PLASTICITY AND AGGLOMERATION IN COAL PYROLYSIS
by
William Shan-chen Fong
B.S. in Chemical Engineering
California Institute of Technology (1981)
Submitted to the Department of Chemical Engineering
in Partial Fulfillment of the Requirements for the
degree of
DOCTOR OF SCIENCE
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
February,
)Massachusetts
1986
Institute
of Technology
1986
Signature of Author
Department of Chemical
Engineering
February 12,
1986
Certified by
"rof. Jack B. Howard
Thesis Supervisor
Certified by
Dr.
William A. Peters
Thesis Supervisor
Accepted by
C
ittee
Of TECHNOLOGY
ARCHIVE-
JUN 0 3 1986
LIBRARIES
Prof. William M. Deen
on Graduate Students
PLASTICITY AND AGGLOMERATION IN COAL PYROLYSIS
by
William S. Fong
Submitted to the Department of Chemical Engineering at the Massachusetts
Institute of Technology in February 1986, in partial fulfillment of the
requirements for the degree of Doctor of Science.
ABSTRACT
Two new techniques were developed to study the plastic behavior
of softening coals at severe conditions, i.e. high heating rates and high
temperatures pertinent to modern coal conversion processes but previously
unattainable with conventional instruments. A Pittsburgh No.8 Seam
bituminous coal was used mainly in this study.
A fast response plastometer was developed to measure the apparent
viscosity of a rapidly pyrolyzing coal. The plastometer determines the
torque required for constant speed rotation of a thin disk embedded in a
thin layer of packed coal particles heated electrically between two
parallel metal plates. Heating rates, final temperatures and sample
residence times at final temperatures can be separately selected and
controlled over the ranges 40 - 1000 K/s, 600 - 1250 K, and 0 - 40s
respectively. The instrument can be operated in hydrogen or inert gas
atmospheres from vacuum to 100 atm.
An electrical screen reactor with rapid sample quenching capabiwas developed to determine quantitatively the intraparticle liquid
lity
content (pyridine extract), and the growth of pyridine insolubles in the
pyrolyzing coal as a function of pyrolysis tempetature-time history in
the same range as in the plastometer study.
The molecular weight distribution of the intraparticle liquid was
by Gel Permeation Chromatography as a function of
determined
also
pyrolysis temperature-time history.
A combination of the above three techniques allows the kinetics
and mechanism of plasticity to be probed. Coal plasticity can be directly
viscosity change. As a Pittsburgh Seam bituminous coal
described by its
softens to form a solid/liquid suspension the
coal
the
is heated,
from more than one million poise to minimum
decreases
which
viscosity of
poise. Resolidification of the interthousand
a
few
values of order
by a rapid rise in viscosity.
indicated
is
phase
liquid
mediate
Initial softening of the coal is by physical melting of an
estimated 25% by weight of the coal. Up to 40% by weight of the coal is
further converted to liquids by pyrolytic bond breaking. These liquids
can condense or polymerize to give higher molecular weight liquids and
solids, and crack to give gas and coke, or escape from the coal particle
to give tar. The result is that the liquid weight averaged molecular
weight can increase from 570 to above 800, and then decrease to around
600.
Pyrolysis of a heat-treated coal tar (similar to the intraparticle liquid in a pyrolyzing bituminous coal) exhibited a similar
change in molecular weight with pyrolysis time, and a slower rate of
solids growth. In the pyrolysis of a slightly caking wood, or a nonsoftening lignite, little or no liquids were formed. Pre-extraction of
the raw bituminous coal with pyridine reduced both the plastic period and
the intra-particle liquid content during pyrolysis. In vacuum or 35 atm
inert gas atmosphere, the duration of plasticity is less than that under
1 atm inert gas pressure.
Experimental viscosity and intraparticle liquid formation/depletion data were described by a global model that accounts for melting,
generation and depletion of the liquid, and volatiles formation. The
kinetic parameters find quantitative application in modeling coal
particle agglomeration and swelling behavior under rapid heating conditions. Predictions of models here developed were in qualitative accord
with literature data on swelling and provide a possible explanation for
coal agglomeration observed within a pulverized coal flame.
Thesis supervisors:
Professor Jack B. Howard, Department of Chemical Engineering
Dr. William A. Peters, Energy Laboratory
- 2 -
- 3 -
ACKNOWLEDGEMENTS
I wish to thank my advisors for providing an interesting and
challenging thesis topic, and their guidance and support during the
course of this work. Professor Howard and Dr. Peters have educated me not
only in the area of coal research, but in the approach to scientific
research in general. For example, they have taught me with extreme
patience how to write technical papers and present ideas in conferences.
I have enriched my experience at MIT through learning from them.
Professors Adel F. Sarofim and Edward W. Merrill have been kind
to be on my thesis committee. Their comments and suggestions are appreciated.
My sincere appreciations are due to my colleagues Dr. Jean-Louis
Saint-Romain, Jaanpyng Hsu, and Yehia F. Khalil for their contributions
on coal tar
pyrolysis, agglomeration calculations and coal pyrolysis
experiments respectively; to Prisca Chen, Alice F. Chow and Anne Reeves
for helping with the experiments in the summer of 1985, and to Dr. Myongsook Oh, Dr. Michael Serio, Michael Boroson, Alain L. Bourhis, Steven
C.K. Lai, Cliff Chang for helpful discussions.
Our able machinist
machinist Richard D. Jankins
this work. They have taught
build most of my reactors and
Mario J. daSilva, and the late project
have been very helpful during the course of
me the basics of machining so that I can
plastometer myself.
I also like to thank my friends Kowk Chow, Alex Liu, Sze-hang Lo,
Steven E. Wirth, Joanna Cheung, Stella Cheung, Francoise Saint-Romain,
Cathy daSilva, Yu-shin and Ru-Huai Yang, and Patrick Cheung. Their
friendship meant a lot to me through my stay at MIT.
Financial support of this work from the U.S. Department of Energy
under Contract No's DE-AC21-82MC-19207
and DE-RA21-85MC-22049
is
grate-
fully acknowledged.
Finally, I like to thank my parents and my sisters for their love
and support.
- 4 -
Table of Contents
Page
List of Figures
6
List of Tables
11
1.
2.
3.
4.
Summary
1.1
Background and objectives
12
1.2
Experimental
15
1.3
Results
23
1.4
Discussion of experimental results
38
1.5
Modeling
45
1.6
Potential Applications
61
1.7
Conclusions
63
1.8
References for this chapter
64
Introduction and background
2.1
Introduction
66
2.2
Background
69
2.3
Thesis objectives
78
Experimental technique and apparatus
3.1
Introduction
80
3.2
Apparent viscosity measurement
81
3.3
Liquids formation and depletion measurement
100
3.4
Other techniques and experiments
105
Results of experiments
4.1
Plastometer results
110
4.2
Screen reactor data
122
4.3
Pyrolysis of other materials
124
4.4
GPC data
134
4.5
Other experiments
134
- 5 -
5.
Discussion of experimental results
6.
Mathematical modeling
143
6.1
Coal plasticity kinetics
154
6.2
Modeling of coal particle agglomeration
163
6.3
Other possibilities of modeling
6.3.1
Swelling of pulverized coal particles
183
6.3.2
Fragmentation of large coal particles
192
7.
Conclusions, potential applications and recommendations
197
8.
References
202
- 6 -
List of Figures
Figures for Summary Chapter
Papj
Figure
Title
1
Physical changes of bituminous coal upon heating
13
2
Schematic of coal plastometer
16
3
Details of plastometer hot stage
16
4
Electrical circuits for sample heating
16
5
Reproducibility of plasticity curves
19
6a,b
Behavior of different materials in the plastometer
19
7
Rapid quenching screen reactor
20
8
Screen reactor temperature histories
20
9
Raw data from plastometer, in 1 atm of helium
24
10
Raw data from plastometer, in 35 atm of helium
indicating different stages of plastic behavior
24
11
Viscosity curves for constant heating rate runs
25
12
Viscosity curves for holding temperature runs
25
13
Effects of pressure on plastic period
27
14
Effects of extraction on plasticity
27
15a,b
Volatiles, pyridine extractables and insolubles yield
for bituminous coal as a function of pyrolysis time
28
Pyridine extractables generation and depletion at
different temperatures
28
Volatiles, pyridine extractables and insolubles yield
for lignite and wood as functions of pyrolysis time
30
18
Pyridine extractables for different materials
31
19
Pyrolysis of a heat-treated tar
32
16
17a,b
- 7
20
Comparison of pyridine extracts for char from raw coal
and pre-extracted coal
32
21
GPC curves for pyridine extracts of different materials
34
at different pyrolysis times
22a,b
Normalized molecular weight distributions of pyridine
extracts at different pyrolysis times
for bituminous coal
35
Normalized molecular weight distributions of pyridine
extracts at different pyrolysis times for coal tar
36
Changes in average molecular weights for bituminous
coal and heat treated coal tar
37
Changes in averaged molecular weights during pyrolysis
of a heat treated coal tar
37
26
Comparison of viscosity and extract data
39
27a,b
Calculated viscosity curves for 450 K/s heating rate
to different holding temperatures, and comparison with
observed plastic period (solid circles)
47
Calculated viscosity curves for conditions of Figures
11 and 12, using liquid formation kinetics and
depletion kinetics
51
23
24
25
28
29
Calculated viscosity curves for conditions of Figures
11 and 12, with k 2 fitted
53
30
A pair of agglomerated particles
55
31a,b
Liquid content and swelling ratio for two temperature
histories
55
Weight fraction of unagglomerated particles at
different intensities of turbulence
55
31e,f
Weight fractions of different size agglomerates formed
58
31g,h
Weight fractions of agglomerate vs final temperature
58
32
Limiting cases of swelling (a) single bubble swelling
(b) reacting and solidifying foam
60
Liquid content and swelling ratio, same temperature
histories as in Figures 31a,b
62
31c,d
33
- 8 -
Figures for Main Text
2.1.1
Physical changes of bituminous coal upon heating
68
2.2.1
Role of plasticity in coke formation
72
2.2.2
Plasticity data from Gieseler plastometer
74
3.2.1
Schematic of coal plastometer
84
3.2.2
Details of plastometer hot stage
85
3.2.3
Electrical circuits for sample heating
87
3.2.4
Device for loading coal into shearing chamber
89
3.2.5
Raw data from plastometer, in 1 atm of helium
93
3.2.6
Raw data from plastometer, in 35 atm of helium
indicating different stages of plastic behavior
94
3.2.7
Behavior of quartz and lignite in plastometer
97
3.2.8
Behavior of pitch, coal tar and bituminous coal
in plastometer
98
3.2.9
Reproducibility of plasticity runs
99
3.3.1
Rapid quenching screen reactor
101
3.3.2
Screen reactor temperature histories
103
3.3.3
Weight loss, pyridine extract and insolubles yield
of bituminous coal at different pyrolysis times
104
3.3.4
New box screen reactor
106
3.4.1
Schematic of Gel Permeation Chromatography system
107
4.1.1
Temperature histories of plastometer runs
111
4.1.2
Viscosity curves for constant heating rate runs
112
4.1.3
Viscosity curves for holding temperature runs
113
4.1.4
Raw data for holding temperature runs
114
4.1.5
Plastic period for runs at 450 K/s to different
holding temperatures
115
- 9 -
4.1.6
Rates of softening and resolidification for
450 K/s runs
116
4.1.7
Plastic period at different heating rates
118
4.1.8
Softening rate at different heating rates
118
4.1.9
Effects of pressure on plastic period
119
4.1.10
Resolidification rates at different ambient pressures
120
4.1.11
Effects of extraction on plasticity
121
4.2.1
Pyridine extractables generation and depletion at
different temperatures
123
4.2.2
Arrhenius plot for extract depletion rate constant
125
4.2.3
Yield of pyridine extract + volatiles at different
temperature histories
126
Comparison of pyridine extracts for chars from raw
bituminous coal and pyridine pre-extracted coal
127
Comparison of pyridine extracts for raw bituminous
coal, tetralin soaked coal, and pyridine
pre-extracted coal
128
Volatiles, pyridine extractables and insolubles yield
for (a) wood ,(b) bituminous coal, and (c) lignite as
functions of pyrolysis time
129
4.3.2
Pyridine extractables for different materials
131
4.3.3
Pyrolysis of a heat-treated tar
132
4.3.4
Pyrolysis of a heat-treated tar at different
temparatures
133
4.4.1
Raw GPC curves for bituminous coal, wood and lignite
pyridine extract
135
4.2.4
4.2.5
4.3.la-c
4.4.2a-c
4.4.3
4.4.4a,b
4.4.5a,b
Raw GPC curves for pyridine extract of bituminous coal
0
pyrolyzed to (a) 5550C, (b) 658 0 C and (c) 749 C
136
GPC curves for pyridine extracts of different
materials at different pyrolysis times
137
Normalized molecular weight distributions of
pyridine extracts at different pyrolysis times
for bituminous coal
138
Normalized molecular weight distributions of pyridine
extracts at different pyrolysis times for coal tar
139
-
4.4.6
4.4.7
10 -
Changes in average molecular weights for bituminous
coal and heat treated coal tar
140
Changes in averaged molecular weights during
pyrolysis of a heat treated coal tar
141
5.la,b
Comparison of viscosity and extract data
6.1.1
Calculated viscosity curves for 450 K/s heating rate
to different holding temperatures, and comparison
with observed plastic period (solid circles)
144,5
155
6.1.2
Calculated vs experimental volatiles, pyridine extract
159
and insolubles yield for two temperature histories
6.1.3
Calculated viscosity curves for conditions of
Figures 11 and 12, using liquid formation kinetics
and depletion kinetics
6.1.4
161
Calculated viscosity curves for conditions of Figures
11 and 12, with k 2 fitted
162
6.2.1
A pair of agglomerated particles
165
6.2.2
Liquid content and swelling ratio for two
temperature-time histories
174
Weight fraction of unagglomerated particles at
different intensities of turbulence
177
Weight fractions of different size
agglomerates formed
178
6.2.5
Weight fractions of agglomerates vs final temperature
179
6.2.6
Weight fractions of agglomerate vs final temperature
for (a) 1000 K/s heating rate, (b) laminar flow
181
Limiting cases of swelling (a) single bubble swelling
(b) reacting and solidifying foam
185
6.3.2
Expansion of plastic foams (Benning, 1969)
191
6.3.3
Liquid content and swelling ratio, same temperature
histories as in Figures 31a,b
193
6.3.4
Possible model for coal particle fragmentation
194
6.3.5
Equations for fragmentation model
196
6.2.3
6.2.4
6.3.1
-
11 -
List of Tables
Title
Page
Characteristics of coal studied
18
3.1.1
Summary of experiments
82
3.2.1
Characteristics of coal studied
92
Table
1
- 12 -
SUMMARY
CHAPTER 1.
BACKGROUND AND OBJECTIVES
1.1
coals upon
Bituminous
that can escape
unstable liquid
melts and
heating
by evaporation
from the coal
to
give
leaving behind a solid
light oil or crack to form light gases,
tar and
decomposes to form an
residue. Polymerization or cross-linking of the liquid molecules to form
solids also competes with these other liquid removal
depleted,
is eventually
liquid
leaving
a
porous
kinetics of these processes are strongly temperature
of changes in physical
schematic
is a
structure
processes.
coke
residue. The
dependent.
and
All the
Figure 1
liquid volume
fractions in coal.
The
transient
occurrence
of
the
liquids,
commmonly
referred
to as the plastic phase, can strongly influence the thermal conversion
and refinement of bituminous coals. Irrespective of reactive atmospheres,
presence
the
of solvents or feed
size,
understanding and subsequent
plasticity kinetics is
important
for
modelling of different bituminous coal
pyrolysis,
gasification,
combustion,
processes.
Some existing
technical
carbonization
problems
such
and
as the
liquefaction
serious caking
difficulties encountered in fluid bed gasifiers are direct consequence of
plasticity.
of the bed,
plastic
Agglomeration
and
coal
of
plastic particles
plugged pipelines. Bubbles
mass,
in
turn affecting
can cause solidification
of volatiles
can swell the
the devolatilization
rate
(Oh,
1985). The yields and qualities of pyrolysis products are affected since
the transport of volatiles through the viscous coal melt affords oppor-
UPON
CHANGE
PHYSICAL
_
HEATING
m
S
raw
coai
L --- *
so
E' tnr
MC
n
w
char
S
.
M Cng
4- %cobusio
Ca
0
1.0
-- void
1 quid
* *
Osolid*
0.0
heatIngq
time
Figure 1. Physical changes of bituminous coal upon heating
- 14 -
tunities
for intra-particle secondary
also affects the physical
and
efficiency
process
(Serio,1984). Swelling
reactions
hence the burnout time
structure of the char,
in
subsequent
combustion.
The
and
development
disappearence of plasticity is closely related to formation and destruction of liquids. Study of plasticity kinetics can therefore shed light on
the mechanisms of coal liquefaction. Therefore, plasticity study contributes to both basic understanding in coal science, and to improved design
in several important coal conversion technologies.
under
Plasticity kinetics
rather extensively
carbonization
studied (Loison et al.,1963,
conditions has
Habermehl
et al.,1981).
to a moderate final
These involve low heating rate experiments
been
tempera-
ture, typically at 0.05 K/s to 800 K. Very limited studies have been made
to date outside this range,
although rather uncertain extraploration of
these data to other conditions have been made in
(Attar,
1978;
Melia
and
1983).
Bowman,
In
some modelling studies
modern
and
conversion
coal
combustion processes, heating rates of order 104 K/s, and final temperatures to 1700 K are possible.
This
coal
thesis
plasticity
modern
coal
involves
at high
conversion
a
heating
systematic
rates
processes
(Fong
and
approach
to
the
study
pertinent
temperatures
Specific
et
al.,1985).
and
instruments
of
to
thesis
objectives are:
(1)
Design
apparent
and
develop
viscosities
new techniques
and
rapidly pyrolyzing coal.
intra-particle
liquid
for
measuring
inventories
of
a
- 15 -
(2)
the
Study
effects
of
process
pressure
on
plasticity
kinetics
aspects
of
the
of
nature
variables
and
as
such
investigate
plasticity
and
temperature
some
softening
and
qualitative
coal
pyrolysis
reactions.
(3) Develop global models of plasticity kinetics and models of agglomeraand
tion
and
particle swelling,
plasticity
possibly other
related
phenomena.
1.2 EXPERIMENTAL
Two
new experimental
methods
were developed
for
the
present
study. To measure the rapidly changing apparent viscosities of pyrolyzing
coals,
a
holding
fast
response
temperatures
was
plastometer
developed.
liquids content within pyrolyzing
for
high
sample
To
determine
heating
rates
quantitatively
and
the
coal, an electrical screen pyrolyzer
with rapid sample heating and quenching capability was also developed.
(a) Apparent viscosity measurement
The
plastometer
determines
the torque
required
for
constant
speed rotation of a thin disk embedded within a thin layer of coal heated
electrically
to a
between
rotating disk
sample residence
and
two parallel
Its geometry is
metal plates.
viscometer. Heating
rates,
final
temperatures
times at final temperatures can be separately
controlled over
the ranges
respectively. The instrument
40-1000 K/s,
can be
600-1250
operated in
atmospheres from near vacuum to 100 atm.
similar
K,
selected
and
0-40
hydrogen or inert
Figure 2 is
and
a schematic
s
gas
of the
To transducer
INOEPENENT
vARiAeLES
_
THIENG
HE AT ING
CIuT
MOTOR SPEED f
CONTROL
HYDROGEN
HELIUM .ETC
3.75 cm
FAST REPOSE
4
HOLDING TEMPERATURE
AND HEATING RATE
SHEAR RATE
TOTALPRESSURE
AND
REACTIVE GASES
TORQUE AND
(TEMPERATURE READING)
RAW
DATA
Front Sectional View
Figure 2. Schematic of coal plastometer
Radius =0375cm
b-
1.35 cm
--. 4
Side Sectional View
Figure 3 Plastometer Hot Stage
Figure 4. Electrical heating circuit
- 17 -
plastometer system.
Figure 3 shows the hot stage of the plastometer.
characteristics
and performance
Design specifications
of the instrument
have been published (Fong et al.,1985).
pulses of D.C.
the metal
heating
circuit. Rapid changes
by
duration and magnitude
current of preselected
electrical
through
recorded
sequentially passing two large constant
is achieved by
Heating
plates. Figure
in
4
is
a
the torque and
electronic
a two-channel
schematic
instrument with viscous liquid standards,
of the dual current
were
temperature
recorder. By
viscosities
each
calibrating the
absolute units
in
are obtained from the experimental torque curves.
The
heating rate
for
an
a
to a final temperature
extended
of
period
to
made
test
the
No.8
of
a constant
constant
which was then maintained
time. Constant
heating
Seam
histories were
runs consist
runs
rate
refer
to
at a contant rate for the duration of
experiments with the
the experiment. Several
Pittsburgh
volatile
Two types of temperature-time
1).
increasing the sample temperature
were
high
the study. Holding temperature
in
employed
(Table
coal
bituminous
was
studied
coal
reproduceability
of
same temperature
history
the
5).
data
(Figure
For
comparison purpose, plasticity curves for some other materials were also
obtained.
These include
a viscous heat treated
coal
tar,
a low melting
point pitch, a non-softening lignite, and quartz (Figures 6a and 6b).
(b) Liquids formation and depletion measurement
The
electrical
screen reactor
(Figure
7)
was
developed
from
-
STUDIED
CPAL
Proximate analysis ( wt.%
Moisture+
Ash
H
68.8
4.9
+
,dry basis )
FC
39.4
49.1
Ultimate analysis ( wt.%
C
,
VM
11.5
1.4
al~
Bituminous~i
Volatile
High
18 -
,
dry basis )
0
N
S
Ash
8.2
1.3
5.4
11.5
as-received basis
Table 1. Characteristics of Coal Studied
1
-
19 -
7.2
0
0
>1
43
T.4
0
0
3-6
00
1.
2.o
3.0
4.0
5.0
S eo
Figure 5.
1
>n d 3
Reproducibility of Plasticity Curves
7.2
4
5
0
94
E
z
'4
CV0
3.6
b
V4
1.5
12.
14
0
0
1.0
2.0
(A)
3.0
5.0
4.0
S--0
" dm
Seconds
10
(B)
Figures 6a,b. Behavior of Different Materials in the Plastometer
- 20 -
seam
b I ttb-m I
-
90
1
atm
ii
coalI
y&m
heliumiy
to temp. recorder
to timer
12 p VDC power
to vacuum Pump
to helium SUPPly
to timer & liquid nitrogen
Figure 7. Rapid quenching screen reactor
11
Fj1
i~p
-4~~i
t:1bz~
r!i
Ri
4
ilI
I
11. 7r
'T
I IIIR11
-i
i"T
I'I~
n~~i~nnhM!mliii Lft11jJILjU
. -- III- z
I
Ell
'i'i
Vertical scale Is two times thermocouple reading in mV
Horizontal scale in 0 Chromel-alumel
10 seconds
thermocouple used
Corresponding to 660 - 670 K/s heating rate,
and holding temperatures 697 - 828 K, quenching at 4500 K/s
Figure 8.
Reactor temperature histories
- 21 -
the
basic
described
version
screen to sample weight
layer
et
Anthony
were made. Larger reactor and
improvements
coal
by
during
extra-particle
the
al.
The
(1974).
screen
size.
following
A much
larger
ratio was utilized to ensure a single particle
experiments.
secondary
This
reactions
opportunities
minimize
of
volatiles,
and
for
improves heat
transfer to the particles. A special sample loading device was developed
for
uniform
charging
of
coal
onto
the
screen.
Addition
of
a
liquid
nitrogen quenching mechanism allowed rapid cooling and precise control of
the pyrolysis residence times. The electrodes holding the hot screen were
spring-loaded to keep the screen flat upon heating.
About 20 mg of 75-90 ym diameter particles were spread evenly in
the central
region of two layers of
steel screen mounted
between
10.5 x
5.3 cm,
two electrode blocks.
400 mesh
stainless
Heating was achieved
by sequentially passing two constant pulses of D.C. electrical current of
preselected
duration and magnitude
through the electrodes.
The volatile
products readily escaped from the screen and were quenched and diluted by
the ambient
interval,
reactor
gas at 0.85 atm.
pressurized
automatically
opening
liquid
a
At the end of the preset
nitrogen was
cooling valve,
sprayed
resulting
heating
onto the screen
in
a
quenching
by
rate
of 1500 K/s. In some runs cooling rates exceeding 6000 K/s were achieved
using a different
valve. Figure 8 shows thermocouple records of typical
temperature-time histories.
Coal samples were pyrolyzed for different times. The measurements
include weight loss of the sample,
corresponding to the volatiles
(tar,
light oil and gases) yield, and the weight of extractable material in the
- 22 -
quenched char, determined by 2 hour Soxhlet extraction using pyridine and
reflecting
the intra-particle
inventory
of
liquids
at
the
time
of
quenching. The yield data are reported on a dry basis.
(c) Other techniques and experiments
Information
pertinent
to
the
reactions
occurring
liquid was obtained using Gel Permeation Chromatography
mine
the molecular
pyrolysis.
weight
of
the extract
as
A Waters Associates ALC/GPC 201
the
(GPC) to deter-
a function
system with
within
of
500
A
extent
of
and 100 A
y-styragel columns in series was employed4.
Reactions
in
the coal
during
plastic
the
phase
further
were
studied using a heat-treated coal tar as a model. (This phase of the work
was conducted in collaboration with Dr. Jean Louis Saint Romain). The tar
was
a
viscous
liquid
initially
containing
3% by
weight
of
pyridine
insolubles. Pyrolysis and plasticity measurements on this "model"
allow intra-particle
phenomena
such
as polymerization,
system
liquids cracking
and evaporation to be studied more directly since certain complications
of the initial
liquid generation steps in
coal itself
are eliminated.
