MIT LIBRARIES DUPL 1
I lll l 1
1N 111IIil
11uN
3 9080 00576822 8
1
PYROLYSIS OF DIFFERENT COAL TYPES
by
Glen H. Ko
B.A.Sc in Chemical Engineering
The University of British Columbia (1984)
Submitted to the Department of Chemical Engineering
in Partial Fulfillment of the Requirements for the
Degree of
DOCTOR OF PHILOSOPHY
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
August, 1988
@ Massachusetts Institute of Technology 1988
Signature of Author
Department of~2IThe-ical Engineering
August, 1988
Certified by
Prof. Jack B. Howard
Thesis Supervisor
Certified by
Dr. William A. Peters
Thesis Supervisor
Accepted by
Prof. Robert G). Armstrong
Chairman, Departmental Graduate Committee
AASSAG4USETTS liSTI-ilnE
NOV 14 1988
LIBRARIES
xyehlveS
PYROLYSIS OF DIFFERENT COAL TYPES
by
Glen H.
Ko
Submitted
to
the
Department
of
Chemical
Engineering
at
the
Massachusetts Institute of Technology in August
1988, in partial
fulfillment of the requirements for the degree of Doctor of Philosophy
in Chemical Engineering.
Abstract
Coal-type effects on rapid pyrolysis behavior were investigated.
The experimental phase of this study examined the pyrolysis behavior of
six (6)
selected coals ranging from lignites
to low-volatile bituminous
coals, and under conditions where mass transport resistances are small
(1 atm and ~C100 pm particle dia.).
To experimentally establish coaltype effects, time-resolved product evolution measurements were made
for tars,
gases and total
volatiles
using a constant heating rate of
1000 C/s up to a maximum temperature of 1050 C.
The modeling phase of
this work derived kinetic information from the experimental data, using
the multiple independent parallel reaction (MIPR) and extended MIPR
models, and attempted to relate the kinetic information to measurable
properties of the coal.
Among the six coals studied, higher rank coals generally produced
tars at higher temperatures, and over a narrower range of temperatures.
Consequently, a larger mean and a narrower distribution of global
activation energies were obtained using the MIPR model for coals of
increasing rank.
Predicted tar yields from the extended MIPR model
agreed well with experimental values for a wide range of coal types
(lignites to low-volatile bituminous coal; non-softening and softening)
and pressures (0.001-10 atm).
The parameter values used to make
predictions are within the range of expected values.
A quantitative
correlation, developed to independently relate tar yield limits to coal
type and pressure, was tested against a large set of experimental data
representing a wide range of coals and pressures.
Good agreement
between the predicted and experimental yields were obtained for all
coals and pressures, with a standard error of estimate of ±3wt% dmmf.
In general, no discernable coal-type effects on the apparent rate
of gas production was observed.
Consequently, kinetic parameters of
the MIPR model for measured gas species were only slightly
affected by
coal type. Higher rank coals generally produced less carbon oxides and
pyrolytic water, but more methane.
The ethylene and ethane yields are
small and their absolute yields were less affected by coal type.
Total volatiles
evolve at higher temperatures and over a narrower
range of temperatures for higher rank coals.
Thus as expected, a
larger mean and a narrower distribution of global activation energies
were obtained using the MIPR model for coals of increasing rank.
The
total volatiles yield limit is fairly constant among the lignites,
2
subbituminous and high-volatile bituminous coals, but is significantly
less for low-volatile bituminous coal.
The high-volatile bituminous
coals produced significantly more reactive volatiles (total volatilesH 2 0 and CO 2 ) than other coals.
Thesis supervisors:
Professor Jack B. Howard, Department of Chemical Engineering
Dr. William A. Peters, Energy Laboratory
3
Acknowledgements
My
supervisors,
Prof.
J.B.
4
Howard
and
Dr.
W.A.
Peters,
have
generously provided many hours of valuable discussions and suggestions.
Their
always
careful
guidance
and
constant
encouragement
during
this
Profs.
J.M.
work are deeply appreciated.
Helpful
comments
from my
thesis
committee
members,
Be6r, J.P. Longwell and A.F. Sarofim, are gratefully acknowledged.
Several
M.I.T.
undergraduate
students
have
made
valuable
contributions in this work: Thomas Kronenberger - coal structure study,
Steve
Beaudoin
modeling,
-
experimental
and Debbie
Sanchez
set-up,
-
Jefferey
experimental
runs.
-
Hwang
computer
Special
thanks
to
Debbie for the help during the many months of collecting data.
Sharing
ideas,
friendship,
and many
other
things with
colleagues
and friends met during my stay at M.I.T. have been a very valuable part
of
my
learning
S.Choi,
experience
G.Darivakas,
here
Dr.W.Fong,
-
B.Barat,
T.Griffin,
Dr.M.Boroson,
A.Bouhris,
Dr.M.Hajaligol,
P.Kwon,
F.Lam, T.McKinnon, D.Mess, Dr.C.Vaughn, P.Utthoff, H.Yoon, and J.Xiaothanks.
The
financial
support provided
by
the
1967
Natural
Sciences
and
Engineering Research Council Scholarship, and by the U.S. Department of
Energy
under
Contract
No.
DE-RA21-85MC-22049
are
gratefully
acknowledged.
My family and family-in-law-to-be
have given me a
lot
My fianc6e, Susie, has been very understanding and patient.
all your love, help, and encouragement.
of support.
Thanks for
Table of Contents
List of Figures
5
.................................................
7
List of Tables ..................................................
12
1. Summary
1.1. Introduction ............................................
14
1.2. Experimental studies ......................................18
1.3. Modeling studies
1.4 . Conclusions
......................................... 35
..............................................
1.5. References for summary
61
...................................
62
2 . Introduction ...................................................
65
3. Background
3.1. Coal characteristics .......................................68
3.1.1. Chemical structure ................................. 71
3.1.2. Physical structure ................................. 76
3.2. Reaction chemistry and mass transport ..................... 77
3.2.1. Reaction chemistry ................................ 77
3.2.2. Mass transport ....................................
85
3.3. Experimental studies .......................................90
3.3.1. Effect of coal type ............................... 90
3.3.2. Effect of pressure ................................. 96
3.3.3. Effect of particle size ........................... 100
3.3.4. Effect of temperature-time history
................ 100
3.4. Modeling studies ......................................... 104
3.4.1. Global models ..................................... 104
3.4.2. Detailed chemistry models ......................... 107
3.4.3. Models with explicit description of
mass transport ................... .................. 112
4. Experimental
6
4 .1. Coal selection ........................................... 117
4.2. Experimental apparatus .................................... 118
4.3. Experimental conditions .................................. 121
4.4. Experimental procedures .................................. 124
5.
Experimental results and discussion
5.1. Coal-type effects on tar production ...................... 128
5.1.1. Observed rate of tar production ................... 128
5.1.2. Tar yield limit ................................... 133
5.2. Coal-type effects on gas production ...................... 144
5.3. Coal-type effects on total volatiles production .......... 154
5.4. Pressure effects ......................................... 156
5.5. Experimental uncertainties ............................... 161
6. Modeling results and discussion
6.1. Multiple independent parallel reaction model ............. 163
6.1.1. Mathematical description ........................... 163
6.1.2. Results and discussion ............................ 164
6.2. Extended multiple independent parallel reaction model
....199
6.2.1. Mathematical formulation ........................... 199
6.2.1. Results and discussion ............................ 217
7. Conclusions and recommendations ................................ 228
8 . References ..................................................... 232
9 . Appendix ....................................................... 240
7
List of Figures
Figures
page
1.2-1
The reactor assembly, product collection, and data
acquisition system ........................................... 16
1.2-2
Details of the electrical screen-heater reactor
1.2-3
Tar collection assembly ...................................... 17
1.2-4
Experimental yields of pyrolysis tar versus peak
temperature ...............................................
...........
17
20
1.2-5
Characteristic yield temperature for tar production versus
elemental carbon content .................................... 20
1.2-6
Correlation of tar yields at different pressures with
28
XTAR.......................................................
1.2-7
Comparison of experimental tar yields ....................... 28
1.2-8
Comparison of the yield limit of gaseous products versus
the elemental carbon content ................................ 29
1.2-9
Comparison of characteristic yield temperatures versus the
elemental carbon content (gasous products) .................
31
1.2-10 Comparison of total and reactive volatiles yield limit
versus the elemental carbon content ......................... 34
1.2-11 Comparison of characteristic yield temperatures for total
volatiles production ......................................
34
1.3-1
Hypothetical molecular structure of coal ...................
37
1.3-2
Chemical and physical mechanism of tar formation
..........
37
1.3-3
Tar yields versus peak and holding temperatures ............
43
1.3-4
Best-fitted values of (a) E0 and (b) a for predicting
atmospheric tar production using the MIPR model versus the
elemental carbon content of the coal ........................ 44
1.3-5
Methane yields versus peak and holding temperatures
1.3-6
Ethylene yields versus peak and holding temperatures
1.3-7
Ethane yields versus peak and holding temperatures
1.3-8
.......
46
......
47
........
48
Carbon monoxide yields versus peak and holding
temperatures ..............................................
49
1.3-9
Carbon dioxide yields versus peak and holding
temperatures ................................................. 50
1.3-10 Best-fitted values of (a) E0 and (b) a for predicting gas
evolution rates versus elemental carbon contents of the
coal ................................-........................ 51
1.3-11 Total volatiles yield versus peak and holding
temperatures ... .............................................
52
1.3-12 Best-fitted values of (a) E0 and (b) a for predicting
atmospheric pressure total volatiles evolution using the
MIPR model versus the elemental carbon contents of the
coal . ......................................................
53
1.3-13 Tar yields versus peak temperatures ......................... 55
1.3-14 Ep avg and E 0 ~n versus temperature for (a) a non-softening
coal and (b) a softening coal ............................... 56
1.3-15 Best-fitted values of (a) E05 and (b) us for predicting tar
evolution using the extended MIPR model versus the
elemental carbon content of coal ............................ 58
1.3-16 Best-fitted values of transport parameters for predicting
tar evolution using the extended MIPR model versus the
elemental carbon content of the coal ........................ 60
3.1-1
Hypothetical coal structure ................................. 72
3.1-2
Aromatic carbon, aliphatic carbon, and etheric carbon versus
elemental carbon content ................................... 72
3.2-1
Diffusivity versus pore size for regimes of configurational,
Knudsen, and regular diffusion ............................... 86
3.3-1
Comparison of product yields and distributions from different
coal types: (a) vacuum, (b) 1 and 69 atm .................... 92
3.3-2
Comparison of experimental and predicted pyrolysis liquid
results .. .................................................... 94
3.3-3
Comparison of calculated weight losses with experimental
results [Kobayashi et al. (1977)] ........................... 94
3.3-4
Effect of pressure on product yields from lignite pyrolyzed
different peak temperatures ................................. 98
3.3-5
Effect of pressure on yield of tar from Pittsburgh Seam
bituminous coal pyrolyzed at different peak temperatures ...
3.3-6
99
Total volatiles yield versus particle size for a German
lignite .. .................................................... 99
8
3.4-1
9
Simulated and experimental (a) weight loss and tar yield,
and (b) hydrocarbon gases from the pyrolysis of a
bituminous coal ............................................
110
4.2-1
The reactor assembly, product collection, and data acquisition
sy stem ..................................................... 122
4.3-1
Details of the electrical screen-heater reactor ............
125
4.3-2
New tar collectors in electrical screen-heater reactor .....
125
5.1-1
Experimental yields of pyrolysis tar versus: (a) peak
temperature and (b) holding temperature for the six coals
selected in this study .......................................129
5.1-2
Characteristic yield temperatures for atmospheric tar production
versus elemental carbon content for the six coals studied .. 131
5.1-3
Estimates of the structural quantities in Eq.(5.1-1) .......
5.1-4
Correlation of tar yields at different pressures with
5.1-5
Comparison of experimental tar yields with those predicted by
Eq.(5.1-7) using the pressure-correlated parameters from
Eqs.(5.1-23)-(5.1-30) ........................................ 142
5.2-1
Comparison of the yield limit of gaseous products versus the
elemental carbon content at ambient pressure: (a) hydrocarbons;
(b) carbon oxides and pyrolytic water ....................... 145
5.2-2
Comparison of methane production rate at 1 atm. (a) combined
plot of yields versus peak and holding temperatures; (b)
characteristic yield temperatures versus the elemental carbon
content .................................................... 147
5.2-3
Comparison of ethylene production rate at 1 atm. (a) combined
plot of yields versus peak and holding temperatures; (b)
characteristic yield temperatures versus the elemental carbon
content .................................................... 148
5.2-4
Comparison of ethane production rate at 1 atm. (a) combined
plot of yields versus peak and holding temperatures; (b)
characteristic yield temperatures versus the elemental carbon
content .................................................... 149
5.2-5
Comparison of carbon monoxide production rate at 1 atm. (a)
combined plot of yields versus peak and holding temperatures;
(b) characteristic yield temperatures versus the elemental
carbon content ...............................................150
5.2-6
Comparison of carbon dioxide production rate at 1 atm. (a)
combined plot of yields versus peak and holding temperatures;
(b) characteristic yield temperatures versus the elemental
carbon content ...............................................151
XTAR
138
.
140
10
5.3-1
Comparison of total and reactive volatiles yield limit versus
the elemental carbon content ................................ 155
5.3-2
Comparison of characteristic yield temperatures for total
volatiles production at 1 atm .............................. 155
5.4-1
Effect of pressure on tar yield limit for different coals
5.4-2
Decrease in the tar yield limit relative to the 'vacuum'
yield. (a) coals from this study; (b) coals from Suuberg
..
157
(1977 ) ............... .......... ......... ................... 157
5.4-3
Effect of pressure on total volatiles yield limits for
different coals ..............................................159
5.4-4
Effect of pressure on gas yield limits for Montana lignite and
Pittsburgh Seam bituminous coal ............................. 160
6.1-1
Tar yields versus peak and holding temperatures
6.1-2
Best-fitted values of E0 and a for predicting atmospheric tar
evolution using the MIPR model versus the elemental carbon
content of the coal ..........................................169
6.1-3
Methane yields versus peak and holding temperatures
6.1-4
Ethylene yields versus peak and holding temperatures
6.1-5
Ethane yields versus peak and holding temperatures
6.1-6
Carbon monoxide yields versus peak and holding temperatures
182
6.1-7
Carbon dioxide versus peak and holding temperatures
185
6.1-8
Best-fitted values of (a) E. and (b) a for predicting
atmospheric pressure gas evolution using the MIPR model
versus carbon contents of the coal .......................... 188
6.1-9
Total volatiles yield versus peak and holding temperatures
............
165
........
173
.......
176
.........
179
........
. 192
6.1-10 Best-fitted values of (a) E0 and (b) a for predicting
atmospheric pressure total volatiles evolution using the
MIPR model versus carbon contents of the coal ..............
...........
197
201
6.2-1
Chemical and physical mechanism of tar formation
6.2-2
Comparison of the relative time scales for external and
internal transport rates of tar ............................. 201
6.2-3
Tar yields versus peak temperatures ......................... 218
6.2-4
(a) Ep avg and Ecn, versus temperature for a non-softening
versus temperature for a softening
coal; (b) Epavg and EC,,
coal ......................................................
223
6.2-5
Best-fitted
values of E0 , and a, for predicting tar
evolution
using the extended MIPR model versus the elemental carbon
content of the coal .........................................225
6.2-6
Best-fitted values of transport parameters for predicting tar
evolution using the extended MIPR model versus the elemental
carbon content of the coal ................................. 226
11
12
List of Tables
pages
Tables
1.2-1
Ultimate and proximate analysis of the six selected coals
in this study ..............................................
15
1.2-2
Summary of experimental conditions employed .................
15
1.2-3
Characteristics of coals and experimental tar yields used
in the tar yield limit correlation ........................... 23
1.2-4
Equations to compute XTAR
1.2-5
Best-fit parameters of pressure dependent coefficients a
.................................
25
and # .....................................................
27
1.3-1
Model parameters for the extended MIPR model ...............
41
3.1-1
A.S.T.M. classification of coals by rank .................... 69
3.1-2
Approximate values of some coal properties in different
rank ranges ...............................................
70
3.1-3
Aromaticity measurement techniques .......................... 74
3.1-4
Initial pore-size distributions for various ranks of coals.
3.3-1
Effect of pressure on pyrolysis product yields from Montana
.97
lignite and Pittsburgh Seam bituminous coal ................
3.3-2
Effect of particle size on pyrolysis product yields from
Pittsburgh Seam bituminous coal ...........................
101
....................
108
3.4-1
Elementary reactions of coal pyrolysis
3.4-2
Values of kinetic parameters used by Gavalas et al.
in their detailed chemistry model of coal pyrolysis
4.1-1
Ultimate and proximate analysis of the six selected coals
78
(1981b)
....... 110
of this study ............................................. 119
4.1-2
Ultimate and proximate analysis of the coals investigated
by Suuberg (1977) ...........................................120
4.1-3
Summary of experimental conditions employed in this study . 123
5.1-1
Characteristics of coals and experimental tar yields used in
the tar yield limit correlation ............................. 143
6.1-1
Best-fitted values of E0 and a of the MIPR model for tar
production ................................................
6.1-2
Best-fitted values of E0 and a of the MIPR model for gas
171
evolution: (a) methane, (b) ethylene, (c) ethane, (d) carbon
monoxide, (e) carbon dioxide ..............................189
6.1-3
6.2-1
6.2-2
Best-fitted values of E0 and a of the MIPR model for total
volatiles evolution ....................................--.
198
Equations and physical properties used to compute the
relative transport time scales in Fig.6.2-2 ...............
208
Model parameters for the extended MIPR model ...............221
13
14
1. Summary
1.1. Introduction
Investigation of the pyrolysis behavior of different coal types is
as
different
coal
properties,
all
vast
the
important
types
stages of almost
involves
pyrolysis
Coal
liquefaction.
decomposition
thermal
complex
and
gasification
combustion,
including
processes,
physical
and
chemical
varying
and since pyrolysis occurs during initial
conversion
coal
widely
with
many
of
consist
U.S.
the
in
reserves
coal
reactions coupled with multicomponent mass transport in a molten liquid
solid depending
or porous
on whether
softening
a
is
the coal
type or
not.
a
requires
observed
Quantitative
behavior.
a
for
measurements
and a
effects,
wide
range
pyrolysis
on
type
coal
model to explain
mathematical
needed
are
types
coal
the
evolution
product
time-resolved
of
behavior
the kinds and
to determine
data base
experimental
reliable
of coal-type
extents
of
influence
the
Understanding
to
experimentally establish coal-type effects, but such data are currently
lacking.
In
response,
the pyrolysis behavior
very
the
of six coals
high-rank bituminous
observed behavior
phase
of
experimental
coals.
of
evolution
under
relatively
transport
are
explicitly
includes
The
conditions
unimportant,
approximate
relate
model
the kinetics
the
effects
whereas
the
descriptions
the
experimental
reaction (MIPR)
where
to
The modeling
from the
former model describes
examines
lignites
to
the coal.
this work derives kinetic information
the extended MIPR model.
made
are
Attempts
to measurable properties
study
from low-rank
ranging
data using the multiple independent parallel
product
of this
phase
of
and
of
physical
latter
model
of transport effects,
and
Table 1.2-1: Ultimate and proximate analysis of the six selected
coals in this studya
coal
coal-rankb
Lower
Wilcox
L
Beulah
Zap
L
56.0
4.2
1.1
0.7
19.9
20.3
60.2
4.0
1.0
1.1
21.6
15.0
HVB
#6
HVB
Lower
Kittanning
LVB
74.9
5.0
1.4
0.8
13.7
4.5
67.4
4.4
1.3
3.9
8.7
15.6
82.5
4.5
1.3
1.2
2.4
8.9
Smith
Roland
SB
Blue
62.0
4.6
1.0
1.1
19.5
13.0
Illinois
Ultimate
analysis
wt%, dry
C
H
N
S
0
ash
Proximate
analysis
wt%,dry
moisturec
volatile
matter
fixed
carbon
ash
3.0
45.3
3.0
42.0
3.0
45.2
4.0
43.3
4.0
35.7
1.0
16.3
34.4
43.0
41.8
52.2
48.7
74.8
20.3
15.0
13.0
4.5
15.6
8.9
a
analyzed by Huffman Laboratories,
b
L = lignite, SB = subbituminous, HVB
Inc.
=
high-volatile bituminous,
LVB = low-volatile bituminous.
C partially vacuum dried.
Table 1.2-2: Summary of experimental conditions employed in this
study
reactor
variables:
varied (v)
or fixed (f)
range
covered
coal type
v
lignites to
low-volatile
bituminous coals,
elemental carbon
content ranges
72-92 wt% dmmf.
temperature-time
history
f
pressure
1000 C/s heat-up,
200-1000 C/s cooldown, 1050 C max.
temperature.
v
10-3
10 atm
particle
size
f
75-90
pm dia.
15
16
TEMPERATURE-11ME HISTORY
REACTOR
(He)
GAS
CHROMATOGRAPH
VACUUM
Figure 1.2-1
system.
The reactor assembly, product collection, and data acquisition
17
REACTOR
SAMPLE
THERMOCOUPLE
Figure 1.2-2
Details of the electrical screen-heater reactor.
SCREEN
ELECTRODE
0
0
0
0
TAR COLLECTION
ASSEMBLY
GLASS FUNNEL
FILTER DISC
Figqure
1.2-3
-
Tar collection assembly
18
thus is applicable over a wider range of operating conditions.
1.2. Experimental studies
1.2.1. Experimental procedures
for this
The six chosen coals
TX (lignite A);
A); Lower Wilcox,
and Lower
A);
bituminous
(low-volatile bituminous).
PA
analysis of the selected
Table 1.2-1 gives the ultimate and proximate
experimental
The
coals.
are
study
this
in
employed
conditions
B);
IL (high-volatile
Illinois #6,
Kittanning,
(lignite
(subbituminous
WY
Smith Roland,
(high-volatile bituminous C);
NM
Blue,
ND
Beulah Zap,
study are:
summarized in Table 1.2-2.
type reactor
An electrically heated screen-heater
(Fig.l.2-1) was
used to measure the apparent evolution kinetics and the yield limit of
extensively
past
in
used
the
for
products
volatile
Suuberg,
1977; Fong,
kinetic
studies
1986),
including
pyrolysis
as it
reactor
This
six coals.
type
(Anthony et
studies
temperature
1974;
al.,
important in
offers many advantages
reliable
has been
of
measurement
the
sample over a wide range of heating rates, rapid quenching and dilution
of
ability
to
work over
(10-3
a wide
to
1
pressure
run
particles
spread thinly in
400
mesh
stainless
leaving
upon
products
volatile
steel
particles,
pressure
runs
smaller
the central
screen
(10
pressure reactor.
sample
atm) since
20
about
atm),
mg
smaller
and
In
a typical
low
75-90
of
pm
diameter
region of 10 cm x 5 cm,
(Fig.l.2-2)
sizes
surface,
particle
of pressures.
range
controlled temperature-time history.
coal
coal
the
To
(~5
mg)
screens
pyrolyzed under a
are
ensure
folded
thin well
dispersed
had to be used in
are
high
used in the high
The sample temperature is measured using a very thin
Chromel-Alumel
thermocouple
The reactor gas,
screen near the coal particles.
remains
(99.999%),
and quenching
foil) placed within the folded 19
in.
(0.0005
ultra high purity He
and provides rapid dilution
near room temperature,
of volatiles as soon as they are evolved from the coal
thus presenting minimal opportunity for further reactions
surface,
of
volatiles outside the particle.
light hydrocarbon gases,
Tars,
as
condense
in
the
sum
products
volatile
all
of
and water are
the
Tars are operationally
from coal pyrolysis.
major volatile products
defined
carbon oxides,
water)
(except
that
and were collected using
the reactor at room temperature,
The gas yields were measured
the tar trap assembly shown in Fig.1.2-3.
using a Perkin Elmer Sigma 2B Gas Chromatograph equipped with thermal
conductivity and flame ionization detectors.
1.2.2. Experimental results and discussion
Coal-type
atmospheric pressure
studied.
coals
respectively
1000
Figure
on tar production.
effects
for
tar yield versus peak temperatures
Heating
and
200-1000
and
cooling
C/s
in
rates
with
these
holding
no
shows
1.2-4
points
in
Fig.l.2-4
were
were
at
peak
time
hand-drawn
the six
runs
temperatures, and with a maximum peak temperature of 1050 C.
data
the
to
The lines
indicate
through
the
trends.
Individual plots with model predictions are given in Fig.l.3-
3.
coal
Qualitatively, the
type
figure shows that there
on both the apparent
is a clear effect of
rate of tar production
and the yield
limit, defined as the asymptotic yield at high peak temperatures (> 800
C).
Low-rank coals
(ZP,LW,SR) tend
to
initiate
and
achieve
given
extents of tar production at lower temperatures compared to higher rank
20
30
28
26
24
22
20
0
18
16
0
-J
U
5:
I-
14
12
10
8
6
4
300
500
700
900
1 100
TEMPERATURE (C)
Experimental yields of pyrolysis tar versus peak temperature
Figure 1.2-4
for the six selected coals in this study. Carbon: LW < ZP < SR < BL < IL <
LK. Abbreviations: LW = Lower Wilcox lignite, ZP -= Beulah Zap lignite, SR =
Smith Roland subbit., BL - Blue high-volatile bit., IL = Illinois highvolatile bit., LK = Lower Kittanning low-volatile bit.
760 740 720 700 660 660 U
TU
D:
640 620 600 580 560 -
0
540 520 500 480
460
440
70
74
78
0
82
86
90
94
ELEMENTALCARBONCONTENT (WT% DMMF)
T50
A T25
V
T75
production versus
Characteristic yield temperatures for tar
Figure 1.2-5
the
studied (Tx denotes
elemental carbon content for the six coals
temperature at which the yield reaches x% of the maximum yield). Carbon: LW
< ZP < SR < BL < IL < LK.
Abbreviations: see Fig.1.2-4.
These points 21
coals (BL,IL,LK); abbreviations are defined in Fig.1.2-4.
are reinforced by quantitative observations on the apparent rate of tar
which compares
Fig.l.2-5,
in
production presented
50% (T50),
which the tar yield reaches 25% (T25),
determined
The
3).
approximate
of the
temperatures
characteristic
three
The
dmmf.
in wt%
contents
and 75% (T75)
at
the six coals represented by their elemental carbon
limit for
yield
the temperatures
were
from the tar data fitted with the MIPR model (see Fig.1.3difference
and
T75
between
(T75-T25)
T25
represents
an
whereas TSO roughly corresponds
spread of the yield curve,
to the temperature at which the observed tar evolution rate is maximum.
Comparing
T50
shows
a
almost
represented by the elemental
coal
with
increase
monotonic
rank
indicating a
carbon content of the coal,
shift in the yield curve to higher temperatures for higher rank coals.
T50
ranges
from 550
maximum difference
shows
(T75-T25)
less spread in
ranges
C for
ZP
to
685
C
for LK, an increase in the
of about 135 C among the coals studied.
Comparing
for higher rank coals,
indicating
a decreasing trend
the yield curve for higher rank coals.
C for ZP to 85 C for LK,
from 175
The difference
a reduction in
the maximum
difference of about 90 C.
An
description
exact
transport
available.
phenomena
Thus,
of
involved
the
in
complex
tar
reaction
chemistry
and
is
currently
not
production
interpretation of the observed tar evolution rate
behavior for different coal types, depends on the assumed mechanism for
tar formation.
'tar
A frequently assumed mechanism is
precursors'
in
the
coal
via
multiple
the decomposition of
first-order
parallel reactions (Serio, 1984; Ko et al., 1988a)
independent
22
first-order decomposition
Tar precursors in coal -------------------------------transport
+
all
that
assumption
further
description and with the
global
a
such
Under
effects.
transport
physical
by
influenced
are
reaction
decomposition
global
this
for
parameters
model
The
Tar
the
have
coals
same preexponential factor in the Arrhenius rate constant, a higher T50
of
assumptions
description, higher
global
this
to
appear
coals
rank
the
under
Thus,
energies.
activation
apparent
of
distribution
a wider
implies
(T75-T25)
larger
a
Similarly,
energies.
activation
apparent
with greater
from reactions
produced
are
tars
that
implies
produce tars from reactions with apparent activation energies that have
a higher mean but a narrower distribution.
quantitatively
coal
capability
of
data
experimental
wide
range
coal.
(37
data
minimal
Table
represent
influence
the
=
predictive
large
a
to
gives
of tar
100-1500
secondary
the
C/s,
reactions
particle (small sample mass and particle sizes).
set
a
the elemental
for
generated
max.
of
anthracites)
specified pressure
maximum amount
(heating rate
from
a
1.2-3
of
representing
from lignites
ranging
90 atm).
against
literature,
and the
tar yield under
rapid devolatilization
with
study
coals,
to
('vacuum'
and measured
The
this
from
to
The
1988b).
tested
is
approach
properties
to measurable
1987,
al.,
correlation
new
of coals
and pressures
analysis
the
et
(Ko
below
given
is
tar yield
the
relate
new
A
to
enough
not
is
alone
trend.
observed
the
explain
quantitatively
study and from the literature
this
information
coal-rank
that
indicate
from
limit data
yield
tar
The
each
during
T ~ 1000 C)
outside
the
coal
Table 1.2-3
Characteristics of
the tar yield limit correlation.
coals
and experimental
tar yields
used in
Tar yield (wt% domf; symbols used in Figs.
at pressures (MPa) of
Elemental Analysis
(wt
dmmf)
Coala
Freihaut and
Montana L
Wlyodak SB 1
Seery (1981)
Freihaut
et al. (1982)
Loison and
Chauvin (1964)
Colorado B
Pittsburgh B
Faulquemont B
Wendel III B
Emma B
Bergrannsqluck B
Oignies
Mlaigre
B
Flenus do Bruay B
Pittsburgh B
Prosper II B
Schlaegel U. Eisen B
B
Oh (1985)
Arendt and
van Meek
(1981)
Uulfen
Suuberg
(1977)
Leopold P
Pittsburgh B
Nontana L
This Study
Iyodak
Sesser
Suuberg
et al. (1985)
Bautista
(1984)
Socc
4.6
4.9
5.2
5.5
5.5
5.4
4.6
2.6
5.5
5.4
5.1
5.3
5.0
0.7
0.4
0.6
0.8
0.6
4.7
0.5
82.2
91.5
90.1
25.5
18.1
17.0
13.9
11.2
9.4
8.2
1.9
11.2
9.4
12.8
8.3
6.4
5.4
5.2
4.4
6.4
10.0
2.7
3.7
4.6
3.8
5.1
5.9
0.7
0.4
0.5
1.9
0.7
0.5
87.6
5.7
Colstrip L
Lower Wilcox L
Illinois B
Blue SB
Beulah Zap L
Smith Roland L
Pocahantas B
North Dakota L
Illinois B
Bruceton B
North Dakota L
Bruceton B
Pittsburgh B
88.4
88.5
89.0
91.9
86.7
84.2
81.0
72.2
73.1
82.9
76.8
72.0
83.2
79.1
S1
SB
Lower Kittanning
Suuberg
et al. (1987)
I
68.3
75.4
75.5
78.2
81.0
82.0
85.0
93.7
81.0
82.0
80.8
86.1
Wyodak SB 2
Utah U)
Colorado B
Pittsburgh B
AlabAma B
Anthricite
Lens-Lievin B
Cosway (1981)
Reitzen (1978)
0b
S
Investigator
B
72.4
91.9
72.7
91.3
74.5
78.6
85.1
75.4
85.1
84.7
8.4
4.4
4.3
4.7
19.8
5.7
5.7
4.6
6.1
10.3
17.4
5.4
4.9
20.9
5.6
5.4
5.3
9.7
22.0
9.8
14.1
21.6
1.7
20.9
4.1
4.8
5.0
7.6
5.3
4.6
4.1
5.4
5.6
4.1
5.6
7.9
5.8
20.5
14.6
7.6
19.1
0.6
1.9
0.5
0.7
0.6
1.9
0.5
0.3
0.01
0.4
0.6
0.6
0.5
9
A19.8df
A28.ld
19.2f
A26.0
'120.0
124.3
20. 5
17.6
15.1
2.1
h39.8
L29.3
* 9.9
.17.7
.26.5
.28.4
V37.0
y
8.4
U16.8
030.1
027.7
0 9.1
*14.0
014.8
0.2
0.5
2.0
0.5
0.5
0.5
0.7
6.9
A20.0
A21 .0
A27.0
A26.0Of
A39,0
A25.0
A 2.0
0.4
0.4
0.5
0.4
2.0
1
0.1
A18.0f
0.6
0.9
1.9
0.8
5.1-4 and 5.1-5)
A 6.7
A37.7
$38.6
.
25.7
E
26.5
V13 8
f 3.2
.6.5
19. 3
21. 5
11.5
13.1
9.9
024.8
021.2
815.0
814.5
4.1
Mu 6. 5h
11i.1
030.0
C9.7( 0.2)9
07.2
[110. 7
S12.9
9.9
5.6
22.8
26.5
3.7
7.5
14.1
15.2
6.6
12.2
19. 3
19.8
25.2(0.7)
23.7(1.0) d
21.1(1.5)
20.6(2.4)
a B-bituminous; L-lignite; SB-subbituminous
b By Difference
c Estimated as half the total sulphur content when organic sulphur not
reported (Loison and Chauvin, 1964; Arendt and van Heek. 1981: Cosway,
1981; Reitzen, 1978; Suuberg et al., 1985, 1987; Bautista, 1984; this
study).
d Obtained by interpolation between 0.0007 and 0.013 MPa in Freihaut
et al. (1982), and between 0.7 and 1.5 MPa in Bautista (1984).
e The tar yield (6.5 wtS demf) reported for Sesser SB seemed low and
fasssubstituted by the 21.5 wt% dmmf measured in this study.
and Montana L from Freihaut and Seery (1981), and Freihaut
Colorado
et al. (1982) were not used because possible errors in tar yield
measurement are suspected.
g Indicates pressure in MPa.
h This value is slightly lower than the previously reported value (7.2
domf) in Ko at al. (1988b)
B
wt
N)
24
Formulation of correlation: treatment of coal-type effects.
Tar is assumed
(1) Chemical and physical mechanism of tar production.
to
be
van
by
suggested
first
mechanism
global
the
via
generated
Krevelen (1961):
[2] transport
[1] thermolysis
Coal --------------------- Metaplast -----------------
Tar
of bridges
(2)
identities
The
structures.
chemical
Important
and numbers
of
bridges between aromatic clusters of the coal and the concentration of
hydrogen
available
scission reactions
structural chemical
important
are
the transport process,
easily
not
in
features
the assumed mechanism,
are
correlated
via
effect
transport
the
identifiable,
[2]
in tar
factors
Since the structural
generation without transport effects.
important in
created by bridge
radicals
the free
to stabilize
is
empirical parameters obtained from best-fit analyses of existing data.
(3) Formulation of coal-specific parameter.
XTAR,
A coal-specific parameter,
proposed to correlate tar yields with coal type is
XTAR = (no. of labile bridges)(amt. of abstractable hydrogen)/
(1.2-1)
(no. of cross-linked bridges)
(4) Estimation of identified structures.
are
structures
estimates
were
information.
quantities in
unavailable
generally
made
Table
for
each
1.2-4
quantity
gives
Since the necessary molecular
for
coals,
most
based on currently
to
procedures
estimate
reasonable
available
three
the
XTAR.
Formulation of correlation: treatment of pressure effects.
Tar yield limit at a given pressure is linearly correlated with the
coal-type parameter derived above:
Tar yield limit (wt% dmmf)
The
pressure
dependent
=
a(P)
coefficients
(1.2-2)
+ P(P)XTAR
a
and
#
are
obtained by
best
Table 1.2-4
Equations to compute XTAR
e number of labile bridges
assumption:
=
((1-fa)[C]/12)1.-8
Labile bridges are only aliphatic, and their concentration is
assumed to be proportional to the aliphatic carbon content of
of the raw coal. The fraction (1-fa) also contains
contributions from carboxyl, carbonyl and ether carbons, but
these are assumed to be small. The exponent 1.8 is a best-fit
parameter obtained by applying multivariable fitting routines
to obtain the best correlation between tar yields and XTAR-
* number of cross-linked
bridges
[0]/16 + [So]/32.0 6 6
if [0] > 3.5 wt% dmmf
3.5/16 + [So]/32.0 6 6
if [0] s 3.5 wt% dmmf
Cross-linked bridges consist only of ether and thioether
structures, whose concentration is assumed to be proportional
to the sum of elemental oxygen and organic sulphur contents of
the raw coal. A constant [0] was needed for coals with low
elemental oxygen contents because the number of cross-linked
bridge is highly sensitive to coal elemental oxygen contents
below about 4 wt% dmmf, and uncertainties in oxygen measurement
can easily exceed ±1 wt% dmmf.
assumption:
* amount of
abstractable
hydrogen
=
[H]/l - [OH]/17
Abstractable hydrogen is the hydrogen attached to aliphatic
carbons. Its concentration is proportional to the amount of
elemental hydrogen in the raw coal, minus a slight correction
to account for experimental observations that OH groups may
compete for the abstractable hydrogen (Suuberg, 1977).
assumption:
Notations:
25
[C] = the elemental carbon content (wt% dmmf)
[0] = the elemental oxygen content (wt% dmmf)
[So] = the organic sulphur content (wt% dmmf)
[H] = the elemental hydrogen content (wt% dmmf)
f
= aromaticity
2
0.830526 - 2.008147([C]/100) + 2.241218 ([C]/100)
(polynomial best-fit of fa versus [C] using data from
Gerstein et al., 1982)
[OH] = the hydroxyl group content (wt% dmmf)
= 33.2 - 0.35 [C] (Given, 1976)
=
experimental tar yield data either for a specified pressure or 26
fitting
range
pressure
or for
(Table 1.2-5a),
(Table 1.2-5b).
all pressures
Figure 1.2-6 compares measured maximum tar yields with those predicted
using
coefficients
pressure-specific
the
1.2-5a].
Table
[Eq.(1.2-2),
The predicted yields are within ± 5 wt% dmmf of the observed values for
all
tested
coals
error
standard
pressures
four
the
at
of estimate
and
ranges.
The
2.8 wt% dmmf.
The
pressure
of the prediction was
standard error of estimate was computed using the definition
n
standard error
=
of estimate
L
(Yield,1,exp'l - Yield.i.pre'd)
n-k
I
1 /2
2
J
(1.2-3)
j=1
where n is
fitted
the number of data points
used
parameters
in
the
(j),
and k the number of best-
correlation.
Figure
1.2-7
compares
experimental data for all pressures with predictions obtained using the
parameters
pressure-correlated
[Eq.(1.2-2),
predicted yields are within ±6 wt%
1.2-5b].
Table
dmmf for all
for all pressures between 10 Pa to 9 MPa,
Use of the
coals.
pressure-correlated parameters has the advantage that it
but suffers
The
is
applicable
from a slightly
greater standard error of estimate of 3.1 wt% dmmf.
Coal-type
effects
gas
on
production.
Figure
1.2-8
compares
the
yield limit of gaseous products versus the elemental carbon content for
the six coals investigated in
this study,
and the two coals studied by
Suuberg (1977) under similar but not identical experimental conditions.
Higher
water,
rank coals
but more methane;
0.4-9.9,
ethane
generally produce
2.4-16,
yields
are
less carbon oxides and pyrolytic
the ranges for CO,
1.6-4.3
wt%
dmmf
small
and
their
C0 2 , 1H20,
respectively.
absolute
CH 4
are 0.9-11.0,
The
ethylene
yield values
are
and
less
27
Best-fit parameters of pressure dependent coefficients a and
Table 1.2-5
for use in Eq.(1.2-2)
(a) pressure-specific
coefficients
<115
XTAR
29,<
(a
/ #)
TAR>3
Xyg
> 31
5 s5 Xyg
TAR : 331
15
10-100 Pa
2 / 0
-30.8125 / 2.1825
37 / 0
0.1 MPa
2 / 0
-22.375 / 1.625
28 / 0
1 MPa
2 /0
-16.75 /1.25
22/
2.5-9 MPa
2 / 0
-10.1875 / 0.8125
15 / 0
(b)
0
pressure-correlated coefficients
X
2
a
0
p
L
< 5 115
=
s
XTAR s
31
1/(0.021533 + 0.028651L)
- 36
0.508030 + 0.696487L
- 0.06959LP2
> 31
X TA
11.24071
+ 9.743707L2
0.91326LP2
0
t
-log 10 P + 1
reactor pressure in MPa for P s 2.5 MPa (1 MPa = 10 atm)
P
fixed at 2.5 MPa for reactor pressure above 2.5 MPaa.
a This was justified since pressure has negligible effects on tar yield
above 2.5 MPa. Bautista (1984) observed that tar yield did not decline
with increasing pressure above =2 MPa, and the present work (Fig.l.2-6)
found a close agreement between predictions and data using 2.5 MPa to
represent pressures from 2.5-9 MPa.
#
LL
35
M
M-
V7
30
1 MPG
25
0j
-J
&
0.1 MPG
30
-
0
'0 0.1 MPc
*
+l
a
MPa
2.5-9
0
eB
Agj
0
20
2.5-9 MPa
15
z
10
Li
5
Luj
28
10-100 Pa
10-100 Pa A I k T
LiJ
I
0
I
I
I
5
I
I
I
I
1
10
1
I
I
20
COAL-TYPE PARAMETER,
Figure
1.2-6
Correlation
Symbols: see Table 1.2-3.
of tar
I I
25
I
15
i
I
I
i
i
30
i
i
35
X
yields
at
different pressures
with XTAR.
Lines are from Eq.(1.2-2) and Table 1.2-5a.
40
P (MPa)
-35-
30
H-
~. 5K
25I
A
10-4-10-s5
0.01
Ao
0-1
0.2-0.7
oAo
000
1
M .
Q,4 V
1.5-6.9
9
Ol'e>
20L
A
15 L
H-
10wL
a-
5
Li@
E
e
w
0
5
10
15
20
25
30
.35
40
PREDICTED TAR YIELD (WT % DMMF)
Figure 1.2-7
Comparison of experimental tar yields with those predicted by
Eq.(1.2-2) and Table 1.2-5b. Symbols: see Table 1.2-3.
4.5
a
CH
4.0
29
4
3.5
2.5 2.0 -j
0A
C
+
Ap
1.5 -
C2H
1.0 --
0.5 0.0
70
94
90
86
82
78
74
ELEMENTAL CARBON CONTENT (WT% DMMF)
DN
+6
CH4
C2H6
0 *
C2H4
16 -15
b
V
14
13 -
V
12
I-
11
10
9
LL
8
-'J
7
6
w
5
4
3
V
2
0 -70
I
I
7
74
I
78
I
82
I
I
86
91
90
I
94
ELEMENTAL CARBON CONTENT (WT7 DMMF)
CO
X @ C02
v
H20
AA
Figure 1.2-8
Comparison of the yield limit of gaseous products versus the
elemental carbon content: (a) hydrocarbons; (b) carbon oxides and pyrolytic
water.
