 6780/1 Fax (82-2) 322 6782 E-mail: kyowa2@ktnet.co.kr.

ANDERE LANDER/OTHER COUNTRIES/ AUTRES PAYS

Bitte wenden Sie sich an ein Büro Ihrer Wahl / Please contact the sales office of your choice / Veuillez vous adresser au bureau de vente de votre choix

NOTICE TO THE READER

All scientific and technical reports published by the European Commission are announced in the periodical 'euro abstracts', published every two months, and in the R&TD publications database of CORDIS, the Community Research and Development Information Service. For subscription (1 year: ECU 65) please write to the address below.

o CO >

o

Price (excluding VAT) in Luxembourg: ECU 31.50 ISBN ^ 2 - 0 5 7 - ^ 5 7 - 7

• * • OFFICE FOR OFFICIAL PUBLICATIONS * EUR -k OF THE EUROPEAN COMMUNITIES • o p •

* • * L-2985 Luxembourg 9 '789282"792278 >