Some pyrolysis and extraction experiments with wood and lignite were also
performed
and
compared
to bituminous
coal
data. To
initial
study the
softening behavior of coals, some raw coal samples were pre-extracted and
afterwards subjected to the plasticity and pyrolysis measurements.
A set of prompt volatiles retention experiments were performed in
order to obtain
additional information on secondary
reactions.
Coal tar
- 23 -
samples sealed inside stainless steel foils were pyrolyzed in the screen
in
reactor
products
these
manner. Volatiles
the usual
to
remain
trapped
escape was
inside
the
suppressed,
foil
and
forcing
undergo
more
extensive secondary reactions.
1.3 RESULTS
(a) Plastometer results
typical raw data obtained under 1 atm
Figures 9 and 10 present
and 35 atm of helium respectively, at a disk rotational speed of 0.67 rpm
(corresponding to an average shear rate of 1.32 s-1 over the area of the
In Figure 9 the sample was heated at 461 K/s to a final tempera-
disk).
ture
of
874
K and
then
held
at
this
temperature for
6.0
s. As
the
temperature increased, the torque decreased to a low value due to liquid
formation. After a period of low viscosity, the torque value increased
due
to
progressive
resolidification
of
the
melt.
However,
continued
rotation of the the disk broke up the coke formed, and the torque vs time
signal showed a net decrease.
are
illustrated
The stages of physical change of the coal
in Figure 10. For the
given shear rate,
the
measured
torque is related to the absolute viscosity by 1.0 x 10-2 Nm = 2.6 x 104
Pa s, obtained by calibrating the plastometer with viscosity standards at
room temperature.
Figures 11 and 12 show typical viscosity vs time curves for this
coal for different constant heating rates, and different holding temperatures,
all under 1 atm of helium.
Measured apparent
viscosities were in
HEATING RATE 461 K/S
IN 0.1 MPa HELIUM
5
4
E
a)
z
3
N
4-
0
b
h.
a)
0.
E
2
a)
b.
0
iI
0
Seconds
Figure 9. Raw Data from Plastometer, in 1 atm of Helium
5
4
a)
h.
E
z
WN
3
0
h.
a)
0.
b
E
a)
O)
IN.
0
O0
Figure 10.
2
4
Seconds
6
8
10
Raw Data from PLastometer, in 35 atm of Helium
Indicating Different Stages of Plastic Behavior
- 24 -
7.2
0
0
3.6
U
CL
CL
0
2
4
6
8
10
Seconds
Start heating at constant rates
Figure 11. Viscosity Curves for Constant Heating Rate Runs
Heating Rates Indicated on Curves
7.2
0.
0
o3.6
CL
4
01
A
t
2
4
10
Seconds
Start heating - 450Ks"
Figure 12. Viscosity Curves for Holding Temperature Runs
Final Temperatures Indicated on Curves
- 25 -
- 26 -
the range of a few thousand to one million poises.
Some runs were made under 0.05 torr of ambient helium atmosphere,
and some under 35 atm of hydrogen or helium (Figure 13). Comparison with
1 atm data shows that the plastic period is shorter for runs in vacuum or
less than 900 K. In view of the
35 atm pressure and holding temperatures
Figure 11,
heating rate effect seen in
the difference
the 1 atm
between
curve and the other curves in Figure 13 would be seen to be even larger
if
the 1 atm experiments had also been conducted
K,
there
is
little
effect
the softening rate is
of
pressure
on
about the same as in
the
at
Above 900
350 K/s.
plastic
the 1 atm case,
period.
While
the resolidi-
fication rates for both the vacuum and 35 atm runs are higher for holding
less than 900 K. Above 900 K,
temperatures
the difference
is
small.
At
about 800 OC runs under 35 atm of helium or hydrogen and 1 atm of helium
all give the same plastic period, suggesting at this high temperature any
effects of
inert gas pressure or coal hydrogen reactions are insignifi-
cant compared
to other
chemical
and
physical
phenomena. However,
from
the present data base, an extension of the plastic period at 5500C due to
The effect
hydrogen (vs helium) at high pressures can not be ruled out.
of pre-extraction
of the raw coal
sample on its
plasticity
is
shown in
Figure 14. Also shown in Figures 6a and 6b are the different behavior of
other materials in the plastometer, note especially coal tar and pitch.
(b) Screen reactor data
Figures 15a and 15b show the yields
extracts,
and the pyridine
of the volatiles,
pyridine
insoluble char as a function of the heating
- 27 1 0
')
0
0
B
6
10
0
4
2
0
600
SOO
BOO
700
HCo 1 d IaC
*0raNtu"re
teCDMpe
900
#--
* Time interval when viscosity is less than 3.6x104 Pa s
Figure 13.
Effects of Pressure on Plastic Period
3
E
a
2
0
Figure 14.
2
4
Secmnds
6
Effects of Extraction on PLasticity
Both Runs at 410 K/s to 900 K
C
0=
.25
100
1.25
A
T
Volatiles
I
80
80
S60
@40 -
40
SPyridine insoluble
20
O
5
0
15
-
20
20
Seconds
I
0
0.5
t.d
1.5
2.0
Seconds
908 K
I
813,
K
C.9
0S
Figure 15a,b Volatiles, Pyridine Extracts and
Insolubles Yield for Bituminous
Coal vs Pyrolysis Time
$4
,i
'2
(heating rate,holding temperature,cooling rate)
(
K/s
K
,
K/s
)
*
o
(470,813,1100)
(446,858,1100)
(514,992.1100)
Figure 16. Pyridine Extractables Generation and Depletion
at Different Temperatures
- 28 -
- 29 -
for two sets of experiments with rapid quenching.
time,
The correspond-
ing temperature-time histories for each set are also shown. For example,
dotted lines connect two data points from a single experi-
the vertical
with the temperature-time
ment,
the
indicated by
history of this experiment
bold line. The solid curves in
Figures 15a and 15b are calculated
according to a kinetic model described in the modeling section.
An initial pyridine extractables yield of about 25 % was obtained
from the raw coal. As the coal was heated, the amount of extract increased
to
a
maximum
then decreased
and
at
a
rate
strongly dependent
on
temperature, as indicated in Figure 16. The time interval over which the
extract
concentration
ture. The
first
was
order
high decreased
depletion
rate
rapidly
constants
with holding temperafor
the
extracts
are
indicated. The maximum amount of fluid material (extract plus volatiles)
was as high
as 80 % by weight
of the raw coal.
However,
this quantity
decreased with increasing holding temperature, or with increasing time at
a given holding temperature, primarily due to conversion of the extractables to coke.
Pyrolysis
in
the
screen
and wood were also studied. Figure
reactor
of
a
non-softening
17 shows the evolution
lignite
of the vola-
tiles, pyridine extracts and the pyridine insoluble char as a function of
heating time. Figure 18 is a comparison
of
the
amount of
extracts of
bituminous coal, wood and lignite, all subjected to the same temperature
-time history. Subsequent reactions of the liquid after softening of the
bituminous
coal can be
inferred
from
the pyrolysis
behavior of a heat
treated coal tar. Figure 19 shows the weight loss, pyridine extracts and
z
0
0L
0
I
0
0
0
Q
0L
0
2
6
4
SECONDS
no
pyridine
t
extrac
4-
SECONOS
-Cobse * rv -ED
Figures 17a,b Volatiles, Pyridine Extractables and Insolubles Yield
For Lignite and Wood as Function of Pyrolysis Time
- 31 -
Bituminous Coal
ca
0
*
20-
100
Wood
A
-------.... Lignite
LL
0.25
0.5
1.0
2
4
8
Seconds
Samples heated at 736 K/s to 555*C
and allowed to cool by convection
Figure 18. Pyridine Extractables for Different Materials
- 32 -
100
80
4.)
60
3
40
4.)
20
0
5
0
10
15
Se~conids
A - weight loss of sample,B - pyridine extractable portion of material
left on screen
C - pyridine insoluble portion of material
left on screen
Figure 19. Pyrolysis of a Heat Treated Tar
IN 0.1 MPa HELIUM
100
0
0
U
80
-
60
-
HEATING RATE 3300Ks~'
HOLDING TEMP
908K
COOLING RATE 4500Ks~'
3
0
h.
S
Ola
3
*0
@1
r
*
.
RC
470 Ks'
813 K
1 100 KS-'
RC
o
0,
40
U
0
h.
PX
20
~...
w
PX
.
N
,
.,
0
t,
2
Start heating
4
.
Seconds
6
8
25
RC - RAW COAL
PX - PYRIDINE EXTRACTED COAL
Figure 20. Comparison of Pyridine Extracts for Chars from
Raw Coal and Pre-extracted Coal
- 33 -
pyridine insoluble char as a function of temperature-time history for the
tar.
The effect
of pre-extracting
the raw coal sample on its
extract
yield is shown in Figure 20, at two different temperature-time histories
removed from raw
% of extract
sample (25
for the pyridine pre-extracted
coal) and the raw coal. For the pre-extracted coals the extract yield is
much less.
(c) GPC Data
The
were
molecular
determined
for
weight
bituminous
distributions
coal,
coal
different extents of pyrolysis. Figure
tion vs
elution
volume)
for
different
21
differences
transformed
in their
tar,
the
pyridine
lignite
and
extracts
wood
at
shows GPC curves (concentra-
pyrolysis
materials showed different distributions
ural
of
times.
Different
raw
and weights reflecting struct-
pyridine extractables. The
and normalized with respect to the total
data
were
also
yield of extract
(Figures 22 and 23). Changes in the moments of the distribution, such as
number and weight averaged molecular weights are shown in Figures 24 and
25. For both bituminous coal and coal tar pyrolysis, the extract average
molecular weights increased to a maximum and then decreased, as a result
of polymerization, evaporation and cracking reactions.
(d) Other experiments
GPFC
CuxArv/Esp-c3,
elu
coc:>:./i.v
:
volume
%At--IoCp"N/
1m
U)
(n
14
.0
C.,-)
.0
Elution volume, ml
Elution volume, ml
14
CL
0
0
0.
U
0)
14
4)
0
Elution volume, ml
Elution volume, ml
Figure 21. GPC Curves for Pyridine Extracts of Different
Materials at Different Pyrolysis Times
- 35 -
Bitiuimincusw- Ccal
K
828
13
12
11
10
.5
9
a
a
7
a
5
4
3
2
0
0
0.2
0.4
0.6
BitAminc3s
0.8
1
1.2
(Thousands)
molecular weight
1.4
Ccoal
1.6
931
1.8
2
K
13
12
11
10
C
.2
9
a
I
I
7
6
5
4
3
2
0
0
0.2
0.4
0.6
1.2
0.8
1
(Thousands)
molecular weight
1.4
1.6
1.5
2
Figures 22a,b Normalized Molecular Weight Distributions
of Pyridine Extracts at Different Pyrolysis
for Bituminous Coal
Hek
a t'
Tre
-Ca tedC-
Cral
C c>
828
Tar
K
2019 1817 -16 ----15 -
0.0
i-
a
'
-
0.60 s
736 K/s heating to 828 K
1.50 s
s
,'3.00
14.80 s
14 -
C
S
0.25 s
0
12-
.'
10 0 8
2
0.0
0.2
0.4
0.6
0.8
(Thousands)
molecular weight
1.0
1.2
1.4
Figure 23 Normalized Molecular Weight Distributions of Pyridine
Extracts at Different Pyrolysis Times for Coal Tar
qW
Bituminous Coal
735 K
A
555"C
781 K
828 K
40O0
102
K
697 K
65B C
*350
cE
0
10
5
T~
749'C
A
15
0Sec
d
Z1400-
Coal tar
3
658*C
B
A
735 K
a
A
0.25
3L
011
102
697 K
Co
5
10
0.5
1.0
Holding temperatures
kip
250
0
-749C
K-
-A78
300
o
z
A
1
'Tr
--
15
Figure 25 Changes in Averaged Molecular Weights
during Pyrolysis of a Coal Tar
2
4
indicated on curves
A
heating rate 465 K/s
*
heating rate 455 K/s
S
heating rate 736 K/s. natural cooling
5000 K/s
quenching
Figure 24 Changes in Extract Average
Molecular Weights with Time
- 38 -
As
mentioned
above,
volatiles
are
trapped
in
foils in
some
pyrolysis experiments, thereby promoting secondary reactions(Serio,1984),
and concomitantly, their effects on plasticity and pyrolysis. A pyridine
obtained in
yield exceeding 40 % was obtained vs the 29 %
insolubles
screen experiments, for the pyrolysis of coal tar at around 1000 K. This
increase was observed at temperatures above 950 K.
1.4. DISCUSSION OF EXPERIMENTAL RESULTS
plasticity and
The
shown in
respective
extract
times
heating conditions.
rapid
26, or by a comparison of Figures
Figure
for
initial
generation
and
first such
curves are the
defined
data obtained under well
experimental
As
pyridine
near
12 and
final
16, the
of
depletion
extractables within the coal correspond well with the time for initiation
and
termination
of
the
coal's
phase.
plastic
This
that
suggests
(a)
plasticity results from the formation of a significant amount of liquid
material from coal, and (b) the pyridine extract of the rapidly quenched
coal
of the amount of
is a reasonably accurate measure
liquid product
within the pyrolyzing coal at the time of quenching.
However, the
raw coal
can
fact
be extracted
that
even without
means
that
heating, 25
"extract"
and
wt
"liquids"
% of
the
are
not
identical materials. It is likely that the initial extract corresponds to
some
loosely
bonded,
low melting
point
material.
Upon
heating
to
a
sufficiently high temperature, this material undergoes physical melting.
Our observation with the plastometer
Ts occurs around 570 K independent
that initial
softening
temperature
of heating rate (Figure 11)
supports
IN 0.1 MPa HELIUM
7-2
100
z
0
CL
80
U
v0
3
a
60
X
160U
40
PC
@1
CL
20
0.
0.
0
2
4
Seconds
68
Start heating at ~450Ks~'
Figure 26.
Comparison of Viscosity and Extract Data
10
L~J
- 40 -
the above hypothesis.
amount of material is
this initial
If
pre-extract-
ed, and afterwards the coal is subjected to pyrolysis, the total extract
racted coal is
and
formed at a later time (Figure 20).
less and is
formed is
the
expected to soften at a later time (higher
of
duration
plasticity
should
be
shorter,
as
The pre-exttemperature),
is
observed
in
Figure 14. Other workers found the Gieseler plasticity (low heating rate
softening
during the early
type experiments)
stage to be reversible as
the temperature is raised or lowered (Waters, 1962), thus supporting the
picture that some part of the liquid is formed by physical melting.
higher
At
significant
back to
coal
quantities
resolidified
of
18).
yield
The
bituminous
or
80 % daf.
coal,
lignite
generates
converted
from
The maximum yields of extractables
pyrolyzed
the
similar experiments on a lignite,
In
up to 7 % was observed
wood
breaking
be reversibly
no extract was obtained.
which does not soften,
extract
bond
which cannot
liquid
coal.
approached
studied
pyrolytic
temperatures,
only a small
the pyrolysis of wood
in
at
Similarly,
a
reflecting the structural
much
faster
rate
(Figure
than
the
difference of the fuels. The
"foil" experiments for coal tar showed more liquids are retained than in
open screen
viscosity
pyridine
pyrolysis. Figure
exceeded
26 shows
the corresponding
extractables
inventory,
probably reflects higher mass
that at 910 K,the period of low
period
inferred
from
of
high
intra-particle
screen heater
data. This
transfer resistance for volatiles
escape
away from the plastometer.
Clear
bituminous
coal
evidence
of
the
resolidification
reactions
of
softened
at longer pyrolysis times or higher temperatures
is
the
- 41 -
increase in pyridine insolubles. Studies of the pyrolysis behavior of the
heat-treated coal tar using the screen reactor showed similar evidence of
the competition between coke formation via polymerization, cracking, and
volatiles escape via transport away from the reacting substrate
19).
However,
pitch and
viscous
differences
bituminous
liquid
in
coal
containing
plastic
97
between tar,
behavior are expected
since the raw tar is
(Figure 6a)
% pyridine
solubles,
and
(Figure
already a
the. pitch is
a
solid with a low softening point at 342 K. The fluid behavior of bituminous coal requires that a minimum inventory of liquids capable of initiating and sustaining plasticity throughout
be available within the coal matrix.
three-phase medium
a reacting,
Generation of this liquid requires
that at least some melting or bond breaking precede the onset of plasticity.
further complication
A
plasticizing
agents
is
be more
may
that
heterogeneous
significant
for
reactions
the
coal
of
due to
higher mineral matter content and non-softening organic matter.
these
its
This may
explain the higher ultimate yields of pyridine insolubles in coal than in
coal
tar
(56
% vs 28 %, respectively),
that the resolidification
and is
consistent with the fact
rate for the coal tar or pitch is
slower than
in bituminous coal (Figures 11 and 14).
The
trends
for
extract
average molecular
weight
evolution are
strong functions of temperature-time history. In general, with increasing
time at a given holding temperature, the molecular weight increases to a
very well defined maximum, and then decreases ( Figures 24 and 25). Figure 22a shows that besides formation of pyridine insolubles, some heavier
molecular
weight
material
are
formed
by
polymerization.
The
molecular
weight increase is due to polymerization reactions or release of heavier
- 42 -
upon
material
heating.
possible cause. The decrease
further generation of
sition
of
the
parent
the
of
Evaporation
lighter
the
arises from cracking of
also
is
material
a
since
liquids
low molecular weight liquids by primary decompoceases
coal
severities. At
these
under
higher
holding temperatures, more severe cracking drives the molecular weight to
even
lower
competition
values. The
between
maxima of
evaporation,
for
thus
are
determined
by
and cracking reactions.
polymerization,
weight distributions
The molecular
curves
these
the extract
of coal
before and
after the onset of softening (0.5 and 0.8 s, respectively in Figure 22b)
are identical, suggesting no chemical changes during the initial stage of
plasticity. This lends strong support to the physical melting assumption
The rate of cooling in the
the molecular
may also affect
experiments
weights (Figure 24) and extract inventories. A slow cooling rate experiment
allows
tend
weights
more time
to escape,
for volatiles
to be higher, and
liquid
the
hence
molecular
those in rapid
yield less than
quenching experiments. Also, the heavier molecules in the liquid may have
more time for realignment and cross-linking
to form pyridine insolubles
and hence lower extract yields are observed (compare Figures 16 and 18).
higher
The
molecular
than
that
extracts
rapidly
for
weight
the
coal
(Figure 24). The coal
with continued
thermal
of
the
tar
bituminous
extract,
or
coal
than
and
for wood
extract molecular weight
treatment
those
is
extract
of
much
lignite
declines more
the
coal
tar
extract. This probably reflects two differences expected in their thermal
behavior becauses of differences in
Cracking of
the high
molecular
state.
their structure and initial
weight
presence of surfaces of mineral matter,
coal
extract
especially
unsoftened macerals
in
the
and already
- 43 -
available coke can give more residual coke. A larger liquid molecule also
has a lower vapor pressure, and thus a slower evaporation rate, and thus
has
chance
more
to
remain
the
in
condensed
and
phase
form
pyridine
insolubles. Both are consistent with the lower pyridine insolubles yield
and
the
faster
of weight
rate
the coal tar pyrolysis
in
loss observed
(compare Figures 15 and 19).
Plasticity
following
points
mechanisms
regarding
the
this study
from
inferred
of
behavior
a
coal
suggest
particle
the
during
heating:
(1) Initial softening at low temperature is due to physical melting.
(2) Further
of
generation
liquids is
by
pyrolytic
bond
breaking,
associated with the release of gases and tar from the coal.
depletion
(3) Liquid
insoluble
cracking
is
associated
including
material
(from
reactions
with
the
(from
solids
decreasing
growth
of
plasticity
molecular
pyridine
curves),
weights),
and
tar
evaporation.
(4) A substantial weight fraction of softening coal is
intermediate
liquids within
the
coal
before
transformed
these
liquids
into
are
depleted by the above pathways.
(5)
The
their
structure
and
evaporation
molecular
behavior
and
weight
reactivity
cracking
reactions. Competition among
determine
the observed
yields,
of
quality
these
in
liquids influence
polymerization
these processes
and overall
in
turn
generation
and
can
rate
of devolatilization liquids and gases.
(6)
Plasticity can be modified by changing the mass transfer resistance
for liquids escape (foil experiments) or by pre-treatment of the raw
- 44 -
coal (pre-extraction runs).
(7) Solvent extraction of rapidly quenched char is a promising technique
to determine
the higher
plasticity at
of
ranges
rate
heating
and
temperature (such as combustion conditions) where direct measurements
even with the present plastometer would be difficult.
The following schematic for plasticity mechanism is proposed:
transport
liquid
tar
bond breaking
coal
cracking
polymerization
melting (low T)
heavier
gas
liquids
+
+
coke
coke
This picture is also consistent with the above points as well as
the
observed
yields
seen
pressure
in
effects
associated
suppressed,
gas
discussion).
yield
of
a reduced
Under
increases
and
high
1977).
(Suuberg,
vacuum pyrolysis are consistent with
and the present observation
and
pyrolysis
on
plastic
period
pressures,
according
to
rapid
tar
the
High
tar
evaporation,
(see
Figure 13
evaporation
above
is
picture,
solids yields will also be higher because of augmented time for repolymerization.
Hence a shorter plastic period is
13). Therefore, an optimum pressure for
expected and found (Figure
maximum plasticity would exist
for rapid-heating, high temperature conditions. The effect of pressure on
- 45 -
the
polymerization
in
donors
this
is not
step
clear. Effects of
are
reaction certainly
hydrogen or hydrogen
practical
of
interest. However,
under the range of heating rate and reaction times studied, no difference
in
the
plastic
hydrogen
or
behavior
helium
differences observed
at
in
was
a
for
observed
temperature
the extents
of
in
experiments
8000C,
of liquid
nor
were
generation
35
atm
of
significant
and
depletion
when raw coal sample was soaked with tetralin and then pyrolyzed in the
screen heater.
1.5
MODELING
(a) Plasticity kinetics
The
apparent
viscosity
data
are
first
fitted
with
the
simple
first order kinetic scheme of Fitzgerald (1956):
ki
coal
k2
metaplast
-
coke
The term metaplast refers to the intraparticle liquid content M. Apparent
viscosity V is related to the viscosity of the solids free liquid P* by a
concentrated suspension model (Frankel and Acrivos, 1967)
y/p* = (9/8)/((1 - M)-1/3 _ 1)()
At a late stage of resolidification, the rate of increase of log(viscosity) is approximately
- 46 -
dlny/dt
=
k2
(2)
Thus k 2 can be found from experimental resolidification rates. Values of
y*
are
estimated
pitch (average
from Nazem's
value
work
taken,
1200
(1980)
on
carbonaceous
poise). The
obtained by fitting data for the plastic period.
lated
viscosity curves,
parameter
mesophase
ki
is
then
Figure 27 shows calcu-
and calculated plastic periods vs experimental
values for 450 K/s holding temperature runs.
However,
the
Fitzgerald
kinetic
scheme
does
not
fit
runs
at
different heating rates, because physical melting of part of the coal is
not taken into account.
The schematic
presented
in
Section
4 is
a
starting point for a
more realistic model. Both primary bond breaking and physical melting are
important
in
coal softening.
Coke formation in
the primary reaction can
be neglected as a first approximation since most of the organic material
of
the
coal
passes
macerals such as
merization in
tion.
liquid
phase,(i.e.,
are a small
the liquid phase can result in
non-softening
fraction
of
the
Poly-
a higher liquid viscosity,
At higher temperatures,
6580C, Figures 22a,b)
For this reason,
the
lower rank coals i-t should be considered.
no solids are formed.
555 0 C and
the
fusinite and inertinite
although in
total mass)
even if
through
(>
6580C,
compare
cracking predominates over polymeriza-
and because the detailed MWD information
required
for a quantitative model is not available, polymerization kinetics in the
liquid phase is not modelled. Both evaporation and cracking are pathways
to resolidification. High pressure or closed reactor geometry can prevent
COAL -
KI
K2
METAPLAST -+& COKE
0
31
LU
I:L
0
U
S21
I
-1
I I
2
t
773
973
Holding temperature, K
1173
4
6
8
Seconds
Start heating at 450Ks'
KI a
245 EXP (- 4900 / T) , S'
K2.
2 x10' EXP(-
16350 / T) , S'
Figures 27a,b Calculated Viscosity Curves for 450 K/s Heating Rate
to Different Holding Temperatures, and Comparison with
Observed PLastic Period (solid circles)
- 48 -
tar evaporation. Since no individual tar or gas species were collected in
the
both
study,
present
reaction with volatiles
as
products.
evaporation
into one
lumped
(tar and gas that leaves the particle) and coke
reaction
Further
cracking are
and
liquid
intraparticle
between
and
coke
or mineral matter are not considered.
Here the global model assumes the coal to undergo both physical
melting and pyrolytic bond breaking to form an intermediate
a
coke and inert mineral matter. The liquid
of unreacted coal,
suspension
liquid in
forms a volatile product which, according to the model, escapes from the
coal
infinitely
coke
to
rapidly, and leaves
volatiles
thus formed
is
coke. The weight ratio of
a solid
be
assumed to
(=
constant
a). The
initial 25 % pyridine extract from raw coal is assumed to be the maximum
that can
be
and
structure,
assumed
melting,
obtained
is
decomposition of
chemical
without
generated
by
melting. This
physical
to melt over a narrow range
the
temperatures. To
of
initial
material
account
is
for
a gaussian distribution of melting points with a mean at 623 K
and a standard deviation of 30 K are assumed in
the calculations.