Open or non-circled symbols are from this study; closed or circled
symbols are from Suuberg (1977). Carbon: LW < ML < ZP < SR < BL < PB < IL <
LK.
Abbreviations: ML - Montana lignite, PB - Pittsburgh Seam bituminous,
see Fig.l.2-4 for others.
affected by coal type; they range from 0.6 to 1.6 wt% dmmf for ethylene 30
The higher carbon oxides and
and from 0.2 to 0.7 wt% dmmf for ethane.
water
have
yields
been
associated
with
higher
concentrations
of
carboxyl and hydroxyl groups respectively in lower rank coals (Suuberg,
However,
1977).
an exact reaction mechanism is
quantitatively rationalize
the
relationship.
not yet available
to
Methane production has
been postulated to occur via bond dissociation of alkyl groups to yield
methyl radicals,
which upon abstracting hydrogen form methane
et al.,
But applying such a mechanism to explain the observed
1981).
(Gavalas
trend for methane yields is difficult due to the lack of the necessary
quantitative
structural
information,
e.g.,
in
particular
the
concentration of alkyl groups.
Figure 1.2-9 compares the apparent evolution rates of (a)
C2 H 4 ,
C2 HA,
(c)
Each figure
and T75)
(d)
CO,
and
(e)
CO 2
for
the
six coals
respectively)
versus
Comparing T50 shows
the
elemental
data
carbon
(Figs.
(T25, T50,
1.3-5 to
content
of
a slightly increasing trend with
(b)
investigated.
shows three characteristic yield temperatures
obtained from the experimental
CH4 ,
the
1.3-9
coal.
coal rank
for
methane and ethane (Fig.l.2-9a,c), but almost no observable effect for
ethylene and carbon oxides
curve as
for
all
(Fig.l.2-9b,d,e).
indicated by (T75-T25)
gases,
except
appears
for carbon
The spread of the yield
to be unaffected by coal type
dioxide, which
shows
a
decreasing
trend for higher rank coals.
Reasons
for
the
lack
of
observable
coal-type
apparent rate of gas production are currently unclear.
is
that
the kinetics
of gas
(Solomon and Hamblen, 1985).
effects
on
the
One hypothesis
production are unaffected by coal type
Gaseous products
are claimed to evolve
n
-1
BOO
860340-
820800
O0
- - - - - - - - -- - -
780 760
740 720 700
0
8
82
78
74
70
90
94
70
a
0
80
ICV
-
94
90
94
a
O20
1020
1000
M
90
ELEMENTAL
CARBON
CONTENT
(WTXOMMF)
T75
V
723
T30
i
85C0
88
82
78
74
(WT DMMF)
CONTENT
CARBON
ELEMENTAL
T75
7
T25
T50
d
-
960 940 970 900 -
740
2
03
0
aw -
ago w
0
-
a2a -
640-
Soo
700 --
620
-
760 -
600
0
800
780
760
82
78
74
70
86
740
90
4
94
70
(WTODMMF)
ELEMENTAL
CARBON
CONTENT
T75
T25
T50
A
V
74
78
0
82
86
OMMF)
7 T75
ELEMENTAL
CARONCONTENT
(WT7
r50
T25
A
2e'
740
720
700
580
680
620
-
-LMNA
ABNCNET(r MF
7 7
% T5
5
580
-60
540
520
500
70
Figure
1.2-9
72
78
82
Comparison
B
9
94
of characteristic yield
temperatures
versus
elemental carbon content: (a) CHG4 , (b) C2 H4 , (c) C 2 H 6 , (d) CO, (e) CO 2 .
Abbreviations: see Fig.l.2-4.
Carbon: LW < ZP < SR < BL < IL < LK.
the
from decomposition of specific functional groups,
is
assumed to be produced from ether groups in
and thus is
group,
a
such
in
problem
and the
other H 2 0
temperatures
But a
type.
in
the
following
above
750
C, phenol
one of which gives CO and a C5
and benzene
former pathway is
The
(1975a,b)].
type of functional
illustrated
is
picture
simple
decomposes along two parallel pathways,
moiety,
The rate of
asserted to be independent of coal
Upon rapid pyrolysis at
example.
carbon monoxide 32
the coal.
depend only on the
assumed to
is
gas production
e.g.,
and
[Cypres
Bettens
a base-catalyzed reaction,
(1974),
and thus
from minerals
is
expected to be promoted by strong solid base materials
in
the coal such as CaO generated by calcite decomposition (Franklin et
al., 1981).
+
CO
+
H2 0
H
OH
H
H
assuming this mechanism applies for the decomposition of phenolic
Thus,
groups
in coal,
species,
and
the phenol
the
group
concentration
can produce
of
several different
gas
minerals
can
in
base-catalysts
strongly influence the relative extent of the two reaction paths.
An
alternative
plausible
and more
observable coal-type effects in
for
explanation
this study,
is
the
that differences
lack
in
of
the
apparent gas production rates are less than or comparable to scatter in
the data
for
this
caused by experimental uncertainties.
explanation
comes
from a
recent
A supporting evidence
study
of
Burnham
et
al.
(1988),
in
which
lignites
to
low-volatile
bituminous coals were pyrolyzed at low heating rates
(<
1 C/s) under
atmospheric
eight
pressure.
evolution rate
coals
They
is maximum)
ranging
from
observed
that
Tmax
(T
at
generally increases with coal
which
the
rank, with
maximum differences ranging from 18 to 33 C among light hydrocarbons
(CH4 ,C2 H 4 ,C2 H6 ).
Such differences
slow heating apparatus which is
within ±5 C (Burnham et al.,
are more
clearly resolved in the
able to measure the sample temperature
1988).
In rapid heating studies such as
the present one, uncertainties in the temperature measurement are much
higher (~+25
C), and are comparable to the reported differences
caused
by coal-type effects in the low-heating experiment.
Coal-type effects on total volatiles production.
compares the yield limit of total and 'reactive'
Figure 1.2-10
volatiles versus the
elemental carbon content for the six coals investigated in
and the two coals studied by Suuberg (1977).
this study
Reactive volatiles are
defined as total volatiles minus water and carbon dioxide yields.
The
total yield limit ranges from 41 to 55 wt% dmmf among lignites, and
subbituminous and high-volatile
dmmf
for
compare
the
is
bituminous
low-volatile
reactive
coals
bituminous coals,
bituminous
coal.
volatile yields,
which
A
show
but drops to 22 wt%
useful
quantity
to
that high-volatile
(BL,PB,IL) produce significantly more than other coal
types.
Figure
1.2-11
compares
the
characteristic yield
temperatures of
total volatiles production at atmospheric pressure for the six coals.
Plots of the total yield versus
shown in
Fig.l.3-ll.
temperature for individual coals
are
The characteristic temperatures tend to increase
for higher rank coals, indicating a shift in the yield curve to higher
33
34
60
50
40
30
20
10
0
70
0U
74
82
78
TOTAL VOLATILES
86
94
90
ELEMENTALCARBON CONTENT (WT% DMMF)
00
REACTIVEVOLATILES
Figure 1.2-10 Comparison of total and reactive volatiles yield limit versus
Open symbols are from this study; closed
the elemental carbon content.
symbols from Suuberg (1977). Carbon: LW < ML < ZP < SR < BL < PB < IL < LK.
Abbreviations: see Figs. 1.2-4 and 1.2-8.
800 - 780
V
V
760
V
740
tJ
V
720
V
700 a
0
680
Id
ftc
D
660
0
A
640 0
Id
I--
620 600 A
580
A
A
560
540
-
A
520 -
A
500
I
70
-
78
0
Figure
1.2-11
volatiles
I
I
I
74
Comparison
I
I
I
I
86
62
90
94
ELEMENTALCARBON CONTENT (WT7 DMMF)
V
T75
A
T25
T50
of
characteristic
production at 1 atm.
Abbreviations:
I
see Fig.l.2-4.
Carbon:
yield
LW <
temperatures
ZP <
SR < BL <
for
total
IL < LK.
temperatures.
Comparing
the
spread
of
the yield curve,
measured by
(T75-T25), shows a small decreasing trend with increasing rank.
trends
are
consistent with
These
the expected behavior from combining the
observed coal-type effects on the rate of tar and gas production.
a consistency together with
helps
to
verify
a good product mass
the experimentally
balance
observed coal-type
Such
(90-110 %)
effects
on the
apparent rate of product evolution.
1.3. Modeling studies
1.3.1. Model description
The MIPR model has been widely used to describe the evolution rate
of tar
(Serio,
1984;
Ko et al.,
1988a),
gaseous products
Ngan, 1979), and total volatiles (Anthony et al.,
1979;
the
Sprouse and Schuman,
MIPR model
is
1981).
expressed
(Weimer and
1974; Ciuryla et al.,
The rate of volatiles evolution in
as the
sum of the
contributions
from a
large number of first-order independent parallel reactions,
dY/dt = X k i exp(-Ei/RT) (Y* -Yi)
where i
for
all
denotes one reaction.
reactions,
i.e.,
(1.3-1)
The same preexponential factor is used
k0 i = k0 , and the activation
described by a Gaussian distribution
function
f(E)
energies
are
with mean EO
and
standard deviation a
f(E)
[u(2r)1/ 2 ]-1 exp[-(E-E 0 )2 /2, 2 ]
=
(1.3-2)
The probability of finding a reaction with activation energy between E
and E+dE
f(E)
is
given by f(E)dE,
= Y*i/Y* and Y*
input parameters
is
equal
required in
where for a large number of reactions,
to the
sum of the Y*
the model are Y*,
E0 ,
for
all
i.
a, and k0 .
The
The
notation 'Y' here is equivalent to 'V' in earlier descriptions of this
35
model (Anthony et al., 1974; Howard, 1981).
36
The extended MIPR model increases the range of applicability of the
MIPR model by explicitly including descriptions
secondary reactions.
The main objective in formulating this model was
to be able
to describe
conditions
(coal
using as
of mass transport and
tar production over a wide range of operating
type, heating
rate,
few difficult-to-obtain
requiring a minimal computational
pressure,
and
particle
size),
physical
parameters
effort.
The chemistry of the model
as possible and
assumes a hypothetical molecular structure of coal shown in Fig.1.3-1.
Figure 1.3-2 gives a schematic diagram of the proposed mechanism where
the
tar
is
produced
via
the
hydrogenation and transport.
are cross-linking,
sequential
steps
or
a
polymerization,
describes
combination
and
explains
operating variables
size.
This
of
the
scission,
and tar cracking reactions,
these
three
is
reactions uniquely
observed effects
heating rate,
mechanism
all of
As will be shown below,
competing
experimentally
- coal type,
proposed
bridge
Competing with the tar production pathway
which lead to the formation of char + gas.
each
of
pressure,
assumed
in
the
of main
and particle
mathematical
formulation of the extended MIPR model described below.
For non-softening coals,
the rate of tar
(Y)
leaving the particle
of radius R is
dY/dt = XEpi Ec,,
kti(V*
1 -Vi)
(1.3-3)
where
dVi/dt
EPi
kti(V*i-Vi)
=
(1.3-4)
rate of scission
=
rate of scission + polymerization
=
k
/
(k8
+kP)
(1.3-5)
37
X
-
(B-PA-B)n-
y
PAC = represents repeating nuclear units of polyaromatic and
hydroaromatic clusters
B = bridging molecules
X = side groups suspected to be responsible for cross-linking
Y = non-cross-linking side groups
n = number of repeating units
Figure
1.3-1
Hypothetical molecular
formulating the extended MIPR model.
COAL
NON-X-LINKED
COAL
3
2
structure
PRIMARY TAR
of
4
.
coal
SECONDARY TAR
5
1/
CHAR
CHAR
CHAR
GAS
GAS
GAS
1 = X-UNKiNG
Figure 1.3-2
2 = SCISSION, HYDROGENATION
3 = POLYMERIZATION
assumed
4 = TRANSPORT
5 = CRACKING
Chemical and physical mechanism of tar formation.
in
Ec,,n=
2exp(-m.R)/[l+exp(-2m.,R)]
=
For a
x-linked
component i,
Eq.(1.3-4)
fraction
coal
of
(1.3-6)
describes
(V*j)
in
a
the rate at which the non-
reacts,
cumulative amount of the reacted material.
reaction
38
rate of tar production with transport limitation
rate without transport limitation
where
Vi
represents
the
The subscript i denotes one
multiple independent parallel
reaction scheme,
in
which
each reaction describes the thermal scission of a bridge bond with its
specific
chemical
scission
reactions
E0 5
The
are described by a
activation energies
for
Gaussian distribution
these
with mean
(1.3-5) and (1.3-6) represent the fraction of the reacted
[Eq.(1.3-4)]
between 0
and 1,
production
which survives polymerization and cracking reactions
The
respectively.
rate
strength.
and standard deviation a,.
Equations
coal
bond
values
where
of each
0 represents
and 1 represents
constants,
k, ,
cracking reactions
of
these
are
respectively,
quantities
are
bound
the most severe limitation on tar
no limitation.
kg , kc
two
for
The
first-order Arrhenius
scission,
and kt,
=
ks,,
polymerization,
+ kP
tar
The kinetics of
.
cross-linking are not considered here since this process
is assumed to
occur at relatively mild temperatures before other reactions proceed to
any appreciable extent.
is
the
Thiele
(kc/Dgeff)1
of tar.
2
The dimensionless
modulus
for
and D,,eff is
quantity mnR in
non-softening
the effective
The transport description
coals,
Eq.
(1.3-6)
n,
where
=
gas phase binary diffusivity
assumed in
deriving Ecns
considers
steady-state transport in macropores where the tar enters the pore from
the
center of the particle, and neglects external transport resistance
and
convective
contributions.
A
characteristic
time
analysis
supports
the
transport
external
negligible
and
steady-state
resistance 39
The assumption that tars enter the pore from the center
assumptions.
not strictly valid, but making this
is
of the particle
approximation
considerably simplifies the mathematics without seriously hindering the
model's ability to capture the effects of the main operating variables.
a similar derivation procedure
Applying
for softening coals gives
the rate of tar leaving the particle surface as
dY/dt
Ei
=
EC'
(1.3-7)
kti(V*i-Vi)
where
2
E,,=
exp(-mSReff)/[l+exp(-
2
msReff)]
(1.3-8)
The quantity msReff is the Thiele modulus for softening coals, where ms
=
(kc/DL) 1 /2;
molten coal;
DL represents
and Vi
coals given in
is
the liquid phase diffusivity of tar in the
and EP,5
Eqs.(1.3-4)
are the same as those for non-softening
and (1.3-5)
respectively.
assumed to have a shape of a cenospherical
thus Reff'I
shell (Sung,
the effective diffusion length scale,
of the shell thickness
(Griffin, 1988;
The molten coal
1978),
and
assumed to be half
is
Hsu, 1988).
Based on recent
data from ~40 pm rad. particles pyrolyzed at 1 atm (Griffin, 1988), the
shell thickness
coal.
and
is
assumed to be roughly 20% of the radius of the raw
An exact explanation for the experimentally
particle-size
currently
not
property values
effect
established
on
tar
due
production for
to
of the molten coal
large
(Oh,
observed pressure
softening coals,
uncertainties
1985).
excluding external and bubble transport effects
In
in
is
physical
this formulation,
leaves the possibility
that the shell thickness is a function of pressure and particle size as
the only viable explanation to describe
the observed behavior.
until more conclusive explanation becomes available,
Thus
the present model
that the shell
assume
will
related to the pressure
is
thickness
and 40
particle size in the form of
Re f
where
=
0.1 R x 10~'* (P/1) 1 /3 (R/40)i/ 3 cm
R is
atm.
the particle
work
The
ym,
(1988)
of Griffin
experimental
quantitative
in
radius
and P the reactor pressure
is
currently
seeking
to
in
provide
the effect of pressure
to examine
data
(1.3-9)
and
particle size on the shell thickness.
For
a given coal,
input parameters:
Dgeff
V*max
DL.
or
to
studied (see below,
predict
gives
tar
the
Fig.l.3-13).
is
V*max
k, ,
k0 ,,
of
experimental
data
estimates
(Gavalas,
estimates
are
E,,
found
ko C
in
1984)
and
the
EC
rates
the
are
gas-phase
E0 ,
obtained
(Serio,
model
six
coals
tar yield
either
1984),
methods,
reactions
and
of
(vacuum).
low pressures
literature
for
for
the experimental
using thermochemical
strictly valid
values
estimated
evolution
limit obtained with rapid heating under
values
kop, E,, k0e,
ko,, Eo,, o,
,
1.3-1
Table
used
parameters
(X V*j)
a total of 9
MIPR model requires
the extended
or
The
from
from
although such
only.
These
parameters were assumed not to vary significantly among different coal
types.
This assumption was mainly made because information necessary
to assign coal-type dependent values for these parameters is presently
not
available, but we
do not
constant for all coal types.
imply that these parameters
Any errors generated from this assumption
will affect the values of best-fitted parameters.
be
sufficiently
If the error cannot
compensated by the fitted parameters,
will be reflected in the model's predictive capability.
parameters,
EO,,
O,,
experimental tar data.
and
are truly
Dgeff
or
DoL
were
then the error
The remaining
best-fitted
from
Model parameters for the extended MIPR model.
Table 1.3-1
41
(a) Coal-type dependent parametersa
wt% dmmf
e/rc
,
E,
V*max
Coalb
kcal/mole
or
DoL d
kcal/mole
16.8
53.8
7.0
10-2.81
9.1
52.8
9.4
l0-3 .23
Smith Roland SB
14.8
51.7
6.3
10-2.70
Blue HVB
27.7
54.6
5.3
10~2.90
Illinois HVB
30.1
54.4
4.4
10~5.67
Lower Kittanning LVB
14.0
56.8
3.5
1~5. 41
Lower Wilcox L
Beulah Zap L
(b) Fixed parameterse:
scission
ko,,
s-
polymerization
k 0 p,
s-
1
10,
EP, kcal/mole
35.5
cracking
k0 , s-
1
1014
E,, kcal/mole
55.0
a V* max
is
obtained
from vacuum tar
1014
yield data;
E0 ,,
,,
e/r
or DoL are
best-fitted from the data.
b Coals are listed in the order of increasing elemental carbon contents in
dmmf basis. Elemental analysis is given in Table 1.2-1.
1 5
(1/P) cm2 /s.
.
= (e/r) 0.1 (T/273)
C Geometrical factor in Dg,,ff
2
d Liquid phase diffusivity, DL = DoL (T/298)
cm /s.
e See text for sources.
42
Modeling results and discussion
compares
Figure 1.3-3
MIPR model.
tar yields from the MIPR model
the experimental and predicted
for the six coals investigated in
The model predictions were made with ko fixed at 1014
study.
this
S-1,
Y*
obtained from the measured maximum tar yield, and E0 and a best-fitted
a multivariable non-linear regression
the experimental data using
to
In
routine.
all
the
cases,
experimental values;
agree well
predicted yields
standard error of the estimate
the
with
the
[Eq.(1.2-3)]
ranges from 6.5 to 10 % of the maximum tar yield.
Figure
1.3-4 plots
production versus
rank coals,
the
best-fitted values
the elemental
indicated by higher
of
carbon contents
far
E0
elemental carbon contents,
the variation
explainable
conditions employed in this
study, there appears
tar
Higher
generally
Such differences
by experimental
i1 kcal/mole for both E0 and a.
+
estimated to be
for
Maximum differences
and a are 7.1 and 3.6 kcal/mole respectively.
exceed
and a
of the coal.
gave greater values of E0 and smaller values of a.
in
E0
uncertainties,
Therefore,
under the
to be a convincing
coal-type effect on the MIPR model rate parameters for tar production.
The trends
for both E0 and a in
Fig.l.3-4 appear to be more scattered
among low-rank coals, where the Beulah Zap lignite shows a considerably
and higher a compared to the Lower Wilcox lignite and Smith
lower E0
Roland subbituminous coal.
One property that appears to distinguish
the different behavior of the low-rank coals is the elemental hydrogen
in
content;
and
Smith
dmmf basis,
Roland
respectively.
the Zap has 4.8 wt% whereas the Lower Wilcox
have noticeably
Therefore in
larger values
estimating E0
of 5.6
and 5.3 wt%
and a from Fig.l.3-4 in
the
43
1403
10-
M
9-
LU
2-
300
500
700
900
1100
300
1000
a00
15 -
26
14 -
24
0
500
700
900
1100
00
1000
-
22
13
20
10
1a
9
14
12
2
300
500
900
700
1000
00
1100
300
50
700
900
1100
700
900
1100
00
1000
12
28
26
0
11
0
24
-0
10 -
22
20
Q
9 -
1Is
N
is
4
2-
E
P00
A 700
T
00
PEAK1
TEAIPKOAISK
(C)
1100
T00
1
(000
HoLoimNG
ro~PAI1-l0nc (C)
300
500
(C)
PEAK
TEMPERA1TJRE
00
1000
(C)
Hot..DMGTMPFtRTURE
Figure 1.3-3
Tar yields versus peak and holding temperatures (5 s hold).
Symbols represent experimental data; lines represent MIPR model predictions.
(a) LW, (b) ZP, (c) SR, (d) BL, (e) IL, (f) LK. Abbreviations: see Fig.l.24.
53
55 -
44
54 -
00
O
E
52 -
o
/
/
50
49 -
70
74
78
82
86
90
94
7.5
b
7.0 -
6.5
o
6.0 -
E
0
-Y 5.5
5.0 -
4.5 -0
4.0
0
0
3.5-
1
70
74
78
82
86
90
94
ELEMENTAL CARBON CONTENT (WT7 DMMF)
Figure 1.3-4
Best-fitted values of (a) E0 and (b) a for predicting
atmospheric tar production using the MIPR model, versus the elemental carbon
content of the coal.
ko was fixed at 1014 s~1 for all coals; Y* was
obtained from experimental data for each coal.
Dashed lines are for coals
with [H] < 5 wt% dmmf; solid lines for [H] > 5 wt% dmmf.
Carbon: LW < ZP <
SR < BL < IL < LK. Abbreviations: see Fig.l.2-4.
low-rank region,
the dashed lines are recommended for coals with the 45
elemental hydrogen content of < 5 wt% dmmf, and the solid lines for
coals with the elemental hydrogen of t 5 wt% dmmf.
Figures 1.3-5 through 1.3-9 compare the experimental and predicted
gas yields from the MIPR model for the six coals investigated in
The model
study.
above
for
tar.
predictions
For
all
this
were made using the same procedure
gas
species,
the
agreement
between
as
the
predicted and experimental yields is generally good; the standard error
of estimate ranges from 4 to 15 % of the maximum yield.
plots the best-fitted values of E0
and a for the measured gas species
versus the elemental carbon contents of the coal.
E.
range from almost none for C2 H4
trend for higher rank coals
shows
a
concave
bituminous
comparable
downward
range.
to
in
estimated
uncertainties, which
Coal-type effects on
and CO2 , to a slightly increasing
cases of CH4
trend
However,
Figure 1.3-10
with
a
and C2 H 6 .
minimum
these
variations
errors
produced
range from ±
0.5
to
The E0
near
are
high-volatile
small
from
1 kcal/mole.
trend for a is decreasing values for higher rank coals.
effect is strong for C02,
errors
and
are
experimental
The general
The coal-type
but for other gas species the effect is much
Except for C02 , variations
weaker.
of CO
produced from experimental
in
a are comparable to estimated
uncertainties,
which range from ± 1
to 1.5 kcal/mole.
Figure
1.3-11
compares
the
experimental
and
predicted
volatile yields from the MIPR model for the six coals studied.
cases,
the predicted yields
standard
yield.
total
In all
agree well with experimental values;
error of the estimate
ranges
from 6 to 10
the
% of the maximum
Figure 1.3-12 plots the best-fitted values of E0
and a versus
2.2
46
1.8
2.0
1.51.41.6
1.3-
1.4
1.2-
bD
1.2
0.90
2
0.80. T
0.8
0.6
0.8
0.4
0.4-
0.2
0.2-
0.30.1-
0.0
400
600
00
1000
800
I
I
1000
400
2.6
3.0
2.4
2.8
2.
22
WO
1000
am8
I
I
I
am8
1008
0
-
2.0
2.2
2
1.5
1.4
10-
9
1.0
102
0.a
2
0.13
0.0
0.4
0.4
02
400
800
S00
1000
00
1000
400
4.0
4.5
3.5 -
4.0-
S00
800
l000
800
1000
000
low0
3.5-
3.0
3.0
0
2.5
2.5
2-
0
2.0
1.5
1.0
1.0 0.5
0.0
0.0
400
500
800
1000
(C)
TEMPERATURE
PEAK
B0
1000
)
HOLDING
TIEMPERATURE
(C)
400
00
00
PEAK
TEMPERARE
(C)
1000
HO.DING
TMPERATURE
(C)
Figure 1.3-5
Methane yields versus peak and holding temperatures (5 s
hold).
Symbols represent experimental data; lines represent MIPR model
predictions.
(a) LW,
(b) ZP,
(c) SR,
(d) BL, (e) IL, (f) LK.
Abbreviations: see Fig.1.2-4.
47
1.2
0.9
z
0.7
0.0
L
0.5
400
&00
800
1000
500
1000
400
S00
800
1000
800
1000
600
800
1000
800
1000
2
1.4 2.0-
1.4
-
1.8 1.2-
0.7
-
z
0.5-
0.4 -
0.50.4 -
0.3
0.2
0.1-
0.0 - -400
0.2 -
600
1000
800
800
400
1000
0.7
0.6-
00
0.9 -
0.80.7-
0.5-
0.8
9
0.30.4
0.2-
0.3 0.20.
0.1-
10.0Y400
600
800
1000
(C)
PEAK
TEMPERATURE
800
1000
HOLDING
TEMPERATURE
(C)
400
600
800
1000
PEAK
TEMPERATURE
(C)
800
1000
HOLDING
TEMPERtATURE
(C)
Figure 1.3-6
Ethylene yields versus peak and holding temperatures (5 s
hold).
Symbols represent experimental data; lines represent MIPR model
predictions.
(a) LW,
(b) ZP,
(c) SR, (d) BL,
(e) IL, (f) LK.
Abbreviations: see Fiz.l.2-4.
48
j
0.20
00
0.20
0
0.15
0
0
0
0.10 -
0.05 -
400
000
800
1000
.0600
S00
1000
500
1000
400
S00
$00
0.200
0.50
0.40 -
0.33
-0
0.300.25 0
0.200.15 0.10 -0
0.05 -O
0.00
800
1000
400
600
00
1000
800
1000
0.8
0
800
1000
800
1000
400
PEAK
TEMPERAlJRE
(C)
600
800
HOLDNG
TEMPERATURE
(C)
PEAK
TEJPERATURE
(C)
1000
TP
1000
U
H0LDIN0r]EIPERATJRE
CC)
Figure 1.3-7 Ethane yields versus peak and holding temperatures (5 s hold).
Symbols represent experimental data; lines represent MIPR model predictions.
(a) LW, (b) ZP, (c) SR, (d) BL, (e) IL, (f) LK. Abbreviations: see Fig.l.24.
13
b
12
11
10
9
a
0
400
600
1000
S00
S00
1000
5-
400
B00o
800
1000
600
1000
10
0
z
0
2
ZA
M
400
1000
8c
600
500
400
1000
0
w0
00w
00
1000
1.0 -
4.0
3.5
0.7 -3.0
0-.7
0.4-
2.0
a
0
0.2
0
1.0
I
400
S00
.
m
1000
(C)
PEAKTEMPERATURE
800
1000
HO.DINGTEMPERATunE
(C)
400
00
I
T
I000
PEAK
TEMPERATURE
(C)
00
1000
HOL.DING
TEMPERATURE
(C)
Figure 1.3-8 Carbon monoxide yields versus peak and holding temperatures (5
Symbols represent experimental data; lines represent MIPR model
s hold).
(c) SR, (d) BL, (e) IL, (f) LK.
(b) ZP,
(a) LW,
predictions.
Abbreviations: see Fig.l.2-4.
)
C
7-
a
9
w
aX
0
0
300
500
700
900
11I
800
1000
300
500
700
900
1100
800
1000
500
700
900
1100
g00
1000
4.5
3 -
9
w
X
0
0
a
z0
1.5
0
0.5
a
0
300
500
700
900
1100
500
092
1000
300
2.2
2.0 1.8 -
F
0.4
z
0.2
1.2-
x
X
z
1.00.8
0.6
0.4-
0.0.0
300
,
,0i
500
700
900
PEAK
TEMPERATURE
(C)
1100
a0m
1000
)
HOLDING
TIEMPERATUR1E
(C)
300
500
700
900
PEAK
TEMPERATURE
(C)
5009P1000
1100
(C)
MOLDING
TEMPERATURE
Figure 1.3-9 Carbon dioxide yields versus peak and holding temperatures (5
Symbols represent experimental data; lines represent MIPR model
s hold).
(c) SR, (d) BL, (e) IL, (f) LK.
(b) ZP,
(a) LW,
predictions.
Fig.l.2-4.
see
Abbreviations:
70
51
68
64
0
E
62
U
0
W
60 -
58 -
56 -
54
70%
10
82
7
90
86
94
-T
b
9
x
x
x
A
8
0
A
0
E
7
Co
x
+
6
++
0
D CH 4
~
+0
CO 2
5 _
+
*C2H6
0
4 -
0
C2 H4
3 -70
O
CH4
I
I
74
+
I
78
82
86
ELEMENTAL CARBON CONTENT (WTX DMMF)
C2H4
0
C2H6
a
CO
I
90
-__T_
X
C02
94
Figure 1.3-10
Best-fitted values of (a)
E 0 and (b) a for predicting gas
evolution rates at 1 atm using the MIPR model, versus elemental carbon
contents of the coal.
ko was fixed at 1014 s-1 in all
cases, and Y* was
obtained from experimental data for each coal.
Carbon: LW < ZP < SR < BL <
IL < LK.
Abbreviations: see Fig.l.2-4.
52
w
80
K
40
9
30
20
0
0
0
000
10
300
500
700
900
1100
800
1000
300
500
700
900
1100
800
1000
50
~~0
50
30
30
0
20
300
500
700
900
1100
800
20
1000
500
26
700
000
800
1100
1000
f
24 -
0
O
22
20
50
9
9
30
C3a
12
01
C,
a
o20
10
20
300
500
700
900
(C)
PEAK
TEMPERA11RE
1100
800
1000
HOL1NGTEMPERAnfJRE
(C)
300
500
700
900
PEAK
TEMPERATURE
(C)
I100
800
MoONG
1000
--EMPERATURE
(C)
Figure 1.3-11 Total volatiles yield versus peak and holding temperatures (5
Symbols represent experimental data; lines represent MIPR model
s hold).
(b) ZP,
(c) SR, (d) BL, (e) IL, (f) LK.
(a) LW,
predictions.
see
Fig.l.2-4.
Abbreviations:
60 0)
59 -
53
58 57 -
E
0
W
56 13
55 -
0
54 -
03
03
53 52 51 50
70
74
~1
A
9.5
b
78
82
86
90
94
0
9.0 8.5 -
0
8.0 -
E
7.5 -
03
0
7.0 -
E3
6.5 6.0 5.5 13
5.0
i
70
74
78
82
86
90
94
ELEMENTAL CARBON CONTENT (WT7. DMMF)
Best-fitted values of (a) E0 and (b) a for predicting
Figure 1.3-12
atmospheric pressure total volatiles evolution using the MIPR model
versus
the elemental carbon contents of the coal.
ko was fixed at 1014 sin all
Carbon:
cases, and Y* was obtained from experimental data for each coal
LW < ZP < SR < BL < IL < LK.
the elemental carbon contents of the coal.
Generally, higher rank
coals show increasing values of E0 with a maximum difference of about 6
The E0 of Lower Wilcox lignite appears to be high compared
kcal/mole.
to the other two low-rank coals, but is within estimated uncertainties
of ±1 kcal/mole.
Comparing the a shows a decreasing trend for higher
rank coals, but with much scatter.
in
a is
moderately
kcal/mole.
The
greater
trends
A maximum difference of 4 kcal/mole
than
the
the
rate
in
measured total volatiles production
estimated
uncertainty
parameters
confirm the
for
of
±1.5
independently
trends
observed
for
individual products.
The relatively modest coal-type dependence of the
MIPR rate
for total volatiles
parameters
reflect the combined effects
of a strong coal-type dependence for tars and a much weaker dependence
for
gases.
Also,
the
general
trend
for
all
products
are
always
consistent - higher E, and lower a for increasing coal rank.
Extended MIPR model.
the
experimental
and
0.001, 1 and 10 atm.
predicted
measured
predicted behavior
low
yields
at
all
three
pressures-
is
especially encouraging
since,
unlike the
the yield limits were predicted without having to rely on
experimentally
fairly
tar
The accurate prediction of the yield limits over
a wide range of pressures
MIPR model,
Figure 1.3-13 shows a good agreement between
in
temperatures
values
which
at
the
(~C 550
different
rate
C),
pressures.
is unaffected
and
that
the
Also,
by pressure
yields
the
at
'level-off'
earlier (i.e., at lower temperatures) as pressure is increased, closely
resembles
the
experimentally
observed
behavior
reported
by
Suuberg
The
figure
versus temperature,
where
(1977).
Figure 1.3-14 helps to illustrate how the model works.
plots
(a)
Epavg, E,
and (b)
Epavg
E
54
55
20
12
Is
P.atm
0.001
10
10-
a
Patm,
S0.001
12
4CC
23
0
00
15
351
tot
-
700
1 100
900
30050
E
700
900
Pat,,
10
300
aP.a
0.0001
~
10
00
1Ow
110
0.0
00
1030SN
1 100
go0
-
10
1
.
0
01
15 -
20
0
2
100
-
31
Ita
9
201
0I
1102
5 -
5
0,
300
.
w00
I4
i 1
700
,
0
900
1100
TEMPERATUREC
(C)
Tar yields versus peak
experimental data: U - 0.001 atm, C - 1
extended MIPR model predictions.
Goals:
0.001 and 10 atm data points
IL, (f) LK.
Figure
runs.
1.3-13
Abbreviations:
see Fig.l.2-4.
300
D
000
700
900
1100
7EMJPERATURE
(C)
Symbols represent
temperatures.
atm, EJ - 10 atm. Lines represent
(a) LW, (b) ZP, (c) SR, (d) BL, (e)
represent avera~ed values from 1-3
1
1.0
a
0.9
01-oo
56
atm
0.7
0.6
0
Ep,avg
0.5
Ec,ns
P =10atm
0.4
1 atm
0.3
0.2
0.1 0.0
-
100
300
500
700
900
1.0
b
0.9 0.8 0.7 0.001
atm
V)
0.6 -E
Ep,avg
w
0
0
Ec,,
0.5 -
1<1
P=1o at
0.4
latm
0.3 -
0.2 0.1 -
0.0
I
-
100
300
500
700
900
TEMPERATURE (C)
Figure 1.3-14
(a) Ep,avg and Ecns versus temperature for a non-softening
coal (LW). (b) E
and Ec's versus temperature for a softening coal (IL).
E av = X E
if(t)A, E ,- from Eq.(1.3-5), Ec,ns and Ecs from Eqs. (1.36) ang (1.3- 8) respectively. Abbreviations: see Fig.l.2-4.
E p,
E,,i
=
f(E)AE.
Recall
that the
rate
of tar production for 57
both non-softening and softening coals is represented as the product of
the
total
rate
at which
the
non-x-linked
fraction
reacts
and the two
'E' factors, and that the values of these E factors range between 0 and
1.
The decrease
in tar production at higher pressures
the smaller E ,n, or
and EC,,
EC,,
as the pressure increases.
is explained by
At vacuum, Ecn
are near 1 indicating negligible mass transport resistance; at
(~
high pressures
10
atm),
the
values
are
much
lower,
indicating
a
substantial transport resistance.
Another
explain
important
feature
of
this
model
is
the experimentally observed heating-rate
1 atm.
Niksa
(1981)
that
it
effects
is
able
to
at vacuum and
observed an increasing volatiles
(implying tar)
production at higher heating rates under vacuum, whereas Anthony et al.
(1974)
and
product
Suuberg
yields
at
contradictory,
shows
that
(1977)
1
observed negligible
atm.
The
but this model
the
sets
of
effect,
indicated
At
transport
1
atm,
effects, higher heating
where
by
effects
on
first
appear
Figure
1.3-14
Epavg,
is
more
The non-x-linked fraction of coal reacts
at higher temperatures as the heating rate is
mass
resulvs
can explain the results.
polymerization
severe at lower temperatures.
two
heating-rate
mass-transport
produced
at
higher
secondary
tar
cracking reactions.
temperatures
rates
effects
are
experience
Thus,
increased.
Thus without
enhance tar
not
a
production.
negligible,
greater
the increased
tar
extent
tars
of
production
at higher temperatures is 'off-set' by more cracking reactions.
Figure 1.3-15 plots the best-fitted values of EO,
elemental
carbon
contents
of
the coal.
As before
and a. versus the
for
the MIPR model
(Fig.l.3-4), higher rank coals generally gave greater values of E0 , and
58
58
a
57
56 -
55
E
0
U)
54
00
53
52
51
50
70
74
78
82
86
90
94
74
78
82
86
90
94
10
b
9_-
8 -
E
o
7-
6
5
4
70
ELEMENTAL CARBON CONTENT (WT% DMMF)
Figure 1.3-15
Best-fitted values
evolution using the extended MIPR
of the coal.
Carbon: LW < ZP <
Fig.l.2-4. Other model parameters
of (a) Eos and (b) as for predicting tar
model versus the elemental carbon content
SR < BL < IL < LK.
Abbreviations: see
are given in Table 1.3-1.
smaller values of as,
coals
have
bond
implying that bridging molecules
dissociation
narrower distribution.
energies
with
a
of higher rank 59
greater
mean
and
a
The best fitted values of E0 , and a. are within
the range of expected values for the scission of bibenzyl type bridges
(Ph-CH2 - CH 2 -Ph).
Figure 1.3-16 plots the best-fitted values of e/r or DOL versus the
elemental carbon contents of the coal.
For non-softening coals (LW,ZP,
SR,BL),
range
the
best-fitted
These values
values
of e/r
imply that the tortuosity
the void fraction (e) is around 0.1.
order
of
magnitude
higher
than
values
within the
range
Sezen, 1985).
of DoL
typical
of
In
values
10~6*1
values of the fitted transport parameters
the ratio k. c /(e/r)
10-2.70.
about 100 assuming that
10-5.41
Changing the value of ko
0
actually f itted is
to
reported
for porous
For softening coals (IL,LK), the
are between
of expected values
is
10-3.23
Such values for r are at least an
solids (Froment and Bischoff, 1979).
best-fitted
(r)
from
and 10-5.67,
(Oh, 1985;
which
are
Suuberg and
will directly influence
the
since the quantity that is
or k. c /Do L.
applications employing coals other than those studied here,
use
of the model parameters obtained from the experimental data specific to
the coal of interest would give the most reliable performance.
If such
experimental information is not available, use of the parameter values
estimated from the coal-type dependent trends established in this study
is expected to give the next best performance.
using the estimated values is
that the trends were established from a
fairly small number of coals (6),
some
here.
A note of caution in
and thus there is a possibility that
'unusual' coals may behave very differently
from those studied
60
10
10
~e/7-
0
10
04
DD
100
DOL
70
74
78
82
o
86
90
94
ELEMENTAL CARBON CONTENT (WT% DMMF)
Figure 1.3-16 Best-fitted values of transport parameters for predicting tar
evolution using the extended MIPR model versus the elemental carbon content
of the coal. e/r is for non-softening coals (LW,ZP,SR,BL),
and DoL is for
softening coals (IL,LK).
Carbon: LW < ZP < SR < BL < IL < LK.
Abbreviations: see Fig.l.2-4.
Other model parameters are given in Table
1.3-1.
1.4. Conclusions
1)
Among the six coals studied, higher rank coals generally produced
tars at higher temperatures, and over a narrower range of temperatures.
Consequently,
a larger mean and a narrower distribution
of global
activation energies were obtained using the MIPR model for coals of
increasing rank.
2)
A quantitative correlation, developed to independently relate
tar
yield limits to coal type and pressure, was tested against a large set
of experimental data representing a wide range of coals (37 coals,
ranging from lignites to anthracites) and pressures ('vacuum' to 90
atm).
A good agreement between the predicted and experimental yields
was obtained for all
coals and pressures, with a standard error of
estimate of ±3 wt% dmmf.
3)
In general, no discernable coal-type effects on the apparent rate
of gas production were observed.
A probable explanation for this
is
that variations in the rate caused by different coal types
are
comparable
to
those
caused
by
uncertainties
in
experimental
measurements.
Consequently, the kinetic parameters of the MIPR model
for measured gas species were only slightly affected by coal type.
4)
Higher rank coals generally produced less carbon oxides and
pyrolytic water, but more methane.
The ethylene and ethane yields were
small and their absolute yield values were less affected by coal type.
5)
Total volatiles
evolve at higher temperatures and over a narrower
range of temperatures for higher rank coals.
Thus for coals of
increasing rank, a larger mean and a narrower distribution
of global
activation energies were obtained using the MIPR model.
These trends
are consistent with the expected behavior from combining the observed
coal-type effects on the rate of tar and gas production.
6)
The total volatiles yield limit is fairly constant among the
lignites,
and subbituminous and high-volatile bituminous coals (41-55
wt% dmmf),
but is significantly less for the low-volatile bituminous
coal (22 wt% dmmf).
The high-volatile bituminous coals produced
significantly more reactive volatiles than other coals (38-45 versus
19-28 wt% dmmf); reactive volatiles are defined as total volatiles
minus water and carbon dioxide yields.
7)
Predicted tar yields from the extended MIPR model agreed well with
experimental values for a wide range of coal types (lignites
to lowvolatile bituminous coal; non-softening and softening) and pressures
(0.001-10 atm).
8)
The best-fitted
values of EO, and a. for bridge scission are within
the range of expected values for the scission of bibenzyl type bridges
(Ph-CH 2 -CH 2 -Ph).
The best-fitted values of e/r imply a tortuosity (r)
that is about an order of magnitude greater than typical values
reported for porous solids, and those of DoL are within the range of
values reported in the literature.
61
1.5. References for summary
62
Anthony, D.B., Howard, J.B., Hottel, H.C., and Meissner, H.P., "Rapid
Devolatilization of Pulverized Coal, "
Fifteenth Symp.
(Int.)
on
Combustion, The Combustion Institute, Pittsburgh, 1303, 1974.
Arendt,
P.A. and van Heek, K.J. , "Comparative Investigations of Coal
Pyrolysis Under Inert Gas and Hydrogen at Low and High Heating Rates
and Pressures Up to 10 MPa," Fuel, 60, 779, 1981.
Bautista,
J.R,
"Time-Resolved Pyrolysis
Product Distributions
of
Softening Coals," Ph.D. Thesis, Dept. Chem. Eng., Princeton University,
Princeton, NJ, 1984
Burnham, A.K.,
Oh, M.S.,
and Crawford, R.W.,
"Activation Energy
Distributions and Related Chemistry for Pyrolysis
of the Argonne
Premium Coals," submitted for publication in Energy and Fuels, 1988.
Ciuryla, V.T., Weimer, R.F., Bivans, D.A.,
and Motika, S.A.,
"Ambientpressure thermogravimetric characterization of four different coals and
their chars," Fuel,
Cosway,
R.G.,
58,
748,
S.M. Thesis,
1979.
Dept.
Chem.
Eng.,
M.I.T.,
Cambridge, MA,
1981.
Cypres, R.
and Bettens, B.,
Tetrahedron, 1253, 30, 1974.
Cypres, R. and Bettens, B.,
Tetrahedron, 353,
31, 1975a.
Cypres, R. and Bettens, B.,
Tetrahedron, 359,
31, 1975b.
Fong, W.S.,
"Plasticity
and Agglomeration in
Coal Pyrolysis,"
Thesis, Dept. Chem. Eng., M.I.T., Cambridge, MA, 1986.
Sc.D.
Franklin, H.D., Peters, W.A.,
Cariello, F., and Howard, J.B.,
"Effects
of Calcium Minerals on the Rapid Pyrolysis of a Bituminous Coal," Ind.
Eng. Chem. Proc. Des. & Dev., 20, 670, 1981.
Freihaut,
J.D.
and Seery,
D.J.,
"An Investigation of Yields and
Characteristics of Tars Released During the Thermal Decomposition of
Coal," Am. Chem. Soc. Div. Fuel Chem. Prepr., 26 (2), 133. 1981.
Freihaut, J.D.,
Zabielski, M.F.,
and Seery, D.J.,
"A Parametric
Investigation of Tar Release in
Coal Devolatilization,"
Nineteenth
Symposium (Int.) on Combustion, The Combustion Institute,
Pittsburgh,
1159, 1982.
Froment, G.F. and Bischoff, K.B.,
Chemical Reactor Analysis and Design,
J. Wiley & Sons, New York, 1979.
Gavalas, G.R., Cheong, P.H.K., and Jain, R.,
"Model of Coal Pyrolysis.
1.Qualitative Development," Ind. Eng. Chem. Fundam., 20, 113, 1981.
Gavalas,
G.R.,
Coal
Pyrolysis,
Elsevier
Scientific
Publishing
Co.,
Amsterdam, 1984.
63
Gerstein, B.C., Murphy, P.D., and Ryan, L.M., "Aromaticity in Coal,"
Coal Structure, R.A. Meyers, ed., Academic Press, New York, 1982.
in
Given, P.H., "The Organic Chemistry of Coal Macerals," Penn State Short
Course on Coal, The Pennsylvania State University, June 1976.
Griffin, T.P., "Intra-Particle
Ph.D. Thesis, Dept. Chem.
Eng.,
Secondary Reactions in Coal Pyrolysis,"
M.I.T., Cambridge, MA, in preparation
1988.
Howard, J.B. , "Fundamentals of Coal Pyrolysis and Hydropyrolysis, " in
Chemistry of Coal Utilization, 2nd Suppl. Vol.,
M.A. Elliott,
ed., J.
Wiley & Sons, New York, 1981.
Hsu, J., Ph.D. Thesis,
preparation 1988.
Dept.
Chem.
Eng.,
Ko, G.H.,
Peters, W.A.,
and Howard, J.B.,
from Rapid Pyrolysis with Coal Type and
M.I.T.,
Cambridge,
MA,
in
"Correlation of Tar Yields
Pressure," Fuel,
66,
1118,
1987.
Ko, G.H., Peters, W.A., and Howard, J.B., "Comparison of Tar Evolution
Rate Predictions in Coal Pyrolysis from Multiple Independent Parallel
Reaction Model and Functional Group Model Over a Wide Range of Heating
Rates," Energy and Fuels, in press 1988a.
Ko, G.H.,
Sanchez, D.M.,
Peters, W.A., and Howard, J.B.,
"Correlations
for Effects of Coal Type and Pressure on Tar Yields from Rapid
Devolatilization,"
Twenty-Second Symposium (Int.)
on Combustion,
in
press 1988b.
Loison,
R.
and Chauvin,
F.,
"Pyrolyse
Rapide
Du Charbon,"
Chem.
Ind.,
(Paris), 91, 269, 1964.
Niksa,
S.,
"Time-Resolved Kinetics of Rapid Coal Devolatilization,"
Ph.D. Thesis, Dept. Chem. Eng., Princeton University, Princeton, NJ,
1981.
Oh, M.S.,
"Softening
Coal Pyrolysis,"
Sc.D.
Thesis,
Dept.
Chem.
Eng.,
M.I.T., Cambridge, MA, 1985.
Reitzen, R.G.,
S.M. Thesis, Dept.
Chem. Eng.,
M.I.T.,
Cambridge, MA,
1978.
Serio,
M.A.,
"Secondary Reactions of Tar in
Coal Pyrolysis,"
Thesis, Dept. Chem. Eng., M.I.T., Cambridge, MA, 1984.
Solomon, P.R. and Hamblen, D.G.,
Pyrolysis, in Chemistry
Conversion, R.H. Schlosberg, ed., Plenum Press, N.Y., 1985.
of
Ph.D.
Coal
Sprouse, K.M. and Schuman, M.D.,
"Predicting Lignite Devolatilization
with
the
Multiple
Parallel
and
Two-Competing
Reaction
Models,"
Combustion and Flame, 43, 265-271, 1981.
64
Sung, W.F. , "The Study of the Swelling Property of Bituminous Coal,"
S.M. Thesis, Dept. Chem. Eng., M.I.T., Cambridge, 1978.
Suuberg, E.M.,
"Rapid Pyrolysis and Hydropyrolysis of
Thesis, Dept. Chem. Eng., M.I.T., Cambridge, MA, 1977.
Suuberg,
E.M.
and Sezen, Y.,
Processes in Coal Pyrolysis,"
Sydney, 913, 1985.
"Competitive Reaction
Proc.
1985 Int. Conf.
Coal,"
Sc.D.
and Transport
Coal Science,
Suuberg, E.M., Lee, D., and Larsen, J.W.,
"Temperature dependence
crosslinking processes in pyrolysing coals," Fuel, 64, 1668, 1985.
of
Suuberg, E.M., Unger, P.E., and Larsen, P.E., "Relation between Tar and
Extractables Formation and Cross-Linking during Coal Pyrolysis," Energy
and Fuels, 1, 305, 1987.
Van Krevelen, , D.W., Coal, Elsevier Publishing Co., Amsterdam, 1961.
Weimer,
R.F.
and Ngan, D.Y.,
"Rates of Light Gas Production by
Devolatilization of Coal and Lignite,"
Am. Chem. Soc. Div. Fuel Chem.
Prepr., 24 (3), 129, 1979.
2. Introduction
Coal
projected
to
be
an
The vast coal reserve
future.
coal
is
65
types
with
Understanding
widely
the
important
in
the U.S.
varying
source
energy
in
the
consists of many different
chemical
relationship between
of
and
the
physical
coal
properties.
properties
and
its
conversion behavior is essential for efficient utilization of the coal.
Combustion,
utilization.
these
gasification and liquefaction are main routes for coal
Coal pyrolysis
conversion processes,
processing
steps.
devolatilization
occurs during
initial stages
and thus impacts
For
instance,
influence
combustion
in all
of
the course of subsequent
tars
generated
behavior
since
from
they
coal
affect
ignition and flame stability, soot and PAH formation, heat release, and
overall burning efficiency.
and
composition
of
the
In
coal gasification,
product
gas
depend
the heating value
on
the
yields
and
distributions of pyrolysis products.
Coal
pyrolysis
involves
complex
thermal
decomposition
reactions
coupled with multicomponent mass transport in a molten liquid or porous
solid depending
decomposition
on whether
reactions
are
the coal is
generally
secondary reactions (Serio, 1984).
produce
refers
volatile
a softening type or not.
The
distinguished
and
as
primary
The former refers to reactions that
products directly
from
the coal,
to further reactions of primary products.
while
the
latter
In addition to the
mass transport, coal type exerts a significant influence on the extent
of both primary and secondary reactions (Howard, 1981).
The focus of this
study is
to improve the understanding of coal-
type effects under conditions of practical interest.
Many modern coal
conversion
particles
processes
rapidly
heat
small
coal
under
atmospheric
or
higher
reactor
Low
pressures.
pressures
and
small 66
particle sizes minimize mass transport resistances, and thus the extent
of secondary reactions within or immediately adjacent to the pyrolyzing
The
coal particle.
secondary
reactions
effect of heating rate on the extent primary and
remains to be established as sufficient intrinsic
kinetic information on the relative rates of the two reaction types is
not
available.
For this reason,
extrapolating kinetic measurements
over a wide range of heating rates can be a difficult procedure (Ko et
al.,
1988a).
Therefore,
investigated using
in
this
small particles
study,
under
coal-type
effects
rapid heating rates
are
over a
wide range of pressures.
Understanding
requires
extents
the
pyrolysis
behavior
a reliable experimental
of coal-type
observed
effects,
behavior.
among
different
data base to determine
coal
types
the kinds and
and a mathematical model to explain the
Quantitative
time-resolved
product
evolution
measurements for a wide range of coal types under conditions of minimal
mass
transport
limitations,
coal-type effects,
are
needed
to
experimentally
but such data are currently lacking.
establish
In response,
the experimental phase of this study examines the pyrolysis behavior of
six coals
ranging from low-rank lignites
to very high-rank bituminous
coals under conditions where mass transport resistances are small.
The modeling
the experimental
to
measurable
utilized
model
in
phase of this work derives kinetic information
data,
(MIPR) model
describes
and attempts to relate the kinetic information
properties
this
kinetics
from
study:
of
the
the
and the
coal.
multiple
Two
different
independent
extended MIPR model.
of product
evolution under
models
parallel
The
are
reaction
former model
conditions
where
the
effects
of physical
transport processes
relatively
unimportant.
approximate
descriptions
and
thus
is
applicable
The
latter
and
secondary
model
reactions
explicitly
are
includes
of transport and secondary reaction effects,
over a wider
range
of operating conditions.
More rigorous models often require detailed information on chemical and
physical
properties
of
the
coal,
many
of
which
are
difficult
to
estimate or experimentally measure with currently available techniques.
Such limitations are greatly magnified when one needs to consider many
different coals.
67
3. Background
68
This chapter provides selected background information pertinent to
the main focus of this study - investigating the effect of coal type on
pyrolysis
behavior.
structural
possible,
properties
trends
Section
Section
3.2
chemistry
among
and
mass
Section 3.3 reviews
discusses
different
that relate
describes
3.1
coal
chemical
types,
and
physical
noting
wherever
the structural properties
the
current
transport
understanding
phenomena
experimental
relevant
to
of
to coal type.
the
coal
data on coal pyrolysis,
reaction
pyrolysis.
specifically
the effect of main operating variables on pyrolysis behavior.
to
quantitatively
model
the
experimentally
observed
Efforts
behavior
are
described in Section 3.4.
3.1. Coal characteristics
According to the A.S.T.M. classification scheme shown in Table 3.11,
coals
are
anthracites,
contents,
are
as
depending
on
lignites,
the
subbituminous,
fixed
carbon
and the heating value of the coal.
further
bituminous,
A,B,C.
ranked
classified
into
medium-volatile
Table 3.1-2
different
bituminous,
shows variations
in
and
bituminous,
volatile
or
matter
Coals within each rank
groups,
e.g.,
low-volatile
and high-volatile
bituminous
some frequently used chemical
and physical properties among different coal types. Coals of increasing
rank (lignites --carbon,
+
anthracites)
aromaticity,
reflectance,
and
average
calorific
tend to have higher values of elemental
number
value;
and
of
lower
oxygen, carboxyl, hydroxyl, and volatile matter.
benzene
amounts
rings/layer,
of
elemental
High-volatile
Table 3.1-1
A.S.T.M. classification of coals by ranka.
from Singer, 1981.]
Class and Group
Fixed Carbon
Limits, %
(Dry, MineralMatter-Free Basis)
Volatile Matt
Limits, %
(Dry, Mineral
Matter-Free Basis)
Calorific Value Limits,
Btullb (Moist,"
Mineral-MalterFree Basis)
Equal or
Greater
Than
Equal or
Greater
Than
Equal or
Greater
Than
Less
Than
Less
Than
Less
Than
[Reproduced
Agglomerating
Character
I. Anthracitic
1. Meta-anthracite
98
2. Anthracite
3. Semianthracite'
92
86
98
2
8
92
8
14
II. Bituminous
1. Low-volatile
bituminous coal
78
86
14
22
2. Medium volatile
bituminous coal
69
78
22
31
.. .
69
31
.. .
14,000"
. ..
...
...
. . .'
13,000"
14,000
11,500
10,500
13,000
11,5()0
...
10,500
11,500
2
nonagglomterating
3. High-volatile
A bituminous coal
commonly
agglomerating"
4. High-volatile
B bituminous coal
5. High-volatile
C bituminous coal
S.
.
agglomerating
III. Subbituminous
1. Subbituminous
A coal
.
2. Subbituminous
B coal
.
.. .
9,500
10,500
3. Subbituminous
Ccoal
...
...
8,300
9,500
.. .
6,300
8,300
6,300
..
IV. Lignitic
1. Lignite A
2. Lignite B
nonagglomuerating
"This classification does not include a few coals, principally nonbanded varieties, which have unusual physical and chemical properties and
which come within the limits of fixed carbon or ca lorific value of the high-volatile bituminous and subbitiminous ranks. All of these coals either
contain less than 48% dry, mineral-matter-free fixed carbon or have more than 15,500 moist. mineral-mat ter-free B1tu per pouni.
'Moist refers to coal containing its natural inherent moisture but not including visible water on the surface of tihe coal.
If agglomerating. classify in low-volatile group of the bituminous class.
"Coals having 69% or more fixed carbon on the dry, mineral-moatter-froe basis shall be classified by fixed carbon, regardless of calorific value.
It is recognized that there may be nonagglomerating varieties in these groups of the bmitouminOus class, and there are notable exceptions in
high-volatile C bituminous group.
Reprinted from ASTM Stondords D 388, Classification of Coals by Rank.
69
Approximate values of some coal properties
Table 3.1-2
in different
rank ranges. [Reproduced from Franklin, 1980; data from Given, 1977.]
Lignite
Subbit.
C
High Vol. Bit.
B
A
65-72
72-76
76-78
78-80
%H
4.5
5
5.5
5.5
%0
30
18
13
10
% 0 as COOH
13-10
5-2
0
0
% 0 as OH
15-10
12-10
9
Aromatic C atoms % of
total C
50
65
?
Av. no.,
benz. rings/layer
1-2
?
40-50
35-50
35-45
?
31-40
0.2-0.3
0.3-0.4
0.5
0.6
0.6-1.0
%C (min. matter free)
Volatile matter, %
Reflectance,
%, Vitrinite
--
80-87
89
5.5
4.5
3.5
10-4
3-4
3
2
0
0
0
0
?
7-3
1-2
0-1
?
75
85-65
85-90
90-95
& 57
>25?
31-20
20-10
<10
1.4
1.8
4
15,000
15,800
15,200
2-3
Density
90
93
2.5
0
minimum
Total surface area
minimum
Plasticity and coke formation
only
Calorific value, moist,
min. matter free,
BTU/lb.
Bituminous
Medium Vol. Low Vol.. Anthracite
7000
10,000
12,000
13,500
14,500
70
bituminous
coals
have
detailed descriptions
maximum
values
of chemical
of
elemental
hydrogen.
More 71
and physical properties of coal are
given below.
3.1.1 Chemical structure
Detailed chemical
structure of coal
is
reviewed in
coal utilization (e.g., Given, 1976; Whitehurst et al.,
1984).
many books
on
1980; Gavalas,
Tingey and Morrey (1973) have compiled chemical structural data
reported
in the
literature
similar work up to 1977.
up
to
1973, and Suuberg
(1977) has done
Much more data have accumulated since then as
a result of improvements made in analytical techniques and a resurgence
of scientific
and commercial interest in coal utilization during the
70's and early 80's.
A brief survey of more recent literature was done
by Ko (Howard et al.,
1987a).
Figure
3.1-1
the literature
gives
survey,
a molecular
where
description of coal inferred from
the coal is
postulated to be made-up of
clusters of condensed and hydroaromatic rings (nuclei) held together by
bridge
groups.
Peripheral groups are
postulated structure identifies
peripheral
groups
as
main
(1)
attached to
nuclei,
(2)
building-units
the nuclei.
bridge groups,
of
a
coal
This
and (3)
molecule.
Differences within each of these three components reflect variations in
chemical
properties
description is
among
convenient in
different
coal
types.
Such
a
unitary
explaining the effect of coal type on the
pyrolysis behavior (Chapter 5), and in formulating a quantitative model
(Chapter 6).
Within nuclei, structural features reckoned
role
in pyrolysis are
the aromaticity, nucleus
to play an important
size,
and heteroatom
72
CH
4C
CH2~
nucleus
bridge
peripheral group
Figure 3.1-1
Hypothetical coal structure.
001
z
80
0
0
60
z
0
40
z
w
0
20
L
45
0
50
I
55
I
I
I
60
65
70
I
75
~
80
85
Ii~
90
95
PERCENT MAF CARBON IN THE COAL
Figure
versus
(1980) .]
Aromatic carbon, aliphatic carbon, and etheric carbon
3.1-2
[Reproduced from Whitehurst et al.
elemental carbon content.
content.
coal,
Aromaticity is a measure of the amount of aromatic carbon in 73
and is
defined as
fa = aromaticity = number of aromatic carbon atoms
(3.1-1)
total number of carbon atoms
Table 3.1-3 lists different methods to probe the aromatic structure of
nuclei and
scatter,
gives a brief description of each method.
Figure
3.1-2
shows
that higher rank coals
aromatic.
From the aromaticity
aliphatic
carbons
and
are
some
tend to be more
information, one can
hydrogens
Despite
present.
infer how many
These
aliphatic
quantities are believed to play key roles in thermal scission and freeradical
stabilization
type
reactions.
The
expressed as number of rings per nucleus.
size
are
1-2
anthracites
for
lignites,
(Given, 1977).
3-4
for
Winans
nucleus
size
is
often
Rough estimates of the ring
bituminous
et al.
coals,
and
> 4
(1988) recently suggested
that the ring size of bituminous coals could be as low as 1 or 2,
that
growth during
thermal
degradation
reactions may be a
reason for observing larger ring sizes for these coals.
is
an important property in
bridging
molecules
stabilization
reactions.
extent
of
a
of benzyl
resonance
is
it
radicals
affects
formed
the
The ring size
degree
of
from typical bridge
stabilization is
difficult to observe
on
the ring
structure
resonance
scission
of coal tars
(Wornat,
1988)
e.g.,
since
the
the position
(Stein and Golden, 1977).
Heteroatoms within the aromatic rings include N,
al.,
possible
determining bond dissociation energies of
also influenced by other factors,
CH2 ' radical
studies
and
But, a simple correlation between the ring size and the
stabilization
of
because
for
S,
and 0.
and model compounds
Pyrolysis
(Briunsma
et
1988) containing nitrogen heteroatoms indicate that the heteroatom
enhances the reactivity.
Unfortunately,
data on the types and amounts
Table 3.1-3
Aromaticity measurement techniques.
74
(1) NMR
Nuclear magnetic resonance arises from the interaction of the magnetic
component of electromagnetic radiation with the very small magnetic
moments possessed by certain nuclei ( 1 3 C, 1H).
Structural information
is obtained from measuring the magnetic field required to resonate and
the number of resonating nuclei.
For more information see Bartle and
Jones (1978), Retcofsky (1982).
(2) IR and FTIR
Infrared spectroscopic techniques measure the frequency and intensity
of light reflection emission and absorption due to the stretching and
bending motions of molecules. Fourier Transform techniques improve the
resolution and sensitivity of this technique.
(3) ESR
Electron spin resonance spectroscopy is based upon the absorption of
microwave radiation by an unpaired electron when it is exposed to a
strong magnetic field.
The unpaired electrons of free radicals in
coals are related to the condensed aromatic system.
This method has
provided only qualitative information on the aromatic coal structure.
For more information see Retcofsky et al. (1981).
(4) X-Ray Diffraction
Structural information is inferred by comparing the x-ray spectrum of
coal to x-ray spectra of known aromatic crystallites.
For more
information see Hirsch (1954), and Kwan and Yen (1976).
(5) Chemical Methods
Fluorination of aromatic rings (Huston and Studier, 1981) appears to be
the only reliable chemical method for measuring aromaticity.
(6) Optical Methods
Aromaticity is measured by comparing reflectance
those of model compounds (Van Krevelen, 1961).
of coal macerals
to
(7) Density Methods
The distance between C-C bonds is related to aromaticty as well as the
density. Empirical correlations between the density and aromaticity of
known compounds are used to measure the aromaticity of coal.
of heteroatom functionalities in coals are very scarce.
Many pyrolysis
the
scission
1981;
models either explicitly or implicitly assume that
of
decomposition
bridge
of coal
groups
1981a,b;
see Chapter
bridges may contribute
in
spectroscopic
have
bridge
studies
the
the
main
route
presence
thermal
Unger and
Suuberg,
Further degradation of thermolyzed
light gas production.
of
for
extended multiple independent parallel
6).
produced
types are present in
indicate
is
(Niksa and Kerstein, 1985;
Gavalas et al.,
reaction model,
75
evidence
the coal.
ethylene,
carbons) linkages (Deno et al.,
Various chemical and
that
a
wide
range
of
Oxidative degradation studies
butyl
ether
or
polymethyl
(>4
1981). Reductive alkylation studies of
Ignasiak and Gawlak (1977) support the presence of ether type bridges.
From C1 3 CP/MAS NMR and acetylation measurements, Yoshida et al. (1984)
report
higher
coals.
The presence of methylene, ethylene, aliphatic ether, and ether
linkages
has
concentrations
been
1978; Whitehurst,
inferred
of
ether
from
1978; Poutsma,
type
liquefaction
bridges
for
studies
1980; Benjamin et al.,
based on currently available data,
lower
(Mayo
1978).
no reliable quantitative
rank
et al.,
However
trends can
be observed on types and amounts of bridge groups among different coal
types.
Peripheral
groups
are
postulated
to
contribute
production and to influence the stability of nuclei.
latter postulate
indicates
groups)
that
are
comes from the
substituted
less
stable
thermal degradation.
there is
little
tars
(i.e.,
unsubstituted
tars
tars
light
gas
Evidence for the
recent work of Wornat
coal
than
in
(1988), which
with
when
peripheral
subjected
to
A further observation from Wornat (1988) is that
difference in
the stability of substituted tars of the
same
ring
number
(or
size).
It
is
convenient
to
subdivide
peripheral group into (1) oxygen containing species, e.g.,
C=O,
(2)
-OCH 3 ;
alkyl chains,
species, -NH2,
-COOH
the
-SH.
-CH3 ,
-C 2 H5 ; and (3)
-OH, -COOH,
N or S containing
Reasonable estimates of the amount of -OH
groups present in
different coals,
data of Yarzab et al.
(1980)
the
and
can be made respectively from
and Blom
(1960).
Based on product
yield data from coal pyrolysis,
pyrolytic water is
often hypothesized
to be produced from -OH
and carbon dioxide
from -COH groups.
groups,
Higher rank coals generally have less oxygen containing groups.
only work found in
Deno et al. (1981).
the literature
on peripheral
alkyl groups is
The
from
Their study shows that aryl-methyl and aryl-ethyl
groups account for approximately 1% and 0.1-0.3% of the total carbon
respectively.
Larger alkyl groups were not detected.
different coals were used,
no information on effects
be
on the
inferred.
Discussions
Although four
of coal type can
subgroup (3) will be omitted here
since the experimental program of this study excludes measuring gaseous
nitrogen and sulphur compounds.
3.1.2 Physical structure
Distribution of pore sizes
coal
pyrolysis.
is an important physical property in
In non-softening
coals,
the
pore
size
is
a
key
parameter in describing the transport of volatiles (see Section 3.4.3).
In softening coals, the
structure.
But,
'melting' destroys much of the
porosity information can still
initial pore
be valuable for these
coals. For example, Oh et al. (1988) utilized the macropore volume data
to
estimate
the
initial
number
density
of bubbles
transport model of softening coal pyrolysis.
in
their bubble
76
Literature data on coal pore structure (Gan et al.,
all three pore types, micropores
(12-300
A) and macropores
Table 3.1-4 gives initial
coal.
In
structure
general,
in
Gavalas
A - 1 pm),
are
present
in
raw coal.
pore-size distributions for various ranks of
micro-
and macropores
transitional pores
bituminous coals.
A), transitional pores
(pore dia. <12
appear
lignites and medium-volatile
coals, whereas
C),
(300
1972) show that
to dominate the pore
and low-volatile bituminous
are most abundant in high-volatile
Upon pyrolysis at relatively mild temperatures (500
and Wilks
(1980)
report that the coal retains its
general
structure, though they observed a slight increase of pores above 0.015
pm and an elimination of pores below that size.
A review by Suuberg
(1985) report that apparent porosity may increase from an initial value
of 10% to a final value of 50% at the end of pyrolysis, and that this
may be attributed to increases in micro- and macropores.
3.2.
Reaction chemistry and mass transport
3.2.1. Reaction chemistry
The reaction chemistry of coal pyrolysis
is
extremely complex.
A
complete quantitative description of reaction mechanisms remains to be
established.
However, recent
reviews by Stein
(1981,
1985) provide
valuable semi-quantitative analysis of reaction chemistry pertinent to
coal conversion.
This section discusses Stein's reviews in the context
of coal pyrolysis.
Free-radical reaction path
Free-radical
reactions
are
believed
to
be
the
primary
path
for
77
78
Table 3.1-4
coalsa.
Initial
pore-size
V|
Rank
S.mplC
Anthracite
LVII
NIVII
I IVA bituminous
I IVI bituminous
II VC bituminous
IIVC-bituminous
IV13 bituminous
IICV bituminous
I'SOC-80
PSOC-127
ISOC- 135
PSOC-4
PSOC-105A
Rand
PSOC-26
P10C-197
PSOC-190
PSOC-141
PSOC-87
PSOC-89
1.ignitc
I.ignite
L.ignitc
distributions
for
various
ranks
VE4
VV2'1
(cm'/g)
(cm '/g)
(cm '/g)
(cn'/gj
Id%)
'A%)
V,(%)
0.076
0.052
0.042
0.033
0.144
0.083
0.158
0.105
0.232
0.114
0.105
0.073
0.009
0.0 1
.0K0
0.00O
0.(XO
0065
0.027
0.061
0.013
0.122
0.004
0.(xx)
0.0)
0.057
75.0
73.0
61.9
48.5
29.9
47.0
41.8
66.7
30.2
19.3
40.9
12.3
13.1
0
0
0
45.1
32.5
38.6
12.4
52.6
3.5
0
0
11.9
27.0
38.1
51.5
25.0
20.5
19.6
20.9
17.2
77.2
59.1
87.7
0.014
0.016
0.017
0036
0.017
0.031
0.022
0.040
0.088
0.062
0.064
0.038
0.026
0.016
0.013
0.039
0.066
0.070
0.070
0.022
0.043
0.009
a Data from Gan et al. (1972); Table reproduced from Suuberg (1985).
b
VT = total porosity.
c Vi = macroporosity (300 A - 1 pm).
d
transitional porosity (12-300 A).
e V 2 = microporosity (4-12 A).
V =
3
of
thermal
this
decomposition
view
reactions
In
comes
of coal
from
the
(Stein,
general
control the pyrolysis
addition,
1985).
Supporting evidence for 79
observation
that
free-radical
chemistry of most organic substances.
resonance-stabilized
aromatic
and
hydroaromatic
units
derived from coal tars and liquids are formed and react readily at coal
decomposition temperatures (> 350 C).
organic
model
mechanisms
compound
studies
as possible pathways
Gavalas (1984) has reviewed some
that
suggest
to explain
concerted
the results.
reaction
But he also
points out that results from other studies have been explained solely
by free-radical mechanisms.
Thus, he concludes that "more experimental
work is needed to determine which of the pyrolysis reactions proceed by
concerted mechanisms and which by the more widely accepted free-radical
mechanisms".
In
this
study,
all
reactions
in
coal
pyrolysis
are
assumed to occur via the latter mechanism.
Applying gas-phase rate constants to condensed-phase reactions
Most of the reaction in primary coal pyrolysis occurs in condensedphase
(liquid
or
solid),
whereas
much
of
the
information as well as rate estimation methods
experimental
[e.g.,
kinetic
thermo-chemical
kinetic methods, Benson (1976)] are for gas-phase reactions.
Thus, the
question on the applicability of gas-phase reaction rates to condensedphase reactions needs to be examined.
Stein (1981) states that using gas-phase rate constants for liquidphase reactions is
a good approximation in
the absence of "significant
differences
in
products".
Such conditions closely resemble
solvent-molecule
during the 'liquid-like'
phase.
interactions
between
reactants
and
those of softening coals
For non-softening and softening coals
during
the
possible
solid phase,
"cage"
applying
effects,
where
the
the
analysis
is less valid due
restricted
product species from diffusing apart.
mobility
hinders
to
the
But Stein (1981) points out that
this effect can be significant only for bond homolysis reactions (Rl-R2
--
+
R1- + R2-).
If R1- and R2- radicals are relatively large in size, a
slower net homolysis reaction rate is
would enhance
the
rate
of
expected since the "cage"
reverse recombination
reaction.
effect
Stein's
evidence on the applicability of gas-phase thermochemistry and kinetics
to liquid phase is based on the comparison of equilibrium constants in
the two phases.
First consider unimolecular reactions of type A -- + B,
involving no change in the number of moles (i.e., An = 0),
"cage"
coal
and solvation effects.
pyrolysis
include
scission reactions
see below).
concludes
assumed
to
equilibrium
be
the
bond
homolysis
(within the
equilibrium
nearly
the
same
condition
can
be
transition-state theory.
without
Examples of such reactions pertinent to
and
free-radical
beta-bond
framework of transition-state
On the basis of available experimental
that
and
constants
for
the
related
for
gas
to
data,
theory,
Stein (1981)
these reactions
can be
and liquid phase.
reaction
rates
from
The
the
Represent a unimolecular homolysis reaction
as
AB -------
+
AB# -------
+
A + B
where # indicates the transition state.
Then,
the rate constant for
decomposition of AB is
kAB = (kT/h)K#AB
where K#AB =[AB]#/[AB] and (kT/h) is a universal frequency factor with a
value
of 6.3x10 1 2
s-1 at 300 K.
If
one assumes that the comparable
equilibrium condition between the two phases for "normal" species holds
80
for
transition-state
species,
the
rate
constants
for
bimolecular
in
the
two
phases
should also be about the same.
Similar
absence
analysis
of
products
substantial
state),
believed
to
the reactants
be
addition
reactions
important
in
the
reactants
and
(e.g.,
unity
are
coal
examples of
For
n-paraff ins),
However, when one of
methane, H-),
implying
Hydrogen
pyrolysis.
(e.g.,
and therefore k1 =kg.
exceeds
in
and diffusion limitations.
is a small molecule
considerably
reactions
between
involving two large reactant molecules
Stein concludes that K1 zKg,
Ki/K9
differences
and molecule-radical
reactions
reactions
made
solvation
(or transition
abstraction
such
can be
that
the value of
the
liquid
phase
reaction rate is much greater than that of the gas phase.
Under conditions where solvation and diffusion limiting effects are
non-negligible, using gas-phase reaction constants in coal pyrolysis is
not
strictly valid.
former
effect
is
However,
expected
at coal conversion
to be far
temperatures,
the
smaller due
to reduced hydrogen
bonding, polar, and charge-transfer interactions.
A decrease in the
latter effect is also expected since the rate of diffusion increases at
higher temperatures, whereas
the
rate
of recombination is relatively
independent of temperature.
To summarize, in the absence of significant solvation and diffusion
effects,
liquid-phase
reaction constants
can be adequately estimated
from gas-phase reaction kinetic data or from thermochemical estimation
methods.
reaction
involving
molecules
Applying
is
less
large
of
the
valid,
gas-phase
especially
molecules,
widely
rate
different
and
information
for
bond
bimolecular
size.
The
to
solid-phase
homolysis
reactions
reactions
involving
solvation
and
diffusion
81
effects
are
less
at
higher
temperatures,
and
thus
the
above
approximation becomes more valid.
Elementary reactions
1. Unimolecular reactions
Bond-homolysis.
estimated
Rates
relative
to
of
homolysis
bibenzyl
reactions,
homolysis,
(3.2-1),
(3.2-2),
whose
may
rates
be
and
thermodynamics are well established.
ki
R-X
R
-
+ X-
(3.2-1)
k2
4CH 2 -CH2 0 ---------- 2 qCH 2
Assuming
that
recombination
rate
(3.2-2)
constants
are
the
same
for
all
radicals in a given fluid, the rate of (3.2-1) is
k1 = k 2 exp[-(AH1 -AH 2 )/RT + (ASi 1 -ASi 2 )/R]
(3.2-3)
where Si is intrinsic entropy (excludes rotational symmetry), AH and
AS
are
obtained
1976), and
1985).
=
gas-phase
data or
estimation methods
(Benson,
exp(-66.8 kcal mol-1 /RT) s~1 in tetralin (Stein,
1016.6
Due to "cage" effects, the rate of bond homolysis declines with
increasing
0.4 ±0.1
k2
from
fluid viscosity
(Stein, 1981).
with
Also,
an approximate
an appreciable
relationship,
decrease in
k a
?-
the ethylene
bond strength, and hence an increase in the rate, is expected if benzyl
rings are more heavily substituted
(e.g.,
replaced by polyaromatic clusters.
Such effects can create a spread in
-OH,
-COOH),
or if
they are
bond strengths of "at least" 10 kcal mol- 1 .
Beta-bond scission.
X=Y + Z-)
An example of beta-scission reaction
in coal pyrolysis is
(-X-Y-Z
--
82
OCH 2 CH2 CH2 *
Preexponential
CH=CH 2
- - ---- - -----
factors
for
beta-scission
1014.5± 1 S- 1 compared 1015.5±1 S-1
energies
are
given by AH +
activation
energy
activation
energies
to
6-12
kcal
significantly
for
the
range
mol- 1
reactions
activation
reverse
addition
1-4 kcal mol-
P
(3.2-4)
are
typically
for homolysis reactions.
intrinsic
from
for
+ CH3 -
C-C
energy
(equal
reaction).
1
p
for
dissociation,
lower than those of homolysis
Activation
to the
Intrinsic
C-H dissociation
all
reactions.
of
which
are
The AH of the
reaction is most easily computed from appropriate nonradical reactions
AH(-XYZ--+XY+Z
where
D(Z-H)
) = AH(H-XYZ--+XY+Z-H)
+ D(Z-H)
and D(H-XYZ) represent bond
- D(H-XYZ)
strengths
of
(3.2-5)
Z-H
and H-XYZ
respectively.
2. Bimolecular reactions
Radical-molecule
reactions.
are
radical-molecule
examples
of
abstraction
reaction
heteroatom
factors.
series
IAHI
are
but
are
to
and
Rate
reaction
not
radical
addition
constants
for
the
thermochemistry
and
strongly affected by
steric
The Polyani relation applies for H abstraction for homologous
Thus,
donors.
, and for endothermic
constant a
ranges
15 kcal molparaffins;
1
and a
for
exothermic abstraction,
abstraction,
from 0 to
1.
Eact
=
E0
Eact
+ (1-a)|AH
Kerr and Parsonage
= 0.5 for methyl radical
=
EO
-
a
, where the
(1976) report E. =
abstraction from acyclic
Stein (1985) believes that this value of a may be considered
typical for H-abstraction.
whereas
reactions.
sensitive
(polar) effects,
of H
reactions
H-atom abstraction
involving
they
may
be
Generally, intrinsic activation energies of
carbon-centered
significantly
radicals
lower
for
are
~15
oxygen-
kcal
and
mol~ 1 ,
sulphur-
83
centered
Preexponential
radicals.
factors
are
typically
108.5
M~
s-1
for polyatomic species, and higher for smaller species.
Radical
abstraction
radicals
addition
to
reactions.
are
107.5±1
smaller species.
unsaturated
structures
Preexponential
factors
M~ 1 s 1 ,
and
they
hold
for
addition
polyaromatics,
these
for
tend
with
H-
carbon-centered
to
be
larger
for
For olefinic compounds, activation energies are often
lower for addition than abstraction.
to
also
compete
of
reactive
and an a of
reactions
(Stein,
The Polanyi relation is expected
radicals
~ 0.25 appears
1985).
At
C6 H 5 '
and
CH3~
to
to be a typical value for
coal
pyrolysis
temperatures,
redissociation of radical/molecule adducts is also expected to be very
fast.
Thus, for addition to be effective, rapid irreversible reactions
of the adducts are important.
Molecule-molecule reactions.
molecular
An example of
disproportionation, where
two
this reaction type is a
radicals
are
transfer of a H atom from one molecule to another.
formed
from a
This reaction may
serve as an important source of free radicals, particularly after weak
covalent bonds
have
decomposed, but
no
quantitative
information
is
available.
Radical-radical reactions.
disproportionation.
mode
of
radical
At
These reactions occur via recombination or
low temperatures,
termination.
Without
the former is the dominant
severe steric
effects,
Stein
reports that radicals recombine with a "near-diffusion-controlled" rate
constant
of
109-1010
radicals with weak
#-H
M-s-1.
The
latter
is
bonds (< 50 kcal mol- 1 ).
only
significant
for
84
85
3.2.2. Mass Transport
Mass
transport
limitations can significantly alter the pyrolysis
behavior
predicted
from considering just
Suuberg
(1985)
gives
transfer effects
only
the
the most
comprehensive
in coal pyrolysis.
transport
the chemical
decomposition.
review to
date on mass
The discussion below considers
of high molecular weight volatiles
molecular weight volatiles
(gases)
generally
(tars).
escape rapidly
Low
and are
relatively unreactive.
Internal mass transport
Distinctions need to be made between non-softening and softening
coals
as
their mode of volatiles transport are radically different.
Consider the non-softening case first.
Porosity characteristics among
different coal types have been discussed in Section 3.1.2.
gives
initial
Pores
are
micropores,
transport
pore-size
generally
classified,
transitional
differs
distributions
pores,
for
according
various
to
or macropores.
significantly
in
the
Table 3.1-4
ranks
of
pore
their
coals.
sizes,
as
The mode of diffusive
three
pore
regimes.
In
addition, depending on the regime, diffusivity values vary many orders
of magnitude (Fig.3.2-1).
Configurational
Since
micropores.
diffusion
the
size
is
of
the
dominant
pore
mode
diameter
of
and
transport
diffusing
in
tar
molecules is comparable in this regime, the configurational diffusivity
(assumed binary),
Dc,
is
expected to be activated,
and is
in
the form
of
Dc
=
D00 exp(-Ec/RT)
(3.2-6)
86
D
Regular
2
cm
S
..t TIl
10
I
r-
I 1iI
Ii
1 atm
I[
l
10am
Gases
K nudse n
10
quids
10-4
10-8Configurational
10-10-
10-'2-
Ln
1
10
100
1000
1
10
Angstroms, ym
Figure
3.2-1
Diffusivity
configurational, Knudsen, and
Froment and Bischoff (1979).]
versus
regular
pore
size
diffusion.
for
regimes
of
[Reproduced from
where Dco,
and Ec represent preexponential factor and activation energy
respectively.
pores,
a
regime
species is
is
collision
dominant in
pore
walls
transitional
and
diffusing
is of the form
DK,
4r/3 (2RT/rMA)1 / 2
"Regular"
where
between
Knudsen diffusivity,
(3.2-7)
the pore radius and
molecule.
regime
where
important.
DK =
where r
Knudsen diffusive transport is
MA
molecular
intermolecular
is
the molecular weight of diffusing
diffusion
collisions
dominates
are
in
macropores,
important.
a
Molecular
diffusivity, DM, is of the form
DM
=
DMo (T/273)1 .5 (1/P)
(3.2-8)
where DO is the reference molecular diffusivity at T = 273 K and P = 1
atm,
and
is
inversely
correlated
with
molecular weight of diffusing species
a
[e.g.,
Giddings method given in Reid et al.
fractional
power
see Fuller,
(1977)].
of
the
Schettler and
Convective
transport
contributions can also be important in this pore regime.
In
softening
"bubble"
coals,
growth/escape
transport.
liquid-phase
are
believed
molecular diffusion and volatile
to
be
the
Diffusivity in the molten coal, DL,
is
main
(Oh,
modes
(3.2-9)
is a constant estimated to be
~10-5 cm 2 /s for molten coal, and
y is
the viscosity of the coal melt with a minimum value of
(Oh,
1985).
Based
tar
1985)
2 3
DL = CD T/ps
/
where CD
of
on the particle
radius
length
scale,
=104 poise
Oh
(1985)
asserts that the characteristic time for liquid phase diffusion is too
slow to explain experimentally observed tar release rates in pyrolysis.
She proposes
transport
an alternative mechanism whereby most
occurs
via
the
following
sequential
of the volatiles
steps:
(1) volatiles
diffuse into many small bubbles distributed throughout the molten coal,
87
(2)
bubbles
released
grow as more volatiles
outside
the
particle
diffuse-in,
when
a
and (3)
growing
volatiles
bubble
are
bursts
upon
contacting the particle surface.
Another
postulated mechanism for volatiles
coals proposes
transport
that most of the bubbling phenomena
tars are evolved (Griffin, 1988;
Hsu, 1988).
in
softening
occur before
most
Quantitative information
on the degree of overlap between the two processes is currently being
gathered by the two investigators
in
computing
the
transport, the
than
the
resembles
characteristic
radius.
cenospherical
Pittsburgh Seam bituminous
pressure,
time
'shell' thickness
particle
a
Implicit in this mechanism is that,
Griffin
is
The
scale
for
intra-particle mass
the more appropriate length scale
particle
shell.
For
shape,
~40
pm
radius
coal pyrolyzed at 1000
(1988)
estimates
the
approximately 20% of the particle radius.
based on the shell thickness
is
comparable
after
bubbling,
particles
of
C/s at atmospheric
shell
thickness
to
The diffusion time
be
scale
to observed characteristic
times for pyrolysis at the heating rate of approximately 1000 C/s.
External mass transport
For non-softening
coals,
molecular diffusion and convective
flow
are the main modes of external transport, whereas for softening coals,
surface evaporation is also important.
The molecular diffusivity is of
the
transport
[Eq.(3.2-8)].
flux
convection
same
form
comparing
rates
diffusive
flux,
as
that
of
as
in macropore
pure
in
diffusive
Eq.(3.2-10),
and
Suuberg
(1985)
From
enhanced
concludes
that
convective enhancement is insignificant for small particles (~C100 pm).
pure diffusion/convective enhancement =
4/[l-exp(-
)]
(3.2-10)
88
4
where
= NR/DVCV , N = total
molar surface flux,
R -
particle
radius,
DV = diffusivity of tar in the surrounding gas, and CV = molar density
of the gas phase.