These
values reflect the range of the initial softening temperatures of coals
(Read et
liquids
al.,1985;
Rees and
Pierron, 1954). The rate of
by melting alone would therefore
be
generation
of
the product of the heating
rate and this distribution function. The mineral matter (wt. fraction fa
= 0.13 for the coal studied,
reaction
scheme
fractions of
basis) are
the
and
Franklin,1980)
corresponding
raw coal
(moisture
rate
free,
is
assumed
expressions
to be inert.
based
mineral matter
on
The
weight
containing
- 49 -
ki
coal
k2
o intermediate liquid
-
0 volatiles + coke
-
melting
dC/dt = -ki ( C - fa ) - rm
dM/dt =
kl ( C - fa ) + rm - k2M
dV/dt =
k 2 M ( 1/1 + a )
dE/dt =
k 2M ( a/l + a )
where
C,M,V
and
E
are
the
weight
(3)
fractions
of
the
unreacted
coal,
liquid, volatiles and coke, respectively, and rm is the rate of physical
melting.
The
ion
and
depletion
pyrolysis,
material
model
is
data
first
from
applied
the
screen
when most of the coal has
that
can
physically
melt
to
intraparticle
reacted
has
At
reactor.
done
(C so
(rm
a
liquid
late
fa = 0)
= 0),
format-
stage
of
and all the
the
rate of
liquid depletion is
dM/dt = -k2 M
(4)
The rate constant k 2 can thus be found from a plot of ln(M) vs time for
constant temperatures.
16.
Typical values thus obtained
are
shown in
Figure
An Arrhenius plot gives an activation energy of 42.2 Kcal/mole,
k2 is found to be
k2 = 1.9 x 1010 exp ( -21200/T), s-1
(5)
and
- 50 -
The rate constant ki was obtained from a best
fit to the experimental
data, using values of k2 already found. The result was
ki = 6.6 x 107 exp ( -14500/T), s-1
Figure
ponding
(6)
15 gives values of pyridine insolubles and extractables
to C + E and M, respectively)
corresponding
model
predictions
observed
(solid
in
lines),
(corres-
the experiments and
using
an
"a"
value
of 1.25 (obtained from long holding time data) for two sets of reaction
conditions.
To calculate the apparent viscosities, Equation (1) is again used
with the present scheme and rate constants in Equations (5) and (6). Figure
28 shows calculations corresponding to the experimental conditions
and laboratory results presented in Figures 11 and 12. The rate constants
obtained from the screen reactor describe constant heating rate plastometer runs well but underpredict viscosity for 450 K/s runs at low holding
temperatures.
Three possible reasons are (a)
liquid phase is neglected, which can affect
P*, (b)
the polymerization
the solids
resistance to volatiles escape can exist in
in
the
free viscosity
the plastometer,
and
(c) y* is assumed independent of temperature.
If
a positive
weight,
the solids free viscosity y* is
power of the
as in
polymer
intraparticle liquid
melts,
runs would have a higher
assumed to be proportional to
then
viscosity
from Figure
weight
average molecular
24 the lower temperature
than predicted
since the
liquid has
- 51 -
7.2
45OK/1
II
160K/S
3.60
75KWS
0
2
4
6
B
10
O 7.2---1C32K
936K
3.6874K
826K
0
2
4
6
8
10
Start heating at 450 K/s
Figure 28 Calculated Viscosity Curves for Conditions
of Figures 11 and 12, using Liquid
Formation and Depletion Kinetics
- 52 -
higher average molecular weights. Temperature variation of y* is correlated by the Andrade equation with an
the expression is
However,
(Bhatia et al.,1985).
activation energy of
38 Kcal/mole
not valid over a wide
range of temperatures, and the resulting fit is not improved. If the rate
constant ki is retained (because little volatiles are formed during early
softening) and k 2 allowed to vary, then k2 = 8 x 106 exp(-14000/T)s-1 fits
the data well
A lower activation energy is
(Figure 29).
consistent with
mass transfer resistance to volatiles escape from the plastometer. Therefore
both
polymerization
at lower temperatures
and
resistance
to vola-
tiles escape cannot be ruled out.
(b) Application of plasticity data: coal particle agglomeration
The agglomeration of coal particles suspended in a gaseous medium
can be expressed as,
Ai + Aj
where
Aj
--
are
-
> Ai+j
the
(7)
j-particle
agglomerates with
a population density
nj. Assuming no agglomerates break after they are formed and no particle
loss due to adherance to the reactor walls, the formation rate of Aj is:
N-j
j-i
dnj/dt = 1/2
Ki,j-ininj -
Kig
is
the
(8)
i=1
i=1
where
Ki,jninj
overall
agglomeration
constant,
i.e.
the
product
- 53 -
0
0
0
2
4
6
8
Secon
0
2
4
6
10
rd
8
10
Se3kcon)dna
Figure 29 Calculated Viscosity Curves for Conditions
of Figures 11 and 12, with k2 fitted
- 54 -
of the collision frequency Kc and the probability of forming an agglomeper
rate
E,
collision,
ticle. For
for
the
purposes,
computational
of an
agglomeration
the
largest
that a j-particle
assumed
related
to the initial
collision constant
agglomerate
single particle
Kc of Ai and Aj in
agglomerate
a spherical
has
Aj parsize
is
Typically, N = 10. It
limited to N single particles sticking together.
is
Ai and an
radius by
shape and
aj = ao
is
(j)1 /3 . The
isotropic turbulence is
(Saffman
and Turner,1956),
K
C
10.4
=
where E 0 is
and
v
the
as/0
the
0
/v
(9)
turbulent energy
viscosity
kinematic
of
dissipation
rate
per
unit
the carrier
gas
of particles
the
bonding
volume,
in
the
agglomeration system.
If
the
strength
of
the
material
in
neck
region
formed between the two particles (Figure 30) is a m,(assumed proportional
to liquid content with constant a
stress
ab (Yoshida et al,1979)
), and if a
is larger than the bending
on the agglomerate
pair,
then successful
agglomeration results. Thus,
a
=
where ci
is
aoM(t)
mean
ci
and
a
=
10. 8p/c
/v' /
(10)
which gives a
the ratio of neck radius to particle radius,
critical value
The quantity
>
c1* >2.21(y/a 0 M)1/3(C /v)1/6 required for agglomeration.
ci is
assumed
standard
to be a normally distributed quantity with
deviation
6 ,since
the
particles
in
reality
are
Figure 30.
A Pair of Agglomerated Particles
1000OK/
tc,
15000K
3000K
1.5
/rn
to
1 700K
2.5
Swelling ratio
1.2
1.2
0.9
0.'9
0.6
0.6
Liquid wt
01l
0.2
fraction
0 1
0.6
0.4
S
co
0
0 07
0.04
0 Of.
S e-c Cn"d
ndSn
-4
J
Figures 31a,b Liquid Content and Swelling Ratio
10000OK/
to
3 500OK/m
1 OOOK
1.0
1.00
0.9
0.95
0.3
0.90
0.7
0.3 5
0.6
0.300
to
3 700K
105
106
U
S
04
0
0.5 I
0
i
0.2
i
0.4.
0.75
0
0.6
0.02
0.04
SeCondM
0.06
B4e=o nds
Figures 31c,d Weight Fractions of Unagglomerated Particles
at Different Intensities of Turbulence.
Numbers on Curves are
- 55 -
0 /vsin
o0
s
-2
- 56 -
irregularly
Summing over all agglomerates
shaped.
an overall efficiency
the critical value,
f(t)
with ci greater than
a
which is
obtained,
is
function of the time dependent liquid content and turbulence intensity.
""
,
- cj)2
l1.0
-
exp
-d
c
1*
262
Once the energy dissipation rate and the initial number density of coal
particles are specified, the agglomeration equations can be solved.
To
account
for
of
swelling
the
particles,
a single
volatile
bubble is assumed to grow in a viscous, spherical droplet. The viscosity
of
the coal melt
to the solids fraction by
is related
concentrated suspension model in
mentioned
Equation
the previously
(1).
The
volatiles
are assumed to be in equilibrium with, arbitrarily, 0.8 mole fraction of
the liquid (see following Section (c)).
The
in
Section
liquid
formation
and
A are used. An initial
depletion
particle
rate
parameters
number density
derived
of 2800/cm3 ,
and particle radii of 40 ym are assumed (corresponding to stoichiometric
combustion
computed
heating
in
air
of
pulverized
coal
liquid content and swelling ratio
from room temperature
(a)
particles
of
(radius/original
to a final
size).
The
radius)
for
this
temperature of
1000 K at
10000 K/s and (b) to 1700 K at 15000 K/s are shown in Figures 31a,b. The
mass
fraction of
values of e0 /v
time
(Figures
collision
in
single
turbulent
31c,d).
frequency
unagglomerated
particles, A1 , for
different
of
pyrolysis
flow are plotted
As the intensity
increases
and
of
results
as functions
turbulence
in
more
increases,
the
agglomeration.
- 57 -
the agglomeration probability per
At very intense levels of turbulence,
of the high
low because
collision
is
potential
agglomerates.
Therefore
agglomeration
little
the
imposed by
stress
results.
the coal is
meration essentially ceases when the liquid in
fluid
on
Agglo-
depleted.
Figures 31e,f show the generation of different size agglomerates with e
0
at 106 sec- 2 . The distances between the curves indicate the mass fraction
of feed coal which have agglomerated to form j-particle agglomerates.
Calculated mass fractions of agglomerates (A2 - A 10 ) are plotted
as a
function of
reactor
constant
temperature
(Figures
31
g,h). A
residence time of 0.24 second is specified. Experimentally, this would
correspond to allowing the mixture of coal particles to pyrolyze, collide
and agglomerate for the
given temperature-time
history, then quenching
the mixture infinitely rapidly and collecting and counting the agglomerates.
(c) Swelling of particles
Electron
ed
micrographs
of
flow reactors show existence
large cavity
in
the center)
impart
a
foam-like
which
char
particles
fed
of both cenospheres
through
(particles with a
and particles with numerous small
appearance
to
the
entrain-
particles.
cavities
Equations
for
describing the two cases are given below.
(1) Swelling of a single bubble
A single volatiles-filled bubble is assumed to grow in a viscous,
/v
- 58 1 OOOOK /B
tc
'700K
t c,
1 5OOK/s
I OOOK
A
A
4
-
1
1.00
3.0
0
0.9
0.95
0.8
0.90
C
r
1
0.85
0.7
2
0.80
0.6
13
||
30.75
0.5
0.2
0
S
c
0.02
0
0.6
0.4
0.04
0.06
:
(
scco
B
3 L
n
Md
S
Figure 31e,f Weight Fractions of Different Size
Agglomerates Formed at E0 /v= 106 s10000K/s,
t-
1 500OK/is
O.24is
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
t-
0.24s
106
Jos
0.1
0.1
0
0
600
800
1000
1200
FinIal
1400
600
800
1000
1200
1400
Tempra-tiure,. K
Figure 31g,h Weight Fractions of Agglomeration vs
Final Temperature, No.on Curves are
E /V
, in
s-2
- 59 -
spherical
droplet
(Figure
The volatiles trapped inside the bubble
32a).
are assumed to be in
equilibrium with a constant mole fraction yi of the
metaplast present
the coal.
in
momentum
balance across
particle
radius at time t
change of
Q
liquid
layer. Defining
to the initial
particle
Q
obtained by a
as
the ratio of
radius,
the
rate
of
is given by,
c (T/T
G(1Y)[
dQ
the
Expansion of the bubble is
)
3
2a s
{
R
RJ0(
-(1-)]
4vi
S1
+-
1
Q
2
1
[Q 3 _(I-)
Q
3
1 3
-(12)
4p Q2
Poo ,c
,
T,
s
To , and
y
are the pressure outside the particle, initial
porosity, surface tension, particle temperature, initial temperature and
viscosity of the liquid/solid suspension of the softening coal, respectively.
Volume shrinkage due to pyrolysis weight loss is
The suspension viscosity is
from Equation
(1),
not considered.
with rate constants from
Equations (5) and (6).
The surface tension effect is
pressure is
only significant
when the ambient
or when the particle size is
much less than 1 atm,
much less
than 100 um. Results of calculations are shown in Figures 31a,b.
(b) Reacting Foam Model
The
swelling
of
the foam
layer is
analogous to
the process-
ing of plastic foams. An empirical approach is attempted here to describe
the
swelling
ratio.
Consider
a
reacting
mixture
of
unsoftened
coal,
- 60 -
R
=
original radius of particle
Swelling ratio Q
A.
Sin rI
IE e
IBu
lez1
S
e LLi
=
R 2 /R
ng
unsoftened coal
'liquid (foaming)
coke (resolidified)
B.-
Figure 32.
IR E:,Ca c'ting
1
Fa c
m
Mocdl
1
Limiting Cases of Swelling (a) Single Bubble
Swelling (b) Reacting and Solidifying Foam
- 61 -
liquid, gas and coke, their weight fractions denoted by C,M,G,E respectiThe swelling of the liquid is
vely (Figure 32b).
and is
tion,
assumed
the generation
rate,
to a power $ of
(via constant a)
to be proportional
dG/dt.
due to the gas genera-
inert mineral
The unsoftened coal (including
matter) in the suspension is assumed to have an initial density p 0.
porous
coke
is
formed
continuously
the
from
liquid,
porosity
its
so
The
Therefore, at any
reflects the instantaneous porosity of the liquid.
time, the volumetric swelling ratio Q(t) is given by
t
M/p0 + f [(dE/dt)/pH+E]
C/po + [ 1 + a(dG/dt)
dt
=(13)
Q(t)
(1/p )
The
of
the
numerator.
separate
terms
liquid is
related to P0 byPM+E
Q(t)
=
C, M and
volumetric contributions of
=
instantaneous
The
/[
density
three
of
the
1 + a(dG/dt)6 ]therefore
(dE/dt)[1 + a(dG/dt)
C + [ 1 + a(dG/dt) 6]M +J
the
in
E are
] dt (14)
0
Figure 33 shows predictions from this equation.
The
two
expressions
above
probably
represent
two
limiting
physical picture of coal swelling.
1.6
POTENTIAL APPLICATIONS
(1)
The
results provide
an approach
for
describing
agglomeration
behavior in gasifiers and combustors.
(2)
The
improved
physical
and
chemical
picture
of
the
pyrolysis
of
MW
10000K/s
to:
10OOK
tc,
1500OK/s
3.0
3.0
2.0
2.0
1.0
1.0
1700K
0~i
k)
0.6
0.6
0.3
0.3
Liquid wt.
Liquid wt.
fraction
fraction
0
0
0.0
0.2
0.4
0. b
se~C~conrd s
0.0
0.02
0.04
0.06
S 4a C <, rdsg
Figure 33. Liquid Content and Swelling Ratio, Heating Rate and
Final Temperature Indicated
- 63 -
improved
softening coals can be used in
description of the kinetics
of coal softening, pyrolysis and liquefaction.
showed that
(3) The data
high liquid yields,
suitable thermal pretreatment alone can give
80 % daf,
exceeding
and
is
advantageous
to the
production of liquids from coal.
(4) High temperature, fast response plastometer can be applied to
the study of other thermoplastic materials.
1.7
CONCLUSIONS
(1) A new
changing
apparent
temperatures,
conversion
been developed to measure the rapidly
plastometer has
coal
viscosities
pressures and
processes but
severe
at
heating
previously
i.e.
conditions,
rates pertinent
at
high
to modern coal
unattainable with
conventional
instruments.
(2) Coal plasticity can be described directly by its viscosity change. As
a Pittsburgh Seam bituminous coal is heated, the coal softens to form
a solid/liquid suspension the viscosity of which decreases from more
than one million poise to minimum
values
of
order a
few thousand
poise. Resolidification of the intermediate liquid phase is indicated
by rapid rise in viscosity values.
(3) Characteristic times for softening and resolidification were obtained. The
measured time
interval
of low viscosity
varies
from
one
- 64 -
minute
less
to
depending
a second,
than
on
the
is
by
temperature-time
history.
(4)
of
softening
Initial
a
Pittsburgh
coal
physical
melting
of an estimated 25% by weight of the coal. Up to 40% by weight of the
coal
is
further
to
converted
liquids
by
pyrolytic
bond
breaking.
These liquids can polymerize to give higher molecular weight liquids
crack
solids,
and
to
gas and
give
coke, or escape
from
the coal
particle to give tar. The result is that the average liquid molecular
weight
can increase
from
above 800, and
570 to
then decrease
to
around 600.
kinetics
(5) The
of
intraparticle coal
can be quite accurately determined
with
formation and
liquid
in an
electrical
depletion
screen heater
rapid sample quenching, followed by solvent extraction of the
quenched char.
(6) The new
information on coal
application
in modeling coal
plasticity kinetics
particle
finds quantitative
agglomeration
and
swelling
behavior under rapid heating high temperature conditions. Predictions
of
model here developed
were in
qualitative
accord with
literature
data on swelling and provide a possible explanation for coal agglomeration observed within a pulverized coal flame.
1.8
REFERENCES FOR THIS CHAPTER
Anthony,D.B.,J.B.Howard,H.P. Meissner,and H.C. Hottel,Rev. Sci.
45, 992 (1974)
Instrum.
- 65 -
Attar,A. AIChEJ,24,106 (1978)
Bhatia,G.,E. Fitzer and D. Kompalik, Abstracts of 17th Biennial Conf.
Carbon, Lexington, Kentucky 165-7, June 1985
on
Fitzgerald, D., Fuel 35, 178 (1956)
Fong,
W.S.,W.A.Peters
Howard,
and J.B.
Proc.
2nd Annual Pittsburgh Coal
Conf.,626 (1985)
Fong, W.S.,W.A. Peters and J.B. Howard, Rev. Sci. Instrum. 56, 586 (1985)
Fong, W.S.,Y.F. Khalil,W.A. Peters and J.B. Howard, Fuel, in press (1986)
Fong, W.S., W.A. Peters and J.B. Howrad, Fuel, in press (1986)
Frankel, N.A., and A. Acrivos, Chem. Eng. Sci. 22, 847 (1967)
Franklin, H.D.,"Mineral matter effects in Coal Pyrolysis and Hydropyroly-
sis" Ph.D. Thesis, Dept. Chem. Eng.,M.I.T. Cambridge, Mass (1980)
Habermehl, D., F.Orywal, and H.D. Beyer,"Plastic properties of coal"
Chemistry of Coal Utilization,2nd Suppl Vol,Wiley, New York, 1981
in
Loison,R.,A. Paytavy,A.F. Boyerand R. Grillot, "The plastic properties
of coal" in Chemistry of Coal Utilization,Suppl Vol,Wiley, New York, 1963
Melia,P.F.,
and
C.T. Bowman,
Combustion Science
and
Technology,
31,195
(1983)
Nazem, F.F., Fuel 5 , 851 (180
OhM.S.,"Softening Coal Pyrolysis" Sc.D. Thesis, Dept. Chem. Eng.,
M.I.T. Cambridge, Mass(1985)
ReadR.B., P.J. Reucroft and W.G. Lloyd, Fuel 64, 495 (1985)
Rees, 0.W., and E.D. Pierron, "Plastic and Swelling Properties of
Illinois Coal", Circular No.190, Illinois State Geological Survey 1954
Saffman P.G.,and J.S. Turner, J. Fluid Mech, 1, 16 (1956)
Serio,
M.A."Secondary
Reactions of Tar in
Coal Pyrolysis"
Ph.D. Thesis,
Dept. Chem. Eng. M.I.T.,Cambridge, Mass (1984)
Suuberg E.M.,"Rapid Pyrolysis and Hydropyrolysis
Dept. Chem. Eng., M.I.T. Cambridge Mass (1977)
of Coal"
Sc.D. Thesis,
Waters P.L., Fuel 41, 3 (1962)
Yoshida,
T.,Y. Kousaka
and K. Okuyama,
Power Co. Ltd, Tokyo (1979)
"Aerosol
Science
for Engineers",
- 66 -
INTRODUCTION AND BACKGROUND
CHAPTER 2.
2.1
INTRODUCTION
coals upon
Bituminous
heating
undergo a number
Melting and pyrolytic decomposition
and chemical changes.
physical
of
of significant
parts of the coal form an unstable liquid that can escape from the coal
to give tar and light oil or crack to form light
by evaporation
residue. Polymerization or cross-linking of
leaving behind a solid
molecules to form
liquid
coke
porous
The
residue.
dependent,
temperature
solids also competes with
the
processes. All
removal
gases,
is eventually
liquid
kinetics
of
liquid
leaving a
depleted,
are
strongly
relative contributions
depend on
processes
these
their
therefore
these other
the
the temperature-time histories to which the coal is subjected.
From an engineering perspective,
the transient occurrence of the
liquids in coal, commmonly referred to as plastic behavior, can strongly
influence
the
refinement
of
history,
chemistry
bituminous
reactive
tion,
is
physics
of
the
important
different bituminous
presence
for
coal
of
the understanding
pyrolysis,
phenomena,
and
temperature
or
feed size,
and
subsequent
gasification,
Additional
and reactions with surrounding
conversion
process
solvents
of
carbonization and liquefaction processes.
transfer
thermal
the
Irrespective
coals.
atmospheres,
plasticity kinetics
modelling of
and
combus-
heat and mass
reactive
atmospheres
or solvents can be suitably superimposed to describe the overall behavior
in different processes.
- 67 -
Some existing technical problems include the serious difficulties
encountered in fluid bed gasifiers are direct consequence of plasticity.
affecting
turn
in
qualities
volatiles
products
pyrolysis
of
through
the
rate. The
devolatilization
the
viscous
are
since
melt
affords
and
yields
transport
of
opportunities
for
the
affected
coal
plastic coal
swell the
and plugged pipelines. Bubbles of volatiles can
mass,
of the bed,
of plastic particles can cause solidification
Agglomeration
intra-particle secondary reactions. Swelling also determines the density
and morphology of the char, hence the burnout time and process efficiency
in
subsequent
impinge
and
combustion,
erode
as
well
heat transfer
as
the
propensity
surfaces. The
of
development
char to
the
disapp-
and
earence of plasticity is closely related to formation and destruction of
liquids. Study of
plasticity
light
shed
kinetics can therefore
on the
mechanisms of coal liquefaction,
Therefore, plasticity study contributes to both basic understanding in coal science, and
to improved design in
several
important coal
conversion technologies.
Figure 2.1.1.
is
a schematic
of changes in
coal physical struc-
ture and liquid volume fractions during the onset, duration and destruction of plasticity.
Plasticity
kinetics
under
rather well studied (Loison et al.,
studies involve
carbonization
conditions
1963; Habermehl et al.,
has
1981).
been
These
low sample heating rate experiments to a moderate final
CHANGE
PHYSICAL
HEATING
UPON
do*
-
raw
CcOal
-do _dv _M
S.--
SO
F tenk
-
L
L
Iri
S
mW *a
irig
00'
44-.--
Mm bLI M t I C3 Mri
1.0
vcl
tm m e
cri
Fac
ti
0.0
heat
Figure 2.1.1
I "ig time
Physical changes of bituminous coal upon heating
- 69 -
been made
at 0.05
typically
temperature,
K/s to 800 K. Very
to date outside this range,
these
of
ploration
data
processes,
extrain
made
been
have
some
1981; Attar,
1983; Scharff et al.,
1976). In modern coal conversion and combustion
and Mills,
James
studies
although rather uncertain
conditions have
to other
modelling studies (Melia and Bowman,
1978;
limited
and final tempera-
104 K/s or higher,
heating rates of order
tures to 2000 K are possible. This thesis addresses a systematic approach
of coal plasticity
to the study
at high heating rates and temperatures
pertinent to modern coal conversion processes.
BACKGROUND
2.2
The major reviews on coal plasticity are to be found in Loison et
al.(19 6 3 ) and
Habermehl
et al.(1981). Theories
of
plasticity
coal
are
oriented towards conditions pertinent to metallurgical coke manufacture.
Rather limited information exist for high heating rate, high temperature
cnditins. Bituminous coals
transient
plastic
heated to about 400*C or above will exhibit
properties.
Coal
will
particles
soften,
swell
and
ultimately coalesce upon heating. Fluidity of the coal melt will increase
to a maximum and then drop. Further heating causes the melt to resolidify
to
a
solid
theories of
coke. Lloyd
coal
et al.(19 8 4 ) gave a
plasticity.
Earlier
summary of the
"bitumen"
theory
explains
development as a result of a fusible component
(bitumen)
provide a viscous slurry. The later "metaplast"
theory is
accepted
metaplast
tions,
(Fitzgerald,
is
generated
1956;
Chermin and Van Krevelen,
and destroyed
by
the
following
different
plastic
which melts to
more commonly
1957).
The liquid
pyrolytic
reac-
- 70 -
coking coal
metaplast
-
semicoke
->
coke
-
secondary gas
primary gas
The liquid metaplast concentration determines the fluidity of the
melt. Kinetic constants of formation and destruction of the metaplast can
be calculated from experimental fluidity data
(Lloyd et al., 1980).
The onset of plasticity is generally regarded as being caused by
a depolymerization reaction of some components of the coal matrix during
pyrolysis producing a liquid "metaplast".
a
plasticizer
(around
pressures
system to become
anthracites and
as
fluid. At moderate
normally limited to coals in
plasticity is
1 atm)
range,
bituminous
the entire
and cause
This metaplast can then act as
low rank
the
have molecular
coals
structures that are too highly cross-linked for complete depolymerization
to occur.
Resolidification
coke is
regarded as occuring
from
evaporate
present
or rehardening
in
reactions
the
coal
the system. In
such
temperatures,
as
causing
it
the metaplast
by two processes. Parts of
surface,
reducing
addition,
cracking,
of the coal metaplast to form a
the
amount
the metaplast
condensation
to resolidify.
The
and
of
plasticizers
undergoes
at
crosslinking
phenomena
of
chemical
higher
plasticity
increasing with increasing pressure can be accounted for by the increased
plasticizer
concentration
in the
system caused
by
the lower
rates of
- 71 -
intra-particle and external mass transfer that occur at higher pressures.