Experimental
factor
in
Suuberg
evidence
determining
and
is
extracts
produced
lower
lower
range
than
than
evaporated.
mechanism
(Unger
to
from
be
the
that
of
is
a
transport
is
1984;
of
HVB
(B)
and
similar to
that
of
tar
based
and
For
softening
the MW
(C)
on
(A)
important
Suuberg,
tar).
they observed
extract,
These observations
an
Suuberg,
precursor
vacuum-tar,
very
tar
is
measurements
pyrolyzing
of the
evaporation
of
and
of 727 to 951 K,
that
atmospheric-tar
surface
(MW) distribution
speculated
temperature
is
the
co-workers'
molecular-weight
(which
that
on
1985)
extract
tars
and
in
the
coal
the MW of the tar
of atmospheric-tar
the
MW
the same
is
distribution
tar
of
that was
re-
are used to support the view that tars
from softening coals are produced from selective evaporation of lighter
components of extractable materials, and thus the rate of tar escape is
significantly
influenced
by
evaporation.
However,
the
above
observations also support the view that the internal mass transport may
be the rate controlling step.
fact
that
lighter
species
For example, consistent with
diffuse
through the
molten
they tend to have larger diffusion coefficients, e.g.,
diffusion coefficient based on hydrodynamic
to
the
radius
of
the
solute
molecule
theory is
(Reid et
(A) is the
coal
faster
as
the liquid-phase
inversely related
al.,
1977).
The
observation (B) can also be explained by the fact that atmospheric-tar
is
more
exposed
1984).
The
internal
mass
to
secondary
observation
transport
(C)
since
reactions,
is
difficult
sample
which
to
conditions
decrease
interpret
in
MW
in
(Serio,
terms
of
the re-evaporation
89
90
experiment are likely to be different from those of the molten coal.
Also observed in the molecular weight studies was that tars from a
wide variety of different coals (1 lignite, 2 HVB and 1 LVB coals) show
"somewhat"
similar
interpreted
as to
for non-softening
MW
This
distributions.
observation
suggest that perhaps evaporation
coals.
But, the
important
and a fractional power of
offers an alternative
1977),
also
is
been
relationship between the
inverse
molecular diffusion coefficient [Eq.(3.2-8)]
tar MW (Reid et al.,
has
explanation for the
observed behavior.
3.3. Experimental studies
Coal type, temperature-time history, reactor pressure, and particle
size
are
the
four
main
reactor variables
in
describing the influence of these variables on
is
further
divide
the
distinguish
secondary
rate
processes
and
into
In
pyrolysis behavior, it
secondary
intra-
and
reactions,
and
extra-particle
the
secondary reaction residence
The first two variables affect both the
reactions,
secondary
and
pyrolysis.
are coupled to secondary reactions
of transport determines
time for reactive volatiles.
primary
primary
Mass transport effects
reactions.
as
to
convenient
coal
whereas
the
latter
two
variables
affect mainly the secondary reactions.
3.3.1. Effect of coal type
Figure
3.3-1
distributions
compares
obtained
operating conditions,
pm dia.)
literature
from
different
data
on
coal
product
types
yields
under
where thinly spread small coal particles
and
similar
(~C 100
were pyrolyzed in a screen-heater type reactor under rapid
heating
(~1000
atm)
(~C10- 3
volatiles
to
data
69
atm.
Table
in
for
C at pressures
opportunity
4.2-1 gives
are
higher
The
(1)
the
carbon
clear.
from
(2)
trends
'vacuum'
analysis
observed
methane
and
secondary
of
the
from
and pyrolytic
tar
coals
the
vacuum
water
yields
yields
carbon content of 85-87
for other hydrocarbon
91
of reactive
extra-particle
trends
oxides
maximum for coals with the elemental
(3)
for
elemental
general
rank coals,
wt% daf respectively;
ranging
The rapid dilution and quenching
Fig.3.3-1.
(Fig.3.3-la)
decrease
to < 1000
presented minimal
reactions.
compared
C/s)
reach
a
and 78-86
gases
are
less
More experimental data are needed to establish trends at higher
pressures.
Several
investigators
have
yields to coal properties.
volatile
elemental
yields
composition
conditions
from
coal
total
be
of
and
biomass
closely
the
correlated
coal.
correlations.
pyrolysis,
The
in
the
and tar
coal,
but
to
the
yields
coal
Peters
carbon monoxide,
ratios
relate pyrolysis
product
Neoh and Gannon (1984) observed that total
representative of pulverized
volatiles,
elemental
can
attempted to
reflectance
were
obtained under
combustion.
Using data
(1984)
graphically
yields
to values
established
and
no
compared
of certain
quantitative
Results from Suuberg (1977) suggest that carbon dioxide
and pyrolytic water yields are linked respectively to the carboxyl and
hydroxyl group contents in the
measurements,
weight
olefin
polymethylene
Calkins
yields
et al.
are
coal.
(1984
Based on CP/MAS
a,b,c,d)
proportional
structures present
in
the
report
to
the
parent
13
that
low molecular
amount
coal.
C NMR spectra
of
long-chain
Neavel
et
al.
(1981) correlated liquid yields, from pyrolysis of vitrinite samples in
a
packed-bed
type
reactor,
to
the
elemental
hydrogen
and
organic
40
92
35
30
L
25
0
20
0
w
15
a-
10
5
A
-8
0
70
+
74
78
82
+0
86
90
ELEMENTAL CARBON CONTENT (WT7. DAF)
40
I b:69atm
b: 1atm
-
35-
30 -
U-
0
25 -
20 -
-J
w
a-
15 -
10 -
5 -
0
0
Figure
i
|
70
74
CO
3.3-1
different
+
78
74
78
ELEMENTAL CARBON CONTENT (WT% DAF)
C02
0
H20
A
CH4
Comparison
coal types:
70
of
product
yields
and
X
distributions from
[Data from Loison
(a) vacuum, (b) 1 and 69 atm.
and Chauvin (1964) and Suuberg (1977).]
TAR
sulphur contents of the coal (Fig.3.3-2).
The effect
of coal type on volatiles
Kobayashi et al.
unclear.
93
(1977)
evolution rate is
currently
showed that total weight loss data
from both Montana lignite and Pittsburgh Seam bituminous
coal can be
described with the same set of model parameters, implying no apparent
coal-type effect
The data were obtained at temperatures
(Fig.3.3-3).
ranging from 700 to 1827 C, and heating rates from 180 to 106 C/s.
The
study assumed a competing reaction model consisting of two first-order
single-reactions
ki
|------+
Coal ----
------
+
Volatile 1 (al)
Volatile 2 (a 2 )
+
Residue 1 (1-a)
+
Residue 2 (1-a 2 )
k2
ki and k 2 are pseudo Arrhenius rate constants,
energy of 2x10 5 s-
factor/activation
kcal mol- 1
respectively;
al and a 2
1
/25
with a pseudo frequency
kcal mol-
1
, and 1.3x107 s~1/40
are the asymptotic volatile yields
for each of the two reactions with values 0.3 and 1 respectively.
The
rate
of
total
weight
investigated by Anthony et al.
heating
rates
(400-1100
Kobayashi et al.
example,
C,
(1977).
comparing the
loss
for
(1974),
the
two
coals
were
also
but at lower temperatures
C/s)
102-104
than
those
and
employed
Their results show some differences.
temperature at which the weight loss
by
For
rate is
maximum shows a difference of 60-85 C over the range of heating rates
studied.
The
temperature
is
reached
earlier
for
Pittsburgh
Seam
bituminous coal than for Montana lignite.
For
individual
products,
the
evolution
significantly among different coal types.
rate
appears
to
vary
For example comparing the
94
17.5 LIO.
WT.
15.0
150
DMMF
12.5
0
0 0 00
00
000
0
0
0
00
10.0-
z1
7.5 -
*-
0 8
5.056
242526 27
7 8 9 10 1112131415 16 17 18 1920 22
PYROLYSIS LIOUIDS (ESTIMATE)
Figure 3.3-2 Comparison of experimental and predicted pyrolysis liquid
yields from Neavel et al. (1982).
I
80
i
70
LIGNITE
0 21C
0 19.
LO~
so
50
0
O
oO
0
4030 -
015
o
V 12
0
0
ZAl1
00
o
-i
310
0
0
50
RESIDENCE
80
150
100
200
TIME ( ms)
-
70 60LL:
0
-J
5040
30
20
10
0i
0
50
100
RESIDENCE
150
TIME
200
(ms)
3.3-3
Comparison of calculated weight losses with experimental
results. [Repr oduced from Kobayashi et al. (1977).]
Figure
rate
of
hydrogen
and
pyrolytic
water
evolution
lignite and a Pittsburgh Seam bituminous
between
a
Montana
coal using kinetic data from
Suuberg (1977), revealed respectively over 100 and 200 C difference in
the temperature at which the maximum rate occurs.
earlier
from
bituminous
the
lignite,
coal.
whereas
Differences
water
in
other
Hydrogen evolved
evolved
earlier
from
(CH4 , CO, tar)
products
the
were
somewhat less, below 100 C.
Some
precautions
apparent pyrolysis
need
to
rates.
noted
in comparing
literature data
on
First, direct comparisons of experimental
rate data are valid only when the temperature-time history of the data
sets
being
compared
final temperatures,
are
completely
holding times,
identical,
i.e.,
heating
and cooling rates.
rates,
In many cases,
such conditions are not fully satisfied among data sets from different
investigators.
Second, comparing
model
predictions
under
identical
temperature-time histories is valid only within the range of conditions
where model parameter values are valid.
commonly used global models
model,
is
such as
For example, predictions from
the
first-order
single-reaction
only valid over a narrow range of heating rates from which
the model parameters
are derived
(Howard et al.,
1987b).
Predictions
from the multiple independent parallel reaction model are valid over a
wide range of heating rates only when the range is
covered by the data
from which the model parameters are derived (Ko et al., 1988a).
Discerning coal-type
rate
effects
is more
data reported in the literature
difficult
from
low heating
since they often employ large
sample sizes in a packed-bed type reactor (Juntgen and van Heek, 1977;
Weimer
and
Ngan,
1979;
Campbell
and
Stephens,
1976).
As
discussed
below, significant contributions from extra-particle in-bed secondary
95
96
reactions
reactors
are
expected
in
such
reactors.
Results
from
fluidized bed
are also severely affected by in-bed secondary reactions, and
thus are not discussed here.
3.3.2. Effect of pressure
Table 3.3-1 shows that for both Montana lignite and Pittsburgh Seam
bituminous
coal,
tar,
more
This
behavior
char,
rapid pyrolysis
and generally
has
been
under higher pressures
more
attributed
gaseous
products
to
greater
particle secondary reactions of tar --Contributions
from
extra-particle
+
the
produces
(Suuberg,
extent
less
1977).
of
intra-
gas + char at higher pressures.
secondary
reactions
are
expected
to
be negligibly small because of rapid dilution and quenching of reactive
volatiles
in
screen-heater type reactors.
Figure 3.3-4 shows the effect
production
for
information
for
cases,
no
Montana
tar
lignite
from
appreciable
of pressure
and
Pittsburgh
influence
on the rate of volatiles
Fig.3.3-5
provides
Seam bituminous
of pressure
is
similar
coal.
seen for
In both
temperatures
below 700 C; but above 700-800 C, noticeable effects of pressure on the
apparent rate of product release are observed.
total
volatiles
temperatures
from
(Figs.
overall activation
conditions
for
gas
3.3-4a
pressure
and
runs
3.3-5),
energy behavior
that is
'leveling-off'
indicates
explained by
a
rates
greater
at
extent
different
of
tar
a
lower
at
under
An opposite behavior
pressures
secondary
lower
apparent
typically encountered
of greater mass transport resistance.
production
pressures.
higher
The yields of tar and
(Fig.3.3-4b,c)
reactions
at
is
higher
Effect of pressure on pyrolysis product yields from
Table 3.3-1
Montana lignite and Pittsburgh Seam bituminous coal. [Reproduced from
Howard (1981); data from Suuberg (1977).]
Average particle diameter, 74 pm; heating rate, 1000 C/s; peak
temperatures, lignite, vacuum and 1 atm, 900-1035 C, all other cases,
850-1070 C; holding times at peak temperature, lignite, vacuum and 1
atm, 0.0 s, all other cases, 2-10 s.
Yield, wt % of Coal (as-received)
Bituminous
Lignite
Product
CO
CO 2
H 2O
H2
CH,
C2 H 4
C2 H,
CIH+C:,Hm
Other HC gases
Light HC liquids
Tar
Char
Error (loss)
4
I atm
69 atm
6.1
7.6
17.7
-0.94
0.43
7.1
8.4
16.5
9.0
10.6
13.4
0.21
0.46
0.60
0.81
Vacuum
6.9
55.2
97.0
3.0
Vacuum
0.50
-r
1.3
0.56
2.5
0.55
2.0
1.4
6.8
0.75
1.6
0.45
0.18
0.37
[ 0 .4 7
0.17
0.38
0.21
1.1
0.44
0.71
0.98
1.6
L I
5.4
58.7
99.5
0.5
2.8
59.0
99.7
0.3
31.9
48.5
97.1
2.9
4
I atm
69 atm
2.4
1.2
7.8
2.5
2.5
0.83
9.5
-3.2
0.46
0.51
1.3
1.3
2.4
0.89
0.71
1.6
2.0
1.0
23.0
53.0
97.2
2.8
1.7
12.
62.4
97.0
3.0
a 6.6 x 10-' atm He.
* Includes coal moisture (lignite, 6.8%; bituminous, 1.4%); may include some H2 S.
*Not measured.
97
48
0
44
-
98
40
G
7
T
T
36
TT
o
32
(a)
28
20
8--
>
16 -O
w
12 ..
T
T
24
z
T
0
o
.
CD
o
oT
coU
0
8
w
0
44~
1
12
0 0
0
.(b)
0
-
2-r-
CO
01
6
a
UJ
I-
r...
Wi
C00
2
S0
2
4
0
00
0
)
12
12
8-
*14
_r
0-6
-.
0.4
UA
H-0.2
0
0
I
200
400
PEAK
600
800
TEMPERATURE,
1000
-C
Effect of pressure on product yields from lignite
Figure 3.3-4
Pressures: 1 atm (single points
pyrolyzed different peak temperatures.
and solid curves) and 6.6x10-5 atm (points in circles and dashed
Heating rate - 1000 C/s. Products: T - total (i.e., tar, all
curves).
other
HCs,
H 2 0,
CO 2 ,
and
CO);
open
circles
=
CO;
*
triangles - CH4 ; solid circles = total HCs, including tar.
from Howard (1981); data from Suuberg (1977).]
=
CO 2 ; open
[Reproduced
99
> 28 U
M 24 -
-
Iatm
He
69atm
He
0
0
0
20
-j
816
0
,
12
00
40
0
0
200
600
400
800
1000
PEAK TEMPERATURE,*C
Figure 3.3-5
Effect of pressure on yield of tar from Pittsburgh Seam
bituminous
coal pyrolyzed at different peak temperatures.
Helium
atmosphere; heating rate, 1000 C/s; average particle
diameter, 74 pm.
[Reproduced from Howard (1981); data from Suuberg (1977).]
55
m, = 0.5*C/s
* Experimental P, = 0.1 MPa
9
O Experimental P, = 5 kPa
(0
SExmperimental P, = 1.0 MPa
-
50-
10-5
3
6 10-'
3
6 10-3
3
6 10-2
3
Model predictions, K0 4 = 2.7 X 104 s'-
6
particle diameter (m)
Figure 3.3-6 Total volatiles yield versus particle size for a German
lignite. [Reproduced from Bleik et al. (1985).]
3.3.3. Effect of particle size
100
Table 3.3-2 shows that, on average for a Pittsburgh Seam bituminous
coal, rapid pyrolysis of larger particles generally produces less tar,
more gas,
and more char (Suuberg,
1977).
Figure 3.3-6 shows a similar
trend for a German lignite pyrolyzed at a low heating rate
al.,
1985).
This behavior has again been attributed to the enhancement
of intra-particle secondary reactions
of tar --
particles.
explanation
Consistent
increasing particle
range
of
(Ko
with
this
size should not have
pyrolysis
negligible
(Bleik et
temperatures
et al.,
1988a),
where
+
gas + char for larger
is
the
view
significant effects
the
secondary
that
in
the
reaction
is
~C600 C for Pittsburgh Seam bituminous
coal (Serio, 1984).
No experimental
data are reported on how the rate of pyrolysis is
influenced by the particle
size.
particle
size
secondary
pyrolysis
rate behavior with changing particle
affect
the
However,
since both pressure
reaction
residence
size is
time,
and
the
expected to be
similar to that reported for varying pressure (see Figs. 3.3-4 and 3.35).
Possible non-isothermality
macerals
for different particle-size cuts
particle-size data.
criteria
for large particles
derived
by
and segregation of
complicate
the analysis of
The former complication can be checked by a set of
Hajaligol
et
al.
(1988),
whereas
the
latter
complication is currently being investigated by Griffin (1988).
3.3.4. Effect of temperature-time history
The
effect
most often
of temperature-time history
studied by varying the heat-up
on pyrolysis behavior
is
rate
In
of
the
sample.
Table 3.3-2
Effect of particle
size on pyrolysis product yields
Pittsburgh Seam bituminous coal.
[Reproduced from Howard (1981);
from Suuberg (1977).]
Heating rate, 1000 C/s; peak temperature, 850-1070
peak temperature, 3-10 s; pressure, 1 atm.
C; holding time at
Yield, wt % of Coal (as-received)
Product
CO
CO2
H2 0#
H2
CH 4
CH 4
CH,
C3's
Other HC gases
Light HC liquids
Tar
Char
Error (loss)
Number of runs
53-88 sim
(avg., 74 sm)
<300 s.m"
300-830 sm
830-990 sm
3.2
3.0
1.2
1.3
1.3
2.4
23.0
53.0
97.2
2.7
1.1
5.4
--'
2.9
1.0
0.50
0.92
1.4
2.5
24.2
57.1
99.7
5.3
-'
3.0
1.1
0.55
0.84
1.1
2.6
21.3
56.5
%.7
7.2
0.99
3.2
1.3
0.63
1.1
1.2
2.7
18.4
55.8
96.8
2.8
20
0.3
1
3.3
2
3.2
3
2.4
1.2
7.8
1.0
2.5
0.83
0.51-
1.3
830-990 sm sample ground to pass 29 7 -sm sieve.
*Includes coal moisture (1.4%); may include some H,S.
Not measured.
from
data
101
general,
one observes higher total volatiles yield under rapid heating
conditions
[> 100 C/s,
Loison and Chauvin
(1964)]
heating carbonization conditions
[=
0.01-1 C/s,
(1965)].
of
heating
However,
volatiles
yield
the
is
effect
difficult
to
discern
compared
Peters and
rate
from
to slower
on
the
the
data
Bertling
change
in
because
of
interferences from other experimental conditions, specifically the size
and extent
of dispersion of the
typically
employ
small
carbonization experiments
beds;
thus,
secondary
in
sample.
sample
reactions
sizes
employ much
the latter set-up,
contribute
Rapid heating
thinly
larger
experiments
spread,
sample
sizes
whereas
in packed
volatiles lost due to extra-particle
to
the
decrease
in volatiles
yield
(Howard, 1981).
Even
when
minimized,
the
the
extent
of
intrinsic heating-rate
discern due to interferences
Experiments
extra-particle
using
thinly
secondary
effect
is
from intra-particle
spread
small
coal
still
reactions
is
difficult
to
secondary reactions.
particles
(=70
pm dia.)
under 1 atm reactor pressure report that the yield of total volatiles
as
well
as
individual products
rates between
350 to 15,000
C/s
is
generally
(Anthony,
independent of heating
1974;
Suuberg,
1977).
The
observations were
made for both Montana lignite and Pittsburgh Seam
bituminous
coal.
An exception
rec'd)
the
in
total
weight
heating rate was raised
contrast,
similar
is
loss
the
for
from 650-750
studies
slight increase
the
bituminous
to 10'
done under vacuum
volatiles yield as the heating rate is
C/s
(=2.5 wt% as
coal
(Anthony,
when
1974).
the
In
show noticeably higher
increased
volatiles yield from a Pittsburgh Seam bituminous
(Niksa,
1981);
total
coal increased from
41 to 52 wt% dmmf as the heating rate was raised from 100 to 10,000 C/s
102
at
1.3x10-4
atm pressure and
1000
C
vacuum data show a slight increase
but no further from 3000 to 104
final
in
C/s
temperature.
going from 650-750
(one
Anthony's
to 3000 C/s,
data point at this heating
rate).
The
different
pressure
may
explained
if
heating-rate
first appear
the
results
at
contradictory.
production
of
vacuum
But,
and
the
atmospheric
results
'primary' volatiles
is
can be
enhanced
at
higher temperatures, or heating rates if the heat-up is continuous. The
extent to which primary volatiles further react
(secondary
depends on both the temperature and mass transport rate.
the transport
negligible
rate is
sufficiently
to
over the range of heating rates studied.
enhance
resistance
secondary
is
primary
pyrolysis.
expected to be
reactions
to
small,
obscure
Under vacuum,
fast for secondary reactions
data may be indicating an intrinsic heating-rate
here
reactions)
the
At
1
Thus, the vacuum
effect
atm,
the
--
postulated
mass
but may be still
intrinsic
to be
transport
sufficient
heating-rate
for
effect.
Thus, in this case, one may not observe the heating-rate effect because
an increase in primary volatiles production at higher heating rates is
'off-set' by a greater extent of secondary reactions.
suggests
a
competing
reaction
scheme
producing reactions and 'low-temperature'
between
This explanation
primary
char forming reactions.
a reaction scheme has been proposed by Kobayashi et al.
3.3.1)
and Niksa and Kerstein (1985)
volatiles
(Section 3.4.2),
(1977)
Such
(Section
but their models
are inadequate to fully explain the observed behavior because they do
not
explicitly
include
mass
transport
descriptions.
A
similar
competing reaction concept was adopted in formulating the extended MIPR
model
(Section
6.2.1),
which
includes
an
explicit
mass
transport
103
description.
Section 6.2.2 shows
that quantitative predictions from
the model are consistent with the experimentally observed heating-rate
effects reported by Anthony (1974), Suuberg (1977), and Niksa (1981).
3.4. Modeling studies
Numerous
exist.
coal
pyrolysis
models
of
various
complexity
currently
These models can be broadly classified into (1) global models,
(2) detailed chemistry models, and (3) models with explicit description
of mass transport.
Each of these three classes of models are discussed
below.
3.4.1 Global models
The multiple independent parallel reaction (MIPR) model (Hanbaba et
al.,
1968)
1978)
are
and the functional group
(Solomon and Colket,
two commonly used coal pyrolysis global models
the evolution of
1988a)
(FG) model
(gases:
volatiles
individual products
Weimer and Ngan,
(Anthony et al.,
conditions
where
the
1979;
1974;
effects
(tar:
of
Serio,
Serio et
Sprouse
to describe
1984;
al.,
Ko
1987),
et al.,
and total
and Schuman, 1981).
physical
transport
Under
processes
and
secondary reactions are relatively unimportant but not negligible, both
models approximate the complex chemical decomposition and any transport
effects
by
global
first-order
uniformly throughout the particle.
be used to
that
decomposition
include
mass
transfer.
treated explicitly
or implicitly,
predict
product
reliable
occurring
In addition, both models can also
represent only the chemical
explicitly
reactions
decomposition in
Whether
a successful
evolution
rates
mass
descriptions
transport
model must be able
over
a
wide
range
is
to
of
104
temperature-time histories.
Ko et al.
two
(1988a)
models
to
This
by comparing
experimental
important criterion was examined by
tar evolution rates predicted from the
data from Pittsburgh
Seam bituminous
over a wide range of heating rates (0.05 - 1000 C/s).
coal
The experimental
conditions are limited to 1 atm pressure and small particle sizes where
mass transport limitations are small.
The
study
found
evolution rates
that
the
MIPR
over the range
model
of heating
can
reliably
predict
tar
rates covered by the data
from which the rate parameters used in the model were obtained, but
generally not at heating rates outside this range.
applicability is
Thus, the range of
substantial when the rate parameters are fitted from
data collected at two or more widely different heating rates.
For the
FG model, several sets of parameter values have been published (Solomon
et al.,
al.,
1981; Solomon et al.,
1982; Solomon and Hamblen, 1985; Serio et
1987) without always showing critical comparisons against data and
without providing guidance as to which values are preferred for a given
However,
set of conditions.
regardless
of which of the published sets
of parameter values is used, tar evolution rates predicted from the FG
model
do
not
generally
agree
well
especially at higher heating rates.
with
the
experimental
data,
Also large discrepancies are found
between experimentally observed maximum tar yields and those predicted
by the FG model.
A reason
for the poor performance
of the FG model
against Pittsburgh Seam coal tar evolution data may be that it employs
the same
rate parameters
tar for all coal types,
non-softening.
kinetics
of
The
product
for the evolution of all products
regardless
assertion
evolution
of whether the coal is
that coal
(Solomon
including
softening or
type has no effects on the
et
al.,
1982;
Solomon
and
105
Hamblen, 1985), which forms the basis of the FG model, is investigated
in Chapters 5 and 6.
The
first-order
widely used under
processes
and
approximates
effects
model
conditions where
secondary reactions
the
by
single-reaction
complex
a single
is most useful
is
another
the effects
global
of physical
are relatively small.
chemical
global
model
decomposition
and
model
transport
This model
any
transport
first-order decomposition reaction.
in applications where
minimizing
The
computational
effort is important such as in large combustion or gasification models
that
fully
describe
fluid
mechanics,
heat
and mass
transport,
reaction kinetics;
and in comprehensive devolatilization models
explicitly
include
the
chemistry,
and multicomponent mass
However,
phase environment.
different
rates.
set
of
complex
rate
decomposition
transfer
and
that
and secondary reaction
or
liquid
the model has a major weakness in
that a
parameters
is
in a gaseous
required
at
different
heating
Thus, for a given set of rate parameters, the applicability of
the model
is
confined to a narrow range of heating rates.
A method
developed by Ko et al. (1988b), extends the applicability of the model
over a wide range
of heating rates.
The two rate parameters
in the
model, a preexponential factor and activation energy, are derived in
the form of heating-rate dependent
heating-rate dependent
weight
functions.
rate parameters
Predictions using the
were compared with the
total
loss data from devolatilization of Montana lignite over heating
rates from ~0.1 to 10' C/s, and were found to agree well with the data.
Two typing errors in the published paper (Ko et al.,
1988b) need to be
noted:
(1) in Table 1, the printed a values are lower than correct
values
by
a
factor
of
10,
e.g.,
1.32
kcal/mole
should
read
13.2
106
kcal/mole,
and (2)
the y-axis label in
Fig.2 should read wt fraction,
not wt%.
The first-order competitive reaction model proposed by Kobayashi et
al.
(1977)
has
distinguishing
reaction
already
feature
mechanism,
been
discussed
of this model
it
is
able
in
Section
is that,
through
predict
the
to
3.3.1.
the
A
competing
observed
greater
volatiles yield at higher heating rates/temperatures without having to
adjust model parameters.
of
predicting
total
Figure 3.3-3 shows that the model is capable
weight
loss
data
from
a
Montana
lignite
and
Pittsburgh Seam bituminous coal over a wide range of temperatures and
heating rates.
3.4.2 Detailed chemistry models
The work of Gavalas and co-workers
(1981a,b) provides a detailed
approach to model the reaction chemistry in coal pyrolysis.
represented
as
a
collection
of
reactive
functional
Coal is
groups
whose
concentrations are estimated from elemental analysis and NMR data.
chemical
the
changes
basis
studies
of
are described by 42 elementary reactions
chemical
(Table 3.4-1).
theory
But,
and
as
information
from
selected on
model
pointed out by Gavalas,
The
compound
this large
reaction set is by no means exhaustive as it excludes certain reactions
that are known to be important at high temperatures (> 700 C),
coals of high oxygen content.
and for
For example the reaction set omits the
formation of CO and H2, which are speculated to evolve from phenolic,
ring or ether oxygen,
respectively.
CH2 -Ph',
and from dehydrogenation of hydroaromatic rings
It also neglects dissociation of ether type bonds (Ph-0-
Ph-0-CH 3 ), which are believed to be major constituents in low-
107
108
[Reproduced from
Elementary reactions of coal pyrolysis.
Table 3.4-1
Gavalas et al. (1981a).]
1
no.
reaction
X
no.
Bond Dissociation Producing Two Radicals
H
2
CH 3
3
C2 H5
4
-X
Ph-
21
Ph--
+ X.
Ph-
-Ph'
2
CH2
Ph-C- + -C-Ph'
Ph -C
-
P
I
+ -Ph'
23
24
25
26
27
28
29
H
CH 3
C2H,
H
CH,
C2 H5
Bond Dissociation Producing One Radical and
One Double Bond
7
Ph-C-CH3 -
8
Ph-C-CH2 CH3 -
Ph-C=CH 2 + CH3
9
Ph-6-CH
2 CH3 -
Ph--C=CH 2CH3 + H-
Ph-C=CH 2 + H-
I
10
/
H
h
CH
I
x
CH 3
13
C 2 H,
14
Ph-C-C-Ph'
Ph-
Ph-C=
-C-Ph'
18
>dH + Ha -+>CH,
(# radical)
+ a radical
31
CH,
Ph-C-
32
C 2 H,
(substituents on a carbon can be
33
H
Ph/
H
Ph-C-X
.CH 2-CH
34
35
CH 3
C 2H,
36
H
37
CH 3
2 -C<
38
C2H
+ X-
Ph-C
-
-- Ph' + X-
CH 3
C 2 H,
(9 represents -CH
hydroaromatic
|
2
or >CH of)
structure
Ph--C
Ph-C-C-Ph' -
40
+ X- -
PhX +
-
H, CH 3 , C2 H,)
C 2 + X. C-CH2
Ph
H2
C.CH2
Ph--C-Ph' -
Ph--C
Ph
H
'.-CH
I
2
-
+ *C-Ph'
I
P-h-C -Ph' + X- -
I
PhX +''-h
(substituents on a carbon can be
H, CH 3 , C2 H,)
Phenolic Condensation
Ph-OH + HO--Ph'
H 20
Ph-OH
+ HC--Ph'
-
--
Ph-O-Ph' +
Ph-C-Ph'
I
I
Formation of Carbon Oxides
+ 'Ph'
'C--CH
2
Ph-COOH -PhH
+ CO 2
0
Ph
I
(substituents on a carbon can be
H, CH 3 , C2H,)
| +I
42
20
XH + # radical
-+
Addition-Displacement
41
19
X- + Hp
H
39
17
X. + Ha -+ XH + a radical
30
Ph-CC -Ph
-
CM2'C<
16
-II
Ph-C--Ph'
H
12
15
I-
PhC CPh' -
+ H.
C I
I
11
+ CH2=C2
-
I
CH2
Ph
Ph-
-
Hydrogen Abstraction
-CH
-
I
--C% 2 6M2
a Radical Recombination
Ph
CH2 -C CH2
Ph-- -C-Ph
6
reaction
22
Ph
5
Ph--.
-
X
H2
Ph-C-CM 2
-Ph6H
2
4' CO
+
H2 0
rank coals (e.g., lignites).
Testing
the
109
capability
complicated by mass
of
the
proposed
transport effects.
reaction
they
have
is
to their larger sizes
attached
effect
reactivities
Such assumption
coefficients.
tar due
low
to
the molecule's
in a simple
relatively
condensed
large
diffusion
and reactive peripheral groups
nucleus.
manner, Gavalas assumes
molecules
the
A)
aromatic
is Xr, where
in
and
CH4 , CO and H20
less valid for large molecules like
(~10
generation of tar
is
Gavalas assumes no transport
limitations for "small permanent molecules" such as H2,
since
scheme
r
is
phase
To
account
that the
for
this
actual rate of
the rate of generation of tar
and
X
is
an
empirical
parameter
adjusted to fit the experimental tar data.
Figure
3.4-1 shows a satisfactory
and experimental
bituminous
weight loss,
coal.
agreement between the simulated
tar yield,
However,
it
needs
and hydrocarbon gases for a
to
be
pointed
out
that
the
simulated results were generated using kinetic parameter values that
are considerably different from the best estimated values (Table 3.4-2)
for some reactions
for
this
used in
is
as shown in
that "the
Table 3.4-1.
The author's explanation
sets are different because many of the values
the simulation were
assigned rather arbitrarily before it
was
realized that they could be estimated by group additivity methods".
In
actual application
the
functional
reactions
of this approach,
groups
remains
and
chemistry.
total weight
For example,
loss
kinetic
challenging
information on coal structure,
in
estimating the concentrations of
because
parameters
of
the
of
the
limited
elementary
quantitative
and theoretical information on reaction
Howard
(1981)
pyrolysis at
points
out that the predicted
500 C for 30 s drops by about 10%
Figure 3.4-1 Simulated and experimental (a) weight loss and tar yield,
and (b) hydrocarbon gases from the pyrolysis of a bituminous coal.
[Reproduced from Gavalas et al. (1981b).]
50
3.0
40-
40
y:Exp.
2.5 -
-CH
C2Hr. 510*C
x :Exp. C2H4
600*C
4
2.0-
W 600*C
2~
510*C
-W
C2 H6
T 600*C
30
b
+: Exp. CH4
600*C
15
T 510*C
0
*-
020 -
-j
W: Wt. Loss
1.0
.T 10
T Tor Yield
+ Exo. Wt. Loss, 510*C
x Exp. Tor, 510'C
0.5
GSAA
C H4
+
600*C
24
0-
S
10
20
40
30
50
00
60
0
20
30
51P
40
50
60
Time (s)
Time (s)
Values of kinetic parameters used by Gavalas et al.
Table 3.4-2
(1981b) in their detailed chemistry model of coal pyrolysis [Reproduced
from Gavalas et al. (1981b).]
values used in
simulation
reaction
log A
E, kcal/
g-mol
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
15.5
15.3
14.9
14.9
14.4
14.4
12.8
12.8
12.8
12.8
12.0
12.0
12.0
12.0
12.3
12.3
12.6
13.6
13.6
12.4
12.4
8.0
11.1
8.0
8.0
11.1
8.0
8.0
8.0
11.0
7.0
7.0
11.0
7.0
7.0
11.0
7.0
7.0
8.5
84.0
65.0
63.0
63.0
52.0
68.5
50.0
48.0
50.0
50.0
38.0
37.0
37.0
38.0
38.0
38.0
35.0
35.0
35.0
35.0
36.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
20.0
15.0
15.0
15.0
10.0
10.0
10.0
10.0
10.0
10.0
35.0
best estimated
- values
log A
E, kcal/
g-mol
14.9
15.3
15.4
15.4
13.9
14.3
15.1
14.4
15.1
15.1
12.8
12.1
12.1
12.1
12.8
12.1
12.1
13.0
13.0
12.1
14.2
81.3
68.4
65.0
65.0
50.4
76.7
55.7
49.0
56.0
55.7
54.2
47.5
47.4
47.4
34.3
23.6
20.8
7.0
57.6
20.9
13.5
10.0
7.5
7.0
10.3
7.8
7.3
2.3
8.0
8.9
9.7
10.8
13.4
10.4
7.8
7.3
10.4
7.8
7.3
10.4
7.8
7.3
2.0
7.0
9.0
2.0
7.0
9.0
2.0
7.0
9.0
110
(from
-z33
=23%)
to
ethylene
bridge
kcal/mole
(from
if
the
activation
(Ph-CH 2 -CH2 -Ph
--
+
energy
2Ph-CH2-)
48 to 50 kcal/mole).
for
is
the
cleavage
raised
by
of
only
Such a small difference
in
2
the
activation energy can easily be produced by structural factors that are
not easily measured or accounted for.
from additional
1985),
and
groups is
ring aromatics
from activating
2-5 kcal/mole
is
Extra free-radical stabilization
approximately
2-4 kcal/mole
substitutions such as phenolic
(Gavalas
et al.,
1981a).
and ether
These complications
have been clearly acknowledged by Gavalas and co-workers in
the proposed reaction scheme.
the model
discussing
They state that, "In its present state
should not be considered as
final and ready to apply but
rather as a source of mechanistic and kinetic information..."
et al.,
(Stein,
(Gavalas
1981b).
Niksa
and
Kerstein
(1985)
present
an
alternative
chemical
description, where a large number of different chemical entities in the
coal are
units,
tar
classified into three groups consisting of bridges,
and peripheral units.
production
occurs
via a
A unique feature of this model is
competitive
reaction
stabilization or recombination of monomers.
bridge scission reactions.
Etar
>
Echar
temperatures
trend is
greater
favors
when mass
involving
According to the reaction scheme, assigning
tar production
transport
yields
pathway
that
Monomers are produced by
over
char
limitations
formation
are
obtained
at
at
higher
can be neglected.
consistent with experimental observations
volatiles
aromatic
higher
(Niksa,
1981)
heating
pyrolysis of small particles under low pressures (vacuum).
rates
This
that
in
Including a
mass transport description would enhance the range of applicability of
the model.
III
112
3.4.3 Models with explicit description of mass transport
Separate
intra-particle mass
transport
are
descriptions
required
for non-softening and softening coals as their transport mechanisms are
radically
different.
conservation
volatiles
In
the
equations
flux
Malinauskas,
terms
1983),
non-softening
inside
derived
and
the
particle
from
global
the
Russel
et
al.
(1979)
Dusty
chemical
formation and destruction of volatiles.
coal,
case,
multicomponent
generally
Gas
mass
consist
of
Theory
(Mason and
terms
describing
kinetic
In modeling hydropyrolysis of
considered
a
four
component
system
consisting of reactive volatiles, non-reactive volatiles, hydrogen, and
inert
gas.
In
significant
simplifications
assuming (1)
state,
quantitatively
in the
general
the
model,
they
conservation equation by
(2) external mass transport resistance is negligible,
pyrolysis,
(3) the
its porous solid structure during all stages of
(4) binary diffusivities for all volatiles are equal, and
particles
are spatially isothermal.
two assumptions
Justifications
are based on characteristic
for the first
time analysis.
assumption is valid only for non-softening coals.
The third
Improvements on the
fourth assumption can be made by assigning separate diffusivities
high
molecular
volatiles.
100 pm dia.)
C/s)
under
model is
made
concentration and pressure profiles are at a pseudosteady
coal particle retains
(5)
formulating
weight
(tars)
and
low
molecular
weight
for
(gases)
The last assumption generally holds when small particles (~C
are pyrolyzed
atmospheric
at moderately
pressure.
difficult to judge
rapid heating rates
The predictive
since the model is
capability
(< 1000
of the
strictly applicable to
non-softening coals, but predictions were compared to experimental data
from softening coals.
113
The model proposed by Gavalas and Wilks (1980) considers a ternary
system consisting of tars, gases and inert carrier gas, and as before,
applies
pseudosteady
state
conservation
external mass transport resistance.
equations
with
negligible
A major advancement made in
this
work was that the model incorporates pore-size distribution information
obtained from experimental measurements.
of a subbituminous coal,
size
Over the course of pyrolysis
they observed only minor changes in
distribution, specifically slight pore
enlargement.
the poreThe model
predicts that at low pressures, product yields depend on particle size
only,
while
particle
at
high
size.
pressures,
These
trends
they
were
depend
shown
on
to
both
agree
pressure
with
and
limited
experimental data from a subbituminous coal.
For a two component system consisting of gases and tars,
al.
(1985) relaxed the pseudosteady
transient
mass
again assumed.
Gas
flux
than
to
coupled with
a transient heat
Negligible external mass transfer resistance is
In estimating physical parameters required in the Dusty
equations,
correlations
state assumption, and formulated
conservation equations
conservation equation.
Bleik et
Bleik
that relate
consider
et
al.
(1985)
the parameters
detailed pore
chose
to
to particle
structure.
use
empirical
porosity,
Justifications
rather
for
this
approach are based on findings that macroscopic transport of volatiles
occurs
predominantly
through
Simons and Finson, 1979).
the model
large
pores
heat
compare well with experimental
transfer
and Wilks,
1980;
For non-softening coals, predictions from
temperature-time histories, pressures,
the
(Gavalas
description
data over a wide range of
and particle
significantly
widens
sizes.
the
Including
range
of
applicability of the model, especially for conditions of very rapid
114
heating and/or large particles.
Two different
softening
coals.
transport
is
coal
directions have been taken in
One
approach assumes
a mass boundary
transport resistance
1981).
The validity of this
values
of
physical
characteristic
It
particle.
the
rate
controlled by evaporation from the surface
and diffusion through
internal
that
modeling pyrolysis
time
is
(Zacharias,
assumption is
properties
scale
are
for
layer,
1979;
volatiles
of the molten
and neglects
Unger
any
and
Suuberg,
highly sensitive
to which
chosen
transport
of
of
in
inside
computing
and
the
outside
the
not uncommon to see as much as an order of magnitude
or more variation among values of some physical properties employed by
different
investigators,
diffusivities
evidence
(Oh,
cited
e.g.,
1985).
to
vapor
Further
support
the
assumption have been covered in
pressures
discussions
surface
and
on
liquid
phase
the
experimental
evaporation
controlling
Section 3.2.2.
Predictions
from the
model generally agree well with experimental data at pressures > 1 atm.
The
model
is
not
directly
applicable
at
evaporation becomes very rapid at pressures
lower
pressures
because
significantly less than 1
atm (Unger and Suuberg, 1981).
The
second
approach
to
modeling
pyrolysis
of
softening
coals
includes internal mass transport effects via a mechanism of growth and
escape of volatiles filled bubbles (Lewellen,
1975; Oh et al.,
1988).
The growth process occurs either by liquid-phase molecular diffusion of
volatiles to nearest bubbles or by coalescence of two adjacent bubbles.
The
escape
process
particle surface.
is
assumed
to
occur
when
a bubble
reaches
the
Although contributions are minor, a route for direct
escape
of volatiles
properties,
many
experimentally.
volatiles
surface is
also included in
this
The bubble description requires a large number of physical
approach.
overcome
to the particle
this
of
which
are
An extensive
inherent
yields,
difficult
estimate
or
effort was made by Oh et al.
difficulty.
plasticity,
to
The
and extent
temperatures, pressures, particle
sizes
model
of
predicted
swelling
and heating
measure
(1988)
to
trends
in
under
rates,
various
and
were in encouraging agreement with several literature measurements.
they
115
4. Experimental
Six coals
were
chosen
116
ranging from lignites
in
this
pyrolysis behavior.
fresh
sample
location;
to
low-volatile bituminous
study to investigate
coals
the effect of coal type on
Main selection criteria in choosing the coals were
quality;
good
and commercial
representation
and scientific
of
coal
interest.
rank
and
mine
Section 4.1 gives
more information on the selection procedure, and the properties of the
chosen coals.
An
electrically
heated
screen-heater
type
reactor
was
used
to
measure the apparent evolution kinetics and the yield limit of volatile
products for the
used in
Fong,
six coals.
past pyrolysis
1986),
as it
This reactor type has been extensively
studies
(Anthony
et al.,
1974;
offers many advantages important in
Suuberg,
1977;
kinetic studies
including
reliable temperature measurement of the sample over a wide
range
heating
of
rates,
products upon leaving
over
a
wide
range
rapid
quenching
the coal particle
of
pressures.
and
dilution
surface,
Section
of
volatile
and ability to work
4.2
gives
a
detailed
description of the experimental apparatus.
Coal
are
the
type,
main
temperature-time history, pressure,
independent variables
and particle size
in coal pyrolysis.
Section 4.3
specifies experimental conditions employed in this study.
Tars,
major
light hydrocarbon gases,
volatile
defined
as
condense in
the
products
sum
of
from coal
all
carbon oxides,
pyrolysis.
volatile
Tars
products
the reactor at room temperature;
and water are
are
(except
typically,
the
operationally
water)
that
they consist of
a complex mixture of molecules with molecular weights ranging from 100
to
1500.
Light hydrocarbon gases include
saturated and unsaturated
species
up
to
about
C4 .
Section
4.4
describes
the
experimental
procedure, including tar and gaseous product collection methods.
Additional liquid products can be obtained by solvent extraction of
the
solid residue of pyrolysis
important
in
determining
agglomeration
information
physical
(char).
tendencies
for softening coals
is potentially valuable
(tar
chromatography (GPC).
estimate
transport
and
(Fong,
is
are
made
The MW data are essential
properties
(e.g.,
1986).
vapor
and
Also, the
the
chemical and
believed to originate
Molecular weight
extract)
information is
swelling behavior,
in elucidating
mechanism of tar production as tar
products
extract
transport properties,
from the extractable material.
liquid
The
(MW)
by
measurements on
gel
permeation
input parameters
pressures
and
to
diffusion
coefficients).
In addition, time-resolved MW data can provide valuable
information in
understanding
liquids
inside
the
coal
the formation and depletion mechanism of
particle.
Extraction
yield
data
and
MW
measurements for these products collected in this work are currently
being analyzed by Sanchez (1988).
4.1. Coal selection
The six coals chosen for this study are:
1. Beulah Zap, ND (Lignite A)
2. Lower Wilcox, TX (Lignite A)
3. Smith Roland, WY (Subbituminous B)
4. Blue, NM (High-Volatile Bituminous C)
5. Illinois #6, IL (High-Volatile Bituminous A)
6. Lower Kittanning, PA (Low-Volatile Bituminous)
Table 4.1-1 gives the proximate
and ultimate analysis of the selected
117
coals.
All
coals
were
obtained
from
the
Advanced
Coal
Combustion
Chemistry Research Program sponsored by the Department of Energy (DOE),
Pittsburgh
Energy
Technology
Center
Technologies Research Center (UTRC).
(PETC)
and
managed
by
United
Whenever appropriate, the results
obtained from this study are compared to those from a similar study on
a
Montana
(1977);
lignite
and
Pittsburgh
Seam
bituminous
coal
by
Suuberg
Table 4.1-2 gives the analysis of these two coals.
The main selection criteria were (1) fresh sample quality, (2) good
representation
scientific
changes
of coal rank and mine location,
interest.
in
their
Weathered coals,
(Jacab et al,
Fresh
properties
coal
which
for example, are
1985).
samples
may
and (3)
are
commercial
desired
affect
to
pyrolysis
known to produce
lower
and
minimize
behavior.
tar yields
The above coals were collected recently (all in
1985) and were carefully stored under inert atmosphere until they were
ground and sieved.
As shown in Table 4.1.1,
the six coals have a wide
variation in the elemental compositions (e.g.,
56 wt% < C < 83 wt% dry
basis) and volatile contents (16-42 wt% dry).
Geographically, one coal
is
(IL),
from the east
(PA),
one from the mid-west
two from the west
(WY,ND), one from the southwest (NM), and one from the south (TX).
lignite
and
Illinois
#6 bituminous
coal
are
of
scientific
Zap
interest
since they have been widely used by other coal researchers (e.g., Serio
et al.,
1987;
Illinois #6
fuels
is
Darivakas,
1988;
Suuberg et al.,
1987).
In addition,
potentially important as a steam coal and in
processing.
Smith
Roland
subbituminous
(WY) coal
synthetic
and
Lower
Wilcox lignite (TX) are potentially important gasification feedstocks.
4.2.
Experimental apparatus
118
Table 4.1-1: Ultimate and proximate analysis of the six selected
coals in this studya
coal
Lower
Wilcox
L
Beulah
Zap
L
C
56.0
H
Smith
HVB
#6
HVB
Lower
Kittanning
LVB
62.0
74.9
67.4
82.5
4.0
4.6
5.0
4.4
4.5
1.1
1.0
1.0
1.4
1.3
1.3
S
0.7
1.1
1.1
0.8
3.9
1.2
0
19.9
21.6
19.5
13.7
8.7
2.4
ash
20.3
15.0
13.0
4.5
15.6
8.9
Roland
SB
60.2
4.2
N
coal-rankb
Blue
Illinois
Ultimate
analysis
wt% , dry
Proximate
analysis
wt% , dry
3.0
3.0
3.0
4.0
4.0
volatile
matter
45.3
42.0
45.2
43.3
35.7
16.3
fixed
carbon
34.4
43.0
41.8
52.2
48.7
74.8
ash
20.3
15.0
13.0
4.5
15.6
8.9
moisturec
a
b
C
1.0
analyzed by Huffman Laboratories, Inc.
L = lignite, SB = subbituminous, HVB
LVB = low-volatile bituminous.
partially vacuum dried.
=
high-volatile bituminous,
119
Table 4.1-2: Ultimate and proximate analysis of the coals
investigated by Suuberg (1977)a
coal
coal-rankb
Montana
L
Pittsburgh Seam
HVB
Ultimate
analysis
wt%, dry
C
63.6
68.8
H
4.1
4.9
N
1.0
1.3
S
1.2
5.4
Oc
19.5
8.1
ash
10.6
11.5
6.8
1.4
volatile
matter
39.6
39.5
fixed
carbon
49.8
49.1
ash
10.6
11.5
Proximate
analysis
wt% , dry
moistured
a
b
analyzed by Huffman Laboratories, Inc.
L = lignite, SB = subbituminous, HVB = high-volatile
bituminous,
LVB = low-volatile bituminous.
c
d
by difference.
as-received basis.
120
Figure
4.2-1
apparatus,
and
shows
data
the
reactor
acquisition
assembly,
system.
A
collection1 2 1
product
screen-heater
reactor,
modified from a version originally constructed by Fong (1986), was used
for
'vacuum'
and atmospheric pressure runs.
The
rector vessel has a
cylindrical casing made of two pyrex cylindrical pipes (Corning Pyrex,
22.86 cm x 22.86 cm)
stacked vertically,
and is
closed at the top and
bottom with a stainless steel plate (3/8" thick).
feedthroughs
pump,
gas
for
sampler,
recirculation
screen-heater
inlet/outlet
He supply),
ports
(recirculation pump,
and electrical
ports
vacuum
(thermocouple,
DC
The top plate has a gas inlet/outlet port leading to
power supply).
the
gas
The bottom plate has
pump.
originally
In
higher
pressure
constructed
modified by Griffin (1988) was used;
by
runs,
Anthony
a high-pressure
(1974)
and
later
it is similar to the low-pressure
reactor described above except for the stainless steel casing (rated up
to 200 atm at room temperature).
4.3. Experimental conditions
Coal
are
both
type,
temperature-time history, pressure, and particle
the main experimental variables.
the
primary
and
secondary
size
The first two variables affect
pyrolysis,
whereas
the
latter
two
variables affect mainly the secondary process.
Table
study.
4.3-1
summarizes
experimental
conditions
employed
The experiments varied coal type from lignites to low-volatile
bituminous coals, and reactor pressure from 'vacuum' (~10~3
atm.
in this
atm) to 10
Fixed variables were the particle size at 75-90 pm diameter,
temperature-time
history at 1000 C/s heat-up,
and 1050 C maximum temperature.
and
200-1000 C/s cool-down,
122
TEM4PERATURE-TIME
HISTORY
REACTOR GAS
(He)
GAS
CHROMATOGRAPH
VACUUM
The
Figure 4.2-1
acquisition system.
reactor
assembly,
product
collection,
and
data
123
Table 4.1-3: Summary of experimental conditions employed in this
study
reactor
variables:
varied (v)
or fixed (f)
range
covered
coal type
temperature-time
pressure
particle
his5toryviz
v
lignites to
low-volatile
bituminous coals,
elemental carbon
content ranges
72-92 wt% dmmf.
f
1000 C/s heat-up,
200-1000 C/s cool
down, 1050 C max.
temperature.
v
10-3
10 atm
f
75-90
pm dia.
124
4.4. Experimental procedures
In a typical low pressure run (10-3 to 1 atm), about 20 mg of 75-90
pm diameter particles spread thinly in the central region of 10 cm x 5
cm,
folded
400
mesh
devolatilized under
stainless
a controlled
steel
high pressure runs
high pressure reactor.
(Fig.4.3-1)
temperature-time history.
thin well dispersed coal particles,
be used in
screen
are
To ensure
smaller sample sizes (~5 mg) had to
since
smaller
screens are used in
the
Digital timers connected to a 24 volt DC power
supply control the heating, holding and cooling periods of the reacting
material.
The
sample
temperature
Chromel-Alumel thermocouple
within
the
Consistent
layers
of
folded
is
(Omega K-2
screen
near
measured
type;
the
coal
temperature measurements require
the screen to be minimal;
this
is
using
a
very
thin
0.0005 in. foil) placed
particles
the
gap
(Fig.4.3-1).
between the
two
achieved by keeping the
screen between the electrodes
as tight as possible.
time history of each run is
recorded using a Bascom Turner digital
recorder
(5
ms resolution),
and is
later transferred
interface to an IBM PC/AT for use in
reactor
gas,
temperature,
ultra
high
purity
The temperature-
through an RS232
kinetic analysis of the data.
He
(99.999%),
remains
near
The
room
and provides rapid dilution and quenching of volatiles
as
soon as they are evolved from the coal surface, thus presenting minimal
opportunity for extra-particle secondary reactions.
Reliable
difficult.
solvent-soaked
tar-yield
The
measurements
conventional
tissue
(e.g.,
in
technique
CH2 C12 )
to
screen-heater
usually
carefully
reactors
involves
using
wipe
the
off
are
a
tar
condensed on the inner surfaces of the reactor, followed by evaporating
125
REACTOR
SAMPLE
-THERMOCOUPLE
Figure 4.3-1
Details of the electrical screen-heater reactor.
SCREEN
ELECTR ODE
TAR COLLECTION
GLASS FUNNEL
ASSEMB LY
FILTER DISC
Figure 4.3-2
New tar collectors in electrical screen-heater reactor.
away the solvent
(Bautista, 1984;
Some major
Oh, 1985).
sources
of
error associated with this method are incomplete tar collection due to
the fact that some tar condenses on reactor internals that are poorly
accessible;
incomplete
evaporation
of solvent
arising from a partial
miscibility between the solvent and the tar; and loss of more-volatile
tar
components
significant
during
solvent
evaporation
(Bautista, 1984).
Thus
scatter and large experimental errors can typify screen-
heater tar data.
The
problems
associated
with
the
conventional
technique
were
collection
system for screen-heater reactors.
diminished
in
this
work
by
tar
developing
collection
a
new
tar
Figure 4.3-2 shows the
tar traps, each consisting of a glass funnel connected at its stem to a
small teflon filter disk (0.2 pm pore size).
The mouths of the funnels
completely cover the coal particles between the layers of the screen.
Upon
leaving
the coal
particle,
tar,
convected into one of the two traps.
after the
increase.
reducing
run,
the
into
pressure
recirculation pump
the
(Cole-Parmer Air
is
taken as their combined weight
traps was
downstream
gas,
The traps were weighed before and
and the tar yield was
Convection
together with reactor
of
achieved during the run by
the
Cadet)
traps
using
a
small
for atmospheric pressure
runs,
and a small vent from the trap to atmosphere for higher pressure
runs
(Griffin, 1988).
As
would be
expected,
a sufficient pressure
differential could not be obtained across the trap in vacuum runs.
In
these experiments, the screen was virtually surrounded by tared sheets
of aluminum foil to condense the tar as it evolved from the screen.
The tar yield was
the run.
taken as the increase in
weight of the foils after
Tar yields measured using the tar traps were comparable
to
126
those using the solvent-soaked
tissue method,
but the new method gave
significantly better reproducibility.
Gaseous
C2H 4 ,
and
products
C 2 H6 .
from
The
coal
pyrolysis
gas yields were
Sigma 2B Gas Chromatograph
(GC),
conductivity detector (TCD),
(FID).
Helium was used as a carrier
be
measurements
identified
CO2 , H20, CH 4 ,
measured using a
Perkin Elmer
using
the
were not attempted in
column,
a
and a flame ionization detector
gas.
been previously reported (Suuberg, 1977;
not
CO,
equipped with a spherocarb
thermal
could
include
Some C3 's
and C4 +'s
have
Oh, 1985), but such species
current
set-up.
this study because
Hydrogen
gas
to do so would
require another GC using a carrier gas with a thermal conductivity much
different
from
'missing'
gas species represents
balance
that
of
H2 , typically
Ar.
But,
the
sum
of
these
a small fraction of the overall mass
(~ 1-4 wt% as rec'd basis).
Previous
gas collection procedures also gave reliable measurements
as indicated by good mass balances (Suuberg, 1977), but were slow since
the entire reactor volume was purged out for as much as several hours
(Oh, 1985).
study,
The large number of experimental runs planned for this
demanded more rapid product
known volume of the reactor gas is
collection.
In
the new method,
a
withdrawn from the reactor using a
gas sampler, and then is concentrated by purging the sampler through a
cooled
lipophilic
trap
helium) are collected.
steel tube
where
The
(1/4 in. O.D.)
all
light
gases
except hydrogen
lipophilic trap consists of a stainless
packed with Porapak QS,
liquid nitrogen (b.p. -196 C).
(and
and is cooled by
The total amount of gas in the reactor
is computed from the volumes of the reactor and gas sampler.
127
5. Experimental results and discussion
Experimental
pyrolysis
carried
studies
product
out
to
yields,
using
six
128
investigate
the
compositions,
coals
ranging
effect
and
from
of coal
evolution
lignites
type
rates
to
on
were
low-volatile
bituminous coals.
Under reactor pressures ranging from 'vacuum' (10-3
atm)
coal
to
10
atm,
particles
with
diameters
of
75-90
pm
were
pyrolyzed in a screen-heater type reactor at a heating rate of 1000 C/s
to a maximum temperature of 1050 C.
Sections 5.1,
on
the
5.2,
evolution
and 5.3
behavior
respectively present coal-type effects
of
tars,
individual
volatiles under 1 atm reactor pressure.
gases,
and
total
The atmospheric pressure data
for a given coal represent an overall pyrolysis behavior, that includes
contributions from primary decomposition reactions, and from secondary
reactions
coupled
pressure
is
with
varied
mass
to
transport
infer
the
processes.
extent
of
contributions in the overall pyrolysis behavior.
In
Section
secondary
5.4,
reaction
Section 5.5 discusses
magnitudes of uncertainties associated with experimental data.
5.1. Coal-type effects on tar production
5.1.1. Observed rate of tar production
Figure
5.1-1
shows
the
atmospheric
tar
yield
holding temperatures for the six coals
studied.
rates
1000
in
these runs were
respectively
holding time at final temperature.
1050
C.
At
higher
structures
such
(Kobayashi
et
as
al.,
temperatures,
aromatic
1977).
peak
and
Heating and cooling
and 200-1000
C/s with no
The maximum peak temperature was
the
rings
The
versus
fragmentation
lead
lines
to
through
of more
further
the
weight
data
stable
loss
points
in
129
30 28 -
LWx
ZP
SR
26 -
+
0
24 _
L 8L
X
IL
22 -
7
LK
20 -
a
IL
x
4
18-x
16 -
aj
14 -
LW
14
12 >_
0
10 -
V 0
LK
8 6-+
6+
+
++
+
24-ZP
1
2
300
500
700
900
1 100
TEMPERATURE (C)
Figure 5.1-1 Experimental yields of pyrolysis tar versus peak temperature
for the six coals selected in this study. Carbon: LW < ZP < SR < BL < IL
< LK.
Abbreviations:
LW = Lower Wilcox lignite
BL = Blue high-volatile bituminous
ZP = Beulah Zap lignite
IL = Illinois high-volatile bituminous
SR = Smith Roland subbituminous LK = Lower Kittanning low-volatile bit.
Fig.5.1-1
are hand-drawn to
indicate
trends.
Individual
plots with
model predictions are given in Chapter 6 (Fig.6.1-1).
Qualitatively, the figure
coal
shows
that there is a clear effect of
type on both the apparent rate of tar production and the yield
limit, defined as the asymptotic yield at high peak temperatures (5 800
C).
Low-rank
coals
(ZP,LW,SR)
tend
to
initiate and
achieve
given
extents of tar production at lower temperatures compared to higher rank
coals (BL, IL,LK); abbreviations are defined in Fig.5.1-1.
These
points
are
reinforced by quantitative
apparent rate of tar production presented in
the temperatures
observations
Fig.5.1-2,
at which the tar yield reaches
on
the
which compares
25% (T25),
50% (T50),
and 75% (T75) of the yield limit for the six coals represented by their
elemental
carbon
contents
in
wt%
dmmf.
temperatures were determined from the
model
(see Fig.6.1-1).
represents
roughly
an
evolution rate
is
to
three
characteristic
tar data fitted with the MIPR
The difference between T75 and T25
approximate
corresponds
The
the
maximum.
spread
of
the
temperature
Comparing
yield
at
T50
which
shows
(T75-T25)
curve,
whereas
T50
the
observed
tar
almost
a monotonic
increase with coal rank represented by the elemental carbon content of
the coal, indicating a shift in the yield curve to higher temperatures
for higher rank coals.
an increase
studied.
in
T50 ranges from 545 C for ZP to 675 C for LK,
the maximum difference
of about 130 C among the coals
Comparing (T75-T25) shows a decreasing trend for higher rank
coals, indicating less spread in the yield curve for higher rank coals.
The difference ranges from 175 C for ZP to 85 C for LK, a reduction in
the maximum difference of about 90 C.
A similar trend of greater T50 and smaller
(T75-T25)
as the coal
130
131
760
740 720 -
700 -T75
0
680
660 O
640 -
W
620
v
-
600 -
W
(L
Li
T25
W
580
0
560
540 520 500 480 460 -
A
440
--70
74
78
0
82
86
90
94
ELEMENTAL CARBON CONTENT (WT% DMMF)
T50
A
T25
V
T75
tar
atmospheric
Figure 5.1-2 Characteristic yield temperatures for
studied.
coals
six
the
production versus elemental carbon content for
[(a) T25, (b) T50, (c) T75 ; Tx denotes the temperature at which the
yield reaches x% of the maximum yield]. Carbon: LW < ZP < SR < BL < IL <
Abbreviations: see Fig.5.1-1.
LK.
rank increases,
lignite
(ML)
has been reported in
a previous
study using a Montana
and a Pittsburgh Seam high-volatile bituminous coal
In
(Suuberg, 1977).
relation to the coals studied here,
(PB)
these coals
have elemental carbon contents in the order of LW < ML < ZP < SR < BL <
PB < IL < LK.
The elemental and proximate analysis of these two coals
are given in Table 4.1-2.
bituminous
coal,
T50
In going from Montana lignite to Pittsburgh
increases
decreases from 230 to 165 C.
from
570
to
675
C,
and
(T75-T25)
However, directly comparing these values
to those from this work is not strictly valid due to somewhat different
temperature-time histories employed in the two studies.
An
exact
transport
description
phenomena
available.
Thus,
of
involved
the
complex
in
interpretation
tar
reaction
production
of the
is
chemistry
and
currently
not
observed tar evolution rate
behavior for different coal types, depends on the assumed mechanism for
tar formation.
'tar
A frequently assumed mechanism is
precursors'
in
the
coal via
multiple
the decomposition of
first-order
independent
parallel reactions (Serio, 1984; Ko et al., 1988a)
first-order decomposition
Tar precursors in coal -------------------------------+ Tar
transport
The
fitted
parameters
for
influenced by any physical
this
global
decomposition
transport effects.
Under
reaction
such a
description with the further assumption that all coals have
preexponential
implies
factor
that tars
activation
assumptions
are
energies.
distribution
of
in
the Arrhenius
constant,
global
the same
a higher T50
produced from reactions with greater apparent
Similarly,
apparent
of this
rate
are
global
a larger
activation
(T75-T25) implies
energies.
Thus,
a wider
under
the
description, higher rank coals appear to
132
produce tars from reactions with apparent activation energies that have
a higher mean but a narrower distribution.
5.1.2. Tar yield limit
Figure 5.1-1 shows that the tar yield limit increases from 7-13 wt%
dmmf for the low-rank coals (LW,SR,ZP), to 21-25 wt% dmmf for the highrank coals
(BL,IL),
rank coal
(LK).
variation
is
and then drops to 11 wt% dmmf for the very highDespite this
general trend, however, a significant
also noted among coals of the same rank;
two fold variation for the low-rank coals.
other tar yield data reported in
rank
information
alone
is
not
e.g.,
almost a
This observation along with
the literature,
enough to
indicates
that coal-
quantitatively
explain
the
observed trend.
A new approach to quantitatively relate the tar yield to measurable
properties
of
the
coal
is
to:
(1) assume
a
chemical
and
physical
mechanism of tar production, (2) identify the chemical structures that
are important from the assumed mechanism, (3) formulate a coal-specific
parameter
the
based on the important
structures
(Ko et al.,
The
against
correlation
a
lignites
large
readily) measurable properties
procedure
set
of
discussed
experimental
in
data
detail
from
representing a wide range of coals (37
to
gives
specified
and (4)
relate
of the coal
1987, 1988c).
literature,
5.1-1
to (ideally
chemical structures,
anthracites)
the
and pressures
elemental
pressure
for
analysis
each
coal.
('vacuum'
and measured
The
data
below
this
study
coals,
to
90
tar
is
tested
and
the
ranging from
atm).
Table
yield under
represent
the
a
maximum
amount of tar generated during devolatilization with minimal influence
133
from
secondary reactions
outside
the
coal particle.
The
following
conditions support the fact that the tar data used here represent good
estimates
of the probable upper bounds
on tar production at a given
pressure:
o All the data are from rapid devolatilization (100-1500 K/s) of
small samples of coal particles (~ 20 mg) in the 50-100 pm dia. size
range, under constant pressures ranging from 0.0001 to 90 atm in
screen-heater reactors.
These conditions afford minimal
opportunity for in-bed secondary reactions of newly evolved tars.
o The screen-heater reactors provide rapid dilution and quenching of
tar and other volatiles as soon as they are evolved from the coal
surface, thus presenting minimal opportunities for extra-particle
secondary reactions.
o The final temperature and holding time (<1000 C and <10 s) are
sufficient to drive devolatilization, including tar generation,
essentially to completion.
Formulation of correlation: treatment of coal-type effects
(1) Chemical and physical mechanism of tar production
Tar is assumed to be generated via the global mechanism:
[2] transport
[1] thermolysis
Coal --------------------- Metaplast ------------------ Tar
of bridges
The
above mechanism was first suggested by van Krevelen
similar
versions
have
since
been widely
applied
(1961),
in many
different
pyrolysis models (Unger and Suuberg, 1981; Oh, 1985; Fong, 1986).
(2) Important chemical structures
and
134
The identities and numbers of bridges between aromatic clusters of
the coal and the concentration of hydrogen available to stabilize the
free
radicals
created
structural
chemical
effects.
Since
by
bridge
factors
the
in
structural
scission
tar
reactions
generation
features
are
important
without
transport
important
in
the
latter
process, [2] in the assumed mechanism, are not easily identifiable, the
transport
effect is
correlated via empirical parameters
obtained from
best-fit analyses of existing data.
(3) Formulation of coal-specific parameter
A coal-specific parameter,
XTAR,
proposed to correlate
tar yields
with coal type is
XTAR =
(no. of labile bridges)(amt. of abstractable hydrogen)/
(no. of cross-linked bridges)
(5.1-1)
(4) Estimation of identified structures
Since
the necessary molecular structures are generally unavailable
for most coals, reasonable estimates were made for each quantity based
on currently available information.
Labile
bridges
are
only
aliphatic,
and
their
concentration
is
assumed to be proportional to the aliphatic carbon content of the raw
coal.
This fraction ( 1 -fa)
also contains contributions from carboxyl,
carbonyl, and ether carbons, but these are assumed to be small.
(labile bridges)
=
((l-fa)[C]/12}
Thus,
8
.
(5.1-2)
where [C] is the carbon content of the coal in wt% dmmf, and fa is
aromaticity,
estimated from a polynomial best-fit of fa
to
[C]
the
using
data from Gerstein et al. (1982)
fa
=
0.830526 - 2.008147 ([C]/100) + 2.241218 ([C]/100) 2
(5.1-3)
135
The exponent of 1.8
in
Eq.(5.1-2)
is
a best-fit parameter obtained by
applying multivariable fitting routines to obtain the best correlation
between experimental tar yields and
An alternative and perhaps
XTAR.
physically more appealing rationale for this exponent is to assume that
is proportional
XTAR
8
to (labile bridges)'
and that (labile bridges)
is linearly related to the aliphatic carbon content.
Cross-linked
structures, whose
bridges
consist
only
of
concentration is assumed
ether
and
thioether
to be proportional
sum of elemental and organic sulphur contents of the raw coal.
cross-linked bridges
to the
Thus,
=
[0]/16 + [S0 ]/32.066
if [0]>3.5 wt% dmmf,
=
3.5/16 + [S0 ]/32.066
if [0]<3.5 wt% dmmf
(5.1-4)
where [0] and [SO] are elemental oxygen and organic sulphur contents in
A constant
wt% dmmf respectively.
elemental
highly
dmmf,
[0] was needed for coals with low
oxygen contents because the number of cross-linked bridge is
sensitive
to coal elemental oxygen contents below about 4 wt%
and uncertainties
in
oxygen measurement can easily exceed 1 wt%
dmmf.
Abstractable
carbons.
hydrogen
is
the
hydrogen
attached
to
aliphatic
Its concentration is assumed to be proportional to the amount
of elemental hydrogen
of the
account for experimental
raw coal,
observations
minus a slight correction
that
the abstractable hydrogen (Suuberg, 1977).
-OH
to
groups may compete for
Thus,
abstractable hydrogen = [H]/l - [OH]/17
where [H] is the elemental hydrogen content (wt% dmmf), and [OH]
(5.1-5)
is the
hydroxyl group content (wt% dmmf) obtained from Given (1976)
[OH] = 33.2 - 0.35 [C]
(5.1-6)
136
Figure
5.1-3,
a
plot
of
the
three
key
structural
quantities
computed from Eqs. (5.1-2,4,5) versus the elemental composition, offers
a quick
and convenient way to obtain XTAR with minimal computational
effort.
Formulation of correlation: treatment of pressure effects
Tar yield limit at a given pressure is linearly correlated with the
coal-type parameter derived above:
#(P)XTAR
Tar yield limit (wt% dmmf) = a(P) +
The
pressure
dependent
coefficients
a
and
#
(5.1-7)
are
obtained
by best
fitting experimental tar yield data either for specified pressures and
pressure ranges or for all pressures.
The best-fit coefficient
values
The results are given below.
for pressure-specific
correlation
are as follows:
For coals with
XTAR <
a(10 Pa- 9 MPa)
15,
P(10 Pa- 9 MPa)
For coals with 155
2
=
XTAR
(5.1-8)
0.
=
(5.1-9)
<31
a(10-100 Pa)
=
-30.8125,
a(O.1 MPa)
=
-22.375
a(l MPa)
=
-16.75
a(2.5-9 MPa)
=
-10.1875,
P(10-100 Pa) = 2.1825
(5.1-10,11)
,
#(0.1 MPa)
=
1.625
(5.1-12,13)
,
#(1 MPa)
=
1.25
(5.1-14,15)
P(2.5-9 MPa)
=
0.8125.
(5.1-16,17)
For coals with XTAR > 31,
a(10-100 Pa)
=
37
(5.1-18)
a(0.1 MPa)
=
28
(5.1-19)
a(l MPa)
=
22
(5.1-20)
a(2.5- 9 MPa)
=
15
(5.1-21)
137
138
1.6
0P
0
6
(a)
1.4
(b)
5
--
:3
0
1.2
4-
0
o
.
75
E
o
(C)
4
3
aCD
800
150-
0
o
3
a(
3
0.0
-
t
E-9 -t
~
0.8
so-2.
m
0tam
80.
0.6
41
.0
1.0*
o
0.5
(a)C
-0.64U
wUmm
E
wt% dmmf
dmm
(c (c)
0=.0, wt%
ydogn
absratale
s
brde
vs
0
.4
fo
v.
H]fro
q(.-)
0] from Eq.(5.1-4)
muto
rmEq(.-
vs.0[]cmue
rde
abl
brde
q.5.111.5a
Etimtesofthestrctralquatiiesin
Figre
.1nubro
E.(.1-);(c
nube
o
crssl.5e
P(10 Pa - 9 MPa)
The
best-fit
0.
=
(5.1-22)
coefficients
applicable
for
all
pressures
are
as
follows:
For coals with XTAR <15,
a = 2
P = 0.
,
(5.1-23,24)
For coals with 15! XTAR <31,
a
#
1/(0.021533 + 0.02865lLp)
=
=
(5.1-25,26)
2
- 0.06959L
0.508030 + 0.696487Lp
For coals with
- 36
.
(5.1-27,28)
>31,
XTAR
a = 11.24071 + 9.743707L, -0.91326L,2
(5.1-29)
p
(5.1-30)
=
LP = -log
0.
OP + 1 with
1
10 atm),
and P is
P=reactor pressure in MPa for P s 2.5 MPa (1 MPa
fixed at
2.5 MPa for reactor pressure
above 2.5
This was justified since pressure has negligible effects on tar
MPa.
yield above
2.5 MPa.
Bautista
(1984)
observed that tar yield did not
decline with increasing pressure above
(see
Fig.
5.1-4 below)
found close
=2 MPa, and the present work
agreement between predictions
and
data using 2.5 MPa to represent pressures from 2.5-9 MPa.
Results and discussion
Figure
predicted
[Eqs.
wt%
5.1-4
from
(5.1-8)
dmmf
of
compares
Eq.(5.1-7)
through
the
measured
using
(5.1-22)].
The standard error
wt%
The
dmmf.
definition
standard error
the
tar
yields
pressure-specific
with
those
coefficients
The predicted yields are within ± 5
observed values
pressures.
maximum
for
all
of estimate
of
coals
tested
at
the
four
of the prediction was 2.8
estimate
was
computed
using
the
139
140
40,
Ao
35-
1M00Pa
30- _
AI1hV7
N L17
A Azi*
30~ 0.1 MPa
A 0
o0
1 MPa
25-
Z-
10-100 Pa
10-10Pa
0.1 MPa
EBa *
4
0z
2.5-9 MPa 0
1 MPa
A
20_
W
5:
-J-
15
2.5-9 MPa
<
-__1_-
z
10 -
Lii
x
5
083
5
10
20
15
COAL-TYPE PARAMETER,
25
30
35
X
Figure 5.1-4 Correlation of tar yields at different pressures with
Symbols: see Table 5.1-1. Lines are from Eq.(5.1-7) using
XTAR.
pressure-specific parameters from Eqs. (5.1-8)-(5.1-22).
the
n
standard error
of estimate
141
2 1 /2
(Yieldjexp'l - Yieldi.pre'd)
n-k
J (5.1-31)
=
j=1
where n
is
the number of data points
(j),
and k the number of best-
fitted parameters used in the correlation.
Figure
5.1-5
predictions
compares
obtained
from
experimental
Eq.(5.1-7)
data
for
all
using
the
pressure-correlated
pressures
with
parameters [Eqs. (5.1-23)-(5.1-30)]. The predicted yields are within ±6
wt% dmmf for all coals.
Use of the pressure correlated parameters has
the advantage that it is applicable for all pressures between 10 Pa to
9 MPa, but suffers from a slightly greater standard error of
estimate
of 3.1 wt% dmmf.
The wide range of coal types
(4s XTAR
534)
and pressures
(10 Pa-9
MPa) covered in the present data base (Table 5.1-1) suggests that there
should
be
little
need
domains tested here.
to
extrapolate
the
correlation
outside
the
Predictions based on small extrapolations of XTAR
should be of comparable reliability to those from tested XTAR values.
As a rough guideline,
anthracites;
bituminous
typically ranges
XTAR
coals;
20
to
30
for
and medium-volatile
high-volatile bituminous coals.
the
4
to
12
for
12 to 20 for low tar producing lignites and low-volatile
subbituminous
MPa,
from
2.5-9
MPa
high
tar
producing
bituminous coals;
lignites,
and 30 to
and
34 for
For applications to pressures above 9
correlation
is
expected
to
give
satisfactory
predictions since pressure effects were observed to be negligible at
pressures beyond ~2.5 MPa.
Increasing
rates,
pressure
lowers
tar
thus allowing additional time
reactions.
yield by
slowing
tar
transport
for tar conversion in secondary
The decrease in tar yield with increasing pressure is less
142
40
35-_
10r4 -10-5
3
AG
0.01
L'q LVV
30.
0.1
OAOO
0.2-0.7
eO>
1
Ewe 4
Lg
e
9
-IJ
T_
20 _~j
L1o
A
ry
15_.g
L-
10-
DOLB
nJ
F-
z
5
A(
0
0
I
5
I
10
I
15
20
25
30
35
40
PREDICTED TAR YIELD (WT % DMMF)
Figure 5.1-5 Comparison of experimental tar yields with those
predicted by Eq.(5.1-7) using the pressure-correlated parameters
Eqs. (5.1-23)-(5.1-30). Symbols: see Table 5.1-1.
from
Table 5.1-1 Characteristics of coals and experimental tar yields used in
the tar yield limit correlation.
Tar yield (wt% doamf; symbols used in Figs.
at pressures (MPa) of
Elemental Analysis
(wtI dmnf)
Investigator
Coala
Montana
Freihaut and
seery (1981)
L
Wyodak SB 1
Wyodak SB 2
Utah B
Colorado B
Pittsburgh B
Alabama B
Anthricite
C
Ob
I
68.3
75.4
75.5
78.2
81.0
82.0
85.0
25.5
18.1
17.0
13.9
11.2
4.6
4.9
5.2
5.5
5.5
5.4
4.6
2.6
5.5
5.4
5.1
5.3
93.7
Freihaut
at al. (1982)
Colorado B
Pittsburgh B
81.0
Loison and
Chauvin (1964)
Paulquemont B
B
Wendel III
80.8
86.1
88.4
88.5
89.0
91.9
86.7
82.2
91.5
90.1
87.6
84.2
81.0
72.2
73.1
82.9
76.8
72.0
83.2
79.1
72.4
91.9
72.7
91.3
74.5
78.6
85.1
75.4
85.1
84.7
Oh (1985)
Arendt and
van Heek
(1981)
Suuberg
(1477)
Ceway
(1981)
geltsen (1978)
This Study
Lens-Lievin B
Emma B
Bergmannsgluck B
Maigre oignies B
Flenus de Bruay B
Pittsburgh B
Prosper II B
Schlugel u. Eisen B
Uulfen B
Leopold B
Pittsburgh B
Montana L
SB
Wyodak
Sesser SB
Colstrip L
Lower Wilcox L
Illinois B
Blue SB
Suuber
et al. (1987)
(1985)
8autista
(1984)
C
Bruceton B
Pittsburgh B
8.2
1.9
11.2
9.4
12.8
8.3
6.4
5.4
5.2
4.4
6.4
10.0
2.7
3.7
5.7
8.4
9.7
22.0
19.8
10.3
17.4
20.9
9.8'
14.1
21.6
1.7
20.9
4.1
20.5
14.6
7.6
19.1
7.6
7.9
5.0
4.7
4.6
3.8
5.1
5.9
4.4
4,3
4.7
5.7
5.7
4.6
6.1
5.4
4.9
5.6
5.4
5.3
4.8
5.0
5.3
4.6
4.1
5.4
5.6
4.1
5.6
5.8
0.6
0.6
0.5
0.2
0.5
2.0
10-4-10-5
A20.0
A21.0
A27.0
A26. Of
A39.0
A25.0
A 2.0
0.1
1
6.9
9
A1 9 . 8 d,f
A28.ld
219.2f
A26.0
-20.0
124.3
.0
17.6
S27
15.1
2.1
39.8
230.51
V37.0
* 8.4
-9.9
17.7
26.5
28.4
L29. 3
3.7
e6.6
25 7
26. 5
.6.5
19. 3
7.5
®14 .1
15. 2
e12.2
19.3
19.8
713. 8
W
3.2
21.5e
11.5
13.1
E}24.8
21.2
7.2
016.8
830.1
W27.7
a 9.1
014.0
014.8
10,7
12.9
9.9
5.6
22.8
26.5
0.5
0.5
0.5
0.7
0.01
A 18.0 f
S6.
815.0
814.5
4.1
M
6.5h
M11.1
7
,A37. 7
*38.
8 9.9
6
(>30.0
29.7(0.2)9
26.5 (0. 4)
,25.2(0.7)
23.7 (1.0) d
21.1(1.5)
'20.6(2.4)
bitueinous;
L-lignite; SB-subbituminous
By Oifference
EstbAated as half the total sulphur content when organic sulphur not
reported (Loison and Chauvin, 1964; Arendt and van Heek, 1981; Cosway,
a1., 1985, 1987; Bautista, 1984; this
1981; Reitzen, 1978; Suuberg
S
b
Pocahantas B
North Dakota L
Illinois B
Bruceton D
North Dakota L
subetrg
ecAl.
leulah Zap L
Lower Kittanning B
Sith
Roland L
82.0
9.4
c
0.7
0.4
0.6
0.8
0.6
1.9
0.5
0.7
0.6
1.9
0.5
0.3
0.6
0.5
0.7
0.4
0.5
1.9
0.7
0.5
0.6
0.9
1.9
0.8
0.4
0.4
0.5
0.4
2.0
0.4
5.1-4 and 5.1-5)
at
study).
d Obtained by interpolation between 0.0007 and 0.013 MPa in Freihaut
et al. (1982), and between 0.7 and 1.5 MPa in Bautista (1984).
e The tar yield (6.5 wt% doimf) reported for Sesser S8 seemed low and
was shstituted by the 21.5 wt% damEfmeasured in this study.
Colorado 8 and Montana L from Freihaut and Seery (1981), and Freihaut
at al. (1982) were not used because possible errors in tar yield
measurement are suspected.
g Indicates pressure in MPa.
h This value is slightly lower than the previously reported value (7.2
wt% deamf) in Ko et al. (1988c).
F".
severe for coals with lower XTAR, becoming almost negligible for coals
with
below 15 (Fig.5.1-4).
XTAR
finding are:
secondary
(1)
Two possible
explanations
tars from coals with higher
reactions,
and
(2) faster rate
caused by higher concentrations
coals with higher
XTAR.
of tars
XTAR
for
this
are more reactive to
of tar secondary
reactions
inside the coal particle for
Quantitative rationalization for this behavior
must await further studies of tar reactivity.
Particle
size
is
another
important
variable
which
affects
tar
yield. However, the small data base on particle-size effects (Suuberg,
1977; Bautista, 1984) suggests that the tar yield at 0.1 MPa is almost
unaffected between 50 and 300 pm
(dia.),
and only
slightly
affected
between 300 and 800 pm (dia.).
5.2. Coal-type effects on gas production
Figure
the
elemental
study,
not
5.2-1 compares the yield limit of gaseous products versus
carbon content for
and the two coals
identical
the
six coals
studied by Suuberg
experimental conditions.
(1977)
Higher
produce less carbon oxides and pyrolytic water,
ranges for CO, C0 2 , H 2 0, CH4
dmmf respectively.
investigated
under similar but
rank coals
to
1.6
ethane.
wt%
generally
but more methane;
the
are 0.9-11.0, 0.4-9.9, 2.4-16, 1.6-4.3 wt%
The ethylene and ethane yields are small and their
absolute yield values are less affected by coal type;
0.6
in this
dmmf for ethylene and from 0.2
to
they range from
0.7 wt% dmmf for
The higher carbon oxides and water yields have been associated
with higher concentrations of carboxyl and hydroxyl groups respectively
in
lower
mechanism
rank
is
coals
not
yet
(Suuberg,
available
1977).
However,
to
quantitatively
an
exact
reaction
rationalize
the
144
145
4.5-
cH 4
4.0 -
3.5 -
3.0 -
2.5 -
I
-J
LJ
2.0 1.5 -
1.0
++
1
ED
+
0.5
0.0
c 2 H4
0c
2 H6
70
74
[IN
78
82
86
94
90
ELEMENTALCARBON CONTENT (WT% DMMF)
CH4
+D
C2H4
C2H6
o+
16
b
V
15
14
13
12
11
10
9
I'.
2
T
87
6
5
4
37
2-
0
070
74
78
82
86
90
94
ELEMENTALCARBON CONTENT (WT DMMF)
v
H20
X 0 C02
AA CO
Figure 5.2-1 Comparison of the yield limit of gaseous products
versus
the
elemental
carbon
content
at
ambient
pressure:
(a)
hydrocarbons; (b) carbon oxides and pyrolytic water.
Open or non-circled
symbols are from this study; closed or circled symbols are from Suuberg
(1977). Carbon: LW < ML < ZP < SR < BL < PB < IL < LK. Abbreviations: ML
=
Montana lignite, PB = Pittsburgh Seam high-volatile bituminous, see
Fig.5.1-1 for others.
relationship.
Methane production has been postulated to occur via bond
dissociation
abstracting
of
alkyl
hydrogen
groups
form
to
yield methyl
methane
(Gavalas
radicals,
et
al.,
which upon
1981a).
But
applying such a mechanism to explain the
observed trend for methane
yields
the
necessary quantitative
in particular the
concentration of alkyl
is
difficult
due
to
the
structural information, e.g.,
lack
of
groups.
Figures
the yields
5.2-2,
5.2-3,
and apparent
5.2-4,
5.2-5,
production
and 5.2-6 respectively compare
rates
of CH 4 , C2
CO2 production for the six coals investigated.
4,
C 2 H6 ,
CO,
and
Each figure consists of
(a) a combined plot of yield versus peak and holding temperatures for
all
six
coals,
temperatures
versus
the
and
(b) a
(T25, T50,
plot
showing
and T75;
see
three
Section 5.1.1
elemental carbon content of the
figures are free-drawn
trend lines.
characteristic
coal.
yield
for definitions)
The lines
in the
Model generated curves for yield
versus peak holding temperatures for these products are given in Figs.
6.1-3
through
6.1-7
respectively.
Comparing
T50
shows
increasing trend with coal rank for methane and ethane
4),
but
almost
no observable
(Figs. 5.2-3,5,6).
T25)
appears
effect
for
ethylene
(Figs.
a
slightly
5.2-2 and
and carbon
oxides
The spread of the yield curve as indicated by (T75-
to be unaffected by coal type for all gases,
except for
carbon dioxide, which shows a decreasing trend for higher rank coals.
Reasons
for
the
lack
of
observable
coal-type
apparent rate of gas production are currently unclear.
is
that
effects
on
the
One hypothesis
the kinetics of gas production are unaffected by coal type
(Solomon and Hamblen, 1985).