Temperatures of rehardening increase with increasing heating rate of the
coal
sample, indicating a competition between the condensation and other
reactions of the system. Evidence for
both mechanisms
participation of
according to both theories is substantial.
Kirov
and
(1967)
Stephens
discussed
the
close
relationship
between the fluidity measured by a Gieseler plastometer and the rate of
weight
loss and yield of chloroform
The maximum of fluidity is
extractables for carbonized
reached about 10 -
coals.
20'C after the maximum of
the extraction curve and 20'C before the maximum of rate of weight loss
curve when these data are plotted with the final temperature of carbonization
as
relation
the
independent
between
coals. Recent
Gieseler
studies of molecular weight
in flash
pyrolysis of
mtnlct
transport
in
nlastic
fine
cnAl
(Unger
extraction
yield
distribution and yield of
coal
the
discussed
al. (1983)
solvent
and
fluidity
generated
et
Ouchi
variable.
particles shed
of
tar
light on the
and Suuberg,1983;
Oh, 1984).
Their findings suggest that under rapid heating conditions, plasticity is
determined
by
the
interplay
of
rate
processes
such
as
thermal
bond
breaking, tar and reactive gas transport, and secondary repolymerization
reactions. The complexity of the plasticity phenomena is well demonstrated by Figure 2.2.1. with respect to metallurgical coke formation (Klepel,
1960).
Knowledge of the viscosity of a softened coal would be useful for
studying the mass transfer effects in softening coal pyrolysis. There is
abundant
evidence that mass
transfer
plays a
role in
determining the
- 72 -
Cause
Manifestat ons
Test method
EL
E
Figure 6.14 Relationships between causes and experimental events for the different methods of testing of
coke formation according to Klepel. 5'
Figure 2.2.1
Role of plasticity in coke formation
according to Klepel (from Chemistry of Coal
Utilization, 2nd Suppl. Vol., 1963)
- 73 -
1982)
volatiles obtained
of
yields
with intra-particle
pyrolysis
during
mass
being
transfer
Gavalas,
(Howard, 1981;
many
for
limiting
rate
conditions of practical interest (Oh, 1985). Intra-particle transport of
reactive gases and pyrolysis products can be achieved by diffusion in the
liquid phase and by
and
liquid diffusion coefficient
also
Viscosity
modeling.
such
in
important
are
viscosity
melt
bubble transport. The
influence the dynamics of bubble growth (Oh, 1985; James and Mills, 1976;
Lewellen, 1975).
Rheological
measurements
coal
on
exclusively
are almost
made
under coking conditions, i.e. low heating rates and up to moderately high
temperatures.
Detailed
Kirov
techniques of measurements are discussed in
and Stephens (1967) and ASTM D2639-74 (1974). The Gieseler plastometer is
widely
used.
is
coal
powered
A finely
into
compacted
a
cylindrical
crucible containing a bladed stirrer. A constant torque is applied to the
stirrer as the coal is
heated at 3K/min. The angular velocity (100
dial
divisions/min) is recorded in units of dial divisions per minute (ddpm).
Most plastic coals have a distinct softening stage followed by a resolidthe transient
ification stage. At 3K/min heating rate,
between 400 0 C -
500*C.
The
plasticity
curve.
spatially
complex
Figure 2.2.2.
stirring
flow
field,
shows a Gieseler plastometer and a
subjects
arm
and
varying range of shear rates. Other
plasticity occurs
the
the
softened
measurement
is
coal
over
a
limitations of the Gieseler
to
a
widely
plasto-
meter are discussed by Kirov and Stephens (1967). The coal melt actually
consists of a liquid phase of chemically reacting polymeric substance in
which is
suspended a mixture of gases and solids (including
matter and
possibly non-softening
coal
macerals).
It
is
coal mineral
obvious
such
a
3
-Dial
Wei ght
V2
V
2
_
_
40
3w
O
0A;-N
:0 or
sw
(From Chemistry of Coal Utilization, 2nd Suppl. Vol.)
Figure 2.2.2
Plasticity data from Gieseler Plastometer
- 75 -
melt cannot be Newtonian. Waters (1962) found coal to have pseudoplastic
behavior
to
visco-elastic
a
suitable
such
the early
in
model.
Brabender
as the
cylinder
stages. Fitzgerald
measurements.
viscosity
viscometer
are
1963)
(Stephens,
operate
at
and moderate
final
is
not
plastometers
Other
and the
for
designed
all
reported viscosity measurements with a capillary viscometer.
rate
data
have
Davies et al' (1983)
making measurements under coke-oven conditions.
heating
fitted
plastometer
the Sapozhnikov
plastograph,
(1957)
plastometer
constant-torque
a
Hence
making apparent
for
concentric
softening
temperatures,
Besides low
also
these plastometers
fluidized
shear rates much lower than those experienced in
beds. For higher heating rates (up to 100 K/min), a shock dilatometer can
give
qualitative
data
behavior
plastic
on
as
inferred
swelling
from
behavior. Khan and Jenkins (1984) have used a high pressure microdilatometer to measure plastic properties at heating rates of 65 K/min.
Thus,
previous studies did not meet the need for measuring the rapidly changing
viscosity of coal under high temperature, high heating rate and different
reactive gaseous atmospheres.
Swelling,
direct
consequence
of
the
caking
of
coal
particles is
properties
of
coal.
To
and
agglomeration
plastic
avoid
a
these
problems, there are two approcahes. The first is to decrease the softening properties of coals, which can be achieved by pre-oxidation, partial
pyrolysis and pretreating the coal with chemicals. Many chemical decaking
treatments
second
approach involves
The extent
period,
have been discussed in
of
the literature
(Habermehl,
1981).
The
the selection of suitable process conditions.
agglomeration depends
surface tension and the
on
actual
the
duration
of
the
plastic
viscosity of the particles,
and
- 76 -
the flow field of the continuous
Experimental
medium.
studies of agglo-
meration have been described by McCarthy (1980) and Tyler (1979, 1980). A
study of free thermal agglomeration of two coal particles heated at 3 - 7
K/min was
made by
agglomeration
in
Klose
data
Viscosity
gasifiers.
have
(1981)
al
(1983). Scharff et
in
modeling
modelled
are
usually
assumed or extrapolated from data under coking conditions.
Swelling
of
powered
packed,
particles
coal
heated
in
small
diameter tubes have been studied in dilatometers up to 65 K/min (Khan and
1984). Swelling
Jenkins,
of individual
particles
coal
at
rates have been studied by Hamilton (1980), Culross (1984),
high
heating
Matsunaga et
al.(1978) and more recently by Lowenthal (1985). Models of swelling have
been proposed by Melia and Bowman (1983), Lewellen (1975) and Oh (1985).
The plastic behavior of coal has other important consequences in
coal
conversion processes. One
kinetics.
ngh
Hedden,
1983).
(Whitehurst,
1980)
A
(1982)
on
reactiviLy iU
close
and plasticity has been observed
Neavel
effect
its
is
fILUIiLy 1S desirable f-or Uig
time liquefaction
liquefaction
example
liquefaction
otl
relationship
(Gorin,
cotac
between
1981; Wilhelm
discussed plasticity mechanism
and
from a coal
84
) studied viscosity changes in
liquefaction viewpoint. Okutani et al.(19
coal/oil
pastes during hydrogenation
and
found that the existence of a
maximum in the slurry viscosity was due to extractive distegration of the
coal by the vehicle oil.
A brief summary of the different
factors affecting coal plasti-
city will be given below. Most of the observations are from experimental
- 77 -
results from Gieseler plastometers operating under carbonization (coking)
conditions.
Coal
-
type
plasticity
is
in
normally observed
bituminous
coals. Coals
with 25 - 30 % volatile matter exhibits maximum fluidity.
is
Fluidity
also evident in
coals with carbon content (daf)
in the range 81 - 92 %, with maximum fluidity at 89%. Vitrincomponent
the main petrographic
is
ite
that contributes
to
fluidity (Van Krevelen, 1961)
Heating
rate
-
an
increase
in
the
heating
rate
shifts
Gieseler
the
softening and resolidification points to higher temperatures,
broadens
the
plastic
(Loison, 1963). Swelling
and
region
is found
fluidity
increases the
to increase
with heating
rate at the low heating rates (to 10 K/s). Swelling may
actually decrease at high heating rates or high temperatures
(Pohl et al., 1978).
Particle
si ze - very fine grinding of coal
reduces its
plastic proper-
ties. Swelling of particles decreases with particle size.
Pressure -
the effects of pressure have only been studied rather
recently.
In
inert
gases,
the maximum
fluidity
increases
with pressure to some limit and then levels off. In hydrogen,
the increase is almost linear in the range 0 - 450 psig.
(Lancet and
Sim,
under pressure,
1981;
Kaiho
and Toda,
1979)
the caking of the particles is
When heated
promoted and
- 78 -
the swelling increases.
Chemical pretreatment and additives - pre-oxidation of a bituminous coal
Ignasiak et al.
caking properties.
will destroy its
due to hydroxyl
is
found that the destruction of plasticity
( 1 9 7 4 )
reactions during pyrolysis.
groups undergoing condensation
Alkali salts reduce coal plasticity by promoting crosslinking
reactions in the coal (Crewe et al,
1975). A large number of
behavior.
additives are capable of changing the plastic
itives
plasticity by
the
affect
catalysis, interaction of
coal and additives,
with the
surface
reaction
chemical
of
thermal conductivity
the
plastic
of the plastic
coal
the
between
additive
by changing
or
(Habermehl
mass
Add-
the
et al.,
1981).
2.3
THESIS OBJECTIVES
The thesis objectives include the following:
(1) Conceptualize, design and develop new techniques and instruments for
liquid quantities
measuring apparent viscosities and intra-particle
of a rapidly pyrolyzing coal.
(2)
Study
pressure
the
effects
of
on plasticity
process
kinetics
variables
and
such
investigate
as
temperature
some
and
qualitative
aspects of the nature of plasticity and softening coal pyrolysis.
- 79 -
(3)
Develop
global models of plasticity kinetics
eration and particle swelling,
phenomena.
and models
of agglom-
and possibly other plasticity related
- 80 -
CHAPER 3. EXPERIMENTAL TECHNIQUE AND APPARATUS
3.1
INTRODUCTION
The
rheological
properties of
methods
experimental
non-existence of
for
studying the
under rapid heating conditions
softened coal
have already been discussed. In modern coal conversion processes, heating
rates of 15000 K/s or more can occur.
Final temperatures can exceed
1100 0 C and shear rates can be high (estimated to 1000 s-1).
for a new coal
plastometer therefore require an ability
Requirements
to operate at
(to 1100 0 C) and heating rates to 1000 K/s.
high temperatures
monitor the rapidly
is
changing viscosity
important
Ability to
order to follow
in
the short duration of plastic phase at high temperatures.
of
Observation
heating
reflects only
the
deformation
indirectly the
swelling
and
behavior
during
behavior. A direct method
plastic
would be to determine the resistance to shear of the coal,
or to deter-
mine the amount of liquid present in the coal during pyrolysis.
Two
new
experimental
study. To measure
methods
the apparent
response plastometer for sample
were. developed
for
the
viscosities of pyrolyzing coals,
heating rates
present
a fast
and holding temperatures
of 40-1000 K/s and 600-1250 K, in vacuum or in an ambient gas atmosphere
up to 100 atm was designed and constructed.
the liquids content
in
To determine quantitatively
the pyrolyzing coal,
an electrical
screen pyro-
lyzer with rapid sample heating and quenching capability (up to 1000 K/s,
and
6000 K/s
respectively)
was also designed
and constructed.
From the
- 81 -
data,
pyrolyzer
devolatilization
liquid
and
formation/destruction
plastometer
data. A second
larger pyrolyzer with gas and tar collection capability,
higher heating
were studied and
reactions
and
rates,
for
allowence
compared to
photographic
was
experiments
of
recording
techniques and devices for both plasto-
later. Sample loading
developed
the
meter and pyrolyzer were also developed and tested.
Reactions in the transient liquid phase were studied in terms of
Chromatography,
molecular weight evolution by Gel Premeation
by
growth
solvent
system (a heat treated
liquid
coal
and
tar),
coal
phase fuels
solid
other
solids
a model
of
experiments
extractions. Pyrolysis
and
(wood and lignite) were also performed and compared to that of a plastic
coal.
experiments
additional
Some
etc)
experiments,
also
were
performed.
pre-treatment
experiments,
(foil
is
3.1.1.
Table
a
summary
of
experiments performed.
APPARENT VISCOSITY MEASUREMENT
3.2
A fast
has
been
response,
developed
for
apparent
temperature
high
rapid-heating
measurement.
viscosity
The
plastometer
instrument
determines the torque required for constant speed rotation of a thin disk
embedded within a thin layer of coal
final
temperatures can
ranges
operated
atm.
final temperatures
Heating rates,
plates.
40-1000
in
K/s,
hydrogen
heated
between two parallel metal
and sample residence times at
be separately selected
600-1250
or inert
K,
and
0-40
gas atmospheres
and
s.
controlled
The
from
over the
instrument
can
be
near vacuum
to
100
- 82 -
Table 3.1.1.
EXPERIMENTS
Plastometer
ABBREVIATION
Summary of Experiments
PARAMETERS STUDIED
PO
prototype development and testing
PT
heating rate and holding temperature
PP
pressure effects
PE
effects of pre-extraction
PM
miscellaneous plastometer experiments:
calibration,other materials,H2 effect
Screen reactor
SB
bituminous coal pyrolysis
ST
model liquid (coal tar) pyrolysis
liquid amount
SE
pre-extraction effect on
SF
foil experiments: mass transfer and
secondary reactions
Gel permeation
SM
miscellaneous: wood,lignite pyrolysis
GB
molecular weight distribution (MWD)
evolution in liquid, bituminous coal
chromatography
GT
MWD evolution in coal tar model system
GM
miscellaneous: MWD in wood, lignite
- 83 -
Apparatus description
a schematic of the coal plastometer.
presents
Figure 3.2.1.
that can be evacuated to 5 Pa or charged with hydrogen or
205 atm)
for
steel rated
(316 stainless
unit is enclosed in a high pressure vessel
The
inert gas. One end of the torque transducer shaft is joined to the shaft
of
while the other end is
the shearing disk,
CYQM
(Barber
motor
1207).
LV
(Minarik
23410-61-5,
driven
1.70Nm)
The torque transducer
type torquemeter (S. Himmelstein
to a 12V DC gear-
coupled
by
strain gauge
a non-contact
is
controller
speed
a
MCRT 3L-08T and System 6 readout,0.1%
The response time of the transducer is less than 10 ms.
accuracy).
details
The
of
shearing
the coal
device are
illustrated
in
Figure 3.2.2. The shearing chamber is a 38.1 mm x 10.2 mm x 0.76 mm slot
formed by upper and lower bounding plates.
It
supports a 0.50 mm thick,
3.75 mm radius axially symmetric shearing disk. These parts are machined
from
superalloy
a nickel
resistence
and
strength
(Rene 41)
and
duration,
information.
including
and
ease of
of
its
temperature
high
The disk and
for convenience of instrument
data
inversion
The demanding experimental
shearing
for
and hydrogen corrosion.
to alkali
flat plate design was chosen
selected
refactory
to yield
construction
apparent
conditions of interest
solid particles and
viscosity
for coal,
operation at high
temperature and high heating rates also favored this design.
Desire for
spatially uniform sample heating and minimization of secondary reactions
in
the
coal layer
consistent
with
suggested use of as thin a sample layer
maching tolerances
and accommodation of
as possible
representative
w
w
I NDEPENDENT VARIABLES
HEATING
CIRCUIT
,
HOLDING TEMPERATURE
AND HEATING RATE
MOTOR SPEED
CONTROL
HYDROGEN
HELIUM *ETC
-~
FAST RESPONSE
RECORDER
SHEAR RATE
TOTAL PRESSURE AND
REACTIVE GASES
(
TORQUE AND
TEMPERATURE READING
RAW
Figure 3.2.1
Schematic of coal plastometer
I
DATA
)
t To transducer
3.75cm
Front Sectional View
Radius =0.375cm
1.93 mm:-
1*-
1.35 c m -- 4
Side Sectional
View
(1) aluminum electroes;
Details of shearing chamber:
thermocouple (0.012 mm thick(3)
clamp;
steel
(2) stainless
(coal); (6) nickel supersample
(5)
disk;
shearing
ness); (4)
ceramic
temp.
high
(7)
alloy plate;
Figure 3.2.2
Details of plastometer hot stage
- 85 -
- 86 -
size coal particles. The shearing disk was thus positioned to establish a
symmetrical 0.13 mm thick layer of coal between each of its faces and the
corresponding bounding plate of the chamber, as shown in Figure 3.2.2.
thermal
The
coal
single,
by
be estimated
can
and
m2 s-
to the center plane
estimate
be
would
K/s
since
of the coal
5
thermal
1
sample as
a
, the temperature drop from the surface
layer at heating
diffusivities
across the metal
the
to
primarily
layers of
the two
50 K respectively.
and
due
assuming a low value of coal thermal
disk is
rates of
100 K/s and
is a
conservative
This
5 x 10-6 m2 s-1 at
much
faster since
with
rapidly
increase
coals
of
temperature above 773 K, to values of
conduction
is
arrangement
treating
slab. Then,
0.26 mm thick
diffusivity of 1.5 x 10 -7
1000
this
lag of
1123 K. Heat
the metal
has
a
1
2
thermal diffusivity of about 5 x 10-6 m s- . For the 0.50 mm thick disk,
the
thermal
conduction
time
through
the
disk
would
order
an
be
of
magnitude smaller than through the coal layer. Measurement of temperature
is
by a 0.012 mm thin-foil thermocouple
outside
surface of the upper metal
(Omega C02-K) attached to
the shearing chamber. Its
plate of
stated response time to a stepwise temperature change is
changes
in
electronic
the torque
recorder
and temperature
are
the
recorded by
(Bascom-Turner 4120) capable
2 - 5 ms. Rapid
a
two-channel
of acquiring
data
at
1000 data points/s.
The
coal
sample is
heated by
two
independent DC
circuits as
shown in Figure 3.2.3. Power is supplied from a set of parallel connected
12 V lead acid batteries.
During an experiment,
two essentially constant
pulses of current are sequentially obtained from these two circuits and
qw
MWM
Electrical circuits for independent control of sample heating
time.
rate, final temperature and heating
UP CIRCUIT
VOLTHEATERY
CARBON RHEOSTAT
800A
PLASTOMETER
00
SIX CONTACTORS
IN PARALLEL
HOLDING CIRCUIT
12 VOLT
BATTERY
-
2 0 0A
V
CARBON RHEOSTAT
TIMER
6v
TIMER
CONTACTOR
120 VAC
Figure 3.2.3
Electrical circuits for sample heating
- 88 -
delivered through the metal plates for heating up and holding the sample
Two
respectively.
temperature,
and
(Biddle #410040
carbon rheostats
#410010) are used to adjust the magnitude of these current pulses, which
can exceed
digital
relays
(Agastat
DSB
adjustable
from
timers
These
XX0125).
0.00
to
in
s
99.99
turn
two solid
by
500 A. Timing of the current pulses is
0.01
in
s
increments
heavy
activate
state
DC
duty
contactors (Stancor 124-909) which activate or deactivate the heating and
holding temperature circuits.
The device in Figure 3.2.4. was developed to permit reproducible
loading of the coal particles into the shearing chamber and alignment of
the rotating disk with the axis of the transducer shaft.
Procedure
Approximately 190 mg (
± 2% by wt. max. error) of ground coal are
used in each experiment. Using the
coal
is
uniformly
placed
between
the
lower
plates Ml and M2 while holding the disk in
tion is
blocks D. Some coal
is
and
upper
poured
bounding metal
proper position. This opera-
holding Ml down (on clamp I)
achieved by first
the
loading device (Figure 3.2.4.),
into the channel,
i.e.,
with displacement
the half-slot
of
Ml, and spread into a uniform layer of prescribed thickness by sliding a
grooved smoothing block S along the shoulders of D. The rotating disk K
is then lowered onto this layer of coal, aligned, and secured. More coal
is
poured over K and a second,
along
the shoulders of D,
more shallow smoothing block S'
leaving a uniform
thickness covering the disk. The displacement
layer
is
slid
of coal of a preset
blocks D are then removed,
W
1W
1W
W
W
I'
K
> M2
I
CO
D
DD
Device for uniform loading of coal in shearing chamber:
C - centering rod; D - displacement blocks; 1,I' - stainless
steel/ceramic clamps; K - rotating disk; M ,M
- lower and
upper heating plates; S,S' - sliding blocks; T ing; P - coupling.
Figure 3.2.4
teflon bush-
Device for loading coal into shearing chamber
S"
-
90 -
and M 2 is rotated into alignment with Ml and lowered carefully on Ml. In
the surface of
order to achieve good contacting between the two plates,
Ml must be cleaned of any coal remaining from the loading operation. The
disk
plates,
metal
C-shaped
metal/ceramic
other
each
against
surfaces
I
pieces
their
with
then
between
secured
the
on
pressing
faces
ceramic
two
tightened
and I'. These two pieces are
outer
of the disk
prevents dislocation
Ml and M2 . This procedure
of
are
assembly
and coal
and the metal plates.
The
assembled
the two
between
and
electrodes
loaded
shearing
in the plastometer. The
then
is
chamber
couplings
placed
to
the
transducer are tightened and the thin-foil thermocouple is then attached
to
outer
the
during
face of
M2 . To
protect
startup the shearing disk is
the
rotated
overload
transducer against
manually
until a steady torque in a measurable range (4 -
one
or
two turns
5 Nm) is attained. The
cover of the pressure vessel is then lowered and the vessel is evacuated
to approximately 5 Pa and flushed twice with helium at 0.1 MPa. The test
gas (helium or hydrogen)
reachedThe
drive
is
motor is
then admitted until the desired pressure
started
0.5
s
before
the heating
is
relays are
activated. The torque vs time, and temperature vs time profiles throughout
the
entire run
points/s, on an
subsequent
are measured,
both
electronic recorder and
data processing.
at
an
acquistion
also digitized
of
200
stored
for
speed
and
For experiments under vacuum the plastometer
unit is enclosed in a pyrex chamber to allow direct observations of the
hot stage during operation.
91 -
-
Typical results
temperature-time histories
types of
Two
the
in
were employed
study. Holding temperature runs consist of a constant heating rate to a
temperature
final
which
was
maintained
then
constant
an
for
extended
period of time. Constant heating-rate runs refer to increasing the sample
temperature at a constant rate for the duration of the experiment.
Experiments to date have employed a Pittsburgh Seam bituminous
coal
75 pm particle diameter,
matter and 63 -
of 39 % volatile
vacuum
dried at 383 K for 4 hours. The coal, which was from the same mine as the
3.2.1.
ures
is
Suuberg
ground
freshly
Franklin
(1977),
under nitrogen. Table
a summary of proximate and ultimate analysis of the coal.
Fig-
run
data
a
disk
and
3.2.5.
obtained
was
(1974),
Anthony
and
(1980)
by
this laboratory
previously in
samples used
under 0.1
and
typical
present
3.2.6.
3.5
holding
respectively,
helium,
MPa of
rotational
speed of 0.67 rpm (corresponding
1.32 s-1).
In
temperature
at
to an average shear rate of
the former run the coal was heated at 461 K/s to a final
temperature of 874 K and then held at this temperature for 6.0 s. As the
increases,
temperature
the torque decreased to a low value due to liquid
frormation. Following a period
of
low viscosity,
the
viscosity
of
the
rises rather rapidly due to progressive resolidification of
molten coal
the melt. However, continued rotation of the disk eventually
breaks up
the resultant coke, and the torque vs time curve fluctuates but maintains
an overall
indicated
decrease.
in
Figure
The
stages
3.2.6. For
of
physical
present
change
of
the
purposes the plastic
coal
period
are
is
arbitrarily defined as the interval during which the torque is less than
- 92 -
COAL
Hi-gh
Vol
Nr.8
rghL
Pittsbu
I ED
srUD
til
Bi
Ash
1.4
11.5
68.8
H
4.9
Co.al
, dry basis )
FC
39.4
49.1
,
dry basis )
0
N
S
Ash
8.2
1.3
5.4
11.5
+ as-received basis
Table 3.2.1
-c
VM
Ultimate analysis ( wt.%
C
A
tu -Am inousc:
Proximate analysis ( wt.%
Moisture+
S4Ea m
Characteristics of Coal Studied
HEATING RATE 461 K/S
IN 0.1 MPa HELIUM
5
4
E
z
N
bin.
3
I
0
ha.
C)
am
C)
E
C)
H
2
H
0
0
2
4
6
Seconds
Figure 3.2.5
Raw data from plastometer, in 1 atm of helium
8
10
(JJ
5
4
E
z
b
3
0.
E
Q)
3 2
0
0
2
6
4
8
Seconds
Figure 3.2.6
Raw data from plastometer, in 35 atm of helium
indicating different stages of plastic behavior
10
'
- 95 -
4
1.4 x 10-2 Nm which correspond to a viscosity of 3.6 x 10
Pa s.
The geometry of the shearing chamber can be approximated
two sets
of
parallel
disks. The
rotating disk viscometer
torque
7
can then
standard
equation
be employed to
as
for a concentric
relate the observed
to the apparent Newtonian viscosity 1, for two sets of disks,
7/7rR 3
where in= ,fR/H is
(3.2.1)
the shear rate at a disk radius of R, f1is the disk
angular velocity, and H is the distance between the parallel faces of the
moving and fixed disks. A calibration factor b is
for the deviation of the actual instrument
introduced to account
geometry from two parallel
disks. Thus
b
where
,,b is
(3.2.2)
=0
the viscosity calculated
torque using Equation (1),
and 1. is
from the experimental
values
of
the true viscosity. The quantity b
is found using liquids of known viscosity
(Cannon Instrument R450000)
gives b = 1.20 at a shear rate iR= 1.98 s-1.