Gaseous products
are claimed to
from decomposition of specific functional groups, e.g.,
evolve
carbon monoxide
146
5
I
a
0
LW
O
SR
+
4 -
An BL
X
7
IL
LK
x
x
I--,
147
V
ZP
V
3 -
x
U
z
00
2 -
01
r
+
1 -
0- |
+
1
3V
N
I
600
400
1000
800
1000
800
HOLDING TEMPERATURE
PEAK TEMPERATURE (C)
960 -
b
(C)
V
940 920 -
V
VV
900 880 -
0
860 C-)
2
Ld
840 820 -
0
800
780
760
740
720 -
700 - i
70
I-
78
74
0
I
i
82
86
II
90
94
ELEMENTAL CARBON CONTENT (WT% DMMF)
V
T75
A
T25
T50
Figure 5.2-2 Comparison of methane production rate at 1 atm.
(a) combined plot of yields versus peak and holding temperatures (5 s
hold); (b) characteristic yield temperatures versus the elemental carbon
Abbreviations: see
Carbon: LW < ZP < SR < BL < IL < LK.
content.
Fig.5.1-1.
2.6 -
a
2.4
O LW
2.2 -
+ ZP
0 SR
ils BL
2.0 -
x
7
148
IL
LK
1.8 -
0
A
-1%
1.6 1.4 -
w
A
1.2 -
z
1.0
L5
0.8 -
O
0
X
x
x+
+
0+
0.6 x
XX
0.4 0.2 0.0
V
0A*
A
V
V
X+
+
'7z S
-
400
600
800
1000
800
1000
HOLDING TEMPERATURE (c)
PEAK TEMPERATURE (C)
900
b
880
V
V
V
860
840
0
820
0
IL
00
800
00
0
M
I-
780
760
AA
740
A
720
AA
700
I
70
78
74
0
82
86
I
90
I
I
94
ELEMENTAL CARBON CONTENT (WT% DMMF)
V
T75
A
725
T50
Figure 5.2-3 Comparison of ethylene production rate at 1 atm.
(a) combined plot of yields versus peak and holding temperatures (5 s
hold); (b) characteristic yield temperatures versus the elemental carbon
content.
Carbon: LW < ZP < SR < BL < IL < LK.
Abbreviations: see
Fig.5.1-1.
0.8
-
a
0 LW
0.7 -
x
+
0 SR
ta BL
X IL
V
0.6 -
X
149
9zP
V
X
LK
IL
VA
0.5 -
0.4 -
x
A
A
0
+
0.2
0
+
E
0.1 -
I A
86
400
0.0-
A
0
-
+
600
0
_
1000
800
1000
800
HOLDING TEMPERATURE (C)
PEAK TEMPERATURE (C)
b
840 820860
wb
780
wV
40 -
0
720 700 -
A
680 -
640 620 -
600
70
i
I
74
78
0
I
I
82
1
86
90
94
ELEMENTAL CARBON CONTENT (WT7% DMMF)
T75
V
A
T25
T50
Figure 5.2-4 Comparison of ethane production rate at 1 atm.
(a) combined plot of yields versus peak and holding temperatures (5 s
hold); (b) characteristic yield temperatures versus the elemental carbon
content.
Carbon: LW < ZP < SR < BL < IL < LK.
Abbreviations: see
Fig.5.1-1.
14
13
150
0
12
00
11
++
10
9
-j
w
AA
0
z
0
z
0
X
a3
X
400
600
800
800
1000
HOLDING TEMPERATURE
PEAK TEMPERATURE (C)
1 020
1000
(C)
b
vv
1 000
980
-
960
940
920
U
900
C
880
w
-
0
860
IL
840
-
A
820
800
780
760
740
.
70
74
78
0
82
86
90
94
ELEMENTAL CARBON CONTENT (WT% DMMF)
T50
A
T25
V
T75
Figure 5.2-5 Comparison of carbon monoxide production rate at 1 atm.
(a) combined plot of yields versus peak and holding temperatures (5 s
hold); (b) characteristic yield temperatures versus the elemental carbon
content.
Carbon: LW < ZP < SR < BL < IL < LK.
Abbreviations: see
Fig.5.1-1.
I
14
13 12 IL
11 -
a
LW
ZP
0 SR
A BL
X IL
0
+
7
151
LK
++
10 -
90
+
+
0
0+
5
x
+
0
70
0
0
8 -
+o
65-
z
A
4-
+0 +
2-
A
-
V
0
500
700
VV-V-
900
1100
PEAK TEMPERATURE (C)
800
X
X
X
X XX
1-
A
A
+0
300
A
A
0~
3-
800
1000
HOLDING TEMPERATURE (C)
-
780
bb
760 740
720 700 .1-
680 660 -
L)
:D
a_
10
640
0
620 600 580
560 540 520 500
70
74
78
82
86
90
94
ELEMENTAL CARBON CONTENT (WT% DMMF)
0
T50
A
T25
V
T75
Figure 5.2-6
Comparison of carbon dioxide production rate at 1 atm.
(a) combined plot of yields versus peak and holding temperatures (5 s
hold); (b) characteristic yield temperatures versus the elemental carbon
content.
Carbon: LW < ZP < SR < BL < IL < LK.
Abbreviations: see
Fig.5.1-1.
is
gas
assumed to be produced from ether groups in
and thus
group,
simple
a
such
in
problem
is asserted to be
Upon rapid pyrolysis
example.
is
decomposes along two parallel pathways,
moiety,
and
the
But a
in
the
following
above
750
C, phenol
illustrated
temperatures
at
functional
of coal type.
independent
picture
The rate of
type of
only on the
assumed to depend
is
production
the coal.
one of which gives CO and a C5
other H2 0 and benzene
[Cypres
and
Bettens
(1974),
(1975a,b)].
0
H
+
CO
+
H20
OH
0
H
The former pathway is
H
a base-catalyzed
reaction,
and thus is
expected
the coal
to be promoted by strong solid base materials from minerals in
such as CaO generated by calcite decomposition
Thus,
this mechanism applies
assuming
groups
in coal,
the
the
species,
and
strongly
influence
phenol group
concentration
for
(Franklin et al.,
decomposition of phenolic
can produce
of
several
different
gas
in
minerals
can
base-catalysts
the relative extent
1981).
of the two reaction paths.
A
more likely mechanism may be that gases are formed from a large set of
complex
initially
elementary
present,
reactions
involving,
not
but also many intermediate
just
the
structure
species that are formed
152
during coal decomposition, e.g., methyl radical side chains formed from
scission of ethylene bridges.
However, a better understanding of the
reaction mechanism is needed to apply such a description to rationalize
the observed gas production rate behavior among different coal types.
An
alternative
and
more
plausible
observable coal-type effects in
explanation
this study,
is
for
the
lack
that differences in
of
the
apparent gas production rates are less than or comparable to scatter in
the data caused by experimental uncertainties.
Supporting evidence for
this explanation comes from a recent study of Burnham et al. (1988), in
which
eight
coals
ranging
from
lignites
to
coals were pyrolyzed at low heating rates (<
They observed that Tmax
pressure.
maximum)
generally
low-volatile
1 C/s) under atmospheric
(T at which the evolution rate is
increases with coal rank, with maximum differences
ranging from 18 to 33 C among light hydrocarbons
differences
bituminous
are more
clearly resolved
in the
(CH4 ,C2 H 4, 2).
Such
slow heating apparatus
which is able to measure the sample temperature within ±5 C (Burnham et
al.,
in
1988).
the
In rapid heating studies such as this one, uncertainties
temperature
comparable
measurement
much
to the reported differences
the low-heating experiment.
and thus
are
higher
(~+25
caused by coal-type
of Tmax,
and
are
effects
in
Rates for carbon oxides show multi-peaks,
are more difficult to compare among different
the basis
C),
though comparing
the difference
in
coal types on
the first
peak
(<500 C) of CO2 production showed a difference of 86 C.
Pyrolytic
Despite
an
measurement
water measurements
extensive
technique,
condensation problems
effort
to
in
this
improve
study
the
are highly uncertain.
accuracy
of
the
water
interference from atmospheric water vapor
generated
large
scatter in
and
the measured values.
153
Therefore, no comparisons are made for the rate of water production.
154
5.3. Coal-type effects on total volatiles production
Figure
5.3-1
volatiles
versus
compares
the
the
yield
elemental
limit
carbon
of
total
and
'reactive'
for
the
six
content
coals
investigated in this study and the two coals studied by Suuberg (1977)
under
similar
but
not
identical
experimental
conditions.
Reactive
volatiles are defined as total volatiles minus water and carbon dioxide
yields.
The total yield limit ranges
lignites,
drops
and
subbituminous
from 41
and high-volatile
to 55 wt% dmmf among
bituminous
to 22 wt% dmmf for the low-volatile bituminous coal.
quantity to compare is
volatile
bituminous
reactive volatile yields,
coals
(BL,PB,IL)
produce
coals, but
A useful
which show that high-
significantly
more
than
other coal types.
Figure
5.3-2
compares
the
characteristic
total volatiles production for the six coals.
versus
temperature
for each of the
six coals
yield
temperatures
of
Plots of the total yield
are shown in
Fig.6.1-9.
The characteristic temperatures tend to increase for higher rank coals,
indicating
a
shift
in
the
yield
curve
to
higher
temperatures.
Comparing the spread of the yield curve, measured by (T75-T25), shows a
small
decreasing
trend
with
increasing
rank.
These
trends
are
consistent with the expected behavior from combining the observed coaltype effects on the rate of tar and gas production.
Such a consistency
together with a good product mass balance (Section 5.5), help to verify
the experimentally
product evolution.
observed coal-type effects
on the apparent rate of
155
60
N
0
0
50 I.-
2
2
0
U
40 -
2
30 -
0
w
U
20
-j
0
10 -
0
78
74
70
82
86
90
94
ELEMENTAL CARBON CONTENT (WTx DMMF)
Comparison of total and reactive volatiles yield limit
Figure 5.3-1
carbon content. Open symbols are from this study;
the
elemental
versus
closed symbols from Suuberg (1977). Carbon: LW < ML < ZP < SR < BL < PB <
IL < LK. Abbreviations: see Figs. 5.1-1 and 5.2.1.
800
780
760
740 V
720 700 680 660 640 2
620 600 580 560
A
540
520 500
70
78
74
0
82
ELEMENTAL CARBON CONTENT (WT
T25
A
T50
86
90
94
DMMF)
T75
V
Comparison of characteristic yield temperatures for total
Figure 5.3-2
Carbon: LW < ZP < SR < BL < IL < LK.
volatiles production at 1 atm.
Abbreviations: see Fig.5.1-1.
5.4. Pressure effects
156
The aim of this pressure study is to determine the extent to which
secondary
reactions
contribute
in
measured at atmospheric pressure.
the
pressures
overall
Changes in
pyrolysis
behavior
tar yields at different
are good indicators of the severity of secondary reactions.
Figure 5.4-1 shows tar yield limits for the six coals over the pressure
range
of 10-
Suuberg
eight
(1977)
coals,
to
10
atm;
the
are also plotted
the
increasing
Other investigators
limit.
yields
from the
in the
reactor
two
figure to
pressure
coals studied by
compare.
lowers
For all
the
including Ardent and van Heek
tar
yield
(1981)
and
Bautista (1984) have reported similar findings.
More quantitative observations on pressure effects can be made from
Fig.5.4-2, which shows
indicated pressure
vacuum
For
is
the
the % decrease in the tar yield limit at the
relative
the vacuum yield limit.
least affected by pressure
coals investigated
decrease at 1 atm ranges
pressure
to
sensitive
in this study
The yield at
secondary
(Fig.5.4-2a),
reactions.
the
relative
from 13 to 24 % of the vacuum value.
effect is expected to be
less at lower temperatures,
The
i.e.,
before the yield limit is reached, since secondary reactions of tar are
more severe at higher temperatures.
The extent of the pressure effect
is noticeably smaller for SR, and thus tars from this coal appear to be
less prone to secondary degradation.
However such differences can also
be attributed to experimental errors associated with tar measurements,
which can easily be ±10% of the vacuum yield limit for this coal.
be more conclusive, further studies are
needed in reactors
better suited to investigate tar secondary reactions,
stage flow-reactor (Serio, 1984).
To
that are
e.g.,
the two-
The results from Suuberg
(Fig.5.4-
157
40
35
30
IL
0
25
20
15
10
0
-1
-3
-5
LOG (P/ATM)
Effect of pressure on tar yield limit for different coals.
Figure 5.4-1
Open symbols are from this study; closed symbols from Suuberg (1977).
Carbon: LW < ML < ZP < SR < BL < PB < IL < LK. Abbreviations: see Figs.
5.1-1 and 5.2-1.
70
60
dP
50
0
40
30
Cd
20
10
0
LW
SR
ZP
P -
1 atm
BL
IL
COALS
P SM
LK
ML
PB
10 or 69 atm
Decrease in the tar yield limit relative to the 'vacuum'
Figure 5.4-2
yield. (a) coals from this study; (b) coals from Suuberg (1977).
10-3 atm in (a) and 6.6x10-5 atm in
Y*vac = tar yield limit at 'vacuum',
pressure, p = 10 atm in (a) and 69
given
a
at
limit
yield
tar
=
(b). Y*
< SR < BL < PB < IL <LK.
ZP
<
ML
<
LW
Carbon:
atm in (b).
5.2-1.
and
5.1-1
Figs.
see
Abbreviations:
2b)
show slightly larger pressure effects compared to those obtained in
this study, but the larger effects may be rationalized by lower vacuum
pressures attained in his study, 6.6x10~5 atm compared to 10-3 atm in
this study.
Figure 5.4-3 shows total volatiles yields over the pressure range
atm
and 6.6x10~5 to 69
to 10 atm for the six coals of this study,
of 10-3
for
reduces
the
two
studied by Suuberg.
coals
total volatiles
as
yields
some of
Figure 5.4-4 shows that for both Montana lignite
at higher pressures.
increasing the pressure produced
and Pittsburgh Seam bituminous coal,
yields.
converted to
is
greater gas yields are expected
yields at pressures other than 1 atm,
in
tar
the
Although this study did not measure gas
produce solid char and gases.
large gains
Increasing the pressure
but only small changes
methane yields,
in
carbon oxide
The trends for ethylene and ethane yields are less clear, but
to the overall gas production (~c 1
these species contribute very little
Assuming a similar behavior for the coals studied here
wt% as rec'd).
would suffice as a rough approximation.
Particle size is
the
75-90
sufficiently
induced
secondary reactions
However the small data base on particle-size effects suggests
of tar.
that
another variable that affects
small
secondary
unaffected
pm
diameter
to
avoid
reactions.
particles
major
between 300 and 800 pm (dia.)
in
contributions
The
between 50 and 300 pm
used
tar
(dia.),
(Suuberg,
yield
at
this
from
1
study
are
particle-size
atm
is
almost
and only slightly affected
1977; Bautista,
1984).
Further
studies on particle-size effects are currently being pursued by Griffin
(1988).
158
159
60
I
LI
A
+
0Y
Ai
'V
X
V
55 Ii0
50 -
LW
ML
ZP
SR
BL
PB
IL
LI<
A
x
I-
45 L -----------
0
-J
w
40
-
C,)
w
-j
35 -
0
-J
30 -
I0
I-
25 -
20
i
-5
I
I
-3
-1
1
LOG (P/ATM)
Effect of pressure on total volatiles yield limits for
Figure 5.4-3
Open symbols are from this study; closed symbols from
different coals.
Carbon: LW < ML < ZP < SR < BL < PB < IL < LK.
Suuberg (1977).
Abbreviations: see Figs. 5.1-1 and 5.2-1.
P
100
1 ATM
=
aa
160
90 8070 0
PB
X
50 -
40 o
ML
30
20
>
10
-10-
/
CH4
C2H4
C2H6
P
200
-
CO
C02
CO
C02
69 ATM
b
180 -
ML
160 140 120
C)
PB
-
1
80S
60 -
>
40
20
CH4
C2H4
C2H6
Figure 5.4-4 Effect of pressure on gas yield limits for Montana lignite
(ML) and Pittsburgh Seam bituminous coal (PB).
(a) p = 1 atm; (b) p = 69
atm.
Data from Suuberg (1977).
Y*i,vac = yield of gas species i at
t
Ya
a
6.6x100 -5 atm;
Y~~
= yield of species I at pressure p.
5.5. Experimental uncertainties
Two
main
sources
measurements
consistent
of
experimental
temperatures
(2)
spread the
two
layers of
it
sample
the
uncertainties
and
temperature measurements,
sample sizes,
between the
sample
of
161
arise
To obtain
product yields.
is crucial to
thinly,
screen.
and (3)
The
from
(1) use small
minimize the gap
first two
criteria are
satisfied by evenly spreading -2 20 mg samples on a 5 cm x 10 cm screen,
whereas the last is satisfied by keeping the loaded screen between the
electrodes as tightly as possible.
Under such conditions, the sample
temperature can be consistently measured within
(1985)
show that
temperature
when
the particle temperature
small coal particles
(
+ 20 C.
Studies by Oh
closely follows
100
pm dia.)
the screen
are pyrolyzed
under atmospheric pressure at heating rates of - 1000 C/s.
Under such
conditions,
isothermal
(Hajaligol
the
et
atmospheric,
coal
al.,
particle
1988).
temperature
At
can
be
assumed
pressures
measurements
to
be
considerably
using the current
lower
than
technique are
uncertain due to slower convective heat transfer from the hot screen to
the coal particle (Oh, 1985).
Thus, this study does not use any vacuum
data to obtain kinetic information.
The magnitude of uncertainties from product yield measurements was
estimated to be the maximum difference
of
2-4
runs
repeatability
made
tests,
under
the
similar
in
the measurements
conditions.
uncertainty
for
a
Based
given
from a set
on
these
product
conservatively estimated to be
Products
tar
char
gases except H2 0
H2 0
Measurement uncertainties
±2 wt% of unpyrolyzed coal, dmmf
±2 wt% of unpyrolyzed coal, dmmf
±0.05-0.5 wt% of unpyrolyzed coal, dmmf
±3 wt% of unpyrolyzed coal, dmmf
is
For
runs in which both tars
and gases were collected, mass balances
were calculated to be between 90 and 110 wt% of the original coal mass.
A large portion of the uncertainty is due to large scatter in water
measurements.
95-105
Considerably better closures were
%) when mass balances
pyrolytic water yield.
obtained
(typically
are computed using an averaged value of
162
163
6. Modeling studies
This chapter derives kinetic information from the experimental data
obtained
in this
study
multiple independent
MIPR model.
(Chapter 5) using
parallel
reaction
two
(MIPR)
different
model,
The former model describes kinetics
models:
the
and the extended
of product evolution
under conditions where the effects of physical transport processes and
secondary
reactions
are
relatively
unimportant.
The
latter model
explicitly includes transport and secondary reaction effects, and thus
is applicable over a wider range of operating
conditions.
Sections
6.1.1 and 6.2.1 give mathematical descriptions of the MIPR and extended
MIPR models respectively; the kinetic information obtained with the two
models is discussed in Sections 6.1.2 and 6.2.2 respectively.
6.1. Multiple independent parallel reaction (MIPR) model
The MIPR model has been widely used to describe the evolution rate
of tar
Ngan,
(Serio,
1979;
1984;
Serio,
Ciuryla et al.,
Ko et al.,
1984),
1988a),
gaseous products (Weimer and
and total volatiles
(Anthony
1979; Sprouse and Schuman, 1981).
not include an explicit description of mass
et al.,
1974;
Since the model does
transport,
it
is
strictly
valid only under conditions where mass transfer resistances are small.
Under
such
conditions, the
model
is capable
of describing volatiles
evolution rates over a wide range of heating rates.
6.1.1. Mathematical description
The rate of volatiles evolution in
the
sum
of
the
contributions
independent parallel reactions,
from
the MIPR model is
a
large
number
of
expressed as
first-order
dY/dt = X k0 i exp(-Ei/RT)
where
(Y* -Yi)
i denotes one reaction.
for all
reactions,
described
k0 i
i.e.,
by a Gaussian
(6.1-1)
The same preexponential factor is used
ko,
=
and the
distribution
activation energies
function
f(E)
with mean
E0
are
and
standard deviation a
f(E)
=
[u(2wr) 1 12 ]-1 exp[-(E-E0 )2 /2U 2 ]
(6.1-2)
The probability of finding a reaction with activation energy between E
and E+dE
=
f(E)
is given by f(E)dE, where for a large number of reactions,
Y*j/Y*
and
Y*
equal
to
the
sum
of
the
Y*i
for
all
i.
Integrating Eq.(6.1-1) for any temperature-time history gives
o
(Y*-Y)/Y*
t
exp[-ko
=
o
where Y* , E, , a, and ko
model,
and in
general
exp(-E/RT)dt] f(E)dE
(6.1-3)
0
are the input parameters required in
temperature
(T)
is
the MIPR
a function of time (t).
The
notation 'Y' here is equivalent to 'V' in earlier descriptions of this
model (Anthony et al., 1974; Howard, 1981).
6.1.2. Results and discussion
This
section discusses
the effect of coal type
on the MIPR model
parameters for tar, gas, and total volatiles production.
Coal-type effects on tar production
Figure
6.1-1
compares
the
experimental
and predicted tar
yields
from the MIPR model for the six coals investigated in this study.
model predictions were made with ko fixed at 1014
the
measured
experimental
maximum
tar yield,
and
E0
and
S-1, Y*
a
data using a multivariable non-linear
The
obtained from
best-fitted
to
the
regression routine
164
15
165
14
0
13
0
0
12
01
01
11
10
9
1-1
wL
8
6
5
4
3
2
-j-
0
300
500
700
900
1100
800
1000
8S
b
7an
6-
a
0
01
5_j
0
03
4-
0
0
3-
20
300
500
700
900
PEAK TEMPERATURE (C)
1100
800
1000
HOLDING TEMPERATURE (C)
Figure 6.1-1
Tar yields versus peak and holding temperatures (5 s
hold).
Symbols represent experimental data; lines represent MIPR model
predictions.
(a)
LW, Lower Wilcox lignte; (b) ZP, Beulah Zap lignite;
BL, Blue high-volatile
(d)
Smith
Roland subbituminous;
(c)
SR,
(f)
LK, Lower
high-volatile bituminous;
IL,
Illinois
bituminous; (e)
Kittanning low-volatile bituminous.
166
C
0
0
0
0
0
L
0
0
-J
w
I-
I
I
300
26 -C
24 22
500
700
900
1000
0
H
0
1
U
20 L
|
800
1100
0
0
18
16
14
0
12
10
8
6
42
-
0
i
300
I
E i
500
700
i
900
PEAK TEMPERATURE (C)
1100
800
1000
HOLDING TEMPERATURE (C)
Tar yields versus peak and holding temperatures (5 s
Figure 6.1-1
Symbols represent experimental data; lines represent MIPR model
hold).
predictions. (a) LW, Lower Wilcox lignte; (b) ZP, Beulah Zap lignite;
(d) BL, Blue high-volatile
Smith Roland subbituminous;
(c) SR,
bituminous; (e) IL, Illinois high-volatile bituminous; (f) LK, Lower
Kittanning low-volatile bituminous.
28
00
26
167
0
0
0
24
Do
22
20
LL
18
16
14
-J
12
10
8
6
4
2
0
300
12 -11
500
700
900
1100
1000
800
-
10-
0
3
0
1
9IL
8
7
0j
6
5
I 3
4
3
2
1
0
300
500
700
900
PEAK TEMPERATURE (C)
1100
800
1000
HOLDING TEMPERATURE
(C)
Tar yields versus peak and holding temperatures (5 s
Figure 6.1-1
Symbols represent experimental data; lines represent MIPR model
hold).
predictions. (a) LW, Lower Wilcox lignte; (b) ZP, Beulah Zap lignite;
(d) BL, Blue high-volatile
(c) SR, Smith Roland subbituminous;
bituminous; (e) IL, Illinois high-volatile bituminous; (f) LK, Lower
Kittanning low-volatile bituminous.
(IMSL math
library subroutine
ZXSSQ).
In all
cases,
yields agree well with the experimental values;
the
predicted
the standard error of
the estimate, as defined in Eq.(5.1-31), ranges from 6.5 to 10 % of the
maximum tar yield.
Figure
elemental
6.1-2 plots the best-fitted values
carbon contents
Table 6.1-1.
in
of the coal;
of E0
numerical values
are given in
These values are slightly different than those presented
an earlier report
(Howard
et al.,
1988).
The previously reported
values were obtained with assumed approximate
histories,
and a versus the
whereas
the
current values were
linear temperature-time
obtained with more exact
temperature histories and therefore are expected to be more accurate.
The figure shows that higher rank coals, indicated by higher elemental
carbon contents,
of
a.
generally gave greater values of E0 and smaller values
Maximum
respectively.
differences
both
employed
E0
and
E0
a are 7.1 and 3.6 kcal/mole
Such differences far exceed the variation explainable by
experimental uncertainties,
for
in
and
in this
a
(see
study,
estimated to be approximately ±1 kcal/mole
below).
Therefore,
there appears
to
be
under
the
conditions
a convincing coal-type
effect on the MIPR model rate parameters for tar production.
Uncertainties associated with E0
(AE0 ) and a (Ac)
independently considering experimental
yield
(AY) measurements,
generated
from
experimental
standard
the
error
error
of
two
in
the
temperature measurement
the
'measured' and
and
are
sources.
the
yield
estimate
is
errors
assumed
For
in
to
each
measurement
are estimated by
temperature
be
the
individual
is
[Eq.(5.1-31)].
(AT)
maximum
and
value
product,
the
approximated as
the
The
error
in
the
estimated as the average difference between
'predicted' temperatures
for
given
experimental
168
56 -
169
a
55 -
54 0
0
53 03
E
'N
"a
cJ
52 -
0
wU
51 -
E3/
50 -
49 -
48
I
I
I
I
70
74
I
I
78
I
I
82
I- - - _
I
I
86
1
90
1
13
7.5
1
94
7.0 -
6.5 -
0
6.0 -
E
NY
5.5 -
0
5.0 -
4.5 -
03
4.0 -
3.5
.1
70
I
--
I
I
74
78
0
0
I
82
I
I
86
I
90
94
ELEMENTAL CARBON CONTENT (WTZ DMMF)
Best-fitted values of (a) E0 and (b) a for predicting
Figure 6.1-2
evolution using the MIPR model versus the elemental
tar
atmospheric
1
carbon content of the coal. ko was fixed at 1014 s~ for all coals; Y*
Dashed lines for
was obtained from experimental data for each coal.
Carbon:
coals with [H] < 5 wt% dmmf; solid lines for [H] ;! 5 wt% dmmf.
LW < ZP < SR < BL < IL < LK. Abbreviations: see Fig.6.1-1.
yields
about
using total volatiles
±25
C.
The
data
on
yield data,
total
and
volatiles
is
approximated to
yields
chosen
were
be
in
estimating AT since they have the largest number of data points among
The AEO was estimated to be the maximum deviation
measured products.
from the best-fitted value of E0 caused by the estimated AT and AY; the
deviation was computed at 50% of the final yield for cases where the
temperature and the yield were consistently either too high or too low
by AT/2 and AY/2 respectively.
Similarly, the Au was estimated to be
the maximum deviation from the best-fitted value of a, and is computed
at 25
the
and 75% of the final yield for cases where the temperature
yield were
respectively
consistently too
to produce
high
or
the most and the
too
low by
and
AT/2 and AY/2
least spread
in the yield
curve.
The
trends
for
scattered among
both
E0
and
low-rank coals,
a
in
Fig.6.1-2
appear
where the Beulah Zap
to
be
more
lignite shows a
considerably lower E. and higher a compared to the Lower Wilcox lignite
and Smith Roland subbituminous coal.
the maximum
tar yield
(dmmf) compared
coals
(Y*),
where
to about 13 wt%
(Table 6.1-1).
These
A similar distinction is noted in
the Zap produced
(dmmf) from the
only about 7 wt%
other two
consistent differences
suggest
low-rank
that the
information on coal rank alone is insufficient to explain the observed
different behavior for coals within the same rank.
One property that
appears to distinguish the two types of low-rank coals is the elemental
hydrogen content;
in
dmmf basis,
the Zap has 4.8 wt% whereas the Lower
Wilcox and Smith Roland have noticeably
wt% respectively.
Therefore
the low-rank region,
in
larger values
of 5.6
and 5.3
estimating E0 and a from Fig.6.1-2 in
the dashed curves are recommended for coals with
170
Table 6.1-1
Best-fitted values of E0 and a of the MIPR model for tar
production (ko fixed at 10" s-1 for all coals).
Coala
Y
wt% dmmf
a
kcal/mole
Standard error
kcal/mole
of estimate
wt% dmmf coal
13.1
51.6
5.5
1.3
6.5
48.7
7.2
0.7
Smith Roland SB
12.9
50.4
5.2
1.4
Blue HVB
21.2
52.7
3.7
1.7
Illinois HVB
24.8
53.3
4.5
2.0
Lower Kittanning
LVB
10.7
55.8
3.6
0.7
Lower Wilcox L
Beulah Zap L
a Coals are listed in the order of increasing elemental carbon contents
in dmmf basis. Elemental analysis is given in Table 4.1-1.
171
the
elemental
hydrogen
content
of <
5 wt% dmmf,
for coals with the elemental hydrogen of
and the solid curves
t 5 wt% dmmf.
Coal-type effects on gas production
Figures
gas yields
study.
6.1-3 through 6.1-7 compare the experimental and predicted
from the MIPR model
The model
above for tar,
maximum
using
tar
a
species,
for
the six coals
predictions
were
made
with k, fixed at
1014
s-1,
yield,
and E0
multivariable
using
the
this
same procedure
as
Y* obtained from the measured
and a best-fitted
non-linear
investigated in
to
regression
the
experimental
routine.
For
data
all
gas
the agreement between the predicted and experimental yields is
generally good;
the standard error of the estimate ranges from 4 to 15
% of the maximum yield.
Figure
elemental
Table
6.1-8
carbon
and
3.9
the best-fitted
contents
of
the
For hydrocarbon
6.1-2.
increase in E0
2.7
plots
values
coal;
gases,
of E0
numerical
methane
and a versus
values
are
and ethane
given
show a
the
in
small
for higher rank coals, with maximum variations of about
kcal/mole
almost unaffected.
respectively,
whereas
ethylene
appears
to
be
Comparing a of hydrocarbon gases shows a decreasing
trend as the coal rank increases for methane and ethylene, with maximum
variations
of about
2 and 2.3
are observed for ethane.
small,
and
are
only
from experimental
for EO
kcal/mole
These
slightly
uncertainties,
respectively,
variations
greater
in
the
but no
effects
rate parameters
than estimated
are
errors produced
which range from ± 0.5 to 1 kcal/mole
and from ± 1 to 1.5 kcal/mole
for u.
Large scatters
in
E0
and a
of ethane for the three low-rank coals are noted, and are attributed to
very low yields among these coals (<0.4 wt% dmmf).
172
173
2.2
2.0
1.8
z
1.6
1.4
1.2
0
K
0
1.0
0.8
0.6
0.4
0.2
0.0
400
600
800
1000
800
1000
400
600
800
1000
800
1000
-
1.8
1.7
1.6
1.5
1.4
LL
1.3
1.2
1.1
1.0
0.9
0.8
ELI
z
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
PEAK TEMPERATURE (C)
HOLDING TEMPERATURE
(C)
Figure 6.1-3 Methane yields versus peak and holding temperatures (5 s
hold). Symbols represent experimental data; lines represent MIPR model
predictions.
(a) LW; (b) ZP; (c) SR;
(d) BL; (e) IL; (f) LK.
Abbreviations: see Fig.6.1-1.
174
2.6
2.4
2.2
2.0
IL
M.
1.8
1.6
M
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
400
600
800
1000
800
1000
600
800
1000
800
1000
-
3.0 - C
2.8 2.6 2.4 IL
2.2 2.0 1.8 1.6 -
-J
w
1.4
w
1.2 -
M
1.0 -
z
0.8 0.6 0.4 0.2 0.0 -400
PEAK TEMPERATURE (C)
HOLDING TEMPERATURE (C)
Figure 6.1-3 Methane yields versus peak and holding temperatures (5 s
hold).
Symbols represent experimental data; lines represent MIPR model
predictions.
(a) LW; (b) ZP; (c) SR;
(d) BL;
(e) IL; (f) LK.
Abbreviations: see Fig.6.1-1.
4.0
175
3.5
0
3.0
e-%
LL
2.5
2.0
1.5
1.0
0.5-
0.0 400
600
800
1000
800
1000
-
4.5 - -
0
4.03.5Li~
12
3.0
2.5
12
-J
LI
5:
2.0
LI
z
I
1.5
Li
1.0 -
0.5 0.0 -400
I
600
800
PEAK TEMPERATURE (C)
1000
800
I
I
1000
HOLDING TEMPERATURE (C)
Methane yields versus peak and holding temperatures (5 s
Figure 6.1-3
Symbols represent experimental data; lines represent MIPR model
hold).
(d) BL; (e) IL; (f) LK.
(a) LW; (b) ZP; (c) SR;
predictions.
Abbreviations: see Fig.6.1-1.
1.8
-
-
176
1.7 1.6 1.5 -
01
0
1.4 Ii.
1.3 1.2 1.1 1.0
0.9 0.8 -
z
0.7 -
r
LJ
0.6
0.5 0.4 0.3 0.2 0.1 -
I
0.0 -r400
600
800
1000
I
800
1000
0.8
0
01
0.7
Li.
0
0.6
0.5
0
0.4
z
La
0.3
-J
I
0.2
0.1
0.0
400
600
800
PEAK TEMPERATURE (C)
1000
800
1000
HOLDING TEMPERATURE
(C)
Figure 6.1-4 Ethylene yields versus peak and holding temperatures (5 s
hold).
Symbols represent experimental data; lines represent MIPR model
predictions.
(a) LW; (b) ZP; (c) SR; (d) BL; (e) IL; (f) LK.
Abbreviations: see Fig.6.1-1.
1.5
1 177
1.4
1.3
C
1.2
EL
1.1
C
1.0
0.9
0
0.8
-I
U
0.7
U
0.6
z
U
-j
0.5
0.4
0.3
0.2
0.1
0.0
400
600
800
1000
800
1000
2.2
2.0
1.8
1.6
1.4
1.2
0
EL
1.0
0.8
0
0.6
0.4 0.2 0.0 -f400
600
800
PEAK TEMPERATURE (C)
1000
800
1000
HOLDING TEMPERATURE
(C)
Figure 6.1-4 Ethylene yields versus peak and holding temperatures (5 s
hold). Symbols represent experimental data; lines represent MIPR model
(a) LW; (b) ZP; (c) SR; (d) BL; (e) IL; (f) LK.
predictions.
Abbreviations: see Fig.6.1-1.
1.1
I
178
1.0
0.9
IL
0
0.8
0.7
w
0.6
0.5
z
0.4
0.3
0.2
0.1
0.0
I
400
600
800
1000
I
1000
800
0.7
0.6
LI
L-
0.5
a
0.4
0.3
z
LiJ
-
L
0.2
0.1
0.0
I
400
600
800
PEAK TEMPERATURE (C)
1000
800
I
1000
HOLDING TEMPERATURE (C)
Figure 6.1-4 Ethylene yields versus peak and holding temperatures (5 s
hold).
Symbols represent experimental data; lines represent MIPR model
SR;
(d)
BL; (e)
IL; (f)
LK.
predictions.
(a) LW; (b) ZP;
(c)
Abbreviations: see Fig.6.1-1.
0.40
1 179
I a
0.35 -
0.30 L
0.25 El
El
0.20 -
w
z
0.15 11
0.10 -
0.05 -
0.00
400
600
800
1000
1000
800
-
0.30
0.28
0.26
0.24
r1
0.22
0.20
e-j
0.18
0
0.16
0.14
w
5:
U
0.12
0.10
0.08
0.06
0.04
0.02
0.00
400
600
800
PEAK TEMPERATURE (C)
1000
800
1000
HOLDING TEMPERATURE
(C)
Ethane yields versus peak and holding temperatures (5 s
Figure 6.1-5
Symbols
represent experimental data; lines represent MIPR model
hold) .
(e) IL; (f) LK.
(a) LW; (b) ZP; (c) SR; (d) BL;
predictions.
Fig.6.1-1.
see
Abbreviations:
180
0.50
0.45
0.40
0.35
00
0
0.30
0.25
w
w
0.20
z
0.15
0.10 0.05 0.00 400
600
800
1000
1000
Bo0
-
0.6
0.5
00
0.4
IL
0.3
z
0.2
0.1
0.0
400
600
800
PEAK TEMPERATURE (C)
1000
800
1000
HOLDING TEMPERATURE (C)
Figure 6.1-5
Ethane yields versus peak and holding temperatures (5 s
hold).
Symbols represent experimental data; lines represent MIPR model
predictions.
(a) LW; (b) ZP;
(c) SR; (d) BL; (e) IL; (f) LK.
Abbreviations: see Fig.6.1-1.
181
0.8
0.7
0.6
L
0.5
0.4
-LJ
z
0
0.3
2:
0.2
0.1
0.0
400
600
800
1000
800
1000
0.7
0
0.6
0.5
L
0.4
0.3
z
IL
0.2
0.1
0.0
400
I
600
800
PEAK TEMPERATURE (C)
1000
800
I
1000
HOLDING TEMPERATURE (C)
Figure 6.1-5
Ethane yields versus peak and holding temperatures (5 s
hold).
Symbols represent experimental data; lines represent MIPR model
predictions.
(a) LW; (b) ZP; (c) SR; (d) BL; (e) IL; (f) LK.
Abbreviations: see Fig.6.1-1.
182
13
C
12
C
11
0
IL
10
9
8
w
w
0
0
7
6
z
5
z
4
0
0
M
3
2
0
0
400
600
800
1000
800
1000
12
11
1-1
10
Cl
0
0
-J
z
0
z
1
0
400
600
800
PEAK TEMPERATURE (C)
1000
800
1000
HOLDING TEMPERATURE
(C)
Figure
6.1-6
Carbon monoxide
yields
versus
peak
and holding
temperatures (5 s hold).
Symbols represent experimental data; lines
represent MIPR model predictions.
(a) LW; (b) ZP; (c) SR; (d) BL; (e)
IL; (f) LK. Abbreviations: see Fig.6.1-1.
183
11
10
9
L
8
7
0
6
X
0
z
0
z
0
5
4
3
0
2
1
0
400
600
800
1000
800
1000
8
7
L
6
5
0
4
0
0
3
z
0
0
2
0
|
400
600
I
I
800
1000
PEAK TEMPERATURE (C)
I
I
800
1000
HOLDING TEMPERATURE (C)
Carbon monoxide yields versus peak and holding
6.1-6
Figure
Symbols represent experimental data; lines
temperatures (5 s hold).
(a) LW; (b) ZP; (c) SR; (d) BL; (e)
represent MIPR model predictions.
IL; (f) LK. Abbreviations: see Fig.6.1-1.
184
4.0
3.5
3.0
2.5
0
0
2.0
:
z
0
1.5
0
1 .0
0.5 -
0.0400
600
800
1000
800
1000
400
600
800
1000
800
1000
1.0
0.9
L.
0.8
0.7
0.6
0
0.5
0
0
0.4
z
0.3
0
Mf
0.2
0.1
0.0
PEAK TEMPERATURE (C)
HOLDING TEMPERATURE
(C)
peak and holding
Carbon monoxide yields versus
6.1-6
Figure
Symbols represent experimental data; lines
temperatures (5 s hold).
(a) LW; (b) ZP; (c) SR; (d) BL; (e)
represent MIPR model predictions.
IL; (f) LK. Abbreviations: see Fig.6.1-1.
185
10
0
9
0
8
/-N
7
6
w
5
w
0
0
4
z
3
0
M
2
0
12 -
900
700
500
300
1000
800
1100
b
11-
0
0
10 LL
9-
El0
:
0
82
2
7-
0
0
0
z
0
M
0:
0
6-
5-
0
43-
0
21 --
0
I
300
I
500
I
7
9 1
700
900
PEAK TEMPERATURE (C)
1100
800
1000
HOLDING TEMPERATURE
(C)
and holding
peak
Carbon dioxide yields versus
6.1-7
Figure
Symbols represent experimental data; lines
temperatures (5 s hold).
(a) LW; (b) ZP; (c) SR; (d) BL; (e)
represent MIPR model predictions.
IL; (f) LK. Abbreviations: see Fig.6.1-1.
186
9
0
8
0
0
L
7
6
-j
0
5
0
4
0
z
x)
0
a:
3
2
1
0
300
500
700
900
1100
1000
800
4.5
4.0
Li
3.5
U
0
3.0
L
0
2.5
0
2.0
0
z
1.5
0
0
1.0
0.5
0.0
I
300
500
700
900
PEAK TEMPERATURE (C)
1100
800
I
1000
HOLDING TEMPERATURE (C)
Figure
6.1-7
Carbon dioxide yields versus
peak and
holding
temperatures (5 s hold).
Symbols represent experimental data; lines
represent MIPR model predictions.
(a) LW; (b) ZP; (c) SR; (d) BL; (e)
IL; (f) LK. Abbreviations: see Fig.6.1-1.
2.2
-
187
e
2.0 -
0
L
1.6 1.4 -
x
1.0 0
1.0
0
0
a
z
0
5~
1
.
0.8 1.20.6 0.4 0.2 0.0 -
EP
500
300
0.6
-
700
900
1100
800
1000
900
1100
800
1000
_________________________
f
0.5
IL
0.4
w
w
0.3
0.
0
0
z
0.2
0
0.
B
0.1
300
500
700
PEAK TEMPERATURE (C)
HOLDING TEMPERATURE (C)
holding
and
peak
versus
yields
dioxide
Carbon
6.1-7
Figure
Symbols represent experimental data; lines
temperatures (5 s hold).
LW; (b) ZP; (c) SR; (d) BL; (e)
(a)
represent MIPR model predictions.
IL;
(f) LK. Abbreviations: see Fig.6.1-1.
-
188
A
CO
70
A
A
68 -
CH4
66 00
64 -
0
E
62-
+
o+
.-
+ C2 H4
+
+
0C2H6
uJ
60 -
58 -
x
56 -
C02
-x
x
x
54 82
78
74
70
86
94
90
10 -
b
x
x
x
X
9 -
8 00
E
7c
A
++
+
0
0~
3 -
I
70
0
CH4
74
+
i
I
78
I
I
82
I
I
86
90
ELEMENTAL CARBON CONTENT (WT% DMMF)
A
CO
C2H6
o
C2H4
X
94
C02
Best-fitted values of (a) E0 and (b) a for predicting
Figure 6.1-8
atmospheric pressure gas evolution usin the MIPR model versus carbon
contents of the coal. ko was fixed at 1 0 i4 s-1 in all cases, and Y* was
obtained from experimental data for each coal.
BL < IL < LK.
Abbreviations: see Fig.6.1-1.
Carbon: LW < ZP < SR <
Table 6.1-2
Best-fitted values of E0 and a of the MIPR model for gas
production (ko fixed at 1014 s-1 for all coals).