That b exceeds unity is
reasonable since the drag force experienced by the
disk exceeds that
expected from two sets of parallel disks due to the larger surface area
of the bounding metal plates. At a shaft speed of 0.67 rpm (or iF = 1.98
s-1),
this calibration factor implies that an observed torque of 1.0 x
10-2 Nm is equivalent to a true apparent viscosity of 2.6 x 104 Pa s.
- 96 -
For comparison purposes, plasticity curves for some other materials were also obtained.
is
a
liquid
viscous
(Figure
ture
slight
coal
its
viscosity due to coke formation,
Quartz
dependent.
resistance
The
formation.
coke
to
due
particles
expansion
of
the
do
low melting
point
shearing
torque value although
it
increase
to
it starts
a
constantly
being temperathey
and
quartz
the plasto-
in
shows
tar
offer
high
probably due to a
drop is
lignite
chamber. Montana
does not
and
lignite,
behavior
soften
not
The initial
to the rotating disk.
decreasing
(which
softens very rapidly upon heating, then
meter. The pitch
increasing
Montana
These materials show expected
3.2.7.).
resolidify
a non-softening
a
and
temperature),
room
at
3.2.8) and
pitch,(Figure
coal tar,
These included a heat treated
soften. The
shows
decrease
a
is
probably due to volume decrease upon pyrolysis. Several experiments with
the same temperature-time histories were performed to test the reproducibility
of
data
the
especially
reproduceable,
minima
(Figure
the
3.2.9.). The viscosity
softening
portion
of the curves do show fluctuations,
of
curves are quite
the
curves.
The
probably due to existence of
some single non-softening macerals. The pitch softens rapidly due to its
low melting
point. Coal tar shows a slower
resolidification to
form a
weak coke.
Since plasticity arises from the formation of liquid within the
coal
continuum,
valuable complementary information
can
be
obtained
by
determining the amount of liquid in a rapidly pyrolyzing coal particle. A
technique for the quantitative study of intra-particle liquids formation
and depletion reactions was therefore developed as described below.
1ww1w
5
1w
954K
E
z
10
MONTANA LIGNITE
Cr
U5-
957K
0
0
Figure 3.2.7
Seconds
10
Behavior of quartz and lignite in plastometer
1w
w
'W
w
1w
7.2
01
0
U 3-6
01
00
'-I
0 L
0
1.0
2.0
3.0
S
Figure 3.2.8
5.0
4.0
-
c C n d~s
Behavior of pitch, coal tar and bituminous coal
in plastometer
0
'7.
2
Bitu
C
runs~
-Three
S3.6
coc~al
iminouDAs
all
at
00
0
1.0
2.0
3.0
4.6
S e co
Figure 3.2.9
Reproducibility of plasticity runs
5
nmd s
-
3.3
100 -
LIQUIDS FORMATION AND DEPLETION MEASUREMENT
Coal
an electrical
samples were pyrolyzed
screen reactor
and
residence
for different
the liquid
left in the solid residue
after rapidly quenching the sample was determined by solvent
of
the residue. The coal used was same as
times using
extraction
that studied in the plasto-
meter.
Apparatus
The reactor (Figure 3.3.1.) was developed from the basic version
of the one described by Anthony et al.
(1974). Some important differences
include a larger reactor volume and screen size.
A much larger screen to
sample weight ratio was employed to ensure a single particle coal layer
during
the experiments.
le secondary
reactions,
This minimizes
the possibility
of extra-partic-
and improves heat transfer to the particles.
A
special sample loading device was developed for uniform loading onto the
screen. Addition of a liquid nitrogen quenching mechanism allowed precise
control of the pyrolysis residence
screen were spring-loaded
times. The electrodes holding the hot
to keep the
screen
flat
upon heating.
A more
powerful 12 volt DC heating circuit with improved control electronics was
installed. The reactor is designed for pyrolysis measurements at atmospheric pressure and also under vacuum. The reactor vessel is a cylindrical
pyrex pipe 23 cm in
diameter and 23 cm long.
reactor are two 3/8 inch thick stainless
The top and bottom of the
steel plates,
which are sealed
to the reactor with 0-rings. The bottom plate is equipped with electrical
leads, thermocouple feed and gas inlet/oulet ports, and a liquid nitrogen
RAPID
QUENCHING
SCREEN
REACTOR
tt-S
urg- <
Sea Cm
tu Xrmni Mnos
to temp.
-
90
1
atm
&m
hkieliam
recorder
to timer & 12 VDC power
to vacuum pump
to helium supply
to timer & liquid nitrogen
Figure 3.3.1
Rapid quenching screen reactor
H:
- 102 -
valve.
About 15-20 mg of 75-90 um diameter particles were spread evenly
in the central region of two layers of 10.5 x 5.3 cm, 400 mesh stainless
electrode blocks.
screen mounted between two
steel
The reactor
then
is
closed and evacuated to 1 mm Hg, flushed two times with helium, and then
helium is admitted to 0.85 atm absolute pressure. Heating was achieved by
sequentially
preselected
passing
duration
two
constant
and magnitude
pulses
of
through
electrical
current
the electrodes.
of
The volatile
products readily escaped from the screen and were quenched and diluted by
the
cold
interval,
ambient
gas
pressurized
automatically
at
0.85
liquid
opening a cooling valve,
valve,
or
less
the
nitrogen was
1500 K/s. Other runs featured
different
At
atm.
of
sprayed
resulting in
cooling rates
than
end
200
K/s
the
onto the
heating
screen
by
a quenching rate of
exceeding
when
preset
the
6000 K/s using a
screen
was
allowed
to cool by natural convection. Figure 3.3.2. shows some temperature-time
histories from thermocouple readings acquired at 200 data points/s.
The weight
loaded
loss of
the sample
was
determined
determined
(tar,
by
light
oil
and
gases)
2 hour Soxhlet extraction
using pyridine at its
boiling point
weighing the
This corresponded
screen before and after the experiment.
volatiles
by
The
yield.
extract
to the
yield
was
of the char left on the screen
(1160C),
drying the extracted
char
and screen in a vacuum oven at 110 0 C for four hours and reweighing. This
quantity reflects the intra-particle inventory of liquids at the time of
quenching.
3.3.3.
All
yield
data
reported
shows the volatile yield,
extract
as a function of the heating time,
ries.
are
based
on
a
and pyridine
dry
basis. Figure
insoluble plotted
for two different temperature histo-
scrri-
h
-- Et
re-C-Cac( toC r
S temp
4er ra
tu 1
C-re
ii
s t-Cry
A
-
~1r
717
S
~
H
w
w
v
7L:__
M-7
i
tr
_
p
1
~-~
U~rt 1~
tII~~g,1-i~
Horizontal scale is 0 -
=iEI
4
thrm,~r~rimn
~
n
iJ, m
10 seconds
Chromel-alumel thermocouple used
Corresponding to 660 - 670 K/s heating rate,
and holding temperatures 697 - 828 K, quenching at 4500 K/s
Figure 3.3.2
Screen reactor temperature histories
ss,,
Wt.
&
trct
pyr-idi riE
insol.
100
100
Run
heating rate
80
_U
H
3
holding temp.
80
cooling rate
a
60
60
-
A
a
40
40
20
20
470 K/s
813 K
1100 K/s
3:
B
446 K/s
100 i
10
5
0
e
I
15
I
20
I
0
Scnds
I
0.5
100l1r
1.0
'
.
1.5
Ir
2.0
I
Secod
i
858 K
1100 K/s
C
80
~1
3
80
3300 K/s
60
4500 K/s
8
908 K
60
a
A
0
D
40
*M40
20
20
640 K/s
1018 K
1100 K/s
m
0
2
Figure 3.3.3
4
6
0
0.3
0.6
0.9
1.2
Weight loss, pyridine extract and insolubles yield
of bituminous coal at different pyrolysis times
H
105 -
-
At
the
later
stage of
this
thesis work,
another new
reactor
was developed with the experience from the previous screen reactor. This
reactor is a l'
surrounded
by
x l' x 2' plexiglass box with a hot stage in the center,
tar
collection and
liquid
(Figure
3.3.4.). New features include
graphic
recording
light
oil,
(3)
overall
mass
quenching mechanisms
(1) plexiglass
real-time
(2)
of experiment,
nitrogen
balance
by
walls
for
gas
and
tar
collection -of
collecting
photo-
products
is
possible, (4) more powerful heating circuit for higher heating rates, (5)
horizontal liquid nitrogen jet for sample quenching but preventing sample
loss from the screen.
3.4
OTHER TECHNIQUES AND EXPERIMENTS
Better
understanding
of
the
reactions
occuring
within
the
coal can shed light on the fundamental mechanisms of coal plasticity. An
important measurement is the molecular weight distribution of the extract
as
a
function
of
the
extent
Permeation Chromatography
U6K
is
4120T
detector
molecular
molecular
pyrolysis. This was
obtained
A Waters Associates ALC/GPC
columns in
Gel
by
201 system
Figure
series was employed.
a schematic of the GPC system. The basic system consists of a
injector,
Turner
(GPC).
100 A y-styragel
with 500 A and
3.4.1.
of
and
M-45
pump
electronic
output
and
a
recorder
was
signals. Subsequent
weight
weights
distributions
were
used
data
were
calculated.
440
model
The
to
UV
detector.
digitize
manipulation
made
on
columns
an
IBM
were
and
to
PC,
A Bascom
store
the
convert
and
to
average
calibrated
with
different size fractions of coal and wood tars. A linear calibration line
- 106 -
Figure 3.3.4
New box screen reactor
- 107 -
IS
OLVEIT
RESEr.RVOIR
r
I.
'
I
.
--
COLUMNS
Figure 3.4.1
p-
--
'
r
0
Schematic of Gel Permeation Chromatography system
OUTEr
- 108 -
was obtained (Oh,1985; Boroson, 1985):
4.086 - 0.0965V
log MW =
where
is the elution
V
volume in ml. Pyridine is used as the
solvent for the molecular weight determination.
exclusion
10,000 which
limit of
the literature (150
coal tars in
is
-
well
3000).
carrier
The 500 A column has an
above tne values
The actual
reported
separation
for
done
is
in the second column, which has the separation range of 2000 - 100 MW.
Reactions
in
coal
the
during
studied employing a heat-treated
the
plastic
phase
was
This is
coal tar as a model.
further
a viscous
liquid with initially 3 % weight fraction of pyridine insolubles. Phenomena
such
as
volatiles
escape,
cracking
polymerization and
of
the
metaplast (an equivalent name for the liquid) in coal can be more readily
studied from pyrolysis and plasticity measuremnet on this simpler system,
since
the
as compared
considered,
and
liquid
initial
extraction
generation
to that
experiments for
or
softening
step
of the bituminous coal.
wood
and
lignite
were
need
not
be
Some pyrolysis
performed
and
compared to bituminous coal data. To study the initial softening behavior
of coals,
some raw coal samples were pre-extracted
and afterwards their
plastic and pyrolysis behavior determined.
Mass
investigated
foils
(0.001
transfer
with a
inch
effects
set
of
thickness)
and
secondary
"foil" experiments.
with samples
briefly
reactions
were
Sealed
stainless
inside were
pyrolyzed
steel
in
the
screen reactor. The evaporation of material was suppressed, and volatiles
-
109 -
were trapped inside the foils and allowed to undergo secondary reactions
These experiments offer an approach for determining the relative contribution
of
pyrolysis.
evaporation,
polymerization
and
cracking
in
softening
coal
-
RESULTS OF EXPERIMENTS
CHAPTER 4.
this section, although some preliminary
presented
data have
also been
pyrolysis of
screen heater results,
Chapter 3. Plastometer,
presented in
are
reactors
different
from
results
Experimental
mainly in
110 -
other materials and GPC data are presented in order.
4.1
PLASTOMETER RESULTS
Viscosity curves for the Pittsburgh Seam No. 8 coal as a function
of
heating
different
for
time
constant
heating
experiments,
rate
and
different holding temperature experiments (Figure 4.1.1.) under 1 atm of
helium are
heating
shown
in Figures
4.1.2. and
at
different
determined as the
Ts,
the initial softening temperature,
rates,
runs
4.1.3. For
observed,
was rather
temperature
when the initial drop of the torque is
insensitive
to the heating rate. Apparent viscosities measured were in
the range of a few thousand to one million poises.
are
obtained
signal
from
raw data
by calibration
which
(Figure 4.1.4.),
with viscosity
The viscosity values
standards,
is
giving
torque vs
time
1.0 x 10-2 Nm
=
2.6 x 104 Pa s.
Figure 4.1.5. is a plot
of
the plastic period (time when vis-
cosity is less than 3.6 x 105 poise) against holding temperature for runs
with 450 K/s heating rate. Figure 4.1.6. is the rate of increase/decrease
of
torque
at
the
point
in
time
when
the
torque
softening occurs mostly during the heating period,
=
2.0
oz-in. Since
for the same heating
-
T"r c,- mx:
111
tt xrY-E
-
-
t I m c,
hi s tc>-ry_
t
HCding
t-,E
M
IX- r-nI
emera-1ture
I
IT
t
ctri n
Figure 4.1.1
tri
t
heat
ing-
r -ate
4
ruI
Temperature histories of plastometer runs
w
lw
m
7.2
CL
0
Gm
0
U
3.6
H
HA
0
Q.
M.
0
T-2
4
Seconds
Start heating
Figure 4.1.2
6
8
450 KsI
Viscosity curves for constant heating rate runs
10
9W
MW
7.2
CL
V
0
H
(AJ
0 3.6
0)
C.
C.
2
O
4
6
8
Seconds
Start heating at constant rates
Figure 4.1.3
Viscosity curves for holding temperature runs
10
- 114 -
Figure 4.1.4
Raw data for holding temperature runs
- 115 -
Runs in 1 atm helium, Pittsburgh No. 8 coal
40
U)
30
.9
20
0
10
0
773
973
1173
Holding temperature, K
Figure 4.1.5
Plastic period for runs at. 450 K/s to different holding
temperatures
w
w
w
mw
NW
0
Lz
10
C:
0
io
I0
H
H-
C
0.i
Figure 4.1.6
Rates of softening and resolidification for 450 K/s runs
- 117 -
softening rate is
the
rate,
rapidly
increases
however,
asymptotic
softening
at "I =
with
holding
700 'C.
above
value
alomst constant.
and
temperature
period
The plastic
2.0 Oz-in are presented
in
rate,
The resolidification
and
Figures 4.1.7.
reaches
an
the rate
of
and
4.1.8.
respectively, for a series of constant heating rate runs.
Some runs were made under 0.05 torr of ambient helium atmosphere,
and some under 35 atm of hydrogen or helium (Figure 4.1.9). Comparison
with
1 atm data
shorter for
period is
shows that the plastic
runs in
vacuum or 35 atm pressure and holding temperatures less than 6500C. Above
6500C, there is no clear effect of inert gas or hydrogen pressure on the
plastic
period, as
shown in Figure 4.1.9. While
the softening rate is
about the same as in the 1 atm case, the resolidification rates for both
the vacuum and 35 atm runs are higher for holding temperatures less than
6000C
(Figure
4.1.10.).
Above
6000C,
their
difference
is
small.
Runs
under 35 atm of helium or hydrogen give virtually the same plastic behavior, implying no major effect of hydrogen under these conditions.
effect
The
plasticity is
of
shown in
pre-extraction
Figure 4.1.11.
of
the
raw
Also shown
coal
in
sample
on
Figures 3.2.7.
its
and
3.2.8. are the different behavior of other materials in the plastometer,
note especially coal tar and pitch.
The apparent viscosities measured reflect the resistance to shear
of
a complex,
matter and
reacting
coke in
suspension
of
coal,
mineral
viscous liquid.
Descrip-
partially
a polymerizing and cracking
reacted
tion of the rheological properties of this system in fundamental terms is
- 118 -
Figure 4.1.7
Plastic period at different heating rates
I
Ii
I ATM HELIUM
CONSTANT RATE
OF HEATING
840 ,160'C ARE
HOLDING TEIPERATUFES
REACHED
Eo-FIGURES
0
Z
U)
-6w
a.
92-J
CL
10
0
100
HEATING RATE
K/S
1000
IS
0
S1412-
10
0
0
N
6 --
I ATM HELIUM
I0
o
N
2z
Figure 4.1.8
HEATING RATE
K/S K100
Softening rate at different beating rates
10
0
0
8
-
A
0
A
6
tatm
450
K/
I
350
K/s
35atm
He
350
K/
35atm
H
K/s
v--iaum
350
HE
_
2
0
o0HH
0
2
1
0
500
*
600
700
800
Hc 1 dig ri
temperature
Time interval when viscosity is less than 3.6x104 Pa s
Figure 4.1.9
Effects of pressure on plastic period
900
0
C
0
w
,w
w
w
0
10
3
I
C0
0
N
1.0
0
N
0.1
600 ThOLD
Figure 4.1.10
Resolidification rates at different ambient pressures
RW
vw
w
1W
ww
qw
1w
mw
3
0
E
2
C*4
t
0
CD
L-
0
0
I
0
2
Figure 4.1.11
4
Seconds
6
Effects of extraction on plasticity
H
H
- 122 -
difficult. However,
suspension
to
the
one
can
relate
solids volume
relative viscosity
the
and
fraction,
hence
the
to
of
this
liquids
formation and depletion kinetics in the coal. The existence of bubbles in
the melt has to be taken into account when such a correlation is made.
interpretation of these plasticity curves is
the
Therefore,
more diffi-
cult than those for less complex, non-reacting liquid/solid suspensions.
A direct
to
method
determine
the
content
liquid
of
pyrolyzing
coal
particle is obtained from the screen reactor.
4.2
SCREEN REACTOR DATA
Figures
shows the
evolution
the volatiles,
of
pyridine
as a function of the heating
and the pyridine insoluble char
extracts,
time,
3.3.3.
The correspond-
for four sets of experiments with rapid quenching.
ing temperature-time histories for each set are also shown.
An
extractables
of
yield
coal. As
the coal
was heated,
maximum
and
decreased,
then
about
the amount
the
rate
from the raw
25 % was obtained
of
of
increased
extract
this
being
process
to
a
strongly
dependent on temperature, as indicated in Figure 4.2.1. The time interval
over
which
decreased
rapidly
experiments
are
extract concentration was
the
The
indicated.
with
first
The
increasing
final
order depletion
maximum
amount
of
high
holding
(>
30 wt
temperatures
rate constants
fluid
material
quantity decreased with increasing holding temperature,
of coal)
of
the
for the extracts
volatiles) was as high as 80 % by weight of the raw coal.
ing time at a given holding temperature,
%
(extract
plus
However,
this
or with increas-
primarily due to conversion of
Figure 4.2.1
Pyridine extractables generation and depletion at
different temperatures
'U
a
6.3
'U
0
0
H
'U
'U
6.3
'U
3'
U
'U
16
12
8
4
(heating rate,holding temperature,cooling rate)
(
K/s
A
00
(470, 813, 1100)
(446, 858,1100)
(514,992,1100)
K
,
K/s
)
Semdas
- 124 -
the extractables to coke.
Figure 4.2.2.
rate
constants
of
is
an Arrhenius plot of the first
4.2.1.
Figure
A global
Kcal/mole is obtained. Figure 4.2.3.
table
material
energy
of
42.2
shows the total amount of transpor-
+ volatiles)
(extract
activation
order depletion
different
for
temperature-time
histories.
The effect of pre-extracting the raw coal sample on its pyrolysis
is shown
behavior
for two different
in Figure 4.2.4.
temperature-time
histories. For the pyridine pre-extracted sample (25% of extract removed
from raw coal sample),
the
coal
particles
the yield of extracts
with
tetralin
(coal
is
soaked
much
in
less. Pretreating
tetralin
with 4.8%
weight gain) does not change the mass rate of extract generation significantly but does prolong the extract survival as seen in Figure 4.2.5.
This may
reflect
partial
hydrogenation of
metaplast
by
the
tetralin
resulting in a somewhat more thermally stable extract.
4.3
PYROLYSIS OF OTHER MATERIALS
For comparing
the plastic
tion/destruction kinetics,
which is
behavior and related extract genera-
pyrolysis on the stainless screen of lignite,
a non-softening material,
and wood,
which is
a slightly caking
sweet-gum wood were also studied. Particles size used were 63-75 ym,
90-105 pm repectively.
Figure
4.3.1.
shows the
evolution
of
and
the vola-
tiles, pyridine extracts and the pyridine insoluble char as a function of
heating time.
All
samples were heated
at 736 K/s to
555'C and allowed
- 125 -
20
5
ccJ
*-0
Ci
4-
140
cc
&M
=
a
4.'
a)
a.
a)
-a
14
a)
Cu
4.'
U
Cu
0.
4.'
uJ
0.05
1.1
1.0
Reciprocal
Figure 4.2.2
12
absolute temperature , 10-3 K'
Arrhenius plot for extract depletion rate constant
1W
I
100
0
3
H~
Ctl
0
0
1
Figure 4.2.3
2
3
4 SECONDS
Yield of pyridine extract + volatiles at different
temperature histories
IN 0.1 MPa HELIUM
100
W
0
*
I
I
HEATING RATE 33001Ks~'
HOLDING TEMP
908 K
COOLING RATE 4500Ks'
0
IJ'J
I
I
ale 0..,,,,%
60h-
470 Ks'
813 K
1 100 KS-'
RC
/
RC
-3
40
H
N,
-1
0,e~
0/
0
,edo
U
20
Of.. -t
-U
0
2
Start heating
4
Seconds
PX
I
6
I
8
RC - RAW COAL
PX - PYRIDINE EXTRACTED COAL
Figure 4.2.4
Comparison of pyridine extracts for chars from raw
bituminous coal and pyridine pre-extracted coal
- 128 -
I-
0
1
3
2
Se cDsri
Holding Temperature 626 0 C, heating rate 2000 K/s
cooling rate 4500 K/s
Figure 4.2.5
Comparison of pyridine extracts for raw bituminous coal,
tetralin soaked coal, and pyridine pre-extracted coal
I
L-J
z
-j
0
C-)
a:
LL
H
0
0
II
M
w
.
0
SECONDS
Figure
4
.3.la-c
4
8
12
20-
2
o~
4
6
SECONDS
SECONDS
no
pyridine
Volatiles, pyridine extractables and insolubles yield
for (a) wood ,(b) bituminous coal, and (c) lignite as
functions of pyrolysis time
ex
treet
observed
- 130 -
comparison
lignite,
the
4.3.2.
subjected
all
samples
of
extracts
of
amount
the
of
the screen reactor.
in
to cool by natural convection
bituminous
to the same temperature
are
cooled
amount of liquid was generated
by
in
natural
wood
coal,
coal than in
and
smaller
a
Note
a
Figure
In
time history.
convection.
the bituminous
is
Figure 4.3.2.
the rapid
quenching experiments. These results offer an interesting comparison with
the basic data for bituminous coal.
similar
the
exhibiting
material
A test
to those of the intermediate
pyrolysis of
coal,
softening
is
and
chemical
physical
liquid metaplast
an
industrial coal
properties
generated
during
tar from bitu-
minous coal. The coal tar studied is a heat treated tar with initially 3
%
of
insolubles. Subsequent
pyridine
reactions
softening of the bituminous coal can be
of
the
metaplast
after
studied by observing the pyro-
lysis behavior of the coal tar in the screen reactor.
The advantage of
this system is that the complex softening step (initial liquid generation
step) need not be considered. Therefore the starting material already has
well
defined
properties
(
molecular
weight
distribution
and
solids
content) and their pyrolysis behavior would not be obscured by softening
reactions. Some detailed experiments were performed. Figures 4.3.3. and
4.3.4. show the
weight
loss,
pyridine extracts
char as a function of pyrolysis-time,
and
pyridine insoluble
for different holding temperatures
and heating rates. This phase of work is conducted in collaboration with
Dr. Jean-Louis Saint-Romain.
- 131 -
Bituminous Coal
cc
30-
*4 20eS
."
10-
Wood
Lignite
0.5
0.25
1.0
4
2
8
Seconds
Samples heated at 736 K/s to 555*C
and allowed to cool by convection
Figure 4.3.2
Pyridine extractables for different materials
100
Heating rate 455 K/s
Holding Temp 1022 K
Cooling rate 4500 K/s
80
Volatiles
60OH
I-I
Pyridine Extractables
3:o
40
1
0
20Residual Wt. on screen
after extraction
2
0
4
6
SECONDS
Figure 4.3.3
Pyrolysis of a heat-treated tar
100
80
cd
60
0\1
LA)
40
I
20
0
5
0
15
10
SE
A - weight loss of sample,B -
pyridine extractable
left on screen
co ndsi
;
portion of material
C - pyridine insoluble portion of material
left on screen
Figure 4.3.4
Pyrolysis of a heat treated coal
tar (Saint-Romain, 1985)
- 134 -
4.4
GPC DATA
The
extracts were
different
curves
molecular
determined
for
Different
raw
extractables
Figures 4.4.1.
show different
their
(Figure
coal,
elution volume) at
materials
reflecting
distribution
bituminous
extents of pyrolysis.
(concentration vs
butions
weight
structural
4.4.1.). The
of
coal
tar
to 4.4.3.
different
molecular
differences
molecular
the
weight
pyridine
and
wood
show a few GPC
pyrolysis
weights. and
in
is
at
their
times.
distri-
pyridine
related
to
the
elution volume V by the equation
log(MW)
obtained
=
from a linear calibration
-0.0965 V + 4.184
method from different
size fractions
of coal tars (Oh,1985, Boroson,1985). The data were also transformed and
normalized with respect to the total weight of extract in Figures 4.4.4.
and
4.4.5. Hence the
different
lost
distance between
pyrolysis times indicates
during
the distribution
curves at
the amount of material
the time interval. Changes in
generated
two
or
the moments of MWD, such as
number and weight averaged molecular weights are shown in Figures 4.4.6.
and
4.4.7.(Saint
Romain,1985) For
both bituminous
coal
and
coal
tar
pyrolysis,
the extract average molecular weights increased to a maximum
and
decreased,
then
as
a result
of
polymerization,
evaporation
and
cracking reactions.
4.5
OTHER EXPERIMENTS
A few "foil" experiments were designed
to study the
effects of
C
CD
CD)
0
CD
0
CD
(A
Cn
0)
12)
10
20
30
Elution Volume, ml
Figure 4.4.1
Raw GPC curves for bituminous coal, wood and lignite
pyridine extract
'o
i'i
I
i
0
030
Elution volume, ml
BC-
---
r
I--
I
I
Elution volume, ml
toA
-14't
cooki
0C
i
oo
AI.
.
...
oC3
3o
502.0
Figure 4.4.2a-c
.
Raw GPC curves for pyridine extract of bituminous coal
pyrolyzed to (a) 5550C, (b) 658 0 C and (c) 749 0 C
- 136 -
GPO
C
r
e
co
.
lu
vs
i
n
voluAme-(ml)
U,
A)
S..
Cu
A)
-Q
HwA
Cu
Elution volume, ml
U,
Elution volume, ml
0
0
C',
0
'-4
4
A)
C.)
Li
0
A)
Elution volume, ml
Figure 4.4.3
Elution volume,
ml
GPC curves for pyridine extracts of different materials
at different pyrolysis times
- 138 B3I-ium
828
Fa I
Co
inousc> A ,
K
13
12
11
10
0
U
9
8
e.
7
0'
6
5
4
3
2
1
0
0
0.2
0.4
0.6
Bitu7-11minous
1
0.8
1.4
1.2
1.6
1.8
2
(Thousands)
molecular weight
A -=
Co>al
931
I
K
13
12
11
10
C
0
9
U
0
8
L
7
0'
6
S
0
.5
I..
5
4
3
2
1
0
0
Figure 4.4.4a,b
0.2
0.4
0.6
1.2
1
0.8
(Thousands)
molecular weight
1.4
1.6
1.8
2
Normalized molecular weight distributions of pyridine
extracts at different pyrolysis times for bituminous
coal
- 139
Xtea
HeC3at
-
t-ed
Cocm1
-
0.0
K
828
Tar
2019-
s
0.25 s
0.60 s
1.50 s
3.00 s
14.80 s
1817 --
C
0
c
0L
0'
-
16 15 ----14 13 1211
10 98--
736 K/s heating to 828 1
1-
0-
HeAt
C
0
0
.c
.4-
.4.
0'
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
Treated A
ca
I
1.4
1.2
1.0
Tar-
0.25 8
0.50 s
1.20 S
2.00
2.25
r
9'
I'
1022K
460 K/s to 1022 K
t
* ~
'I
~
4..4.
-. 4-
-~
I
0.0
Figure
0.8
0.6
(Thousands)
molecular weight
0.4
0.2
0.0
4
.4.5a,b
0.2
0.4
0.6
0.8
(Thousands)
molecular weight
1.0
S
1.2
I
I
1.4
Normalized molecular weight distributions of pyridine
extracts at different pyrolysis times for coal tar
- 140 -
cm
03
Lu
Coal tar
4oF
6580C
D
A
3oo
-
200'
-u
-U749
C
-- -
0.25
0.5
1.0
2
4
8
Seconds
Holding temperatures indicated on curves
A
Figure 4.4.6
heating rate 465 K/s
-5000 K/s
quenching
E
heating rate
*
heating rate 736 K/s, natural cooling
455 K/s
Changes in average molecular weights for bituminous
coal and heat treated coal tar
- 141 -
1.)
3
H
400
C)
H
0
E
4:
350
4.)
0
5
10
15
SEco n d s
300
00
.
250
0
5
10
15
S
Figure 4.4.7
nc
rid
Changes of extract weight and number averaged
molecular weights during pyrolysis of a heat
treated coal tar. Temperature histories are the
same as in Figure 4.3.4
(Saint-Romain, 1985)
- 142
-
external mass transfer and secondary reactions on plasticity and pyrolysis. By trapping the volatiles in the foil, secondary reactions of those
volatiles
tar,
were
promoted.
For
the
pyrolysis
of
the
heat
treated
coal
a pyridine insolubles yield exceeding 40 % was obtained vs the 29 %
obtained
in
screen experiments,
for the pyrolysis of coal tar at around
1000 K. This increase was observed at high temperatures above 950 K.
- 143 -
DISCUSSION OF EXPERIMENTAL RESULTS
CHAPTER 5.
plasticity
mechanism of
The
coals
in
is not
well
understood
for rapid heating and high temperature conditions. This present study of
rapid heating, high temperature conditions
plasticity at
systematic
swelling
liquid
Plasticity
approach attempted.
and
agglomeration
plasticizer
direct
are
formation
by
the first
is
and related phenomena
consequences
heating. The
intra-particle
of
present
such as
study shows
that
physical melting of part of the coal first occurs when the temperature is
sufficiently high to release mobile materials by overcoming weak bonds.
Subsequent pyrolytic bond
breaking generates additional
At the same time,
severe conditions.
liquid is
liquid
at more
lost through escape from
the particle, and through repolymerization and cracking reactions to form
light gases and solids inside the liquid phase.
As shown in Figure 5.1.,or by a comparison of plasticity data and
intra-particle liquid formation and depletion data in Figures 4.1.3. and
4.2.1.,
the respective
times
for
initial
generation
and
near
final
depletion of extractables within the coal corresponds well with the time
for initiation and termination of the coal's plastic phase. This supports
our assumption that the pyridine extract of the rapidly quenched coal is
a reasonably accurate measure of the amount of liquid product within the
pyrolyzing coal at the time of quenching.
25 wt % of the
However,
by
yet
pyridine,
temperature.
It
is
the
raw unpyrolyzed
unpyrolyzed
coal
coal
contains
no
can be
liquids
extracted
at
room
quite likely that the initial 25 % extract corresponds
Figure 5.la
Comparison of viscosity and extract data
IN 0.1 MPa
HELIUM
7.2
100
*U%
D
0
U
80
0
0N
0
60
I
3.6 o
40
'I
C
C)
0
ha
0
x
ha
0
0.
0.
20
0
0
2
4
6
Seconds
t
Start heating at ~450Ks~nI
8
10
100
10
VISCOSITY'
0
-- <I
80
0
8
RESIDUAL COAL
4
460
V
-
<60
PYRIDINE INSOLUBLE
-6
Z
N
IM
I
I
1
0Lf
oi 40 -
4 d
a-EXTRACT
tr
I0
20
VOLATILES
0
-2
'0
0
Figure 5.1b
2
4
6
SECONDS
Comparison of viscosity and pyridine insolubles, extracts
growth for the higher temperature case in Figure 5.1a
- 146 -
this material
sufficiently high temperature,
causes
and
plastometer
heating
hypothesis. If
K,
(50-700
Figure
initial
this
consistent
is
material
25 % of
pre-extracted,
and
less and
shorter.
the time duration of a high liquid content inside the coal is
(Figure 4.2.4.). The pre-extracted
of
the above
with
the extract- is
subjected to pyrolysis,
is
the coal
is
the
with
570 K independent
occurs around
4.1.2.)
melting
physical
undergoes
the coal. Our observation
of
initial softening
that
rate
afterwards
softening
initial
point material. Upon heating to a
low melting
to some loosely bonded,
expected to soften at a later
coal is
time (higher temperature), and the duration of plasticity is shorter, as
is observed in Figure 4.1.11. Other workers found the Gieseler plasticity
be
reversible
as
the
stage to
during the early softening
(low heating rate type experiments)
temperature
raised
is
or
supports the picture that some part of the liquid is
lowered.
Again,
this
formed by physical
melting.
At
higher
temperatures,
pyrolytic
bond breaking
liquids which
significant quantities of intra-particle
cannot
generates
be rever-
sibly converted back to resolidified coal. The maximum yields of extractables from the coal studied can be as high as 80
% daf.
In
similar
screen heating experiments on a lignite, which does not soften, almost no
extract
was obtained
only a small extract
wood
from the residual
char
in
a
the screens.
yield up to 7 % was observed in
(Figure 4.3.2.). This wood exhibits
pyrolysis
on
fixed
bed
reactor
Similarly,
the pyrolysis of
slight caking behavior
(Boroson,
1985).
Thus
during
plasticity
behavior is well reflected by the amount of intra-particle liquid content
during pyrolysis.
- 147 -
faster rate than the
the structural
reflecting
bituminous coal (Figure 4.3.1),
composition, and they
remain more or
of
chemical
less a porous solid during pyrois
It
offering less resistance to products escape).
lysis,
difference
are more aliphatic than the bituminous coal in
(they
the fuels
a much
at
pyrolyzed
lignite
The wood or
possible that
biomass, lignite and low rank coals may still undergo an initial conversion
to
very
high
yields
potentially
of
from high rank coals,
sample,
reactive
may be much more
mediates. However, these liquids
liquid
transportable
inter-
than those
and thus are removed by polymerization within the
times so short
and transport away from the sample in
that they
escape detection by the char extraction technique.
Although
transfer
mass
external
internal
transfer
mass
bituminous
for
resistance
predominates
resistance
coal
pyrolysis
the
under
the experimental conditions studied in the screen heater (Oh, 1985), in a
liquids
have
may
a
survival
longer
closed
system,
undergo
further reactions such as polymerization or
coal tar
experiments for a heat treated
time before they
cracking. The foil
indeed showed more liquids are
retained than in open screen pyrolysis (more solids are formed, too) due
to trapping of volatiles.
In
the plastometer at the higher
viscosity
than
Figures 5.la,b,
temperature has
the duration for
reflects the transfer resistance
meter, which
K/s)
is a semi-open
or high temperatures,
plasticity
a high
then it
a longer duration
pyridine extract.
to volatiles
chamber. At
the plasticity curve from
escape
very rapid
of low
This probably
from the
plasto-
heating rates (104
may be more appropriate to infer the
from data on pyridine extract generation/depletion kinetics
- 148 -
from the screen heater reactor than the fast response plastometer.
Another clear manifestation ot the resolidification reactions of
softened bituminous coal at longer holding times and higher temperatures
is
the yield of pyridine insolubles.
the corresponding increase in
dies
Stu-
of the pyrolysis behavior of the heat-treated coal tar using the
screen reactor showed similar evidence of the competition between coke
formation via
polymerization,
cracking,
and volatiles escape via trans-
port away from the reacting substrate (Figure 4.3.4.).
However,
differences
in
plasticity
are expected between
tar,
pitch and bituminous coal (Figure 3.2.8.) since the raw tar is already a
viscous
97
liquid containing
% pyridine
solubles, and the pitch is a
solid with a low softening point at 342 K. The bituminous coal requires
that a minimum inventory of liquids capable of initiating and substaining
plasticity
throughout
a
reacting,
three-phase
suspension
be
available
within the coal matrix. Generation of this liquid requires that at least
some
melting
further
cizing
content
or
bonding breaking precede
complication
is
that
heterogeneous
agents may be more significant
of
char
this char (Serio,
(pyridine
1984).
insolubles)
for
the onset
reactions
of
of
plasticity. A
these plasti-
the coal due to
or higher
its
higher
cracking reactivity of
This may explain the higher ultimate yields of
pyridine insolubles in coal than in coal tar, 56 % vs 28 % respectively,
and is
consistent
with the fact that the resolidification rates for the
coal tar or pitch are slower than for bituminous coal (Figure 3.2.8.).
It
is
thus clear
that
better
understanding
of
the
reactions
- 149 -
occurring within the coal can shed light on the underlying mechanisms of
coal
of the extract
(MWD)
distribution
measurement
important
plasticity. An
is
the
molecular
weight
as a function of the extent of pyro-
lysis.
average
for extract
trends
The
molecular weight
evolution
are
strong functions of temperature-time history. In general, with increasing
the molecular weight increases to a
time at a given holding temperature,
very
well
4.4.7.).
maximum,
defined
4.4.4.
Figure
and
shows
then
that
decreases (
besides
Figures
formation
4.4.6.
of pyridine
and
insol-
ubles, some heavier molecular weight material are formed (1.20 s and 3.00
s in
Figure 4.4.4a,
and 1.10 s in
Figure 4.4.4b) at intermediate times.
These high molecular weight materials are probably
intermediate products
in the solid
formation reaction by polymerization. The decrease arises
from cracking
of the liquids since further
weight
generation
of low molecular
liquids by primary decomposition of the parent coal is
no longer
possible at these severities. At higher holding temperatures, more severe
cracking drives the molecular weight to even lower values. The maxima of
these
curves
are
thus
determined
by
competition
between
evaporation,
polymerization, and cracking reactions. At small heating times (0.8 s in
Figure
4.4.4b.),
the extract
molecular
weight
distribution
from
the
quenched char is identical to that of the extract of unsoftened coal at
0.5s, while the coal has already become plastic from plastometer studies.
This supports our hypothesis that initial softening is due to physical
melting of part of the coal corresponding to the raw coal extract.
The
rate
of
cooling
in
the
experiments may
also affect
the
- 150 -
weights
molecular
and 4.3.2.).
4.2.1.
volatiles
lighter
and
(Figure 4.4.6.)
(compare Figures
extract amount
A slow-cooling rate experiment allows more time for
tend
weights
molecular
the
hence
escape,
to
to
be
higher, and liquid yield less than those in rapid quenching experiments.
Also, the heavier molecules in the liquid may have more time to realign
themselves to form pyridine insolubles.
higher
The
molecular
than
that
for
the
the
of
weight
tar
coal
coal
bituminous
or
extract,
extract
and
wood
for
much
is
lignite
extracts (Figure 4.4.6.). This probably reflects two differences expected
in their thermal behavior becauses of differences in their structure and
initial state. Cracking and repolymerization of the high molecular weight
extract
coal
especially
the
in
presence
of
unsoftened
of
surfaces
macerals and already available coke can give more residual coke and high
molecular weight coke
lower
more
vapor
pressure,
with
the
rate,
and thus a slower evaporation
pyridine
lower
also has a
molecule
liquid
pyridine insolubles in the
chance to form
consistent
precursors. A larger
and thus has
liquid state. Both
and
yield
insolubles
the
are
faster
evaporation rate observed in the coal tar pyrolysis.
Plasticity
following
points
mechanism
regarding
inferred
the
behavior
from
of
this
a
study
coal
suggest
particle
during
heating:
(1) Initial softening at low temperature is due to physical melting.
(2) Further
generation of
liquids
is
by
pyrolytic
bond
associated with the release of gases and tar from coal.
the
breaking,
- 151 -
(3)
associated with the growth of pyridine insoluble
Liquid depletion is
(from
or solids
material
the cracking
reactions
yield and
extract
curves),
plasticity
decreasing molecular weights),
(from
and
and tar
transport away from coal.
(4)
A substantial
transformed
is
coal
softening
of
fraction
weight
before these liquids
into intermediate liquids in the plastic coal
are depleted by the above pathways.
molecular
structure and
(5) The
these
of
weight
liquids
influence
their devolatilization behavior and reactivity in polymerization and
reactions. Competition
cracking
determine
the observed
among
quality,
yields,
processes
these
in
turn
can
generation rate
and overall
of devolatilization liquids and gases.
(6) For conditions where external mass transport is, or is made to become
limiting in
rate
can
intra-particle liquids
be modified by
(foil
runs),
experiments).
and
changing
the
Pretreatment
manipulation of
mass
of
the
depletion, then
liquids escape
transfer for
raw
intra-particle
coal
plasticity
(pre-extraction
mass transfer
rates
can
also change plasticity.
(7) Solvent extraction of rapidly quenched char is a promising technique
to
determine plasticity
at
the higher
ranges
of
heating
rate
and
temperature (such as combustion conditions) where direct measurements
even with the present plastometer would be difficult.
- 152 -
The following schematic for plasticity mechanism is proposed:
transport
liquid
tar
+
-
bond breaking
coal
cracking
polymerization
-6
heavier
melting (low T)
gas
-o
liquids
+
coke
+
coke
This picture is also consistent with the above points as well as
the
observed
pressure
effects
1977). High tar
(Suuberg,
vacuum
on
yields
or high
vacuum
in
pressure
pyrolysis
is
pyrolysis
due to
rapid
transport of tar away from coal, and this also reduces the plastic period
(Figure
4.1.9.).
Under high pressures,
yield increases and according
higher
shorter
plastic
pressure
for
to the above picture,
because of aubmented
also be
period
maximum
is
tar diffusion is
expected
plasticity
suppressed,
gas
solids yields will
time for repolymerization.
and
found.
Therefore,
would
exist
for
an
Hence a
optimum
rapid-heating,
high
temperature conditions. The effect of pressure on the polymerization step
is not clear. Effects of hydrogen or hydrogen donors in
certainly
rate and
is
of practical
reaction
interest.
times studied,
However,
this reaction
under the range of heating
no difference
in
the
plastic behavior
was observed for experiments in 35 atm of hydrogen or 35 atm helium. Pretreating the coal particles with tetralin does not change the mass rate
- 153 -
of extraction but does prolong the survival of extract. This may reflect
partial
hydrogenation
of
metaplast
by
the
tetralin
resulting
in
a
somewhat more thermally stable extract.
In
the
case of
pyrolysis of coal
tar,
a similar scheme
proposed.
transport
tar vapor
tar
polymerization
cracking
heavier
tar
coke
gas + coke
can be
- 154 -
MATHEMATICAL MODELING
CHAPTER 6.
6.1
COAL PLASTICITY KINETICS
The
apparent
are
data
viscosity
fitted
first
with
simple
the
first order kinetic scheme of Fitzgerald (1956):
ki
k2
-hm
coal
metaplast
coke
-
The term metaplast refers to the intraparticle liquid content M. Apparent
viscosity y is related to the viscosity of the solids free liquid y* by a
concentrated suspension model (Frankel and Acrivos, 1967)
=
(9/8)/((l - M)-1/ 3 - 1)
(6.1.1)
At a late stage of resolidification, the rate of increase of
log(viscosity) is approximately
(6.1.2)
dlnuy/dt = k 2
Thus k 2 can be found from experimental
y*
are
estimated from
pitch (average
obtained
by
value
Nazem's work
taken,
fitting data
1200
for the
resolidification rates. Values of
(1980)
on carbonaceous
poise). The
plastic
calculated viscosity curves, and calculated
mesophase
parameter ki
period. Figure
is
6.1.1.
then
shows
plastic periods vs experi-
mental values for 450 K/s holding temperature runs.
MW
KI
COAL -P
METAPLAST
K2
-+
COKE
300
gr240
40
in
oU 180
0
-o
C 3
0
0
U
0i
N
en
*0
0
120
2
6I
0.
60
U
Hn
U,
4-
I3
0
0
f
773
1173
973
Holding temperature, K
Figure 6.1.1
2
4
6
Seconds
1
Start heating at 450Ks
K1 -
1O
8
245 EXP(- 4900/ T),S
-I
K2 = 2 x 108 EXP( - 16350 / T ) , S-
Calculated viscosity curves for 450 K/s heating rate
to different holding temperatures, and comparison with
observed plastic period (solid circles)
- 156 -
scheme
kinetic
Fitzgerald
the
However,
fit
not
does
runs at
different heating rates, because physical melting of part of the coal is
not taken into account.
The plasticity
mechanism presented
Chapter 5 is a starting
in
point for a more realistic model. Both primary bond breaking and physical
melting are important
in
coal
softening.
most of
the
of the coal passes through the liquid phase,(i.e.,
the
first
be neglected as a
reaction can
organic material
macerals such as
non-softening
fusinite
considered. Polymerization in the
liquid viscosity,
658 0 C,
since
approximation
and
are
inertinite
a
small
the total mass) although in lower rank coals it should be
fraction of
(>
the primary
in
Coke formation
no solids are formed.
even if
compare
555 0 C
and
658 0 C, Figures
in
result
liquid phase can
a higher
At higher temperatures,
cracking
4.4.4.a,b)
pre-
dominates over polymerization. For this reason, and because the detailed
MWD
information
polymerization
tar
required
kinetics
in
for
a
the
quantitative
liquid
phase
model
is
not
is
not
available,
modelled. Both
escape and cracking are pathways to resolidification.
High pressure
or closed reactor geometry can prevent tar evaporation. Since no individual
tar
or
gas
species were
collected
in the
present
study,
both
tar escape and cracking are lumped into one reaction with volatiles (tar
and gas that leaves the particle) and coke as products. Further reaction
between intraparticle liquid and coke or mineral matter are not considered.
Here the global model assumes the coal to undergo
both physical
- 157 -
melting and pyrolytic bond breaking to form an intermediate liquid in a
suspension of unsoftened coal, coke and inert mineral matter. The liquid
forms a volatile product which, according to the model, escapes from the
coal
infinitely rapidly,
coke
to
is
formed
thus
to be
assumed
25 % pyridine extract from raw coal is
initial
that
volatiles
a solid coke. The weight ratio of
and leaves
can
obtained
be
structure, and
is
melt
by
generated
over
a
narrow
range
of
melting. This
physical
of
a).
The
assumed to be the maximum
decomposition
chemical
without
(=
constant
the
initial
material
To account
temperatures.
assumed
to
melting,
a gaussian distribution of melting points with a mean
for
at 623 K
the calculations.
and a standard deviation of 30 K are assumed in
is
These
values reflect the range of the initial softening temperatures of coals
(Read et al.,1985;
Rees and Pierron,
liquids by melting alone would
therefore
rate and this distribution function.
= 0.13 for the coal studied,
1954). The
rate of
be the product
of the heating
The mineral matter (wt.
Franklin,1980)
is
scheme
and
corresponding
rate expressions
fractions
of
raw
coal
free,
(moisture
fraction fa
assumed to be inert. The
reaction
the
generation of
mineral
based
matter
on weight
containing
basis) are
ki
coal
-
k2
l intermediate liquid
volatiles + coke
melting
dC/dt = -ki ( C - fa ) - rm
dM/dt =
ki ( C - fa ) + rm - k2 M
dV/dt = k 2M ( 1/1 + a )
dE/dt =
k2M ( a/l + a )
(6.1.3)
- 158 -
where
C,M,V
and
E
are the
weight
fractions
of
the unsoftened
coal,
liquid, volatiles and coke, respectively, and rm is the rate of physical
given
melting,
by
the
product
of
rate
heating
and
the
point
melting
distribution function.
The model
ion and
is
data
depletion
applied
from the
that
intraparticle
reactor. At
has reacted
physically melt
can
to
screen
when most of the coal
pyrolysis,
material
first
has
done
(C so
a
late
fa = 0)
(rm =
format-
liquid
0),
stage
of
and all the
the rate of
liquid depletion is
dM/dt = -k2 M
The
rate
(6.1.4)
constant
k2
can
thus
be
found
from a
plot
of
ln(M)
vs time for constant temperatures. Typical values thus obtained are shown
in
Figure
4.2.1. An
Arrhenius plot
gives an activation
energy of 42.2
Kcal/mole, and k 2 is found to be
k2 = 1.9 x 1010 exp ( -21200/T), s-1
The rate constant kI was obtained from a best fit
(6.1.5)
to the experi-
mental data, using values of k 2 already found. The result was
ki = 6.6 x 107 exp ( -14500/T), s-1
Figure 6.1.2. gives values
ables,
(corresponding
to C +
E
of
and
(6.1.6)
pyridine insolubles and extractM respectively)
observed
in
the
W
14W
M
W
a= 1.25
100
a=
100
1.25
80
80
Cu
cc
(3
cc
60
60
0
40
40
20
20
Z
5
0
10
15 Seconds
0
0.5
1.d
1.5
T
813, K
0
Q
Figure 6.1.2
Calculated vs experimental volatiles, pyridine extract,
and insolubles yield for two temperature histories
2.0
Seconds
- 160 -
corresponding
experiments and
"a" value of 1.25
model
(solid lines),
predictions
(obtained from long holding time data)
using an
for two sets of
reaction conditions.
To
the
calculate
used with the present
apparent
scheme
Equation
viscosities,
in
and rate constants
6.1.1
Equations
is
again
6.1.5 and
6.1.6. Figure 6.1.3. shows calculations corresponding to the experimental
conditions and
lab results presented
in Figures 4.1.2. and
4.1.3. The
rate constants obtained from the screen reactor describe constant heating
rate plastometer runs well but underpredict viscosity for 450 K/s runs at
low holding
temperatures.
Three possible reasons are (a)
zation in the liquid phase is neglected,
free viscosity y*,
(b)
resistance
which
can
the polymeri-
affect
the solids
to volatiles escape can exist in
the
plastometer, and (c) y* is assumed independent of temperature.
If the solids free viscosity P* is assumed to be proportional to
a positive
power of the intraparticle
liquid
weight
average
molecular
weight, as in polymer melts, then from Figure 4.4.4. the lower temperature runs would have a higher viscosity than predicted since the liquid has
higher
average
molecular
weights.
Temperature
variation
of
-p* is
cor-
related by the Andrade equation with an activation energy of 38 Kcal/mole
(Bhatia et al.,1985).
However,
the expression is
not valid over a wide
range of temperatures, and the resulting fit is not improved. If the rate
constant ki is retained (because little volatiles are formed during early
softening) and k 2 allowed to vary, then k 2 = 8 x 106 exp(-14000/T)s- 1 fits
the data well (Figure 6.1.4.). A lower activation energy
with mass
is consistent
transfer resistance to volatiles escape from the plastometer.
- 161 -
7.2
~I)
3.6
0
'.,
2
0
4
6
8
I0
SEco>rId s
0
7.2
4.'
3.6
0S
2
t Start
Figure 6.1.3
4
8
6
10
heating aL 450 K/s
Calculated viscosity curves for conditions of
Figures 11 and 12, with liquid formation and
depletion rate constants(
Eq.