(a) methane
Coala
Y
wt% dmmf
EO
kcal/mole
a
kcal/mole
Standard error
of estimate
wt% dmmf coal
Lower Wilcox L
2.0
64.5
7.0
0.1
Beulah Zap L
1.6
63.8
7.9
0.2
Smith Roland SB
2.2
63.3
7.7
0.2
Blue HVB
2.8
63.5
6.6
0.2
Illinois HVB
3.4
65.5
6.5
0.1
Lower Kittanning
LVB
4.3
66.5
5.9
0.4
a Coals are listed in the order of increasing elemental carbon contents
in dmmf basis. Elemental analysis is given in Table 4.1-1.
(b) ethylene
Coal
Y
wt% dmmf
E0
kcal/mole
a
kcal/mole
Standard error
of estimate
wt% dmmf coal
Lower Wilcox L
1.6
63.0
5.8
0.1
Beulah Zap L
0.68
63.4
6.2
0.07
Smith Roland SB
1.3
62.0
5.6
0.09
Blue HVB
1.6
62.3
5.4
0.2
Illinois HVB
0.91
62.3
4.7
0.08
Lower Kittanning
LVB
0.55
63.3
3.9
0.05
189
Table 6.1-2 (continued)
190
(c) ethane
Coal
Y*
wt% dmmf
Ea
kcal/mole
a
kcal/mole
Standard error
of estimate
wt% dmmf coal
Lower Wilcox L
0.3
59.1
4.2
0.04
Beulah Zap L
0.23
60.5
5.3
0.03
Smith Roland SB
0.36
57.3
3.9
0.05
Blue HVB
0.5
57.1
3.3
0.06
Illinois HVB
0.7
60.0
5.5
0.07
Lower Kittanning
LVB
0.6
61.2
4.7
0.05
Y
wt% dmmf
EO
kcal/mole
a
kcal/mole
Standard error
of estimate
wt% dmmf coal
11.0
69.0
6.6
1.0
9.1
68.3
7.5
1.2
10.0
67.1
8.8
0.8
Blue HVB
7.3
65.9
8.7
0.3
Illinois HVB
3.8
66.1
7.1
0.3
Lower Kittanning
LVB
0.92
69.7
6.4
0.07
(d) carbon monoxide
Coal
Lower Wilcox L
Beulah Zap L
Smith Roland SB
Table 6.1-2 (continued)
(e)
191
carbon dioxide
a
kcal/mole
Y*
wt% dmmf
E0
kcal/mole
Lower Wilcox L
8.5
55.4
9.0
0.8
Beulah Zap L
9.5
55.9
9.4
1.0
Smith Roland SB
8.0
55.7
9.7
0.8
Blue HVB
3.5
54.8
9.2
0.5
Illinois HVB
1.8
57.3
6.4
0.2
Lower Kittanning
LVB
0.42
55.8
4.9
0.03
Coal
Standard error
of estimate
wt% dmmf coal
192
60
50
o
1
LL
C
40
-
30
0
Id
20
10
0
-f
300
60
.
500
700
900
1100
800
1000
1100
800
1000
b
50 01
0
C1
40 -
0
30 -
01
C)
20
0
El
10
n
03
0
|
300
500
700
900
PEAK TEMPERATURE (C)
HOLDING TEMPERATURE (C)
Figure
6.1-9
Total volatiles yield versus
peak and holding
temperatures (5 s hold).
Symbols represent experimental data; lines
represent MIPR model predictions.
(a) LW; (b) ZP; (c) SR; (d) BL; (e)
IL; (f) LK. Abbreviations: see Fig.6.1-1.
60
1
0
o
50
193
IL
40
30
w
-J
0
i
20
-J
0
10
-t-
0
300
500
700
900
1100
10 00
800
50
00
IL
0
40
30
0
Fl
0
0
20
10
0
300
500
700
900
PEAK TEMPERATURE (C)
1100
800
10 00
HOLDING TEMPERATURE (C)
peak and holding
Total volatiles yield versus
6.1-9
Figure
Symbols represent experimental data; lines
temperatures (5 s hold).
(a) LW; (b) ZP; (c) SR; (d) BL; (e)
represent MIPR model predictions.
IL; (f) LK. Abbreviations: see Fig.6.1-1.
60
1 194
e
0
50 -
0
0
0
O
U
0
Li
LL
O
130o
40 -
En
C
-/
o
-I
LLI
30 -
0
o
0
20 03
10 -
-+-
0
300
I
500
I
I
700
I
900
I
1000
800
1100
26
o
24
0
22
LL
20
18
16
Q
14
V)
12
0
10
0
8
6
42
-
0
300
500
700
900
PEAK TEMPERATURE (C)
1100
I
800
I
1000
HOLDING TEMPERATURE
peak and holding
yield versus
Total volatiles
6.1-9
Figure
Symbols represent experimental data; lines
temperatures (5 s hold).
(a) LW; (b) ZP; (c) SR; (d) BL; (e)
represent MIPR model predictions.
IL; (f) LK. Abbreviations: see Fig.6.1-1.
(C)
Comparing E. of carbon monoxide shows a concave upward trend with a
minimum
near
between the
high-volatile
bituminous
coals.
But
the
differences
three low-rank coals and the two high-volatile bituminous
coals are small,
E, = 67-69 kcal/mole
66 kcal/mole for the latter.
for the former compared to E.
~
Such differences are small considering
that about ± 1 kcal/mole deviations
in
EO
of carbon monoxide account
for uncertainties from experimental errors.
A modest increase to about
70 kcal/mole for the low-volatile bituminous coal is of less practical
interest since the carbon monoxide yield for this coal is very small (<
1 wt% dmmf).
Comparing a of carbon monoxide shows a slight decreasing
trend for most coals
as the
are within uncertainties
but the differences
produced from experimental
to be about ±1.5 kcal/mole.
coal-type
coal rank increases,
errors,
estimated
The E0 of carbon dioxide shows almost no
effect, whereas the a
shows
a clear
decreasing trend
higher rank coals, with a maximum difference of 4.5 kcal/mole.
for
The
uncertainties in the rate parameters of carbon dioxide are estimated to
be comparable to those of carbon monoxide.
Coal-type effects on total volatiles production
Figure
6.1-9
compares
the
experimental
and
predicted
total
volatiles yields from the MIPR model for the six coals investigated in
this study.
as above,
The model predictions were made using the same procedure
where ko
was again fixed at 1014
the measured maximum total volatiles
fitted to the experimental data.
In
yield,
s-1,
Y* was obtained from
and EO
all cases,
and a were best-
the predicted yields
agree well with experimental values; the standard error of the estimate
ranged from approximately 6 to 10 % of the maximum yield.
195
Figure 6.1-10 plots the best-fitted values of E0
elemental carbon
Table 6.1-3.
contents of the coal;
and a versus the
numerical values
are given in
Generally, higher rank coals show increasing values of E
0
with a maximum difference of about 6 kcal/mole.
lignite appears
to be high compared to the
The E0 of Lower Wilcox
other two low-rank coals,
but is within estimated uncertainties of ±1 kcal/mole.
Comparing the a
shows a decreasing trend for higher rank coals, but with much scatter.
A maximum difference of 4 kcal/mole in
a is
slightly greater than the
estimated uncertainty of ±1.5 kcal/mole.
The trends in the rate parameters for independently measured total
volatiles
production
products.
The relatively modest coal-type dependence of the MIPR rate
confirm
the
trends
observed
for
individual
parameters for total volatiles reflect the combined effects of a strong
coal-type dependence for tars and a much weaker dependence for gases.
Also,
the
general
trends
for
all
products
are
always
consistent-
higher E, and lower a for increasing coal rank.
Use
of
should be
the MIPR model
parameter
values
obtained
in
this
study
strictly confined to pyrolysis conditions similar to those
employed in this study, where small coal particles (75-90 pm dia.) were
rapidly pyrolyzed (~ 1000 C/s) to a maximum temperature of about 1000 C
under
atmospheric
pressure
volatile products.
reactions
inside
with
rapid
dilution
Under such conditions,
the
expected to be small,
coal
on
the
observed
although not negligible,
the
and
quenching
of
impact of secondary
products
evolution
is
and secondary reaction
effects outside the coal are expected to be unimportant.
In applications employing coals other than those studied here, use
of the model parameters obtained from the experimental data specific to
196
60
I
197
a
59 58 57 0,
E
U
56 55 03
0
wj
E3
54 0
52
51
-
70
74
78
82
86
94
90
9.5
b
3
9.0 8.5 8.0 -
0
0
E
7.5
0:
E3
7.0 -
03
6.5 6.0 5.5 5.0
70
1
01
7
74
7I
78
I
I
82
I
86
91
90
94
ELEMENTAL CARBON CONTENT (WT% DMMF)
Figure 6.1-10
Best-fitted values of (a) E0 and (b) a for predicting
atmospheric pressure total volatiles evolution using the MIPR model
versus carbon contents of the coal. ko was fixed at 1014 s-1 in all
cases, and Y* was obtained from experimental data for each coal.
Carbon: LW < ZP < SR < BL < IL < LK. Abbreviations: see Fig.6.1-1.
198
Table 6.1-3 Best-fitted values of E and a of the MIPR model for total
volatiles production (ko fixed at 10 4 s-1 for all coals).
Coala
Y*
wt% dmmf
E0
kcal/mole
a
kcal/mole
Standard error
of estimate
wt% dmmf coal
Lower Wilcox L
51
55.2
8.0
3.9
Beulah Zap L
45
53.7
9.3
2.9
Smith Roland SB
49
53.2
7.3
3.5
Blue HVB
47
54.3
5.3
4.2
Illinois HVB
51
56.1
7.4
3.9
Lower Kittanning
LVB
22
59.2
6.5
2.3
a Coals are listed in the order of increasing elemental carbon contents
in dmmf basis. Elemental analysis is given in Table 4.1-1.
the coal of interest would give the most reliable performance.
If such
experimental information is not available, use of the parameter values
estimated from the coal-type dependent trends obtained in this study is
expected to give the next best performance - Figs. 6.1-2 for tars, 6.18 for
individual gases, and 6.1-10 for
caution
in
the
using
estimated
total volatiles.
values
is
that
established from a fairly small number of coals (6),
a possibility
that
some
from those studied here.
the number
'unusual'
coals
may behave
the
A note of
trends
were
and thus there is
very
differently
In applications where one wishes to minimize
of input parameters,
use of an averaged value of the rate
parameters is expected to give an adequate approximation over a narrow
But, such an approximation is generally
range of coal types for gases.
not recommended for describing tar or total volatiles production.
6.2. Extended MIPR model
The extended MIPR model increases the range of applicability of the
MIPR model by explicitly including descriptions of mass transport and
secondary reactions.
6.2.1 Mathematical formulation
This
section
presents a
chemical
and physical mechanism of tar
formation, and derives a quantitative model based on the mechanism.
Figure 6.2-1 gives a schematic diagram of the proposed mechanism, where
the solid arrows indicate reaction pathways leading to tar production
and the dashed arrows indicate
gas
formation.
The
chemistry
competing pathways
of
the
model
leading to char and
assumes
a hypothetical
199
molecular structure of coal deduced from the literature survey
200
presented in Section 3.1.1 (Fig.3.1-1)
X
-
(B-PAC-B)n Y
where
PAC
represents
repeating
hydroaromatic clusters,
nuclear
B bridging molecules,
to be responsible
for cross-linking,
and n the
of repeating units.
number
units
subunits are currently not known,
of
polyaromatic
X side groups
Y non-cross-linking
bridging molecules
are
suspected
side groups,
The exact structures of these
but some qualitative information can
be inferred from the literature survey in Section 3.1.1.
suggests
and
of polymethylene
The survey
and polymethylene-
ether type molecules; and side groups are molecules such as -OH, -COH,
-CH3 , and -C2 H 5 , among which the oxygenated side groups are postulated
to
cross-link.
these
The survey also
structures
is
these structures,
indicates
a strong function
that the concentration of
of coal type,
but for most of
reliable quantitative correlations are currently not
available.
The model formulation is based on the mechanism shown in Fig.6.2-1,
where the tar is
produced via the sequential
hydrogenation and transport.
are cross-linking,
steps of bridge scission,
Competing with the tar production pathway
polymerization,
and tar cracking reactions,
which lead the formation of char + gas.
all of
In the literature, the latter
two reactions are often lumped together and globally referred to as the
secondary
reaction
distinguishing
the
of
tar
(or
metaplast).
different competing reactions
The
importance
of
is described below,
where each or a combination of these three competing reactions is shown
201
2
2
NON-X-LINKED
COAL
COAL
4
4
PRIMARY TAR
SECONDARY TAR
5
3
\V
1V
CHAR
CHAR
CHAR
GAS
GAS
GAS
2 = SCISSION, HYDROGENATION
3 = POLYMERIZATION
1 = X-LINKING
Figure 6.2-1
4 = TRANSPORT
5 = CRACKING
Chemical and physical mechanism of tar formation.
4
2
0
C
X
01
-4
-6
1
suuoer
soft.
-
-10
-4
Maiorella
1
-2
0
2
toglo (P/Otm)
Comparison of the relative time scales for external and
Figure 6.2-2
internal transport rates of tar.
to uniquely describe and explain the experimentally observed effects of
operating
main
variables
-
coal
heating
type,
rate,
pressure,
and
particle size.
Cross-linking
The cross-linking
side groups
of two repeating units or nuclei
held together by a 'strong'
that is
a reaction between
in this work is defined as
to yield a coupled unit
We assume that the
bridge bond.
cross-linked molecule is too heavy to be volatile, and thus forms solid
Experimental evidence for this process
residue
(char).
solvent
swelling
where
changes
cross-linking.
studies of pyrolyzed coals
the
in
is
ability
swelling
(Suuberg et al.,
related
to
that have higher
-OH
contents
1987),
degree
the
The observation that the cross-linking is
for lower rank coals
comes from
of
more severe
suggest,
perhaps,
dehydration between two -OH side groups plays an important role
PAC-OH + OH-PAC ----- + PAC-O-PAC + H20
The coupled molecule from the dehydration
reaction is
very stable due to the high activation energy
expected to be
(~ 70 kcal/mole) for the
scission of the ether bond between the two phenyl groups.
lack of direct
evidence
However, a
supporting the dehydration mechanism suggests
that the actual cross-linking reaction pathway is much more complicated
that
than
shown
above
(Stein,
1988).
Carbon
dioxide
is
another
potential by-product of cross-linking, but the fact that it can also be
produced
from
unimolecular
decarboxylation
reactions
this
makes
argument less appealing.
Define
the
fraction
which
survives
cross-linking
as
V*
,
and
assume that this quantity depends only on the chemical structure of the
202
coal.
Disregarding
at
occurs
cross-linking
is
justified since
method to
A
extent.
Under
the
of
conditions
estimate
in form of an
presented in Section 5.1.2
correlation.
other
before
temperatures
mild
relatively
for a given coal is
empirical
of this process
any appreciable
proceed to
reactions
V*max
the kinetics
negligible
polymerization and cracking reactions (steps 3 and 5 in Fig.6.2-1), all
Experimentally, such a maximum yield is
will evolve as tars.
of V*
obtained by applying high heating rates (> 100 C/s) to small particles
(.
under low pressures
100 pm dia.)
to
postulated
polymerization
minimize
1985),
Kerstein,
10-3 atm).
(P
type
Rapid heating is
reactions
and
(Niksa
and low pressures and/or small particle sizes reduce
thereby leaving little
(step 4),
transport limitations
or no time for
cracking reactions.
Scission/polymerization
further heating, the
Upon
macromolecular matrix.
of the
coal
One route is the scission
(V*max) reacts via two competing pathways.
of bridge bonds which release
fraction
non-cross-linked
the PAC units in
Various bridge bonds,
radical
form from the
each with its
specific
chemical bond strength, are assumed to be present in the coal molecule.
To account for this,
the thermal decomposition is
first-order with respect to the
multiple independent parallel reaction,
amount
of
reacting
expressed as the
scission is
reactions,
material
(scission)
ksi
k,
1i
=
exp (-E
k 51
1
remaining
in
the
coal.
sum of the contributions
The
rate
from all
of
the
described by,
each of which is
dV /dt
assumed to occur by
(V*
/RT)
Vi)
(6.2-1)
(6.2-2)
203
where
i
denotes
one
reaction
reacted material by scission.
i.e.,
for all the reactions,
are
described by a Gaussian
deviation a,.
f(E)
=
where f(E)
and Vi
the
cumulative
amount
of
the
The same preexponential factor is used
k 0 ,,1
=
and the activation energies
kes,
distribution with mean E..
and standard
Thus,
[a,(2)/
2
]-1 exp[-(E-E0 , )2 /2a, 2 ]
= V*j/V*max for a large i,
(6.2-3)
and V*
equal to the sum of
is
the V*i for all i.
is
Competing with the scission pathway
the free-radical
initiated
polymerization type reaction assumed to occur first-order with respect
to the amount of unreacted coal
dV1 /dt (polymerization) = kP (V* -Vi)
The
polymerization pathway was
(6.2-4)
postulated based
on
the
experimental
Their
data of Serio (1984), Kobayashi et al. (1977), and Niksa (1981).
data indicate that a fraction of tar (Serio) and coal (Kobayashi et al.
and Niksa) reacts to form char at relatively low temperatures, and the
activation energies
range
for
derived
from the
data
gas phase polymerization reactions
(Serio,
1984)
are of
the
of coal-related aromatic
compounds (Gavalas, 1984).
Adding the two competing reactions, Eqs. (6.2-1) and (6.2-4), gives
the total rate of disappearance for species i
dVi/dt = (k,,1 + k )(Vi*-V )
= kti(V*
1
where kti
= k,
(6.2-5)
Vi)
i + kg .
(6.2-6)
To convert quantities from weight fractions to molar basis, divide
the quantity by the average molecular
weight (MWa,,)
of nuclei.
All
quantities in molar basis will be enclosed by a pair of square brackets
204
and will have the units of g-moles/g of raw coal.
Hydrogenation of the PAC radicals
route to stabilize the radicals.
as primary
tar,
and it
is
(PAC
) is
205
assumed to be the main
We define these stabilized molecules
denoted YH.
The rate of hydrogenation
for
species i is
dYHj/dt = MWavg kh [AH][PAC
]i
(6.2-7)
where [AH] represents the molar concentration of abstractable hydrogen,
and kh
the bimolecular
steady-state
assumption
rate
constant
to
[PAC ]
for hydrogenation.
requires
that
hydrogenation be equal to the rate of scission.
Applying
a
rate
of
the
Equation (6.2-7) can
thus be rewritten in the form of
(V*i-V )
dYH/dt = k,,,
Repolymerization
of
PAC
(6.2-8)
is
another
possible
depletion
pathway,
but
including this reaction would substantially complicate the formulation
and demand input parameters that are difficult to obtain,
To get around this complication,
MWavg
that
[AH],
we have arbitrarily assumed
the
rate
of scission given
in Eq.(6.2-1)
fraction
that
hydrogenates,
the
and
e.g.,
fraction
represents
that
only the
repolymerizes
is
accounted for by the polymerization step.
The
influence
of polymerization
on
tar
production
effectively described by expressing the rate of YHi
of
the
total
rate
of
First, define a quantity,
E,,i
=
disappearance
EP,,
of
the
can be
formation in
non-x-linked
more
terms
fraction.
, as
rate of scission
rate of scission + polymerization
=
k,
/ (k,,i+kP)
Then, combining Eq.s (6.2-5),
dYHj/dt
=
EP,1 kti(V*iVi)
(6.2-9)
(6.2-8), and (6.2-9) gives
(6.2-10)
Summing the above equation over all species
i
gives the total rate of
primary tar production
dYH/dt = X EPi kti(V*i-Vi)
(6.2-11)
Under minimal transport limiting conditions (i.e., low pressures, small
particles),
Eq.(6.2-ll)
heating rates.
describes the rate of tar production for all
Mass transport effects are included next.
Transport/cracking
Upon
hydrogenation,
transported
the
away from the
primary
tar
coal particle
cracked to produce char plus
gases.
is
either
to yield
physically
secondary
To rigorously model
tar,
or
the former
step would require a description of transient transport processes in a
porous solid or liquid phase environment inside the particle,
with transport in
the mass boundary layer outside
approach requires a large number of physical
and
tar, many of which
transient pore-size
vapor
pressures
to
transport
processes
softening
and
computational
deal
difficult to measure
tars.
with
for
in
Such
this
study
coal
often
required
In
in
the
potentially restrict the use of such models
such
as
we
types,
coals.
or estimate, e.g.,
would
where
in
be
need
which
rigorous
coals,
especially
to
describe
include
both
the
large
addition,
can
approach
practical
large combustion or gasification models
This
of both coal
for non-softening
limitations
different
non-softening
effort
the particle.
properties
distribution information
of
difficult
are
coupled
applications
that describe
mechanics, heat and mass transport, and reaction kinetics.
fluid
Therefore,
we felt a need to develop an approximate transport description that is
able to capture the observed effects of pressure and particle size on
206
as
using
production,
tar
few
parameters
input
as
possible,
and
requiring minimal computational effort.
first
The
in formulating an approximate mass
simplification made
that,
is
transport description
softening
for both non-softening and
coals, the transport resistance outside the coal particle is assumed to
be
negligibly
compared
small
characteristic times
that
to
inside.
of
Comparing
the
the two domains shows that this assumption is
in
valid over the range of operating conditions considered in this study.
6.2-2 compares the relative time
Figure
scales
over a wide range of
pressures for both non-softening and softening coals; the equations and
the values of physical properties
given in
Table 6.2-1.
used to compute the time scales are
A similar simplification has been made in more
rigorous models (Russel et al., 1979; Gavalas and Wilks, 1980; Bleik et
al.,
1985).
While escaping, some of the tar reacts to produce char plus gaseous
Serio
products.
cracking
in
(1984) has
shown
that the
rate of homogeneous
tar
the vapor phase can be modeled as a first-order reaction
with respect to the vapor phase tar concentration; a possible catalytic
presently poorly understood,
effect of coal surface is
of tar
cracking
=
(6.2-12)
kcYHavg
where kc is the cracking reaction rate constant, and YHavg
mass
not
The rate of cracking reaction is then expressed as
included.
rate
and thus is
fraction of primary
tar
inside
the
coal.
The
the average
total
rate
of
secondary tar, denoted as Y, leaving the particle surface is given as
dY/dt = I E
kti(V* ji)
where YHavg represents
the
coal.
- kcYHavg
(6.2-13)
the average concentration of primary tar inside
Derivations below
relate YHavg
to
the relative rates
of
207
Equations and physical properties used
Table 6.2-1
relative transport time scales shown in Fig.6.2-2.
Coals
time scalesa
Non-softeningb:
te/t
= e/r
Softening:
te/ti
=
(Re f 2 DLP)/
(Rb DaPvap)
to
compute the
physical properties
e
0.1
r ~2
Reff
Rb
DL
D9
P
Eq.(6.2-26) with R
of 50 pm
= boundary layer
thicknessa
= Eq.(6.2-29) with DoL
of 10-6
=
=
Eq.(6.2-28)
= reactor pressure
Pvap = tar vapor pressurec
a Particle radius is assumed for the external boundary thickness. This
approximation is comparable to the value computed from a rigorous
flow description (Zacharias, 1979).
b e = void fraction, r = tortuosity.
o Vapor pressure correlations of Maiorella (1978), and Unger and
Suuberg (1983) were used.
208
primary tar transport and cracking.
209
Consider the mass transport in non-softening coals first.
the
objective
of
developing
transport description,
a mathematically
concise
but
To meet
effective
the following assumptions were made:
(1) All of the primary tar enters macropores without encountering
any
appreciable
configurational
or
Knudsen
type
diffusion
resistances in the smaller pores.
(2) All of the primary tar enters macropores from the center of
the particle.
(3) Neglect convective contributions.
(4) The concentration of the primary tar in the coal is at steadystate.
Supporting evidence for the first assumption comes from the analysis of
Gavalas and Wilks (1980).
are
The study indicates that most of volatiles
generated within the small pores
(micro-
and transition)
and are
transported to the outside via the large pores rather than directly.
Thus although the diffusivity is extremely low in the small pores, the
transport
resistance
distance may
be
sufficiently
in these pores negligible.
short
The
to make
the
transport
second assumption is not
strictly valid since tars enter from all points along the pore, not
just from one end.
But making this assumption considerably simplifies
the mathematics without seriously hindering our ability to capture the
effects of main operating variables.
assumption is
inside
the
compensate
An important consequence of this
that the apparent average residence time of primary tar
coal would
for
this
diffusivity values.
be
somewhat
effect
would
one
over-estimated,
but
be
slightly
to
assign
way
to
higher
Evidence for the third assumption is based on the
macropores
that
observation
can
be
pore
a bimodal
by
approximated
system with radii 0.05 and 0.5 pm, and that the larger pores are few in
The transport
number and hence poorly inter-connected (Gavalas, 1984).
in
larger pores
the
contributions
convective
greater
to
due
pores
radius 2 ,
pore
[tdiff/tc
,c
approximate
the macropore
system
Suuberg
as
in
the
(1985)].
having
the
two
the smaller
in
much faster than the transport
is
larger pores
can
Thus,
one
types
of pores
connected in series, where the resistance offered by the larger pores
Implicit in this approximation is that the effective
is negligible.
diffusion path is now shorter than the particle radius, but the current
allow this
description will
is
much
to be
absorbed
in the
effective
The last assumption is valid if the particle radius
diffusivity term.
(R)
effect
less
than the
square
root of
the
effective
gas
phase
diffusivity of tar inside the coal (D,,eff) times a characteristic time
scale for pyrolysis (tpyro)
R << (Dgeff
D,,eff is
Dgeff
tpyro)
1
/2
(6.2-14)
related to the gas phase binary diffusivity of tar by
=
(6.2-15)
e/r D
where e and r are the void fraction and tortuosity of pores in the coal
respectively.
steady-state
in
For Dg = 0.1 cm 2 /s, e = 0.1,
r = 2,
and tPYro = 1 s,
assumption holds for R << 700 pm, and thus is
this study where the maximum particle radius is
< 50 pm.
the
applicable
With the
above approximations, a material balance of YH across a thin cylindrical section of a macropore gives (< > indicates mass fraction/unit pore
volume)
d 2 <YH>/dx 2 - (kc/Dgeff)<YH> = 0
(6.2-16)
where x is the coordinate along eh pore axis, and a general solution is
210
<YH> = A exp(m.,x) + A 2 exp(m~yrx)
m..
(kc
/Dgeff )1I/2 ,
to
equal
is
(6.2-17)
and Al , A2
constants
are integration
evaluated at the boundary conditions
d<YH>/dx
(6.2-18)
at x = R
<YH> = 0,
=
E
-
(m,,
(V*,-Vi)/(NAR2)
,iR)2 ,
at x = 0
(6.2-19)
where
are
APNP
the
and
respectively,
macropores
area
cross-sectional
m,,
=
and
the
number
total
(kti/Dgeff)1 /2.
boundary condition has little physical meaning;
The
of
latter
the form was chosen to
be consistent with the assumption (2) that the primary tar enters from
the
center
of
the
particle.
Solving
for
the
integration
constants
gives
<YH>
=
E '(V*j-Vj)(_ns,,R)2
[exp(-mx)-exp(-2mR+mnx)]
[l+exp(-2m.,R)]
N A, R
(6.2-20)
<YH>
Integrating
over
the
radius
R
and
converting
to
mass
fraction
basis gives
YHav
2 Epi(ki/kc)(V*i-Vi)rl+expp(-2msR)-2exp( -mR)1
=
[l+exp(-2mnsR)]
(6.2-21)
Substituting
Eq.(6.2-21)
into
Eq.(6.2-13)
gives
the
rate
of
tar
(secondary) leaving the particle surface as
dY/dt = X EPi Ec,,,
kti(V* -Vi)
(6.2-22)
where
Ec,,,
production with transport limitation
= rate of tar
rate without transport limitation
=
2exp(-mnR)/[l+exp(-2mnR)]
(6.2-23)
211
The quantity mn.R is
the Thiele modulus for non-softening coals.
three multiplication
factors in
Ecn,
represent
an
respectively account
Eq.(6.2-22),
important
for the
result
effects
(X V*), E ,
V*max
of
this
model,
of cross-linking,
The
and
,,
as
they
polymerization
and cracking on tar production.
Their values are each bounded between
0
the
and
1,
where
0
represents
most
severe
limitation
on
tar
production and 1 represents no limitation.
In softening coals, the liquid phase diffusion is assumed to be the
dominant
mode
of
transport
inside
the
The
coal.
possibility
of
volatile bubbles enhancing the rate of tar escape is neglected based on
the recent
experimental
evidence
that at least
in
some cases most of
the bubbling phenomena occur before any appreciable amount of tars are
produced
(Hsu, 1988;
Griffin, 1988).
Applying a similar derivation
procedure as above for non-softening coals gives the rate of secondary
tar escaping the particle surface as
dY/dt
Ei
=
(V* -Vi)
E,, kt
(6.2-24)
where
E ,=
2
exp(-mSRff)/[l+exp(- 2 msRff)]
(6.2-25)
The quantity msReff is the Thiele modulus for softening coals, where
m, equal to (kc/DL)1
tar in
scale.
2
; DL represents
the molten coal;
Often,
(Sung, 1978),
and Reff
the liquid phase diffusivity of
is the
effective diffusion length
the molten coal has a shape of a cenospherical
prior to
extensive
tar release
(Hsu, 1988).
shell
Thus an
appropriate diffusion length scale for softening coals is assumed to be
half of
the
shell thickness.
with a particle
radius of
For
a Pittsburgh Seam bituminous
~40 pm pyrolyzed under atmospheric
coal
reactor
212
pressure, Griffin (1988) reports the shell thickness to be about 20% of
the particle radius of the raw coal.
for
the
two
softening
coals
This approximate value is assumed
in this
study
(Illinois HVB
and Lower
Kittanning LVB).
An exact explanation
for
the observed
pressure and particle-size
effect on tar production for softening coals,
is
currently unclear due
to large uncertainties in physical parameter values of the molten coal
(see
3.2.2
Sections
and
3.4.3).
In
this
formulation,
excluding
leaves the possibility that the
external and bubble transport effects
shell thickness is a function of pressure and particle size as the only
viable
explanation
the observed
to describe
conclusive explanation becomes available,
that the
shell thickness
is
behavior.
Until a more
the present model will assume
related to the pressure and particle size
in the form of
Reff
=
0.1 R x 10~'
the particle
where R is
atm.
(P/1)
1
/3(R/40)
radius
experimental
quantitative
data
/ 3
cm
(6.2-26)
and P the reactor pressure
pm,
(1988)
work of Griffin
The
in
1
is
currently
to examine
seeking
the effect
in
to provide
of pressure
and
particle size on the shell thickness (6.2-26).
Model parameters for the extended MIPR model
For a given coal, the model requires a total of 9 input parameters:
V*max
ko,, E,, o-,, k0 p, E,, k0 c, EC,
the kinetic parameters
discussions below,
are
used
available.
phase
in
and Dgeff
cases
where
no
reliable
or
DL .
In
estimated by Gavalas
experimental
the
(1984)
measurements
are
But, these estimated parameters are strictly valid for gas-
reactions
where
the
interference
from
neighboring
molecules
213
(e.g.,
thus
"cage" and solvation effects; Stein, 1981) can be neglected, and
one may
question
reactions occurring in
Justifications
for
the
validity of
applying
these
a condensed phase environment such as in
taking
this
interaction effects decrease
approach
at higher
are
C);
(2)
coal.
solvent-molecule
temperatures
(Stein, 1981) so
temperatures
experimentally measured homolysis rate
gas and liquid phase are approximately
to
(1)
that they may be small at typical coal conversion
1000
estimates
the same,
constants
(400in
the
within a factor of 2
(Stein, 1981); and (3) no alternative methods are currently available.
The
non-cross-linked
experimental
tar
yield
pressures (vacuum).
fraction
limit
of
obtained
the
coal,
is
the
under
low
V*max,
with rapid heating
If experimental values are not available, the low-
pressure tar correlation developed in
Section 5.1.2 can be used;
from
just the elemental composition information of the coal, the correlation
predicts V*
In
coals;
with a standard error of estimate of ±3 wt% dmmf.
the
scission
this
value
reaction,
is
within
log(kO,/s'1)
the
was
fixed
at
14
for
all
range estimated by Gavalas et
al.
(1981a) for homolysis of ethylene (13.9) and methylene (14.3) bridges.
Estimating a priori the two activation energy related parameters,
and as,
is
more difficult because (1)
the activation energy is highly
Gavalas et al.
for
ethylene
(1981a,b)
unlike the preexponential
sensitive
to
the bridge
type,
E0 .
factor
e.g.,
estimates the dissociation activation energy
and methylene
bridges between
two unsubstituted
phenyl
groups to be 56.4 and 80.7 kcal/mole respectively; and (2) quantitative
information on the type of bridges
in
the coal is
currently lacking.
Therefore, E0 , and a, were best-fitted using the experimental tar data.
The best-fitted values of E05
and a, respectively
range from 51.7 to
214
These
56.8 and from 3.5 to 9.4 kcal/mole among the six coals studied.
for
expected values
range of
within the
are
values
scission of
the
bibenzyl type bridges (Ph-CH2 -CH2 -Ph ) (Gavalas, 1984).
for
energy
activation
The
the
was
value
This
types.
coal
all
for
was
and was assumed to the
estimated to be approximately 35.5 kcal/mole
same
EP,
reaction,
polymerization
obtained from using
the
relationship (Rempp and Merrill, 1986)
(6.2-27)
EP = E (propagation) + 1/2 E (initiation)
-
1/2 E (termination)
steps
assumed
addition to an unsaturated side
to
to occur by a
group of a phenyl molecule,
by
is very close
the vapor phase
(1984) for
reported by Serio
the 35.3 kcal/mole
termination
and
The estimated E
recombination of two benzyl radicals.
respectively
kcal/mole
scission,
bridge
ethylene
by
initiation
0
The propagation reaction was assumed
(Gavalas, 1984).
radical
and
56.4
7.3,
about
be
to
and termination
initiation,
energies for propagation,
with activation
secondary reaction of the most reactive fraction of tar (preexponential
1.43x108
=
factor
s-1,
modeled
as
a
single
first-order
Applying the same procedure gave k0 p of about 1012
value
greatly over-predicted
s- 1,
reaction).
but using this
the extent of polymerization.
When ko
was allowed to be best-fitted from the data, values ranging from 106.8
s-1 were obtained for the six coals studied.
to 107.3
Such a variation
1
is small, and thus kop was fixed at 107 s- for all six coals.
The
experimental
tar
intra-particle
Using
Serio's
reactions
about
information
cracking
(1984) data
reactions
on homogenous
from Pittsburgh Seam bituminous
1014
s-1
and
69
kcal/mole
and E,
to estimate ko,
required
is
currently
not
for
available.
extra-particle
tar cracking
gave ko
0 ,
and E. to be
coal,
respectively
[first-order
single-
215
reaction model was assumed, see Howard (1981) for description].
Among
the three different reactive tar fractions distinguished by Serio, the
reactive
moderately
used
in
were
reaction,
of the polymerization
fraction
data
to
used
compute
the
kinetic
The reaction rate for the most reactive fraction is in the
parameters.
range
fraction
and that
for the
is virtually negligible at temperatures below
the
extended
MIPR
under-predicted
substantially
the
model,
the
extent
estimated
of
least
1000
rate
secondary
reactive
C.
When
parameters
tar
cracking
reactions, implying that the rate of cracking reactions inside the coal
is
possibly much faster
that
lowering
than that
outside.
to 55 kcal/mole gave
E0 0
Preliminary tests
good predictions
showed
on the extent
of tar cracking reactions, and thus this value was assumed for all six
coals.
The
diffusivity
(Suuberg et al.,
inside
1979;
the
non-softening
coal
is
assumed
e
is
the
be
Froment and Bischoff, 1979)
Dgeff = e/r 0.1 (T/273)'. 5 (1/P) cm 2 /s
where
to
internal void
fraction
and
(6.2-28)
r the
tortuosity
factor.
Since no measured values of e and r are readily available for different
coals,
the
ratio
data.
The diffusivity
e/r
is best-fitted
from the
experimental
inside the softening coal
is
tar
yield
assumed to be (Oh,
1985; Reid et al.,1977)
DL
=
DoL
(T/298) cm2 /s
where DoL is of order 10-5
and
inversely
molten coal.
related to
(6.2-29)
to 10~7
a
(Oh, 1985;
fractional
power
Suuberg and Sezen, 1985),
of
the viscosity of the
The exact value of DoL for a given coal was obtained from
applying best-fit regression routines with the experimental tar data.
216
6.2. Results and discussions
Figure
6.2-3
compares
217
the
experimental
from the extended MIPR model for the
study.
six coals investigated in this
Table 6.2-2 gives the values of model parameters used to make
The four parameters
the predictions.
are
and predicted tar yields
V*max,
EOs,
a,, and
e/r
or
assumed to vary with coal type
The V*max
DoL.
experimentally measured tar yield limit
the remaining
at vacuum
inputted
with
(~C 0.001 atm),
and
three parameters were best-fitted from the experimental
The parameters,
tar data.
was
k 0 ,,
k,,O, E,
, and EC were assumed not
k0
to vary significantly among different coal types.
This assumption was
mainly made because information necessary to assign coal-type dependent
values for these parameters is
currently not available,
but we do not
imply that these parameters are truly constant for all coal types.
Any
errors generated from this assumption will affect the values of bestfitted parameters.
If the error cannot be sufficiently compensated by
the fitted parameters, then the error will be reflected in the model's
predictive capability.
Figure
6.2-3
shows
a good agreement between the experimental
predicted tar yields at all three pressures
important
observation
to be
made
from
- 0.001,
the
figure
1 and 10 atm.
is
the
is
especially
encouraging
since,
unlike
An
accurate
prediction of the yield limits over a wide range of pressures.
result
and
This
the MIPR model, the
maximum yields are predicted without having to rely on experimentally
measured values at different pressures,
does not need a pressure-specific Y*.
which the rate is
550
C),
and
that
i.e.,
the extended MIPR model
Also, the predicted behavior in
unaffected by pressure at fairly low temperatures
the yields
'level-off' earlier
(i.e.,
at
(~C
lower
218
20-
a
18
P,atm
0.001
16
14 -
2-
o12
10
w
5:
KI
10
8
6
4
2
0-B
300
12
500
700
900
1 100
-
b
11
10
P,atm
0.001
8L
-
7-
0
6
L
5
H
4
25
010
01
LI
2
0--T
300
500
700
TEMPERATURE
Tar
Figure 6.2-3
experimental data:
represent extended
represent averaged
SR, (d) BL, (e) IL,
900
1 100
(C)
Symbols represent
yields versus peak temperatures.
Lines
N - 0.001 atm, E - 1 atm, E - 10 atm.
MIPR model predictions. 0.001 and 10 atm points
values from 1-3 runs.
Coals: (a) LW, (b) ZP, (c)
(f) LK. Abbreviations: see Fig.6.1-1.
219
20
C
18-
L
16
P, atm
14
0.001
1
12
12010
10
-
6
-
D6
4 00
2
35 -
1 100
900
700
500
300
d
30 -
P, atm
0.001
25 L
01
S
0
20 -
15 -
g1
100
0
20-
0
a:o
5-
0300
500
700
900
1,100
TEMPERATURE (C)
Tar
Figure 6.2-3
experimental data:
represent extended
represent averaged
SR, (d) BL, (e) IL,
Symbols represent
yields versus peak temperatures.
Lines
10 atm.
0
atm,
1
El
atm,
0.001
m points
atm
10
and
0.001
MIPR model predictions.
LW, (b) ZP, (c)
Coals: (a)
values from 1-3 runs.
Fig.6.1-1.
see
Abbreviations:
(f) LK.
40-
220
e
35 -
P, atm
0.001
30 25-0
000
20 -
10
w
15 100
10 -0
03
0
5
0300
18 -
500
700
900
1 100
f
16
P,atm
0.001
14
12
0r
1
1
10
8-
10
Wi
6
4
-
2
0
0
300
500
700
900
1 100
TEMPERATURE (C)
Symbols represent
Tar yields versus peak temperatures.
Lines
N - 0.001 atm, C - 1 atm, 0 - 10 atm.
points
10
atm
and
0.001
predictions.
model
MIPR
extended
represent
Coals: (a) LW, (b) ZP, (c)
represent averaged values from 1-3 runs.
SR, (d) BL, (e) IL, (f) LK. Abbreviations: see Fig.6.1-1.
Figure 6.2-3
experimental
data:
Table 6.2-2
Model parameters for the extended MIPR model.
221
(a) Coal-type dependent parameters:
Coala
V*max
O
E,
e/r
or
DoL
kcal/mole
kcal/mole
16.8
53.8
7.0
10-2.81
-
9.1
52.8
9.4
10-3.23
-
Smith Roland SB
14.8
51.7
6.3
10-2.70
-
Blue HVB
27.7
54.6
5.3
10-2.90
-
Illinois HVB
30.1
54.4
4.4
-
10-5-67
Lower Kittanning
LVB
14.0
56.8
3.5
-
10-5.41
wt% dmmf
Lower Wilcox L
Beulah Zap L
(b) Fixed parameters:
scission
k0 s, s-1
1014
polymerization
key,
107
EP,
kcal/mole
35.5
cracking
k00 , s-1
1014
E,, kcal/mole
55.0
s~1
a Coals are listed in the order of increasing elemental carbon contents
in dmmf basis. Elemental analysis is given in Table 4.1-1.
temperatures)
pressure
experimentally
is
increased,
observed behavior
for
tars
closely
[Suuberg
resembles
(1977),
the
Fig.3.3-5]
and for total volatiles [Suuberg (1977), Fig.3.3-4a; Niksa (1981)].
Figure
plots (a)
6.2-4
Ecn,
Epavg,
Ep avg= X EPi
both
helps
explain how
and (b)
f(E)AE.
non-softening
represented
to
Epavg,
Recall
the
E,
that the
[Eq.(6.2-22)]
and
as the product of the
model
The
figure
versus temperature,
where
rate of
softening
total
works.
tar production for
coals
[Eq.6.2-24)]
rate at which
is
the non-x-linked
fraction reacts and the two 'E' factors, and that the values of these E
factors
range
(6.2-9),
(6.2-23),
production
ES
,as
between
at
0 and
(6.2-25)
pressures
the pressure
increases.
10
atm),
the values
particle
1
is
At vacuum,
by
Ecn,
given
in
decrease
the
smaller
and EC,,
resistance;
lower,
are
The
explained
transport
are much
important
feature
the experimentally
atm.
Recall
from
increasing volatiles
at
EC,,
Eqs.
in
tar
E ,,s
or
are near 1
at high pressures
indicating a
substantial
(>
intra-
transport resistance.