6.1.5,
6.1.6)
- 162 -
7.2
U)
(0
3.6
0
0
'
2
4
6
SC,-
0
-H
8
I0
ds
7.2
4C
3.6
0
2
4
8
6
Se
Figure 6.1.4
10
cds
Calculated viscosity curves for conditions of Figures
11 and 12, with k2 fitted
- 163 -
Therefore both
at
polymerization
lower
and
temperatures
to
resistance
volatiles escape cannot be ruled out.
6.2
MODELING OF COAL PARTICLE AGGLOMERATION
Heating
coals to
of
practical
soften. As a result,
agglomerates
and
regimes
under
selected
agglomeration
utilization.
For
can
contacting
in
conditions.
desired
fixed-
or
stable agglomerate
obtaining a mechanically strong product
ture. Conversely,
agglomerate
cause
many
bituminous
pyrolyzing particles can swell
result
example,
interest
buildup
Softening,
unwanted
effects
formation
is
and form
swelling
in
coal
critical
to
in metallurgical coke manufac-
in
fluidized-,
entrained-,
and
(more properly moving-) bed gasifiers or combustors can degrade
process operability by bed bogging, defluidization, plugging of chambers
and
transfer
surfaces.
lines,
Lightman
and
erosion
and Street
(1968)
combustion of pulverized coal in
Few
been
studies on
reported
agglomeration
K/min)
in
of
the
of
refractories
and
observed
agglomerates
coal
literature. Klose
and final temperatures
transfer
in
the
large flames.
the agglomeration kinetics of
stationary
heat
particles
(around
and
at
Lent
low
softening coal have
(1984)
heating
investigated
rates
(2.5
-
6
520 C) and modelled their results
in terms of a viscous sintering mechanism Their study is pertinent to low
temperature coal carbonization processes. Tyler (1979, 1980) and McCarthy
(1980) measured extents of coal agglomeration due to flash pyrolysis in a
laboratory scale fluidized bed and an entrained flow reactor, respective-
- 164 -
et al.
Scharff
ly.
(1981)
a detailed coal agglomeration model
developed
binding energies for the
based on surface
particle aggregates. However
none of these studies has treated plasticity
and agglomeration kinetics
such a way that their influence on pulverized coal combustion
in
can be
predicted.
The essential
features of the agglomeration model are presented by
treating the case of formation of an agglomerate
first
by collision
of
two spherical particles of softening coal of constant radii, a. Extensions to generation of larger multi-particle agglomerates and to account for
particle swelling are then presented.
(A) Two single particles agglomerating in isotropic turbulence
For a system initially composed solely of single spherical particles
agglomeration is
of radius a,
modeled by assuming there are only binary
collisions of single particles to form two-particle agglomerates (Figure
6.2.1) at a rate per unit volume of,
dn /dt
where
=
Kc
-
is
agglomerate
the
per
collision
collision,
unit volume at any time t.
an isotropic
(6.2.1)
K&n2
constantE the
and
n the
number
probability
of
single
of
forming
particles
an
per
The particles are assumed to be entrained in
turbulent flow field.
For the particle
size and turbulent
intensity ranges in this study, the particles are much smaller than the
Kolmogorov microscale of turbulence. Permanent sticking of two particles
- 165 -
Figure 6.2.1
A pair of agglomerated particles
- 166 -
to form an agglomerate is proposed to occur in the following sequence:
(1) After Saffman and Turner (1956), particles collide with a dimensional
collision constant, Kc
K
C
=
10. 4 a 3 /F7/v
(6.2.2)
0
whereEe is the turbulent energy dissipation rate per unit volume,and v is
the kinematic viscosity of the entrainment gas.
(2) In reality the particles are irregularly shaped. Thus, two particles
may
approach
at
many
different
number of possible contact
by
assuming each
spheres
of
areas.
two-particle
radii
sionless
neck
radius,
quantity
is
normally
For simplicity this is
a neck
areas is
defined
resulting
in
a
large
dealt
with
agglomerate consists of two distorted
a, joined by
possible collision contact
orientations
as
distributed
of
diameter
2x. A
range
of
treated by introducing a dimenx/a,
with
and
a
further
assuming
mean 7 1 and
a
that
standard
deviation 6.
(3)
The strength of the neck region am at the instant when a potential
agglomerate is formed, is assumed to be
a
m
a M(t)
o
=
(6.2.3)
where a0
is an empirical constant and M is the time-dependent weight
fraction
of
liquid
plasticizer
in
the
coal
(i.e.
only
the
liquid
- 167 -
assumed to adhere effectively).
portion of the contacting surface is
(4) Among the different stresses imposed by the flow field on
pair of particles
(Figure
6.2.1)
of
radii a,
a fused
the maximum stress is
the bending stress ab (Yoshida et al.,1979),
ab
=
10. 8Y/%e/v /c1 3
(6.2.4)
where v is the viscosity of the carrier gas.
(5) If the bending stress is larger than the strength of the bonding neck
region,
no agglomeration
E =1. Thus from Eqs.
will
take
6.2.3 and 6.2.4,
place,
successful
i.e.
C =0,
otherwise
agglomeration colli-
sions will only occur for
C1 > 2.21(p/a M) 1/3 (C /v)
1
=
c *
(6.2.5)
(6) For any temperature-time history of the agglomerating coal particles,
their instantaneous metaplast content M(t) can be calculated from the
global kinetic model for liquid generation and depletion as described
in Section 6.1.
(7) In practice the neck radius cannot exceed the particle radius, so the
result in Equation 6.2.5 complies with
- 168 -
E(c1) =
0
((ci) =
1
0 < c1 < c
replacing
< 1.0
c 1 < c1 *
These constraints
are included
in
(6.2.6)
the agglomeration
kinetics
by
Eq. 6.2.1 with an effective probability of forming a
in
successful agglomerate E(t),
obtained
by averaging
E(c1) over all
allowed values of ci. Since the probability of having a given value
described by a continuous distribution function,
of ci is
and since
the collision times are much smaller than the plastic period, the
required averaging can be performed as an integral.
1. 0
((t)
P(c
=
(6.2.7)
)dc 1
)(c
and using Equation 6.2.6
J
1
((t)
=
-
/zr
where P(ci) is
6
1.0
exp
-
dc1
-
cx*
2 62
(6.2.8)
the normalized gaussian distribution function for ci
values. The quantity
E(t) is
time and flow field dependent
because
the lower limit of the integral depends on M, and one 0 /v(Eq. 6.2.5)
Thus for a given turbulent energy dissipation rate and initial number
density of coal particles,the binary agglomeration equation can be
solved by substituting Eqs. 6.2.2, 6.2.5, 6.2.8, into Eq. 6.2.1.
(B) Multiple agglomeration process
The preceding discussion is for agglomeration of two single particl-
- 169 -
es of equal radii.
In reality,
two-particle agglomerates can grow larger
by colliding with other single particles or agglomerates. The process can
be expressed as:
Ai + Ag
Ai+j
--
(6.2.9)
where Aj are the j-particle agglomerates
with a population density nj.
Assuming no agglomerates breakup after permanent agglomerates are formed
and neglecting particle loss due to adherance to the reactor walls, the
formation rate of Aj is:
= -dt
where
N-j
1 j-1
dnj
Ki,j-ininj-i -
is
overall agglomeration constant,
the
frequency,Kc,
collision
E(t),
collision,
(6.2.10)
i=l
2 i=l
Ki,j
Ki,ninj
for
and probability
the
agglomeration
agglomerate
purpose,
a
maximum
together
is
allowed in
size
the present
of
agglomerate
=
a (0) 1/3
per
forming
an
of
Ai
Ag. For computation
ten
and
single
and related
It
particles
sticking
is assumed that all
to the
initial
single
particle radius by
a
of
of
calculations.
j-particle agglomerates are spherical
i.e. the product
(6.2.11)
The collision constant Kc for Ai with Aj is (Saffman and Turner,1956)
- 170 -
K
=
C
1. 3 (a.
.)
+ a
i
3
/
still
The quantity E(t) is
00
/v
(6.2.12)
6.2.8 assuming the distribution
given by Eq.
of ci values is unchanged.
(C) Effect of particle swelling
Significant swelling of softened bituminous coal particles can occur
during
rapid
pyrolysis
due
to
expansion
of
the plastic
coal
mass
by
pyrolysis-derived vapors. This changes the collision cross section of the
particles. To account
to
grow in
for swelling,
a viscous,
spherical
a single volatile bubble is
droplet.
particle radius at time t to the initial
Q
Defining
as
particle radius,
assumed
the ratio of
it
is
shown in
Section 6.3.1. that,
d
(T/T
)
2a
ig3(1-Y
d[Q 3 -(1-c)]
dQ
1
initial
,
,
viscosity
of
the
/3
Q
(6.2.13)
Q3
respectively are the ambient pressure,
surface tension,
liquid-solid
1
~ 1
a s * T, T 0 ,
porosity,
[Q 3 _(l-_)]
R
41Q2{ Q 3
_(11)
where P.
11
temperature,
suspension
of
initial
temperature
the softening coal.
and
Volume
shrinkage due to pyrolysis weight loss is not considered. Because of the
presence of mineral matter and non softening macerals, the softened coal
is
in
reality
a
solid/liquid
suspension.
related to the liquid content by Eq. 6.1.1.
Its
apparent
viscosity
is
- 171 -
During
the
agglomeration
process,
the
particle
or
agglome-
rate radius is thus modified according to,
a. =
J
a
o
(j)1/3
(6.2.14)
Q
(D) Agglomeration in laminar flow
In a laminar flow particle particle collision rate is expected to be
lower
than
gradient
in
the
direction
Smoluchowski (1917)
travel,
K
The
under turbulent
=
4(a 1
conditions. Defining
perpendicular
to
the
Y as
the
direction
velocity
of
particle
obtained
(6.2.15)
+ a )3y/3
corresponding bending stress
for a
pair of agglomerated
spheres
given by Yoshida (1979),
ab =
30py/c 1 3
(6.2.16)
and the allowed range of ci values for successful agglomeration becomes
3. 11(pya
M)1/3
<
c
<
1. 0
ESTIMATION OF PARAMETERS
(A) Parameters in agglomeration equations
(6.2.17)
is
- 172 -
Depending on each particular process, the initial particle size and
number density of the feed can vary. For stoichiometric combustion of 40
of
feed
an initial
air,
in
coal particles
pm diameter pulverized
2800
particles/cm 3 is required. The turbulent energy dissipation rate per unit
can
volume
also
vary
widely
between
different
processes
different regions of the same process. For example,
injection
adjoining
nozzles
combustion chamber. Values of
combustors
estimate
but
can
be
can
greatly
exceed
and
values for regions
E
those
for
the
e 0 can be as high as 1010 cm2 /sec
lower in
significantly
e0
of the average value of
in
within
fluidized
in MHD
A rough
beds.
pipe
turbulent
3
main
flow with a
volume average velocity U is (Leslie, 1983)
where
f
0
=
(2/d)U 3 f
is
the
standard
(6.2.18)
Fanning
friction
factor
and
d
is
the
pipe
diameter. For values of U up to a few meters/sec and with a pipe diameter
of order 0.1 m, the range of c 0 is between 103 and 106 cm 2 /sec
ic
viscosities of entrainment
depending
gradient
on
y is
the
gas are
temperature. In
in
3
the range of 0.1 -
the laminar
flow case,
. Kinemat1 cm2 /sec,
the
velocity
estimated from the stress at the wall of laminar flow in
pipe. The range of
for practical
y
flows is
a
estimated at from 1 to 100
sec- 1 . The viscosity of the entrainment gas (helium) is calculated by the
Chapman-Enskog theory. The
estimated
from data
of
empirical
Sheomaker
et
constant
Go
al. (1976). The
in
Eq.
ultimate
6.2.3
is
strength
in uniaxial compression for a Pittsburgh bituminous coal in the softening
range
decreases
with
temperature.
At
343 &C, the
value
reaches
5x10 5
- 173 -
2
5
dynes/cm 2 . A value of 10 dynes/cm is taken for the present computation-
The parameters for the distribution of neck sizes c7
and 6 are arbitra-
rily chosen to be 0.1 and 0.01, respectively.
parameters
Kinetic
formation and
for metaplast
depletion
obtained
from rapid pyrolysis experiments with an electrical screen heater reactor
are
is
given above.
to
assumed
be
In
the particle swelling equation,
porosity
initial
the
and
50 dynes/cm
the surface tension
The
0.3.
is
which represents the mole fraction of volatiles dissolved in
quantity yi,
the metaplast, is arbitrarily taken to be 0.8.
(B)
Process variables
Results
Pittsburgh
are
No.8
presented
Seam bituminous
feed of 40 pm particles
initial
an
for
coal
at
a
particle number
density
of
of
2800 particles/cm 3 , heated at either 10000 K/s or 15000 K/s from 298 K to
constant
spheric
final temperatures between
600 -
1700
pressure. The largest agglomerate size is
K, in
helium
at
atmo-
limited to a
10-particle aggregate, A1 0 .
DISCUSSION OF RESULTS
Computed
swelling ratios
heating from room temperature
and intra-particle
(a)
to a final
liquid inventories
temperature
of
for
1000 K at
10000 K/s, and (b) to 1700 K at 15000 K/s are shown in Figures 6.2.2ab.
In both cases, the particle shows a maximum swelling of more than 1.3
times its
original diameter.
For the 1000 K final temperature,
there is
0
tc>
10000K/
1000K
A
1.5
1500oK/s
to
B
1.5
1.2
1.2
0.9
0.9
1700K
0'
:
:3
(0
0.6
0.6
0.3
0.3
100
0
0.2
0.4
0.6
Se
cond
0
0.02
0.04
0.06
SecCrnds
9.4
Figure 6.2.2
Liquid content and swelling ratio for two temperature
histories
- 175 -
more time for the particle to contract after maximum swelling, before the
its
of
cessation
at
plastic stage while
duration
1700 K the short
of
plasticity results in essentially no contraction. These swelling predicton
particles of larger diameter rapidly pyrolyzed under
vitrain
individual
measurements
recently reported
qualitative accord with
ions are in
10 atm helium (Lowenthal et al., 1985).
Maximum liquid yield is less for the higher heating rate case where
the
and depletion reactions
generation
liquid
before
are completed
the
final temperature is reached.
The growth of agglomerates containing from 2 to 10 of the original
particles (i.e. A2 , A3 , A4 ,...) in turbulent flow under the same temperahistories
ture-time
is
plotted
the
originally
Figure
6.2.3a
of
in
example
at
coal
fed
0.6s
that
andc
e 0 /v
values of
different
Figures 6.2.3a,b.
function of heating time in
fraction
for
The ordinate is
remains
/v= 105
as a
the mass
unagglomerated.
For
20% of
the
s-2
about
original coal is predicted to exist as two to ten particle agglomerates.
As
the
of
intensity
increases
and
values of c 0 /v
results
turbulence
in
However,
more agglomeration.
probability
agglomeration
,the
the collision
increases,
at
per collision,
frequency
very
E
high
is low
because of the high shear stress imposed on potential agglomerates by the
entrainment
gas and
less agglomeration
occurs.
By definition agglome-
ration must cease when the metaplast in the coal is depleted and Figures
6.2.3a,b bear
out
this
picture.
At
the
higher
heating
rates
and
final
temperatures, the duration of plasticity is much shorter, resulting in a
decrease in the extent of agglomeration.
1OOOOK/s
t
0o
OOOK
1500OK/s
to
A
1.0
1700K
B
1.00
104
15
0
0.9
0.95
0.8
0.90
106
H
0
0~i
'H
0
(U
0
4
0.7
0.85
0.6
0.80
0.5
0.75
a
C)
4.4
0
0.2
0.4
0.6
S e
Figure 6.2.3
co rids
I
SI
0
0.02
0.04
0.06
s econrIds
Weight fraction of unagglomerated particles at different
intensities of turbulence, numbers on curves indicate
E.// ,in s-2
- 177 -
Figures 6.24a,b show the time evolution of different size agglomerates with
two
EO/v
at 106 sec~2 , with the vertical separation between any
depicting
curves
which
particles
to
combined
has been
of
the
originally
form
the
indicated
fraction
mass
the
coal
fed
agglomerte,
Ag. Significant amounts of two, three, four and five-particle agglomeralarger agglomerates constitute
or
tes are formed. Six-particle
a small
mass fraction of the total initial feed at either heating rate.
Effects of
final temperature
are plotted in
agglomerates
Figures 6.2.5a,b for heating
Agglomeration
15000 K/s respectively.
K/s and
collision
efficiency,
does not
of
rates of 10000
stop until
the
and the
e 0 /v
of
which depends on the magnitude
metaplast content, reaches a small value.
fractions
mass
predicted
on the
Hence, to allow comparison of
the different cases, a residence time, 0.24 s is specified. Experimentally,
this would correspond
and
collide,
pyrolyze,
to allowing
the mixture of coal
agglomerate
under
the
given
particles to
temperature-time
history, then quenching the mixture infinitely rapidly and collecting and
counting the agglomerates.
effect
The
identical
c
/v
heating
of
in
rate can be
seen
by
comparing
Figures 6.2.6a,b. At lower heating rate,
the plastic period is
curves
(10000 K/s)
predicted.
longer and more agglomeration is
of
Figure
6.2.6a shows predictions for a much lower heating rate of 1000 K/s to the
same temperatures. A longer residence time (0.9 s) is imposed so that the
highest
final
rate. Even
temperature will
more agglomeration
still
is
be reached
at
this lower heating
predicted. Particle
agglomeration
in
to
1500oK/s
loooK
to
1ooooK/s
BBA
A
1700K
A- 4
-1
0
1.00
1.0
0.
0
0
4.4
0.9
0.95
0.8
0.90
H
-
0
C
0.85
0.7
0
0.6
0.80
0.5
0.75
4.4
0
0.2
0.4
s
Figure 6.2.4
0
0.6
C
rids
0.02
0.04
0.06
S I Co
C> rds
Weight fractions of different size agglomerates formed
at E./v equal to 106 s-2
0.24s
B
A
0.5
t-
15000K/s,
0.24s
t-
lOOOOK/s,
0.5
0
0
P4o
0.4
0.4
0.3
0.3
(0
1-
C
0.2
0.2
0
106
-.0.1
0.1
105
44
10
0
600
4
0
I800
1200
1000
Final
Figure 6.2.5
1400
Temp
600
a-ture,
1000
800
1200
K
Weight fractions of agglomerates vs final temperature
2
Numbers on curves indicate E/Y ,in s-
1400
- 180 -
laminar flow is much less, as seen in Figure 6.2.6b.
Effects of feed particle size distribution on agglomeration behavior
are
of
interest
since
Particles 250 -
sizes.
combustion.
practical
500 urm or
devices
larger
utilize
are of
sizes of 10 -
On the other hand,
35
a distribution
interest
to
fluid
of
bed
um command attention in
research on coal fired diesel engines. Assuming a single particle size
and a constant coal/gas mass ratio, then ao3 n = constant. From Eqs. 6.2.1
and 6.2.2, it can be seen that the fraction of agglomeration is independent
of
particle
agglomeration is
size
particle
Eq.
6.2.10
is
not sensitive to the initial
coal/gas feed ratio.
may be larger
ao. When
entrainment
eddy
is
not
the
fraction
of
particle size for the same
For larger particles,
than the
solved,
(>
200 um)
the particle size
size at the higher range
possible. The
collision
of
. Hence
F 0 /V
frequency
will
be
less and so is the extent of agglomeration.
The
agglomeration
Opportunities
swelling
for improvement
that
volatile
model
elaborates
bubble. Electron
pyrolyzed
coal
suggest
presented
include
upon
the
simple
a better
picture
micrographs
that a
is
of
satisfying.
description
of
char
"foam-coated
yet
expansion
particles
bubble"
of particle
of
a
from
model
may
single
rapidly
be
more
realistic. Particle fragmentation can also occur in pyrolysis or combustion processes, and
density.
For
thus
simplicity
contribute
this
to
changes
possibility
and
in the
the
particle number
detailed
motion
of
agglomerating pairs (spinning, rotation) are not considered here. At very
intense
levels
of
turbulence,
flow
of volatiles away from the coal;
field
effects
on
mass
and for very fluid coals,
transfer
on recircu-
t0.05,
AOOOK/
0
0
lOOO
9
A
K/s,
ciaE
laminr
0.5B
0.8
0.04
0
0.6
0.03
40
HOD~
H-
0.4
0
0.02
0
0.2
0
0.01
I
I
600
800
1000
1200
Firia
Figure 6.2.6
0I
I
'-
1400
Te
0Mpera
600
800
X C1tureAX-C3,
I
I -
1000
1200
K
Weight fractions of agglomerate vs final temperature
for (a) 1000 K/s heating rate, (b) laminar flow
Numbers on curves indicate fo/v ,in s-2 or velocity
gradient Y,in s-2
1400
- 182 -
lation
within
effects
could modify
the molten
the
coal,
would
global
have
to
be accounted
for. Both
plasticity kinetics. Collision
due
to
sedimentation of different size particles can also be important at lower
levels of turbulence.
The
ration
model
predictions
are important
in
show that
explaining
both
changes in
swelling and
particle
agglome-
size distri-
butions in coal combustion and gasification. The model provides a priori
prediction
of a maximum
The
temperature.
(>
rates
fields
also
6-particle) is
of
improving
form is
model
practical
its
in extent
of
of
agglomeration
predicts that
relatively
formation of
minor importance
interest. Although
physical
and chemical
applicable to laminar
with
there
features,
and turbulent
are
increasing
large agglomefor many
opportunities
the model
in
its
flow
for
present
entrained flow pyrolysis or
combustion of pulverized coal particles.
Devolatilization kinetics
and depletion
local
(and
rates
are important for agglomeration
metaplast
pressure,
including
inventories as
particle size,
type).
metaplast
formation
modelling. They determine
a function of
and coal
of
temperature-time history
The metaplast,
or
liquid
formation/depletion kinetics influence at least two phenomena, namely the
swelling dynamics due to effects on melt viscosity and volatiles release
into
bubbles;
and
the
agglomeration
strength of the agglomerate pairs.
dynamics
due
to
effects
on
the
- 183 -
6.3 OTHER POSSIBILITIES OF MODELING
Other chemical
thermal
processing
and physical behavior
can also be modelled
of bituminous coal during
with the
liquid
generation and
destruction kinetics. A few possibilities are briefly mentioned.
6.3.1. Swelling of pulverized coal particles
Electron
ed
micrographs
of
char
flow reactors show existence of
large
cavity
in
the center)
particles
fed
through
both cenospheres
and particles with
entrain-
(particles with a
numerous small
cavities
which impart a foam-like appearance to the particles.
Two
large
cavity
equations
for
(bubble),
and
describing
the
separately
swelling
of
the
the
swelling
of
liquid material
the
surr-
ounding the bubble are given below.
(a) Swelling of the large bubble
The
droplet with
model
here is
for the
a gas bubble trapped
expansion
of
a spherical,
inside. For pulverized
coal
viscous
particles
with diameters about 100 pm or less, the temperature inside the particle
can be assumed to be spatially uniform at heating rates of up to
K/s.
An
increase
in
the
particle
temperature
causes
(1)the
1000
coal
to
soften, (2)the trapped gas to expand, and (3) the softened melt to evolve
volatiles,
creating
a
higher
gas
pressure
inside
the
particle. These
events contribute to the swelling of the particle. Besides the assumption
- 184 -
of
uniform
temperature
it
particle,
the whole
inside
is
also assumed
that the flow of the softened coal melt is incompressible and Newtonian,
and
the
loss
of liquid due to evaporation
liquid phase does not change the volume of
and diffusion away
from the
the liquid to a significant
extent. The last assumption should be true at the early stage of swelling
when relatively little devolatilization has occurred.
Consider the following coordinate system (Figure 6.3.1.), where
Pb = pressure inside bubble
= pressure outside particle
P
RO = original radius of particle
Ri = radius of bubble inside particle
R2 = radius of softened particle
e = initial porosity of particle
To = initial temperature of particle
T
= temperature of particle
u
= viscosity of coal melt
as = surface tension of melt
By symmetry the velocity field
(vr, v0 , v4
)
=
(vr,
0
)
Applying the continuity equation for the liquid,Vov
r 2 vr
1 2R
=
R
22
=
C (t)
(6.3.1)
=
0 gives
(6.3.2)
- 185 -
R
=
original radius of particle
Swelling ratio Q
A.
Singl
3j
IBub-L
a<j
DI
l L<
=
R /R
Sw E 1 1 LirI
unsoftened coal
-liquid (foaming)
coke (resolidified)
IB.-
Figure 6.3.1
1;,
Rea
c-t,-j
in
c
Fo>
m
Moc:d eL
Limiting cases of swelling (a) single bubble swelling
(b) reacting and solidifying foam
- 186 -
At the interface r = Rl, a momentum balance gives
avv
Pb
IR1
a
+ -S
IR
(6.3.3)
ar1
and similarly at r = R2,
a
P O=
+
- 2V
y
P
3r
2
S
(6.3.4)
R2
andP jR2 are the pressure inside the liquid at r = Ri and r
here P IR
=
R2 , respectively.