Another
explain
mass
Ec,ns,
respectively].
higher
indicating negligible
[E ,1,
1
vacuum,
of
this
model
observed heating-rate
Section
3.3.4
that
Anthony
(1974)
temperatures
transport
effect,
The
as
effects,
indicated
non-x-linked
the
it
is
effects at
Niksa
and
negligible heating- rate effects at 1 atm.
temperatures.
that
(1981)
able
to
vacuum and
observed
an
(implying tar) production at higher heating rates
whereas
polymerization
is
heating
rate
higher heating
by
(1977)
observed
Figure 6.2-4 shows that the
Epavg,
fraction
is
Suuberg
is
of
more
coal
increased.
rates enhance
tar
severe
reacts
Thus
at
at
lower
higher
without
production.
mass
At 1
atm, where mass-transport effects are not negligible, tars produced at
222
1.0
223
0.9 -
0.8 0.7 0.6 -
0
0.5E
avg
-Ec,na
-p,
0.5
p =1oatm
0.4 1 atm
0.3 -
0.2 0.1
-
0.0
--
1.0 -
900
700
500
300
100
b
0.9 0.8 0.7 0.001
atm
0.6 -Epavg
0
0.5 P=io at
IL
*
0.4
1atm
0.3
0.2
0.1 -
900
700
500
300
100
TEMPEfRATURE (C)
Figure 6.2-4
softening coal
(a) Ep avg and Ec,ns
(LW). (b) Epavg and
softening
(IL).
coal
Ep,avg
Ecns
and
Ec's
from
Eqs.
Abbreviations: see Fig.6.1-1.
-
versus temperature for a nonEc s versus temperature for a
Y E
1 fE)AE,
(6.2-23)
and
Ep i
from
(6.2-25)
Eq.(6.2-9),
respectively.
higher
temperatures
cracking
experience
reactions.
Thus,
a
the
greater
extent
increased
tar
of
secondary
tar
production at higher
temperatures is 'off-set' by more cracking reactions.
Figure 6.2-5 plots the best-fitted values of E0 . and a.
elemental
carbon contents
of the coal.
As before
versus the
for the MIPR model
(Fig.6.1-2), higher rank coals generally gave greater values of E0 . and
smaller values of a,,
coals
have
narrower
bond
implying that bridging molecules
dissociation
distribution.
energies
with
a
The noticeably larger
of higher rank
greater
mean
and
E0 . for LW compared
a
to
the other low-rank coals (ZP,SR), is more representative of higher rank
coals
(BL,IL);
abbreviations
are
defined
behavior has also been observed in
volatiles
using the MIPR model
in
Fig.6.1-1.
A
similar
describing the production of total
(Fig.6.1-10a).
The higher a, for ZP
among the low rank coals is similar to the trend observed in describing
tar production
Section 6.1.2,
using
the
MIPR model
(Fig.6.1-2b),
global E0
discussed
in
this may be a typical behavior for low-rank coals with
small elemental hydrogen contents (<5wt% dmmf).
values of EO,
As
For a given coal, the
are slightly but consistently greater than those of the
for the MIPR model
(tar).
mass transport effects implicit in
The difference
EO,
is
attributed to
and confirms that the transport
resistance at 1 atm is small but not negligible.
Figure 6.2-6 plots the best-fitted values of e/r or DoL versus the
elemental carbon contents of the coal.
For non-softening coals (LW,ZP,
SR,BL),
range
the
These values
best-fitted
values
of e/r
imply that the tortuosity
the void fraction (e) is around 0.1.
order
of
magnitude
higher
than
(r)
is
from 10-3.23
to
10-2.70.
about 100 assuming that
Such values for r are at least an
typical
values
reported
for porous
224
225
58
Ia
57 -
56 -
55 -
01
54
0
0
53 -
0
52 -
51 -
50
70
10
74
78
82
86
90
94
b
-
9
8
-
7
-
6
5
4
3
70
74
78
82
86
90
g4
ELEMENTAL CARBON CONTENT (WT% DMMF)
Figure 6.2-5
Best-fitted values Eos and as for predicting tar
evolution using the extended MIPR model versus the elemental carbon
content of the coal.
Carbon:
LW < ZP < SR < BL < IL < LK.
Other model parameters are given in
Abbreviations: see Fig.6.1-1.
Table 6.2.2.
226
10
-2
-
eDo
-j
0
100
10-5
0
DOL
70
74
78
82
86
90
94
ELEMENTAL CARBON CONTENT (WT% DMMF)
Figure 6.2-6 Best-fitted values of transport parameters for predicting
tar evolution using the extended MIPR model versus the elemental carbon
content of the coal. e/r is for non-softening coals (LW,ZP,SR,BL), and
DoL is for softening coals (IL,LK).
Carbon: LW < ZP < SR < BL < IL <
LK. Abbreviations: see Fig.6.1-1. Other model parameters are given in
Table 6.2.2.
solids (Froment and Bischoff, 1979).
best-fitted values of DoL
within the range
are between 10~5-41
of expected values
of e/r or DoL assume that k00
k00
will
parameters
koc/(e/r)
directly
since
influence
the
For softening coals (IL,LK), the
is
of 10-6±1.
about 1014
the
quantity that
values
and 10-5.67,
which
are
The best-fitted values
S-1.
of
is actually
Changing the value of
the
fitted
fitted is
transport
the
ratio
or koc/DoL.
In applications employing coals other than those studied here, use
of the model parameters obtained from the experimental data specific to
the coal of interest would give the most reliable performance.
If such
experimental information is not available, use of the parameter values
estimated from the coal-type dependent trends obtained in this study is
expected to give the next best performance
and Fig.6.2-6
for e/r or DoL.
- Fig.6.2-5 for E0 , and o- ,
As before for the MIPR model parameters,
a note of caution in using the estimated values is that the trends were
established from a fairly small number of coals (6),
a possibility that
some
from those studied here.
and thus there is
'unusual' coals may behave very differently
227
7. Conclusions and recommendations
The
objective
coal-type effects
this
228
of this study was
in
to
improve
rapid coal pyrolysis.
study examined the pyrolysis behavior
lignites to
low-volatile bituminous
transport resistances are small
dia.).
atm.
(1
Selected high-temperature
the
understanding
of
The experimental phase of
of six coals ranging
from
coals under conditions where mass
atm pressure
and < 100 pm particle
runs were also made at 0.001 and 10
The modeling phase of this work derived kinetic information from
the experimental
data using
the MIPR and extended MIPR models,
and
attempted to relate the kinetic information to measurable properties of
the coal.
7.1. Conclusions
Tar production
1.
Among the
six coals
studied,
higher rank coals generally produced
tars at higher temperatures, and over a narrower range of temperatures.
Consequently,
a
larger
mean
and
a
narrower
distribution
activation energies were obtained using the MIPR model
of
global
for coals
of
increasing rank.
2.
High-volatile bituminous
dmmf),
coals produced
the most
followed by lignites and a subbituminous
coal
tar
(7-13
(21-25 wt%
wt% dmmf),
and a low-volatile bituminous coal (11 wt% dmmf).
3.
A quantitative correlation, developed to independently relate
tar yield limits to coal type and pressure, was tested against
a large
set of experimental data representing a wide range of coals (37 coals,
ranging from
atm).
lignites
Good agreement
to anthracites) and pressures
('vacuum' to
between the predicted and experimental
90
yields
were obtained for all
coals and pressures, with a standard error of
estimate of ±3 wt% dmmf.
Gas production
4.
In general, no discernable coal-type effects on the apparent
rate of gas production were observed.
is
that
variations
comparable
to
in
those
measurements.
the
A probable explanation for this
rate caused by different
caused
Consequently,
by
uncertainties
kinetic
parameters
coal
in
of
types
are
experimental
the
MIPR
model
parameters for measured gas species were only slightly affected by coal
type.
5.
Higher
rank
coals
generally
pyrolytic water, but more methane.
produced
less
carbon
oxides
and
The ethylene and ethane yields are
small and their absolute yield values are less affected by coal type.
Total volatiles production
6.
Total volatiles
range
of
evolve at higher temperatures
temperatures
for
higher
rank
coals.
and over a narrower
Thus
as
expected, a
larger mean and a narrower distribution of global activation energies
were obtained using the MIPR model for coals of increasing rank.
trends
are
consistent
with
the expected
behavior
from
The
combining the
observed coal-type effects on the rate of tar and gas production.
7.
The total volatiles yield limit is fairly constant among the
lignites,
wt% dmmf),
coal
(22
and subbituminous and high-volatile bituminous coals
but is significantly less
wt%
dmmf).
The
for the low-volatile bituminous
high-volatile
significantly more reactive volatiles
(41-55
bituminous
than other coals
coals
produced
(38-45 versus
229
19-28
wt%
dmmf);
reactive volatiles
are
defined
as
total volatiles
minus water and carbon dioxide yields.
Pressure effects
8.
Increasing the pressure gave less tar and total volatiles for
the coals investigated in this study. In absolute values, the
pressure effect was more severe for coals that produced more
tar.
Extended MIPR model
9.
Predicted tar yields from the extended MIPR model agreed well
with experimental
low-volatile
values for
bituminous
a wide range of coal types
coal;
non-softening
and
(lignites to
softening)
and
pressures (0.001-10 atm).
10.
For a given coal,
parameters
(E0 ,
and
the model requires just three adjustable
a. for bridge scission, and either a geometrical
factor e/r for non-softening coals or DoL
quantity
V*max
that is
for softening coals),
fitted values of E0 , and
or
The best-
a. for bridge scission are within the range of
expected
values
bridges.
The best-fitted values of e/r imply a tortuosity (r)
the
scission
about an order of magnitude
porous solids,
plus a
either directly obtained from experiment
estimated from the tar correlation developed in this study.
for
input
greater
of
bibenzyl
(Ph-CH2 -CH2 -Ph)
than typical values
type
that is
reported for
and those of DoL are within the range of values reported
in the literature.
7.2. Recommendations
230
1.
More
fundamental
studies
quantitative
understanding
mechanism
coal
in
pyrolysis
of
are
the
(e.g.,
recommended
coal
to
improve
and
the
structure,
studies using model
the
reaction
compounds
and
chemically modified coals).
2.
this
Analyze the MW and extract data collected in conjunction with
study to determine
the effect
of coal
type on chemical changes of
tar and extracts during pyrolysis.
3.
Investigate the validity of the assumed relationship between
the shell thickness of softening coals,
and pressure and particle size
[Eq.(6.2-26)].
4.
Provide
experimentally
measured
intra-particle
reaction rates to the extended MIPR model.
tar
secondary
231
8. References
232
Anthony,
D.B.,
"Rapid
Pulverized Coal," Sc.D.
devolatilization
and
Hydrogasification
of
Thesis, Dept. Chem.
Eng., M.I.T.,
Cambridge,
MA, 1974.
Anthony, D.B., Howard, J.B., Hottel, H.C.,
and Meissner, H.P.,
"Rapid
Devolatilization
of Pulverized Coal,"
Fifteenth Symp. (Int.)
on
Combustion, The Combustion Institute, Pittsburgh, 1303, 1974.
Arendt, P.A. and van Heek, K.J.,
"Comparative Investigations of Coal
Pyrolysis Under Inert Gas and Hydrogen at
Low and High Heating Rates
and Pressures Up to 10 MPa," Fuel, 60, 779, 1981.
Bartle, K.D. and Jones, D.E.,
"NMR Spectroscopy" in Analytical Methods
for Coal and Coal Products, Vol.2, Clarence Karr Jr., ed., 1978.
Bautista,
J.R,
"Time-Resolved
Pyrolysis
Product
Distributions
of
Softening Coals," Ph.D. Thesis, Dept. Chem. Eng., Princeton University,
Princeton, NJ, 1984.
Benjamin, B.M.,
C.J.,
"Thermal
Structure,"
Raaen, V.F., Maupin,
Cleavage of Chemical
Fuel,
Benson, S.W.,
57,
269,
P.H., Brown, L.L.,
Bonds in
Selected
1978.
Thermochemical Kinetics, John Wiley, New York, 1976.
Bleik, A.,
van Poelje,
W.M.,
van Swaaij,
W.P.M.,
F.P.H.,
"Effects
of Intraparticle
Heat
and Mass
Devolatilization of a Single Coal Particle,"
AIChE J.,
Blom, 1960, taken from van Krevelen,
Co., Amsterdam, 1961.
Briunsma,
and Collins,
Coal-related
O.S.L.,
Geertsma,
R.S.,
D.S.,
R.S.,
Coal,
Bank,
P.,
and van Beckum,
Transfer
During
31, 1666, 1985.
Elsevier
Publishing
and Moulijn,
J.A.,
Fuel, 67, 327, 1988.
Burnham, A.K.,
Oh, M.S.,
and Crawford, R.W.,
"Activation Energy
Distributions and Related Chemistry for Pyrolysis of the Argonne
Premium Coals," submitted for publication in Energy and Fuels, 1988.
Calkins, W.H., Hagman, E., and Zaldes, H., "Coal Flash Pyrolysis 1. An
Indication of the Olefin Precursors
in Coal by CP/MAS 1 3 C NMR
Spectroscopy,"
Fuel,
63,
113,
1984a.
Calkins,
W.H.,
and
Tyler,
R.J.,
"Coal
Flash
Pyrolysis
Polymethylene Compounds in Low Temperature Flash Pyrolysis Tars,"
2.
Fuel,
63, 1119, 1984b.
Calkins, W.H.,
"Coal Flash Pyrolysis
Polymethlylene Moieties in Coal," Fuel,
3. An Analytical Method
63, 1125, 1984c.
Calkins, W.H.,
Housepian, B.K.,
Dyrkez, G.R., Bloomquist, C.A.A.,
Ruscic,
L.,
"Coal Flash Pyrolysis 4. Polymethylene Moities in
for
and
Coal
Macerals," Fuel, 63, 1226, 1984d.
233
Campbell, J.H., and Stephens, D.R. "Kinetic Studies
During Pyrolysis of Subbituminous Coal," Am. Chem.
Chem. Prepr., 21 (7), 94, 1976.
of Gas Evolution
Soc. , Div. Fuel
Ciuryla, V.T., Weimer, R.F., Bivans, D.A.,
and Motika, S.A.,
"Ambient
pressure thermogravimetric characterization of four different coals and
their chars," Fuel, 58, 748, 1979.
Cosway, R.G.,
1981.
S.M. Thesis, Dept. Chem. Eng.,
M.I.T.,
Cambridge, MA,
Cypres, R.,
and Bettens, B., Tetrahedron, 1253, 30, 1974.
Cypres, R.,
and Bettens, B., Tetrahedron, 353, 31, 1975a.
Cypres, R.,
and Bettens, B.,
Tetrahedron, 359, 31, 1975b.
Darivakas, D.G.,
"Effects of Reaction Conditions and Pre-treatment
Methods on Liquid Yields from Low Rank Coals," Ph.D. Thesis, Dept.
Chem. Eng., M.I.T., Cambridge, MA, in preparation, 1988.
Deno, N.C., Curry, K.W., Jones, D.A., Keegan, K.R., Rakitsky, W.G.,
Ritcher, C.A., and Minard, R.D., "Linear Alkane Chains in Coals," Fuel,
60, 210, 1981.
Fong, W.S., "Plasticity and Agglomeration in Coal Pyrolysis,"
Thesis, Dept. Chem. Eng., M.I.T., Cambridge, MA, 1986.
Sc.D.
Franklin, H.D.,
"Mineral Matter
Effects
in
Coal
Pyrolysis
and
Hydropyrolysis," Ph.D. Thesis, Dept. Chem. Eng., Cambridge, MA, 1980.
Franklin, H.D., Peters, W.A., Cariello, F., and Howard, J.B., "Effects
of Calcium Minerals on the Rapid Pyrolysis of a Bituminous Coal," I&EC
Proc. Des. & Dev., 20, 670, 1981.
Freihaut,
J.D. , and Seery, D.J. , "An Investigation of Yields and
Characteristics of Tars Released During the Thermal Decomposition of
Coal," Am. Chem. Soc. Div. Fuel Chem. Prepr. , 26 (2), 133. 1981.
Freihaut, J.D.,
Zabielski, M.F.,
and Seery, D.J.,
"A Parametric
Investigation of Tar Release in Coal Devolatilization,"
Nineteenth
Symposium (Int.) on Combustion, The Combustion Institute, Pittsburgh,
1159, 1982.
Froment, G.F., and Bischoff, K.B.,
Design, John Wiley, New York, 1979.
Gan H., Nandi, S.P., and Walker Jr.,
American Coals," Fuel, 51, 272, 1972.
Chemical
P.L.,
Reactor
Analysis
and
"Nature of the Porosity in
Gavalas, G.R. , and Wilks , K.A. , "Intraparticle Mass Transfer
Pyrolysis," AIChE J., 26, No.2, 201, 1980.
in
Coal
234
Gavalas, G.R., Cheong, P.H.K., and Jain, R., "Model of Coal Pyrolysis.
1.Qualitative Development," Ind. Eng. Chem. Fundam., 20, 113, 1981a.
Gavalas, G.R., Cheong, P.H.K., and Jain, R., "Model of Coal Pyrolysis.
2.Quantitative Formulation and Results," Ind. Eng. Chem. Fundam., 20,
122, 1981b.
Gavalas, G.R., Coal
Amsterdam, 1984.
Pyrolysis,
Elsevier
Scientific
Publishing
Co.,
Gerstein, B.C., Murphy, P.D., and Ryan, L.M., "Aromaticity in Coal," in
Coal Structure, R.A. Meyers, ed., Academic Press, New York, 1982.
Given, P.H., "The Organic Chemistry of Coal Macerals," Penn State Short
Course on Coal, The Pennsylvania State University, June 1976.
Given, P.H.,
"What are the Important Characteristics of Coal?," Penn
State Short Course on Coal, The Pennsylvania State University, Oct.
1977.
Griffin, T.P., "Intra-Particle Secondary Reactions in Coal Pyrolysis,"
Ph.D. Thesis, Dept. Chem. Eng., M.I.T., Cambridge, MA, in preparation
1988.
Hajaligol, M.R.,
Peters, W.A.,
and Howard, J.B., "Intra-Particle Heat
Transfer Effects in Coal Pyrolysis," Energy and Fuels, in press 1988.
Hanbaba,
P.,
Juntgen,
H.,
and
Peters,
W.,
"Nicht-isotherme
Reactionskinetik der Kohlenpyrolyse,Teil II. Erweiterung der Theorie
der Gasabspaltung und Experimentelle
Bestatigung an Steinkohlen,"
Brennstoff-Chemie, 49, 368-376, 1968.
Hirsch, P.B.,
"X-ray Scattering from Coals,"
A226, 143, 1954.
Proc.
Roy.
Soc.
London,
Howard, J.B. ,"Fundamentals of Coal Pyrolysis and Hydropyrolysis,"
Chemistry of Coal Utilization, 2nd Suppl. Vol., M.A. Elliott, ed.,
Wiley & Sons, New York, 1981.
in
J.
Howard, J.B., Ko, G.H., and Peters, W.A., Quarterly Technical Progress
Report submitted to DOE/METC, 1987a.
Howard,
J.B.,
Fong,
W.S.,
and
Peters,
W.A.,
"Kinetics
of
Devolatilization,"
in Fundamentals
of the
Physical-Chemistry of
Pulverized Coal Combustion, Lahaye, J. and Prado, G., eds., Martinus
Nijhoff Publishers, Dordrecht, 1987b.
Howard, J.B., Griffin, T.P., Ko, G.H., and Peters, W.A.,
"Fundamental
Mechanisms of Coal Gasification",
presented at the Eighth Annual
Gasification and Gas Stream Cleanup Systems Contractors Review Meeting,
1988.
Hsu, J.,
Ph.D. Thesis,
Dept.
Chem. Eng.,
M.I.T.,
Cambridge,
MA,
in
preparation 1988.
235
Huston, J.L. and Studier, M.H.,
"Aromaticities of Various Coals
Measured by Reaction with Fluorine and TGA-DTA studies of the Reaction
Products,"
Fuel,
60,
1149,
1981.
Ignasiak, B.S. and Gawlak, M.,
"Polymeric Structure
Ether Bonds in Constitution of High-Rank Vitrinite,"
1977.
of Coal. l.Role of
Fuel,
56, 216,
Jacab, E., McClennen, W.H., Hoesterey, B., Meuzelaar, H.L.C., and Hill
G.R., "Weathering Effects on Yield and Composition of Pyrolytic Coal
Tars," Am. Chem. Soc. Div. Fuel Chem. Prepr., 30 (3), 137, 1985.
Juntgen, H. and van Heek, K.H. , "Research in the Field of Pyrolysis at
Bergbau-Forschung during the Last Fifteenth Years," presented at the
Meeting on Coal Fundamentals, Stoke Orchard, England, 1977.
Kerr, J.A. and Parsonage, M.J., Evaluated Kinetic Data on Gas Phase
Hydrogen Transfer Reactions in Methyl Radicals, Butterworths, London,
1976.
Ko, G.H., Peters, W.A., Howard, J.B., "Correlation of Tar Yields
from Rapid Pyrolysis with Coal Type and Pressure,"
Fuel,
66,
1987.
1118,
Ko, G.H.,
Peters, W.A., and Howard, J.B., "Comparison of Tar Evolution
Rate Predictions in Coal Pyrolysis from Multiple Independent Parallel
Reaction Model and Functional Group Model Over a Wide Range of Heating
Rates," Energy and Fuels, in press 1988a.
Ko, G.H., Sanchez, D.M.,
Peters, W.A., and Howard, J.B., "Application
of First-Order, Single-Reaction Model for Coal Devolatilization Over a
Wide Range of Heating Rates," Am. Chem. Soc. Div. Fuel Chem. Prepr., 33
(2),
112,
1988b.
Ko, G.H.,
Sanchez, D.M., Peters, W.A.,
and Howard, J.B.,
"Correlations
for Effects of Coal Type and Pressure on Tar Yields from Rapid
Devolatilization,"
Twenty-Second Symposium (Int.)
on Combustion,
in
press 1988c.
Kobayashi, H., Howard, J.B., and Sarofim, A.F.,
"Coal Devolatilization
at High Temperatures,"
Sixteenth Symp.
(Int.)
on Combustion, The
Combustion Institute, Pittsburgh, 411, 1977.
Kwan, J.T. and Yen, T.F., "Aromaticity Determination of Coal,
and their
Derivatives by X-Ray Diffraction," Am. Chem. Soc.
Chem. Prepr., 21 (7),
Lewellen, P.C.,
Oil Shale
Div. Fuel
67, 1976.
"Product Decomposition Effects in Coal Pyrolysis," S.M.
Thesis, Dept. Chem. Eng., M.I.T., Cambridge, MA, 1975.
Loison, R. and Chauvin,
(Paris), 91, 269, 1964.
F.,
"Pyrolyse Rapide Du Charbon,"
Chem.Ind.,
236
Maiorella,
B.L. , "Behavior of Liquid Subbituminous Coal Tars Upon
Heating, as it Relates to In-Situ Gasification," S.B. Thesis, Dept.
Chem. Eng., M.I.T., Cambridge, MA, 1978.
Mason, E. A. and Malinauskas, A.P., "Gas Transport in Porous Media: The
Dusty Gas Model." Chem. Eng. Monograph 17, Elsevier, Amsterdam, 1983.
Mayo, F.R., Huntington, J.B.,
and Kirshen, N.A.,
"Chemistry of Coal
Liquefaction,"
in Organic Chemistry of coal, J.W. Larsen, ed., Am.
Chem. Symposium Series No. 71, Am. Chem. Soc., 126, 1978.
Neavel, R.C., Smith, S.E.,
Classification of Coals,"
Dusseldorf, 1-9, Sept.1981.
Hippo, E.J., and Miller, R.N.,
Proc. of Int. Conf. on Coal
"Optimum
Science,
Neoh, K.G. and Gannon, R.E., "Coal Volatile Yield and Element Partition
in Rapid Pyrolysis," Fuel, 63, 1347, 1984.
Niksa,
S.,"Time-Resolved Kinetics of Rapid Coal Devolatilization,"
Ph.D. Thesis, Dept. Chem. Eng., Princeton University, Princeton, NJ,
1981.
Niksa, S. and Kerstein, A.R., "The Distributed-Energy Chain Model for
Rapid Coal Devolatilization Kinetics.
Part I: Formulation, Part II:
Transient Weight Loss Correlations,"
Sandia National Lab.
Report,
SAND84-8870, 1985.
Oh, M.S.,
"Softening Coal Pyrolysis,"
M.I.T., Cambridge, MA, 1985.
Sc.D.
Thesis,
Dept.
Chem.
Eng.,
Oh, M.S.,
Peters, W.A.,
and Howard, J.B.,
"A Mathematical Model of
Softening
Coal
Pyrolysis,"
AIChE J.,
submitted for
review for
publication, 1988.
Peters, W.A.,
"Studies of the Pyrolysis Behavior of Condensed Phase
Fuels with Applications to Fuel Conversion Technology," in Proc., Fifth
Int. Symp. on Analytical Pyrolysis, K.J. Voorhees, ed., Butterworths,
Borough Green, Seven Oaks, England, 340, 1984.
Peters,
coals,"
W. and Bertling, H.,
Fuel, 44, 317, 1965.
"Kinetics of the Rapid Degasification of
Poutsma, M.L.,
"Free Radical Model for Coal Conversions: Effect of
Conversion Level and Concentration on Thermolysis of Bibenzyl," Fuel,
59, 535, 1980.
Reid, R.C., Prausnitz, J.M., and Sherwood, T.K., The Properties
Gases and Liquids, 3rd ed., McGraw-Hill, New York, 1977.
Reitzen, R.G.,
1978.
Remmp,
P.
S.M. Thesis, Dept. Chem. Eng.,
and Merrill,
E.W.,
Polymer Synthesis,
M.I.T.,
of
Cambridge, MA,
Huethig & Wepf Verlag,
Basel, 1986.
Retcofsky,
237
H.L.,
Hough,
Maguire,
M.R.,
M.M. , and Clarkson,
R.B.,
Adv.
Chem. Ser., 192, 37, 1981.
Retcofsky, H.G., "Magnetic Resonance Studies of Coal," in Coal Science,
Vol.1, M.L. Gorbaty, J.W. Larsen, I. Wender, eds., 43, 1982.
Russel,
W.B., Saville,
D.A.,
and Greene,
M.I.,
"A
Residence Time Hydropyrolysis of Single Coal Particles,"
Model for Short
AIChE J., 25,
Vol.1, 65, 1979.
Sanchez, D.M., S.B. Thesis, Dept. Chem. Eng., M.I.T.,
preparation 1988.
Cambridge, MA, in
Serio, M.A.,
"Secondary Reactions of Tar in
Coal Pyrolysis,"
Thesis, Dept. Chem. Eng., M.I.T., Cambridge, MA, 1984.
Ph.D.
Serio, M.A., Hamblen, D.G., Markham, J.R., and Solomon, P.R., "Kinetics
of Volatile
Product Evolution in Coal Pyrolysis:
Experiment and
Theory,"
Simons,
Energy & Fuels,
G.A.
and Finson,
1, 138,
M.L.,
1987.
"The
Structure
of Coal Chars.
1:
Pore
Branching," Combst. Sci. and Tech., 19, 227, 1979.
Singer, J.G., Combustion Fossil Power Systems, Combustion Engineering,
Inc., Windsor, CT, 1981.
Solomon,
P.R. and Colket, M.B., "Coal Devolatilization,"
Symposium (Int.)
on Combustion, The Combustion Institute,
Seventeenth
Pittsburgh,
113, 1978.
Solomon, P.R., Hamblen,
D.G.,
and Carangelo,
R.M.,
presented at AIChE meeting, New Orleans, Nov. 1981.
"Coal
Pyrolysis,"
Solomon, P.R., Hamblen, D.G., Carangelo, R.M., and Krause, J.L.,
"Coal Thermal Decomposition in an Entrained Flow Reactor: Experiments
and Theory,"
Nineteenth Combustion Symposium (Int.),
The Combustion
Institute, Pittsburgh, 1139, 1982.
Solomon, P.R. and Hamblen, D.G.,
Pyrolysis, in Chemistry
Conversion, R.H. Schlosberg, ed., Plenum Press, N.Y., 1985.
of
Coal
Sprouse, K.M. and Schuman, M.D.,
"Predicting Lignite Devolatilization
with
the
Multiple
Parallel
and
Two-Competing
Reaction
Models,"
Combustion and Flame, 43, 265-271, 1981.
Stein, S.E. and Golden, D.M.,
J. Org. Chem.,
42, 839, 1977.
Stein,
S.E.,
"A Fundamental
Chemical
Kinetics
Approach
to
Coal
Conversion,"
in
New Approaches in
Coal Chemistry,
Blaustein,
B.D.,
Bockrath, B.C., Friedman, S., eds., Am. Chem. Soc. Symp. Ser. 169,
1981.
Stein, S.E.,
"Free radicals in Coal Conversion," in Chemistry of Coal
Conversion, Schlosberg, R.H., ed., Plenum Publishing Co., 1985.
Stein, S.E., private communication, 1988.
"The Study of the Swelling Property of Bituminous
Sung, W.F.,
S.M. Thesis, Dept. Chem. Eng., M.I.T., Cambridge, 1978.
Coal,"
"Rapid Pyrolysis and Hydropyrolysis of Coal,"
Suuberg, E.M.,
Thesis, Dept. Chem. Eng., M.I.T., Cambridge, MA, 1977.
Sc.D.
E.M.
and Sezen, Y.,
Suuberg,
Processes in Coal Pyrolysis,"
Sydney, 913, 1985.
"Competitive Reaction
Proc.
1985 Int. Conf.
and Transport
Coal Science,
"Temperature dependence
Lee, D., and Larsen, J.W.,
Suuberg, E.M.,
crosslinking processes in pyrolysing coals," Fuel, 64, 1668, 1985.
of
Suuberg, E.M., "Mass Transfer Effects in Pyrolysis of Coals: A Review
of Experimental Evidence and Models," in Chemistry of Coal Conversion,
R.H. Schlosberg, ed., Plenum Press, New York, 1985.
Suuberg, E.M., Unger, P.E., and Larsen, P.E., "Relation between Tar and
Extractables Formation and Cross-Linking during Coal Pyrolysis," Energy
and Fuels, 1, 305, 1987.
"Coal Structure and Reactivity," a
Tingey,
G.L. and Morrey, J.A.,
Batelle Energy Program Report, Batelle Pacific Northwest Laboratories,
Richland, WA, 1973.
Unger, P.E. and Suuberg, E.M., "Modelling the Devolatilization Behavior
of a Softening Bituminous Coal," Eighteenth Symp. (Int.) Combustion,
Pittsburgh, The Combustion Institute, 1203, 1981.
Unger, P.E. and Suuberg, E.M., Am. Chem. Soc. Div. Fuel Chem. Prepr. 28
(4), 278, 1983.
Unger, P.E. and Suuberg, E.M., "Molecular weight distributions of tars
produced by flash pyrolysis of coals," Fuel, 63, 606, 1984.
Van Krevelen, , D.W., Coal, Elsevier Publishing Co., Amsterdam, 1961.
"Rates of Light Gas Production by
and Ngan, D.Y.,
Weimer,
R.F.,
Devolatilization of Coal and Lignite," Am. Chem. Soc. Div. Fuel Chem.
Prepr., 24 (3), 129, 1979.
Whitehurst, D.D., "A Primer on the Chemistry and Constitution of Coal,"
in Organic Chemistry of Coal, J.W. Larsen, ed., Am. Chem. Soc.
Symposium Series No. 71, Am. Chem. Soc., 1, 1978.
Whitehurst, D.D., Mitchell, T.O., and Farcasio, M., Coal LiquefactionThe Chemistry and Technology of Thermal Processes, Academic Press, New
York, 1980.
238
Winans, R.E., Hayatsu, R., McBeth, R.L., Scott, R.G., and Botto, R.E.,
"Aromatic Structures in Whole Coals and Coal Macerals," Am. Chem. Soc.,
Div. Fuel Chem. Prepr., 33 (1), 407, 1988.
"Pyrolysis-Induced
M.J.,
Wornat,
Polycyclic Aromatic Compounds from
in
Changes
a Bituminous
Composition of
the
Coal," Ph.D. Thesis,
Dept. Chem. Eng., M.I.T., Cambridge, MA, 1988.
Yarzab, R.F., Abdul-Baset, Z.,
Coals: New Data and Statistical
and Given, P.H., "Hydroxyl Contents of
et
Cosauo Chim.
Analysis," Geochim.
Acta., 43, 281, 1980.
Yoshida, T., Tokuhashi, K., Narita, H., Hasegawa, Y., and Maekawa, Y.,
Coals by Carbon -13 CP/MAS NMR
"Determination of Ether Oxygen in
Spectrometry and Acetylation,"
Fuel,
63,
282,
1984.
"Analysis of Product Yields from Rapid Pyrolysis of
M.W.,
Zacharias,
Bituminous Coal," S.M. Thesis, Dept. Chem. Eng., M.I.T., Cambridge, MA,
1979.
239
240
APPENDIX 1:
EXPERIMENTAL DATA
Note:
1. All data are in wt% dmmf basis.
241
2. Vacuum and 10 atm runs represent averaged values of 1-3 runs.
Coal type: Lower Kittanning
Atmospheric peak temperature runs:
T (C)
total wt.
tar
CO
CHG
4
loss
0.0
4.0
5.3
4.4
0.00
0.09
0.12
0.07
0.00
0.11
0.05
0.10
10.6
-
0.15
0.93
19.4
10.6
0.66
3.23
-
0.20
1.41
9.8
9.5
11.2
2.9
2.6
3.2
1.0
0.0
0.0
1.4
6.8
7.9
11.0
0.36
3.09
476
644
678
659
0.0
4.6
4.6
10.6
732
975
840
913
1009
854
619
619
649
218
267
359
514
698
688
834
-
20.3
17.3
21.4
1.0
0.9
4.3
0.1
0.0
0.7
4.2
10.5
11.3
17.0
Atmospheric holding temperature runs (5
T (C)
total wt. tar
CO
CHG4
loss
0.90
4.17
850
23.1
10.7
10.4
0.94
4.34
950
21.2
0.91
4.32
1050
20.5
10.2
9.1
850
24.0
10.0
920
22.6
10.5
1050
24.2
CO2
C2 H4
C 2H 6
0.00
0.16
0.55
0.15
0.36
0.55
0.35
0.45
0.00
0.02
0.02
0.02
0.16
0.67
0.31
0.48
0.00
0.10
0.02
0.04
0.30
0.53
0.44
0.56
C2 H4
C2 H 6
0.55
0.56
0.62
0.59
0.58
s hold):
CO 2
0.42
0.43
0.41
0.001 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
14.0
22.0
1000
10 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
7.2
21.1
1000
Coal type: Lower Wilcox
Atmospheric peak temperature runs:
T (C)
total wt. tar
CO
CH4
loss
0.12
0.04
3.2
0.0
463
0.46
0.05
19.5
611
0.09
0.64
19.7
650
0.21
1.28
26.0
668
1.13
3.69
837
41.9
1.14
3.31
43.5
824
1.90
8.45
12.5
53.2
1042
0.36
1.14
40.0
760
2.00
0.95
827
42.4
11.5
4.92
1.23
906
53.4
3.3
10.3
520
2.4
10.3
445
3.5
10.4
545
4.1
6.5
545
8.7
17.8
661
11.7
37.6
739
12.1
35.9
739
0.0
1.5
267
1.5
0.7
359
4.5
13.9
514
10.5
27.2
590
13.5
44.7
763
8.1
16.4
636
9.6
19.5
636
6.2
10.0
610
242
CO 2
C 2 H4
C 2H 6
0.54
2.73
3.28
5.05
7.59
7.21
8.23
6.56
6.74
0.01
0.04
0.08
0.19
1.04
1.05
1.60
0.45
0.76
0.69
0.01
0.01
0.03
0.09
0.31
0.31
C2 H4
C2 H6
1.47
1.76
1.42
0.28
0.26
0.26
Atmospheric holding temperature runs (5 s hold):
T (C)
total wt. tar
CO
CH4
CO2
loss
52.5
9.68
850
11.8 10.76
1.79
2.17
52.9
13.2 12.40
8.81
950
52.0
12.6 11.81
1.99
8.58
1050
850
50.1
13.0
57.1
11.9
1000
49.9
900
13.2
0.001 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
16.8
56.8
1000
10 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
9.9
48.5
1000
0.13
0.21
0.22
Coal type: Beulah Zap
Atmospheric peak temperature runs:
T (C)
total wt. tar
CO
CH4
loss
0.00
0.11
1.1
8.5
419
0.04
0.39
3.7
13.0
581
0.31
1.48
5.1
23.5
650
0.64
2.36
6.6
34.9
753
0.82
3.14
5.8
36.9
836
1.18
4.26
38.9
833
1.45
6.95
6.8
45.4
1024
0.90
3.49
6.6
41.5
906
0.15
0.71
4.7
26.9
674
1.36
6.69
6.4
47.3
1032
0.80
1.65
6.2
36.2
827
1.44
6.10
7.8
46.6
920
2.9
13.9
518
28.6
648
6.2
41.3
864
4.2
545
10.8
545
3.6
10.4
545
16.5
3.3
621
4.6
667
28.7
6.6
720
2.4
359
1.3
6.7
514
5.4
20.6
590
31.4
6.9
698
40.5
881
4.8
20.4
610
3.8
16.8
559
32.6
7.5
667
243
CO2
C 2 H4
C 2H 6
1.15
2.87
6.11
6.65
7.77
8.98
9.29
0.00
0.01
0.12
0.26
0.45
0.48
0.75
0.55
0.10
0.62
0.31
0.11
0.00
0.01
0.07
0.13
0.17
0.23
0.21
0.21
0.06
0.25
0.14
0.19
C2 H4
C2 H6
0.76
0.62
0.74
0.24
0.17
0.19
4.68
10.76
6.21
9.11
Atmospheric holding temperature runs (5 s hold):
T (C)
total wt. tar
CO
CH4
CO 2
loss
11.00
1.60
45.1
9.02
850
7.1
8.07
1.48
6.9
9.27
46.3
950
9.63
1.74
6.1 11.52
45.7
1050
13.0
50.1
850
11.9
57.1
1000
13.2
49.9
900
0.001 atm runs (total wt. loss and tar only):
T (C)
1000
total wt. loss
tar
51.2
9.1
10 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
1000
41.3
4.1
Coal type: Blue
Atmospheric peak temperature runs:
T (C)
total wt. tar
CO
CH4
loss
451
1.0
0.0
0.09
0.00
0.51
0.13
9.1
15.9
589
665
21.4
13.1
0.84
0.32
1.12
0.48
14.8
20.5
682
6.74
2.76
47.8
1024
4.17
2.11
49.1
850
1.22
2.13
20.5
40.6
787
0.19
0.01
0.9
1.9
526
22.2
46.4
746
20.0
43.0
757
19.7
47.2
900
5.2
11.0
559
6.6
10.0
567
19.5
36.2
693
17.6
34.3
684
20.9
45.6
864
6.9
619
649
7.8
36.7
22.0
739
0.6
0.0
359
2.6
1.1
514
20.3
13.1
590
698
14.7
32.9
610
29.4
662
32.2
244
CO2
C 2 H4
C 2H 6
0.58
1.75
2.09
2.73
4.40
2.64
2.95
1.22
0.00
0.07
0.16
0.26
2.02
1.57
0.89
0.00
0.00
0.07
0.13
0.22
0.46
0.55
0.45
0.00
C2 H4
C2 H6
1.20
1.68
1.47
0.52
0.47
0.49
Atmospheric holding temperature runs (5 s hold):
T (C)
total wt. tar
CO
CH4
CO 2
loss
850
6.16
2.40
3.43
23.3
48.1
6.87
3.57
48.4
21.4
2.79
950
2.62
19.9
3.61
48.8
7.31
1050
850
48.4
19.7
920
45.8
22.1
21.7
1050
44.9
0.001 atm runs (total wt. loss and tar only):
T (C)
1000
total wt. loss
tar
27.7
10 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
14.5
1000
Coal type: Smith Roland
Atmospheric peak temperature runs:
T (C)
total wt. tar
CO
CH4
loss
0.0 0
0.10
5.5
457
0.1 3
0.82
8.2
18.3
562
0.2 8
1.24
28.4
692
1.30
0.38
12.9
656
1.22
0.29
28.5
666
1.39
12.6
3.72
47.3
877
11.2
4.78
1.50
50.0
1000
4.83
1.60
48.1
827
3.21
1.16
44.7
802
0.30
1.23
35.7
13.0
687
0.01
2.5
0.07
11.1
502
1.88
5.45
12.3
49.5
1009
3.1
13.7
549
8.1
31.2
631
12.2
43.8
693
14.5
48.7
988
12.9
46.9
860
6.6
15.9
599
24.5
649
10.7
11.0
30.4
661
1.2
3.9
359
3.8
8.1
478
21.6
8.7
590
11.2
30.5
698
6.7
20.4
610
245
CO2
C2 H4
C2 H 6
0.90
3.17
3.93
5.02
4.05
6.18
0.00
0.07
0.16
0.18
0.17
1.09
1.05
1.06
0.87
0.20
0.00
1.11
0.00
0.05
0.10
0.15
0.11
0.37
0.27
0.34
0.35
0.15
0.00
0.46
C2 H4
C2 H6
1.28
1.30
1.46
0.35
0.30
0.33
5.05
6.48
4.95
0.60
7.35
Atmospheric holding temperature runs (5 s hold):
T (C)
total wt. tar
CO
CH4
CO2
loss
10.5 10.40
2.15
8.55
49.6
850
11.1
2.18
47.1
9.35
7.39
950
12.5 10.13
2.19
8.04
48.8
1050
12.8
50.9
850
920
11.7
51.2
49.8
11.7
1050
0.001 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
14.8
51.8
1000
10 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
11.1
46.6
1000
Coal type: Illinois #6
Atmospheric peak temperature runs:
T (C)
total wt. tar
CO
CH4
loss
0.44
0.53
17.9
32.6
701
0.40
0.74
18.1
28.7
677
1.50
1.71
24.6
39.7
849
2.54
3.16
24.8
53.3
936
2.25
2.52
54.5
913
0.7
1.29
25.0
40.1
802
2.74
2.60
26.7
48.4
987
3.13
3.50
27.9
48.9
982
0.84
0.53
23.0
36.1
761
0.00
0.00
1.9
502
7.5
8.4
567
16.7
23.6
695
12.7
645
8.0
586
4.7
549
7.9
14.6
10.5
619
19.5
33.5
761
7.5
636
13.3
16.8
610
19.4
33.9
641
18.8
25.5
682
28.2
40.9
829
246
CO2
C 2 H4
C 2H 6
1.11
0.52
1.29
1.80
1.36
1.40
1.72
1.79
1.29
0.00
0.17
0.14
0.63
0.91
1.89
0.52
1.01
0.91
0.32
0.00
0.23
0.20
0.53
0.60
0.63
0.50
0.63
0.60
0.38
0.00
Atmospheric holding temperature runs (5 s hold):
T (C)
total wt. tar
CO
CH4
C 2 H4
CO 2
loss
1.73
0.80
3.32
26.4
3.20
48.3
850
3.82
3.40
2.02
0.91
49.4
950
25.9
49.1
1050
24.5
50.7
850
24.7
920
51.2
53.3
1050
0.001 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
30.1
56.5
1000
10 atm runs (total wt. loss and tar only):
tar
total wt. loss
T (C)
15.0
47.4
1000
C 2H 6
0.71
0.69
-