The Navier -Stokes equation gives
3V
3V
+
p (r
at
Substituting
r)
vr
r
C(t)
ap
=
(6.3.5)
-
-
r
3r
from
Equation
into
6.3.2
and
6.3.5,
integrating
from Ri to R 2 gives
dC(t)
P R
2
1
=
(-
p
dt
1
-)
r
C2 (t)
+
2 r4
I
(6.3.6)
By mass conservation one finds
R2 3 -
Subtracting
R13
=
R
3(i -
Equation 6.3.4
(6.3.7)
C)
from 6.3.3,
and
subtracting
IR1 -
Equation 6.3.6 into the resulting expression gives an equation in
IR2from
terms
187 -
of C(t), Ri and R2 . Defining R* = R 2 /Ro and making use of Equation 6.3.7,
finally we arrive at,
1
1
P b-
P
2
pR
(2R*R*2 + R*2R*)(
3
[R*
1
1
+ -(R*
2
2
+ 4yR*
2
R*
2
R*
4
[R*
3
_(1
E))4
-
by
the
3
R*
3
-
(
)3
_3
1
-)
-
R*
)
-
3
(6.3.8)
+ --
(
)
[R* 3 -(1-c)11/3
0
The
e)]1
1
R*(
R*
2a
-S
R
-
,
)(
1
+
-(1
first
term
viscous term
on
the
and
the
the
is
right
surface
term,
inertia
tension term. For
an
followed
order
of
magnitude analysis, let
h
RO
= 100 um
as
= 50 dynes/cm
p
= 1 gm/cm 3
y
= 104 poise
= 1 atm
-CO
~
Rewriting Equation 6.3.8 as
PPb
b-
P O4
~_e_4
=
a
_
+
+
a
pR 2 2
PR 2
t
0(1010)
0(a 1 )
0
0(10
8
)0(a
2a
- -a
pR 3 3
0
2
)
(6.3.9)
0(108)0(a
3)
where al, a2 and a3 respectively represent the terms in braces, from left
to right in Equation 6.3.8.
It
is
the term a2,
easy to see O(a3) ~ 0(100).
This forces 0(a2) ~0(102).
In
- 188 -
1
1
R*2(-
R* 3 -
R*
(1-e)
Hence 0(R*) ~ 0(102
-
0(101
-
3
100)
(6.3.10)
_ 103). This implies O(ai) ~ 0(106).
So,
the final
numbers are,
Pb
-
04P
a
=
PR 2
+
pR
1
T
0(1010)
0(106)
-
if Pb
T
The
P
=
2a
+ ---a
pR 3 3
2
T(6.3.11)
0(1010)
0(108)
1 atm
above shows
that
(1) inertia
terms
can be
neglected,
(2)
surface tension effect cannot be neglected if the pressure difference is
small (less than 0.1 atm).
To
evaluate
Pb
-
P
consider
no
to
be
the
number
of
moles
of gas (assumed inert) initially trapped inside the particle occupying a
volume
heating,
4/3(rrR 0 3)
at
temperature
this volume of gas expands,
a pressure
To and
Also,
Po (=
P.).
Upon
volatiles diffuse across r
=
Ri.
dn /dt
n (0)
=
4rR 1 2 k (y
=
no
-
yb
(6.3.12)
yi is the gas phase mole fraction of volatiles at the interface and yb is
the mole fraction of volatiles
inside the bubble.
ky is the local mass
transfer coefficient based on the gas phase. Solving for n in Equation
6.3.12 and using the ideal gas law allows P b
limiting cases exist.
-
PC, to be calculated. Two
- 189 -
(1) ky
very slow volatiles transfer across
0
--
interface.
(2) ky
y
volatiles in the bubble attain
equilibrium with volatiles in the
liquid melt
=
Define AP*
b
O
then for the case (1),
C (T/TO)
AP*
(6.3.13)
=
R*
3
_ 1
_
(1-)
and for case (2),
E (T/T )
(1-y )[ R*
In
principle,
function
of
3
(6.3.14)
_(1_e)]
yi can be estimated
the metaplast
from Henry's law constants and is
concentration
in
the
melt.
Hence,
the
a
final
equation would be,
e (T/T )
2a
1
47rR*
2
1
[
_
R*
3
-(1
-E)
R*
3
(6.3.15)
The viscosity y can be obtained from the plastometer data or calculated
from
the
plasticity
model
as
described
global model for coal plasticity kinetics,
in Section
6.1. Recalling our
190 -
-
coal
metaplast
-
coke
-
+
volatiles
melting
The parameters
related
k1 , k 2
to metaplast
are given
in
Section 6.1.
concentration M (wt.
The viscosity
fraction of coal)
is
then
by Equation
6.1.1.
Figures
6.2.2a,b
shows
calculated
results
of
swelling
for
40 um diameter coal particles heated at (a) 10000 K/s to 1000 K, and (b)
15000 K/s to 1700 K.
(b) Swelling of the foam layer
The
ing
swelling
of plastic
quite similar
foams.
of
the
foam
Expansion
of
layer is analogous
foams by chemical
to swelling of a devolatilizing
cellular structure at the end of expansion.
coal
to the
blowing
liquid.
processagents is
Both give a
Figure 6.3.2 shows expansion
of a cross-linked polyethylene. During the expansion process the polymer
has a low viscosity. As the degree of cure (as indicated by gel formation)
increases,
elasticity.
swelling
coke,
6.3.1).
the foam is
An
emprical
stabilized
approach
by the increases of viscosity and
is
attempted
ratio. Consider a reacting
their weight
fractions
mixture
denoted
The swelling of the liquid is
by
of
C,M,G,E
here
coal,
to
describe
liquid,
respectively
due to the gas generation,
gas
the
and
(Figure
and is
assumed to be proportional to a a power of the generation rate, dG/dt via
a constant a . The unreacted coal in
the suspension is
assumed having an
-
191 -
1.0
0.9
0.8
D
80
0
0.7
70
0.6 -
60
>0.5 -
Ln
-50
E
0.4 0.3 -
- 40
:-
30
0.2 -
- 20
0.10
0
10
5
Figure 6.3.2
10
15
Time at 175'C, min
20
0
25
Schematic diagram of expansion of
a polyethylene foam (Benning,1969)
- 192 -
initial
density
liquid,
so its
p. . The
porosity
liquid. Therefore,
at
any
coke
porous
reflects
is
continuously
the instantaneous
time, the volumetric
formed
from
the
of
the
porosity
swelling ratio Q(t)
is
given by
8t
C/p0 + [ 1 + a(dG/dt)
] M/p0 + f [(dE/dt)/p +E] dt
Q(t)
0
M(6.3.16)
(1/p 0 )
The
separate
volumetric
terms
of
contributions
of
numerator.
The
the
M+E
liquid is related top 0 by
"
Q(t) = C + [ 1 + a(dG/dt)a]M +t
p
C,
M and
E are
instantaneous
/[ 1 + a(dG/dt)e
in
the
density
three
of
the
],therefore
(dE/dt)[1 + a(dG/dt)
] dt
(6.3.17)
ratios for
the same
0
Figure
6.3.3
shows
calculated
swelling
conditions of Figures 6.2.2.
The
two expressions above
probably
represent
the
two
limiting
coal
particle
cases of the physical picture of coal swelling.
6.3.2.
Fragmentation of large coal particles
"possible model"
When
a
large
(a
few
mm
or
larger)
softening
is rapidly heated, the non-isothermal temperature distribution creates a
plastic
layer
(Figure 6.3.4).
that
is
progressing
towards the center
of
the
particle
15000K/s
to
00K 1000K
10OOOO0K/:s
3.0
3.0
2.0
2.0
1.0
1.0
0.6
0.6
0.3
0.3
to
L'70OK
170
H
Liquid wt. fraction
Liquid wt.
fraction
0
0
0.0
0.2
0.4
0.6
sEc3 <) C) M d S
Figure 6.3.3
0.0
0.02
0.06
0.04
S C
C)
Liquid content and swelling ratio, same temperature
histories as in Figures 31a,b
M
i as
- 194 -
Progressing plastic zone
unreacted coal
(non-porous)
plastic layer (mostly liquid)
coke layer (pyrolyzed porous layer)
r-
a
liquid
wt.
%
raw coal
Outline of computation
UNSTEADY STATE HEAT TRANSFER
hA
-
PLASTICITY KINETICS
CALCULATE LIQUID FORMATION RATE
PRESSURE DIFFERENCE FROM Darcy's Law
STRESS ON OUTER COKE SHELL
Figure 6.3.4
r
-progressing plastic zone
Possible model for coal particle fragmentation
- 195 -
As
the liquid
results. Initial
layer is formed,
density
of
the coal
an increase in
may be
specific volume
between 1.2
while the liquid may have a density near 1.0 g/cm3 .
-
1.4 g/cm
3
,
Hence an increase
of 20 - 30 % by volume can happen in the initial softening step. Both the
liquid and the volatiles formed must flow outwards through the outer coke
layer,
since
sity. The
the inner unreacted
volatiles can
core of coal
has a much smaller
poro-
escape much faster than the liquid, while the
liquid undergoes coking and secondary reactions
as it
flows through the
porous coke layer.
Even
large
though
differences
in
the
volatiles
their viscosities
that the pressure generated
viscous
liquid.
formed
(
have a
larger volume,
0.02 cp
vs
by the flow os species is
The volumetric
formation
rate
of
>1000
p)
the
means
mainly due to the
the
liquid
can
be
calculated from plasticity kinetics, this is equated to the flow through
the porous coke layer by Darcy's law to obtained the pressure difference
required for flow. The stress on the outer coke shell is then calculated
from
the
strength
pressure difference.
of
coke,
the outer
If
the
shell
stress
exceeds
cracks, and
a
the
critical
few fragments
formed. The equations to be solved are given in Figure 6.3.5.
are
- 196 -
r
jar + r
a -
ar
a (T)
T(Or) -
0
-
1
T(t,a)
1I(t,O)
-
-k1(C
am
k
-
fM) -
)
rm
-v
-kM+r
M
C(r,o)
2
m
LIQUID
-
Q(t) -
(
FORMATION
rar
loss term due to flow, assumed small
1.0
M(r,O)
TRANSFER
0
(C -f
1
at
HEAT
0
r* t)
)
r*
4wr2( 4) dr
at
00
P
is
the outer radius at which M is
(1 aP(t)
l).
-
)
viscosity of liquid
88
L (a/r*)3
2
-
CALCULATE
-
FLOWRATE
less than some small value
PRESSURE DROP
permeability of coke layer
1]
STRESS ON
COKE LAYER
the
particle breaks up if stress is greater than
strength of the coke
Figure 6.3.5
Equations for fragmentation model
- 197 -
CHAPTER 7. CONCLUSIONS, POTENTIAL APPLICATIONS AND
RECOMMENDATIONS FOR FUTURE WORK
CONCLUSIONS
(1) A
plastometer has
new coal
changing
apparent
temperatures,
conversion
been developed
viscosities
pressures
processes
and
but
at
severe
heating
conditions,
rates
previously
to measure
pertinent
the rapidly
i-.e.
at
high
to modern coal
unattainable with
conventional
instruments.
(2) Coal plasticity can be described directly by its viscosity change. As
a Pittsburgh Seam bituminous coal is heated, the coal softens to form
a solid/liquid suspension the viscosity of which decreases from more
than
one million poise to minimum
values
of
order
a few thousand
poise. Resolidification of the intermediate liquid phase is indicated
by rapid rise in viscosity values.
(3)
Characteristic
times
for
softening
tained.
The measured time interval
minute
to
less
than a second,
and
resolidification
were
ob-
of low viscosity varies from one
depending on
the temperature-time
history.
(4) Initial softening of a Pittsburgh coal is
estimated 25% by weight of the coal.
is
further
liquids
can
converted to
polymerize
liquids
to
give
by
by physical melting of an
Up to 40% by weight of the coal
pyrolytic
higher
bond breaking. These
molecular
weight
liquids
- 198 -
and
solids, crack
to
give
gas
and
coke,
escape
or
from the
coal
particle to give tar. The result is that the average liquid molecular
weight
can
increase
from 570
above
to
and
800,
decrease
then
to
around 600.
(5) The
can
intraparticle coal
kinetics
of
be
accurately
quite
electrical
screen
for
determined
heater with
liquid
rapid
a
formation and depletion
coal
bituminous
in
followed
sample quenching,
an
by
solvent extraction of the quenched char.
(6) The new
information
application
in
on coal plasticity kinetics finds quantitative
modeling coal
particle
agglomeration
and
swelling
behavior under rapid heating high temperature conditions. Predictions
of model here developed were in
qualitative
accord with
literature
data on swelling and provide a possible explanation for coal agglomeration observed within a pulverized coal flame.
POTENTIAL APPLICATIONS
(1)
The results provide
an
approach
for
describing
agglomeration
behavior in gasifiers and combustors.
(2)
The
improved
physical
and
softening coals can be used in
chemical
picture
of
the
pyrolysis
of
improved description of the chemical
kinetics and transport processes that determine the rates and extents
of coal softening, pyrolysis and liquefaction.
(3) The
data showed that
suitable thermal
treatment
followed by
rapid
-
quenching
can
give
high
199
-
liquid
yields,
exceeding
80 % daf
from
bituminous coal without addition of catalyst or hydrogen. A possible
application is improved production of liquids from coal.
(4) The high temperature, fast response plastometer can be applied to
the study of other thermoplastic materials.
- 200 -
RECOMMENDATIONS FOR FUTURE WORK
The present
points
study opens some interesting
for future research
in
softening coal pyrolysis:
1. Further
experiments
the effects of
on
heating
rate on plasticity is
thesis:
viscosity
needed
extract
curves,
reactive gases and
pressure,
(using the approach
generation
in
the present
depletion
and
study
in
conjunction with the characterization of the extract).
2. The above approach can be applied to a range of coal types (lignite,
sub-bituminous and bituminous coals of different
fluidity)
for a better
as
determining
understanding of coal pyrolysis in general.
3.
Chemical
the H/C
characterization
ratio and
of
the
extracts,
aromatic content will
understanding of the chemistry of coal
such
be useful
plasticity,
for a more
including
detailed
runs
with
samples pretreated with tetralin.
4.
More
work
is
needed
in
the
study
of
resolidification
reactions
of pre-treated coals, or for raw coal in the presence of (a) heterogenous
surfaces,
(b) catalysts,
(c) hydrogen
donors, and
(d) secondary
reac-
tions, such as increased contact with its pyrolysis products in a closed
system.
5. More detailed molecular weight information can be obtained from the
newly completed
screen heater
reactor, since the escaping tar is
also
- 201 -
can
and
collected
be analyzed.
light
shed
This can
on the mode of tar
secondary reaction and transport in coal.
conditions
6. For
pertinent
kinetics
can be determined by
passing
through
coal
to pulverized
studying
the extract
plasticity
yields of particles
rapidly quenched
tube furnace and
a drop
combustion,
on a cold
collector.
7.
By
combining
single
the
bubble
and
foam
swelling
a
models,
realistic description of particle swelling can be achieved.
more
This with
other modifications in the description of particle sticking (or fragmentation) can lead to a better understanding of the behavior of a system of
softening, pyrolyzing coal particles under realistic conditions.
8.
in
The effective diffusivity of pyrolysis
a
viscous
(assuming
no
coal
melt
internal
can
motion
be
of
products or gaseous
estimated
the
melt).
reactants
from its viscosity
This
is
useful
values
for
the
modeling of softening coal pyrolysis.
9. The present work suggests that an optimum solids residence time exists
at each reaction temperature for maximum liquids yield in the absence of
hydrogen. A
two-stage
coal
liquefaction
process with
first
thermal
treatment followed by rapid dissolution in a solvent seems possible. The
quality of the liquids should be studied as mentioned in Point 3.
- 202 -
REFERENCES
Anthony, D.B., J.B. Howard, H.P. Meissner, and H.C. Hottel, Rev. Sci. In-
strum., 45, 992 (1974)
ASTM "Standards on Coal and Coke" 1974 Annual Book of ASTM Standards,
American Society for Testing and Materials, Philadelphia, PA (1974)
Attar, A., "Bubble Nucleation in Viscous Material Due to Gas Formation by
a Chemical Reaction, Application to Coal Pyrolysis" AIChE J. 24, 106
(1978)
Bhatia, G., E. Fitzer, and D. Kompalik, "Rheological Characteristics of
Mesophase Forming Pitch Materials up to 500 OC" Abstract, 17 th Biennial
Conf. on Carbon, Lexington, Kentucky pp 165-7 June 1985
Benning, C. J., "Plastic Foams" Vol 1, p310. Wiley, New York 1969
Boroson, M.,
unpublished results, 1985
Chermin, H.A.G. and D.W. Van Krevelen, "Chemical Structure and Properties
of Coal XVII - Mathematical Model of Coal Pyrolysis" Fuel, 36, 85 (1957)
U. Gat,
and V.K. Dhir, "Decaking
Crewe, G.F.,
alkaline solutions" Fuel, 54., 20 (1975)
of
bituminous
coals
by
Culross, C.C., L.R. Clavenna and R.D. Williams, "Technique for Measuring
Swelling Tendency and Coke Density for Catalytic Coal Gasification" ACS
Div. Fuel Chem. 29, 181 (1984)
-
Davies, C.E., S.T. Pemberton and J. Abrahamson, "Capillary Viscometry of
a New Zealand Coal at high pressures and shear stresses" Fuel, 62, 417
(1983)
Fitzgerald, D., "The Kinetics of Coal Carbonization in the Plastic State"
Fuel, 35, 178 (1956)
Fong, W.S., W.A. Peters, and J.B. Howard, "Basic studies of coal plasticity" Proc. 2nd Annual Pitteburgh Coal Conf., pp 626-637 (1985)
Fong, W.S., W.A. Peters, and J.B. Howard, "Apparatus for determining high
temperature high pressure coal plastic behavior under rapid heating
conditions" Rev. Sci. Instrum., 56 (4), 586 (1985)
Fong, W.S., Y.F. Khalil, W.A. Peters, and J.B. Howard, "Plastic behavior
of coal under rapid heating high temperature conditions" in press, Fuel,
(1986)
"Kinetics of generation
W.A. Peters, and J.B. Howard,
Fong, W.S.,
and destruction of pyridine extractables in a rapidly pyrolyzing bituminous coal" in press, Fuel, (1986)
Frankel,
N.A.,
and
A. Acrivos,
"On
the
Viscosity
of
a
Suspension of Solid Spheres" Chem. Eng. Sci. 22, 847 (1967)
Concentrated
- 203 -
Franklin, H.D., "Mineral Matter Effects in Coal Pyrolysis and Hydropyrolysis", Ph.D. Thesis, Dept. of Chem. Eng., M.I.T., Cambridge, MA (1980)
"Fundamentals of Coal
2nd Suppl. Vol., Wiley,
Gorin, E.,
Utilization,
Liquefaction" in
New York, 1981
Chemistry
Coal
of
Habermehl, D., F. Orywal, and H.D. Beyer, "Plastic Properties of Coal" in
Chemistry of Coal Utilization, 2nd Suppl. Vol., Wiley, New York, 1981
"A preliminary account of char structures produced from
pyrolyzed at various heating rates" Fuel, 59, 112
Hamilton, L.H.,
Liddell vitrinite
(1980)
"Fundamentals of Coal Pyrolysis and Hydropyrolysis"
J.B.,
Howard,
in Chemistry of Coal Utilization, 2nd Suppl. Vol., Wiley, New York, 1981
Ignasiak, B.S., A.J. Szladow, and D.S. Montgomery, "Oxidation studies on
coking coal related to weathering. 3. The influence of acidic hydroxyl
groups created during oxidation, on the plasticity and dilatation of
coking coal" Fuel,
53, 12 (1974)
and A.F. Mills, "Analysis of
James, R.K.,
Letters in Heat and Mass Transfer, 3_, 1 (1976)
Coal
Particle
Pyrolysis"
Kaiho, M., and Y. Toda, "Change in thermoplastic properties of coal under
pressure of various gases" Fuel, 58., 397 (1979)
and R.G. Jenkins, "Thermoplastic properties of
M.R.,
Khan,
elevated pressures:
1."Evaluation of a high-pressure microdilatometer" Fuel,
coal
at
63,
109
Fuel,
64,
(1984)
preoxidation
2."Low-temperature
of a Pittsburgh Seam Coal"
189 (1985)
Kirov, N.Y., and J.N. Stephens, "Physical Aspects of Coal Carbonization"
Kingsway Printers, Caringbah, Australia, 1967
Klepel, G., Freiburg. Forsch-H, A 141, 59 (1960)
particles
Klose, W., and M. Lent, "Agglomeration kinetics of coking coal
during the softening phase" Fuel, 64, 193 (1985)
Lancet,
M.S.,
and F.A.
Sim,
"The
Effect of Pressure and Gas Composition
on the Fluidity of Pittsburgh No. 8 Coal" ACS Div. of Fuel Chem.,
(1981)
Leslie, D.C., "Developments in
sity Press, 1983
Lewellen,
P.C.,
"Product
26, 3
the theory of turbulence" Oxford Univer-
Decomposition
Effects
in
Coal
Pyrolysis"
M.S. Thesis, Dept. of Chem. Eng., M.I.T., Cambridge, MA (1975)
Lightman,
P.,
and
P.S.
Street,
"Microscopical
Examination
of
Heat
- 204 -
Treated Pulverized Coal Particles" Fuel,
47, 7 (1968)
Llyod, W.G., H.E. Francis, M.R. Yewell Jr, R.O. Kushida, and V.D. Sankur,
"A model for the isothermal behavior of coals" ACS Div. Fuel Chem., 25
(2),
128 (1980)
Llyod,
W.G.
and
others,
"Predictors
of Plasticity
Bituminous Coals"
in
Final Technical Report DE-FG22-81PC40793 (1984)
Lowenthal, G., W. Wanzl, and K.H. Van Heek,
Science, pp965 Pergaman Press (1985)
Proc.
1985 Int'l
Coal
Conf.
"Gasification of
behaviour of a
Y. Nishiyama, H. Sawabe, and Y. Tamai,
Matsunaga, T.,
coals treated with non-aqueous solvents. 3. Swelling
single coal particle" Fuel, 57, 562 (1978)
McCarthy, D.J., "Design and characterization of an apparatus for agglomeration testing during flash pyrolysis in entrained flow reactors" Fuel,
59, 563 (1980)
Melia,
P.F., and C.T. Bowman, "A Three Zone Model
Swelling" Combustion Sci. and Tech., 31, 195 (1983)
for
Coal
Particle
Montero-Garcia, J.H.,"Molecular weight characterization of the tar
products of coal pyrolysis using the technique of Gel Permeation Chromatography" S.B. Thesis, Dept. of Chem. Eng., M.I.T., Cambridge, MA (1983)
Nazem,
F.F.,
"Rheology
of
carbonaceous
mesophase
pitch"
Fuel,
851
59,
(1980)
R.C., "Coal PLasticity Mechanism: Inferences from Liquefaction
Neavel,
Coal Science, Vol. 1, M.L. Gorgaty, J.W. Larsen and
Studies" in
I. Wender (eds) Academic Press, New York 1982
Oh,
M.,
"Softening
Coal Pyrolysis",
Sc.D. Thesis,
of Chem.
Dept.
Eng.,
M.I.T., Cambridge, MA (1985)
"Viscosity
T.,
S. Yokoyama, and Y. Maekawa,
Okutani,
164
(1984)
63,
Fuel,
hydrogenation"
during
paste
changes
in
coal
Ouchi, K., T. Kazuyuki, M. Makabe and H. Itoh, "Relation between fluidity
and solvent extraction yield of coals" Fuel, 62, 1227 (1983)
Pierron, E.D., and O.W. Rees, "Solvent extract and the plastic properties
of coal" Illinois State Geological Survey Circular 288 (1960)
Pohl, J.H., H. Kobayashi, and A. F. Sarofim, "The effect of temperature
and time on the swelling of pulverized coal particles" Presented at
Combustion Institute Technical Meeting, Boulder, Colorado (1978)
R.B.,
P.J. Reucroft and W.G. Lloyd, "Rheological
Read,
coals" Fuel, 64, 495 (1985)
bituminous
selected
properties
of
- 205 -
Saffman,
P.G.,
clouds"
J. Fluid Mech., 1, 16 (1956)
and J.S. Turner,
"On
the collision of drops in
turbulent
Saint Romain, J.L., private communications, 1985
Scharff, M.F. and others."Computer modeling of mixing and agglomeration
in coal conversion reactors" Final Report for USDOE, DE-AC21-78ET10329,
June 1981
Shoemaker, H.D., L.Z. Shuck, R.R. Haynes and S.H. Advani,
"Directional
Viscoelastic Properties of the Pittsburgh Coal at Elevated Temperatures
in Compression and Shear" MERC report, RI 76/5 USDOE, August 1976
Smoluchowski, M.,Z. Phys. Chem., 92, 129 (1917)
Stephens, J.N.,"A new coal plastometer" Fuel, 42, 175 (1963)
Suuberg, E.M.,"Rapid pyrolysis and Hydropyrolysis of Coal" Sc.D. Thesis,
Dept. of Chem. Eng., M.I.T., Cambridge, MA (1977)
Tyler, R.J., "Flash pyrolysis of coals. 1. Devolatilization of a Victorian brown coal in a small fluidized-bed reactor" Fuel, 58, 680 (1979)
2. Devolatilization of bituminous coals in a small fluidized-bed reactor"
Fuel, 59, 218 (1980)
Unger, P.E., and E.M. Suuberg, "Molecular weight distributions
produced by flash pyrolysis of coals" Fuel, 63, 606 (1984)
Waters,
P.L.,
"Rheological
of tars
Properties of Coal during the Early Stage of
Thermal Softening" Fuel, 41, 3 (1962)
Whitehurst,
D.D.,
"Coal liquefaction fundamentals" ACS Symposium Series,
139, 1980
Wilhelm,
A.,and K. Hedden,
"Noniosthermal extraction of
solvents in liquid and supercritical state" pp 6-9, Proc. Int'l
Coal Science, Pittsburgh, August 1983
Van Krevelen, D.W.,
"Coal: Typology, Chemistry,
Elsevier Publishing Co. Amsterdam, 1961
Yoshida, T., Y. Kousaka, and K. Okuyama,
Power Co. Ltd., Tokyo, Japan 1979
Physics,
coal with
Conf. on
Constitution"
"Aerosol Science for Engineers"